average 123.3 Fresh: 25.6 Fresh: 18.7 -

other physical and chemical properties of the soil " in [11] " .

**Table 3.** Influence of mill by-products on cane and sugar yields

(Tc/ha)

Cane Sugar (Estimated)

**Table 5.** Nutrient status of crop residues and their availability to the succeeding crops

### **3.3. Renewable energy potential**

### *3.3.1. Ethanol productivity*

Sugarcane, given its energy balance advantage, is likely to be beneficial if promoted as bio-fuel feedstock as this is likely to increase sugarcane prices to the benefit of the small scale farmer.

Promoting sugarcane as a feedstock for ethanol is likely to improve rural livelihood and also minimize on forest encroachment since energy output per unit land area is very high for sugarcane.

In Brazil for example the production of sugar cane for ethanol only uses 1% of the available land and the recent increase in sugar cane production for bio-fuels is not large enough to explain the displacement of small farmers or soy production into deforested zones " in [13] ".

To minimize competition over land, it is advisable to grow sugarcane that has high yields with higher energy output compared to other biofuel crops. High yielding bio-fuels are preferable as they are less likely to compete over land "in [16]".

### *3.3.2. Bioelectricity generation*

Hydropower contributes about 90 per cent of electricity generated in Uganda with sugarcane based bagasse bioelectricity, fossil fuel and solar energy among other sources of power. Although the current generation of 800 MW "in [18] ". has boosted industrial growth, the capacity is still lagging behind the demand that is driven by the robust growth of the economy.

The low pressure boilers of 45 bar currently generate 22 MW of which 10 MW is connected to the grid. However, Kakira sugar estate has a target of generating 50 MW of electricity with the installation of higher pressure boilers of 68 bar in 2013. This target can be surpassed given the abundance of the bagasse (Table 6 and 7).


**average** 

**Particulars Plant Ratoon 1 Ratoon 2 Ratoon 3 Total/** 

Cane harvest area (ha) 4,890.4 4,748.6 3,869.9 1,824.1 15,333.0

\*This electric power generation is calculated based on using low pressure boilers of 45 bar

**Table 8.** Combined (Estate + Outgrowers cane production/productivity, Electrical power generation\*

The competition for resources between sugarcane and food crops is apparent with foreseen consequent increased food insecurity. Fifty percent of the arable land area good for food

A farm household that allocates all of its one hectare of land to sugarcane is expected to earn 359 \$ at high input level, 338 \$ at intermediate level and 261 \$ at low input level " in [4] ". The 391 \$ required to purchase maize meal is well above the net margins from one ha. This shows that proceeds from one hectare cannot sustain a household of 5. It is further revealed that maize produced from 0.63 ha can sustain a household nutritionally; however considering the annual household expenditure (760.8 \$; "in [17] "), about three hectares of land under sugarcane are required at low input level to support a household "

However, this study reveals that sugarcane sales accrued from ethanol under a scenario of a flourishing bio-fuel industry is associated with increased income that is likely to support households (Table 9). An ethanol gross sale per person per day is 1.6 dollars; an indication that the cultivation of sugarcane based biofuel is likely to contribute to alleviation of household poverty. A trickle-down effect on household income is expected from a foreseen expansion of bagasse-based electricity generation beyond the estate into the national

475,307.50 452,205.60 325,711.05 153,976.15 1,407,200.4

97.19 95.22 84.17 84.41 91.78

18.75 17.75 18.00 16.00 17.65

5.18 5.36 4.67 5.27 5.20

38.88 38.09 33.67 33.76 36.71

77.76 76.18 67.34 67.52 73.42

15.55 15.24 13.47 13.50 14.68

Total cane supply

Average cane yield

Average harvest age

Cane productivity

Bagasse production

Electric power generation (Mwh/ha)

at Kakira estate

Steam generation (tons

norms (mean for 2008 – 2012)

**3.4. Household income and food security** 

crop production is equally good for sugarcane.

(tons)

(tc/ha)

(months)

(tc/ha/m)

/ha

/ha)

in [4]".

electricity grid.

\*This electric power generation is calculated based on using low pressure boilers of 45 bar at Kakira estate



\*This electric power generation is calculated based on using low pressure boilers of 45 bar at Kakira estate

Tc = Tons of cane

**Table 7.** Mean outgrowers cane production/productivity, electrical generation\* norms (mean for 2008 – 2012)

Putting into consideration the productivity norms at Kakira estate and outgrowers (Table 8), with a potential of producing 908.9 m tons of sugarcane, Uganda has a potential of producing bio-electricity that surpasses the nation's demand by far. Much of this electrical power can be exported to the region, greatly expanding on Uganda's export base.


\*This electric power generation is calculated based on using low pressure boilers of 45 bar at Kakira estate

**Table 8.** Combined (Estate + Outgrowers cane production/productivity, Electrical power generation\* norms (mean for 2008 – 2012)

### **3.4. Household income and food security**

364 Biomass Now – Cultivation and Utilization

Electric power generation

(Mwh/ha)

Total cane supply

Average cane yield

Average harvest age

Cane productivity

Electric power generation (Mwh/ha)

Tc = Tons of cane

Steam generation (tons

(tons)

(tc/ha)

(months)

(tc/ha/m)

/ha)

2012)

**Particulars Plant Ratoon 1 Ratoon 2 Ratoon 3 Total/** 

Cane harvest area (ha) 1,461.1 1,541.7 1,535.2 392.8 4,930.8 Total cane supply (tons) 155,207.3 162,132.3 137,436.4 40,138.9 494,915.9 Average cane yield (tc/ha) 106.23 105.16 89.52 87.28 100.37 Average harvest age (months) 19.20 18.15 17.98 16.50 17.96 Cane productivity (tc/ha/m) 5.53 5.79 4.97 5.29 5.40 Bagasse production /ha 42.49 42.06 35.80 34.90 40.14 Steam generation (tons /ha) 84.98 84.12 71.60 69.80 80.28

\*This electric power generation is calculated based on using low pressure boilers of 45 bar at Kakira estate

**Table 6.** Mean estate cane production/productivity, electrical generation\* norms (2008 – 2012)

Particulars **Plant Ratoon 1 Ratoon 2 Ratoon 3 Total/** 

Cane harvest area (ha) 3,429.3 3,206.9 2,334.7 1,431.3 10,402.2

Bagasse production /ha 37.30 36.18 32.25 31.81 35.08

\*This electric power generation is calculated based on using low pressure boilers of 45 bar at Kakira estate

power can be exported to the region, greatly expanding on Uganda's export base.

**Table 7.** Mean outgrowers cane production/productivity, electrical generation\* norms (mean for 2008 –

Putting into consideration the productivity norms at Kakira estate and outgrowers (Table 8), with a potential of producing 908.9 m tons of sugarcane, Uganda has a potential of producing bio-electricity that surpasses the nation's demand by far. Much of this electrical

17.00 16.82 14.32 13.96 16.05

320,100.20 290,073.29 188,274.65 113,837.25 912,285.49

93.34 90.45 80.64 79.53 87.70

18.50 17.50 18.00 16.00 17.50

5.05 5.17 4.48 4.97 5.01

74.60 72.36 64.50 63.62 70.16

14.90 14.50 12.90. 12.70 14.00

**average** 

**average** 

The competition for resources between sugarcane and food crops is apparent with foreseen consequent increased food insecurity. Fifty percent of the arable land area good for food crop production is equally good for sugarcane.

A farm household that allocates all of its one hectare of land to sugarcane is expected to earn 359 \$ at high input level, 338 \$ at intermediate level and 261 \$ at low input level " in [4] ". The 391 \$ required to purchase maize meal is well above the net margins from one ha. This shows that proceeds from one hectare cannot sustain a household of 5. It is further revealed that maize produced from 0.63 ha can sustain a household nutritionally; however considering the annual household expenditure (760.8 \$; "in [17] "), about three hectares of land under sugarcane are required at low input level to support a household " in [4]".

However, this study reveals that sugarcane sales accrued from ethanol under a scenario of a flourishing bio-fuel industry is associated with increased income that is likely to support households (Table 9). An ethanol gross sale per person per day is 1.6 dollars; an indication that the cultivation of sugarcane based biofuel is likely to contribute to alleviation of household poverty. A trickle-down effect on household income is expected from a foreseen expansion of bagasse-based electricity generation beyond the estate into the national electricity grid.


practices reported in this study. Consequently this reduces the need to expand land acreage under cane while releasing land for use in food crop productivity. The high biomass returned to the ground sequesters carbon thereby offering the opportunity for sugarcane based farmers to earn extra income through the sale of carbon credits. Trickle down effects are expected to increase household income through the production and marketing of cane

These developments are expected to improve the farmers purchasing power, making

*Faculty of Natural Resources and Environment, Namasagali Campus, Busitema University, Kamuli,* 

*Research and Dev't Section - Agricultural Department Kakira Sugar Limited, Jinja, Uganda* 

*Katholieke Universiteit Leuven, Department of Earth and Environmental Sciences, Celestijnenlaan* 

Financial support for the various studies reported in this chapter was provided by UNEP GRID-Arendal, Kakira Sugar Limited and Belgian Technical Cooperation (BTC). The National Environment Management Authority and the National Agricultural Research Organization in Uganda gave the facilities and technical support that enabled the

[1] Bahiigwa B. A. Godfrey, 1999 Household Food Security In Uganda: An Empirical

[2] Uganda government, 2002b Uganda Food and Nutrition Policy, Ministry of Agriculture, Animal Industry and Fisheries, Ministry of Health, Kampala, Uganda

Analysis, Economic Policy Research Center, Kampala, Uganda

*National Environment Management Authority, Kampala, Uganda* 

households to be less dependent on the land and more food secure financially.

based biofuel and electricity.

**Author details** 

Moses Isabirye\*

*Uganda* 

D.V.N Raju

M. Kitutu

V. Yemeline

*Leuven, Belgium* 

**7. References** 

Corresponding Author

 \*

*UNEP/GRID-Arendal, Norway* 

J. Deckers and J. Poesen

**Acknowledgement** 

accomplishment of studies reported here.

**Table 9.** Sugarcane productivity, sales and potential land-use conflict

### **4. Further research**

The expansion of cane production is largely driven by market forces oblivious to the detrimental impact the industry is likely to have on food, livelihood security and the status of biodiversity. In addition to lack of appropriate policies to support the small-scale cane farmer, the policies are largely sectoral with no linkages with other relevant policies. Information is required to support the sustainable development of the cane industry with minimal negative impact on food and livelihood security and the status of biodiversity.

### **5. Relevant questions to explore among others include**

Can food crop productivity be improved in the context of a sugarcane-based farming system?

Can the understanding of the dimensions of food and livelihood security in sugarcanebased farming systems inform the synergistic development and review of relevant policies in the food, agriculture, health, energy, trade and environment sectors? What are the social impacts of the industry in light of the various agro-ecological zones of the country? What is the gender based livelihood strategies with special emphasis on labor exploitations- child labor etc?

What do people consider as possible options for improving food and livelihood security in a sugarcane-based farming system? Do these options differ between different actors (local women and men, NGOs and government)? How do families cope with food inadequacy, inaccessibility and malnutrition?

Can the study inform the carbon credit market initiative for farming systems in Uganda through the climate smart agriculture concept? Are the proposed assessment tools appropriate for Ugandan situations and the cane-based systems in particular?

### **6. Conclusion**

Driven by the need to meet the increasing local and regional sugar demand, and fossil fuel import substitution, cane expansion has potential negative impact on food security and biodiversity. However, this negative impact parallels the benefits related to cane cultivation. Cane biomass yield can be improved and sustained through the integrated use of various practices reported in this study. Consequently this reduces the need to expand land acreage under cane while releasing land for use in food crop productivity. The high biomass returned to the ground sequesters carbon thereby offering the opportunity for sugarcane based farmers to earn extra income through the sale of carbon credits. Trickle down effects are expected to increase household income through the production and marketing of cane based biofuel and electricity.

These developments are expected to improve the farmers purchasing power, making households to be less dependent on the land and more food secure financially.

### **Author details**

366 Biomass Now – Cultivation and Utilization

Cane

**4. Further research** 

system?

labor etc?

**6. Conclusion** 

inaccessibility and malnutrition?

Production / year Gross sales Conflict

ton litres /ha/year % 908.9 m 75.4 1869 22161 1.6 50.0 14.0 4.3

Sugarcane= USD 21/ton: projected population of 33 m in 2009 is used

**5. Relevant questions to explore among others include** 

**Table 9.** Sugarcane productivity, sales and potential land-use conflict

production Billion Farm Ethanol Capita Food Gazetted Forest

The expansion of cane production is largely driven by market forces oblivious to the detrimental impact the industry is likely to have on food, livelihood security and the status of biodiversity. In addition to lack of appropriate policies to support the small-scale cane farmer, the policies are largely sectoral with no linkages with other relevant policies. Information is required to support the sustainable development of the cane industry with minimal negative impact on food and livelihood security and the status of biodiversity.

Can food crop productivity be improved in the context of a sugarcane-based farming

Can the understanding of the dimensions of food and livelihood security in sugarcanebased farming systems inform the synergistic development and review of relevant policies in the food, agriculture, health, energy, trade and environment sectors? What are the social impacts of the industry in light of the various agro-ecological zones of the country? What is the gender based livelihood strategies with special emphasis on labor exploitations- child

What do people consider as possible options for improving food and livelihood security in a sugarcane-based farming system? Do these options differ between different actors (local women and men, NGOs and government)? How do families cope with food inadequacy,

Can the study inform the carbon credit market initiative for farming systems in Uganda through the climate smart agriculture concept? Are the proposed assessment tools

Driven by the need to meet the increasing local and regional sugar demand, and fossil fuel import substitution, cane expansion has potential negative impact on food security and biodiversity. However, this negative impact parallels the benefits related to cane cultivation. Cane biomass yield can be improved and sustained through the integrated use of various

appropriate for Ugandan situations and the cane-based systems in particular?

Moses Isabirye\* *Faculty of Natural Resources and Environment, Namasagali Campus, Busitema University, Kamuli, Uganda* 

D.V.N Raju *Research and Dev't Section - Agricultural Department Kakira Sugar Limited, Jinja, Uganda* 

M. Kitutu *National Environment Management Authority, Kampala, Uganda* 

V. Yemeline *UNEP/GRID-Arendal, Norway* 

J. Deckers and J. Poesen *Katholieke Universiteit Leuven, Department of Earth and Environmental Sciences, Celestijnenlaan Leuven, Belgium* 

### **Acknowledgement**

Financial support for the various studies reported in this chapter was provided by UNEP GRID-Arendal, Kakira Sugar Limited and Belgian Technical Cooperation (BTC). The National Environment Management Authority and the National Agricultural Research Organization in Uganda gave the facilities and technical support that enabled the accomplishment of studies reported here.

### **7. References**


<sup>\*</sup> Corresponding Author

[3] USCTA (2001) The Uganda Sugarcane Technologist's Association, Fourth Annual Report, 2001, Kakira, Uganda.

**Chapter 16** 

© 2013 Ratanakhanokchai et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2013 Ratanakhanokchai et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,

distribution, and reproduction in any medium, provided the original work is properly cited.

*Paenibacillus curdlanolyticus*

Khanok Ratanakhanokchai, Rattiya Waeonukul, Patthra Pason, Chakrit Tachaapaikoon, Khin Lay Kyu,

Kazuo Sakka, Akihiko Kosugi and Yutaka Mori

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/51820

**1. Introduction** 

**Strain B-6 Multienzyme Complex:** 

**A Novel System for Biomass Utilization** 

To develop a bio-based economy for sustainable economic growth, it is necessary to produce chemicals and fuels from renewable resources, such as plant biomass. Plant biomass contains a complex mixture of polysaccharides, mainly cellulose and hemicellulose (mainly xylan), and other polysaccharides (Aspinall, 1980). The hemicelluloses, as well as the aromatic polymer lignin, interact with the cellulose fibrils, creating a rigid structure strengthening the plant cell wall. Therefore, complete and rapid hydrolysis of these polysaccharides requires not only cellulolytic enzymes but also the cooperation of xylanolytic enzymes (Thomson, 1993). Many microorganisms that produce enzymes capable of degrading cellulose and hemicellulose have been reported and characterized. Two enzyme systems are known for their degradation of lignocellulose by microorganisms. In many aerobic fungi and bacteria, endoglucanase, exoglucanase, and ancillary enzymes are secreted individually and can act synergistically on lignocellulose. The most thoroughly studied enzymes are the glycosyl hydrolases of *Trichoderma reesei* (Dashtban et al., 2009). On the other hand, several anaerobic cellulolytic microorganisms such as *Clostridium thermocellum* (Lamed & Bayer, 1988), *C. cellulovorans* (Doi et al., 2003), *C. josui* (Kakiuchi et al., 1998) and *C. cellulolyticum* (Gal et al., 1997) are known to produce a cell-associated, large extracellular polysaccharolytic multicomponent complex called the cellulosome, in which several cellulolytic and xylanolytic enzymes are tightly bound to a scaffolding protein (core protein). Thus, the cellulosome provides for a large variety of enzymes and attractive enzymatic properties for the degradation of recalcitrant plant biomass. So far, anaerobic microorganisms have been identified as producing the multienzyme complex, cellulosome


http://www.Bio-fuel-news.com/magazine\_store.php?issue\_id=34


### *Paenibacillus curdlanolyticus* **Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization**

Khanok Ratanakhanokchai, Rattiya Waeonukul, Patthra Pason, Chakrit Tachaapaikoon, Khin Lay Kyu, Kazuo Sakka, Akihiko Kosugi and Yutaka Mori

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/51820

### **1. Introduction**

368 Biomass Now – Cultivation and Utilization

Uganda

experience

Report, 2001, Kakira, Uganda.

International volume 3, issue 10,2009,

Research Station, Uganda

*Annual Technical Conference.*

Vision, Publish Date: Oct 09, 2012

[3] USCTA (2001) The Uganda Sugarcane Technologist's Association, Fourth Annual

[4] Isabirye (2005) Land Evaluation around Lake Victoria: Environmental Implications for Land use Change, PhD Dissertation, Katholieke Universiteit, Leuven, Belgium [5] MEMD (2007) The Renewable Energy Policy for Uganda. MEMD, Kampala, Uganda. [6] MEMD (2010) The Uganda Energy Balance Report., MEMD, Kampala, Uganda.

[7] Bio-fuel-news (2009) Uganda to produce cellulosic ethanol in a year. Bio-fuel

[8] Meteorology Department (1961) Climate data, Meteorological Department, Entebbe

[9] Chenery (1960) Introduction to the soils of the Uganda Protectorate, Memoirs of the Research Division, Series 1- Soils, Number 1, Department of Agriculture, Kawanda

[10] FAO (1983) Guidelines: Land evaluation for rainfed agriculture. FAO Soils Bulletin 52,

[11] Antwerpen R. V. (2008) Organic wastes as an alternative source of nutrients. The link

[13] Xavier M.R. (2007) The Brazilian sugarcane ethanol experience. Issue Analysis, no. 3,

[14] Carr Carr, A.P, Carr, D.R, Carr, I.E, Wood, A.W and Poggio, M. (2008) Implementing sustainable farming practices in the Herbert: The Oakleigh farming company

[15] Raju, D.V.N and Raju, K.G.K (2005) Sustainable sugarcane production through integrated nutrient management. *In: Uganda Sugarcane Technologists' Association 17TH*

[16] Pesket Leo, Rachel Slater, Chris Steven, and Annie Dufey (2007) Biofuels, Agriculture and Poverty Reduction. Natural Resource Perspectives 107, Overseas Development

[17] UBOS (2001) Uganda National household survey 1999/2000; Report on the socioeconomic. Uganda Bureau of Statistics, Entebbe, Uganda. www.ubos.org [18] Ibrahim Kasita (2012) Strategic plan to increase power supply pays dividends, New

[12] FAOStat (2012) http://faostat.fao.org/site/339/default.aspx Sunday, May 06, 2012

http://www.Bio-fuel-news.com/magazine\_store.php?issue\_id=34

Food and Agricultural Organization of the United Nations, Rome

published by SASRI. Vol. 17 No. 2: May 2008. 8-9.

Washington, USA, Competitive Enterprise Institute. 11 p.

Institute, 111 Westminster Bridge Road, London SE1 7JD

To develop a bio-based economy for sustainable economic growth, it is necessary to produce chemicals and fuels from renewable resources, such as plant biomass. Plant biomass contains a complex mixture of polysaccharides, mainly cellulose and hemicellulose (mainly xylan), and other polysaccharides (Aspinall, 1980). The hemicelluloses, as well as the aromatic polymer lignin, interact with the cellulose fibrils, creating a rigid structure strengthening the plant cell wall. Therefore, complete and rapid hydrolysis of these polysaccharides requires not only cellulolytic enzymes but also the cooperation of xylanolytic enzymes (Thomson, 1993). Many microorganisms that produce enzymes capable of degrading cellulose and hemicellulose have been reported and characterized. Two enzyme systems are known for their degradation of lignocellulose by microorganisms. In many aerobic fungi and bacteria, endoglucanase, exoglucanase, and ancillary enzymes are secreted individually and can act synergistically on lignocellulose. The most thoroughly studied enzymes are the glycosyl hydrolases of *Trichoderma reesei* (Dashtban et al., 2009). On the other hand, several anaerobic cellulolytic microorganisms such as *Clostridium thermocellum* (Lamed & Bayer, 1988), *C. cellulovorans* (Doi et al., 2003), *C. josui* (Kakiuchi et al., 1998) and *C. cellulolyticum* (Gal et al., 1997) are known to produce a cell-associated, large extracellular polysaccharolytic multicomponent complex called the cellulosome, in which several cellulolytic and xylanolytic enzymes are tightly bound to a scaffolding protein (core protein). Thus, the cellulosome provides for a large variety of enzymes and attractive enzymatic properties for the degradation of recalcitrant plant biomass. So far, anaerobic microorganisms have been identified as producing the multienzyme complex, cellulosome

© 2013 Ratanakhanokchai et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Ratanakhanokchai et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

(Doi & Kosugi, 2004; Demain et al., 2005). However, when compared with aerobic enzymes, production of those enzymes by anaerobic culture presents a high cost because of the high price of medium, slow rate of growth and low yield of enzyme, while only a little information has been reported on cellulosome-like multienzyme complex produced by aerobic bacteria (Kim & Kim, 1993; Jiang et al,, 2004; van Dyk et al., 2009). Therefore, the multienzyme complexes, cellulosomes, produced by aerobic bacteria show great potential for improving plant biomass degradation. A facultatively anaerobic bacterium, *P. curdlanolyticus* strain B-6, is unique in that it produces extracellular xylanolytic-cellulolytic multienzyme complex under aerobic conditions (Pason et al., 2006a, 2006b; Waeonukul et al., 2009b). In the following years, the characteristics, function, genetics and mechanism of the xylanolytic-cellulolytic enzymes system of this bacterium has been the subject of considerable research. In light of new findings in this field, this review will describe the state of knowledge about the multienzyme complex of strain B-6 and its potential biotechnological exploitations.

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 371

and Van der Waals forces. Hydrogen bonding within a cellulose microfibril determines 'straightness' of the chain but inter-chain hydrogen bonds might introduce order (crystalline) or disorder (amorphous) into the structure of the cellulose (Klemm et al., 2005). In the latter conformation, cellulose is more susceptible to enzymatic degradation (Pérez et al., 2002). In nature, cellulose appears to be associated with other plant compounds and this

**Hemicelluloses:** Hemicelluloses are the second most abundant polymers and differ from cellulose in that they are not chemically homogeneous. Hemicelluloses are branched, heterogenous polymers of pentoses (xylose, arabinose), hexoses (mannose, glucose, galactose) and acetylated sugars. They have lower molecular weight compared to cellulose and branches with short lateral chains that are easily hydrolysed (Saha, 2003; Scheller & Ulvskov, 2010). Hemicelluloses differ in composition. Hemicelluloses in agricultural biomass like straws and grasses are composed mainly of xylan, while softwood hemicelluloses contain mainly glucomannan. In many plants, xylans are heteropolysaccharides with backbone chains of 1,4-linked β-D-xylopyranose units. In addition to xylose, xylan may contain arabinose, glucuronic acid, or its 4-*O*-methyl ether, acetic acid, ferulic and *p*-coumaric acids. Hemicelluloses are bound via hydrogen bonds to the cellulose microfibrils in the plant cell wall, crosslinking them into a robust network. Hemicelluloses are also covalently attached to lignin, forming together with cellulose to

**Lignin:** Lignin is the third most abundant polymer in nature. It is present in plant cell walls and confers a rigid, impermeable, resistance to microbial attack and oxidative stress. Lignin is a complex network formed by polymerization of phenyl propane units and constitutes the most abundant non-polysaccharide fraction in lignocelluloses (Pérez et al., 2002; Sánchez, 2009). The three monomers in lignin are *p*-coumaryl alcohol, coniferyl alcohol and sinapyl alcohol; they are joined through alkyl–aryl, alkyl–alkyl and aryl–aryl ether bonds. Lignin embeds the cellulose thereby offering protection against microbial and enzymatic degradation. Furthermore, lignin is able to form covalent bonds to some hemicelluloses, e.g. benzyl ester bonds with the carboxyl group of 4-*O*-methyl-Dglucuronic acid in xylan. More stable ether bonds, also known as lignin carbohydrate complexes, can be formed between lignin and arabinose, or between galactose side groups

Several biological methods for lignocellulose recycling based on the enzymology of cellulose, hemicelluloses, and lignin degradation have been developed. To date, processes that use lignocellulolytic enzymes or microorganisms could lead to promising, environmentally friendly technologies. The relationship between cellulose and hemicellulose in the cell walls of higher plants is much more intimate than was previously thought. It is possible that molecules at the cellulose-hemicellulose boundaries, and those

within the crystalline cellulose, require different enzymes for efficient hydrolysis.

association may affect its biodegradation.

form a highly complex structure.

in xylans and mannans.

**1.2. Biodegradation of lignocellulosic biomass** 

### **1.1. Composition of lignocellulosic biomass**

Lignocellulosic biomass is composed mainly of plant cell walls, with the structural carbohydrates, cellulose and hemicelluloses and heterogeneous phenolic polymer lignin as its primary components (Fig. 1). However, their proportions vary substantially, depending on the type, the species, and even the source of the biomass (Aspinall et al., 1980; Pérez et al., 2002; Pauly et al., 2008).

**Figure 1.** Structure of lignocellulosic plant biomass. (This figure is adapted from Tomme et al., 1995).

**Cellulose:** Cellulose, the main constituent of the plant cell wall, is a polysaccharide composed of linear glucan chains linked together by β-1,4-glycosidic bonds with cellobiose residues as the repeating unit at different degrees of polymerization, depending on resources. The cellulose chains are grouped together to form microfibrils, which are bundled together to form cellulose fibers. The cellulose microfibrils are mostly independent but the ultrastructure of cellulose is largely due to the presence of covalent bonds, hydrogen bonds and Van der Waals forces. Hydrogen bonding within a cellulose microfibril determines 'straightness' of the chain but inter-chain hydrogen bonds might introduce order (crystalline) or disorder (amorphous) into the structure of the cellulose (Klemm et al., 2005). In the latter conformation, cellulose is more susceptible to enzymatic degradation (Pérez et al., 2002). In nature, cellulose appears to be associated with other plant compounds and this association may affect its biodegradation.

**Hemicelluloses:** Hemicelluloses are the second most abundant polymers and differ from cellulose in that they are not chemically homogeneous. Hemicelluloses are branched, heterogenous polymers of pentoses (xylose, arabinose), hexoses (mannose, glucose, galactose) and acetylated sugars. They have lower molecular weight compared to cellulose and branches with short lateral chains that are easily hydrolysed (Saha, 2003; Scheller & Ulvskov, 2010). Hemicelluloses differ in composition. Hemicelluloses in agricultural biomass like straws and grasses are composed mainly of xylan, while softwood hemicelluloses contain mainly glucomannan. In many plants, xylans are heteropolysaccharides with backbone chains of 1,4-linked β-D-xylopyranose units. In addition to xylose, xylan may contain arabinose, glucuronic acid, or its 4-*O*-methyl ether, acetic acid, ferulic and *p*-coumaric acids. Hemicelluloses are bound via hydrogen bonds to the cellulose microfibrils in the plant cell wall, crosslinking them into a robust network. Hemicelluloses are also covalently attached to lignin, forming together with cellulose to form a highly complex structure.

**Lignin:** Lignin is the third most abundant polymer in nature. It is present in plant cell walls and confers a rigid, impermeable, resistance to microbial attack and oxidative stress. Lignin is a complex network formed by polymerization of phenyl propane units and constitutes the most abundant non-polysaccharide fraction in lignocelluloses (Pérez et al., 2002; Sánchez, 2009). The three monomers in lignin are *p*-coumaryl alcohol, coniferyl alcohol and sinapyl alcohol; they are joined through alkyl–aryl, alkyl–alkyl and aryl–aryl ether bonds. Lignin embeds the cellulose thereby offering protection against microbial and enzymatic degradation. Furthermore, lignin is able to form covalent bonds to some hemicelluloses, e.g. benzyl ester bonds with the carboxyl group of 4-*O*-methyl-Dglucuronic acid in xylan. More stable ether bonds, also known as lignin carbohydrate complexes, can be formed between lignin and arabinose, or between galactose side groups in xylans and mannans.

### **1.2. Biodegradation of lignocellulosic biomass**

370 Biomass Now – Cultivation and Utilization

biotechnological exploitations.

al., 2002; Pauly et al., 2008).

**1.1. Composition of lignocellulosic biomass** 

(Doi & Kosugi, 2004; Demain et al., 2005). However, when compared with aerobic enzymes, production of those enzymes by anaerobic culture presents a high cost because of the high price of medium, slow rate of growth and low yield of enzyme, while only a little information has been reported on cellulosome-like multienzyme complex produced by aerobic bacteria (Kim & Kim, 1993; Jiang et al,, 2004; van Dyk et al., 2009). Therefore, the multienzyme complexes, cellulosomes, produced by aerobic bacteria show great potential for improving plant biomass degradation. A facultatively anaerobic bacterium, *P. curdlanolyticus* strain B-6, is unique in that it produces extracellular xylanolytic-cellulolytic multienzyme complex under aerobic conditions (Pason et al., 2006a, 2006b; Waeonukul et al., 2009b). In the following years, the characteristics, function, genetics and mechanism of the xylanolytic-cellulolytic enzymes system of this bacterium has been the subject of considerable research. In light of new findings in this field, this review will describe the state of knowledge about the multienzyme complex of strain B-6 and its potential

Lignocellulosic biomass is composed mainly of plant cell walls, with the structural carbohydrates, cellulose and hemicelluloses and heterogeneous phenolic polymer lignin as its primary components (Fig. 1). However, their proportions vary substantially, depending on the type, the species, and even the source of the biomass (Aspinall et al., 1980; Pérez et

**Figure 1.** Structure of lignocellulosic plant biomass. (This figure is adapted from Tomme et al., 1995).

**Cellulose:** Cellulose, the main constituent of the plant cell wall, is a polysaccharide composed of linear glucan chains linked together by β-1,4-glycosidic bonds with cellobiose residues as the repeating unit at different degrees of polymerization, depending on resources. The cellulose chains are grouped together to form microfibrils, which are bundled together to form cellulose fibers. The cellulose microfibrils are mostly independent but the ultrastructure of cellulose is largely due to the presence of covalent bonds, hydrogen bonds Several biological methods for lignocellulose recycling based on the enzymology of cellulose, hemicelluloses, and lignin degradation have been developed. To date, processes that use lignocellulolytic enzymes or microorganisms could lead to promising, environmentally friendly technologies. The relationship between cellulose and hemicellulose in the cell walls of higher plants is much more intimate than was previously thought. It is possible that molecules at the cellulose-hemicellulose boundaries, and those within the crystalline cellulose, require different enzymes for efficient hydrolysis.

**Cellulase:** Cellulases responsible for the hydrolysis of cellulose are composed of a complex mixture of enzymes with different specificities to hydrolyze the β-1,4-glycosidic linkages (Fig. 2A). Cellulases can be divided into three major enzyme activity classes (Goyal et al., 1991; Rabinovich et al., 2002). These are endoglucanases or endo-1-4-β-glucanase (EC 3.2.1.4), exoglucanase or cellobiohydrolase (EC 3.2.1.91), and β-glucosidase (EC 3.2.1.21). Endoglucanases, are thought to initiate attack randomly at multiple internal sites in the amorphous regions of the cellulose fiber, which opens-up sites for subsequent attack by the cellobiohydrolases. Cellobiohydrolases remove cellobiose from the ends of both sides of the glucan chain. Moreover, cellobiohydrolase can hydrolyze highly crystalline cellulose. βglucosidase hydrolyzes cellobiose and in some cases short chain cellooligosaccharides to glucose.

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 373

**Hemicellulase:** Xylan is the main carbohydrate found in hemicelluloses. Its complete degradation requires the cooperative action of a variety of hydrolytic enzymes (Fig. 2B). Xylanases are frequently classified according to their action on distinct substrates: endo-1,4 β-xylanase (endoxylanase) (EC 3.2.1.8) generates xylooligosaccharides from the cleavage of xylan while 1,4-β-xylosidase (EC 3.2.1.37) produces xylose from xylobiose and short chain xylooligosaccharides. In addition, xylan degradation needs accessory enzymes, such as α-Larabinofuranosidase (EC 3.2.1.55), α-4-*O*-methyl-D-glucuronidase (EC 3.2.1.39), acetyl xylan esterase (EC 3.1.1.72), ferulic acid esterase (EC 3.1.1.73), and *p*-coumaric acid esterase (EC 3.1.1.-), acting synergistically, to efficiently hydrolyze wood xylans. In the case of acetyl-4-*O*methylglucuronoxylan, which is one of the most common hemicelluloses, four different enzymes are required for degradation: endo-1,4-β-xylanase, acetyl esterase (EC 3.1.1.6), αglucuronidase, and β-xylosidase. The degradation of *O*-acetyl galactoglucomannan starts with the rupture of the polymer by endomannanase (EC 3.2.1.78). Acetylglucomannan esterase (EC 3.1.1.-) removes acetyl groups, and α-galactosidase (EC 3.2.1.22) eliminates galactose residues. Finally, β-mannosidase (EC 3.2.1.85) and β-glucosidase break down the endomannanase-generated oligomeric β-1,4 bonds (Thomson, 1993; Li et al., 2000; Pérez et

The enzyme systems for the lignocellulose degradation by microorganisms can be generally regarded as non-complexed or complexed enzymes (Lynd et al., 2002). In the case of aerobic fungi and bacteria, the cellulase enzymes are free and mostly secreted. In such organisms, by the very nature of the growth of the organisms, they are able to reach and penetrate the cellulosic substrate and, hence, the secreted cellulases are capable of hydrolyzing the substrate. The enzymes in these cases are not organized into high molecular weight complexes and are called non-complexed (Fig. 3A). The polysaccharide hydrolases of the aerobic fungi are largely described based on the examples from *Trichoderma, Penicillum, Fusarium, Humicola, Phanerochaete,* etc., where a large number of the cellulases are encountered (Dashtban et al., 2009; Sánchez, 2009). In contrast, various cellulases and hemicellulases from several anaerobic cellulolytic microorganisms, are tightly bound to a scaffolding protein, as core protein and organized to form structures on the cell surfaces; these systems are called complexed enzymes or cellulosomes (Fig. 3B). The cellulosome is thought to allow concerted enzyme activities in close proximity to the bacterial cell, enabling optimum synergism between the enzymes presented on the cellulosome. Concomitantly, the cellulosome also minimizes the distance over which hydrolysis products must diffuse, allowing efficient uptake of these oligosaccharides by the host cells (Bayer et al., 1994;

Biotechnological applications in terms of hydrolysis efficiency for complexed enzyme systems might have an advantage over non-complexed enzyme systems. The high efficiency of the cellulosome has been attributed to (i) the correct ratio between catalytic domains that optimize synergism between them, (ii) appropriate spacing between the individual

al., 2002).

**1.3. Multienzyme complex cellulosome** 

Schwarz, 2001; Lynd et al., 2002).

**Figure 2.** Enzyme systems involved in the degradation of cellulose (A) and xylan (B). (This figure is adapted from Aro et al., 2005).

**Hemicellulase:** Xylan is the main carbohydrate found in hemicelluloses. Its complete degradation requires the cooperative action of a variety of hydrolytic enzymes (Fig. 2B). Xylanases are frequently classified according to their action on distinct substrates: endo-1,4 β-xylanase (endoxylanase) (EC 3.2.1.8) generates xylooligosaccharides from the cleavage of xylan while 1,4-β-xylosidase (EC 3.2.1.37) produces xylose from xylobiose and short chain xylooligosaccharides. In addition, xylan degradation needs accessory enzymes, such as α-Larabinofuranosidase (EC 3.2.1.55), α-4-*O*-methyl-D-glucuronidase (EC 3.2.1.39), acetyl xylan esterase (EC 3.1.1.72), ferulic acid esterase (EC 3.1.1.73), and *p*-coumaric acid esterase (EC 3.1.1.-), acting synergistically, to efficiently hydrolyze wood xylans. In the case of acetyl-4-*O*methylglucuronoxylan, which is one of the most common hemicelluloses, four different enzymes are required for degradation: endo-1,4-β-xylanase, acetyl esterase (EC 3.1.1.6), αglucuronidase, and β-xylosidase. The degradation of *O*-acetyl galactoglucomannan starts with the rupture of the polymer by endomannanase (EC 3.2.1.78). Acetylglucomannan esterase (EC 3.1.1.-) removes acetyl groups, and α-galactosidase (EC 3.2.1.22) eliminates galactose residues. Finally, β-mannosidase (EC 3.2.1.85) and β-glucosidase break down the endomannanase-generated oligomeric β-1,4 bonds (Thomson, 1993; Li et al., 2000; Pérez et al., 2002).

### **1.3. Multienzyme complex cellulosome**

372 Biomass Now – Cultivation and Utilization

adapted from Aro et al., 2005).

glucose.

**Cellulase:** Cellulases responsible for the hydrolysis of cellulose are composed of a complex mixture of enzymes with different specificities to hydrolyze the β-1,4-glycosidic linkages (Fig. 2A). Cellulases can be divided into three major enzyme activity classes (Goyal et al., 1991; Rabinovich et al., 2002). These are endoglucanases or endo-1-4-β-glucanase (EC 3.2.1.4), exoglucanase or cellobiohydrolase (EC 3.2.1.91), and β-glucosidase (EC 3.2.1.21). Endoglucanases, are thought to initiate attack randomly at multiple internal sites in the amorphous regions of the cellulose fiber, which opens-up sites for subsequent attack by the cellobiohydrolases. Cellobiohydrolases remove cellobiose from the ends of both sides of the glucan chain. Moreover, cellobiohydrolase can hydrolyze highly crystalline cellulose. βglucosidase hydrolyzes cellobiose and in some cases short chain cellooligosaccharides to

**Figure 2.** Enzyme systems involved in the degradation of cellulose (A) and xylan (B). (This figure is

The enzyme systems for the lignocellulose degradation by microorganisms can be generally regarded as non-complexed or complexed enzymes (Lynd et al., 2002). In the case of aerobic fungi and bacteria, the cellulase enzymes are free and mostly secreted. In such organisms, by the very nature of the growth of the organisms, they are able to reach and penetrate the cellulosic substrate and, hence, the secreted cellulases are capable of hydrolyzing the substrate. The enzymes in these cases are not organized into high molecular weight complexes and are called non-complexed (Fig. 3A). The polysaccharide hydrolases of the aerobic fungi are largely described based on the examples from *Trichoderma, Penicillum, Fusarium, Humicola, Phanerochaete,* etc., where a large number of the cellulases are encountered (Dashtban et al., 2009; Sánchez, 2009). In contrast, various cellulases and hemicellulases from several anaerobic cellulolytic microorganisms, are tightly bound to a scaffolding protein, as core protein and organized to form structures on the cell surfaces; these systems are called complexed enzymes or cellulosomes (Fig. 3B). The cellulosome is thought to allow concerted enzyme activities in close proximity to the bacterial cell, enabling optimum synergism between the enzymes presented on the cellulosome. Concomitantly, the cellulosome also minimizes the distance over which hydrolysis products must diffuse, allowing efficient uptake of these oligosaccharides by the host cells (Bayer et al., 1994; Schwarz, 2001; Lynd et al., 2002).

Biotechnological applications in terms of hydrolysis efficiency for complexed enzyme systems might have an advantage over non-complexed enzyme systems. The high efficiency of the cellulosome has been attributed to (i) the correct ratio between catalytic domains that optimize synergism between them, (ii) appropriate spacing between the individual components to further favor synergism, (iii) the presence of different enzymatic activities (cellulolytic or hemicellulolytic enzymes) in the cellulosome that can remove "physical hindrances" of other polysaccharides in heterogeneous plant cell materials (Lynd et al., 2002), and (iv) the presence of carbohydrate-binding modules (CBMs) that can increase the rate of hydrolysis by bringing the cellulosome into intimate and prolonged association with its recalcitrant substrate (Shoseyov et al., 2006). Thus, the complexed enzyme system, cellulosome, may provide great potential for the degradation of plant biomass.

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 375

**Figure 4.** Simplified schematic of general cellulosome components and connection with cell surface based on knowledge of *Clostridium* cellulosome. (This figure is adapted from Bayer et al., 1994).

lort et al., 2012

Ponpium et al., 2000

1990

al.,2002

1997

al., 1994

Lamed et al., 1983

Blair and Anderson, 1999b

2000

2010

*Bacteria Bacteria*

*alkalithermophilum* Soil Watthanalarm-

*fibrisolvens* Rumen Berger et al.,

*acetobutylicum* Soil Sabathé et

*cellobioparum* Rumen Lamed et al.,

*cellulolyticum* Compost Pagès et al.,

*cellulovorans* Fermenter Sleat et al., 1984 *Clostridium josui* Compost Kakiuchi et al.,

*papyrosolvens* Paper mill Pohlschröder et

*Ruminococcus albus* Rumen Ohara et al.,

*flavefaciens* Rumen Ding et al., 2001

*xylanilyticum* BT14 Soil Phitsuwan et al.,

Sewage soil

Anaerobic digester

*cellulolyticus* Sewage Ding et al., 1999 *Bacillus circulans*

*cellulosolvens* Sewage Ding et al., 2000 *Paenibacillus* 

*Acetivibrio* 

*Bacteroides* 

*Bacteroides* sp. strain P-1

*Butyrivibrio* 

*Clostridium* 

*Clostridium* 

*Clostridium* 

*Clostridium* 

*Clostridium* 

*Clostridium thermocellum* 

*Eubacterium* 

*Ruminococcus* 

*Tepidimicrobium* 

*cellulosolvens* Rumen

*Amorocellulobacter* 

Anaerobic Aerobic Microorganism Source Ref. Microorganism Source Ref.

F-2

*Bacillus* 

*curdlanolyticus* B-6

*Sorangium* 

*Streptomyces* 

*Chaetomium* sp.

<sup>1987</sup>*Actinomycetes* 

<sup>1998</sup>*Fungi* 

Potato starch granules

Anaerobic digester

*cellulosum* Soil Hou, et al.,

*olivaceoviridis* E-86 Soil Jiang et al.,

Nov. MS-017 Rotted wood Ohtsuki et

*licheniformis* SVD1 Bioreactor van Dyk et

Kim and Kim, 1993

al., 2009

Pason et al., 2006b

2006

2004

al., 2005

**Figure 3.** Simplified schematic of the hydrolysis of amorphous and microcrystalline celluloses by noncomplexed (A) and complexed (B) cellulase systems. (This figure is adapted from Lynd et al., 2002).

The cellulosome was first identified in 1983 from the anaerobic, thermophilic, spore-forming *Clostridium thermocellum* (Lamed et al., 1983). The cellulosome of *C. thermocellum i*s commonly studied along with cellulosomes from the anaerobic mesophiles, *C. cellulovorans* (Doi et al., 2003), *C. josui* (Kakiuchi et al., 1998) and *C. cellulolyticum* (Gal et al., 1997). All cellulosomes share similar characteristics, they all contain a large distinct protein, referred to as the scaffoldin which allows binding of the whole complex to microcrystalline cellulose via CBM. Also, the cellulosome scaffoldin expresses type I cohesins which allow binding of a wide variety of cellulolytic and hemicellulolytic enzymes within the complex via the expression of complementary type I dockerins on enzymes. Similarly, at the C-terminal the scaffoldin expresses type II cohesins, which allow the binding of the cellulosome to the cell through type II dockerins on surface layer homology proteins (SLH) (Fig. 4).

Cellulosomes are produced mainly by anaerobic bacteria, mostly from the class clostridia, and some anaerobic fungi such as genus *Neocallimastix* (Dalrymple et al., 1997), *Piromyces* (Teunissen et al., 1991) and *Orpinomyces* (Li et al., 1997). However, evidence suggests the presence of cellulosomes or cellulosome-like multienzyme complexes in a few aerobic microorganisms (Table 1). It is speculated that several other cellulolytic bacteria may also produce cellulosomes not yet described.

components to further favor synergism, (iii) the presence of different enzymatic activities (cellulolytic or hemicellulolytic enzymes) in the cellulosome that can remove "physical hindrances" of other polysaccharides in heterogeneous plant cell materials (Lynd et al., 2002), and (iv) the presence of carbohydrate-binding modules (CBMs) that can increase the rate of hydrolysis by bringing the cellulosome into intimate and prolonged association with its recalcitrant substrate (Shoseyov et al., 2006). Thus, the complexed enzyme system,

**Figure 3.** Simplified schematic of the hydrolysis of amorphous and microcrystalline celluloses by noncomplexed (A) and complexed (B) cellulase systems. (This figure is adapted from Lynd et al., 2002).

The cellulosome was first identified in 1983 from the anaerobic, thermophilic, spore-forming *Clostridium thermocellum* (Lamed et al., 1983). The cellulosome of *C. thermocellum i*s commonly studied along with cellulosomes from the anaerobic mesophiles, *C. cellulovorans* (Doi et al., 2003), *C. josui* (Kakiuchi et al., 1998) and *C. cellulolyticum* (Gal et al., 1997). All cellulosomes share similar characteristics, they all contain a large distinct protein, referred to as the scaffoldin which allows binding of the whole complex to microcrystalline cellulose via CBM. Also, the cellulosome scaffoldin expresses type I cohesins which allow binding of a wide variety of cellulolytic and hemicellulolytic enzymes within the complex via the expression of complementary type I dockerins on enzymes. Similarly, at the C-terminal the scaffoldin expresses type II cohesins, which allow the binding of the cellulosome to the cell

Cellulosomes are produced mainly by anaerobic bacteria, mostly from the class clostridia, and some anaerobic fungi such as genus *Neocallimastix* (Dalrymple et al., 1997), *Piromyces* (Teunissen et al., 1991) and *Orpinomyces* (Li et al., 1997). However, evidence suggests the presence of cellulosomes or cellulosome-like multienzyme complexes in a few aerobic microorganisms (Table 1). It is speculated that several other cellulolytic bacteria may also

through type II dockerins on surface layer homology proteins (SLH) (Fig. 4).

produce cellulosomes not yet described.

cellulosome, may provide great potential for the degradation of plant biomass.

**Figure 4.** Simplified schematic of general cellulosome components and connection with cell surface based on knowledge of *Clostridium* cellulosome. (This figure is adapted from Bayer et al., 1994).



*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 377

Adhesion of cells to insoluble substances (%)

Zymogram analysis

Xylanase band

band

xylanases in the form of a high molecular weight complex. Thus, strain B-6 exhibits great promise bacterium in the production of multienzyme complex under aerobic conditions. Some properties of bacterial cells and cellulase and xylanase from strain B-6 compared with

> Enzyme binding ability to insoluble substances (%)

Avicel grown 0.16 1.12 57.1 64.3 28.0 39.9 11 13 Xylan grown 0.12 7.19 39.1 51.5 13.6 74.7 9 15

Avicel grown 0.15 1.10 50.0 50.0 0 0 2 2 Xylan grown 0.09 4.23 31.1 38.5 0 0 1 3

Avicel grown 0.15 0.90 43.4 49.1 0 0 3 2 Xylan grown 0.09 4.49 37.9 45.8 0 0 2 3

Avicel grown 0 0 0 0 0 0 0 0 Xylan grown 0.05 3.29 29.2 45.0 0 0 0 2

Avicel grown 0 0 0 0 0 0 0 0 Xylan grown 0.06 3.19 29.6 36.1 0 0 0 2

Avicel grown 0 0 0 0 0 0 0 0 Xylan grown 0.04 3.10 28.2 38.2 0 0 0 2 **Table 2.** Production of carboxymethyl cellulase (CMCase) and xylanase by *Bacillus* strains; binding ability of enzymes to insoluble substances; adherence of bacterial cells to insoluble substances; and

*P. curdlanolyticus* strain B-6 was a facultative, spore-forming, Gram-positive, motile, rod-shaped organism and produced catalase. Thus, this bacterium was identified as a member of the genus *Bacillus* according to Bergey's Manual of Systematic Bacteriology (Sneath, 1986). The bacterium was also identified by 16S rRNA gene sequence analysis. The use of a specific PCR primer designed for differentiating the genus *Paenibacillus* from other members of the *Bacillaceae* showed that this strain had the same amplified 16S rRNA gene fragment as a member of the genus *Paenibacillus*. Based on these observations, it is reckoned that this strain was transferred to the genus *Paenibacillus* (Shida et al., 1997). The 16S rDNA sequence of this strain had 1,424 base

pairs and 97% similarity with *Paenibacillus curdlanolyticus* (Innis & Gelfand, 1990).

CMCase Xylanase Avicel Xylan Avicel Xylan CMCase

other *Bacillus* spp. are shown in Table 2.

Specific activity (U/mg protein)

Strain (*Bacillus* sp.) and growth condition

1. Strain B-6

2. Strain H-4

3. Strain S-1

4. Strain X-11

5. Strain X-24

6. Stain X-26

zymograms analysis in culture supernatant.

**Table 1.** Cellulosome and cellulosome-like multienzyme complexes from anaerobic and aerobic microorganisms. (This table is adapted from Doi & Kosugi, 2004).

### **2. Novel multienzyme complex system from** *P. curdlanolyticus* **strain B-6**

Efficient enzymatic degradation of lignocellulosic biomass requires a tight interaction between the enzymes and their substrates, and the cooperation of multiple enzymes to enhance the hydrolysis due to the complex structure. Multienzyme complexes, cellulosomes from anaerobic cellulolytic microorganisms, are dedicated to hydrolyzing lignocellulosic substances efficiently because of a large variety of cellulases and hemicellulases in complexes, useful enzymatic properties, and binding ability to insoluble cellulose and/or xylan via CBMs (Bayer et al., 2004; Doi and Kosugi, 2004; Schwarz et al., 2001; Shoham et al., 1999). When compared with aerobic enzymes, they produce several individual enzymes, but microorganisms are not binding to insoluble substrates. However, *P. curdlanolyticus* B-6 was found to produce a multienzyme complex under aerobic conditions (Pason et al., 2006a, 2006b). Little information has been reported on cellulosome-like multienzyme complexes produced by aerobic bacterium (Kim & Kim, 1993; Jiang et al., 2004; van Dyk et al., 2009). Therefore, the multienzyme complex produced by strain B-6 is critical for improving plant biomass degradation.

### **2.1. Selection of multienzyme complex-producing bacteria under aerobic cultivation**

Among several *Bacillus* strains, isolated from various sources and cultivated under aerobic conditions, *P. curdlanolyticus* strain B-6 shows important evidences for multienzyme complex producing bacterium (Pason et al., 2006a) as follows: high production of cellulase and xylanase, presence of CBMs that have ability to bind to insoluble substances, adhesion of bacterial cells to insoluble substances, and production of multiple cellulases and xylanases in the form of a high molecular weight complex. Thus, strain B-6 exhibits great promise bacterium in the production of multienzyme complex under aerobic conditions. Some properties of bacterial cells and cellulase and xylanase from strain B-6 compared with other *Bacillus* spp. are shown in Table 2.

376 Biomass Now – Cultivation and Utilization

*patriciarum* Rumen Dalrymple et

*Orpinomyces joyonii* Rumen Qiu et al., 2000 *Orpinomyces PC-2* Rumen Borneman et al.,

*Piromyces equi* Rumen Teunissen et al.,

*Piromyces E2* Faeces Teunissen et al.,

microorganisms. (This table is adapted from Doi & Kosugi, 2004).

*Thermoanaerobacterium thermosaccharolyticum* 

biomass degradation.

**cultivation** 

NOI-1

*Fungi Neocallimastix* 

Anaerobic Aerobic Microorganism Source Ref. Microorganism Source Ref.

Soil Chimtong et al., 2011

al., 1997

1989

1991

1991 **Table 1.** Cellulosome and cellulosome-like multienzyme complexes from anaerobic and aerobic

**2.1. Selection of multienzyme complex-producing bacteria under aerobic** 

Among several *Bacillus* strains, isolated from various sources and cultivated under aerobic conditions, *P. curdlanolyticus* strain B-6 shows important evidences for multienzyme complex producing bacterium (Pason et al., 2006a) as follows: high production of cellulase and xylanase, presence of CBMs that have ability to bind to insoluble substances, adhesion of bacterial cells to insoluble substances, and production of multiple cellulases and

**2. Novel multienzyme complex system from** *P. curdlanolyticus* **strain B-6** 

Efficient enzymatic degradation of lignocellulosic biomass requires a tight interaction between the enzymes and their substrates, and the cooperation of multiple enzymes to enhance the hydrolysis due to the complex structure. Multienzyme complexes, cellulosomes from anaerobic cellulolytic microorganisms, are dedicated to hydrolyzing lignocellulosic substances efficiently because of a large variety of cellulases and hemicellulases in complexes, useful enzymatic properties, and binding ability to insoluble cellulose and/or xylan via CBMs (Bayer et al., 2004; Doi and Kosugi, 2004; Schwarz et al., 2001; Shoham et al., 1999). When compared with aerobic enzymes, they produce several individual enzymes, but microorganisms are not binding to insoluble substrates. However, *P. curdlanolyticus* B-6 was found to produce a multienzyme complex under aerobic conditions (Pason et al., 2006a, 2006b). Little information has been reported on cellulosome-like multienzyme complexes produced by aerobic bacterium (Kim & Kim, 1993; Jiang et al., 2004; van Dyk et al., 2009). Therefore, the multienzyme complex produced by strain B-6 is critical for improving plant


**Table 2.** Production of carboxymethyl cellulase (CMCase) and xylanase by *Bacillus* strains; binding ability of enzymes to insoluble substances; adherence of bacterial cells to insoluble substances; and zymograms analysis in culture supernatant.

*P. curdlanolyticus* strain B-6 was a facultative, spore-forming, Gram-positive, motile, rod-shaped organism and produced catalase. Thus, this bacterium was identified as a member of the genus *Bacillus* according to Bergey's Manual of Systematic Bacteriology (Sneath, 1986). The bacterium was also identified by 16S rRNA gene sequence analysis. The use of a specific PCR primer designed for differentiating the genus *Paenibacillus* from other members of the *Bacillaceae* showed that this strain had the same amplified 16S rRNA gene fragment as a member of the genus *Paenibacillus*. Based on these observations, it is reckoned that this strain was transferred to the genus *Paenibacillus* (Shida et al., 1997). The 16S rDNA sequence of this strain had 1,424 base pairs and 97% similarity with *Paenibacillus curdlanolyticus* (Innis & Gelfand, 1990).

### **2.2. Characteristics of** *P. curdlanolyticus* **B-6 multienzyme complex**

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 379

**Figure 6.** Proteins and enzymes patterns of multienzyme complex in culture supernatant at the late stationary growth phase; Native-PAGE (lane 1), SDS-PAGE (lane 2), and zymograms analysis of

**2.3. Effect of carbon sources on the induction of multienzyme complex in** *P.* 

*olivaceoviridis* E-86 (Jiang et al., 2004) were affected by carbon sources in the media.

formation by strain B-6 must be different from that of other microorganisms.

Many investigators have reported that the synthesis of cellulosome assemblies requires the presence of crystalline cellulose under anaerobic conditions, and that synthesis hardly occurs in growth on glucose or other soluble carbohydrates (Nochur et al., 1992; Blair & Anderson; 1999a; Bayer 2004; Doi & Kosugi, 2004). Some strains of *C. thermocellum* (Bayer et al., 1985; Bhat et al., 1993), however, can induce cellulosome synthesis when grown on cellobiose. *P. curdlanolyticus* B-6 differs from most cellulosome-producing microorganisms in that it produces multienzyme complex when grown on both polymeric substances and soluble sugars under aerobic conditions. Therefore, the mechanism of multienzyme complex

The effect of polymeric substances such as cellulose, xylan, corn hull, and sugarcane bagasse, and of soluble sugars such as L-arabinose, D-galactose, D-glucose, D-xylose, and cellobiose on the induction of multienzyme complexes in a facultatively anaerobic bacterium, *P. curdlanolyticus* B-6, was investigated under aerobic conditions (Waeonukul et al., 2008; 2009b). Cells grown on each carbon source adhered to cellulose. Hence strain B-6 cells from all carbon sources must have an essential component responsible for anchoring the cells to the substrate surfaces. Native–PAGE, SDS–PAGE, zymograms analysis, and enzymatic assays revealed that many proteins having xylanolytic and cellulolytic activities from *P. curdlanolyticus* B-6 grown on each carbon source were produced as multienzyme complex into the culture supernatants. These results indicated that strain B-6 produced multienzyme complexes when grown on both polymeric substances and soluble sugars. However, the subunits expressed in the multienzyme complex of strain B-6 depended on the carbon sources. These observations are consistent with previous reports that the enzymatic activities and enzyme compositions of the cellulosomes of *C. thermocellum* (Bayer et al., 1985; Bhat et al., 1993; Nochur et al., 1993), *C. cellulolyticum* (Mohand-Oussaid et al., 1999), and *C. cellulovorans* (Kosugi et al., 2001; Han et al., 2004; 2005) and the xylanosome of *S.* 

xylanase activity (lane 3), and CMCase activity (lane 4).

*curdlanolyticus* **B-6** 

During growth of *P. curdlanolyticus* B-6 on Berg's mineral salt medium containing 0.5% xylan as carbon sources, the protein concentration in the medium was low up to the late stationary growth phase. CMCase and xylanase activities could be detected in the culture medium after the late exponential phase (Pason et al., 2006b). At the declining growth phase, the extracellular xylanase and CMCase rapidly increased due to the release of enzymes from the cell surfaces into the culture medium. These phenomena were different from the growth patterns of other aerobic bacteria, which grew and produced extracellular enzymes into culture supernatant immediately, but similar to those of the anaerobic bacteria which produced multienzyme complexes (cellulosomes) around the cell surfaces and adhered to these substrates and secreted into culture supernatant later (Bayer & Lamed, 1986; Lamed & Bayer, 1988). The observation of cell surfaces at the late exponential growth phase by scanning electron microscopy (SEM) revealed that the cells adhered to xylan (Fig. 5A), similar to the cells of the cellulosome producing anaerobic bacterium, *C. thermocellum*, which is a cell associated entity that mediates the adhesion of the bacterium to cellulose (Lamed et al., 1987; Mayer et al., 1987), whereas the surface of the cells of strain B-6 at the late stationary growth phase lacked such structures because the multienzyme complex was released into the medium from the cell surfaces (Fig. 5B). In addition, the pattern of multienzyme complex in the culture medium at the late stationary growth phase was determined. Native-polyacrylamide gel electrophoresis (native-PAGE) exhibited a high molecular weight band at the top of the gel (Fig. 6, lane 1). This protein band was dissociated into major and minor components through treatment by boiling in sodium dodecyl sulphate (SDS) solution, showing at least 18 proteins with molecular masses in the range of 29 to 280 kDa (Fig. 6, lane 2). Among those protein bands, at least 15 bands showed xylanase activities (Fig. 6, lane 3) and at least 9 bands showed CMCase activities (Fig. 6, lane 4) on zymograms. These multiple cellulases and xylanases are assembled into the high molecular weight complexes and released from the cell surfaces into medium at the late stationary growth phase. In *C. thermocellum*, the cellulosome consisted of many different types of glycosyl hydrolases, including cellulases, hemicellulases, and carbohydrate esterases, which served to promote their synergistic action (Lamed et al., 1983). These evidences confirm that the strain B-6 can produce xylanolytic-cellulolytic enzyme system that exists as multienzyme complex under aerobic conditions.

**Figure 5.** SEM of the cell surfaces of *P. curdlanolyticus* B-6 harvested at the late exponential growth phase showing adhesion of cell to xylan (A) and the cell harvested at the late stationary growth phase showing no adhesion of cell to xylan (B).

**2.2. Characteristics of** *P. curdlanolyticus* **B-6 multienzyme complex** 

system that exists as multienzyme complex under aerobic conditions.

showing no adhesion of cell to xylan (B).

**Figure 5.** SEM of the cell surfaces of *P. curdlanolyticus* B-6 harvested at the late exponential growth phase showing adhesion of cell to xylan (A) and the cell harvested at the late stationary growth phase

During growth of *P. curdlanolyticus* B-6 on Berg's mineral salt medium containing 0.5% xylan as carbon sources, the protein concentration in the medium was low up to the late stationary growth phase. CMCase and xylanase activities could be detected in the culture medium after the late exponential phase (Pason et al., 2006b). At the declining growth phase, the extracellular xylanase and CMCase rapidly increased due to the release of enzymes from the cell surfaces into the culture medium. These phenomena were different from the growth patterns of other aerobic bacteria, which grew and produced extracellular enzymes into culture supernatant immediately, but similar to those of the anaerobic bacteria which produced multienzyme complexes (cellulosomes) around the cell surfaces and adhered to these substrates and secreted into culture supernatant later (Bayer & Lamed, 1986; Lamed & Bayer, 1988). The observation of cell surfaces at the late exponential growth phase by scanning electron microscopy (SEM) revealed that the cells adhered to xylan (Fig. 5A), similar to the cells of the cellulosome producing anaerobic bacterium, *C. thermocellum*, which is a cell associated entity that mediates the adhesion of the bacterium to cellulose (Lamed et al., 1987; Mayer et al., 1987), whereas the surface of the cells of strain B-6 at the late stationary growth phase lacked such structures because the multienzyme complex was released into the medium from the cell surfaces (Fig. 5B). In addition, the pattern of multienzyme complex in the culture medium at the late stationary growth phase was determined. Native-polyacrylamide gel electrophoresis (native-PAGE) exhibited a high molecular weight band at the top of the gel (Fig. 6, lane 1). This protein band was dissociated into major and minor components through treatment by boiling in sodium dodecyl sulphate (SDS) solution, showing at least 18 proteins with molecular masses in the range of 29 to 280 kDa (Fig. 6, lane 2). Among those protein bands, at least 15 bands showed xylanase activities (Fig. 6, lane 3) and at least 9 bands showed CMCase activities (Fig. 6, lane 4) on zymograms. These multiple cellulases and xylanases are assembled into the high molecular weight complexes and released from the cell surfaces into medium at the late stationary growth phase. In *C. thermocellum*, the cellulosome consisted of many different types of glycosyl hydrolases, including cellulases, hemicellulases, and carbohydrate esterases, which served to promote their synergistic action (Lamed et al., 1983). These evidences confirm that the strain B-6 can produce xylanolytic-cellulolytic enzyme

**Figure 6.** Proteins and enzymes patterns of multienzyme complex in culture supernatant at the late stationary growth phase; Native-PAGE (lane 1), SDS-PAGE (lane 2), and zymograms analysis of xylanase activity (lane 3), and CMCase activity (lane 4).

### **2.3. Effect of carbon sources on the induction of multienzyme complex in** *P. curdlanolyticus* **B-6**

The effect of polymeric substances such as cellulose, xylan, corn hull, and sugarcane bagasse, and of soluble sugars such as L-arabinose, D-galactose, D-glucose, D-xylose, and cellobiose on the induction of multienzyme complexes in a facultatively anaerobic bacterium, *P. curdlanolyticus* B-6, was investigated under aerobic conditions (Waeonukul et al., 2008; 2009b). Cells grown on each carbon source adhered to cellulose. Hence strain B-6 cells from all carbon sources must have an essential component responsible for anchoring the cells to the substrate surfaces. Native–PAGE, SDS–PAGE, zymograms analysis, and enzymatic assays revealed that many proteins having xylanolytic and cellulolytic activities from *P. curdlanolyticus* B-6 grown on each carbon source were produced as multienzyme complex into the culture supernatants. These results indicated that strain B-6 produced multienzyme complexes when grown on both polymeric substances and soluble sugars. However, the subunits expressed in the multienzyme complex of strain B-6 depended on the carbon sources. These observations are consistent with previous reports that the enzymatic activities and enzyme compositions of the cellulosomes of *C. thermocellum* (Bayer et al., 1985; Bhat et al., 1993; Nochur et al., 1993), *C. cellulolyticum* (Mohand-Oussaid et al., 1999), and *C. cellulovorans* (Kosugi et al., 2001; Han et al., 2004; 2005) and the xylanosome of *S. olivaceoviridis* E-86 (Jiang et al., 2004) were affected by carbon sources in the media.

Many investigators have reported that the synthesis of cellulosome assemblies requires the presence of crystalline cellulose under anaerobic conditions, and that synthesis hardly occurs in growth on glucose or other soluble carbohydrates (Nochur et al., 1992; Blair & Anderson; 1999a; Bayer 2004; Doi & Kosugi, 2004). Some strains of *C. thermocellum* (Bayer et al., 1985; Bhat et al., 1993), however, can induce cellulosome synthesis when grown on cellobiose. *P. curdlanolyticus* B-6 differs from most cellulosome-producing microorganisms in that it produces multienzyme complex when grown on both polymeric substances and soluble sugars under aerobic conditions. Therefore, the mechanism of multienzyme complex formation by strain B-6 must be different from that of other microorganisms.

### **3. The feature of** *P. curdlanolyticus* **B-6 multienzyme complex**

Recently, the structures and mechanisms for assembly of multienzyme complexes, cellulosomes, in anaerobic cellulolytic microorganisms are clear (Bayer et al., 2004, 2007; Doi & Kosugi, 2004). Generally, the key feature of the cellulosome is a scaffoldin that integrates the various catalytic subunits into the complex by self-assembly by cohesion-dockerin interaction. However, the structure and mechanism of the multienzyme complex produced by a facultatively anaerobic bacterium, such as *P. curdlanolyticus* B-6 is still unknown. In order to describe features of the multienzyme complex system produced by strain B-6, the multienzyme complex was purified by four kinds of chromatography (cellulose affinity, gel filtration, anion-exchange and hydrophobic-interaction chromatographys) (Fig. 7).

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 381

Protein

subunits Ref.

(kDa)

*Paenibacillus curdlanolyticus* B-6 1450 11 Pason et al.,2006b *Bacillus circulans* F-2 669 7 Kim and Kim, 1993 *Bacillus licheniformis* SVD1 2000 12 van Dyk et al., 2009 *Sorangium cellulosum* 1000-2000 10 Hou et al., 2006 *Streptomyces olivaceoviridis* E-86 1200 5 Jiang et al., 2004 *Chaetomium* sp. Nov. MS-017 468 12 Ohtsuki et al., 2005 

*Clostridium acetobutylicum* 665 11 Sabathé et al.,2002 *Clostridium cellulolyticum* 600 14 Gal et al., 1997 *Clostridium cellulovorans* 900 10 Shoseyov & Doi 1990 *Clostridium josui* 700 14 Kakiuchi et al., 1998 *Clostridium popyrosolvens* 600 15 Pohlschröder et al., 1994 *Clostridium thermocellum* 2100 14 Lamed et al., 1983 *Ruminococcus albus* 1500 15 Ohara et al., 2000 **Table 3.** Molecular weights and protein subunits of multienzyme complexes from aerobic and

Elucidation of the purified multienzyme feature of *P. curdlanolyticus* strain B-6 was followed by anion-exchange and hydrophobic-interaction chromatographys (Pason et al., 2010). The complex G1 from gel filtration chromatography (1,450 kDa) was purified by anion-exchange chromatography and showed at least five large protein complexes or aggregates, namely F1-F5. Among the fractions obtained from anion-exchange chromatography, F1 was apparently the most suited fraction to study on the organization and function of the multienzyme system of strain B-6 because F1 formed one clear band on the top of native PAGE, had the highest xylanase activity, and its subunit composition was clearly shown on SDS-PAGE. In the final step, complex F1 was separated to one major complex (H1) and two minor protein components (H2 and H3) by hydrophobicinteraction chromatography. The multienzyme complex (H1) was composed of a 280 kDa protein with xylanase activity, a 260 kDa protein that is a truncated form on the Cterminal side of the 280 kDa protein, two xylanases of 40 and 48 kDa, and 60 and 65 kDa proteins having both xylanase and CMCase activities (Fig. 8). The two components (280 and 40 kDa) of the multienzyme complex has characteristics similar to the cellulosome of *C. thermocellum* in that it is composed of a scaffolding protein and a catalytic subunit (Bayer et al., 1998; Demain et al., 2005). The 280 kDa protein resembled the scaffolding proteins of the multienzyme complex based on its migratory behavior in polyacrylamide gels and as a glycoprotein. The 280 kDa protein and a 40 kDa major xylanase subunit are

the key components of multienzyme complex of the strain B-6.

Multienzyme complex Mol. Mass

Aerobic microorganisms

Anaerobic microorganisms

anaerobic microorganisms.

**Figure 7.** Isolation and purification of multienzyme complex of *P. curdlanolyticus* strain B-6.

The multienzyme complex of *P. curdlanolyticus* strain B-6 with molecular mass of 1,450 (G1) was isolated from culture supernatant at the late stationary growth phase through cellulose affinity and Sephacryl S-300 gel filtration chromatographys (Pason et al., 2006b). Basically, the individual cellulosomes from anaerobic bacteria show 600 kDa to 2.1 MDa complexes size and show cohesion-dockerin domain as a signature protein (Bayer et al., 2004; Doi & Kosugi, 2004). While, multienzyme complexes from aerobic microorganisms, were range in mass from about 468 kDa to 2 MDa (with contained 5-12 protein subunits) (Table 3) and has no report of cohesion-dockerin domain. Here, the multienzyme complex produced by strain B-6 under aerobic conditions was the first report on characterization.


*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 381

380 Biomass Now – Cultivation and Utilization

**3. The feature of** *P. curdlanolyticus* **B-6 multienzyme complex** 

filtration, anion-exchange and hydrophobic-interaction chromatographys) (Fig. 7).

**Figure 7.** Isolation and purification of multienzyme complex of *P. curdlanolyticus* strain B-6.

B-6 under aerobic conditions was the first report on characterization.

The multienzyme complex of *P. curdlanolyticus* strain B-6 with molecular mass of 1,450 (G1) was isolated from culture supernatant at the late stationary growth phase through cellulose affinity and Sephacryl S-300 gel filtration chromatographys (Pason et al., 2006b). Basically, the individual cellulosomes from anaerobic bacteria show 600 kDa to 2.1 MDa complexes size and show cohesion-dockerin domain as a signature protein (Bayer et al., 2004; Doi & Kosugi, 2004). While, multienzyme complexes from aerobic microorganisms, were range in mass from about 468 kDa to 2 MDa (with contained 5-12 protein subunits) (Table 3) and has no report of cohesion-dockerin domain. Here, the multienzyme complex produced by strain

Recently, the structures and mechanisms for assembly of multienzyme complexes, cellulosomes, in anaerobic cellulolytic microorganisms are clear (Bayer et al., 2004, 2007; Doi & Kosugi, 2004). Generally, the key feature of the cellulosome is a scaffoldin that integrates the various catalytic subunits into the complex by self-assembly by cohesion-dockerin interaction. However, the structure and mechanism of the multienzyme complex produced by a facultatively anaerobic bacterium, such as *P. curdlanolyticus* B-6 is still unknown. In order to describe features of the multienzyme complex system produced by strain B-6, the multienzyme complex was purified by four kinds of chromatography (cellulose affinity, gel

> **Table 3.** Molecular weights and protein subunits of multienzyme complexes from aerobic and anaerobic microorganisms.

Elucidation of the purified multienzyme feature of *P. curdlanolyticus* strain B-6 was followed by anion-exchange and hydrophobic-interaction chromatographys (Pason et al., 2010). The complex G1 from gel filtration chromatography (1,450 kDa) was purified by anion-exchange chromatography and showed at least five large protein complexes or aggregates, namely F1-F5. Among the fractions obtained from anion-exchange chromatography, F1 was apparently the most suited fraction to study on the organization and function of the multienzyme system of strain B-6 because F1 formed one clear band on the top of native PAGE, had the highest xylanase activity, and its subunit composition was clearly shown on SDS-PAGE. In the final step, complex F1 was separated to one major complex (H1) and two minor protein components (H2 and H3) by hydrophobicinteraction chromatography. The multienzyme complex (H1) was composed of a 280 kDa protein with xylanase activity, a 260 kDa protein that is a truncated form on the Cterminal side of the 280 kDa protein, two xylanases of 40 and 48 kDa, and 60 and 65 kDa proteins having both xylanase and CMCase activities (Fig. 8). The two components (280 and 40 kDa) of the multienzyme complex has characteristics similar to the cellulosome of *C. thermocellum* in that it is composed of a scaffolding protein and a catalytic subunit (Bayer et al., 1998; Demain et al., 2005). The 280 kDa protein resembled the scaffolding proteins of the multienzyme complex based on its migratory behavior in polyacrylamide gels and as a glycoprotein. The 280 kDa protein and a 40 kDa major xylanase subunit are the key components of multienzyme complex of the strain B-6.

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 383

family

Mol. Mass (kDa) GenBank accession No.

Enzyme Modular structure GH

surface layer homology domain; U, unhomology sequence

(unpublished data).

(Waeonukul et al., 2009a).

**Table 4.** Modular structure xylanases of *P. curdlanolyticus* strain B-6.

S1 U 91 -

Xyn10A 10 142 EU418764

Xyn10B 10 40 AB570291

Xyn10C 10 35 AB688987

Xyn10D 10 61 AB600191

Xyn11A 11 40 FJ956758

Abbreviations: CBM, carbohydrate-binding module; Fn, fibronectin homology module; GH, glycosyl hydrolase; SLH,

**S1 protein:** From the early research, the 280 kDa subunit (S1) plays a role of scaffoldin in assembling the enzyme complex and shows xylanase activity (Pason et al., 2010). The *S1* gene consists of 2,589 nucleotides and encodes 863 amino acids with a molecular weight of 91,000 Da, indicating that the 280 kDa subunit is highly glycosylated. Sequence analysis revealed that S1 did not have significant homology with any proteins in the databases except for two surface layer homology (SLH) domains in its N-terminal region. Surprisingly, the recombinant S1 exhibits xylanase activity, and cellulose- and xylan-binding ability, suggesting that the S1 should be a novel xylanase and CBM(s) with new functions

**Xylanase Xyn10A**: The *xyn10A* gene consists of 3,828 nucleotides encoding a protein of 1,276 amino acids with a predicted molecular weight of 142,726 Da. Xyn10A is a multidomain enzyme comprised of nine domains in the following order: three family-22 CBMs, a family-10 catalytic domain of glycosyl hydrolases (GH), a family-9 CBM, a glycine-rich region, and three SLH domains. Xyn10A can effectively hydrolyze insoluble xylan and natural biomass without pretreatment such as sugarcane bagasse, corn hull, rice bran, rice husk and rice straw. Xyn10A binds to various insoluble polysaccharides such as cellulose, xylan and chitin. The SLH domains functioned in Xyn10A by anchoring this enzyme to the cell surfaces of *P. curdlanolyticus* B-6. Removal of the CBMs from Xyn10A strongly reduced the ability of binding and plant cell wall hydrolysis. Therefore, the CBMs of Xyn10A play an important role in the hydrolysis of native biomass materials

**Xylanase Xyn10B:** The *xyn10B* gene consists of 1,047 nucleotides encoding a protein of 349 amino acids with a predicted molecular weight of 40,480 Da. Xyn10B consists of only a

family-10 catalytic of GH. Xyn10B is an intracellular endoxylanase (Sudo et al., 2010).

**Figure 8.** Native-PAGE (A) and SDS-PAGE (B) in isolated complex from culture supernatant at the late stationary growth phase (lane Cr), affinity column (lane A1), gel filtration column (lane G1), anionexchange column (lane F1) and hydrophobic-interaction column (lane H1). All samples contained 200 g of protein.

These apparently propose that *P. curdlanolyticus* B-6 produced multienzyme complex, which consisted of many subunit compositions. The large protein (280 kDa) may function as a scaffoldin-like protein that allowed the enzyme subunits, majority is 40 kDa, binding to form a multienzyme complex. The key components, 280 and 40 kDa, are identified in the next topic.

### **4. Molecular structure of important xylanases**

*P. curdlanolyticus* B-6 produces an extracellular xylanolytic-cellulolytic multienzyme complex mainly comprised of xylanases under aerobic conditions. To understand the xylanase system, a genomic library of the strain B-6 was constructed and screened for high xylanase activity. Recently, six xylanase genes, *S1* (Pason et al., 2010)*, xyn10A* (Waeonukul et al., 2009a), *xyn10B* (Sudo et al., 2010), *xyn10C* (unpublished data), *xyn10D* (Sakka et al., 2011) and *xyn11A* (Pason et al., 2010) were cloned, and the translated products were characterized (Table 4).


Abbreviations: CBM, carbohydrate-binding module; Fn, fibronectin homology module; GH, glycosyl hydrolase; SLH, surface layer homology domain; U, unhomology sequence

**Table 4.** Modular structure xylanases of *P. curdlanolyticus* strain B-6.

382 Biomass Now – Cultivation and Utilization

g of protein.

next topic.

(Table 4).

**Figure 8.** Native-PAGE (A) and SDS-PAGE (B) in isolated complex from culture supernatant at the late stationary growth phase (lane Cr), affinity column (lane A1), gel filtration column (lane G1), anionexchange column (lane F1) and hydrophobic-interaction column (lane H1). All samples contained 200

These apparently propose that *P. curdlanolyticus* B-6 produced multienzyme complex, which consisted of many subunit compositions. The large protein (280 kDa) may function as a scaffoldin-like protein that allowed the enzyme subunits, majority is 40 kDa, binding to form a multienzyme complex. The key components, 280 and 40 kDa, are identified in the

*P. curdlanolyticus* B-6 produces an extracellular xylanolytic-cellulolytic multienzyme complex mainly comprised of xylanases under aerobic conditions. To understand the xylanase system, a genomic library of the strain B-6 was constructed and screened for high xylanase activity. Recently, six xylanase genes, *S1* (Pason et al., 2010)*, xyn10A* (Waeonukul et al., 2009a), *xyn10B* (Sudo et al., 2010), *xyn10C* (unpublished data), *xyn10D* (Sakka et al., 2011) and *xyn11A* (Pason et al., 2010) were cloned, and the translated products were characterized

**4. Molecular structure of important xylanases** 

**S1 protein:** From the early research, the 280 kDa subunit (S1) plays a role of scaffoldin in assembling the enzyme complex and shows xylanase activity (Pason et al., 2010). The *S1* gene consists of 2,589 nucleotides and encodes 863 amino acids with a molecular weight of 91,000 Da, indicating that the 280 kDa subunit is highly glycosylated. Sequence analysis revealed that S1 did not have significant homology with any proteins in the databases except for two surface layer homology (SLH) domains in its N-terminal region. Surprisingly, the recombinant S1 exhibits xylanase activity, and cellulose- and xylan-binding ability, suggesting that the S1 should be a novel xylanase and CBM(s) with new functions (unpublished data).

**Xylanase Xyn10A**: The *xyn10A* gene consists of 3,828 nucleotides encoding a protein of 1,276 amino acids with a predicted molecular weight of 142,726 Da. Xyn10A is a multidomain enzyme comprised of nine domains in the following order: three family-22 CBMs, a family-10 catalytic domain of glycosyl hydrolases (GH), a family-9 CBM, a glycine-rich region, and three SLH domains. Xyn10A can effectively hydrolyze insoluble xylan and natural biomass without pretreatment such as sugarcane bagasse, corn hull, rice bran, rice husk and rice straw. Xyn10A binds to various insoluble polysaccharides such as cellulose, xylan and chitin. The SLH domains functioned in Xyn10A by anchoring this enzyme to the cell surfaces of *P. curdlanolyticus* B-6. Removal of the CBMs from Xyn10A strongly reduced the ability of binding and plant cell wall hydrolysis. Therefore, the CBMs of Xyn10A play an important role in the hydrolysis of native biomass materials (Waeonukul et al., 2009a).

**Xylanase Xyn10B:** The *xyn10B* gene consists of 1,047 nucleotides encoding a protein of 349 amino acids with a predicted molecular weight of 40,480 Da. Xyn10B consists of only a family-10 catalytic of GH. Xyn10B is an intracellular endoxylanase (Sudo et al., 2010).

**Xylanase Xyn10C:** The *xyn10C* gene consists of 957 nucleotides and encodes 318 amino acid residues with a predicted molecular weight of 35,123 Da. Xyn10C is a single module enzyme consisting of a signal peptide and a family-10 catalytic module of GH (unpublished data).

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 385

xylanase subunit, Xyn11A. In addition, strain B-6 also produces cell bound multimodular xylanase Xyn10A that contains the numerous CBMs and SLH domains. Xyn10A can bind to the plant cell wall through CBM, whereas the catalytic module (GH10) is able to access its target substrate. Thus, the CBM greatly increases the concentration of the enzyme in the vicinity of the substrate, leading to the observed increase in polysaccharide hydrolysis. Besides, the presence of the functional CBMs and SLH domains in Xyn10A allows the cells to attach to substrate. Although, the overall structure of the enzyme complex system of the strain B-6 is not entirely clear, the enzyme complex has unique characteristics distinct from multienzyme complex cellulosome of anaerobic microorganisms. However, the mechanism for complex formation, interaction between the S1 protein as scaffoldin and enzyme

**5. Biotechnological uses of** *P. curdlanolyticus* **B-6 multienzyme complex** 

**Figure 10.** The multienzyme complex of *P. curdlanolyticus* strain B-6 for biotechnological applications.

Biological conversion of lignocellulosic materials has been proposed as a renewable and sustainable route for the production of value-added products (Bayer et al., 2007, Doi et al., 2003). There is much interest in exploiting the properties of multienzyme complexes for practical purposes. The facultative bacterium, *P. curdlanolyticus* strain B-6 produces a unique extracellular multienzyme system under aerobic conditions that effectively degrade cellulose and hemicellulose by gaining access through the protective matrix surrounding the cellulose microfibrils of plant cell walls. Therefore, the multienzyme complex from strain B-6 is a promising enzyme which can potentially be used in many applications, such as enhancing extraction and production of value-added bioproducts by saccharification of cell wall components and application for construction of the modular

subunits, needs to be further investigated.

enzymes creation (Fig. 10).

**Xylanase Xyn10D:** The *xyn10D* gene consists of 1,734 nucleotides and encodes 577 amino acid residues with a calculated molecular weight of 61,811 Da. Xylanase Xyn10D is a modular enzyme consisting of a family-10 catalytic module of the GH, a fibronectin type-3 homology (Fn3) module, and family-3 CBM, in that order, from the N terminus. The CBM3 in Xyn10D has an affinity for cellulose and xylan, and plays an important role in hydrolysis of arabinoxylan and native biomass materials (Sakka et al., 2011).

**Xylanase Xyn11A:** The *xyn11A* gene consists of 1,150 bp and encodes a protein of 385 amino acids with a molecular weight of 40,000 Da. Xyn11A is composed of two major functional domains, a catalytic domain belonging to family-11 GH and a CBM classified as family-36. A glycine- and asparagine-repeated sequences existed between the two domains. Xyn11A has been identified to be one of the major xylanase subunit in the multienzyme complex of strain B-6 (Pason et al., 2010).

**Figure 9.** Simplified schematic view of the interaction between the *P. curdlanolyticus* B-6 multienzyme complex system and its substrate, and its connection to the cell surface via an associated anchoring protein. (Abbreviations: CBM, carbohydrate-binding module; CD, catalytic domain, En, enzyme subunit; SLH, surface layer homology domain).

Based on both biochemical and molecular biological findings, a simplistic schematic view of the enzyme system from *P. curdlanolyticus* B-6 and its interaction with substrate and cell surface was created and presented in Fig. 9. In this assessment, the S1 protein did not have significant homology with any proteins in the databases except for two S-layer homology domains in its N-terminal region. However, the S1 protein that exhibits xylanase activity and cellulose- and xylan-binding ability, and contains cell anchoring function, seems remarkable. The multifunctional protein S1 is also responsible for forming the enzyme subunits into the complex and anchoring the complex into cell surface via the SLH domain. The interaction between S1 protein and enzyme subunits should be a mechanism distinct from the cohesion-dockerin interaction known in cellulosome of anaerobic microorganisms, since cohesion- or dockerin– like sequences were not observed in the S1 protein or the major xylanase subunit, Xyn11A. In addition, strain B-6 also produces cell bound multimodular xylanase Xyn10A that contains the numerous CBMs and SLH domains. Xyn10A can bind to the plant cell wall through CBM, whereas the catalytic module (GH10) is able to access its target substrate. Thus, the CBM greatly increases the concentration of the enzyme in the vicinity of the substrate, leading to the observed increase in polysaccharide hydrolysis. Besides, the presence of the functional CBMs and SLH domains in Xyn10A allows the cells to attach to substrate. Although, the overall structure of the enzyme complex system of the strain B-6 is not entirely clear, the enzyme complex has unique characteristics distinct from multienzyme complex cellulosome of anaerobic microorganisms. However, the mechanism for complex formation, interaction between the S1 protein as scaffoldin and enzyme subunits, needs to be further investigated.

384 Biomass Now – Cultivation and Utilization

strain B-6 (Pason et al., 2010).

subunit; SLH, surface layer homology domain).

**Xylanase Xyn10C:** The *xyn10C* gene consists of 957 nucleotides and encodes 318 amino acid residues with a predicted molecular weight of 35,123 Da. Xyn10C is a single module enzyme consisting of a signal peptide and a family-10 catalytic module of GH (unpublished data).

**Xylanase Xyn10D:** The *xyn10D* gene consists of 1,734 nucleotides and encodes 577 amino acid residues with a calculated molecular weight of 61,811 Da. Xylanase Xyn10D is a modular enzyme consisting of a family-10 catalytic module of the GH, a fibronectin type-3 homology (Fn3) module, and family-3 CBM, in that order, from the N terminus. The CBM3 in Xyn10D has an affinity for cellulose and xylan, and plays an important role in hydrolysis

**Xylanase Xyn11A:** The *xyn11A* gene consists of 1,150 bp and encodes a protein of 385 amino acids with a molecular weight of 40,000 Da. Xyn11A is composed of two major functional domains, a catalytic domain belonging to family-11 GH and a CBM classified as family-36. A glycine- and asparagine-repeated sequences existed between the two domains. Xyn11A has been identified to be one of the major xylanase subunit in the multienzyme complex of

**Figure 9.** Simplified schematic view of the interaction between the *P. curdlanolyticus* B-6 multienzyme complex system and its substrate, and its connection to the cell surface via an associated anchoring protein. (Abbreviations: CBM, carbohydrate-binding module; CD, catalytic domain, En, enzyme

Based on both biochemical and molecular biological findings, a simplistic schematic view of the enzyme system from *P. curdlanolyticus* B-6 and its interaction with substrate and cell surface was created and presented in Fig. 9. In this assessment, the S1 protein did not have significant homology with any proteins in the databases except for two S-layer homology domains in its N-terminal region. However, the S1 protein that exhibits xylanase activity and cellulose- and xylan-binding ability, and contains cell anchoring function, seems remarkable. The multifunctional protein S1 is also responsible for forming the enzyme subunits into the complex and anchoring the complex into cell surface via the SLH domain. The interaction between S1 protein and enzyme subunits should be a mechanism distinct from the cohesion-dockerin interaction known in cellulosome of anaerobic microorganisms, since cohesion- or dockerin– like sequences were not observed in the S1 protein or the major

of arabinoxylan and native biomass materials (Sakka et al., 2011).

### **5. Biotechnological uses of** *P. curdlanolyticus* **B-6 multienzyme complex**

Biological conversion of lignocellulosic materials has been proposed as a renewable and sustainable route for the production of value-added products (Bayer et al., 2007, Doi et al., 2003). There is much interest in exploiting the properties of multienzyme complexes for practical purposes. The facultative bacterium, *P. curdlanolyticus* strain B-6 produces a unique extracellular multienzyme system under aerobic conditions that effectively degrade cellulose and hemicellulose by gaining access through the protective matrix surrounding the cellulose microfibrils of plant cell walls. Therefore, the multienzyme complex from strain B-6 is a promising enzyme which can potentially be used in many applications, such as enhancing extraction and production of value-added bioproducts by saccharification of cell wall components and application for construction of the modular enzymes creation (Fig. 10).

**Figure 10.** The multienzyme complex of *P. curdlanolyticus* strain B-6 for biotechnological applications.

Biological treatment and saccharification using microorganisms and their enzymes selectively for degradation of lignocellulosic residues has the advantages of low energy consumption, minimal waste production, and environmental friendliness (Schwarz, 2001). The catalytic components of the multienzyme complex release soluble sugars, simple 5- and 6-carbon, from lignocellulose providing the primary carbon substrates, which can be subsequently converted into fuels by microorganisms. For enzyme saccharification, the close proximity between cellulolytic and xylanolytic enzymes is key to concerted degradation of the substrate, whereby the activities of the different enzymes facilitate the activities of their counterparts by promoting access to appropriated portions of the rigid insoluble substrates, since the release of sugar products was high. The synergistic action of the combination of enzymes by different modes of actions (xylanases and cellulases) and the presence of xylan- or cellulose-binding ability on lignocellulose enhanced soluble sugars released from the plant cell walls. In practicality, the multienzyme complex produced from *P. curdlanolyticus* B-6 allows access to lignocellulosic substrate and produces reducing sugar more than non-complexed enzymes from fungi (*T. viride* and *Aspergillus niger*) when the same cellulase activity (0.1 unit) was applied for degradation of corn hull and rice straw residues (unpublished data). In addition, the multienzyme complex of strain B-6 has been used to improve the extraction of plant food such as making low-cyanide-cassava starch by using multienzyme complex to enhance linamarin released by allowing more contact between linamarase and linamarin (Sornyotha et. al., 2009). Also, extraction of volatile compounds such as sea food-like flavor from seaweed, served for food supplement. Consequently, enzymatic treatment has advantages for the preparation of β-glucan and acidic α-glucan-protein complex from the fruiting body of mushroom, *Pleurotus sajor-caju* because the specificity of the multienzyme complex and gentle conditions allow for the recovery of high purity glucans in their native forms with minimal degradation (Satitmanwiwat et al., 2012a,b).

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 387

cellulolytic multienzyme complex capable of efficient degradation of plant biomass materials under aerobic conditions. The production of strain B-6 multienzyme complex under aerobic conditions has several advantages: (i) a simple process, (ii) low price of medium, (iii) high growth rate, (iv) large quantities of extracellular enzymes yields, and (v) safe use with regard to health and environmental aspects. Thus, strain B-6 and its multienzyme complex is a promising tool for an industrial process employing direct hydrolysis for the bioconversion of cellulose as well as hemicellulose in biomass. This review shows that strain B-6 multienzyme complex is a novel enzymatic system known at the biochemical, genetic, and mechanism level. It also stresses that some points still need to be further investigated, mainly (i) the elucidation of scaffolding protein functions, (ii) the characterization of others key enzyme subunits, (iii) the assembly mechanism of the multienzyme complex, (iv) improvement of the efficiency in degradation of biomass of the multienzyme complex, and (v) improvement of the production of the multienzyme complex.

Khanok Ratanakhanokchai, Rattiya Waeonukul, Patthra Pason, Chakrit Tachaapaikoon and

Aro, N.; Pakula, T. & Penttilä, M. (2005). Transcriptional Regulation of Plant Cell Wall Degradation by Filamentous Fungi. *FEMS Microbiology Reviews,* Vol.29, No.4,

Aspinall, G.O. (1980). Chemistry of Cell-Wall Polysaccharides, In: *The Biochemistry of Plants*. *A Comprehensive Treatise,* Vol.3, Preiss J., pp. (473–500), Academic Press, ISBN 0-12-

Bayer, E.A.; Belaich, J.P.; Shoham, Y. & Lamed, R. (2004). The Cellulosomes: Multienzyme Machines for Degradation of Plant Cell Wall Polysaccharides. *Annual Review of* 

Bayer, E.A.; Chanzyt, H.; Lamed, R. & Shoham, Y. (1998). Cellulose, Cellulases and Cellulosomes. *Current Opinion in Structural Biology*, Vol.8, No.5, (October, 1998), pp. 548-

Bayer, E.A. & Lamed, R. (1986). Ultrastructure of the Cell Surface Cellulosome of *Clostridium thermocellum* and Its Interaction with Cellulose. *Journal of Bacteriology*, Vol.167, No.3,

*Microbiology*, Vol.58, (October, 2004), pp. 521-554, ISSN 0066-4227

The latter will certainly represent a challenge for future research.

*King Mongkut's University of Technology Thonburi, Thailand* 

*Japan International Research Center for Agricultural Sciences, Japan* 

(September, 2005), pp. 719-739, ISSN 0168-6445

(September, 1986), pp. 828-836, ISSN 0021-9193

**Author details** 

Khin Lay Kyu

Kazuo Sakka

*Mie University, Japan* 

**7. References** 

Akihiko Kosugi and Yutaka Mori

675403-9, New York

557, ISSN 0959-440X

Typically, most plant cell wall degrading enzymes are composed of a series of separate modules (modular enzymes). These domains may fold and function in an independent manner and are normally separated by short linker. *P. curdlanolyticus* B-6*,* produces a number of glycosyl hydrolase (GHs) families and CBM families which have different substrates recognition affinity and increase amorphous regions of cellulose by H-bond elimination. Interestingly, modular architecture created by chimeric proteins creation with various tandem CBMs, GHs, and SLH-specific, should make it possible to construct effective lignocellulosic degrading enzymes, strongly binding, targeting enzyme to their substrates and bacterial cell surfaces for enhancing a variety of substrates hydrolysis. The strong carbohydrate-binding property of the cellulose-binding domain and xylan-binding domain, specific degradative activities exhibit important properties of the lignocellulosic material degrading enzymes that can be used in biotechnology.

### **6. Conclusion**

A facultatively anaerobic bacterium *P. curdlanolyticus* strain B-6, isolated from an anaerobic digester fed with pineapple wastes, is unique in that it produces extracellular xylanolyticcellulolytic multienzyme complex capable of efficient degradation of plant biomass materials under aerobic conditions. The production of strain B-6 multienzyme complex under aerobic conditions has several advantages: (i) a simple process, (ii) low price of medium, (iii) high growth rate, (iv) large quantities of extracellular enzymes yields, and (v) safe use with regard to health and environmental aspects. Thus, strain B-6 and its multienzyme complex is a promising tool for an industrial process employing direct hydrolysis for the bioconversion of cellulose as well as hemicellulose in biomass. This review shows that strain B-6 multienzyme complex is a novel enzymatic system known at the biochemical, genetic, and mechanism level. It also stresses that some points still need to be further investigated, mainly (i) the elucidation of scaffolding protein functions, (ii) the characterization of others key enzyme subunits, (iii) the assembly mechanism of the multienzyme complex, (iv) improvement of the efficiency in degradation of biomass of the multienzyme complex, and (v) improvement of the production of the multienzyme complex. The latter will certainly represent a challenge for future research.

### **Author details**

386 Biomass Now – Cultivation and Utilization

degradation (Satitmanwiwat et al., 2012a,b).

degrading enzymes that can be used in biotechnology.

**6. Conclusion** 

Biological treatment and saccharification using microorganisms and their enzymes selectively for degradation of lignocellulosic residues has the advantages of low energy consumption, minimal waste production, and environmental friendliness (Schwarz, 2001). The catalytic components of the multienzyme complex release soluble sugars, simple 5- and 6-carbon, from lignocellulose providing the primary carbon substrates, which can be subsequently converted into fuels by microorganisms. For enzyme saccharification, the close proximity between cellulolytic and xylanolytic enzymes is key to concerted degradation of the substrate, whereby the activities of the different enzymes facilitate the activities of their counterparts by promoting access to appropriated portions of the rigid insoluble substrates, since the release of sugar products was high. The synergistic action of the combination of enzymes by different modes of actions (xylanases and cellulases) and the presence of xylan- or cellulose-binding ability on lignocellulose enhanced soluble sugars released from the plant cell walls. In practicality, the multienzyme complex produced from *P. curdlanolyticus* B-6 allows access to lignocellulosic substrate and produces reducing sugar more than non-complexed enzymes from fungi (*T. viride* and *Aspergillus niger*) when the same cellulase activity (0.1 unit) was applied for degradation of corn hull and rice straw residues (unpublished data). In addition, the multienzyme complex of strain B-6 has been used to improve the extraction of plant food such as making low-cyanide-cassava starch by using multienzyme complex to enhance linamarin released by allowing more contact between linamarase and linamarin (Sornyotha et. al., 2009). Also, extraction of volatile compounds such as sea food-like flavor from seaweed, served for food supplement. Consequently, enzymatic treatment has advantages for the preparation of β-glucan and acidic α-glucan-protein complex from the fruiting body of mushroom, *Pleurotus sajor-caju* because the specificity of the multienzyme complex and gentle conditions allow for the recovery of high purity glucans in their native forms with minimal

Typically, most plant cell wall degrading enzymes are composed of a series of separate modules (modular enzymes). These domains may fold and function in an independent manner and are normally separated by short linker. *P. curdlanolyticus* B-6*,* produces a number of glycosyl hydrolase (GHs) families and CBM families which have different substrates recognition affinity and increase amorphous regions of cellulose by H-bond elimination. Interestingly, modular architecture created by chimeric proteins creation with various tandem CBMs, GHs, and SLH-specific, should make it possible to construct effective lignocellulosic degrading enzymes, strongly binding, targeting enzyme to their substrates and bacterial cell surfaces for enhancing a variety of substrates hydrolysis. The strong carbohydrate-binding property of the cellulose-binding domain and xylan-binding domain, specific degradative activities exhibit important properties of the lignocellulosic material

A facultatively anaerobic bacterium *P. curdlanolyticus* strain B-6, isolated from an anaerobic digester fed with pineapple wastes, is unique in that it produces extracellular xylanolyticKhanok Ratanakhanokchai, Rattiya Waeonukul, Patthra Pason, Chakrit Tachaapaikoon and Khin Lay Kyu *King Mongkut's University of Technology Thonburi, Thailand* 

Kazuo Sakka *Mie University, Japan* 

Akihiko Kosugi and Yutaka Mori *Japan International Research Center for Agricultural Sciences, Japan* 

### **7. References**


Bayer, E.A.; Lamed, R. & Himmel, M.E. (2007). The Potential of Cellulases and Cellulosomes for Cellulosic Waste Management. *Current Opinion in Biotechnology*, Vol.18, No.3, (June, 2007), pp. 237-245, ISSN 0958-1669

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 389

Ding, S.Y.; Bayer, E.A.; Steiner, D.; Shoham, Y. & Lamed, R. (1999). A Novel Cellulosomal Scaffoldin from *Acetivibrio cellulolyticus* that Contains a Family 9 Glycosyl Hydrolase. *Journal of Bacteriology,* Vol.181, No.21, (November, 1999), pp. 6720-6729, ISSN 0021-9193 Ding, S.Y.; Bayer, E.A.; Steiner, D.; Shoham, Y. & Lamed, R. (2000). A Scaffoldin of the *Bacteroides cellulosolvens* Cellulosome that Contains 11 Type II Cohesins. *Journal of* 

Doi, R.H. & Kosugi, A. (2004). Cellulosomes: Plant-Cell-Wall-Degrading Enzyme Complexes. *Nature Reviews Microbiology,* Vol.2, No.7, (July, 2004), pp. 541–551, ISSN

Doi, R.H.; Kosugi, A.; Murashima, K.; Tamaru, Y. & Han, S.O. (2003). Cellulosomes from Mesophilic Bacteria. *Journal of Bacteriology,* Vol.185, No.20, (October, 2003), pp. 5907–

Gal, L.; Pages, S.; Gaudin, C.; Belaich, A.; Reverbel-Leroy, C.; Tardif, C. & Belaich, J.P. (1997). Characterization of the Cellulolytic Complex (Cellulosome) Produced by *Clostridium cellulolyticum*. *Applied and Environmental Microbiology*, Vol.63, No.3, (March, 1997), pp.

Goyal, A.; Ghosh, B. & Eveleig, D. (1991). Characterization of Fungal Cellulases. *Bioresource Technology*, Vol.36, No.1, (Special Issue for Enzymatic Hydrolysis of Cellulose, 1991), pp.

Han, S.O.; Cho, H.Y.; Yukawa, H.; Inui, M. & Doi, R.H. (2004). Regulation of Expression of Cellulosomes and Noncellulosomal (Hemi)Cellulolytic Enzymes in *Clostridium cellulovorans* During Growth on Different Carbon Sources. *Journal of Bacteriology*,

Han, S.O.; Yukawa, H.; Inui, M. & Doi, R.H. (2005). Effect of Carbon Source on the Cellulosomal Subpopulations of *Clostridium cellulovorans*. *Microbiology*, Vol.151, No.5,

Hou, P.; Li, Y.; Wu, B.; Yan, Z.; Yan, B. & Gao, P. (2006). Cellulolytic Complex Exists in Cellulolytic Myxobacterium *Sorangium*. *Enzyme and Microbial Technology*, Vol.38, No.1-2

Innis, M.A. & Gelfand, D.H. (1990). Optimization of PCRs, In: *PCR protocols: A Guide to Methods and Application*, Innis, M.A.; Gelfand, D.H.; Sninsky, J.J. & White, T.J. pp. (3-12),

Jiang, Z.Q.; Deng, W.; Li, L.T.; Ding, C. H.; Kusakabe, I. & Tan, S.S. (2004). A Novel, Ultra-Large Xylanolytic Complex (Xylanosome) Secreted by *Streptomyces olivaceoviridis*.

*Biotechnology Letters,* Vol.26, No.5, (March, 2004), pp. 431-436, ISSN 0141-5492 Kakiuchi, M.; Isui ,A.; Suzuki, K.; Fujino, T.; Fujino, E.; Kimura, T.; Karita, S.; Sakka, K. & Ohmiya, K. (1998). Cloning and DNA Sequencing of the Genes Encoding *Clostridium josui* Scaffolding Protein CipA and Cellulase CelD and Identification of Their Gene

Vol.186, No.13, (July, 2004), pp. 4218-4227, ISSN 0021-9193

(May, 2005), pp. 1491-1497, ISSN 1350-0872

(January, 2006), pp. 273–278, ISSN 0141-0229

Academic Press, ISBN 0-12-372180-6, San Diego

*Bacteriology,* Vol.182, No.17, (September, 2000), pp. 4915-4925, ISSN 0021-9193 Ding, S.Y.; Rincon, M.T.; Lamed, R.; Martin, J.C.; Mccrae, S.I.; Aurilia, V.; Shoham,Y.; Bayer E.A. & Flint, H.J. (2001). Cellulosomal Scaffoldin-Like Proteins from *Ruminococcus flavefaciens*. *Journal of Bacteriology,* Vol.183, No.6, (March, 2001), pp. 1945–1953, ISSN

0021-9193

1740-1526.

5914, ISSN 0021-9193

903–909, ISSN 0099-2240

37-50, ISSN 0960-8524


Ding, S.Y.; Bayer, E.A.; Steiner, D.; Shoham, Y. & Lamed, R. (1999). A Novel Cellulosomal Scaffoldin from *Acetivibrio cellulolyticus* that Contains a Family 9 Glycosyl Hydrolase. *Journal of Bacteriology,* Vol.181, No.21, (November, 1999), pp. 6720-6729, ISSN 0021-9193

388 Biomass Now – Cultivation and Utilization

ISSN 0167-7799

8925

559, ISSN 0021-9193

pp. 242-249, ISSN 008-4166

1066-1073, ISSN 0099-2240

ISSN 1092-2172

1999) pp. 547-554, ISSN 1075-9964

2007), pp. 237-245, ISSN 0958-1669

Bayer, E.A.; Lamed, R. & Himmel, M.E. (2007). The Potential of Cellulases and Cellulosomes for Cellulosic Waste Management. *Current Opinion in Biotechnology*, Vol.18, No.3, (June,

Bayer, E.A.; Morag, E. & Lamed, R. (1994). The Cellulosome—a Treasure-Trove for Biotechnology. *Trends in Biotechnology*, Vol.12, No.9, (September, 1994), pp. 379-386,

Bayer, E.A.; Setter, E. & Lamed, R. (1985). Organization and Distribution of the Cellulosome in *Clostridium thermocellum*. *Journal of Bacteriology*, Vol.163, No.2, (August, 1985), pp. 552-

Berger, E.; Jones, W.A.; Jones, D.T. & Woods, D.R. (1990). Sequencing and Expression of a Cellodextrinase (*ced1*) Gene from *Butyrivibrio fibrisolvens* H17c Cloned in *Escherichia coli*. *Molecular and General Genetic,* Vol.223, No.2, (September, 1990), pp. 310-318, ISSN 0026-

Bhat, S.; Goodenough, P.W.; Owen, E. & Bhat, M.K. (1993). Cellobiose: A True Inducer of Cellulosome in Different Strains of *Clostridium thermocellum*. *FEMS Microbiology Letters*,

Blair, B.G. & Anderson, K.L. (1999a). Regulation of Cellulose Inducible Structures of *Clostridium cellulovorans*. *Canadian Journal of Microbiology,* Vol.45, No.3, (March, 1999),

Blair, B.G. & Anderson, K.L. (1999b). Cellulose-Inducible Ultrastructural Protuberances and Cellulose-affinity Proteins of *Eubacterium cellulosolvens*. *Anaerobe,* Vol.5, No.5, (October,

Borneman, W.S.; Akin, D.E. & Ljungdahl, L.G. (1989). Fermentation Products and Plant Cell Wall Degrading Enzymes Produced by Monocentric and Polycentric Anaerobic Ruminal Fungi. *Applied and Environmental Microbiology,* Vol.55, No.5, (May, 1989), pp.

Chimtong, S.; Tachaapaikoon, C.; Pason, P.; Kyu, K.L.; Kosugi, A.; Mori, Y. & Ratanakhanokchai, K. (2011). Isolation and Characterization of Endocellulase-Free Multienzyme Complex from Newly Isolated *Thermoanaerobacterium thermosaccharolyticum* Strain NOI-1. *Journal of Microbiology and Biotechnology*, Vol.21,

Dalrymple, B.P.; Cybinski, D.H.; Layton I.; McSweeney,C.; Xue, G.P.; Swadling Y.J. & Lowry, J.B. (1997). Three *Neocallimastix patriciarum* Esterases Associated with the Degradation Complex Polysaccharides are Members of a New Family of Hydrolases.

Dashtban, M.; Schraft, H. & Qin, W. (2009). Fungal Bioconversion of Lignocellulosic Residues; Opportunities & Perspectives. *International Journal of Biological Sciences*, Vol.5,

Demain, A.L.; Newcomb, M. & Wu, J.H.D. (2005). Cellulase, *Clostridium*, and Ethanol. *Microbiology and Molecular Biology Reviews*, Vol.69, No.1, (March, 2005), pp. 124–154,

*Microbiology,* Vol.143, No.8, (August, 1997) pp. 2605-2614, ISSN 1350-0872

Vol.111, No.1, (July, 1993), pp. 73-78, ISSN 0378-1097

No.3, (March, 2011), pp. 284-292, ISSN 1017-7825

No.6, (September, 2009), pp. 578-595, ISSN 1449-1907


Products as Major Components of the Cellulosome. *Journal of Bacteriology,* Vol.180, No.16, (August, 1998), pp. 4303–4308, ISSN 0021-9193

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 391

Nochur, S.V.; Roberts, M.F. & Demain, A.L. (1993). True Cellulase Production by *Clostridium thermocellum* Grown on Different Carbon Sources. *Biotechnology Letter*, Vol.15, No.6,

Ohara. H.; Karita, S.; Kimura, T.; Sakka, K. & Ohmiya, K. (2000). Characterization of the Cellulolytic Complex (Cellulosome) from *Ruminococcus albus*. *Bioscience, Biotechnology,* 

Ohtsuki, T.; Suyanto, S.; Yazaki, S.U. & Mimura, A. (2005). Production of Large Multienzyme Complex by Aerobic Thermophilic Fungus *Chaetomium* sp. nov. MS-017 Grown on Palm Oil Mill Fibre. *Letters in Applied Microbiology,* Vol.40, No.2, (February,

Pagès, S.; Bélaïch, A.; Bélaïch, J.P.; Morag, E.; Lamed, R.; Shoham, Y. & Bayer E.A. (1997). Species-Specificity of the Cohesin–Dockerin Interaction Between *Clostridium thermocellum* and *Clostridium cellulolyticum*: Prediction of Specificity Determinants of the Dockerin Domain. *Proteins: Structure, Function, and Bioinformatics*, Vol.29, No.4,

Pason, P.; Chon, G.H.; Ratanakhanokchai, K.; Kyu K.L.; Jhee, O.H.; Kang, J.; Kim, W.H.; Choi, K.M.; Park, G.S.; Lee, J.S., Park, H.; Rho, M.S. & Lee, Y.S. (2006a). Selection of Multienzyme Complex Producing Bacteria Under Aerobic Cultivation. *Journal of Microbiology and Biotechnology,* Vol.16, No.8, (August, 2006), pp. 1269-1275, ISSN 1017-

Pason, P.; Kosugi, A.; Waeonukul, R.; Tachaapaikoon, C.; Ratanakhanokchai, K.; Arai, T.; Murata, Y.; Nakajima, J. & Mori, Y. (2010). Purification and Characterization of a Multienzyme Complex Produced by *Paenibacillus curdlanolyticus* B-6. *Applied Microbiology and Biotechnology*, Vol.85, No.3, (January, 2010), pp. 573-580, ISSN 0175-7598 Pason, P; Kyu, K.L. & Ratanakhanokchai, K. (2006b). *Paenibacillus curdlanolyticus* Strain B-6 Xylanolytic-Cellulolytic Enzyme System That Degrades Insoluble Polysaccharides. *Applied and Environmental Microbiology*, Vol.72, No.4, (April, 2006), pp. 2483-2490, ISSN

Pauly, M. & Keegstra, K. (2008) Cell Wall Carbohydrates and Their Modification as a Resource for Biofuels. *The Plant Journal*, Vol.54, No.4, (May, 2008), pp. 559–568, ISSN

Pérez, J.; Muñoz-Dorado, J.; De-la-Rubia, T. & Martínez, J. (2002). Biodegradation and Biological Treatments of Cellulose, Hemicellulose and Lignin: an Overview.

Ponpium, P.; Ratanakhanokchai, K. & Kyu, K.L. (2000). Isolation and Properties of a Cellulosome-Type Multienzyme Complex of the Thermophilic *Bacteroides* sp. Strain P-1.

*International Microbiology*, Vol.5, No.2, (June, 2002), pp. 53-63, ISSN 1139-6709 Phitsuwan, P.; Tachaapaikoon, C.; Kosugi, A.; Mori, Y.; Kyu, K.L. & Ratanakhanokchai, K. (2010). A Cellulolytic and Xylanolytic Enzyme Complex from an Alkalothermoanaerobacterium, *Tepidimicrobium xylanilyticum* BT14. *Journal of Microbiology and Biotechnology*, Vol.20, No.5, (May, 2010), pp. 893-903 ISSN 1017-7825 Pohlschröder, M.D.; Leschine, S.B., & Canale-Parola, E. (1994). Multicomplex Cellulase– Xylanase System of *Clostridium papyrosolvens* C7. *Journal of Bacteriology,* Vol.176, No.1,

*and Biochemistry*, Vol.64, No.2, (February, 2000), pp. 254–260, ISSN 0916-8451

(June, 1993), pp. 641-646, ISSN 0141-5492

2005), pp. 111-116, ISSN 0266-8254

7825

0099-2240

0960-7412

(December, 1997), pp. 517-527, ISSN 0887-3585

(January, 1994), pp. 70-76, ISSN 0021-9193


Nochur, S.V.; Roberts, M.F. & Demain, A.L. (1993). True Cellulase Production by *Clostridium thermocellum* Grown on Different Carbon Sources. *Biotechnology Letter*, Vol.15, No.6, (June, 1993), pp. 641-646, ISSN 0141-5492

390 Biomass Now – Cultivation and Utilization

No.16, (August, 1998), pp. 4303–4308, ISSN 0021-9193

No.1, (July, 1993), pp. 83-94, ISSN 0885-4513

pp. 3358-3393, ISSN 1433-7851

pp. 3792-3800, ISSN 0021-9193

(November, 1983), pp. 828-836, ISSN 0021-9193

Vol.27, No.1-2, (July, 2000), pp. 89-94, ISSN 0141-0229

(December, 1987), pp. 2785-2792, ISSN 0099-2240

9193

3, San Diego

Products as Major Components of the Cellulosome. *Journal of Bacteriology,* Vol.180,

Kim, C.H. & Kim, D.S. (1993). Extracellular Cellulolytic Enzymes of *Bacillus circulans* Are Present as Two Multi-Protein Complexes. *Applied Biochemistry and Biotechnology*, Vol.42,

Klemm, D.; Heublein, B. & Fink, H.P. (2005). Cellulose: Fascinating Biopolymer and Sustainable Raw Material. *Angewandte Chemie International*, Vol.44, No.22, (May, 2005),

Kosugi, A.; Murashima K. & Doi, R.H. (2001). Characterization of Xylanolytic Enzymes in *Clostridium cellulovorans*: Expression of Xylanase Activity Dependent on Growth Substrates. *Journal of Bacteriology*, Vol.183, No.11, (June, 2001), pp. 7037-7043, ISSN 0021-

Lamed, R. & Bayer, E.A. (1988). The Cellulosome of *Clostridium thermocellum*, In: *Advances in Applied Microbiology*, Vol.33, Laskin, A.I., pp. (1–46), Academic Press, ISBN 0-12-002633-

Lamed, R.; Naimark. J.; Morgenstern, E. & Bayer, E.A. (1987). Specialized Cell Surface Structure in Cellulolytic Bacteria. *Journal of Bacteriology*, Vol.169, No.8, (August, 1987),

Lamed, R.; Setter, E. & Bayer, E.A. (1983). Characterization of a Cellulose-Binding Cellulase-Containing Complex in *Clostridium thermocellum*. *Journal of Bacteriology*, Vol.156, No.2,

Li, X.L.; Chen, H. & Ljungdahl, L.G. (1997). Two Cellulases, CelA and CelC, from the Polycentric Anaerobic Fungus *Orpinomyces* Strain PC-2 Contain N-terminal Docking Domains for a Cellulase-Hemicellulase Complex. *Applied and Environmental* 

Li, K.; Azadi, P.; Collins, R.; Tolan, J.; Kim, J.S. & Eriksson, K.E.L. (2000). Relationship Between Activities of Xylanases and Xylan Structures. *Enzyme and Microbial Technology*,

Lynd, L.R.; Weimer, P.J.; van Zyl, W.H. & Pretorius, I.S. (2002). Microbial Cellulose Utilization: Fundamentals and Biotechnology. *Microbiology and Molecular Biology* 

Mayer, F.; Coughlan, M. P.; Mori, Y., & Ljungdahl, L.G. (1987). Macromolecular Organization of the Cellulolytic Enzyme Complex of *Clostridium thermocellum* as Revealed by Electron Microscopy. *Applied and Environmental Microbiology*, Vol.53, No.12,

Mohand-Oussaid, O.; Payot, S.; Guedon, E.; Gelhaye, E.; Youyou, A. & Petitdemange, H. (1999). The Extracellular Xylan Degradative System in *Clostridium cellulolitycum* Cultivated on Xylan: Evidence for Cell-Free Cellulosome Production. *Journal of* 

Nochur, S.V.; Demain, A.L. & Roberts, M.F. (1992). Carbohydrate Utilization by *Clostridium thermocellum*: Importance of Internal pH in Regulating Growth. *Enzyme and Microbial* 

*Microbiology*, Vol.63, No.12, (December, 1997), pp. 4721-4728, ISSN 0099-2240

*Reviews,* Vol.66, No.3, (September, 2002), pp. 506-577, ISSN 1092-2172

*Bacteriology*, Vol.181, No.13, (July, 1999), pp. 4035-4040, ISSN 0021-9193

*Technology*, Vol.14, No.5, (May, 1992), pp. 338-349, ISSN 0141-0229


*Enzyme and Microbial Technology*, Vol.26, No.5-6, (March, 2000), pp. 459-465, ISSN 0141- 0229

*Paenibacillus curdlanolyticus* Strain B-6 Multienzyme Complex: A Novel System for Biomass Utilization 393

Shoseyov, O. & Doi, R.H. (1990). Essential 170-kDa Subunit for Degradation of Crystalline Cellulose by *Clostridium cellulovorans* Cellulase. *Proceedings of the National Academy of Sciences of the United States of America,* Vol.87, No.6, (March, 1990), pp. 2192-2195, ISSN

Shoseyov, O.; Shani, Z. & Levy, I. (2006). Carbohydrate Binding Modules: Biochemical Properties and Novel Applications. *Microbiology and Molecular Biology Reviews*, Vol.70,

Sleat, R.; Mah, R. A. & Robinson, R. (1984). Isolation and Characterization of an Anaerobic, Cellulolytic Bacterium, *Clostridium cellulovorans* sp. nov. *Applied and Environmental* 

Sneath, P.H.A. (1986). Endospore-Forming Gram-Positive Rods and Cocci, In: *Bergey's Manual of Systematic Bacteriology*, Vol. 2, Sneath, P.H.A.; Mair, N.S.; Sharpe, M.E. & Holt,

Sornyotha, S.; Kyu K.L. & Ratanakhanokchai K. (2010). An Efficient Treatment for Detoxification Process of Cassava Starch by Plant Cell Wall-Degrading Enzymes. *Journal of Bioscience and Bioengineering*. Vol.109, No.1, (January, 2010), pp. 9-14, ISSN 1389-1723 Sudo, M.; Sakka, M; Kimura,T.; Ratanakhanokchai, K. & Sakka, K. (2010). Characterization of *Paenibacillus curdlanolyticus* Intracellular Xylanase Xyn10B Encoded by the *xyn10B*  gene. *Bioscience, Biotechnology, and Biochemistry*, Vol.74, No.11, (November, 2010), pp.

Teunissen, M.J.; Op den Camp, H.J.M.; Orpin, C.G.; Huis in't Veld, J.H.J. & Vogels, G.D. (1991). Comparison of Growth Characteristics of Anaerobic Fungi Isolated from Ruminant and Non-Ruminant Herbivores During Cultivation in a Defined Medium. *Journal of General Microbiol*ogy, Vol.137, No.6, (June, 1991), pp. 1401-1408, ISSN 0022-

Thomson, J.A. (1993). Molecular Biology of Xylan Degradation. *FEMS Microbiology Letters*,

Tomme, P.; Warren, R.A.J. & Gilkes; N.R. (1995). Cellulose Hydrolysis by Bacteria and Fungi, In: *Advances in Microbial Physiology,* Vol.37, Poole, R.K., pp. (1-81), Academic

van Dyk J.S.; Sakka, M.; Sakka, K. & Pletschke, B.I. (2009). The Cellulolytic and Hemi-Cellulolytic System of *Bacillus licheniformis* SVD1 and the Evidence for Production of a Large Multi-Enzyme Complex. *Enzyme and Microbial Technology,* Vol.45, No.5,

Waeonukul, R.; Kyu, K.L.; Sakka, K. & Ratanakhanokchai, K. (2008). Effect of Carbon Sources on the Induction of Xylanolytic-Cellulolytic Multienzyme Complexes in *Paenibacillus curdlanolyticus* Strain B-6. *Bioscience, Biotechnology, and Biochemistry*, Vol.72,

Waeonukul, R.; Pason, P.; Kyu, K.L.; Sakka, K.; Kosugi, A.; Mori, Y, & Ratanakhanokchai, K. (2009a). Cloning, Sequencing, and Expression of the Gene Encoding a Multidomain Endo-β-1,4-Xylanase from *Paenibacillus curdlanolyticus* B-6, and Characterization of the Recombinant Enzyme. *Journal of Microbiology and Biotechnology*, Vol.19, No.3, (March,

Vol.10, No.1-2, (January, 1993), pp.65–82, ISSN 0378-1097

*Microbiology*, Vol.48, No.1, (July, 1984), pp. 88-93, ISSN 0099-2240

J. G., pp. (1104– 1207), William & Wilkins, ISBN 0-68307-893-3, Baltimore

No.2, (June, 2006), pp. 283-295, ISSN 1092-2172

0027-8424

2358-2360, ISSN 0916-8451

Press, ISBN 0-12-027737-9, London

2009), pp. 277-285, ISSN 1017-7825

(November, 2009), pp. 372-378, ISSN 0141-0229

No.2, (February, 2008), pp. 321-328, ISSN 0916-8451

1287


Shoseyov, O. & Doi, R.H. (1990). Essential 170-kDa Subunit for Degradation of Crystalline Cellulose by *Clostridium cellulovorans* Cellulase. *Proceedings of the National Academy of Sciences of the United States of America,* Vol.87, No.6, (March, 1990), pp. 2192-2195, ISSN 0027-8424

392 Biomass Now – Cultivation and Utilization

126, ISSN 0378-1119

5423-5430, ISSN 0021-8561

(June, 2010), pp. 263-289, ISSN 1543-5008

(November, 2002), pp. 15-22, ISSN 0378-1097

No.12, (June, 2011), pp. 4260-4263, ISSN 0099-2240

0229

6838

8451

ISSN 0175-7598

*Enzyme and Microbial Technology*, Vol.26, No.5-6, (March, 2000), pp. 459-465, ISSN 0141-

Qiu, X.; Selinger, B.; Yanke, L.J. & Cheng, K.J. (2000). Isolation and Analysis of Two Cellulase cDNAs from *Orpinomyces joyonii*. *Gene,* Vol.245, No.1, (March, 2000), pp. 119-

Rabinovich, M.L.; Melnik, M.S. & Bolobova, A.V. (2002). Microbial Cellulases (Review). *Applied Biochemistry and Microbiology*, Vol.38, No.4, (July, 2002), pp. 305-322, ISSN 0003-

Sabathé, F.; Bélaïch, A. & Soucaille, P. (2002) Characterization of the Cellulolytic Complex (Cellulosome) of *Clostridium acetobutylicum. FEMS Microbiology Letters,* Vol.217, No.1

Saha, B.C.; (2003). Hemicellulose Bioconversion. *Journal of Industrial Microbiology and* 

Sakka, M.; Higashi, Y.; Kimura, T.; Ratanakhanokchai, K. & Sakka K. (2011). Characterization of *Paenibacillus curdlanolyticus* B-6 Xyn10D, a Xylanase That Contains a Family 3 Carbohydrate-Binding Module. *Applied and Environmental Microbiology*, Vol.77,

Sánchez, C. (2009). Lignocellulosic Residues: Biodegradation and Bioconversion by Fungi. *Biotechnology Advances*, Vol.27, No.2, (March-April, 2009), pp. 185–194, ISSN 0734-9750 Satitmanwiwat, S.; Ratanakhanokchai, K.; Laohakunjit, N.; Chao, L.K.; Chen, S-T.; Pason, P.; Tachaapaikoon, C. & Kyu, K.L. (2012a) Improved Purity and Immunostimulatory Activity of β-(1→ → 3)(1 6)-Glucan from *Pleurotus sajor-caju* Using Cell Wall-Degrading Enzymes. *Journal of Agricultural and Food Chemistry*, Vol.60, Issue 21, (May, 2012), pp.

Satitmanwiwat, S.; Ratanakhanokchai, K.; Laohakunjit, N.; Pason, P.; Tachaapaikoon, C. & Kyu, K.L. (2012b) Purification and Partial Characterization of an Acidic α-Glucan-Protein Complex from the Fruiting Body of *Pleurotus sajor-caju* and Its Effect on Macrophage Activation. *Bioscience, Biotechnology, and Biochemitry*, in press, ISSN 0916-

Scheller, H.V. & Ulvskov, P. (2010). Hemicelluloses. *Annual Review of Plant Biology*, Vol.61,

Schwarz, W.H. (2001). The Cellulosome and Cellulose Degradation by Anaerobic Bacteria. *Applied Microbiology and Biotechnology*, Vol.56, No.5-6, (September, 2001), pp. 634-649,

Shida, O.; Takagi, H.; Kadowaki, K.; Nakamura L.K. & Komagata, K. (1997). Transfer of *Bacillus alginolyticus*, *Bacillus chondroitinus*, *Bacillus curdlanolyticus*, *Bacillus glucanolyticus*, *Bacillus kobensis*, and *Bacillus thiaminolyticus* to the Genus *Paenibacillus* and Emended Description of the Genus *Paenibacillus*. *International Journal of Systematic and Evolutionary* 

Shoham, Y.; Lamed, R. & Bayer, E.A. (1999). The Cellulosome Concept as an Efficient Microbial Strategy for the Degradation of Insoluble Polysaccharides. *Trends in* 

*Microbiology,* Vol.47, No.2, (April, 1997), pp. 289-298, ISSN 1466-5026

*Microbiology,* Vol.7, No.7, (July, 1999), pp. 275-281, ISSN 0966-842X

*Biotechnology,* Vol*.*30, No.5, (May, 2003), pp. 279-291, ISSN 1367-5435


Waeonukul, R.; Kyu, K.L.; Sakka, K. & Ratanakhanokchai, K. (2009b). Isolation and Characterization of a Multienzyme Complex (Cellulosome) of the *Paenibacillus curdlanolyticus* B-6 Grown on Avicel Under Aerobic Conditions. *Journal of Bioscience and Bioengineering*, Vol.107, No.6, (June, 2009), pp. 610-614, ISSN 1389-1723

**Chapter 17** 

© 2013 Rulík et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

© 2013 Rulík et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

**Methanogenic System of a Small Lowland** 

Martin Rulík, Adam Bednařík, Václav Mach, Lenka Brablcová, Iva Buriánková,

Methane (CH4) is an atmospheric trace gas present at concentration of about 1.8 ppmv, that represents about 15% of the anthropogenic greenhouse effect (Forster et al. 2007). The atmospheric CH4 concentration has increased steadily since the beginning of the industrial revolution ( 0.7 ppmv) and is stabilized at 1.8 ppmv from 1999 to 2005 (Forster et al. 2007). An unexpected increase in the atmospheric growth of CH4 during the year 2007 has been recently reported (Rigby et al. 2008), indicating that the sources and sinks of atmospheric CH4 are dynamics, evolving, and not well understood. Freshwater sediments, including wetlands, rice paddies and lakes, are thought to contribute 40 to 50 % of the

The river hyporheic zone, volume of saturated sediment beneath and beside streams containing some proportion of water from surface channel, plays a very important role in the processes of self-purification because the river bed sediments are metabolically active and are responsible for retention, storage and mineralization of organic matter transported by the surface water (Hendricks 1993; Jones & Holmes 1996, Baker et al. 1999, Storey et al. 1999, Fischer et al. 2005). The seemingly well-oxygenated hyporheic zone contains anoxic and hypoxic pockets ("anaerobic microzones") associated with irregularities in sediment surfaces, small pore spaces or local deposits of organic matter, creating a 'mosaic' structure of various environments, where different microbial populations can live and different microbially mediated processes can occur simultaneously (Baker et al. 1999, Morrice et al. 2000, Fischer et al. 2005). Moreover, hyporheic-surface exchange and subsurface hydrologic flow patterns result in solute gradients that are important in microbial metabolism. Oxidation processes may occur more readily where oxygen is replenished by surface water infiltration, while reduction processes may prevail where surface-water exchange of oxygen

annual atmospheric methane flux (Cicerone & Oremland 1988; Conrad 2009).

**Stream Sitka, Czech Republic** 

Pavlína Badurová and Kristýna Gratzová

http://dx.doi.org/10.5772/52718

**1. Introduction** 

Additional information is available at the end of the chapter

Watthanalamloet, A.; Tachaapaikoon, C.; Lee, Y. S.; Kosugi, A.; Mori, Y.; Tanasupawat, S.; Kyu, K.L. & Ratanakhanokchai, K. (2012). *Amorocellulobacter alkalithermophilum* gen. nov., sp. nov. an Anaerobic Alkalithermophile, Cellulolytic-Xylanolytic Bacterium Isolated from Soil in a Brackish Area of a Coconut Garden. *International Journal of Systematic and Evolutionary Microbiology,* doi: 10.1099/ijs.0.027854-0

**Chapter 17** 

### **Methanogenic System of a Small Lowland Stream Sitka, Czech Republic**

Martin Rulík, Adam Bednařík, Václav Mach, Lenka Brablcová, Iva Buriánková, Pavlína Badurová and Kristýna Gratzová

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/52718

### **1. Introduction**

394 Biomass Now – Cultivation and Utilization

Waeonukul, R.; Kyu, K.L.; Sakka, K. & Ratanakhanokchai, K. (2009b). Isolation and Characterization of a Multienzyme Complex (Cellulosome) of the *Paenibacillus curdlanolyticus* B-6 Grown on Avicel Under Aerobic Conditions. *Journal of Bioscience and* 

Watthanalamloet, A.; Tachaapaikoon, C.; Lee, Y. S.; Kosugi, A.; Mori, Y.; Tanasupawat, S.; Kyu, K.L. & Ratanakhanokchai, K. (2012). *Amorocellulobacter alkalithermophilum* gen. nov., sp. nov. an Anaerobic Alkalithermophile, Cellulolytic-Xylanolytic Bacterium Isolated from Soil in a Brackish Area of a Coconut Garden. *International Journal of* 

*Bioengineering*, Vol.107, No.6, (June, 2009), pp. 610-614, ISSN 1389-1723

*Systematic and Evolutionary Microbiology,* doi: 10.1099/ijs.0.027854-0

Methane (CH4) is an atmospheric trace gas present at concentration of about 1.8 ppmv, that represents about 15% of the anthropogenic greenhouse effect (Forster et al. 2007). The atmospheric CH4 concentration has increased steadily since the beginning of the industrial revolution ( 0.7 ppmv) and is stabilized at 1.8 ppmv from 1999 to 2005 (Forster et al. 2007). An unexpected increase in the atmospheric growth of CH4 during the year 2007 has been recently reported (Rigby et al. 2008), indicating that the sources and sinks of atmospheric CH4 are dynamics, evolving, and not well understood. Freshwater sediments, including wetlands, rice paddies and lakes, are thought to contribute 40 to 50 % of the annual atmospheric methane flux (Cicerone & Oremland 1988; Conrad 2009).

The river hyporheic zone, volume of saturated sediment beneath and beside streams containing some proportion of water from surface channel, plays a very important role in the processes of self-purification because the river bed sediments are metabolically active and are responsible for retention, storage and mineralization of organic matter transported by the surface water (Hendricks 1993; Jones & Holmes 1996, Baker et al. 1999, Storey et al. 1999, Fischer et al. 2005). The seemingly well-oxygenated hyporheic zone contains anoxic and hypoxic pockets ("anaerobic microzones") associated with irregularities in sediment surfaces, small pore spaces or local deposits of organic matter, creating a 'mosaic' structure of various environments, where different microbial populations can live and different microbially mediated processes can occur simultaneously (Baker et al. 1999, Morrice et al. 2000, Fischer et al. 2005). Moreover, hyporheic-surface exchange and subsurface hydrologic flow patterns result in solute gradients that are important in microbial metabolism. Oxidation processes may occur more readily where oxygen is replenished by surface water infiltration, while reduction processes may prevail where surface-water exchange of oxygen

© 2013 Rulík et al., licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Rulík et al., licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

is less, and the reducing potential of the environment is greater (Hendricks 1993). As water moves through the hyporheic zone, decomposition of the organic matter consumes oxygen, creating oxygen gradients along the flow path. Thus,compared to marine or lake surface sediments, where numerous studies on O2 profiles have showed that O2 concentrations become zero within less than 3 mm from the surface, the hyporheic sediment might be welloxygenated habitats even up to the depth of 80 cm (e.g. Bretschko 1981, Holmes et al. 1994) . The extent of the oxygen gradient is determined by the interplay between flow path lengh, water velocity, the ratio of surface to ground water, and the amount and quality of organic matter. Organic matter decomposition in sediments is an important process in global and local carbon budgets as it ultimately recycles complex organic compounds from terrestrial and aquatic environments to carbon dioxide and methane. Methane is a major component in the carbon cycle of anaerobic aquatic systems, particularly those with low sulphate concentrations. Since a relatively high production of methane has been measured in river sediments (e.g. Schindler & Krabbenhoft 1998, Hlaváčová et al. 2005, Sanders et al. 2007, Wilcock & Sorrell 2008, Sanz et al. 2011), we proposed that river sediments may act as a considerable source of this greenhouse gas which is important in global warming (Hlaváčová et al. 2006).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 397

Methane oxidation can occur in both aerobic and anaerobic environments; however, these are completely different processes involving different groups of prokaryotes. Aerobic methane oxidation is carried out by aerobic methane oxidizing bacteria (methanotrophs, MOB), while anaerobic methane oxidizers, discovered recently, thrive under anaerobic conditions and use sulphate or nitrate as electron donors for methane oxidation (e.g. Strous & Jetten 2004). MOB are a physiologically specialized group of methylotrophic bacteria capable of utilizing methane as a sole source of carbon and energy, and they have been recognized as major players in local and global elemental cycling in aerobic environments (Hanson & Hanson 1996, Murrell et al. 1998, Costelo & Lidstrom 1999, Costelo et al. 2002, McDonald et al. 2008). Aerobic MOB have been detected in a variety of environments, and in some they represent significant fractions of total microbial communities (e.g. Henckel et al. 1999; Carini et al, 2005, Trotsenko & Khmelenina 2005, Kalyuzhnaya et al. 2006). However, the data on the diversity and activity of methanotrophic communities from the river ecosystems are yet fragmentary. Methanotrophs play an important role in the oxidation of methane in the natural environment, oxidizing methane biologically produced in anaerobic environments by the methanogenic archaea and thereby reducing the amount

The present investigation is a part of a long-term study focused on organic carbon and methane dynamics and microbial communities in hyporheic zone of a Sitka, small lowland stream in Czech Republic. The overall purpose of this research was to characterize spatial distribution of both methanogens and methanotrophs within hyporheic sediments and elucidate the differences in methane pathways and methane production/consumption as well as methane fluxes and atmospheric emissions at different sites along a longitudinal

The sampling sites are located on the Sitka stream, Czech Republic (Fig. 1). The Sitka is an undisturbed, third-order, 35 km long lowland stream originating in the Hrubý Jeseník mountains at 650 m above sea level. The catchment area is 118.81km2, geology being composed mainly of Plio-Pleistocene clastic sediments of lake origin covered by quaternary sediments. The mean annual precipitation of the downstream part of the catchment area varies from 500 to 600 mm. Mean annual discharge is 0.81 m3.s-1. The Sitka stream flows in its upper reach till Šternberk through a forested area with a low intensity of anthropogenic effects, while the lower course of the stream naturally meanders through an intensively managed agricultural landscape. Except for short stretches, the Sitka stream is unregulated with well-established riparian vegetation. River bed sediments are composed of gravels in the upper parts of the stream (median grain size 13 mm) while the lower part, several kilometres away from the confluence, is characterised by finer sediment with a median grain size of 2.8 mm. The Sitka stream confluences with the Oskava stream about 5 km north of Olomouc. More detailed characteristics of the geology, gravel bar, longitudinal

of methane released into the atmosphere.

profile of the stream.

**2.1. Study site** 

**2. Material and methods** 

Breakdown of organic matter and gas production are both results of well functioned river self-purification. This degrading capacity, however, requires intensive contact of the water with biologically active surfaces. Flow over various morphological features ranging in size from ripples and dunes to meanders and pool-riffle sequences controls such surfacesubsurface fluxes. Highly permeable streambeds create opportunities for subsurface retention and long-term storage, and exchange with the surface water is frequent. Thus, study of the methane production within hyporheic zone and its subsequent emission to the atmosphere can be considered as a measure of mineralization of organic matter in the freshwater ecosystem and might be used in evaluation of both the health and environmental quality of the rivers studied.

Methane (CH4) is mostly produced by methanogenic archaea (Garcia et al. 2000, Chaban et al. 2006) as a final product of anaerobic respiration and fermentation, but there is also aerobic methane formation (e.g. Karl et al. 2008). Methanogenic archaea are ubiquitous in anoxic environments and require an extremely low redox potential to grow. They can be found both in moderate habitats such as rice paddies (Grosskopf et al. 1998a,b), lakes (Jürgens et al. 2000, Keough et al. 2003) and lake sediments (Chan et al. 2005), as well as in the gastrointestinal tract of animals (Lin et al. 1997) and in extreme habitats such as hydrothermal vents (Jeanthon et al. 1999), hypersaline habitats (Mathrani & Boone 1995) and permafrost soils (Kobabe et al. 2004, Ganzert et al. 2006). Rates of methane production and consumption in sediments are controlled by the relative availability of substrates for methanogenesis (especially acetate or hydrogen and carbon dioxide). The most important immediate precursors of methanogenesis are acetate and H2/CO2. The acetotrophic methanogens convert acetic acid to CH4 and CO2 while the hydrogenotrophic methanogens convert CO2 with H2 to CH4 (Conrad 2007).

Methane oxidation can occur in both aerobic and anaerobic environments; however, these are completely different processes involving different groups of prokaryotes. Aerobic methane oxidation is carried out by aerobic methane oxidizing bacteria (methanotrophs, MOB), while anaerobic methane oxidizers, discovered recently, thrive under anaerobic conditions and use sulphate or nitrate as electron donors for methane oxidation (e.g. Strous & Jetten 2004). MOB are a physiologically specialized group of methylotrophic bacteria capable of utilizing methane as a sole source of carbon and energy, and they have been recognized as major players in local and global elemental cycling in aerobic environments (Hanson & Hanson 1996, Murrell et al. 1998, Costelo & Lidstrom 1999, Costelo et al. 2002, McDonald et al. 2008). Aerobic MOB have been detected in a variety of environments, and in some they represent significant fractions of total microbial communities (e.g. Henckel et al. 1999; Carini et al, 2005, Trotsenko & Khmelenina 2005, Kalyuzhnaya et al. 2006). However, the data on the diversity and activity of methanotrophic communities from the river ecosystems are yet fragmentary. Methanotrophs play an important role in the oxidation of methane in the natural environment, oxidizing methane biologically produced in anaerobic environments by the methanogenic archaea and thereby reducing the amount of methane released into the atmosphere.

The present investigation is a part of a long-term study focused on organic carbon and methane dynamics and microbial communities in hyporheic zone of a Sitka, small lowland stream in Czech Republic. The overall purpose of this research was to characterize spatial distribution of both methanogens and methanotrophs within hyporheic sediments and elucidate the differences in methane pathways and methane production/consumption as well as methane fluxes and atmospheric emissions at different sites along a longitudinal profile of the stream.

### **2. Material and methods**

### **2.1. Study site**

396 Biomass Now – Cultivation and Utilization

(Hlaváčová et al. 2006).

quality of the rivers studied.

convert CO2 with H2 to CH4 (Conrad 2007).

is less, and the reducing potential of the environment is greater (Hendricks 1993). As water moves through the hyporheic zone, decomposition of the organic matter consumes oxygen, creating oxygen gradients along the flow path. Thus,compared to marine or lake surface sediments, where numerous studies on O2 profiles have showed that O2 concentrations become zero within less than 3 mm from the surface, the hyporheic sediment might be welloxygenated habitats even up to the depth of 80 cm (e.g. Bretschko 1981, Holmes et al. 1994) . The extent of the oxygen gradient is determined by the interplay between flow path lengh, water velocity, the ratio of surface to ground water, and the amount and quality of organic matter. Organic matter decomposition in sediments is an important process in global and local carbon budgets as it ultimately recycles complex organic compounds from terrestrial and aquatic environments to carbon dioxide and methane. Methane is a major component in the carbon cycle of anaerobic aquatic systems, particularly those with low sulphate concentrations. Since a relatively high production of methane has been measured in river sediments (e.g. Schindler & Krabbenhoft 1998, Hlaváčová et al. 2005, Sanders et al. 2007, Wilcock & Sorrell 2008, Sanz et al. 2011), we proposed that river sediments may act as a considerable source of this greenhouse gas which is important in global warming

Breakdown of organic matter and gas production are both results of well functioned river self-purification. This degrading capacity, however, requires intensive contact of the water with biologically active surfaces. Flow over various morphological features ranging in size from ripples and dunes to meanders and pool-riffle sequences controls such surfacesubsurface fluxes. Highly permeable streambeds create opportunities for subsurface retention and long-term storage, and exchange with the surface water is frequent. Thus, study of the methane production within hyporheic zone and its subsequent emission to the atmosphere can be considered as a measure of mineralization of organic matter in the freshwater ecosystem and might be used in evaluation of both the health and environmental

Methane (CH4) is mostly produced by methanogenic archaea (Garcia et al. 2000, Chaban et al. 2006) as a final product of anaerobic respiration and fermentation, but there is also aerobic methane formation (e.g. Karl et al. 2008). Methanogenic archaea are ubiquitous in anoxic environments and require an extremely low redox potential to grow. They can be found both in moderate habitats such as rice paddies (Grosskopf et al. 1998a,b), lakes (Jürgens et al. 2000, Keough et al. 2003) and lake sediments (Chan et al. 2005), as well as in the gastrointestinal tract of animals (Lin et al. 1997) and in extreme habitats such as hydrothermal vents (Jeanthon et al. 1999), hypersaline habitats (Mathrani & Boone 1995) and permafrost soils (Kobabe et al. 2004, Ganzert et al. 2006). Rates of methane production and consumption in sediments are controlled by the relative availability of substrates for methanogenesis (especially acetate or hydrogen and carbon dioxide). The most important immediate precursors of methanogenesis are acetate and H2/CO2. The acetotrophic methanogens convert acetic acid to CH4 and CO2 while the hydrogenotrophic methanogens

The sampling sites are located on the Sitka stream, Czech Republic (Fig. 1). The Sitka is an undisturbed, third-order, 35 km long lowland stream originating in the Hrubý Jeseník mountains at 650 m above sea level. The catchment area is 118.81km2, geology being composed mainly of Plio-Pleistocene clastic sediments of lake origin covered by quaternary sediments. The mean annual precipitation of the downstream part of the catchment area varies from 500 to 600 mm. Mean annual discharge is 0.81 m3.s-1. The Sitka stream flows in its upper reach till Šternberk through a forested area with a low intensity of anthropogenic effects, while the lower course of the stream naturally meanders through an intensively managed agricultural landscape. Except for short stretches, the Sitka stream is unregulated with well-established riparian vegetation. River bed sediments are composed of gravels in the upper parts of the stream (median grain size 13 mm) while the lower part, several kilometres away from the confluence, is characterised by finer sediment with a median grain size of 2.8 mm. The Sitka stream confluences with the Oskava stream about 5 km north of Olomouc. More detailed characteristics of the geology, gravel bar, longitudinal

physicochemical (e.g. temperature, pH, redox, conductivity, O2, CH4, NO3, SO4) patterns in the sediments and a schematic view of the site with sampling point positions have been published previously (Rulík et al. 2000, Rulík & Spáčil 2004). Earlier measurements of a relatively high production of methane, as well as potential methanogenesis, confirmed the suitability of the field sites for the study of methane cycling (Rulík et al. 2000, Hlaváčová et al. 2005, 2006).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 399

clay

gravelsand

**Figure 2.** Graphic depiction of the thalweg of the Sitka stream with sampling localities. The main source of pollution is an effluent from Šternberk city sewage water plant, located just in the middle between

**Variable/ Locality I. II. III. IV. V.**  elevation above sea-level [m] 535 330 240 225 215 distance from the spring [km] 6,9 18,2 25,6 30,9 34,9 channel width [cm] 523 793 672 444 523 average flow velocity [m.s-1] 0,18 0,21 0,46 0,42 0,18 stretch longitude [km] 12,6 9,3 6,3 4,7 2,3 stretch surface area [km2] 0,043 0,06 0,043 0,024 0,012 stretch surface area (%) 24 32 24 13 7

dominant substrate composition gravel gravel gravel silt-

grain median size [mm] 12,4 12,9 13,2 0,2 5,4 surface water PO43- [mg L-1] 0,15 0,24 7,0 2,6 1,8

surface water N - NH4+ [mg L-1] 0,39 0,26 0,66 0,72 0,61 surface dissolved oxygen saturation [%] 101,7 110,0 105,8 108,5 103,5

surface water temperature [°C] 8,1 9,7 10,7 11,1 8,9

**Table 1.** Longitudinal physicochemical patterns of the Sitka stream (annual means*)*. Hyporheic water

A few randomly selected subsamples (1 mL) were used for extraction of bacterial cells and, consequently, for estimations of bacterial numbers; other sub-samples were used for

means mix of interstitial water taken from the depth 10 up to 50 cm of the sediment depth

[mg L-1] 0,01 0,21 1,2 0,5 0,18

107,5 127,5 404,8 394,0 397,7 115,3 138,3 414,5 506,5 416,2

2,47 0,81 2,62 2,69 3,74 2,05 1,31 2,71 5,76 2,62

stretch II and III.

surface water N - NO3-

surface water conductivity [µS.cm-1] hyporheic water conductivity [µS.cm-1]

surface water DOC [mg L-1] hyporheic water DOC [mg L-1]

**Figure 1.** A map showing the location of the Sitka stream. Black circles represents the study sites (1-5)

### **2.2. Sediment sample collecting and sample processing**

Five localities alongside stream profile were chosen for sampling sediment and interstitial water samples based on previous investigations (Figure 2, Table 1). Hyporheic sediments were collected with a freeze-core using N2 as a coolant (Bretschko & Klemens 1986) throughout summer period 2009-2011. At each locality, three cores were taken for subsequent analyses. After sampling, surface 0-25 cm sediment layer and layer of 25-50 cm in depth were immediately separated and were stored at a low temperature whilst being transported to the laboratory. Just after thawing, wet sediment of each layer was sieved and only particles < 1 mm were considered for the following microbial measurements and for all microbial activity measurements since most of the biofilm is associated with this fraction (Leichtfried 1988).

al. 2005, 2006).

(Leichtfried 1988).

physicochemical (e.g. temperature, pH, redox, conductivity, O2, CH4, NO3, SO4) patterns in the sediments and a schematic view of the site with sampling point positions have been published previously (Rulík et al. 2000, Rulík & Spáčil 2004). Earlier measurements of a relatively high production of methane, as well as potential methanogenesis, confirmed the suitability of the field sites for the study of methane cycling (Rulík et al. 2000, Hlaváčová et

**Figure 1.** A map showing the location of the Sitka stream. Black circles represents the study sites (1-5)

Five localities alongside stream profile were chosen for sampling sediment and interstitial water samples based on previous investigations (Figure 2, Table 1). Hyporheic sediments were collected with a freeze-core using N2 as a coolant (Bretschko & Klemens 1986) throughout summer period 2009-2011. At each locality, three cores were taken for subsequent analyses. After sampling, surface 0-25 cm sediment layer and layer of 25-50 cm in depth were immediately separated and were stored at a low temperature whilst being transported to the laboratory. Just after thawing, wet sediment of each layer was sieved and only particles < 1 mm were considered for the following microbial measurements and for all microbial activity measurements since most of the biofilm is associated with this fraction

**2.2. Sediment sample collecting and sample processing** 

**Figure 2.** Graphic depiction of the thalweg of the Sitka stream with sampling localities. The main source of pollution is an effluent from Šternberk city sewage water plant, located just in the middle between stretch II and III.


**Table 1.** Longitudinal physicochemical patterns of the Sitka stream (annual means*)*. Hyporheic water means mix of interstitial water taken from the depth 10 up to 50 cm of the sediment depth

A few randomly selected subsamples (1 mL) were used for extraction of bacterial cells and, consequently, for estimations of bacterial numbers; other sub-samples were used for measurement of microbial activity and respiration, organic matter content determination, etc. Sediment organic matter content was determined by oven-drying at 105 oC to constant weight and subsequent combustion at 550 oC for 5 hours to obtain ash-free dry weight (AFDW). Organic matter values were then converted to carbon equivalents assuming 45 % carbon content of organic matter (Meyer et al. 1981). Sediment from another freeze-core was oven-dried at 105 °C and subjected to granulometric analysis. Grain size distribution and descriptive sediment parameters were computed using the database SeDi (Schönbauer & Lewandowski 1999).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 401

Henry's law. The saturation ratio (R) was calculated as the measured concentration of gas divided by the concentration in equilibrium with the atmosphere at the temperature of the

The rate of methane production (methanogenesis) was measured using the PMP method (Segers 1998). C-amended solutions (flushed for 5 minutes with N2) with acetate Ca(CH3COO)2 (100 mg C in the incubation flask) were used for the examination of methanogenic potential. All laboratory sediment incubations were performed in 250-mL dark glass flasks, capped with rubber stoppers, using approximately 100 g (wet mass) of sediment (grain size < 1 mm) and 180 mL of amended solution or distilled water. The headspace was maintained at 20 mL. Typically, triplicate live and dead (methanogenesis was inhibited by addition of 1.0 mM chloroform) samples from each depth were stored at 20°C in the dark and the incubation time was 72 hours; however, subsamples from the headspace atmosphere were taken every 24 hours. Gas production was calculated from the difference between final and initial headspace concentration and volume of the flask; results are expressed per volume unit of wet sediment (CH4 mL-1 WW hour-1) or per unit dry weight of sediment per one day (g CH4 kg-1 DW day-1). Rate of potential methane oxidation (methanotrophy) was measured using modified method of methane oxidation in soil samples from Hanson (1998). Briefly, 50 mL of methane was added by syringe to the closed incubation flask with the sieved sediment and then the pressure was balanced to atmospheric pressure. All laboratory sediment incubations were performed in 250-mL dark glass flasks, capped with rubber stoppers, using approximately 100 g (wet mass) of sediment (grain size < 2 mm). Typically, triplicate live and dead (samples killed by HgCl2 to arrest all biological activity) samples from each depth were stored at 20°C in the dark, and incubation time was 72 hours; however, subsamples from the headspace atmosphere were taken every 24 hours. Potential CH4 oxidation rates at the different concentrations were obtained from the slope of the CH4 decrease with time (r2 > 0.90; methane oxidation was calculated from the difference between final and initial headspace concentration and volume of the flask; results are expressed per volume unit of wet sediment (CH4 mL-1 WW hour-1) or

water sample using the solubility data of Wiesenburg & Guinasso (1979).

**2.4. Methanogenic potential and methanotrophic activity** 

per unit dry weight of sediment per one day (mg CH4 kg-1 DW day-1).

Fluxes of methane across the sediment-water interface were estimated either by direct

The methane fluxes across the sediment-water interface were measured using the method of benthic chambers (e.g. Sansone et al. 1998). Fluxes were measured during the summer months (VII, VIII, IX). The plexiglas chamber (2.6 dm3) covered an area 0.0154 m2. The chambers (n = 7) were installed randomly and gently anchored on the substrate without

measurement with benthic chambers or calculated by applying Fick´s first law.

**2.5. Fluxes of methane across the sediment-water interface** 

*Benthic fluxes* 

### **2.3. Water samples and analysis of methane**

Surface water was collected from running water at a depth of 10 cm below the surface level in autumn 2009 at each study site. Interstitial water samples were collected using a set of 5–6 minipiezometers (Trulleyová et al. 2003) placed at a depth of about 20-50 cm randomly in sediments at each study site. The initial 50–100 mL of water was used as a rinse and discarded. As usual, two subsamples of interstitial water from each minipiezometer were collected from a continuous column of water with a 100 mL polypropylene syringe connected to a hard PVC tube, drawn from a minipiezometer and injected into sterile, clear vials (40 mL) with screw-tops, covered by a polypropylene cap with PTFE silicone septa (for analysis of dissolved gasses) and stored before returning to the laboratory. All samples were taken in the morning between 9 a.m. and 12 noon. All measurements were done during the normal discharge levels (i.e. no spates or high flood levels were included). Interstitial water temperature, dissolved oxygen (mg L-1 and percent saturation) and conductivity were measured in the field with a portable Hanna HI 9828 pH/ORP/EC/DO meter. Dissolved organic carbon (DOC) was measured by Pt-catalysed high temperature combustion on a TOC FORMACSHT analyser. Long term observation of interstitial water temperature was carried out using temperature dataloggers Minikin (EMS Brno, Czech Republic) buried in the sediment depth of 25-30 cm for a period of one year. Dissolved ferrous iron (Fe2+) concentration was measured using absorption spectrophotometry after reaction with 1,10 phenanthroline. Concentrations of organic acids were meausred using capilary electrophoresis equipped with diode array detector HP 3D CE Agilent (Waldbron, Germany). Limits of detection for particular organic acids were set as following: LOD (acetate) = 6,2 µmol L-1; LOD (propionate) = 4,8 µmol L-1; LOD (butyrate) = 2,9 µmol L-1; LOD 32 (valerate) = 1,8 µmol L-1.

Concentrations of dissolved methane in the stream and interstitial water were measured directly using a headspace equilibration technique. Dissolved methane was extracted from the water by replacing 10 mL of water with N2 and then vigorously shaking the vials for 15 seconds (to release the supersaturated gas from the water to facilitate equilibration between the water and gas phases). All samples were equilibrated with air at laboratory temperature. Methane was analysed from the headspace of the vials by injecting 2ml of air sub-sample with a gas-tight syringe into a CHROM 5 gas chromatograph, equipped with the flame ionization detector (CH4 detection limit = 1µg L-1) and with the 1.2m PORAPAK Q column (i.d. 3 mm), with nitrogen as a carrier gas. Gas concentration in water was calculated using Henry's law. The saturation ratio (R) was calculated as the measured concentration of gas divided by the concentration in equilibrium with the atmosphere at the temperature of the water sample using the solubility data of Wiesenburg & Guinasso (1979).

### **2.4. Methanogenic potential and methanotrophic activity**

The rate of methane production (methanogenesis) was measured using the PMP method (Segers 1998). C-amended solutions (flushed for 5 minutes with N2) with acetate Ca(CH3COO)2 (100 mg C in the incubation flask) were used for the examination of methanogenic potential. All laboratory sediment incubations were performed in 250-mL dark glass flasks, capped with rubber stoppers, using approximately 100 g (wet mass) of sediment (grain size < 1 mm) and 180 mL of amended solution or distilled water. The headspace was maintained at 20 mL. Typically, triplicate live and dead (methanogenesis was inhibited by addition of 1.0 mM chloroform) samples from each depth were stored at 20°C in the dark and the incubation time was 72 hours; however, subsamples from the headspace atmosphere were taken every 24 hours. Gas production was calculated from the difference between final and initial headspace concentration and volume of the flask; results are expressed per volume unit of wet sediment (CH4 mL-1 WW hour-1) or per unit dry weight of sediment per one day (g CH4 kg-1 DW day-1). Rate of potential methane oxidation (methanotrophy) was measured using modified method of methane oxidation in soil samples from Hanson (1998). Briefly, 50 mL of methane was added by syringe to the closed incubation flask with the sieved sediment and then the pressure was balanced to atmospheric pressure. All laboratory sediment incubations were performed in 250-mL dark glass flasks, capped with rubber stoppers, using approximately 100 g (wet mass) of sediment (grain size < 2 mm). Typically, triplicate live and dead (samples killed by HgCl2 to arrest all biological activity) samples from each depth were stored at 20°C in the dark, and incubation time was 72 hours; however, subsamples from the headspace atmosphere were taken every 24 hours. Potential CH4 oxidation rates at the different concentrations were obtained from the slope of the CH4 decrease with time (r2 > 0.90; methane oxidation was calculated from the difference between final and initial headspace concentration and volume of the flask; results are expressed per volume unit of wet sediment (CH4 mL-1 WW hour-1) or per unit dry weight of sediment per one day (mg CH4 kg-1 DW day-1).

### **2.5. Fluxes of methane across the sediment-water interface**

Fluxes of methane across the sediment-water interface were estimated either by direct measurement with benthic chambers or calculated by applying Fick´s first law.

### *Benthic fluxes*

400 Biomass Now – Cultivation and Utilization

Lewandowski 1999).

32 (valerate) = 1,8 µmol L-1.

**2.3. Water samples and analysis of methane** 

measurement of microbial activity and respiration, organic matter content determination, etc. Sediment organic matter content was determined by oven-drying at 105 oC to constant weight and subsequent combustion at 550 oC for 5 hours to obtain ash-free dry weight (AFDW). Organic matter values were then converted to carbon equivalents assuming 45 % carbon content of organic matter (Meyer et al. 1981). Sediment from another freeze-core was oven-dried at 105 °C and subjected to granulometric analysis. Grain size distribution and descriptive sediment parameters were computed using the database SeDi (Schönbauer &

Surface water was collected from running water at a depth of 10 cm below the surface level in autumn 2009 at each study site. Interstitial water samples were collected using a set of 5–6 minipiezometers (Trulleyová et al. 2003) placed at a depth of about 20-50 cm randomly in sediments at each study site. The initial 50–100 mL of water was used as a rinse and discarded. As usual, two subsamples of interstitial water from each minipiezometer were collected from a continuous column of water with a 100 mL polypropylene syringe connected to a hard PVC tube, drawn from a minipiezometer and injected into sterile, clear vials (40 mL) with screw-tops, covered by a polypropylene cap with PTFE silicone septa (for analysis of dissolved gasses) and stored before returning to the laboratory. All samples were taken in the morning between 9 a.m. and 12 noon. All measurements were done during the normal discharge levels (i.e. no spates or high flood levels were included). Interstitial water temperature, dissolved oxygen (mg L-1 and percent saturation) and conductivity were measured in the field with a portable Hanna HI 9828 pH/ORP/EC/DO meter. Dissolved organic carbon (DOC) was measured by Pt-catalysed high temperature combustion on a TOC FORMACSHT analyser. Long term observation of interstitial water temperature was carried out using temperature dataloggers Minikin (EMS Brno, Czech Republic) buried in the sediment depth of 25-30 cm for a period of one year. Dissolved ferrous iron (Fe2+) concentration was measured using absorption spectrophotometry after reaction with 1,10 phenanthroline. Concentrations of organic acids were meausred using capilary electrophoresis equipped with diode array detector HP 3D CE Agilent (Waldbron, Germany). Limits of detection for particular organic acids were set as following: LOD (acetate) = 6,2 µmol L-1; LOD (propionate) = 4,8 µmol L-1; LOD (butyrate) = 2,9 µmol L-1; LOD

Concentrations of dissolved methane in the stream and interstitial water were measured directly using a headspace equilibration technique. Dissolved methane was extracted from the water by replacing 10 mL of water with N2 and then vigorously shaking the vials for 15 seconds (to release the supersaturated gas from the water to facilitate equilibration between the water and gas phases). All samples were equilibrated with air at laboratory temperature. Methane was analysed from the headspace of the vials by injecting 2ml of air sub-sample with a gas-tight syringe into a CHROM 5 gas chromatograph, equipped with the flame ionization detector (CH4 detection limit = 1µg L-1) and with the 1.2m PORAPAK Q column (i.d. 3 mm), with nitrogen as a carrier gas. Gas concentration in water was calculated using

The methane fluxes across the sediment-water interface were measured using the method of benthic chambers (e.g. Sansone et al. 1998). Fluxes were measured during the summer months (VII, VIII, IX). The plexiglas chamber (2.6 dm3) covered an area 0.0154 m2. The chambers (n = 7) were installed randomly and gently anchored on the substrate without disturbing the sediment. Samples to determine of initial concentration of CH4 were collected from each chamber before the beginning of incubation. Incubation time was 24 hours. Samples of water were stored in 40 ml glass vials closed by cap with PTFE/silicone septum until analysis.

### *Diffusive fluxes*

Fluxes of methane between the sediment and overlying water were calculated from Fick´s first law as described by Berner (1980):

$$J = -D\_{\mathbb{S}} \times \Phi \times \left(\Lambda \mathbb{C} / \Delta \mathbb{x}\right) \tag{1}$$

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 403

where *F* is a gas flux in mg m-2day-1; *cI* is a concentration of particular gas in the chamber headspace in g L-1; *cR* is a concentration of particular gas in background air; *V* is volume of the chamber in L; *t* is time of incubation in hr; *p* is an area of chamber expressed in m2 . For each chamber, the fluxes were calculated using linear regression based on the concentration change as a function of time, regardless of the value of the coefficient of determination (cf.

In order to assess emissions produced from a total stream area, the stream was divided into five stretches according to the channel width, water velocity and substrate composition. For each stretch we have then chosen one representative sampling site (locality I-V) where samples of both stream and interstitial waters and sediments, respectively, were repeatedly taken. Localities were chosen in respect to their character and availability by car and measuring equipments. For calculation of whole-stream gases emissions into the atmosphere, the total stream area was derived from summing of 14 partial stretches. The area of these stretches was caculated from known lenght and mean channel width (measured by a metal measuring type). Longitudinal distance among the stretches was evaluated by using ArcGIS software and GPS coordinates that have been obtained during the field measurement and from digitalised map of the Sitka stream. The total area of the Sitka stream was estimated to be 181 380 m2 or 0.18 km2. Stretches have differed in their

percentual contribution to this total area and also by their total lenght (Table 1).

The total annual methane emissions to the atmosphere from the five segments of the Sitka stream, *E*a (kg yr-1) were derived from seasonal average, maximum or minimum emissions measured on every locality and extrapolated to the total area of the particular segment. The total methane emissions produced by the Sitka stream annualy were then calculated

where *E*a is average, maximal or minimal assess of emission of methane from the total stream area in kilograms per year; *pi* is an area of stretch (in m2) representing given locality; *Fi* is average, maximal or minimal assess of the methane from a given locality expressed in

**2.7. Carbon isotopic composition of dissolved methane and carbon dioxide in** 

Interstitial water samples for carbon isotopic analysis of methane and carbon dioxide were collected in 2010 - 2011 through three courses at study site. Sampling was performed by set of minipiezometers placed in a depth of 20 to 60 cm randomly in a sediment. After sampling, refrigerated samples were transported (within 72 hours) in 250 mL bottles to laboratory at the Department of Plant Physiology, Faculty of Science University of South Bohemia in Ceske Budejovice, which are equipped with mass spectrometry for carbon isotopes measurements. Firstly both water samples, for methane and for carbon dioxide, were extracted to helium headspace. After relaxation time isotopic equilibrium was

a ii *E* p \* F \* 365 / 1 000 000 (5)

Duchemin et al. 1999, Silvenoinen et al. 2008).

according to the following formula:

mg m-2day-1.

**sediments** 

where *J* is the diffusive flux in µg m-2 s-1, *Ф* is the porosity of the sediment, *DS* is the bulk sediment diffusion coefficient in cm-2 s-1, *∆C/∆x* is the methane concetrations gradient in µg cm-3 cm-1. Bulk sediment diffusion coefficient (*DS*) is based on diffusion coefficient for methane in the water (*D0*) and tortuosity (*θ*) according to the formula:

$$D\_{\mathbb{S}} = D\_0 \theta^{-2} \tag{2}$$

Tortuosity (*θ*) is possible calculate from porosity according to equation (Boudreau 1996):

$$
\boldsymbol{\theta}^{-2} = \boldsymbol{1} - \ln(\boldsymbol{\Phi}^2) \tag{3}
$$

Diffusive fluxes of CH4 were determined at all five study sites along the longitudinal profile of the Sitka stream.

### **2.6. Measurement of emissions**

Gas flux across the air-water interface was determined by the floating chamber method four times during the year period in 2005 – 2006. The open-bottom floating PE chambers (5L domes with an area of 0.03 m2) were maintained on the water's surface by a floating body (Styrene) attached to the outside. The chambers (n = 4 – 5) were allowed to float on the water's surface for a period of 3 hours. Previous measurements confirmed that time to be quite enough to establish linear dependence of concentration change inside the chambers on time for the gas samples collected every 30 min over a 3 hour period. Due to trees on the banks, the chambers at all study sites were continuously in the shade. On each sampling occasion, ambient air samples were collected for determining the initial background concentrations. Samples of headspace gas were collected through the rubber stopper inserted at the chamber's top, and stored in 100mL PE gas-tight syringes until analysis. Emissions were calculated as the difference between initial background and final concentration in the chamber headspace, and expressed on the 1m2 area of the surface level per day according to the formula:

$$\mathbf{F} = \begin{bmatrix} \begin{pmatrix} \mathbf{c}\_{\mathrm{I}} - \mathbf{c}\_{\mathrm{R}} \end{pmatrix} \text{ \* } \mathbf{V} \text{ \* 24 / \text{ t} \* 1000} \end{bmatrix} / \text{ \* } \mathbf{p} \tag{4}$$

where *F* is a gas flux in mg m-2day-1; *cI* is a concentration of particular gas in the chamber headspace in g L-1; *cR* is a concentration of particular gas in background air; *V* is volume of the chamber in L; *t* is time of incubation in hr; *p* is an area of chamber expressed in m2 . For each chamber, the fluxes were calculated using linear regression based on the concentration change as a function of time, regardless of the value of the coefficient of determination (cf. Duchemin et al. 1999, Silvenoinen et al. 2008).

402 Biomass Now – Cultivation and Utilization

first law as described by Berner (1980):

until analysis.

*Diffusive fluxes* 

of the Sitka stream.

**2.6. Measurement of emissions** 

per day according to the formula:

disturbing the sediment. Samples to determine of initial concentration of CH4 were collected from each chamber before the beginning of incubation. Incubation time was 24 hours. Samples of water were stored in 40 ml glass vials closed by cap with PTFE/silicone septum

Fluxes of methane between the sediment and overlying water were calculated from Fick´s

where *J* is the diffusive flux in µg m-2 s-1, *Ф* is the porosity of the sediment, *DS* is the bulk sediment diffusion coefficient in cm-2 s-1, *∆C/∆x* is the methane concetrations gradient in µg cm-3 cm-1. Bulk sediment diffusion coefficient (*DS*) is based on diffusion coefficient for

> <sup>2</sup> *D D <sup>S</sup>* <sup>0</sup>

2 2

Diffusive fluxes of CH4 were determined at all five study sites along the longitudinal profile

Gas flux across the air-water interface was determined by the floating chamber method four times during the year period in 2005 – 2006. The open-bottom floating PE chambers (5L domes with an area of 0.03 m2) were maintained on the water's surface by a floating body (Styrene) attached to the outside. The chambers (n = 4 – 5) were allowed to float on the water's surface for a period of 3 hours. Previous measurements confirmed that time to be quite enough to establish linear dependence of concentration change inside the chambers on time for the gas samples collected every 30 min over a 3 hour period. Due to trees on the banks, the chambers at all study sites were continuously in the shade. On each sampling occasion, ambient air samples were collected for determining the initial background concentrations. Samples of headspace gas were collected through the rubber stopper inserted at the chamber's top, and stored in 100mL PE gas-tight syringes until analysis. Emissions were calculated as the difference between initial background and final concentration in the chamber headspace, and expressed on the 1m2 area of the surface level

I R F c – c \* V \* 24 / t \* 1000 / p

(4)

Tortuosity (*θ*) is possible calculate from porosity according to equation (Boudreau 1996):

methane in the water (*D0*) and tortuosity (*θ*) according to the formula:

/ *<sup>S</sup> J D Ф C x* (1)

(2)

1 – *ln Ф* (3)

In order to assess emissions produced from a total stream area, the stream was divided into five stretches according to the channel width, water velocity and substrate composition. For each stretch we have then chosen one representative sampling site (locality I-V) where samples of both stream and interstitial waters and sediments, respectively, were repeatedly taken. Localities were chosen in respect to their character and availability by car and measuring equipments. For calculation of whole-stream gases emissions into the atmosphere, the total stream area was derived from summing of 14 partial stretches. The area of these stretches was caculated from known lenght and mean channel width (measured by a metal measuring type). Longitudinal distance among the stretches was evaluated by using ArcGIS software and GPS coordinates that have been obtained during the field measurement and from digitalised map of the Sitka stream. The total area of the Sitka stream was estimated to be 181 380 m2 or 0.18 km2. Stretches have differed in their percentual contribution to this total area and also by their total lenght (Table 1).

The total annual methane emissions to the atmosphere from the five segments of the Sitka stream, *E*a (kg yr-1) were derived from seasonal average, maximum or minimum emissions measured on every locality and extrapolated to the total area of the particular segment. The total methane emissions produced by the Sitka stream annualy were then calculated according to the following formula:

$$E\_{\mathbf{a}} = \left(\sum \mathbf{p}\_{\mathbf{i}} \, ^\* \, \, \mathbf{F}\_{\mathbf{i}} \, ^\* \, 365\right) \, / \, 1000000 \, \tag{5}$$

where *E*a is average, maximal or minimal assess of emission of methane from the total stream area in kilograms per year; *pi* is an area of stretch (in m2) representing given locality; *Fi* is average, maximal or minimal assess of the methane from a given locality expressed in mg m-2day-1.

### **2.7. Carbon isotopic composition of dissolved methane and carbon dioxide in sediments**

Interstitial water samples for carbon isotopic analysis of methane and carbon dioxide were collected in 2010 - 2011 through three courses at study site. Sampling was performed by set of minipiezometers placed in a depth of 20 to 60 cm randomly in a sediment. After sampling, refrigerated samples were transported (within 72 hours) in 250 mL bottles to laboratory at the Department of Plant Physiology, Faculty of Science University of South Bohemia in Ceske Budejovice, which are equipped with mass spectrometry for carbon isotopes measurements. Firstly both water samples, for methane and for carbon dioxide, were extracted to helium headspace. After relaxation time isotopic equilibrium was achieved and four subsamples of gas were determined by GasBanch (ThermoScientific) and IRMS DeltaplusXL equiped by TC/EA (ThermoFinnigan) for analysis of δ13CO2. Afterwards δ13CO2 of water samples were calculated from gaseous δ13CO2 by fractionation factor from a linear equation (Szaran 1997):

$$\left(\varepsilon\_{\rm o}\right)^{13}\text{C}=-\left(0.0954\pm0.0027\right)\text{T[C]}+\left(10.41\pm0.12\right)\tag{6}$$

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 405

least 20 microscopic fields were counted in three replicates from each locality. TCN was

The methanogenic archaea, three selected methanogen families (*Methanobacteriaceae*, *Methanosetaceae* and *Methanosarcinaceae*) and methanotrophic bacteria belonging to groups I and II were detected using FISH (Fluorescence in situ hybridization) with 16S rRNAtargeted oligonucleotide probe labelled with indocarbocyanine dye Cy3. The prokaryotes were hybridized according to the protocol by Pernthaler et al. (2001). Briefly, the supernatants which were used also for TCN were filtered onto polycarbonate membrane filters (0.2 µm GTTP; Millipore), filters were cut into sections and placed on glass slides. For the hybridization mixtures, 2 µL of probe-working solution was added to 16 µL of hybridization buffer in a microfuge tube. Hybridization mix was added to the samples and the slides with filter sections were incubated at 46 °C for 3 hours. After incubation, the sections were transferred into preheated washing buffer (48 °C) and incubated for 15 minutes in a water bath at the same temperature. The filter sections were washed and airdried. The DAPI staining procedure followed as previously described. Finally, the samples were mounted in a 4:1 mix of Citifluor and Vecta Shield. The methanogens and methanotrophs were counted in three replicates from each locality and the relative proportion of bacteria, archaea, methanogens and methanotrophs to the total number of

**2.11. Nucleic acid extraction and Denaturing gradient gel electrophoresis** 

Nucleic acids were extracted from 0,3 g of sieved sediment with a Power Soil DNA isolation kit (MoBio, Carlsbad, USA) according to the manufacturer's instructions. 16S rRNA gene fragments (~350 bp) were amplified by PCR using primer pair specific for methanogens. Primer sequences are as follows, 0357 F-GC 5'-CCC TAC GGG GCG CAG CAG-3' (GC clamp at 5'-end CGC CCG CCG CGC GCG GCG GGCGGG GCG GGG GCA CGG GGG G) and 0691 R 5'- GGA TTA CAR GAT TTC AC -3' (Watanabe et al. 2004). PCR amplification was carried out in 50 µL reaction mixture contained within 0.2 mL, thin walled micro-tubes. Amplification was performed in a TC-XP thermal cycler (Bioer Technology, Hangzhou, China). The reaction mixture contained 5 µL of 10 × PCR amplification buffer, 200 µM of each dNTP, 0,8 µM of each primer, 8 µL of template DNA and 5.0 U of FastStart Taq DNA polymerase (Polymerase dNTPack; Roche, Germany). The initial enzyme activation and DNA denaturation were performed for 6 min at 95°C, followed by 35 cycles of 1 min at 95°C, 1 min at 55°C and 2 min at 69°C and a final extension at 69°C for 8 min (protocol by Watanabe et al. 2004). PCR products were visualised by electrophoresis in ethidium

DGGE was performed with an INGENYphorU System (Ingeny, Netherlands). PCR products were loaded onto a 7% (w/v) polyacrylamide gel (acrylamide: bisacrylamide, 37.5:1). The

expressed as bacterial numbers per 1 mL of wet sediments.

**2.10. Procaryotic community composition** 

DAPI stained cells was then calculated.

bromide stained, 1.5% (w/v) agarose gel.

**(DGGE)** 

Stable isotope analysis of 13C/12C in gas samples was performed using preconcentration, kryoseparation of CO2 and gas chromatograph combustion of CH4 in PreCon (ThermoFinnigan) coupled to isotope ratio mass spectrometer (IRMS, Delta Plus XL, ThermoFinnigan, Brehmen, Germany). After conversion of CH4 to CO2 in the Finningan standard GC Combustion interface CO2 will be tranfered into IRMS. The obtained 13C/12C ratios (R) will be referenced to 13C/12C of standard V-PDB (Vienna-Pee-Dee Belemnite)(Rs), and expressed as δ13C = (Rsample/Rstandard – 1) x 1000 in ‰. The standard deviation of δ13C determination in standard samples is lower than 0.1‰ with our instrumentation. From our data, we also calculated an apparent fractionation factor αC that is defined by the measured δCH4 and δCO2 (Whiticar et al. 1986):

$$a\_{\mathbb{C}} = \left(\delta \text{CO}\_2 + 10^3\right) \mid \left(\delta \text{CH}\_4 + 10^3\right) \tag{7}$$

This fractionation factor gives rough idea of magnitude of acetoclastic and hydrogenotrophic methanogenesis.

### **2.8. Abundance of microbial cells and microbial community composition**

For measuring of microbial parameters, formaldehyde fixed samples (2 % final conc.) were first mildly sonicated for 30 seconds at the 15 % power (sonotroda MS 73, Sonopuls HD2200, Sonorex, Germany), followed by incubation for 3 hours under mild agitation with 10 mL of detergent mixture (Tween 20 0.5%, vol/vol, tetrasodium pyrophosphate 0.1 M and distilled water) and density centrifugation (Santos Furtado & Casper 2000, Amalfitano & Fazi 2008). For density centrifugaton, the non-ionic medium Nycodenz (1.31 g mL-1; Axis- Shield, Oslo, Norway) was used at 4600 G for 60 minutes (Rotofix 32A, Hettich, Germany). After the preparation processes, a 1 mL of Nycodenz was placed underneath 2 ml of treated slurry using a syringe needle (Fazi et al. 2005). 1 ml of supernatant was then taken for subsequent analysis.

### **2.9. Total cell numbers (TCN)**

The supernatant was filtered onto membrane filters (0.2 µm GTTP; Millipore Germany), stained for 10 minutes in cold and in the dark with DAPI solution (1 mg/ ml; wt/ vol; Sigma, Germany) and gently rinsed in distilled water and 80 % ethanol. Filters were air-dried and fixed in immersion oil. Stained cells were enumerated on an epifluorescence microscope (Olympus BX 60) equipped with a camera (Olympus DP 12) and image analysis software (NIS Elements; Laboratory Imaging, Prague, Czech Republic). At least 200 cells within at least 20 microscopic fields were counted in three replicates from each locality. TCN was expressed as bacterial numbers per 1 mL of wet sediments.

### **2.10. Procaryotic community composition**

404 Biomass Now – Cultivation and Utilization

linear equation (Szaran 1997):

δCH4 and δCO2 (Whiticar et al. 1986):

hydrogenotrophic methanogenesis.

**2.9. Total cell numbers (TCN)** 

analysis.

achieved and four subsamples of gas were determined by GasBanch (ThermoScientific) and IRMS DeltaplusXL equiped by TC/EA (ThermoFinnigan) for analysis of δ13CO2. Afterwards δ13CO2 of water samples were calculated from gaseous δ13CO2 by fractionation factor from a

Stable isotope analysis of 13C/12C in gas samples was performed using preconcentration, kryoseparation of CO2 and gas chromatograph combustion of CH4 in PreCon (ThermoFinnigan) coupled to isotope ratio mass spectrometer (IRMS, Delta Plus XL, ThermoFinnigan, Brehmen, Germany). After conversion of CH4 to CO2 in the Finningan standard GC Combustion interface CO2 will be tranfered into IRMS. The obtained 13C/12C ratios (R) will be referenced to 13C/12C of standard V-PDB (Vienna-Pee-Dee Belemnite)(Rs), and expressed as δ13C = (Rsample/Rstandard – 1) x 1000 in ‰. The standard deviation of δ13C determination in standard samples is lower than 0.1‰ with our instrumentation. From our data, we also calculated an apparent fractionation factor αC that is defined by the measured

3 3

This fractionation factor gives rough idea of magnitude of acetoclastic and

For measuring of microbial parameters, formaldehyde fixed samples (2 % final conc.) were first mildly sonicated for 30 seconds at the 15 % power (sonotroda MS 73, Sonopuls HD2200, Sonorex, Germany), followed by incubation for 3 hours under mild agitation with 10 mL of detergent mixture (Tween 20 0.5%, vol/vol, tetrasodium pyrophosphate 0.1 M and distilled water) and density centrifugation (Santos Furtado & Casper 2000, Amalfitano & Fazi 2008). For density centrifugaton, the non-ionic medium Nycodenz (1.31 g mL-1; Axis- Shield, Oslo, Norway) was used at 4600 G for 60 minutes (Rotofix 32A, Hettich, Germany). After the preparation processes, a 1 mL of Nycodenz was placed underneath 2 ml of treated slurry using a syringe needle (Fazi et al. 2005). 1 ml of supernatant was then taken for subsequent

The supernatant was filtered onto membrane filters (0.2 µm GTTP; Millipore Germany), stained for 10 minutes in cold and in the dark with DAPI solution (1 mg/ ml; wt/ vol; Sigma, Germany) and gently rinsed in distilled water and 80 % ethanol. Filters were air-dried and fixed in immersion oil. Stained cells were enumerated on an epifluorescence microscope (Olympus BX 60) equipped with a camera (Olympus DP 12) and image analysis software (NIS Elements; Laboratory Imaging, Prague, Czech Republic). At least 200 cells within at

**2.8. Abundance of microbial cells and microbial community composition** 

 

C2 4 CO 10 / CH 10 (7)

C 0.0954 0.0027 T[ C] (10.41 0.12) (6)

<sup>13</sup>

The methanogenic archaea, three selected methanogen families (*Methanobacteriaceae*, *Methanosetaceae* and *Methanosarcinaceae*) and methanotrophic bacteria belonging to groups I and II were detected using FISH (Fluorescence in situ hybridization) with 16S rRNAtargeted oligonucleotide probe labelled with indocarbocyanine dye Cy3. The prokaryotes were hybridized according to the protocol by Pernthaler et al. (2001). Briefly, the supernatants which were used also for TCN were filtered onto polycarbonate membrane filters (0.2 µm GTTP; Millipore), filters were cut into sections and placed on glass slides. For the hybridization mixtures, 2 µL of probe-working solution was added to 16 µL of hybridization buffer in a microfuge tube. Hybridization mix was added to the samples and the slides with filter sections were incubated at 46 °C for 3 hours. After incubation, the sections were transferred into preheated washing buffer (48 °C) and incubated for 15 minutes in a water bath at the same temperature. The filter sections were washed and airdried. The DAPI staining procedure followed as previously described. Finally, the samples were mounted in a 4:1 mix of Citifluor and Vecta Shield. The methanogens and methanotrophs were counted in three replicates from each locality and the relative proportion of bacteria, archaea, methanogens and methanotrophs to the total number of DAPI stained cells was then calculated.

### **2.11. Nucleic acid extraction and Denaturing gradient gel electrophoresis (DGGE)**

Nucleic acids were extracted from 0,3 g of sieved sediment with a Power Soil DNA isolation kit (MoBio, Carlsbad, USA) according to the manufacturer's instructions. 16S rRNA gene fragments (~350 bp) were amplified by PCR using primer pair specific for methanogens. Primer sequences are as follows, 0357 F-GC 5'-CCC TAC GGG GCG CAG CAG-3' (GC clamp at 5'-end CGC CCG CCG CGC GCG GCG GGCGGG GCG GGG GCA CGG GGG G) and 0691 R 5'- GGA TTA CAR GAT TTC AC -3' (Watanabe et al. 2004). PCR amplification was carried out in 50 µL reaction mixture contained within 0.2 mL, thin walled micro-tubes. Amplification was performed in a TC-XP thermal cycler (Bioer Technology, Hangzhou, China). The reaction mixture contained 5 µL of 10 × PCR amplification buffer, 200 µM of each dNTP, 0,8 µM of each primer, 8 µL of template DNA and 5.0 U of FastStart Taq DNA polymerase (Polymerase dNTPack; Roche, Germany). The initial enzyme activation and DNA denaturation were performed for 6 min at 95°C, followed by 35 cycles of 1 min at 95°C, 1 min at 55°C and 2 min at 69°C and a final extension at 69°C for 8 min (protocol by Watanabe et al. 2004). PCR products were visualised by electrophoresis in ethidium bromide stained, 1.5% (w/v) agarose gel.

DGGE was performed with an INGENYphorU System (Ingeny, Netherlands). PCR products were loaded onto a 7% (w/v) polyacrylamide gel (acrylamide: bisacrylamide, 37.5:1). The polyacrylamide gels were made of 0.05% (v/v) TEMED (N,N,N,N-tetramethylenediamine), 0.06% (w/v) ammonium persulfate, 7 M (w/v) urea and 40 % (v/v) formamide. Denaturing gradients ranged from 45 to 60%. Electrophoresis was performed in 1×TAE buffer (40 mM Tris, 1 mM acetic acid, 1 mM EDTA, pH 7.45) and run initially at 110V for 10 min at 60°C, afterwatds for 16 h at 85 V. After electrophoresis, the gels were stained for 60 min with SYBR Green I nucleic acid gel stain (1:10 000 dilution) (Lonza, Rockland USA) DGGE gel was then photographed under UV transilluminator (Molecular Dynamics). Images were arranged by Image analysis (NIS Elements, Czech Republic). A binary matrix was created from the gel image by scoring of the presence or absence of each bend and then the cluster tree was constructed (programme GEL2k; Svein Norland, Dept. Of Biology, University of Bergen).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 407

phylogenetic tree was constructed by the maximum likelihood method (Kimura 2 parameter model). The tree topology was statistically evaluated by 1000 bootstrap replicates

The physicochemical sediment and interstitial water properties of the investigated sites showed large horizontal and vertical gradients. Sediment grain median size decreased along a longitudinal profile while organic carbon content in a sediment fraction < 1 mm remained unchaged (Table 2). Generally, interstitial water revealed relatively high dissolved oxygen saturation with the exceptions of localities IV and V where concentration of dissolved oxygen sharply decreased with the depth, however, never dropped below 10%. Vice versa, these two localities were characterized by much higher concentrations of ferrous iron and dissolved methane (Table 2) compared to those sites located upstream. Concentration of the ferrous iron reflects anaerobic conditions of the sediment and showed the highest concentration to occur in the deepest sediment layers (40-50cm). Average annual temperatures of interstitial water at localities in downstream part of the Sitka stream were about 2.5 oC higher compared to localities upstream and may probably promote higher methane production occuring here. Precursors of methanogenesis, acetate, propionate and butyrate were found to be present in the interstitial water at all study sites, however, only acetate was measured regularly at higher concentration with maximum concentration

**Variable/ Locality I II III IV V**  particulate organic C in sediment < 1 mm [%] 0.9 0.9 0.6 1.3 0.7 interstitial dissolved O2 saturation [%] 80.5 88.1 82.3 38.5 50.9 ferrous iron [mg L-1] < 1 < 1 1.8 8.1 4.2 acetate [mmol L-1] 0.21 0.34 0.52 1.87 0.29 interstitial CH4 concentration [µg L-1] 4.9 0.7 8.1 2 480.2 42.8

[pM CH4 mL-1 WW hour-1] 6.6 1.9 2.9 80.7 9.7

[nM CH4 mL-1 WW hour-1] 0.3 1.3 28.5 30.3 25.1 average daily interstitial water temperature [oC] 8.7 9.4 11.6 11.2 11.4 **Table 2.** Selected physicochemical parameters (annual means) of the hyporheic interstitial water and

Methanogenic potential (MP) was found to be significantly higher in the upper sediment layer compared to that from deeper sediment layer. Generally, average MP varied between 0.74-158.6 pM CH4 mL-1 WW hour-1 with the highest values found at site IV. Average

**3.2. Methanogenic potential and methanotrophic activity of sediments** 

(maximum likelihood) and 2000 bootstrap replicates (neighbour joining).

**3. Results** 

**3.1. Sediment and interstitial water** 

reached usually during a summer period.

sediments of studied localities taken from the depth 25-30 cm.

methanogenic potential

methanotrophic activity

### **2.12. PCR amplification, cloning and sequencing of methyl coenzyme M reductase (mcrA) gene**

Fragments of the methanogen DNA (~470 bp) were amplified by PCR using *mcrA* gene specific primers. Primer sequences for mcrA gene are as follows, mcrA F 5'- GGTGGTGTACGGATTCACACAAGTACTGCATACAGC-3',mcrA R 5'- TTCATTGCAGTAGTTATGGAGTAGTT-3'. PCR amplification was carried out in 50 µl reaction mixture contained within 0.2 mL thin walled micro-tubes. Amplification was performed in a TC-XP thermal cycler (Bioer Technology, Hangzhou, China). The reaction mixture contained 5 µL of 10 x PCR amplification buffer, 200 µM of each dNTP, 0.8 µM of each primer, 2 µL of template DNA and 2.5 U of FastStart Taq DNA polymerase (Polymerase dNTPack; Roche, Mannheim, Germany). The initial enzyme activation and DNA denaturation were performed for 6 min at 95°C, followed by 5 cycles of 30s at 95°C, 30s at 55°C and 30s at 72°C, and the temperature ramp rate between the annealing and extension segment was set to 0.1°C/s because of the degeneracy of the primers. After this, the ramp rate was set to 1°C/s, and 30 cycles were performed with the following conditions: 30 s at 95°C, 30 s at 55°C, 30s at 72°C and a final extension at 72°C for 8 min. PCR products were visualised by electrophoresis in ethidium bromide stained, 1.5% (w/v) agarose gel.

Purified PCR amplicons (PCR purification kit; Qiagen, Venlo, Netherlands) were ligated into TOPO TA cloning vectors and transformed into chemically competent *Escherichia coli*  TOP10F' cells according to the manufacturer's instructions (Invitrogen, Carlsbad, USA). Positive colonies were screened by PCR amplification with the primer set and PCR conditions described above. Plasmids were extracted using UltraClean 6 Minute Plasmid Prep Kit (MoBio, Carlsbad, USA), and nucleotide sequences of cloned genes were determined by sequencing with M13 primers in Macrogen company (Seoul, Korea). Raw sequences obtained after sequencing were BLAST analysed to search for the sequence identity between other methanogen sequences available in the GenBank database. Then these sequences were aligned by using CLUSTAL W in order to remove any similar sequences. The most appropriate substitution model for maximum likelihood analysis was identified by Bayesian Information Criterion implemented in MEGA 5.05 software. The phylogenetic tree was constructed by the maximum likelihood method (Kimura 2 parameter model). The tree topology was statistically evaluated by 1000 bootstrap replicates (maximum likelihood) and 2000 bootstrap replicates (neighbour joining).

### **3. Results**

406 Biomass Now – Cultivation and Utilization

Bergen).

**reductase (mcrA) gene** 

polyacrylamide gels were made of 0.05% (v/v) TEMED (N,N,N,N-tetramethylenediamine), 0.06% (w/v) ammonium persulfate, 7 M (w/v) urea and 40 % (v/v) formamide. Denaturing gradients ranged from 45 to 60%. Electrophoresis was performed in 1×TAE buffer (40 mM Tris, 1 mM acetic acid, 1 mM EDTA, pH 7.45) and run initially at 110V for 10 min at 60°C, afterwatds for 16 h at 85 V. After electrophoresis, the gels were stained for 60 min with SYBR Green I nucleic acid gel stain (1:10 000 dilution) (Lonza, Rockland USA) DGGE gel was then photographed under UV transilluminator (Molecular Dynamics). Images were arranged by Image analysis (NIS Elements, Czech Republic). A binary matrix was created from the gel image by scoring of the presence or absence of each bend and then the cluster tree was constructed (programme GEL2k; Svein Norland, Dept. Of Biology, University of

**2.12. PCR amplification, cloning and sequencing of methyl coenzyme M** 

Fragments of the methanogen DNA (~470 bp) were amplified by PCR using *mcrA* gene specific primers. Primer sequences for mcrA gene are as follows, mcrA F 5'- GGTGGTGTACGGATTCACACAAGTACTGCATACAGC-3',mcrA R 5'- TTCATTGCAGTAGTTATGGAGTAGTT-3'. PCR amplification was carried out in 50 µl reaction mixture contained within 0.2 mL thin walled micro-tubes. Amplification was performed in a TC-XP thermal cycler (Bioer Technology, Hangzhou, China). The reaction mixture contained 5 µL of 10 x PCR amplification buffer, 200 µM of each dNTP, 0.8 µM of each primer, 2 µL of template DNA and 2.5 U of FastStart Taq DNA polymerase (Polymerase dNTPack; Roche, Mannheim, Germany). The initial enzyme activation and DNA denaturation were performed for 6 min at 95°C, followed by 5 cycles of 30s at 95°C, 30s at 55°C and 30s at 72°C, and the temperature ramp rate between the annealing and extension segment was set to 0.1°C/s because of the degeneracy of the primers. After this, the ramp rate was set to 1°C/s, and 30 cycles were performed with the following conditions: 30 s at 95°C, 30 s at 55°C, 30s at 72°C and a final extension at 72°C for 8 min. PCR products were visualised by electrophoresis in ethidium bromide stained, 1.5% (w/v) agarose gel.

Purified PCR amplicons (PCR purification kit; Qiagen, Venlo, Netherlands) were ligated into TOPO TA cloning vectors and transformed into chemically competent *Escherichia coli*  TOP10F' cells according to the manufacturer's instructions (Invitrogen, Carlsbad, USA). Positive colonies were screened by PCR amplification with the primer set and PCR conditions described above. Plasmids were extracted using UltraClean 6 Minute Plasmid Prep Kit (MoBio, Carlsbad, USA), and nucleotide sequences of cloned genes were determined by sequencing with M13 primers in Macrogen company (Seoul, Korea). Raw sequences obtained after sequencing were BLAST analysed to search for the sequence identity between other methanogen sequences available in the GenBank database. Then these sequences were aligned by using CLUSTAL W in order to remove any similar sequences. The most appropriate substitution model for maximum likelihood analysis was identified by Bayesian Information Criterion implemented in MEGA 5.05 software. The

### **3.1. Sediment and interstitial water**

The physicochemical sediment and interstitial water properties of the investigated sites showed large horizontal and vertical gradients. Sediment grain median size decreased along a longitudinal profile while organic carbon content in a sediment fraction < 1 mm remained unchaged (Table 2). Generally, interstitial water revealed relatively high dissolved oxygen saturation with the exceptions of localities IV and V where concentration of dissolved oxygen sharply decreased with the depth, however, never dropped below 10%. Vice versa, these two localities were characterized by much higher concentrations of ferrous iron and dissolved methane (Table 2) compared to those sites located upstream. Concentration of the ferrous iron reflects anaerobic conditions of the sediment and showed the highest concentration to occur in the deepest sediment layers (40-50cm). Average annual temperatures of interstitial water at localities in downstream part of the Sitka stream were about 2.5 oC higher compared to localities upstream and may probably promote higher methane production occuring here. Precursors of methanogenesis, acetate, propionate and butyrate were found to be present in the interstitial water at all study sites, however, only acetate was measured regularly at higher concentration with maximum concentration reached usually during a summer period.


**Table 2.** Selected physicochemical parameters (annual means) of the hyporheic interstitial water and sediments of studied localities taken from the depth 25-30 cm.

### **3.2. Methanogenic potential and methanotrophic activity of sediments**

Methanogenic potential (MP) was found to be significantly higher in the upper sediment layer compared to that from deeper sediment layer. Generally, average MP varied between 0.74-158.6 pM CH4 mL-1 WW hour-1 with the highest values found at site IV. Average methanotrophic activity (MA) varied between 0.02– 31.3 nM CH4 mL-1 WW hour-1 and the highest values were found to be at the downstream localities while sediment from sites located upstream showed much lower or even negative activity. Similar to MP, values of MA were significantly higher in sediments from upper layers compared to those from deeper layers (e.g. Figs. 3c, 3d).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 409

Usually, both the surface and interstitial water were found to be supersaturated compared to the atmosphere with locality IV displaying saturation ratio *R* to be almost 195 000. This high supersaturation greatly promote diffusive fluxes of methane to the atmosphere across air-water interface and is also an important mechanisms for loss of water column

Stable carbon isotope signature of carbon dioxide (δ13C-CO2) measured in the interstitial water ranged from -19.8 ‰ to -0.8 ‰, while carbon isotope signature of methane (δ13C-CH4) ranged between -72 ‰ to -19.8 ‰. This relatively high variation in the methane isotopic values could be caused due to consequential fractionation effects preferring light carbon isotopes like methane oxidation or fractionation through diffusion and through flow of an interstitial water. Contrary, the narrow range of the δ13C-CH4 was found in the sediment depth of 40-60 cm where a high methane production has occured. Here, the δ13C-CH4 values varied only from -67.9 ‰ to -72 ‰. Apparent fractionation factor (αC) varied also greatly from 1,004 to 1,076. Usually values of αC > 1.065 and αC < 1.055 are characteristic for environments dominated by hydrogenothropic and acetoclastic methanogenesis, respectively. Our measurements indicate predominant occurrence of a hydrogenothropic methanogenesis in the high methanogenic zones where the most amount of methane is produced and δ13C of CO2 values were markedly depleted (i.e. 13C enriched). This could be caused by enhanced carbon dioxide consumption by hydrogenothrophic methanogens, strongly preferring light isotopes. Nevertheless, both acetoclastic and hydrogenotrophic pathways take part in the methanogenesis along the

**Figure 3.** Vertical distribution of methane concentration in the interstitial water at study site IV,

CH4.

longitudinal profile of the Sitka stream.

horizontal bars indicate 1 SE

**Methanogenic potential (g CH4 kg-1 DW day-1)** 

### **3.3. Methane concentration along the longitudinal profile, vertical and temporal pattern, stable isotopes**

Methane concentrations ranged between 0.18 – 35.47 µg L-1 in surface water and showed no expected trend of gradual increase from upstream localities to those laying downstream. However, significant enhancement of CH4 concentration was found on locality IV and V, respectively. Concentrations of dissolved CH4 inboth surface and interstitial waters peaked usually during summer and autumn period (Hlaváčová et al. 2005, Mach et al. in review).

Generally, methane concentrations measured in interstitial water were much higher compared to those from surface stream water and on a long-term basis ranged between 0.19 - 11 698.9 µg L-1. Due to low methane concentrations in interstitial water at localities I and II, vertical distribution of its concentrations was studied only at the downstream located sites III-V. Significant increase of the methane with the sediment depth was observed at the localities IV and V, respectively. Namely locality IV proved to be a methane pool, methane concentrations in a depth of 40 cm were found to be one order of magnitude greater than those from the depth of 20 cm (Tab. 3). Recent data from locality IV show much lower methane concentrations in the upper sediment horizons compared to those from deeper layers (Fig. 3a). Considerable lowering of methane concentration in upper sediment horizons is likely caused by oxidizing activity of methanotrophic bacteria (Fig. 3d). while dissolved oxygen concentration sharply decreased with the sediment depth (Fig. 3b).


**Table 3.** Average concentrations of methane in the vertical sediment profile at localities III-V compared to those from surface water at the same sites

Usually, both the surface and interstitial water were found to be supersaturated compared to the atmosphere with locality IV displaying saturation ratio *R* to be almost 195 000. This high supersaturation greatly promote diffusive fluxes of methane to the atmosphere across air-water interface and is also an important mechanisms for loss of water column CH4.

408 Biomass Now – Cultivation and Utilization

deeper layers (e.g. Figs. 3c, 3d).

**pattern, stable isotopes** 

2005, Mach et al. in review).

to those from surface water at the same sites

methanotrophic activity (MA) varied between 0.02– 31.3 nM CH4 mL-1 WW hour-1 and the highest values were found to be at the downstream localities while sediment from sites located upstream showed much lower or even negative activity. Similar to MP, values of MA were significantly higher in sediments from upper layers compared to those from

**3.3. Methane concentration along the longitudinal profile, vertical and temporal** 

Methane concentrations ranged between 0.18 – 35.47 µg L-1 in surface water and showed no expected trend of gradual increase from upstream localities to those laying downstream. However, significant enhancement of CH4 concentration was found on locality IV and V, respectively. Concentrations of dissolved CH4 inboth surface and interstitial waters peaked usually during summer and autumn period (Hlaváčová et al.

Generally, methane concentrations measured in interstitial water were much higher compared to those from surface stream water and on a long-term basis ranged between 0.19 - 11 698.9 µg L-1. Due to low methane concentrations in interstitial water at localities I and II, vertical distribution of its concentrations was studied only at the downstream located sites III-V. Significant increase of the methane with the sediment depth was observed at the localities IV and V, respectively. Namely locality IV proved to be a methane pool, methane concentrations in a depth of 40 cm were found to be one order of magnitude greater than those from the depth of 20 cm (Tab. 3). Recent data from locality IV show much lower methane concentrations in the upper sediment horizons compared to those from deeper layers (Fig. 3a). Considerable lowering of methane concentration in upper sediment horizons is likely caused by oxidizing activity of methanotrophic bacteria (Fig. 3d). while

dissolved oxygen concentration sharply decreased with the sediment depth (Fig. 3b).

**Locality Profile (depth) CH4 [µg L-1]** 

III. Surface water 1.8

IV. Surface water 5.52

V. Surface water 4.72

Interstitial water (depth 20cm) 1.44 Interstitial water (depth 40 cm) 1.52

Interstitial water (depth 20 cm) 1 523.9 Interstitial water (depth 40 cm) 11 390.54

Interstitial water (depth 20 cm) 6.92 Interstitial water (depth 40 cm) 24.4

**Table 3.** Average concentrations of methane in the vertical sediment profile at localities III-V compared

Stable carbon isotope signature of carbon dioxide (δ13C-CO2) measured in the interstitial water ranged from -19.8 ‰ to -0.8 ‰, while carbon isotope signature of methane (δ13C-CH4) ranged between -72 ‰ to -19.8 ‰. This relatively high variation in the methane isotopic values could be caused due to consequential fractionation effects preferring light carbon isotopes like methane oxidation or fractionation through diffusion and through flow of an interstitial water. Contrary, the narrow range of the δ13C-CH4 was found in the sediment depth of 40-60 cm where a high methane production has occured. Here, the δ13C-CH4 values varied only from -67.9 ‰ to -72 ‰. Apparent fractionation factor (αC) varied also greatly from 1,004 to 1,076. Usually values of αC > 1.065 and αC < 1.055 are characteristic for environments dominated by hydrogenothropic and acetoclastic methanogenesis, respectively. Our measurements indicate predominant occurrence of a hydrogenothropic methanogenesis in the high methanogenic zones where the most amount of methane is produced and δ13C of CO2 values were markedly depleted (i.e. 13C enriched). This could be caused by enhanced carbon dioxide consumption by hydrogenothrophic methanogens, strongly preferring light isotopes. Nevertheless, both acetoclastic and hydrogenotrophic pathways take part in the methanogenesis along the longitudinal profile of the Sitka stream.

**Figure 3.** Vertical distribution of methane concentration in the interstitial water at study site IV, horizontal bars indicate 1 SE

### **3.4. Fluxes of methane across the sediment-water and the air-water interfaces**

Methane diffusion rate from deeper sediment layers depends on a methane concentration gradient whilst is affected by oxidation and rate of methanotrophic bacteria consumption. When diffusion fluxes are positive (positive values indicate net CH4 production), then surface water is enriched by methane which in turn may be a part of downstream transport or is further emitted to the atmosphere (Fig. 4).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 411

deeper sediments upward is either transported by advection through sediments downstream or is probably almost completely oxidized by methanotrophic bacteria due to increasing oxygen supply from the surface stream. As a consequence, very low or no benthic fluxes were recorded during the time of high flow discharge. Compared to calculated diffusive fluxes it is clear that fluxes obtained by direct measurement were approximately 15× higher than the fluxes calculated with using Fick´s first law. Thus, direct benthic fluxes were used for a

Gaseous fluxes from surface water to the atmosphere were found at all localities except locality I, where emissions were not mesured directly but were calculated lately using a known relationships between concentrations of gases in surface water and their emissions to the atmosphere found at downstream laying localities II-V. Methane showed an increase in emissions toward downstream where highest surface water concentrations have also occured (Table 4). Methane emissions measured at localities II-V ranged from 0 – 167.35 mg m-2 day-1 and no gradual increase in downstream end was found in spite of our expectation. However, sharp increase in the amount of methane emitted from the surface water was measured at lowermost localities IV and V (Tab. 4). We found positive, but weak correlation between surface water methane concentrations and measured emissions (rs = 0.45, p <

> **Locality/Gas CH4 [mg m-2day-1]** Locality I. 2.39

Locality II. 0.25 (0 – 0.6) *n* = 9

locality III. 1.3 (0 – 5.01) *n* = 10

Locality IV. 32.1 (7.3 – 87.9) *n* = 8

Locality V. 36.3 (2.8 – 167.4) *n* = 12 **Table 4.** Average emissions to the atmosphere and their range in parenthesis and from all localities e*x*cept locality I. Emissions values for the locality I were calculated using a known relationships between concentrations of methane gas in surface water and its emissions to the atmosphere found at

Depending on the time of year we measured the emissions, values of *E*a ranged from 430 to 925 kg year-1 for methane. Annually, approximately 0.7 tonne of methane was emitted to the atmosphere from the water level of the Sitka stream (total area ca 0.2 km2). The majority of annual methane emissions (90 %) occured in the lower 7 km of the stream (stretch IV and V) that represents only 1/5 of the total stream area. In addition, contribution of methane emissions to the total annual emissions was found to be the highest during spring-summer

calculation of water column CH4 budget.

downstream laying localities II-V. *n* means sample size

**3.5. Whole-stream emissions Ea** 

period (Mach et al. in review).

0.05)(Fig. 5).

**Figure 4.** Possible fate of the methane within hyporheic zone and two kinds of chambers for measurement of methane fluxes. Providing that some sites along the longitudinal stream profile should be sources of methane for the stream water, we chose locality IV to be suitable for benthic fluxes measurements.

On the contrary, when the fluxes of methane across the sediment-water interface are negative then all methane produced in the sediments is likely oxidized and consumed by methanotrophic bacteria here or transported via subsurface hyporheic flow.

Calculated diffusive fluxes of CH4 ranged from 0.03 to 2307.32 µg m-2 day-1 along the longitudinal profile. The lowest average values of diffusive fluxes were observed at study site II (0.11 ± 0.05 µg m-2 day-1) while the highest average values were those observed at study site IV (885.81 ± 697.54 µg m-2 day-1). Direct benthic fluxes of CH4 using the benthic chambers were measured at study site IV only and ranged from 0.19 to 82.17 mg m-2 day-1. We observed clear negative relationships between benthic methane fluxes and the flow discharge. During higher discharges when the stream water is pushed into sediments, methane diffusing from deeper sediments upward is either transported by advection through sediments downstream or is probably almost completely oxidized by methanotrophic bacteria due to increasing oxygen supply from the surface stream. As a consequence, very low or no benthic fluxes were recorded during the time of high flow discharge. Compared to calculated diffusive fluxes it is clear that fluxes obtained by direct measurement were approximately 15× higher than the fluxes calculated with using Fick´s first law. Thus, direct benthic fluxes were used for a calculation of water column CH4 budget.

Gaseous fluxes from surface water to the atmosphere were found at all localities except locality I, where emissions were not mesured directly but were calculated lately using a known relationships between concentrations of gases in surface water and their emissions to the atmosphere found at downstream laying localities II-V. Methane showed an increase in emissions toward downstream where highest surface water concentrations have also occured (Table 4). Methane emissions measured at localities II-V ranged from 0 – 167.35 mg m-2 day-1 and no gradual increase in downstream end was found in spite of our expectation. However, sharp increase in the amount of methane emitted from the surface water was measured at lowermost localities IV and V (Tab. 4). We found positive, but weak correlation between surface water methane concentrations and measured emissions (rs = 0.45, p < 0.05)(Fig. 5).


**Table 4.** Average emissions to the atmosphere and their range in parenthesis and from all localities e*x*cept locality I. Emissions values for the locality I were calculated using a known relationships between concentrations of methane gas in surface water and its emissions to the atmosphere found at downstream laying localities II-V. *n* means sample size

### **3.5. Whole-stream emissions Ea**

410 Biomass Now – Cultivation and Utilization

measurements.

or is further emitted to the atmosphere (Fig. 4).

**3.4. Fluxes of methane across the sediment-water and the air-water interfaces** 

**Figure 4.** Possible fate of the methane within hyporheic zone and two kinds of chambers for

methanotrophic bacteria here or transported via subsurface hyporheic flow.

measurement of methane fluxes. Providing that some sites along the longitudinal stream profile should be sources of methane for the stream water, we chose locality IV to be suitable for benthic fluxes

On the contrary, when the fluxes of methane across the sediment-water interface are negative then all methane produced in the sediments is likely oxidized and consumed by

Calculated diffusive fluxes of CH4 ranged from 0.03 to 2307.32 µg m-2 day-1 along the longitudinal profile. The lowest average values of diffusive fluxes were observed at study site II (0.11 ± 0.05 µg m-2 day-1) while the highest average values were those observed at study site IV (885.81 ± 697.54 µg m-2 day-1). Direct benthic fluxes of CH4 using the benthic chambers were measured at study site IV only and ranged from 0.19 to 82.17 mg m-2 day-1. We observed clear negative relationships between benthic methane fluxes and the flow discharge. During higher discharges when the stream water is pushed into sediments, methane diffusing from

Methane diffusion rate from deeper sediment layers depends on a methane concentration gradient whilst is affected by oxidation and rate of methanotrophic bacteria consumption. When diffusion fluxes are positive (positive values indicate net CH4 production), then surface water is enriched by methane which in turn may be a part of downstream transport

> Depending on the time of year we measured the emissions, values of *E*a ranged from 430 to 925 kg year-1 for methane. Annually, approximately 0.7 tonne of methane was emitted to the atmosphere from the water level of the Sitka stream (total area ca 0.2 km2). The majority of annual methane emissions (90 %) occured in the lower 7 km of the stream (stretch IV and V) that represents only 1/5 of the total stream area. In addition, contribution of methane emissions to the total annual emissions was found to be the highest during spring-summer period (Mach et al. in review).

**Figure 5.** Relationships between atmospheric emissions and surface water concentrations of the methane. Each point represents the mean of five replicate emission measurements and the two replicates of stream water methane concentrations at all

### **3.6. Sitka stream water column CH4 budget for the experimental stretch of a stream**

The potentially important source and sinks terms for dissolved methane in the water column of the Sitka stream are shown in Figure 6. Previously calculated rates of inputs (benthic fluxes) and loss of dissolved CH4 through evasion to the atmosphere can be combined together with advection inputs and losses to yield a CH4 dynamics (budget) for any particular section of the stream.

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 413

with an area being 200 m2. Positive fluxes of CH4 were found to occur at 30.9 % of the study area. Assuming that average benthic flux of methane across the sediment-water interface was 15.40 mg m-2 day-1, the benthic flux of 3081.39 mg CH4 day-1 should occur from the whole area of 200 m2. Average emission flux of CH4 across the water-air interface for all study sites was determined to be 14.47 ± 4.73 mg CH4 m-2 day-1. This value is slightly lower than the direct benthic flux of CH4 and suggests that some portion of methane released from the bottom sediments may contribute to increasing concentration of CH4 in the surface water. Average flow of the Sitka stream during time of benthic fluxes measurements was 0.351 m3s-1 (i.e. 351 L-1s-1). Therefore, we may expect that water column was enriched at least by 187.4 mg (i.e. 0.006 µg L-1) of CH4 from sediment at 45 m long section near study site IV during one day. Next study site V is located some 4 km downstream from the site IV. Average CH4 concentration difference in the stream water between these study sites was found to be 3.2 µg L-1 of CH4 indicating that CH4 supply exceeds slightly CH4 removal. Methane fluxes from the sediment would contribute to this concentration difference only by 0.6 µg L-1, thus, the immediate difference in the CH4 budget found between two studied sites IV and V indicates that there must likely be other sources of methane supply to the stream water (Fig. 7). This "missing source" seems to be relatively small (0.9 mg CH4 0.351 m-3s-1), however, net accumulation of CH4 in the stream water during 4 km section of the Sitka

**Figure 7.** CH4 budget in mol day-1 for a section of the Sitka stream between study sites IV and V (lenght

Both methanogenic archaea and aerobic methanotrophs were found at all localities along the longitudinal stream profile. The proportion of these groups to the DAPI-stained cells was quite consistent and varied only slightly but a higher proportion to the DAPI-stained cells in deeper sediment layer 25-50 cm was observed. On average 23,4 % of DAPI-stained cells were detected by FISH with a probe for methanogens while type I methanotrophs reached 21,4 % and type II methanotrophs 11,9 %, respectively. All three groups also revealed nonsignificant higher proportion to the TCN in deeper sediment layer; the abundance of

stream below study site IV was almost 78 g CH4 per one day.

ca 4 km). The arrows correspond to those depicted in Figure 6.

**3.7. Fluorescence in situ hybridization (FISH)** 

**Figure 6.** Simple box model used to calculate a CH4 budget for the Sitke stream experimental section; advection in + supply = advection out + removal (box adjusted after de Angelis & Scranton 1993)

The CH4 budget determined for the 2011 sampling period in an experimental stream section is summarized in Figure 7. Benthic fluxes were measured along a stream section 45 m long with an area being 200 m2. Positive fluxes of CH4 were found to occur at 30.9 % of the study area. Assuming that average benthic flux of methane across the sediment-water interface was 15.40 mg m-2 day-1, the benthic flux of 3081.39 mg CH4 day-1 should occur from the whole area of 200 m2. Average emission flux of CH4 across the water-air interface for all study sites was determined to be 14.47 ± 4.73 mg CH4 m-2 day-1. This value is slightly lower than the direct benthic flux of CH4 and suggests that some portion of methane released from the bottom sediments may contribute to increasing concentration of CH4 in the surface water. Average flow of the Sitka stream during time of benthic fluxes measurements was 0.351 m3s-1 (i.e. 351 L-1s-1). Therefore, we may expect that water column was enriched at least by 187.4 mg (i.e. 0.006 µg L-1) of CH4 from sediment at 45 m long section near study site IV during one day. Next study site V is located some 4 km downstream from the site IV. Average CH4 concentration difference in the stream water between these study sites was found to be 3.2 µg L-1 of CH4 indicating that CH4 supply exceeds slightly CH4 removal. Methane fluxes from the sediment would contribute to this concentration difference only by 0.6 µg L-1, thus, the immediate difference in the CH4 budget found between two studied sites IV and V indicates that there must likely be other sources of methane supply to the stream water (Fig. 7). This "missing source" seems to be relatively small (0.9 mg CH4 0.351 m-3s-1), however, net accumulation of CH4 in the stream water during 4 km section of the Sitka stream below study site IV was almost 78 g CH4 per one day.

**Figure 7.** CH4 budget in mol day-1 for a section of the Sitka stream between study sites IV and V (lenght ca 4 km). The arrows correspond to those depicted in Figure 6.

### **3.7. Fluorescence in situ hybridization (FISH)**

412 Biomass Now – Cultivation and Utilization

**Figure 5.** Relationships between atmospheric emissions and surface water concentrations of the methane. Each point represents the mean of five replicate emission measurements and the two

Water concentration [g CH4 L-1]

**3.6. Sitka stream water column CH4 budget for the experimental stretch of a** 

The potentially important source and sinks terms for dissolved methane in the water column of the Sitka stream are shown in Figure 6. Previously calculated rates of inputs (benthic fluxes) and loss of dissolved CH4 through evasion to the atmosphere can be combined together with advection inputs and losses to yield a CH4 dynamics (budget) for

**Figure 6.** Simple box model used to calculate a CH4 budget for the Sitke stream experimental section; advection in + supply = advection out + removal (box adjusted after de Angelis & Scranton 1993)

The CH4 budget determined for the 2011 sampling period in an experimental stream section is summarized in Figure 7. Benthic fluxes were measured along a stream section 45 m long

replicates of stream water methane concentrations at all

any particular section of the stream.

**stream** 

Both methanogenic archaea and aerobic methanotrophs were found at all localities along the longitudinal stream profile. The proportion of these groups to the DAPI-stained cells was quite consistent and varied only slightly but a higher proportion to the DAPI-stained cells in deeper sediment layer 25-50 cm was observed. On average 23,4 % of DAPI-stained cells were detected by FISH with a probe for methanogens while type I methanotrophs reached 21,4 % and type II methanotrophs 11,9 %, respectively. All three groups also revealed nonsignificant higher proportion to the TCN in deeper sediment layer; the abundance of methanogens and methanotrophs remained almost unchanged with increasing sediment depth. The average abundance of methanogens (0,88 *±* 0,28 and 1,07 *±* 0,23 x 106 cells mL-1 in the upper and deeper layer, respectively) and type II methanotrophs (0,44 ± 0,14 x 106 cells mL-1 and 0,56*±* 0,1 x 106 cells mL-1) increased slightly with the sediment depth , while type I methanotrophs revealed average abundance 0,98 *±* 0,23 x 106 cells mL-1 in the deeper layer being lower compared to abundance 1,07*±* 0,28 x 106 cells mL-1 found in upper sediment layer (Buriánková et al. 2012). Very recently, however, using the FISH method we found that abundance of methanogens belonging to three selected families reached their maximum in the sediment depth of 20-30 cm and had closely reflected vertical distribution of acetate concentrations. Species of family *Methanobacteriaceae* grow only with hydrogen, formate and alcohols (except methanol), *Methanosarcinaceae* can grow with all methanogenic substrates except formate, and members of *Methanosaetaceae* grow ecxlusively with acetate as energy source. All three families also showed similar proportion to the DAPI stained cells, ranging in average (depth 10-50 cm) from 9.9% (*Methanosarcinaceae*) to 12.3% (*Methanobacteriaceae*) (Fig. 8).

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 415

**3.8. Denaturing gradient gel electrophoresis and cloning** 

ferrous iron compared to other localities (cf. Table 2).

bands found at different sediment depths

Methanogenic communities associated with hyporheic sediments at two different depths (0- 25 cm and 25-50 cm) along the longitudinal stream profile were compared based on the DGGE patterns. As shown in Fig. 9, the DGGE patterns varied highly among study localities (Fig. 9A), irrespective of the depth (Fig. 9B). However, presence of the bands in all samples indicates that methanogens may occur up to 50 cm of the sediment depth. The number of DGGE bands of the methanogenic archaeal communities was compared either among localites or among different sediment depths. A total of 22 different bands were observed in the DGGE image ranging from 4 (locality II) to 16 (locality IV) in the samples (Fig. 9A).

The number of DGGE bands also ranged from 2 to 10 for the samples from upper layer (0-25 cm) and from 2 to 11 for the samples from deeper layer (25-50 cm), respectively (Fig. 9B). We found no clear trend in the number of DGGE bands with increasing depth (Fig. 9B). Locality IV appears to be the richest in number of DGGE bands. We suppose that this might be due to most favorable conditions prevailing for the methanogens life as indicated by a relatively low grain median size, lower dissolved oxygen concentration or higher concentration of the

The methanogenic community diversity in hyporheic sediment of Sitka stream was also analysed by PCR amplification, cloning and sequencing of methyl coenzyme M reductase (*mcrA*) gene. A total of 60 *mcrA* gene sequences revealed 26 different *mcrA* gene clones.

**Figure 9.** Number of DGGE bands associated with hyporheic sediments at two different depths along the longitudinal stream profile. A – Total number of all bands detected at each locality; B – number of

Most of the clones showed low affiliation with known species (< 97% nucleotide identity) and probably represented genes of novel methanogenic archeal genera/species, but all of them were closely related to uncultured methanogens from environmental samples (> 97% similarity) retrieved from BLAST. The 25 clones were clustered to four groups and were confirmed to be affiliated to *Methanosarcinales, Methanomicrobiales* and *Methanobacteriales*  orders and other unclassified methanogens. The members of all three orders and novel methanogenic cluster were detected to occur in a whole bottom sediment irrespective of a depth, nevertheless, the richness of methanogenic archaea in the sediment was slightly

**Figure 8.** The percentage of chosen methanogenic families as compared to the total bacterial cell numbers found in different sediment layers at locality no. IV, horizontal bars indicate 1 SE

### **3.8. Denaturing gradient gel electrophoresis and cloning**

414 Biomass Now – Cultivation and Utilization

(Fig. 8).

methanogens and methanotrophs remained almost unchanged with increasing sediment depth. The average abundance of methanogens (0,88 *±* 0,28 and 1,07 *±* 0,23 x 106 cells mL-1 in the upper and deeper layer, respectively) and type II methanotrophs (0,44 ± 0,14 x 106 cells mL-1 and 0,56*±* 0,1 x 106 cells mL-1) increased slightly with the sediment depth , while type I methanotrophs revealed average abundance 0,98 *±* 0,23 x 106 cells mL-1 in the deeper layer being lower compared to abundance 1,07*±* 0,28 x 106 cells mL-1 found in upper sediment layer (Buriánková et al. 2012). Very recently, however, using the FISH method we found that abundance of methanogens belonging to three selected families reached their maximum in the sediment depth of 20-30 cm and had closely reflected vertical distribution of acetate concentrations. Species of family *Methanobacteriaceae* grow only with hydrogen, formate and alcohols (except methanol), *Methanosarcinaceae* can grow with all methanogenic substrates except formate, and members of *Methanosaetaceae* grow ecxlusively with acetate as energy source. All three families also showed similar proportion to the DAPI stained cells, ranging in average (depth 10-50 cm) from 9.9% (*Methanosarcinaceae*) to 12.3% (*Methanobacteriaceae*)

**Figure 8.** The percentage of chosen methanogenic families as compared to the total bacterial cell numbers found in different sediment layers at locality no. IV, horizontal bars indicate 1 SE

Methanogenic communities associated with hyporheic sediments at two different depths (0- 25 cm and 25-50 cm) along the longitudinal stream profile were compared based on the DGGE patterns. As shown in Fig. 9, the DGGE patterns varied highly among study localities (Fig. 9A), irrespective of the depth (Fig. 9B). However, presence of the bands in all samples indicates that methanogens may occur up to 50 cm of the sediment depth. The number of DGGE bands of the methanogenic archaeal communities was compared either among localites or among different sediment depths. A total of 22 different bands were observed in the DGGE image ranging from 4 (locality II) to 16 (locality IV) in the samples (Fig. 9A).

The number of DGGE bands also ranged from 2 to 10 for the samples from upper layer (0-25 cm) and from 2 to 11 for the samples from deeper layer (25-50 cm), respectively (Fig. 9B). We found no clear trend in the number of DGGE bands with increasing depth (Fig. 9B). Locality IV appears to be the richest in number of DGGE bands. We suppose that this might be due to most favorable conditions prevailing for the methanogens life as indicated by a relatively low grain median size, lower dissolved oxygen concentration or higher concentration of the ferrous iron compared to other localities (cf. Table 2).

The methanogenic community diversity in hyporheic sediment of Sitka stream was also analysed by PCR amplification, cloning and sequencing of methyl coenzyme M reductase (*mcrA*) gene. A total of 60 *mcrA* gene sequences revealed 26 different *mcrA* gene clones.

**Figure 9.** Number of DGGE bands associated with hyporheic sediments at two different depths along the longitudinal stream profile. A – Total number of all bands detected at each locality; B – number of bands found at different sediment depths

Most of the clones showed low affiliation with known species (< 97% nucleotide identity) and probably represented genes of novel methanogenic archeal genera/species, but all of them were closely related to uncultured methanogens from environmental samples (> 97% similarity) retrieved from BLAST. The 25 clones were clustered to four groups and were confirmed to be affiliated to *Methanosarcinales, Methanomicrobiales* and *Methanobacteriales*  orders and other unclassified methanogens. The members of all three orders and novel methanogenic cluster were detected to occur in a whole bottom sediment irrespective of a depth, nevertheless, the richness of methanogenic archaea in the sediment was slightly higher in the upper sediment layer 0-25 cm (15 clones) than in the deeper sediment layer 25- 50 cm (11 clones)(Buriánková et al. in review). The clones affiliated with *Methanomicrobiales* predominated in the deeper layer while *Methanosarcinales* clones dominated in the upper sediment layer. This prevalence of *Methanosarcinales* in the upper sediment layer was also confirmed by our FISH analyses as has been mentioned above.

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 417

methane, being usually in the range -27 ‰ up to -100 ‰ (Conrad 2004; Michener & Lajtha 2007). Whiticar et al. (1986) demonstrated that methane in freshwater sediments is isotopically distinguished by being relatively enriched in 13C (δ13C = -65 to -50‰) in contrast to marine sediments (-110 to -60‰). Accordingly, the two precursors of methane, namely acetate and CO2/H2, yield methane with markedly different δ13C values; methane from acetate is relatively enriched in 13C. Average minimum in the carbon isotopic composition of CH4 (-61.4 ‰) occurred deeper in sediments (60 cm) while average maximum in δ13C-CH4 occured in the lower sediment depth of 30 cm. Enrichment of 13C in CH4 probably reflects aerobic CH4 oxidation because oxidation would result in residual CH4 with δ13C-CH4 values less negative than the source CH4 (Barker & Fritz 1981; Chanton et al. 2004). However, this

Our working hypotesis suggested that along with the longitudinal profile of a stream, slope and flow conditions also change together with corresponding settling velocity, sediment composition and organic matter content. Thus, according to this prediction, sediment with prevalence of fine-grained particles containg higher amount of organic matter should dominate at the downstream stretches. Moreover, due to prevalence of anoxic environment, production of methane and its emissions was expected to be also higher here compared to that from upstream stretches. Based on our findings, it seems that this presumption is valid for the methane. In addition, we found higher methane concentrations in both the surface and interstitial water at the uppermost locality I compared to lower situated locality II. Similar situation with high methane concentration in the upstream part with subsequent decline further downstream was also reported from USA by Lilley et al. (1996). Dissimilarity of this first stretch is apparent in a comparison with the next, downstream laying stretch (locality II), represented by profile with steep valley and high slope. Generally, there were found very low methane concentrations either in surface or interstitial water and fluxes of

Flux rates of gaseous emissions into atmosphere depend on partial pressure of particular gas in the atmosphere and its concentration in a water, water temperature and further on the water depth and flow velocity. Thus, maximum peak of emissions may be expected during summer period and in well torrential stretch of the river. Silvennoinen et al. (2008), for example, found that the most upstream river site, surrounded by forests and drained peatlands, released significant amounts of CO2 and CH4. The downstream river sites surrounded by agricultural soils released significant amounts of N2O whereas the CO2 and CH4 concentrations were low compared to the upstream site. When consider seasonal distribution of methane emissions, it is clear, in concordance with above mentioned presumption, that majority of methane emissions was relesed during a warm period of the year (81%). Effect of temperature on methane production was also observed in southeastern USA where the most methane reased to the atmosphere during warm months (Pulliam 1993). In addition, close correlation between methane emissions and temperature was reported also from south part of Baltic Sea; the temperature has been found to be a key

effect has been observed only at the study site IV.

emissions to atmosphere were also very low.

factor driving methane emissions (Heyer & Berger 2000).

**4.3. Spatial and temporal distributions of emissions** 

### **4. Discussion**

### **4.1. Occurence of methane in stream water and sediments**

In spite of commonly held view of streams as well-oxygenated habitats, we found both surface and interstitial water to be supersaturated with methane compared to the atmosphere at all five localities (Mach et al. in review). Availability of interstitial habitats for bacteria and archaea carrying out anaerobic processes has been confirmed by our previous (Hlaváčová et al. 2005, 2006; Cupalová & Rulík 2007) and contemporary findings. During this study we found relatively well developed populations of methanogenic archaea at all localities and that all localities also showed positive methanogenic potential. Emissions of methane from water ecosystems results from complex microbial activity in the carbon cycle (production and consumption processes), which depends upon a large number of environmental parameters such as availability of carbon and terminal electron acceptors, flow velocity and turbulence, water depth. In our previous paper (Hlaváčová et al. 2006), we suggested that surface water concentrations, and as a consequence methane gas emissions to the atmosphere would result from downstream transport of gases by stream water (advection in/out), and moreover, from autochthonous microbial metabolism within the hyporheic zone. If so, surface water is continually saturated by gases produced by hyporheic metabolism, leading to supersaturation of surface water and induced diffusion of these gases out of river water (volatizing). Moreover, the run-off and drainage of adjacent soils can also contribute greatly to the degree of greenhouse gas supersaturation (De Angelis & Lilley 1987, Kroeze & Seitzinger 1998, Worral & Lancaster 2005, Wilcock & Sorrell 2008). For example, CH4 in the estuarine waters may come from microbial production in water, sediment release, riverine input and inputs of methane-rich water from surrounding anoxic environments (Zhang et al. 2008b). For the European estuaries, riverine input contribute much to the estuarine CH4 due to high CH4 in the river waters and wetlands also play important roles. However, low CH4 in the Changjiang Estaury (China) may be resulted from the low CH4 in the Changjiang water together with the low net microbial production and low input from adjacent salt marshes (Zhang et al. 2008b). Dissolved methane concentrations in a surface water of Sitka stream is consistent with literature data on methane in rivers published by Middelburg et al. (2002) and Zhang et al. (2008b).

### **4.2. Stable carbon isotopes**

A knowledge of the stable carbon isotopic ratio of methane δ13C-CH4 in natural systems can be useful in studies of the mechanisms and pathways of CH4 cycling (Sansone et al. 1997). Values of carbon isotope signature of methane (δ13C-CH4) indicate biogenic nature of the methane, being usually in the range -27 ‰ up to -100 ‰ (Conrad 2004; Michener & Lajtha 2007). Whiticar et al. (1986) demonstrated that methane in freshwater sediments is isotopically distinguished by being relatively enriched in 13C (δ13C = -65 to -50‰) in contrast to marine sediments (-110 to -60‰). Accordingly, the two precursors of methane, namely acetate and CO2/H2, yield methane with markedly different δ13C values; methane from acetate is relatively enriched in 13C. Average minimum in the carbon isotopic composition of CH4 (-61.4 ‰) occurred deeper in sediments (60 cm) while average maximum in δ13C-CH4 occured in the lower sediment depth of 30 cm. Enrichment of 13C in CH4 probably reflects aerobic CH4 oxidation because oxidation would result in residual CH4 with δ13C-CH4 values less negative than the source CH4 (Barker & Fritz 1981; Chanton et al. 2004). However, this effect has been observed only at the study site IV.

### **4.3. Spatial and temporal distributions of emissions**

416 Biomass Now – Cultivation and Utilization

**4.2. Stable carbon isotopes** 

**4. Discussion** 

higher in the upper sediment layer 0-25 cm (15 clones) than in the deeper sediment layer 25- 50 cm (11 clones)(Buriánková et al. in review). The clones affiliated with *Methanomicrobiales* predominated in the deeper layer while *Methanosarcinales* clones dominated in the upper sediment layer. This prevalence of *Methanosarcinales* in the upper sediment layer was also

In spite of commonly held view of streams as well-oxygenated habitats, we found both surface and interstitial water to be supersaturated with methane compared to the atmosphere at all five localities (Mach et al. in review). Availability of interstitial habitats for bacteria and archaea carrying out anaerobic processes has been confirmed by our previous (Hlaváčová et al. 2005, 2006; Cupalová & Rulík 2007) and contemporary findings. During this study we found relatively well developed populations of methanogenic archaea at all localities and that all localities also showed positive methanogenic potential. Emissions of methane from water ecosystems results from complex microbial activity in the carbon cycle (production and consumption processes), which depends upon a large number of environmental parameters such as availability of carbon and terminal electron acceptors, flow velocity and turbulence, water depth. In our previous paper (Hlaváčová et al. 2006), we suggested that surface water concentrations, and as a consequence methane gas emissions to the atmosphere would result from downstream transport of gases by stream water (advection in/out), and moreover, from autochthonous microbial metabolism within the hyporheic zone. If so, surface water is continually saturated by gases produced by hyporheic metabolism, leading to supersaturation of surface water and induced diffusion of these gases out of river water (volatizing). Moreover, the run-off and drainage of adjacent soils can also contribute greatly to the degree of greenhouse gas supersaturation (De Angelis & Lilley 1987, Kroeze & Seitzinger 1998, Worral & Lancaster 2005, Wilcock & Sorrell 2008). For example, CH4 in the estuarine waters may come from microbial production in water, sediment release, riverine input and inputs of methane-rich water from surrounding anoxic environments (Zhang et al. 2008b). For the European estuaries, riverine input contribute much to the estuarine CH4 due to high CH4 in the river waters and wetlands also play important roles. However, low CH4 in the Changjiang Estaury (China) may be resulted from the low CH4 in the Changjiang water together with the low net microbial production and low input from adjacent salt marshes (Zhang et al. 2008b). Dissolved methane concentrations in a surface water of Sitka stream is consistent with literature data on

methane in rivers published by Middelburg et al. (2002) and Zhang et al. (2008b).

A knowledge of the stable carbon isotopic ratio of methane δ13C-CH4 in natural systems can be useful in studies of the mechanisms and pathways of CH4 cycling (Sansone et al. 1997). Values of carbon isotope signature of methane (δ13C-CH4) indicate biogenic nature of the

confirmed by our FISH analyses as has been mentioned above.

**4.1. Occurence of methane in stream water and sediments** 

Our working hypotesis suggested that along with the longitudinal profile of a stream, slope and flow conditions also change together with corresponding settling velocity, sediment composition and organic matter content. Thus, according to this prediction, sediment with prevalence of fine-grained particles containg higher amount of organic matter should dominate at the downstream stretches. Moreover, due to prevalence of anoxic environment, production of methane and its emissions was expected to be also higher here compared to that from upstream stretches. Based on our findings, it seems that this presumption is valid for the methane. In addition, we found higher methane concentrations in both the surface and interstitial water at the uppermost locality I compared to lower situated locality II. Similar situation with high methane concentration in the upstream part with subsequent decline further downstream was also reported from USA by Lilley et al. (1996). Dissimilarity of this first stretch is apparent in a comparison with the next, downstream laying stretch (locality II), represented by profile with steep valley and high slope. Generally, there were found very low methane concentrations either in surface or interstitial water and fluxes of emissions to atmosphere were also very low.

Flux rates of gaseous emissions into atmosphere depend on partial pressure of particular gas in the atmosphere and its concentration in a water, water temperature and further on the water depth and flow velocity. Thus, maximum peak of emissions may be expected during summer period and in well torrential stretch of the river. Silvennoinen et al. (2008), for example, found that the most upstream river site, surrounded by forests and drained peatlands, released significant amounts of CO2 and CH4. The downstream river sites surrounded by agricultural soils released significant amounts of N2O whereas the CO2 and CH4 concentrations were low compared to the upstream site. When consider seasonal distribution of methane emissions, it is clear, in concordance with above mentioned presumption, that majority of methane emissions was relesed during a warm period of the year (81%). Effect of temperature on methane production was also observed in southeastern USA where the most methane reased to the atmosphere during warm months (Pulliam 1993). In addition, close correlation between methane emissions and temperature was reported also from south part of Baltic Sea; the temperature has been found to be a key factor driving methane emissions (Heyer & Berger 2000).

These findings also indicate that we should be very carefull in making any generalization in total emissions estimation for any given stream or river. Even though some predictions can be made based on gas concentrations measured in the surface or interstitial water, results may be very different. From this point, noteworthy was locality IV; enormous concentrations of a methane found in the deep interstitial water were caused probably by very fine, clayed sediment containing high amount of organic carbon, as well as high DOC concentrations. Supersaturation led also to the enrichment of the surface water with methane - such places may be considered as very important methane sources for surface stream and, consequently source of emissons to the atmosphere.

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 419

The presence of relatively rich assemblage of methanogenic archaea in hyporheic river sediments is rather surprising, however it is in accordance with other studies. The number of total different bands (i.e. estimated diversity of the methanoges) observed in the DGGE patterns of the methanogenic archaeal communities was comparable with a number of the DGGE bands found in other studies. For example, Ikenaga et al. (2004) in their study of methanogenic archaeal community in rice roots found 15-19 DGGE bands, while Watanabe et al. (2010) showed 27 bands at different positiosns in the DGGE band pattern obtained from Japanese paddy field soils. Our results from the DGGE analysis are supported by cloning and sequencing of methyl coenzyme M reductase (*mcrA*) gene which also retrieved relatively rich diversity (25 different *mcrA* gene clones) of the methanogenic community in the Sitka stream hyporheic sediments. Similar richness in number of clones was also mentioned in a methanogenic community in Zoige wetland, where 21 different clones were found (Zhang et al. 2008a), while 20 clones were described in the methane cycle of a meromictic lake in France (Biderre-Petit et al. 2011). In addition, soils from Ljubljana marsh (Slovenia) showed 17 clones (Jerman et al. 2009), for example. Both DGGE and *mcrA* gene sequencing results suggest that both hydrogenotrophic and acetoclastic methanogenesis are an integral part of the CH4 - producing pathway in the hyporheic zone and were represented by appropriate methanogenic populations. Further, these methanogenic archaea form important component of a hyporheic microbial community and may

To our knowledge this study is the first analysis of the composition of active methanogenic/methanotrophic communities in river hyporheic sediments. By use of various molecular methods we have shown that both methanogenic archaea and aerobic methanotrophs can be quantitatively dominant components of hyporheic biofilm community and may affect CH4 cycling in river sediments. Their distribution within hyporheic sediments, however, only partly reflects potential methane production and consumption rates of the sediments. Rather surprising is the detection of methanotrophs in the deep sediment layer 25-50 cm, indicating that suitable conditions for methane oxidation occur here. In addition, this work constitutes the first estimation of sources, sinks and fluxes of CH4 in the Sitka stream and in 3rd order stream environment. Fluxes of CH4 from supersaturated interstitial sediments appear to be a main CH4 source toward the water column. Compared with CH4 production rates, the diffusive fluxes are very low due to efficient aerobic oxidation by methanotrophic bacteria, especially during higher flow discharges. Although fluxes to the atmosphere from the Sitka stream seems to be insignificant, they are comparable or higher in comparison with fluxes from other aquatic ecosystems, especially those measured in running waters. Finally, our results suggest that the Sitka Stream is a source of methane into the atmosphere, and loss of carbon via the fluxes of this greenhouse gas out into the ecosystem can participate significantly in river

substantially affect CH4 cycling in the Sitka stream sediments.

**4.5. Methanogens diversity** 

**5. Conclusion** 

self-purification.

### **4.4. Benthic fluxes and potential methane oxidation**

CH4 can be produced and released into overlying near-bottom water through exchange at sediment-water interface. Methane released from the sediments into the overlying water column can be consumed by methanotrophs. Methanotrophs can oxidize as much as 100 % of methane production (Le Mer & Roger 2001). According to the season, 13-70 % of methane was consumed in a Hudson River water column (de Angelis et Scranton 1993). For the Sitka stream, measurement of benthic fluxes into the overlying surface waters indicates that methane consumption by methanotrophic bacteria is likely a dominant way of a methane loss, nevertheless some methane still supports relatively high average methane concentrations in the surface water and, in turn, high emissions to the atmosphere.

The methane production (measured as methanogenic potential) was found to be 3 orders of magnitude lower than the oxidation (methanotrophic activity), thus, almost all methane should be oxidized and consumed by methanotrophic bacteria and no methane would occur within the sediments. However, situation seems to be quite different suggesting that namely methanotrophic activity measured in a laboratory could be overestimated. Since oxidation of methane requires both available methane and oxygen, methanotrophic activity is expected to be high at sites where both methane and dissolved oxygen are available. Therefore, high values of the MA were usually found in the upper layers of the sediments (Segers 1998) or at interface between oxic and anoxic zones, respectively. Relatively high methanotrophic activity found in deeper sediments of the localities III-V indicates that methane oxidation is not restricted only to the surface sediments as is common in lakes but it also takes place at greater depths. It seems likely that oxic zone occurs in a vertical profile of the sediments and that methane diffusing from the deeper layer into the sedimentary aerobic zone is being oxidized by methanotrophs here. Increased methanotrophic activity at this hyporheic oxic-anoxic interface is probably evident also from higher abundance of type II methanotrophs in the same depth layer. Similar pathway of methane cycling has been observed by Kuivila et al. (1988) in well oxygenated sediments of Lake Washington, however, methane oxidation within the sediments would be rather normal in river sediments compared to lakes. All the above mentioned findings support our previous suggestions that coexistence of various metabolic processes in hyporheic sediments is common due to vertical and horizontal mixing of the interstitial water and occurrence of microbial biofilm (Hlaváčová et al. 2005, 2006).

### **4.5. Methanogens diversity**

418 Biomass Now – Cultivation and Utilization

These findings also indicate that we should be very carefull in making any generalization in total emissions estimation for any given stream or river. Even though some predictions can be made based on gas concentrations measured in the surface or interstitial water, results may be very different. From this point, noteworthy was locality IV; enormous concentrations of a methane found in the deep interstitial water were caused probably by very fine, clayed sediment containing high amount of organic carbon, as well as high DOC concentrations. Supersaturation led also to the enrichment of the surface water with methane - such places may be considered as very important methane sources for surface

CH4 can be produced and released into overlying near-bottom water through exchange at sediment-water interface. Methane released from the sediments into the overlying water column can be consumed by methanotrophs. Methanotrophs can oxidize as much as 100 % of methane production (Le Mer & Roger 2001). According to the season, 13-70 % of methane was consumed in a Hudson River water column (de Angelis et Scranton 1993). For the Sitka stream, measurement of benthic fluxes into the overlying surface waters indicates that methane consumption by methanotrophic bacteria is likely a dominant way of a methane loss, nevertheless some methane still supports relatively high average methane

The methane production (measured as methanogenic potential) was found to be 3 orders of magnitude lower than the oxidation (methanotrophic activity), thus, almost all methane should be oxidized and consumed by methanotrophic bacteria and no methane would occur within the sediments. However, situation seems to be quite different suggesting that namely methanotrophic activity measured in a laboratory could be overestimated. Since oxidation of methane requires both available methane and oxygen, methanotrophic activity is expected to be high at sites where both methane and dissolved oxygen are available. Therefore, high values of the MA were usually found in the upper layers of the sediments (Segers 1998) or at interface between oxic and anoxic zones, respectively. Relatively high methanotrophic activity found in deeper sediments of the localities III-V indicates that methane oxidation is not restricted only to the surface sediments as is common in lakes but it also takes place at greater depths. It seems likely that oxic zone occurs in a vertical profile of the sediments and that methane diffusing from the deeper layer into the sedimentary aerobic zone is being oxidized by methanotrophs here. Increased methanotrophic activity at this hyporheic oxic-anoxic interface is probably evident also from higher abundance of type II methanotrophs in the same depth layer. Similar pathway of methane cycling has been observed by Kuivila et al. (1988) in well oxygenated sediments of Lake Washington, however, methane oxidation within the sediments would be rather normal in river sediments compared to lakes. All the above mentioned findings support our previous suggestions that coexistence of various metabolic processes in hyporheic sediments is common due to vertical and horizontal mixing of the interstitial water and occurrence of

concentrations in the surface water and, in turn, high emissions to the atmosphere.

stream and, consequently source of emissons to the atmosphere.

**4.4. Benthic fluxes and potential methane oxidation** 

microbial biofilm (Hlaváčová et al. 2005, 2006).

The presence of relatively rich assemblage of methanogenic archaea in hyporheic river sediments is rather surprising, however it is in accordance with other studies. The number of total different bands (i.e. estimated diversity of the methanoges) observed in the DGGE patterns of the methanogenic archaeal communities was comparable with a number of the DGGE bands found in other studies. For example, Ikenaga et al. (2004) in their study of methanogenic archaeal community in rice roots found 15-19 DGGE bands, while Watanabe et al. (2010) showed 27 bands at different positiosns in the DGGE band pattern obtained from Japanese paddy field soils. Our results from the DGGE analysis are supported by cloning and sequencing of methyl coenzyme M reductase (*mcrA*) gene which also retrieved relatively rich diversity (25 different *mcrA* gene clones) of the methanogenic community in the Sitka stream hyporheic sediments. Similar richness in number of clones was also mentioned in a methanogenic community in Zoige wetland, where 21 different clones were found (Zhang et al. 2008a), while 20 clones were described in the methane cycle of a meromictic lake in France (Biderre-Petit et al. 2011). In addition, soils from Ljubljana marsh (Slovenia) showed 17 clones (Jerman et al. 2009), for example. Both DGGE and *mcrA* gene sequencing results suggest that both hydrogenotrophic and acetoclastic methanogenesis are an integral part of the CH4 - producing pathway in the hyporheic zone and were represented by appropriate methanogenic populations. Further, these methanogenic archaea form important component of a hyporheic microbial community and may substantially affect CH4 cycling in the Sitka stream sediments.

### **5. Conclusion**

To our knowledge this study is the first analysis of the composition of active methanogenic/methanotrophic communities in river hyporheic sediments. By use of various molecular methods we have shown that both methanogenic archaea and aerobic methanotrophs can be quantitatively dominant components of hyporheic biofilm community and may affect CH4 cycling in river sediments. Their distribution within hyporheic sediments, however, only partly reflects potential methane production and consumption rates of the sediments. Rather surprising is the detection of methanotrophs in the deep sediment layer 25-50 cm, indicating that suitable conditions for methane oxidation occur here. In addition, this work constitutes the first estimation of sources, sinks and fluxes of CH4 in the Sitka stream and in 3rd order stream environment. Fluxes of CH4 from supersaturated interstitial sediments appear to be a main CH4 source toward the water column. Compared with CH4 production rates, the diffusive fluxes are very low due to efficient aerobic oxidation by methanotrophic bacteria, especially during higher flow discharges. Although fluxes to the atmosphere from the Sitka stream seems to be insignificant, they are comparable or higher in comparison with fluxes from other aquatic ecosystems, especially those measured in running waters. Finally, our results suggest that the Sitka Stream is a source of methane into the atmosphere, and loss of carbon via the fluxes of this greenhouse gas out into the ecosystem can participate significantly in river self-purification.

### **Author details**

Martin Rulík, Adam Bednařík, Václav Mach, Lenka Brablcová, Iva Buriánková, Pavlína Badurová and Kristýna Gratzová *Department of Ecology and Environmental Sciences, Laboratory of Aquatic Microbial Ecology, Faculty of Science, Palacky University in Olomouc, Czech Republic* 

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 421

Chan, O.C., Klaus, P., Casper, P., Ulrich, A., Lueders, T. & Conrad, R., 2005: Vertical distribution of methanogenic archaeal community in Lake Dagow sediment. - Environ

Carini, S., LeCleir, G. & Joye, S.B., 2005: Aerobic methane oxidation and methanotroph community composition during seasonal stratification in Mono Lake, California (USA).

Chanton, J., Chaser, L., Glaser, P.& Siegel, D.I., 2004: Carbon and hydrogen isotopic effects on microbial methane from terrestrial environments. pp. 85-101. In: Flanagan, L.B., Ehleringer, J.R. and Pataki, D.E. (Eds.): Stable Isotopes and Biosphere-Atmosphere

Cicerone, R.J. & Oremland, R.S. 1988: Biogeochemical aspects of atmospheric methane*.* -

Conrad, R., 2004: Quantification of methanogenic pathways using stable carbon isotopic

Conrad, R., 2007: Microbial ecology of methanogens and methanotrophs. – Adv. Agron. 96:

Conrad, R., 2009: The global methane cycle: recent advances in understanding the microbial

Costelo, A. M., Auman, A.J., Macalady, J.L., Scow, K.M. & Lidstrom, M.E., 2002: Estimation of methanotroph abundance in a freshwater lake sediment. - Environ Microbiol 4: 443-

Costelo, A. M. & Lidstrom, M.E., 1999: Molecular characterization of functional and phylogenetic genes from natural populations of methanotrophs in lake sediments. –

Cupalová, J. & Rulík, M., 2007: Bacterial community analysis in river hyporheic sediments – the influence of depth and particle size. - Acta Universitatis Carolinaeana

de Angelis, M. A. & Lilley, M. D., 1987: Methane in surface waters of Oregon eustaries and

de Angelis, M. A. & Scranton, M. I., 1993: Fate of Methane in the Huson River and Estuary. -

Duchemin, E., Lucotte, M. & Canuel, R., 1999: Comparison of static chamber and thin boundary layer equation methods for measuring greenhouse gas emissions fom large

Fazi, S., Amalfitano, S., Pernthaler, J. & Puddu, A., 2005: Bacterial communities associated with benthic organic matter in headwater stream microhabitats. – Environ. Microbiol. 7:

Fischer, H., Kloep, F., Wilczek, S. & Pusch, M.T., 2005: A river's liver - microbial processes within the hyporheic zone of a large lowland river. - Biogeochemistry 76: 349-371. Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey D.W., Haywood, J., Lean, K., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga, G., Schulz, M., Van Dorland, R., 2007: Changes in atmospheric constituents and in radiative forcing. In:

signatures: a review and proposal. Organic Geochemistry 36: 739 – 752.

processes invelved. – Environ. Microbiol. Rep. 1: 285-292

Microbiol 7: 1139-1149.


Global Biogeochem Cy 1: 61-86.

1-63.

450.

1633-1640.

Interactions, Elsevier Press. New York.

Appl. Environ. Microbiol. 65: 5066-5074.

rivers. Limnology and Oceanography 33: 716 – 722.

water bodies*.* - Environ. Sci. Technol. 33: 350-357.

Global Biogeochemical Cycles 7: 509 – 523.

Environmentalica 21: 47-56.

### **Acknowledgement**

This work was supported by the Czech Grant Agency grant 526/09/1639 and partly by the Ministry of Education, Youth and Sports grants 1708/G4/2009 and 2135/G4/2009 and Palacký University IGA grant 913104161/31. We thank Lubomir Čáp and Vítězslav Maier for analyses of dissolved methane and acetic acids, Martina Vašková and Jiří Šantrůček are ackowledged for their help and suggestions when performing stable isotope analysis of 13C/12C in gas samples.

### **6. References**


Chan, O.C., Klaus, P., Casper, P., Ulrich, A., Lueders, T. & Conrad, R., 2005: Vertical distribution of methanogenic archaeal community in Lake Dagow sediment. - Environ Microbiol 7: 1139-1149.

420 Biomass Now – Cultivation and Utilization

Pavlína Badurová and Kristýna Gratzová

oxidation. – Nature 293: 289-291.

Geochim. Cosmochim. Acta 60: 3139–3142.

ordinary. - Can. J. Microbiol. 52: 73-116.

Martin Rulík, Adam Bednařík, Václav Mach, Lenka Brablcová, Iva Buriánková,

*Faculty of Science, Palacky University in Olomouc, Czech Republic* 

streambed sediments. – J. Microbiol. Meth. 75: 237-243.

*Department of Ecology and Environmental Sciences, Laboratory of Aquatic Microbial Ecology,* 

This work was supported by the Czech Grant Agency grant 526/09/1639 and partly by the Ministry of Education, Youth and Sports grants 1708/G4/2009 and 2135/G4/2009 and Palacký University IGA grant 913104161/31. We thank Lubomir Čáp and Vítězslav Maier for analyses of dissolved methane and acetic acids, Martina Vašková and Jiří Šantrůček are ackowledged for their help and suggestions when performing stable isotope analysis of

Amalfitano, S. & Fazi, S., 2008: Recovery and quantification of bacterial cells associated with

Baker, M.A., Dahm, C.N. & Vallet, H.M., 1999: Acetate retention and metabolism in the

Barker, J.F. & Fritz, P., 1981: Carbon isotope fractionation during microbial methane

Boudreau, B. P.,1996: The diffusive tortuosity of fine-grained unlithified sediments. –

Bretschko, G., 1981: Vertical distribution of zoobenthos in an alpine brook of the

Bretschko, G. & Klemens, W.E., 1986: Quantitative methods and aspects in the study of the

Buriánková, I. Brablcová, L., Mach, V., Hýblová, A., Badurová, P., Cupalová, J., Čáp, L. &

Buriánková, I., Brablcová, L., Mach, V. & Rulík, M., in review: Methanogens diversity in hyporheic river sediment by targeting methyl-coenzyme M reductase (*mcrA)* gene. Chaban, B., Ng, S.Y.M. & Jarrell, K.F., 2006: Archaeal habitats – from the extreme to the

RITRODAT-LUNZ study area. - Verh. Internat. Verein. Limnol. 21: 873-876.

hyporheic sediments of a small lowland stream. - Fundam. Appl. Limnol. 181: 87-102.

interstitial fauna of running waters. - Stygologia 2: 297–316.

Rulík, M., 2012: Methanogens and methanotrophs distribution in the

Berner, R. A., 1980: Early Diagenesis: A Theoretical Approach. Princeton University Press. Biderre-Petit, C. , Jézéquel, D., Dugat-Bony, E., Lopes, F., Kuever, J., Guillaume Borrel, G., Viollier, E., Fonty, G. & Peyret, P., 2011: Identification of microbial communities involved in the methane cycle of a freshwater meromictic lake. - FEMS Microbiol. Ecol.

hyporheic zone of a moutain stream. – Limnol. Oceanogr. 44: 1530 – 1539.

**Author details** 

**Acknowledgement** 

13C/12C in gas samples.

**6. References** 

77: 533–545.


Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor & M., Miller, H.L. (eds): Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel in Climate Change. - Cambridge University Press, Cambridge, United Kingdom/New York, NY, USA.

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 423

East Pacific Rise, and identification of *Methanococcus* sp. DSM 4213T as *Methanococcus* 

Jerman, V., Metje, M., Mandic-Mulec, I. & Frenzel, P., 2009: Wetland restoration and methanogenesis: the activity of microbial populations and competition for substrates at

Jones, J.B. & Holmes, R.M., 1996: Surface-subsurface interactions in stream ecosystems. -

Jürgens, G., Glöckner, F.-O., Amann, R., Saano, A., Montonen, L., Likolammi, M. & Münster, U. 2000: Identification of novel Archaea in bacterioplankton of a boreal forest lake by phylogenetic analysis and fluorescent in situ hybridization. - FEMS Microbiol. Ecol. 34:

Kalyuzhnaya, M.G., Zabinsky, R., Bowerman, S., Baker, D.R., Lidstrom, M.E. & Chistoserdova, L., 2006: Fluorescence in situ hybridization-flow cytometry-cell sortingbased method for separation and enrichment of type I and type II methanotroph

Karl, D.M., Beversdorf, L., Björkman, K.M., Church, M.J., Martinez, A. & DeLong, E.F., 2008:

Keough, B.P., Schmidt, T.M. & Hicks, R.E., 2003: Archaeal nucleic acids i n picoplankton

Kobabe, S., Wagner, D. & Pfeifer, E.M., 2004: Characterisation of microbial community composition of a Siberian tundra soil by fluorescence in situ hybridization. - FEMS

Kroeze, C. & Seitzinger, S. P., 1998: The impact of land use on N2O emissions from watersheds draining into the Northeastern Atlantic Ocean and European Seas. -

Kuivila, K.M., Murray, J.W., Devol, A.H., Lidstrom, M.E. & Reimers, C.E., 1988: Methane cycling in the sediments of Lake Washington. – Limnol. Oceanogr. 33: 571-581. Leichtfried, M., 1988: Bacterial substrates in gravel beds of a second order alpine stream (Project Ritrodat-Lunz, Austria). – Verh. Internat. Verein. Limnol. 23: 1325–1332. Le Mer J. & Roger, P., 2001: Production, oxidation, emission and consumption of methane

Lilley, M.D., de Angelis, M.A. & Olson, J.E., 1996: Methane concentrations and estimated fluxes from Pacific Northwest rivers. – Mitt. Internat. Verein. Limnol. 25: 187-196. Lin, C., Raskin, L. & Stahl, D.A., 1997: Microbial community structure in gastrointestinal tracts of domestic animals: comparative analyzes using rRNA-targeted oligonucleotide

Mach, V., Čáp, L., Buriánková, I., Brablcová, L., Cupalová, J. & Rulík, M., in review: Spatial and temporal heterogeneity in greenhouse gases emissions along longitudinal profile of

Mathrani, I.M., Boone, D.R., Mah, R.A., Fox, G.E. & Lau, P.P., 1988: *Methanohalophilus zhilinae*, sp. nov., an alkaliphilic, halophilic, methylotrophic methanogen. - Int. J. Syst.

Aerobic production of methane in the sea. - Nature Geoscience 1: 473-478.

from great lakes on three continents. – Microb. Ecol. 46: 238-248.

*fervens* sp. nov. – Int. J. Syst. Evol. Microbiol. 49: 583–589.

different temperatures. - Biogeosciences 6: 1127-1138.

populations. – Appl. Environ. Microb. 72: 4293–4301.

TREE 11: 239-242.

Microbiol. Ecol. 50: 13-23.

Environm. Pollut*.* 102: 149–158.

by soils. A review. - Eur. J.Soil. Biol*.* 37: 25-50

probes. - FEMS Microbiol. Ecol*.* 22: 281–294.

a small lowland stream. – J. Limnol..

Bacteriol. 38: 139–142.

45–56.


East Pacific Rise, and identification of *Methanococcus* sp. DSM 4213T as *Methanococcus fervens* sp. nov. – Int. J. Syst. Evol. Microbiol. 49: 583–589.

Jerman, V., Metje, M., Mandic-Mulec, I. & Frenzel, P., 2009: Wetland restoration and methanogenesis: the activity of microbial populations and competition for substrates at different temperatures. - Biogeosciences 6: 1127-1138.

422 Biomass Now – Cultivation and Utilization

York, NY, USA.

960–969.

4983–4989.

353.

Press New York, pp. 137-162.

Microb. 65: 1980–1990.

Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor & M., Miller, H.L. (eds): Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel in Climate Change. - Cambridge University Press, Cambridge, United Kingdom/New

Ganzert, L., Jürgens, G., Münster, U. & Wagner, D., 2006: Methanogenic communities in permafrost-affected soils of the Laptev Sea coast, Siberian Arctic, characterized by 16S

Garcia, J.-L., Patel, B.K.C. & Ollivier, B., 2000: Taxonomic, phylogenetic, and ecological

Großkopf, R., Janssen, P.H. & Liesack, W., 1998*a:* Diversity and structure of the methanogenic community in anoxic rice paddy soil microcosms as examined by cultivation and direct 16S rRNA gene sequence retrieval. – Appl. Environ. Microb. 64:

Großkopf, R., Stubner, S. & Liesack, W., 1998*b:* Novel euryarchaeotal lineages detected on rice roots and in the anoxic soil of flooded rice microcosms. – Appl. Environ. Microb. 64:

Hanson, R.S., 1998: Ecology of Methanotrophic Bacteria. - In: Burlage, R.S., Atlas, R., Stahl, D., Geesey, G. & Sayler G. (eds): Techniques in Microbial Ecology. - Oxford University

Hanson, R.S. & Hanson, T.E., 1996: Methanotrophic bacteria. – Microbiol. Rev. 60: 439-471. Henckel, T., Friedrich, M. & Conrad, R., 1999: Molecular analyses of the methane-oxidizing microbial community in rice field soil by targeting the genes of the 16S rRNA, particulate methane mono-oxygenase, and methanol dehydrogenase. – Appl. Environ.

Hendricks, S.P., 1993: Microbial ecology of the hyporheic zone: a perspective integrating

Heyer, J. & Berger, U., 2000: Methane Emissions from the Coastal Area in the Southern Baltic

Hlaváčová, E., Rulík, M. & Čáp, L., 2005: Anaerobic microbial metabolism in hyporheic sediment of a gravel bar in a small lowland stream. – River Res. Appl. 21: 1003 – 1011. Hlaváčová, E., Rulík, M., Čáp, L. & Mach, V., 2006: Greenhouse gases (CO2, CH4, N2O) emissions to the atmosphere from a small lowland stream. – Arch. Hydrobiol. 165: 339 –

Holmes, R.M., Fisher, S.G. & Grimm, N.B., 1994: Parafluvial nitrogen dynamics in a desert

Ikenaga, M., Asakawa, S., Muraoka, Y. & Kimura, M., 2004: Methanogenic archaeal communities in rice roots grown in flooded soil pots: Estimation by PCR-DGGE and

Jeanthon, C., L'Haridon, S., Reysenbach, A.L., Corre, E., Vernet, M., Messner, P. et al., 1999: *Methanococcus vulcanius* sp. nov., a novel hyperthermophilic methanogen isolated from

rRNA gene fingerprints. - FEMS Microbiol. Ecol. 59: 476-488.

diversity of methanogenic Archaea. - Anaerobe 6: 205-226.

hydrology and biology. - J. N. Am. Benthol. Soc. 12: 70-78.

Sea. - Estuarine, Coastal and Shelf Science 51: 13 – 30.

stream ecosystem. – J. N. Am. Benthol. Soc. 13: 468-478.

sequence analyses. - Soil. Sci. Plant Nutr. 50: 701-711.


McDonald, I.R., Bodrossy, L., Chen, Y. & Murrell, C.J., 2008: Molecular ecology techniques for the study of aerobic methanotrophs. - Appl. Environ. Microbiol. 74: 1305-1315

Methanogenic System of a Small Lowland Stream Sitka, Czech Republic 425

(solution) –

Schindler, J.E. & Krabbenhoft, D.P., 1998: The hyporheic zone as a source of dissolved organic carbon and carbon gases to a temperate forested stream. - Biogeochemistry 43:

Schönbauer, B. & Lewandowski, G., 1999: SEDI – a database oriented analysis and evaluation tool for processing sediment parameters. - J. Biol. Stn. Lunz 16: 13–27. Silvennoinen, H., Liikanen, A., Rintala, J. & Martikainen, P.J., 2008: Greenhouse gas fluxes from eutrophic Temmesjoki River and its Estuary in the Liminganlahti Bay (the Baltic

Storey, R.G., Fulthorpe, R.R., Williams, D.D., 1999: Perspectives and predictions on the

Strous, M. & Jetten, M.S.M., 2004: Anaerobic oxidation of methane and ammonium. – Annu.

Trotsenko, Y.A., Khmelenina, V.N., 2005: Aerobic methanotrophic bacteria of cold

Truleyová, Š., Rulík, M. & Popelka, J., 2003: Stream and interstitial water DOC of a gravel bar (Sitka stream, Czech Republic): characteristics, dynamics and presumable origin. –

Watanabe T., Asakawa, S., Nakamura, A., Nagaoka K. & Kimura, M., 2004: DGGE method for analyzing 16S rDNA of methanogenic archaeal community in paddy field soil. -

Watanabe, T., Hosen, Y., Agbisit, R., Llorca, L., Fujita, D., Asakawa, S. & Kimura, M., 2010: Changes in community structure and transcriptional activity of methanogenic archaea in a paddy field soil brought about by a water-saving practice – Estimation by PCR-DGGE and qPCR of 16S rDNA and 16S rRNA. pp. 5-8. – 19th World Congress of Soil

Whiticar, M.J., Faber, E. & Schoell, M., 1986. Biogenic methane formation in marine and freshwater environments: CO2 reduction vs. Acetate fermentation – isotopic evidence. -

Wiesenburg, D.A. & Guinasso, N.L., 1979: Equilibrium solubilities of methane, carbon monoxide, and hydrogen in water and sea water. - J. Chem. Eng. Data 24: 356–360. Wilcock, R.J. & Sorrell, B.K., 2008: Emissions of greenhouse gases CH4 and N2O from lowgradient streams in agriculturally developed catchments. - Water Air Soil Poll. 188: 155-

Worral, F. & Lancaster, A., 2005: The release of CO2 from riverwaters - the contribution of

Zhang, G., Tian, J., Jiang, N., Guo, X. Wang, Y. & Dong, X., 2008a: Methanogen community in Zoige wetland of Tibetan plateau and phenotypic characterization of a dominant

Solutions for a Changing World. 1-6 August 2010, Brisbane, Australia

excess CO2 from groundwater. - Biogeochemistry 76: 299 317.

uncultured methanogen cluster ZC-I. – Environ. Microbiol. 10:1850-60.

microbial ecology of the hyporheic zone. - Freshwat. Biol. 41: 119-130.

Szaran, J., 1997: Achivement of carbon isotope equilibrium in the systém HCO3-

157-174.

170.

Sea*).* - Biogeochemistry 90: 193-208.

CO2 (gas). - Chemical Geology 142, 79 – 86.

ecosystems. - FEMS Microbiol. Ecol. 53: 15-26.

Geochemica et Cosmochimica Acta 50: 693 – 709.

Rev. Microbiol. 58: 99-117.

Arch. Hydrobiol. 158: 407–420.

FEMS Microb. Letters 232: 153-163.


Schindler, J.E. & Krabbenhoft, D.P., 1998: The hyporheic zone as a source of dissolved organic carbon and carbon gases to a temperate forested stream. - Biogeochemistry 43: 157-174.

424 Biomass Now – Cultivation and Utilization

59: 95-119.

30: 207-225.

Am. Bethol. Soc. 19: 593-608.

swamp. - Ecol. Monogr*.* 63: 29–53.

Geophysical Research Letters 35: L22805.

a headwater stream. – Arch. Hydrobiol. 91: 28–44.

edition. Wiley-Blackwell. ISBN: 978-1-4051-2680-9

methanotroph ecology. - FEMS Microbiol. Ecol*.* 27: 103-114.

stream (Sitka, Czech Republic). - Limnologica 30: 359 – 366.

small lowland stream. - Soil Biol. & Biochem. 36: 1653 – 1662.

for agricultural practices. – Freshw. Biol 52: 1176-1186.

methane in water and gas. - Anal. Chem. 69: 40-44.

bay, California. - Estuaries. 21: 66–77.

samples. – J. Microbiol. Meth. 41: 249-257.

McDonald, I.R., Bodrossy, L., Chen, Y. & Murrell, C.J., 2008: Molecular ecology techniques for the study of aerobic methanotrophs. - Appl. Environ. Microbiol. 74: 1305-1315 Meyer, J.L., Likens, G.E. & Sloane, J., 1981: Phosphorus, nitrogen, and organic carbon flux in

Michener, R.& Lajtha, K., 2007: Stable Isotopes in Ecology and Environmental Science, 2nd

Middelburg, J.L., Nieuwenhuize, J., Iversen, N., Høgh, N., De Wilde, H., Helder, W., Seifert, R., & Christof, O., 2002: Methane distribution in European estuaries. - Biogeochemistry

Morrice, J.A., Dahm, C.N., Valett, H.M., Unnikrishna, P. & Campana, M.E., 2000: Terminal electron accepting processes in the alluvial sediments of a headwater stream. – J. N.

Murrell, J.C., McDonald, I.R. & Bourne, D.G., 1998: Molecular methods for the study of

Pernthaler, J., Glockne, FO, Schonhuber, W. and Amann, R. (2001). Fluorescence in situ hybridization (FISH) with rRNA-targeted oligonucleotide probes. Methods Microbiol*.*

Pulliam, W. M., 1993: Carbon dioxide and methane exports from a southeastern floodplain

Rigby, M., Prinn, R.G., Fraser, P.J., Simmonds P.G., Langenfelds, R.L., Huang, J., Cunnold, D.M., Steele, L.P., Krummel, P.B., Weiss, R.F., O'Doherty, S., Salameh, P.K., Wang, H.J., Harth, C.M., Mühle, J., Porter, L.W., 2008: Renewed growth of atmospheric methane. -

Rulík, M., Čáp, L. & Hlaváčová, E., 2000: Methane in the hyporheic zone of a small lowland

Rulík, M. & Spáčil, R., 2004: Extracellular enzyme activity within hyporheic sediments of a

Sanders, I.A., Heppell, C.M., Cotton, J.A., Wharton, G., Hildrew, A.G., Flowers, E.J. & Trimmer, M., 2007: Emissions of methane from chalk streams has ptential implications

Sansone, F.J., Popp, B.N. & Rust, T.M., 1997: Stable carbon isotopic analysis of low-level

Sansone, F. J., Rust, T. M. & Smith S. V., 1998: Methane distribution and cycling in Tomales

Santos Furtado, A.L. & Casper, P., 2000: Different methods for extracting bacteria from freshwater sediment and simple method to measure bacterial production in sediment

Sanz, J.L., Rodríguez, N., Díaz, E.E., Amils, R., 2011: Methanogenesis in the sediments of Rio

Segers, R., 1998: Methane production and methane consumption: a review of processes

Tinto, an extreme acidic river. - Environmental Microbiology 13: 2336-2341.

underlying wetland methane fluxes. - Biogeochemistry 41: 23-51.

	- Zhang, G., Zhang, J., Liu, S., Ren, J., Xu, J., Zhang, F., 2008b: Methane in the Changjiang (Yangtze River) Estuary and its adjacent marine area: riverine input, sediments release and atmospheric fluxes. - Biogeochemistry 91: 71-84.

**Chapter 18** 

© 2013 Chatzistathis and Therios, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is

distribution, and reproduction in any medium, provided the original work is properly cited.

© 2013 Chatzistathis and Therios, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,

**How Soil Nutrient Availability Influences Plant** 

**The Cases of Nutrient Use Efficient Genotypes** 

There are many factors influencing plant biomass, such as soil humidity, soil and air temperature, photoperiod, solar radiation, precipitations, genotype e.t.c. One of the most important factors influencing biomass is soil nutrient availability. Both nutrient deficiency and toxicity negatively affect total biomass and fruit production [1-10]. So, by controlling the optimum levels of nutrient availability in soil, the production of biomass and, of course, the economic benefit (fruit production) for the farmers can be maximized. In the cases of limited nutrient availability in soils, fertilization seems to be the most usual practice adopted by the farmers in order to ameliorate the low nutrient status. However, since: i) during the last two decades the prices of fertilizers have been dramatically increased, and ii) soil degradation and pollution, as well as underground water pollution, are serious consequences provoked by the exaggerate use of fertilizers, a global concern to reduce the use of fertilizers has been developed. So, the best (most economic and ecological) way in our days to achieve maximum yields is by selecting and growing nutrient efficient genotypes, i.e. genotypes which are able to produce high yields (biomass) in soils with limited nutrient availability. Many researchers studied the influence of genotype on biomass and plant growth (nutrient use efficient genotypes) and found impressive results. According to Chapin and Van Cleve (1991) [11], nutrient use efficiency is defined as the amount of biomass produced per unit of nutrient. So, nutrient use efficient genotypes are those having the ability to produce biomass sufficiently under limited nutrient availability. In our research with different olive cultivars, grown under hydroponics, or in soil substrate, we found significant differences concerning

**Biomass and How Biomass Stimulation** 

**Alleviates Heavy Metal Toxicity in Soils:** 

**and Phytoremediators, Respectively** 

Theocharis Chatzistathis and Ioannis Therios

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/53594

**1. Introduction** 

properly cited.

## **How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively**

Theocharis Chatzistathis and Ioannis Therios

Additional information is available at the end of the chapter

http://dx.doi.org/10.5772/53594

### **1. Introduction**

426 Biomass Now – Cultivation and Utilization

Zhang, G., Zhang, J., Liu, S., Ren, J., Xu, J., Zhang, F., 2008b: Methane in the Changjiang (Yangtze River) Estuary and its adjacent marine area: riverine input, sediments release

and atmospheric fluxes. - Biogeochemistry 91: 71-84.

There are many factors influencing plant biomass, such as soil humidity, soil and air temperature, photoperiod, solar radiation, precipitations, genotype e.t.c. One of the most important factors influencing biomass is soil nutrient availability. Both nutrient deficiency and toxicity negatively affect total biomass and fruit production [1-10]. So, by controlling the optimum levels of nutrient availability in soil, the production of biomass and, of course, the economic benefit (fruit production) for the farmers can be maximized. In the cases of limited nutrient availability in soils, fertilization seems to be the most usual practice adopted by the farmers in order to ameliorate the low nutrient status. However, since: i) during the last two decades the prices of fertilizers have been dramatically increased, and ii) soil degradation and pollution, as well as underground water pollution, are serious consequences provoked by the exaggerate use of fertilizers, a global concern to reduce the use of fertilizers has been developed. So, the best (most economic and ecological) way in our days to achieve maximum yields is by selecting and growing nutrient efficient genotypes, i.e. genotypes which are able to produce high yields (biomass) in soils with limited nutrient availability. Many researchers studied the influence of genotype on biomass and plant growth (nutrient use efficient genotypes) and found impressive results. According to Chapin and Van Cleve (1991) [11], nutrient use efficiency is defined as the amount of biomass produced per unit of nutrient. So, nutrient use efficient genotypes are those having the ability to produce biomass sufficiently under limited nutrient availability. In our research with different olive cultivars, grown under hydroponics, or in soil substrate, we found significant differences concerning

© 2013 Chatzistathis and Therios, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2013 Chatzistathis and Therios, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

macro- and micronutrient utilization efficiency among genotypes [12-13]. Possible reasons for differential nutrient utilization efficiency among genotypes may be: i) the genetic material used, i.e. cultivar (differential nutrient uptake, accumulation and distribution among tissues, mechanisms of cultivars/genotypes), ii) differential colonization of their root system mycorrhiza fungi. Chatzistathis et al. (2011) [14] refer that the statistically significant differences in Mn, Fe and Zn utilization efficiency among three Greek olive cultivars ('Chondrolia Chalkidikis', 'Koroneiki' and 'Kothreiki') may be probably ascribed to the differential colonization of their root system by arbuscular mycorrhiza fungus (AMF) (the percentage root colonization by AMF varied from 45% to 73%).

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

and cultivars, are within the aim of the present review. Furthermore, the characteristics that should have the plant species used for phytoremediation (fast-growing, high biomass crops) in heavy metal polluted soils are fully analyzed, and the different strategies that should be adopted in order to enhance plant growth and biomass production under so adverse soil conditions are also discussed under the light of the most important and recent research papers.

**2. Agronomic, environmental and genotypic factors influencing plant** 

Plant growth (i.e. biomass production) is influenced by many (agronomic environmental and others, such as genetic) factors. Some of the most important factors that influence biomass production are: i) soil humidity, ii) soil and air temperature, iii) air humidity, iv) photoperiod, v) light intensity, vi) soil fertility, i.e. soil nutrient availability, and vii)

Soil humidity is a very crucial factor influencing root growth, thus nutrient uptake and total biomass. Many plant species are more sensitive in soil humidity shortage during a particular (crucial) period of their growth. In olive trees, if soil humidity shortage happens early spring, shoot elongation, as well as the formation of flowers and fruits, are negatively influenced. If the shortage happens during summer, shoot thickening, rather than shoot elongation, is influenced. Finally, soil humidity shortage reduces olive tree canopy (in order to reduce the transpiration by leaf surface) and favors root system growth (in order to have the ability to exploit greater soil volume and to search for more soil humidity), so that the ratio canopy/root is significantly reduced [30]. On the other hand, under excess soil humidity conditions (waterlogging), when soil oxygen is limited, the root system may suffer from hypoxia, thus, nutrient uptake is negatively influenced. Under extreme anaerobic soil conditions, the presence of pathogen microorganisms, such as *Phytophthora* sp. may lead to root necrosis. According to Therios (2009) [31], for olive trees the mechanism of tolerance to waterlogging is based on the production of adventitious roots near to the soil surface.

Soil temperature influences root growth, thus nutrient and water uptake and, of course, biomass production. Most nutrients are absorbed with energy consumption (energetic uptake), so, low and very high soil temperatures negatively influence root growth and

Air temperature directly influences photosynthesis, which is the most important physiological function in plants. The optimum temperature for photosynthesis depends on plant species and also on cultivar for the same species. Usually, the optimum temperature

nutrient uptake. Furthermore, low soil temperatures induce a water deficit [32].

**growth** 

**2.1. Soil humidity** 

**2.2. Soil temperature** 

**2.3. Air temperature** 

genotype, and are fully analyzed below.

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 429

Heavy metal (Cu, Zn, Ni, Pb, Mn, Cr, Cd) toxicity is a very serious problem in soils suffering from: i) industrial and mine activities [15], ii) the exaggerate use of fertilizers, fungicides and insecticides, iii) acidity, iv) waterlogging, v) other urban activities, such as municipal sewage sludges, vi) the use of lead in petrols, paints and other materials [16]. Under these conditions, plant growth and biomass are negatively affected [17-20]. According to Caldelas et al. (2012) [19], not only growth inhibition happened, but also root to shoot dry matter partitioning (R/S) modified (increased 80%) at Cr toxic conditions in Iris pseudacorus L. plants. Some plant species, which may tolerate very high metal concentrations in their tissues, can be used as hyper-accumulators and are very suitable in reducing heavy metal concentrations in contaminated soils [21]. These species are able to accumulate much more metal in their shoots, than in their roots, without suffering from metal toxicity [22]. By successive harvests of the aerial parts of the hyper-accumulator species, the heavy metals concentration can be reduced [23]. Phytoremediation is an emerging technology and is considered for remediation of inorganic- and organic-contaminated sites because of its cost-effectiveness, aesthetic advantages, and long-term applicability. This technique involves the use of the ability of some plant species to absorb and accumulate high concentrations of heavy metal ions [17]. Some of these species may be a few ones from Brassicaceae family, such as raya (Brassica campestris L.) [17] and Thlaspi caerulescens [23], or from other families, such as spinach (Spinacia oleracea L.) [17], Sedum plumbizincicola [24], Amaranthus hypochondriacus [25], Eremochloa ophiuroides [26], Iris pseudacorus L. [19], Ricinus communis L., plant of Euphorbiaceae family [18]. Finally, the tree species Genipa Americana L. may be used as one with great ability as phytostabilizer and rhizofilterer of Cr ions, according to Santana et al. (2012) [20]. Basically, there are two different strategies to phytoextract metals from soils: the first approach is the use of metal hyper-accumulator species. The second one is to use fast-growing, high biomass crops that accumulate moderate to high levels of metals in their shoots for metal phytoremediation, such as Poplar (Populus sp.) [27-28], maize (Zea mays), oat (Avena sativa), sunflower (Helianthus annuus) and rice (Oryza sativa L.) [25]. Generally, the more high biomass producing is one plant species, the more efficient is the phytoremediation effect. So, in order to enhance biomass production under metal toxicity conditions, different strategies, such as the application of chemical amendments, may be adopted [21]. Since Fe deficiency symptoms may be appeared under Cu and Zn toxicity conditions in some species of Brassicaceae family used for phytoremediation, a good practice is to utilize Fe foliar sprays in order to enhance biomass, thus the phytoremediation effect [29].

All the above mentioned topics, concerning the influence of nutrient deficiency and metal toxicity on plant biomass, as well as the importance of using nutrient use efficient genotypes and cultivars, are within the aim of the present review. Furthermore, the characteristics that should have the plant species used for phytoremediation (fast-growing, high biomass crops) in heavy metal polluted soils are fully analyzed, and the different strategies that should be adopted in order to enhance plant growth and biomass production under so adverse soil conditions are also discussed under the light of the most important and recent research papers.

### **2. Agronomic, environmental and genotypic factors influencing plant growth**

Plant growth (i.e. biomass production) is influenced by many (agronomic environmental and others, such as genetic) factors. Some of the most important factors that influence biomass production are: i) soil humidity, ii) soil and air temperature, iii) air humidity, iv) photoperiod, v) light intensity, vi) soil fertility, i.e. soil nutrient availability, and vii) genotype, and are fully analyzed below.

### **2.1. Soil humidity**

428 Biomass Now – Cultivation and Utilization

thus the phytoremediation effect [29].

macro- and micronutrient utilization efficiency among genotypes [12-13]. Possible reasons for differential nutrient utilization efficiency among genotypes may be: i) the genetic material used, i.e. cultivar (differential nutrient uptake, accumulation and distribution among tissues, mechanisms of cultivars/genotypes), ii) differential colonization of their root system mycorrhiza fungi. Chatzistathis et al. (2011) [14] refer that the statistically significant differences in Mn, Fe and Zn utilization efficiency among three Greek olive cultivars ('Chondrolia Chalkidikis', 'Koroneiki' and 'Kothreiki') may be probably ascribed to the differential colonization of their root system by arbuscular mycorrhiza fungus (AMF) (the

Heavy metal (Cu, Zn, Ni, Pb, Mn, Cr, Cd) toxicity is a very serious problem in soils suffering from: i) industrial and mine activities [15], ii) the exaggerate use of fertilizers, fungicides and insecticides, iii) acidity, iv) waterlogging, v) other urban activities, such as municipal sewage sludges, vi) the use of lead in petrols, paints and other materials [16]. Under these conditions, plant growth and biomass are negatively affected [17-20]. According to Caldelas et al. (2012) [19], not only growth inhibition happened, but also root to shoot dry matter partitioning (R/S) modified (increased 80%) at Cr toxic conditions in Iris pseudacorus L. plants. Some plant species, which may tolerate very high metal concentrations in their tissues, can be used as hyper-accumulators and are very suitable in reducing heavy metal concentrations in contaminated soils [21]. These species are able to accumulate much more metal in their shoots, than in their roots, without suffering from metal toxicity [22]. By successive harvests of the aerial parts of the hyper-accumulator species, the heavy metals concentration can be reduced [23]. Phytoremediation is an emerging technology and is considered for remediation of inorganic- and organic-contaminated sites because of its cost-effectiveness, aesthetic advantages, and long-term applicability. This technique involves the use of the ability of some plant species to absorb and accumulate high concentrations of heavy metal ions [17]. Some of these species may be a few ones from Brassicaceae family, such as raya (Brassica campestris L.) [17] and Thlaspi caerulescens [23], or from other families, such as spinach (Spinacia oleracea L.) [17], Sedum plumbizincicola [24], Amaranthus hypochondriacus [25], Eremochloa ophiuroides [26], Iris pseudacorus L. [19], Ricinus communis L., plant of Euphorbiaceae family [18]. Finally, the tree species Genipa Americana L. may be used as one with great ability as phytostabilizer and rhizofilterer of Cr ions, according to Santana et al. (2012) [20]. Basically, there are two different strategies to phytoextract metals from soils: the first approach is the use of metal hyper-accumulator species. The second one is to use fast-growing, high biomass crops that accumulate moderate to high levels of metals in their shoots for metal phytoremediation, such as Poplar (Populus sp.) [27-28], maize (Zea mays), oat (Avena sativa), sunflower (Helianthus annuus) and rice (Oryza sativa L.) [25]. Generally, the more high biomass producing is one plant species, the more efficient is the phytoremediation effect. So, in order to enhance biomass production under metal toxicity conditions, different strategies, such as the application of chemical amendments, may be adopted [21]. Since Fe deficiency symptoms may be appeared under Cu and Zn toxicity conditions in some species of Brassicaceae family used for phytoremediation, a good practice is to utilize Fe foliar sprays in order to enhance biomass,

All the above mentioned topics, concerning the influence of nutrient deficiency and metal toxicity on plant biomass, as well as the importance of using nutrient use efficient genotypes

percentage root colonization by AMF varied from 45% to 73%).

Soil humidity is a very crucial factor influencing root growth, thus nutrient uptake and total biomass. Many plant species are more sensitive in soil humidity shortage during a particular (crucial) period of their growth. In olive trees, if soil humidity shortage happens early spring, shoot elongation, as well as the formation of flowers and fruits, are negatively influenced. If the shortage happens during summer, shoot thickening, rather than shoot elongation, is influenced. Finally, soil humidity shortage reduces olive tree canopy (in order to reduce the transpiration by leaf surface) and favors root system growth (in order to have the ability to exploit greater soil volume and to search for more soil humidity), so that the ratio canopy/root is significantly reduced [30]. On the other hand, under excess soil humidity conditions (waterlogging), when soil oxygen is limited, the root system may suffer from hypoxia, thus, nutrient uptake is negatively influenced. Under extreme anaerobic soil conditions, the presence of pathogen microorganisms, such as *Phytophthora* sp. may lead to root necrosis. According to Therios (2009) [31], for olive trees the mechanism of tolerance to waterlogging is based on the production of adventitious roots near to the soil surface.

### **2.2. Soil temperature**

Soil temperature influences root growth, thus nutrient and water uptake and, of course, biomass production. Most nutrients are absorbed with energy consumption (energetic uptake), so, low and very high soil temperatures negatively influence root growth and nutrient uptake. Furthermore, low soil temperatures induce a water deficit [32].

### **2.3. Air temperature**

Air temperature directly influences photosynthesis, which is the most important physiological function in plants. The optimum temperature for photosynthesis depends on plant species and also on cultivar for the same species. Usually, the optimum temperature for maximum photosynthetic activity is around 25oC for most vegetative species. When temperature exceeds 35oC photosynthesis is inhibited, thus biomass production may be restrained. High temperatures are associated with a high vapor pressure deficit between leaves and the surrounding air. The same applies to fruit, where high temperatures may cause fruit drop in olive trees [31]. On the other hand, low temperatures act negatively in photosynthesis function and starch is redistributed and is accumulated in organs protected from frost, such as roots. Very low temperatures (<-12oC) damage the leaf canopy, shoot and branches of trees [31].

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

symptoms: i) General symptoms, such as limited growth and inability of reproduction (flowering and fruit setting), caused by the deficiency of many necessary macro- or micronutrients, and ii) typical, characteristic, deficiency symptoms, such as chlorosis, i.e. yellowing (due to Fe deficiency). In both cases biomass production is depressed. In the study of Msilini et al. (2009) [10], bicarbonate treated plants of *Arabidopsis thaliana* suffered from Fe deficiency displayed significantly lower biomass, leaf number and leaf surface, as compared to control plants, and showed slight yellowing of their younger leaves. Under limited nutrient availability, arbuscular mycorrhiza fungi (AMF) may favor nutrient uptake and thus enhance biomass production. Hu et al. (2009) [35] refer that AMF inoculation of maize plants was likely more efficient in extremely P-limited soils. Generally, root colonization by AMF influences positively plant growth under N, P, or micronutrient

**2.8. Genotypic factors (root morphology and architecture, genetic growth** 

alluvium in north India, concerning plant height, diameter e.t.c.

**3. Physiological roles of nutrients** 

nutrients are described in detail below.

reduced flowering and decreased yield [31].

**3.1. Macronutrients** 

According to Bayuelo-Jimenez et al. (2011) [3], under P deficiency, P-efficient accessions of maize plants (*Zea mays* L.) had greater root to shoot ratio, nodal rooting, nodal root laterals, nodal root hair density and length of nodal root main axis, and first-order laterals. In our experiments, we also found differential root system morphology among three Greek olive cultivars (the root systems of 'Koroneiki' and 'Chondrolia Chalkidikis' were less branched and more lateral, and with less root hair development and density, than that of 'Kothreiki', which was richly-branched and with much greater root hair development and density), something which was probably the main reason for the great genotypic variations in nutrient uptake and growth among the three cultivars (Chatzistathis, unpublished data). Singh et al. (2010) [37] found that great differences existed among 10 multipurpose tree species, grown in a monoculture tree cropping system on the sodic soils of Gangetic

The absolutely necessary nutrients for plant growth are the following: N, P, K, S, Ca, Mg (macronutrients), Zn, Cu, B, Mn, Fe, Mo (micronutrients). Without one of these nutrients, plant organism can not grow normally and survive. The physiological roles of these

Nitrogen: It is a primary component of nucleic acids, proteins, amino acids, purines, pyrimidines and chlorophyll. Nitrogen exerts a significant effect on plant growth, as it reduces biennial bearing and increases the percentage of perfect flowers. In olive trees, lack of N leads to decreased growth, shorter length of annual shoots (<10cm), fewer leaves,

deficiency conditions [36].

**capacity e.t.c.)** 

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 431

### **2.4. Air humidity**

Low atmosphere humidity speeds up transpiration by leaf surface. Increase of the rate of transpiration causes reduction of vegetative tissues water content, thus depression in the rate of growth and biomass production.

### **2.5. Photoperiod**

Photoperiod is the duration of light in 24 hours and it is one of the most important factors influencing vegetative growth. Plant species whose vegetative growth is mostly influenced by long day conditions are *Populus robusta*, *Ulmus Americana* and *Aesculus hippocastanus* [30].

### **2.6. Light intensity**

Light, together with CO2, are the two main factors influencing photosynthetic rate. By increasing light intensity up to an optimum limit the maximum photosynthetic rate, so the greatest biomass production can be achieved.

### **2.7. Nutrient availability**

Limited nutrient availability influences negatively biomass production. Nitrogen deficiency strongly depresses vegetation flush. According to Boussadia et al. (2010) [8], total biomass of two olive cultivars ('Meski' and 'Koroneiki') was strongly reduced (mainly caused by a decrease in leaf dry weight) under severe N deprivation, while in an out-door pot-culture experiment with castor bean plants (*Ricinus communis* L.), conducted by Reddy and Matcha (2010) [9], it was found that among the plant components, leaf dry weight had the greatest decrease; furthermore, root/shoot ratio increased under N deficiency [9]. Phosphorus deficiency caused reduced biomass, photosynthetic activity and nitrogen fixing ability in mungbean (*Vigna aconitifolia*) and mashbean (*Vigna radiata*) [33]. Under P deficiency conditions, genotypic variation in biomass production is evident; according to Pang et al. (2010) [34], who studied in a glasshouse experiment the response of ten perennial herbaceous legume species, found that under low P conditions several legumes produced more biomass than lucerne. Nutrient deficiency may cause physiological and metabolism abnormalities in plants, which may lead to deficiency symptoms. There are two categories of symptoms: i) General symptoms, such as limited growth and inability of reproduction (flowering and fruit setting), caused by the deficiency of many necessary macro- or micronutrients, and ii) typical, characteristic, deficiency symptoms, such as chlorosis, i.e. yellowing (due to Fe deficiency). In both cases biomass production is depressed. In the study of Msilini et al. (2009) [10], bicarbonate treated plants of *Arabidopsis thaliana* suffered from Fe deficiency displayed significantly lower biomass, leaf number and leaf surface, as compared to control plants, and showed slight yellowing of their younger leaves. Under limited nutrient availability, arbuscular mycorrhiza fungi (AMF) may favor nutrient uptake and thus enhance biomass production. Hu et al. (2009) [35] refer that AMF inoculation of maize plants was likely more efficient in extremely P-limited soils. Generally, root colonization by AMF influences positively plant growth under N, P, or micronutrient deficiency conditions [36].

### **2.8. Genotypic factors (root morphology and architecture, genetic growth capacity e.t.c.)**

According to Bayuelo-Jimenez et al. (2011) [3], under P deficiency, P-efficient accessions of maize plants (*Zea mays* L.) had greater root to shoot ratio, nodal rooting, nodal root laterals, nodal root hair density and length of nodal root main axis, and first-order laterals. In our experiments, we also found differential root system morphology among three Greek olive cultivars (the root systems of 'Koroneiki' and 'Chondrolia Chalkidikis' were less branched and more lateral, and with less root hair development and density, than that of 'Kothreiki', which was richly-branched and with much greater root hair development and density), something which was probably the main reason for the great genotypic variations in nutrient uptake and growth among the three cultivars (Chatzistathis, unpublished data). Singh et al. (2010) [37] found that great differences existed among 10 multipurpose tree species, grown in a monoculture tree cropping system on the sodic soils of Gangetic alluvium in north India, concerning plant height, diameter e.t.c.

### **3. Physiological roles of nutrients**

The absolutely necessary nutrients for plant growth are the following: N, P, K, S, Ca, Mg (macronutrients), Zn, Cu, B, Mn, Fe, Mo (micronutrients). Without one of these nutrients, plant organism can not grow normally and survive. The physiological roles of these nutrients are described in detail below.

### **3.1. Macronutrients**

430 Biomass Now – Cultivation and Utilization

branches of trees [31].

**2.4. Air humidity** 

**2.5. Photoperiod** 

*hippocastanus* [30].

**2.6. Light intensity** 

**2.7. Nutrient availability** 

rate of growth and biomass production.

greatest biomass production can be achieved.

for maximum photosynthetic activity is around 25oC for most vegetative species. When temperature exceeds 35oC photosynthesis is inhibited, thus biomass production may be restrained. High temperatures are associated with a high vapor pressure deficit between leaves and the surrounding air. The same applies to fruit, where high temperatures may cause fruit drop in olive trees [31]. On the other hand, low temperatures act negatively in photosynthesis function and starch is redistributed and is accumulated in organs protected from frost, such as roots. Very low temperatures (<-12oC) damage the leaf canopy, shoot and

Low atmosphere humidity speeds up transpiration by leaf surface. Increase of the rate of transpiration causes reduction of vegetative tissues water content, thus depression in the

Photoperiod is the duration of light in 24 hours and it is one of the most important factors influencing vegetative growth. Plant species whose vegetative growth is mostly influenced by long day conditions are *Populus robusta*, *Ulmus Americana* and *Aesculus* 

Light, together with CO2, are the two main factors influencing photosynthetic rate. By increasing light intensity up to an optimum limit the maximum photosynthetic rate, so the

Limited nutrient availability influences negatively biomass production. Nitrogen deficiency strongly depresses vegetation flush. According to Boussadia et al. (2010) [8], total biomass of two olive cultivars ('Meski' and 'Koroneiki') was strongly reduced (mainly caused by a decrease in leaf dry weight) under severe N deprivation, while in an out-door pot-culture experiment with castor bean plants (*Ricinus communis* L.), conducted by Reddy and Matcha (2010) [9], it was found that among the plant components, leaf dry weight had the greatest decrease; furthermore, root/shoot ratio increased under N deficiency [9]. Phosphorus deficiency caused reduced biomass, photosynthetic activity and nitrogen fixing ability in mungbean (*Vigna aconitifolia*) and mashbean (*Vigna radiata*) [33]. Under P deficiency conditions, genotypic variation in biomass production is evident; according to Pang et al. (2010) [34], who studied in a glasshouse experiment the response of ten perennial herbaceous legume species, found that under low P conditions several legumes produced more biomass than lucerne. Nutrient deficiency may cause physiological and metabolism abnormalities in plants, which may lead to deficiency symptoms. There are two categories of

Nitrogen: It is a primary component of nucleic acids, proteins, amino acids, purines, pyrimidines and chlorophyll. Nitrogen exerts a significant effect on plant growth, as it reduces biennial bearing and increases the percentage of perfect flowers. In olive trees, lack of N leads to decreased growth, shorter length of annual shoots (<10cm), fewer leaves, reduced flowering and decreased yield [31].

Phosphorus: P is a component of high-energy substances such as ATP, ADP and AMP; it is also important for nucleic acids and phospholipids. Phosphorus affects root growth and maturation of plant tissues and participates in the metabolism of carbohydrates, lipids and proteins [31].

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

Boron: B plays role in the transfer of sugars along cell membranes, as well as in RNA and DNA synthesis. It also participates to cell division process, as well as to the pectine synthesis

Molybdenum: It is part of the enzyme nitrogenase (capturing of atmospheric N) and nitric

As it is clear from all the above physiological roles of nutrients, the deficiency of even one of them in the mineral nutrition of higher plants depresses their growth, thus biomass production. So, in order to achieve the maximum biomass production, apart from the optimum conditions of all the other environmental and agronomic factors influencing plant growth (temperature, soil humidity, photoperiod, light intensity), it should always be taken care of maintaining the optimum levels of all the necessary soil nutrients. This is usually achieved with the correct fertilization program of the different crops. For example, fruit trees have high demands in K, since fruit production is a K sink and reduces its levels in plant level. According to Therios (2009) [31], potassium plays an important role in olive nutrition. Thus, fruit trees should be periodically fertilized (usually K fertilizers applied during autumn, or winter, and are incorporated into the soils) with enhanced doses of potassium fertilizers (usually K2SO4). Apart from chemical fertilizers, organic amendments can be also applied under limited nutrient conditions in order to enhance plant growth. According to Hu et al. (2009) [35], stem length, shoot and root biomass, as well as crop yield of maize were all greatly increased by the application of organic amendments on a sandy loam soil. Apart from the application of chemical fertilizers, organic amendments e.t.c., another modern method to improve yields and to increase biomass is the irrigation of crops with FFC H2O, a commercial product currently utilized by the agriculture, fishery and food industries in Japan. In the study of Konkol et al. (2012) [39], radish and shirona plants irrigated with FFC H2O developed larger average leaf area by 122% and greater dry weight and stem length by 39% and 31%, respectively, compared to the plants irrigated with deionized H2O. FFC H2O offers agriculturalists a simple and effective tool for the fortification of irrigation waters with

**4. Nutrient utilization efficiency (NUE): The case of nutrient use efficient** 

World population is expected to increase from 6.0 billion in 1999 to 8.5 billion by 2025. Such an increase in population will intensify pressure on the world's natural resource base (land, water, and air) to achieve higher food production. Increased food production could be achieved by expanding the land area under crops and by increasing yields per unit area through intensive farming. Chemical fertilizers are one of the expensive inputs used by farmers to achieve desired crop yields [40]. However, during the last years, the prices of fertilizers have been considerably increased. Furthermore, soil degradation and pollution, as well as underground water pollution, are serious consequences provoked by the exaggerate

to NO2-

[38].

ascorbic acid [38].

micronutrients [39].

**genotypes** 

reductase (transformation of NO3-

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 433

). Mo also participates to the metabolism of

Potassium: K plays a crucial role in carbohydrate metabolism, in the metabolism of N and protein synthesis, in enzyme activities, in the regulation of the opening and closing of stomata, thus to the operation of photosynthesis, in the improvement of fruit quality and disease tolerance, in the activation of the enzymes peptase, catalase, pyruvic kinase e.t.c. [31,38].

Calcium: It is the element that participates in the formation and integrity of cell membranes, in the integrity and semipermeability of the plasmalemma, it increases the activity of many enzymes, it plays a crucial role in cell elongation and division, in the transfer of carbohydrates e.t.c. [31,38].

Magnesium: It is part of chlorophyll molecule, it activates the enzymes of Crebs' cycle and it also plays a role in oil synthesis [38].

Sulphur: Sulphur plays role in the synthesis of some amino-acids, such as cysteine, cystine, methionine, as well as in proteins synthesis. It also activates some proteolytic enzymes, such as papaine, bromeline e.t.c. Finally, it is part of some vitamins' molecule and that of gloutathione [31,38].

### **3.2. Micronutrients**

Iron: Iron plays an important role in chlorophyll synthesis, without being part of its molecule. Furthermore, it participates in the molecule of Fe-proteins catalase, cytochrome a, b, c, hyperoxidase e.t.c. In addition to that, it is found in the enzymes nitric and nitrate reductase, which are responsible for the transformation of NO3 into NH4+, as well as in nitrogenase, which is the responsible enzyme for the atmospheric N capturing [38].

Manganese: Manganese is activator of the enzymes of carbohydrates metabolism, those of Crebs' cycle, and of some other enzymes, such as cysteine desulphydrase, glutamyl transferase e.t.c. It also plays a key-role in photosystem II of photosynthesis, and particularly in the reactions liberating O2. Finally, Mn acts as activator of some enzymes catalyzing oxidation and reduction reactions [38].

Zinc: Zn plays crucial role in tryptophane biosynthesis, which is the previous stage from IAA (auxin) synthesis (direct influence of Zn on plant growth and biomass production). IAA concentration is significantly reduced in vegetative tissues suffering from Zn deficiency. In addition to the above, Zn is part of some metal-enzymes [38].

Copper: Cu is activator of some enzymes, as well as it is part of enzymes catalyzing oxidation and reducing reactions, such as oxidase of ascorbic acid, lactase, nitrate and nitric reductase e.t.c. [38].

Boron: B plays role in the transfer of sugars along cell membranes, as well as in RNA and DNA synthesis. It also participates to cell division process, as well as to the pectine synthesis [38].

432 Biomass Now – Cultivation and Utilization

carbohydrates e.t.c. [31,38].

gloutathione [31,38].

**3.2. Micronutrients** 

reductase e.t.c. [38].

also plays a role in oil synthesis [38].

proteins [31].

[31,38].

Phosphorus: P is a component of high-energy substances such as ATP, ADP and AMP; it is also important for nucleic acids and phospholipids. Phosphorus affects root growth and maturation of plant tissues and participates in the metabolism of carbohydrates, lipids and

Potassium: K plays a crucial role in carbohydrate metabolism, in the metabolism of N and protein synthesis, in enzyme activities, in the regulation of the opening and closing of stomata, thus to the operation of photosynthesis, in the improvement of fruit quality and disease tolerance, in the activation of the enzymes peptase, catalase, pyruvic kinase e.t.c.

Calcium: It is the element that participates in the formation and integrity of cell membranes, in the integrity and semipermeability of the plasmalemma, it increases the activity of many enzymes, it plays a crucial role in cell elongation and division, in the transfer of

Magnesium: It is part of chlorophyll molecule, it activates the enzymes of Crebs' cycle and it

Sulphur: Sulphur plays role in the synthesis of some amino-acids, such as cysteine, cystine, methionine, as well as in proteins synthesis. It also activates some proteolytic enzymes, such as papaine, bromeline e.t.c. Finally, it is part of some vitamins' molecule and that of

Iron: Iron plays an important role in chlorophyll synthesis, without being part of its molecule. Furthermore, it participates in the molecule of Fe-proteins catalase, cytochrome a, b, c, hyperoxidase e.t.c. In addition to that, it is found in the enzymes nitric and nitrate

Manganese: Manganese is activator of the enzymes of carbohydrates metabolism, those of Crebs' cycle, and of some other enzymes, such as cysteine desulphydrase, glutamyl transferase e.t.c. It also plays a key-role in photosystem II of photosynthesis, and particularly in the reactions liberating O2. Finally, Mn acts as activator of some enzymes

Zinc: Zn plays crucial role in tryptophane biosynthesis, which is the previous stage from IAA (auxin) synthesis (direct influence of Zn on plant growth and biomass production). IAA concentration is significantly reduced in vegetative tissues suffering from Zn deficiency. In

Copper: Cu is activator of some enzymes, as well as it is part of enzymes catalyzing oxidation and reducing reactions, such as oxidase of ascorbic acid, lactase, nitrate and nitric

nitrogenase, which is the responsible enzyme for the atmospheric N capturing [38].

into NH4+, as well as in

reductase, which are responsible for the transformation of NO3-

catalyzing oxidation and reduction reactions [38].

addition to the above, Zn is part of some metal-enzymes [38].

Molybdenum: It is part of the enzyme nitrogenase (capturing of atmospheric N) and nitric reductase (transformation of NO3 to NO2- ). Mo also participates to the metabolism of ascorbic acid [38].

As it is clear from all the above physiological roles of nutrients, the deficiency of even one of them in the mineral nutrition of higher plants depresses their growth, thus biomass production. So, in order to achieve the maximum biomass production, apart from the optimum conditions of all the other environmental and agronomic factors influencing plant growth (temperature, soil humidity, photoperiod, light intensity), it should always be taken care of maintaining the optimum levels of all the necessary soil nutrients. This is usually achieved with the correct fertilization program of the different crops. For example, fruit trees have high demands in K, since fruit production is a K sink and reduces its levels in plant level. According to Therios (2009) [31], potassium plays an important role in olive nutrition. Thus, fruit trees should be periodically fertilized (usually K fertilizers applied during autumn, or winter, and are incorporated into the soils) with enhanced doses of potassium fertilizers (usually K2SO4). Apart from chemical fertilizers, organic amendments can be also applied under limited nutrient conditions in order to enhance plant growth. According to Hu et al. (2009) [35], stem length, shoot and root biomass, as well as crop yield of maize were all greatly increased by the application of organic amendments on a sandy loam soil. Apart from the application of chemical fertilizers, organic amendments e.t.c., another modern method to improve yields and to increase biomass is the irrigation of crops with FFC H2O, a commercial product currently utilized by the agriculture, fishery and food industries in Japan. In the study of Konkol et al. (2012) [39], radish and shirona plants irrigated with FFC H2O developed larger average leaf area by 122% and greater dry weight and stem length by 39% and 31%, respectively, compared to the plants irrigated with deionized H2O. FFC H2O offers agriculturalists a simple and effective tool for the fortification of irrigation waters with micronutrients [39].

### **4. Nutrient utilization efficiency (NUE): The case of nutrient use efficient genotypes**

World population is expected to increase from 6.0 billion in 1999 to 8.5 billion by 2025. Such an increase in population will intensify pressure on the world's natural resource base (land, water, and air) to achieve higher food production. Increased food production could be achieved by expanding the land area under crops and by increasing yields per unit area through intensive farming. Chemical fertilizers are one of the expensive inputs used by farmers to achieve desired crop yields [40]. However, during the last years, the prices of fertilizers have been considerably increased. Furthermore, soil degradation and pollution, as well as underground water pollution, are serious consequences provoked by the exaggerate

use of fertilizers during last decades. These two aspects are responsible for the global concern to reduce the use of fertilizers. The best way to do that is by selecting and growing nutrient use efficient genotypes. According to Khoshgoftarmanesh (2009) [41], cultivation and breeding of micronutrient-efficient genotypes in combination with proper agronomic management practices appear as the most sustainable and cost-effective solution for alleviating food-chain micronutrient deficiency.

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

reasons (genotypic ability to absorb and utilize efficiently, or inefficiently, soil nutrients), b) mycorrhiza colonization of the root system, c) differential root exudation of organic compounds favorizing nutrient uptake, d) different properties of rhizosphere, e) other reasons. According to Cakmak (2002) [49], integration of plant nutrition research with plant genetics and molecular biology is indispensable in developing plant genotypes with high genetic ability to adapt to nutrient deficient and toxic soil conditions and to allocate more micronutrients into edible plant products. According to Aziz et al. (2011b) [50], *Brassica* cultivars with high biomass and high P contents, such as 'Rainbow' and 'Poorbi Raya', at low available P conditions would be used in further screening experiments to improve P efficiency in *Brassica*. More specifically, a number of genes have been isolated and cloned, which are involved in root exudation of nutrient-mobilizing organic compounds [51,52]. Successful attempts have been made in the past 5 years to develop transgenic plants that produce and release large amounts of organic acids, which are considered to be key compounds involved in the adaptive mechanisms used by plants to tolerate P-deficient soil conditions [53-55]. However, differential root exudation ability in nature exists among different plant species. According to Maruyama et al. (2005) [56], who made a comparison of iron availability in leaves of barley and rice, the difference in the Fe acquisition ability between these two species was affected by the differential mugineic acid secretion. Chatzistathis et al. (2009) [12] refer that, maybe, a similar mechanism was responsible for the differential micronutrient uptake and accumulation between the Greek olive cultivars 'Koroneiki' and 'Kothreiki'. According to the same authors, differential reduction of Fe3+ to Fe2+, or acidification capacity of root apoplast (which associates with the increase of Fe3+-chelate reductase and H-ATPase activities) among three Greek olive cultivars should not be excluded from possible causes for the significant differences observed concerning Fe uptake [14]. Mycorrhiza root colonization may be another responsible factor for the differential micronutrient utilization efficiency among genotypes. According to Citernesi et al. (1998) [57], arbuscular mycorrhiza fungi (AMF) influenced root morphology of Italian olive cultivars, thus nutrient uptake and accumulation, as well as plant growth. In our study with olive cultivars 'Koroneiki', 'Kothreiki' and 'Chondrolia Chalkidikis', we found significant differences concerning root colonization by AMF (that varied from 45% to 73%), together with great differences in uptake and utilization efficiency of Mn, Fe and Zn among them (particularly, 1.5 to 10.5 times greater amount of Mn, Fe and Zn accumulated by 'Kothreiki', compared to the other two cultivars, but the differences in plant growth parameters between the three cultivars were not impressive; this is why the micronutrient utilization efficiency by 'Kothreiki' was significantly lower, compared to that of the other two ones). Finally, the different properties of rhizosphere among genotypes may be another important factor influencing nutrient uptake and utilization efficiency, and of course biomass production. According to Rengel (2001) [58], who made a review on genotypic differences in micronutrient use efficiency of many crops, micronutrient-efficient genotypes were capable of increasing soil available micronutrient pools through changing the chemical and microbiological properties of the rhizosphere, as well as by growing thinner and longer roots and by

having more efficient uptake and transport mechanisms.

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 435

Nutrient use efficient genotypes are those having the ability to produce high yields under conditions of limited nutrient availability. According to Chapin and Van Cleve (1991) [11] and Gourley et al. (1994) [42], as nutrient utilization efficiency (NUE) is defined the amount of biomass produced per unit of nutrient absorbed. **Nutrient efficiency ratio (NER)** was suggested by Gerloff and Gabelman (1983) [43] to differentiate genotypes into efficient and inefficient nutrient utilizers, i.e. **NER=(Units of Yields, kgs)/(Unit of elements in tissue, kg),** while **Agronomic efficiency (AE)** is expressed as the additional amount of economic yield per unit nutrient applied, i.e. **AE=(Yield F, kg-Yield C, kg)/(quantity of nutrient applied, kg),** where F applies for plants receiving fertilizer and C for plants receiving no fertilizer.

Many researchers found significant differences concerning nutrient utilization efficiency among genotypes (cultivars) of the same plant species [1,12,13,40,44-46] Biomass (shoot and root dry matter production) was used as an indicator in order to assess Zn efficient Chinese maize genotypes, grown for 30 days in a greenhouse pot experiment under Zn limiting conditions [1]. NUE is based on: a) uptake efficiency, b) incorporation efficiency and c) utilization efficiency [40]. The uptake efficiency is the ability of a genotype to absorb nutrients from the soil; however, the great ability to absorb nutrients does not necessarily mean that this genotype is nutrient use efficient. According to Jiang and Ireland (2005) [45], and Jiang (2006) [46], Mn efficient wheat cultivars own this ability to a better internal utilization of Mn, rather than to a higher plant Mn accumulation. We also found in our experiments that, despite the fact that the olive cultivar 'Kothreiki' absorbed and accumulated significantly greater quantity of Mn and Fe in three soil types, compared to 'Koroneiki', the second one was more Mn and Fe-efficient due to its better internal utilization efficiency of Mn and Fe (greater transport of these micronutrients from root to shoots) [12] (Tables 1 and 2). Aziz et al. (2011a) [47] refer that under P deficiency conditions, P content of young leaves in *Brassica* cultivars increased by two folds, indicating remobilization of this nutrient from older leaves and shoot. However, differences in P remobilization among *Brassica* cultivars could not explain the differences in P utilization. Phosphorus efficient wheat genotypes with greater root biomass, higher P uptake potential in shoots and absorption rate of P were generally more tolerant to P deficiency in the growth medium [6]. According to Yang et al. (2011) [48], on average, the K efficient cotton cultivars produced 59% more potential economic yield (dry weight of all reproductive organs) under field conditions even with available soil K at obviously deficient level (60 mg/kg).

The possible causes for the differential nutrient utilization efficiency among genotypes and/or species may be one, or combination of more than one, of the following: a) genetic

micronutrient deficiency.

deficient level (60 mg/kg).

fertilizer.

use of fertilizers during last decades. These two aspects are responsible for the global concern to reduce the use of fertilizers. The best way to do that is by selecting and growing nutrient use efficient genotypes. According to Khoshgoftarmanesh (2009) [41], cultivation and breeding of micronutrient-efficient genotypes in combination with proper agronomic management practices appear as the most sustainable and cost-effective solution for alleviating food-chain

Nutrient use efficient genotypes are those having the ability to produce high yields under conditions of limited nutrient availability. According to Chapin and Van Cleve (1991) [11] and Gourley et al. (1994) [42], as nutrient utilization efficiency (NUE) is defined the amount of biomass produced per unit of nutrient absorbed. **Nutrient efficiency ratio (NER)** was suggested by Gerloff and Gabelman (1983) [43] to differentiate genotypes into efficient and inefficient nutrient utilizers, i.e. **NER=(Units of Yields, kgs)/(Unit of elements in tissue, kg),** while **Agronomic efficiency (AE)** is expressed as the additional amount of economic yield per unit nutrient applied, i.e. **AE=(Yield F, kg-Yield C, kg)/(quantity of nutrient applied, kg),** where F applies for plants receiving fertilizer and C for plants receiving no

Many researchers found significant differences concerning nutrient utilization efficiency among genotypes (cultivars) of the same plant species [1,12,13,40,44-46] Biomass (shoot and root dry matter production) was used as an indicator in order to assess Zn efficient Chinese maize genotypes, grown for 30 days in a greenhouse pot experiment under Zn limiting conditions [1]. NUE is based on: a) uptake efficiency, b) incorporation efficiency and c) utilization efficiency [40]. The uptake efficiency is the ability of a genotype to absorb nutrients from the soil; however, the great ability to absorb nutrients does not necessarily mean that this genotype is nutrient use efficient. According to Jiang and Ireland (2005) [45], and Jiang (2006) [46], Mn efficient wheat cultivars own this ability to a better internal utilization of Mn, rather than to a higher plant Mn accumulation. We also found in our experiments that, despite the fact that the olive cultivar 'Kothreiki' absorbed and accumulated significantly greater quantity of Mn and Fe in three soil types, compared to 'Koroneiki', the second one was more Mn and Fe-efficient due to its better internal utilization efficiency of Mn and Fe (greater transport of these micronutrients from root to shoots) [12] (Tables 1 and 2). Aziz et al. (2011a) [47] refer that under P deficiency conditions, P content of young leaves in *Brassica* cultivars increased by two folds, indicating remobilization of this nutrient from older leaves and shoot. However, differences in P remobilization among *Brassica* cultivars could not explain the differences in P utilization. Phosphorus efficient wheat genotypes with greater root biomass, higher P uptake potential in shoots and absorption rate of P were generally more tolerant to P deficiency in the growth medium [6]. According to Yang et al. (2011) [48], on average, the K efficient cotton cultivars produced 59% more potential economic yield (dry weight of all reproductive organs) under field conditions even with available soil K at obviously

The possible causes for the differential nutrient utilization efficiency among genotypes and/or species may be one, or combination of more than one, of the following: a) genetic reasons (genotypic ability to absorb and utilize efficiently, or inefficiently, soil nutrients), b) mycorrhiza colonization of the root system, c) differential root exudation of organic compounds favorizing nutrient uptake, d) different properties of rhizosphere, e) other reasons. According to Cakmak (2002) [49], integration of plant nutrition research with plant genetics and molecular biology is indispensable in developing plant genotypes with high genetic ability to adapt to nutrient deficient and toxic soil conditions and to allocate more micronutrients into edible plant products. According to Aziz et al. (2011b) [50], *Brassica* cultivars with high biomass and high P contents, such as 'Rainbow' and 'Poorbi Raya', at low available P conditions would be used in further screening experiments to improve P efficiency in *Brassica*. More specifically, a number of genes have been isolated and cloned, which are involved in root exudation of nutrient-mobilizing organic compounds [51,52]. Successful attempts have been made in the past 5 years to develop transgenic plants that produce and release large amounts of organic acids, which are considered to be key compounds involved in the adaptive mechanisms used by plants to tolerate P-deficient soil conditions [53-55]. However, differential root exudation ability in nature exists among different plant species. According to Maruyama et al. (2005) [56], who made a comparison of iron availability in leaves of barley and rice, the difference in the Fe acquisition ability between these two species was affected by the differential mugineic acid secretion. Chatzistathis et al. (2009) [12] refer that, maybe, a similar mechanism was responsible for the differential micronutrient uptake and accumulation between the Greek olive cultivars 'Koroneiki' and 'Kothreiki'. According to the same authors, differential reduction of Fe3+ to Fe2+, or acidification capacity of root apoplast (which associates with the increase of Fe3+-chelate reductase and H-ATPase activities) among three Greek olive cultivars should not be excluded from possible causes for the significant differences observed concerning Fe uptake [14]. Mycorrhiza root colonization may be another responsible factor for the differential micronutrient utilization efficiency among genotypes. According to Citernesi et al. (1998) [57], arbuscular mycorrhiza fungi (AMF) influenced root morphology of Italian olive cultivars, thus nutrient uptake and accumulation, as well as plant growth. In our study with olive cultivars 'Koroneiki', 'Kothreiki' and 'Chondrolia Chalkidikis', we found significant differences concerning root colonization by AMF (that varied from 45% to 73%), together with great differences in

uptake and utilization efficiency of Mn, Fe and Zn among them (particularly, 1.5 to 10.5 times greater amount of Mn, Fe and Zn accumulated by 'Kothreiki', compared to the other two cultivars, but the differences in plant growth parameters between the three cultivars were not impressive; this is why the micronutrient utilization efficiency by 'Kothreiki' was significantly lower, compared to that of the other two ones). Finally, the different properties of rhizosphere among genotypes may be another important factor influencing nutrient uptake and utilization efficiency, and of course biomass production. According to Rengel (2001) [58], who made a review on genotypic differences in micronutrient use efficiency of many crops, micronutrient-efficient genotypes were capable of increasing soil available micronutrient pools through changing the chemical and microbiological properties of the rhizosphere, as well as by growing thinner and longer roots and by having more efficient uptake and transport mechanisms.


How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

Soil Cultivar MnUE FeUE ZnUE **Marl** mg of the total plant d.w./μg of the total per plant

The different letters in the same column symbolize statistically significant differences between the two cultivars in each

**Table 2.** Nutrient utilization efficiency (mg of the total plant d.w. /μg of the total per plant quantity of micronutrient or mg of the total per plant quantity of macronutrient) of the olive cultivars 'Koroneiki' and 'Kothreiki', when each of them was grown in three soils (from parent material Marl, Gneiss schist.

Soil heavy metal contamination has become an increasing problem worldwide. Among the heavy metals, Cu, Zn, Mn, Cd, Pb, Ni and Cr are considered to be the most common toxicity problems causing increasing concern. Growth inhibition and reduced yield are common responses of horticultural crops to nutrient and heavy metal toxicity [2]. Nevertheless, sometimes less common responses happen under metal toxicity conditions. For example, in the case of Pb it has been suggested that inhibition of root growth is one of the primary effects of Pb toxicity through the inhibition of cell division at the root tip [59]. Significant reductions in plant height, as well as in shoot and root dry weight (varying from 3.3% to 54.5%), as compared with that of the controls, were found for *Typha angustifolia* plants in different Cr treatments [60]. Furthermore, according to Caldelas et al. (2012) [19], not only growth inhibition happened (reached 65% dry weight) under Cr toxicity conditions, but also root/shoot partitioning increased by 80%. Under Cr stress conditions, it was found that root and shoot biomass of *Genipa americana* L. were significantly reduced [20]. The biomass reduction of *Genipa americana* trees is ascribed, according to the same authors, to the decreased net photosynthetic rates and to the limitations in stomatal conductance. The disorganization of chloroplast structure and inhibition of electron transport is a possible explanation for the decreased photosynthetic rates of trees exposed to Cr stress [20]. In contrast to the above, Cd and Pb applications induced slight or even significant increase in plant height and biomass. The fact that Cd and Pb addition enhanced Ca and Fe uptake suggests that these two nutrients may play a role in heavy metal detoxification by *Typha angustifolia* plants; furthermore, increased Zn uptake may also contribute to its hyper Pb tolerance, as recorder in the increased biomass over the control plants [60]. According to the

and Peridotite) with different physicochemical properties (Chatzistathis et al., 2009).

**5. The influence of heavy metal toxicity on biomass production** 

**Gneiss schist**

**Peridotite** 

of the three soils, for *P*≤0.05 (n=6) (SPSS; t-test).

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 437

**Kor** 31.85a 1.73a 77.53a **Koth** 18.68b 0.65b 68.08a

**Kor** 39.87a 1.84a 51.04a **Koth** 17.94b 0.44b 49.15a

**Kor** 23.33a 1.19a 61.75a **Koth** 18.00a 0.58b 72.88a

quantity of micronutrient

The different letters in the same column symbolize statistically significant differences between the two olive cultivars in each of the three soils, for *P*≤0.05 (n=6) (SPSS; t-test).

**Table 1.** Distribution (%) of the total per plant quantity of Mn, Fe and Zn in the three vegetative tissues (root, stem and leaves) of the olive cultivars 'Koroneiki' and 'Kothreiki', when each one was grown in three soils (from parent material Marl, Gneiss schist. and Peridotite) with different physicochemical properties (Chatzistathis et al., 2009).

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 437

436 Biomass Now – Cultivation and Utilization

**Gneiss schist** 

**Peridotite** 

**Gneiss schist** 

**Peridotite** 

**Gneiss schist** 

**Peridotite** 

**Marl Mn**

**Marl Fe**

**Marl Zn**

in each of the three soils, for *P*≤0.05 (n=6) (SPSS; t-test).

properties (Chatzistathis et al., 2009).

Soil Cultivar Micronutrient Root Stem Leaves

 **Kor** 56.5b 34.2a 9.3a

 **Kor** 44.0b 44.0a 12.0a

 **Kor** 94.0a 3.7a 2.3a

 **Kor** 90.8a 7.1a 2.1a

 **Kor** 59.1b 26.7a 14.2a

 **Kor** 37.3b 33.9a 28.8a

The different letters in the same column symbolize statistically significant differences between the two olive cultivars

**Table 1.** Distribution (%) of the total per plant quantity of Mn, Fe and Zn in the three vegetative tissues (root, stem and leaves) of the olive cultivars 'Koroneiki' and 'Kothreiki', when each one was grown in three soils (from parent material Marl, Gneiss schist. and Peridotite) with different physicochemical

**Kor** 50.2b 38.0a 11.8a **Koth** 74.1a 12.8b 13.1a

**Koth** 81.3a 10.8b 7.9a

**Koth** 76.0a 12.9b 11.1a 

**Kor** 93.7a 3.9a 2.4a **Koth** 98.0a 0.9b 1.1b

**Koth** 98.8a 0.6b 0.6b

**Koth** 98.3a 0.8b 0.9b 

**Kor** 49.3b 29.6a 21.1a **Koth** 64.4a 15.6b 20.0a

**Koth** 73.7a 14.3b 12.0a

**Koth** 65.3a 18.0b 16.7b


The different letters in the same column symbolize statistically significant differences between the two cultivars in each of the three soils, for *P*≤0.05 (n=6) (SPSS; t-test).

**Table 2.** Nutrient utilization efficiency (mg of the total plant d.w. /μg of the total per plant quantity of micronutrient or mg of the total per plant quantity of macronutrient) of the olive cultivars 'Koroneiki' and 'Kothreiki', when each of them was grown in three soils (from parent material Marl, Gneiss schist. and Peridotite) with different physicochemical properties (Chatzistathis et al., 2009).

### **5. The influence of heavy metal toxicity on biomass production**

Soil heavy metal contamination has become an increasing problem worldwide. Among the heavy metals, Cu, Zn, Mn, Cd, Pb, Ni and Cr are considered to be the most common toxicity problems causing increasing concern. Growth inhibition and reduced yield are common responses of horticultural crops to nutrient and heavy metal toxicity [2]. Nevertheless, sometimes less common responses happen under metal toxicity conditions. For example, in the case of Pb it has been suggested that inhibition of root growth is one of the primary effects of Pb toxicity through the inhibition of cell division at the root tip [59]. Significant reductions in plant height, as well as in shoot and root dry weight (varying from 3.3% to 54.5%), as compared with that of the controls, were found for *Typha angustifolia* plants in different Cr treatments [60]. Furthermore, according to Caldelas et al. (2012) [19], not only growth inhibition happened (reached 65% dry weight) under Cr toxicity conditions, but also root/shoot partitioning increased by 80%. Under Cr stress conditions, it was found that root and shoot biomass of *Genipa americana* L. were significantly reduced [20]. The biomass reduction of *Genipa americana* trees is ascribed, according to the same authors, to the decreased net photosynthetic rates and to the limitations in stomatal conductance. The disorganization of chloroplast structure and inhibition of electron transport is a possible explanation for the decreased photosynthetic rates of trees exposed to Cr stress [20]. In contrast to the above, Cd and Pb applications induced slight or even significant increase in plant height and biomass. The fact that Cd and Pb addition enhanced Ca and Fe uptake suggests that these two nutrients may play a role in heavy metal detoxification by *Typha angustifolia* plants; furthermore, increased Zn uptake may also contribute to its hyper Pb tolerance, as recorder in the increased biomass over the control plants [60]. According to the

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

same authors (Bah et al., 2011), plants have mechanisms that allow them to tolerate relatively

Tzerakis et al. (2012) [2] found that excessively high concentrations of Mn and Zn in the leaves of cucumber (reached 900 and 450 mg/kg d.w., respectively), grown hydroponically under toxic Mn and Zn conditions, reduced the fruit biomass due to decreases in the number of fruits per plants, as well as in the net assimilation rate, stomatal conductance and transpiration rate. However, it was found that significant differences concerning biomass production between different species of the same genus exist under metal toxicity conditions; *Melilotus officinalis* seems to be more tolerant to Pb than *Melilotus alba* because no differences in shoot or root length, or number of leaves, were found between control plants and those grown under 200 and 1000 mg/kg Pb [15]. In addition to the above, genotypic differences between cultivars of the same species, concerning biomass production, under metal toxicity conditions may also be observed; Chatzistathis et al. (2012) [13] found that under excess Mn conditions (640 μΜ), plant growth parameters (shoot elongation, as well as fresh and dry weights of leaves, root and stem) of olive cultivar 'Picual' were significantly decreased, compared to those of the control plants (2 μΜ), something which did not happen in olive cultivar 'Koroneiki' (no significant differences were recorder between the two Mn treatments) (Figure 1). According to the same authors, some factors related to the better tolerance of 'Koroneiki' not only at whole plant level, but also at tissue and cell level, could take place. Such possible factors could be a better compartmentalization of Mn within cells and/or functionality of Mn detoxification systems [13]. Significant growth reductions of several plant species, grown under Mn toxicity conditions, have been mentioned by several

Nickel (Ni) toxicity, which may be a serious problem around industrial areas, can also cause biomass reduction. At high soil Ni levels (>200 mg/kg soil) reduced growth symptoms of *Riccinus communis* plants were observed [18]. According to Baccouch et al. (1998) [66], the higher concentrations of Ni have been reported to retard cell division, elongation, differentiation, as well as to affect plant growth and development. Excess Cd, which causes direct or indirect inhibition of physiological processes, such as transpiration, photosynthesis, oxidative stress, cell elongation, N metabolism and mineral nutrition may lead in growth retardation, leaf chlorosis and low biomass production [67]. According to the same authors, Cd stress could induce serious damage in root cells of grey poplar (*Populus x canescens*). Arsenic (As) toxicity may be another (although less common) problem contributing to soil contamination. Repeated and widespread use of arsenical pesticides has significantly contributed to soil As contamination [4]. According to the same authors, plant growth parameters, such as biomass, shoot height, and root length, decreased with increased As

Soil pollution represents a risk to human health in various ways including contamination of food, grown in polluted soils, as well as contamination of groundwater surface soils [68].

high concentrations of Pb in their environment without suffering from toxic effects.

researchers [61-65].

concentrations in all soils.

**6. Phytoremediation** 

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 439

**Figure 1.** Shoot elongation of olive cultivars 'Picual' (A) and 'Koroneiki' (B), when grown under hydroponics at normal (2 μΜ) and excess Mn conditions (640 μΜ Mn) (Chatzistathis et al., 2012).

same authors (Bah et al., 2011), plants have mechanisms that allow them to tolerate relatively high concentrations of Pb in their environment without suffering from toxic effects.

Tzerakis et al. (2012) [2] found that excessively high concentrations of Mn and Zn in the leaves of cucumber (reached 900 and 450 mg/kg d.w., respectively), grown hydroponically under toxic Mn and Zn conditions, reduced the fruit biomass due to decreases in the number of fruits per plants, as well as in the net assimilation rate, stomatal conductance and transpiration rate. However, it was found that significant differences concerning biomass production between different species of the same genus exist under metal toxicity conditions; *Melilotus officinalis* seems to be more tolerant to Pb than *Melilotus alba* because no differences in shoot or root length, or number of leaves, were found between control plants and those grown under 200 and 1000 mg/kg Pb [15]. In addition to the above, genotypic differences between cultivars of the same species, concerning biomass production, under metal toxicity conditions may also be observed; Chatzistathis et al. (2012) [13] found that under excess Mn conditions (640 μΜ), plant growth parameters (shoot elongation, as well as fresh and dry weights of leaves, root and stem) of olive cultivar 'Picual' were significantly decreased, compared to those of the control plants (2 μΜ), something which did not happen in olive cultivar 'Koroneiki' (no significant differences were recorder between the two Mn treatments) (Figure 1). According to the same authors, some factors related to the better tolerance of 'Koroneiki' not only at whole plant level, but also at tissue and cell level, could take place. Such possible factors could be a better compartmentalization of Mn within cells and/or functionality of Mn detoxification systems [13]. Significant growth reductions of several plant species, grown under Mn toxicity conditions, have been mentioned by several researchers [61-65].

Nickel (Ni) toxicity, which may be a serious problem around industrial areas, can also cause biomass reduction. At high soil Ni levels (>200 mg/kg soil) reduced growth symptoms of *Riccinus communis* plants were observed [18]. According to Baccouch et al. (1998) [66], the higher concentrations of Ni have been reported to retard cell division, elongation, differentiation, as well as to affect plant growth and development. Excess Cd, which causes direct or indirect inhibition of physiological processes, such as transpiration, photosynthesis, oxidative stress, cell elongation, N metabolism and mineral nutrition may lead in growth retardation, leaf chlorosis and low biomass production [67]. According to the same authors, Cd stress could induce serious damage in root cells of grey poplar (*Populus x canescens*). Arsenic (As) toxicity may be another (although less common) problem contributing to soil contamination. Repeated and widespread use of arsenical pesticides has significantly contributed to soil As contamination [4]. According to the same authors, plant growth parameters, such as biomass, shoot height, and root length, decreased with increased As concentrations in all soils.

### **6. Phytoremediation**

438 Biomass Now – Cultivation and Utilization

**Figure 1.** Shoot elongation of olive cultivars 'Picual' (A) and 'Koroneiki' (B), when grown under hydroponics at normal (2 μΜ) and excess Mn conditions (640 μΜ Mn) (Chatzistathis et al., 2012). Soil pollution represents a risk to human health in various ways including contamination of food, grown in polluted soils, as well as contamination of groundwater surface soils [68]. Classical remediation techniques such as soil washing, excavation, and chelate extraction are all labor-intensive and costly [69].

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

Santana et al. (2012) [20] refer that *Genipa americana* L. is a tree species that tolerates high levels of Cr3+, therefore it can be used in recomposition of ciliary forests at Cr-polluted watersheds. According to the same authors, this woody species demonstrates a relevant capacity for phytoremediation of Cr. *Elsholtzia splendens* is regarded as a Cu tolerant and accumulating plant species [77]. Peng et al. (2012) [78] refer that *Eucalyptus urophylla X E.grandis* is a fast growing economic species that contributes to habitat restoration of degraded environments, such as the Pb contaminated ones. On the other hand, concerning Cd phytoextraction ability, only a few plant species have been accepted as Cd hyperaccumulators, including *Brassica juncea, Thlaspi caerulescens* and *Solanum nigrum*. Poplar (*Populus* L.), which is an easy to propagate and establish species and it has also the advantages of rapid growth, high biomass production, as well as the ability to accumulate high heavy metal concentrations, could be used as a Cd-hypaeraccumulator for phytoremediation [27-28,67]. According to Wang et al. (2012) [28], the increase in total Cd uptake by poplar genotypes in Cd contaminated soils is the result of enhanced biomass production under elevated CO2 conditions. Furthermore, *Amaranthus hypochondriacus* is a high biomass, fast growing and easily cultivated potential Cd hyperaccumulator [25]. Another species was found to be a good phytoremediator concerning its phytoaccumulation and tolerance to Ni stress is *Riccinus communis* L. [18]. Finally, *Justicia gendarussa*, which was proved to be able to tolerate and accumulate high concentration of heavy metals (and

Differences between species, or genotypes of the same species, concerning heavy metal accumulation have been found by many researchers. According to Dheri et al. (2007) [17], the overall mean uptake of Cr in shoot was almost four times and in root was about two times greater in rays, compared to fenugreek. These findings, according to the same authors, indicated that family *Cruciferae* (raya) was most tolerant to Cr toxicity, followed by *Chenopodiaceae* (spinach) and *Leguminosae* (fenugreek). Peng et al. (2012) [78] found that cultivar ST-9 of *Eucalyptus urophylla X E.grandis* was shown to accumulate more Pb than

**8. Different strategies adopted in order to enhance biomass production** 

Under elevated CO2 conditions the photosynthetic rate is enhanced, thus biomass production is positively influenced. According to Wang et al. (2012) [28], the increase in total Cd uptake by poplar (*Populus* sp.) and willow (*Salix* sp*.*) genotypes due to increased biomass production under elevated CO2 conditions suggests an alternative way of improving the

The use of fertilizers is another useful practice that should be adopted by the researchers in order to enhance biomass production under extreme heavy metal toxicity conditions. Some Brassica species, which are suitable to be used as phytoremediators, may suffer from Fe or Mn deficiency symptoms under Cu, or Zn toxicity conditions. In that case, leaf Fe and Mn fertilizations should be done in order to increase their biomass production [29], thus their

especially that of Al), could be used as a potential phytoremediator.

efficiency of phytoremediation in heavy metal contaminated soils.

others of the same species, like ST-2, or ST-29.

**under heavy metal toxicity conditions** 

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 441

Phytoremediation of heavy metal contaminated soils is defined as the use of living green plants to transport and concentrate metals from the soil into the aboveground shoots, which are harvested with conventional agricultural methods [70]. The technique is suitable for cultivated land with low to moderate metal contaminated level. According to Jadia and Fulekar (2009) [71], phytoremediation is an environmental friendly technology, which may be useful because it can be carried out *in situ* at relatively low cost, with no secondary pollution and with the topsoil remaining intact. Furthermore, it is a cost-effective method, with aesthetic advantages and long term applicability. It is also a safe alternate to conventional soil clean up [17]. However, a major drawback of phytoremediation is that a given species typically remediates a very limited number of pollutants [24]. For example, a soil may be contaminated with a number of potentially toxic elements, together with persistent organic pollutants [72]. There are two different strategies to phytoextract metals from soils. The first approach is the use of metal hyperaccumulator species, whose shoots or leaves may contain rather high levels of metals [25]. The important traits for valuable hyperaccumulators are the high bioconcentration factor (root-to-soil metal concentration) and the high translocation factor (shoot to root metal concentration) [73]. Another strategy is to use fast-growing, high biomass crops that accumulate moderate levels of metals in their shoots for metal phytoremediation [25]. Phytoextraction ability of some fast growing plant species leads to the idea of connecting biomass production with soil remediation of contaminated industrial zones and regions. This biomass will contain significant amount of heavy metals and its energetic utilization has to be considered carefully to minimize negative environmental impacts [74].

### **7. Plant species used for phytoremediation**

Many species have been used (either as hyperaccumulators, or as fast growing-high biomass crops) to accumulate metals, thus for their phytoremediation ability. Hyperaccumulators are these plant species, which are able to tolerate high metal concentrations in soils and to accumulate much more metal in their shoots than in their roots. By successive harvests of the aerial parts of the hyperaccumulator species, the heavy metals concentration in the soil can be reduced [23]. According to Chaney et al. (1997) [21], in order a plant species to serve the phytoextraction purpose, it should have strong capacities of uptake and accumulation of the heavy metals when it occurs in soil solution. For example, *Sedum plumbizincicola* is an hyperaccumulator that has been shown to have a remarkable capacity to extract Zn and Cd from contaminated soils [75]. In addition, a very good also hyperaccumulator for Zn and Cd phytoextraction is *Thlaspi caerulescens* [23]. *Iris pseudacorus* L. is an ornamental macrophyte of great potential for phytoremediation, to tolerate and accumulate Cr and Zn [19]. Furthermore, many species of *Brassica* are suitable for cultivation under Cu and Zn toxicity conditions and may be used for phytoremediation [29]. *Phragmites australis*, which is a species of *Poaceae* family, may tolerate extremely high concentrations of Zn, Cu, Pb and Cd, thus can be used as heavy metal phytoremediator [76].

all labor-intensive and costly [69].

negative environmental impacts [74].

**7. Plant species used for phytoremediation** 

thus can be used as heavy metal phytoremediator [76].

Classical remediation techniques such as soil washing, excavation, and chelate extraction are

Phytoremediation of heavy metal contaminated soils is defined as the use of living green plants to transport and concentrate metals from the soil into the aboveground shoots, which are harvested with conventional agricultural methods [70]. The technique is suitable for cultivated land with low to moderate metal contaminated level. According to Jadia and Fulekar (2009) [71], phytoremediation is an environmental friendly technology, which may be useful because it can be carried out *in situ* at relatively low cost, with no secondary pollution and with the topsoil remaining intact. Furthermore, it is a cost-effective method, with aesthetic advantages and long term applicability. It is also a safe alternate to conventional soil clean up [17]. However, a major drawback of phytoremediation is that a given species typically remediates a very limited number of pollutants [24]. For example, a soil may be contaminated with a number of potentially toxic elements, together with persistent organic pollutants [72]. There are two different strategies to phytoextract metals from soils. The first approach is the use of metal hyperaccumulator species, whose shoots or leaves may contain rather high levels of metals [25]. The important traits for valuable hyperaccumulators are the high bioconcentration factor (root-to-soil metal concentration) and the high translocation factor (shoot to root metal concentration) [73]. Another strategy is to use fast-growing, high biomass crops that accumulate moderate levels of metals in their shoots for metal phytoremediation [25]. Phytoextraction ability of some fast growing plant species leads to the idea of connecting biomass production with soil remediation of contaminated industrial zones and regions. This biomass will contain significant amount of heavy metals and its energetic utilization has to be considered carefully to minimize

Many species have been used (either as hyperaccumulators, or as fast growing-high biomass crops) to accumulate metals, thus for their phytoremediation ability. Hyperaccumulators are these plant species, which are able to tolerate high metal concentrations in soils and to accumulate much more metal in their shoots than in their roots. By successive harvests of the aerial parts of the hyperaccumulator species, the heavy metals concentration in the soil can be reduced [23]. According to Chaney et al. (1997) [21], in order a plant species to serve the phytoextraction purpose, it should have strong capacities of uptake and accumulation of the heavy metals when it occurs in soil solution. For example, *Sedum plumbizincicola* is an hyperaccumulator that has been shown to have a remarkable capacity to extract Zn and Cd from contaminated soils [75]. In addition, a very good also hyperaccumulator for Zn and Cd phytoextraction is *Thlaspi caerulescens* [23]. *Iris pseudacorus* L. is an ornamental macrophyte of great potential for phytoremediation, to tolerate and accumulate Cr and Zn [19]. Furthermore, many species of *Brassica* are suitable for cultivation under Cu and Zn toxicity conditions and may be used for phytoremediation [29]. *Phragmites australis*, which is a species of *Poaceae* family, may tolerate extremely high concentrations of Zn, Cu, Pb and Cd, Santana et al. (2012) [20] refer that *Genipa americana* L. is a tree species that tolerates high levels of Cr3+, therefore it can be used in recomposition of ciliary forests at Cr-polluted watersheds. According to the same authors, this woody species demonstrates a relevant capacity for phytoremediation of Cr. *Elsholtzia splendens* is regarded as a Cu tolerant and accumulating plant species [77]. Peng et al. (2012) [78] refer that *Eucalyptus urophylla X E.grandis* is a fast growing economic species that contributes to habitat restoration of degraded environments, such as the Pb contaminated ones. On the other hand, concerning Cd phytoextraction ability, only a few plant species have been accepted as Cd hyperaccumulators, including *Brassica juncea, Thlaspi caerulescens* and *Solanum nigrum*. Poplar (*Populus* L.), which is an easy to propagate and establish species and it has also the advantages of rapid growth, high biomass production, as well as the ability to accumulate high heavy metal concentrations, could be used as a Cd-hypaeraccumulator for phytoremediation [27-28,67]. According to Wang et al. (2012) [28], the increase in total Cd uptake by poplar genotypes in Cd contaminated soils is the result of enhanced biomass production under elevated CO2 conditions. Furthermore, *Amaranthus hypochondriacus* is a high biomass, fast growing and easily cultivated potential Cd hyperaccumulator [25]. Another species was found to be a good phytoremediator concerning its phytoaccumulation and tolerance to Ni stress is *Riccinus communis* L. [18]. Finally, *Justicia gendarussa*, which was proved to be able to tolerate and accumulate high concentration of heavy metals (and especially that of Al), could be used as a potential phytoremediator.

Differences between species, or genotypes of the same species, concerning heavy metal accumulation have been found by many researchers. According to Dheri et al. (2007) [17], the overall mean uptake of Cr in shoot was almost four times and in root was about two times greater in rays, compared to fenugreek. These findings, according to the same authors, indicated that family *Cruciferae* (raya) was most tolerant to Cr toxicity, followed by *Chenopodiaceae* (spinach) and *Leguminosae* (fenugreek). Peng et al. (2012) [78] found that cultivar ST-9 of *Eucalyptus urophylla X E.grandis* was shown to accumulate more Pb than others of the same species, like ST-2, or ST-29.

### **8. Different strategies adopted in order to enhance biomass production under heavy metal toxicity conditions**

Under elevated CO2 conditions the photosynthetic rate is enhanced, thus biomass production is positively influenced. According to Wang et al. (2012) [28], the increase in total Cd uptake by poplar (*Populus* sp.) and willow (*Salix* sp*.*) genotypes due to increased biomass production under elevated CO2 conditions suggests an alternative way of improving the efficiency of phytoremediation in heavy metal contaminated soils.

The use of fertilizers is another useful practice that should be adopted by the researchers in order to enhance biomass production under extreme heavy metal toxicity conditions. Some Brassica species, which are suitable to be used as phytoremediators, may suffer from Fe or Mn deficiency symptoms under Cu, or Zn toxicity conditions. In that case, leaf Fe and Mn fertilizations should be done in order to increase their biomass production [29], thus their ability to absorb and accumulate great amounts of heavy metals in contaminated soils, i.e. the efficiency of phytoremediation. According to Li et al. (2012) [25], in order to achieve large biomass crops, heavy fertilization has been practiced by farmers. Application of fertilizers not only provides plant nutrients, but may also change the speciation and mobility of heavy metals, thus enhances their uptake. According to Li et al. (2012) [25], NPK fertilization of *Amaranthus hypochondriacus*, a fast growing species grown under Cd toxicity conditions, greatly increased dry biomass by a factor of 2.7-3.8, resulting in a large increment of Cd accumulation. High biomass plants may be beneficed and overcome limitations concerning metal phytoextraction from the application of chemical amendments, including chelators, soil acidifiers, organic acids, ammonium e.t.c. [21]. Mihucz et al. (2012) [79] found that Poplar trees, grown hydroponically under Cd, Ni and Pb stress, increased their heavy metal accumulation by factor 1.6-3.3 when Fe (III) citrate was used.

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

On the other hand, in heavy metal contaminated soils, many plant species could be used (either as hyperaccumulators, or as fast growing-high biomass crops) in order to accumulate metals, thus to clean-up soils (phytoremediation). Particularly, the use of fast growing-high biomass species, such as Poplar, having also the ability to accumulate high amounts of heavy metals in their tissues, is highly recommended, as the efficiency of phytoremediation reaches its maximum. Particularly, since a given species typically remediates a very limited number of pollutants (i.e. in the cases when soil pollution caused by different heavy metals, or organic pollutants), it is absolutely necessary to investigate the choice of the best species for phytoremediation for each heavy metal. In addition to that, more research is needed in order to find out more strategies (apart from fertilization, the use of different *Bacillus* sp. strains, CO2 enrichment under controlled atmospheric conditions e.t.c.) to enhance biomass production under heavy metal toxicity conditions, thus to ameliorate the phytoremediation

[1] Karim MR, Zhang YQ, Tian D, Chen FJ, Zhang FS, Zou CQ (2012) Genotypic differences in zinc efficiency of Chinese maize evaluated in a pot experiment. J. Sci. Food Agric.

[2] Tzerakis C, Savvas D, Sigrimis N (2012) Responses of cucumber grown in recirculating nutrient solution to gradual Mn and Zn accumulation in the root zone owing to

[4] Quazi S, Datta R, Sarkar D (2011) Effects of soil types and forms of arsenical pesticide

[5] Sandana P, Pinochet D (2011) Ecophysiological determinants of biomass and grain yield

[6] Yaseen M, Malhi SS (2011) Exploitation of genetic variability among wheat genotypes

[7] Zribi OT, Abdelly C, Debez A (2011) Interactive effects of salinity and phosphorus availability on growth, water relations, nutritional status and photosynthetic activity of

barley (*Hordeum vulgare* L.). Plant Biol. DOI : 10.1111/j. 1438-8677.2011.00450.x

excessive supply via the irrigating water. J. Plant. Nutr. Soil Sci. 175: 125-134. [3] Bayuelo-Jimenez JS, Gallardo-Valdez M, Perez-Decelis VA, Magdaleno-Armas L, Ochoa I, Lynch JP (2011) Genotypic variation for root traits of maize (*Zea mays* L.) from the Purhepecha Plateau under contrasting phosphorus availability. Field Crops Research

on rice growth and development. Intern. J. Environ. Sci. Tech. 8: 445-460.

for tolerance to phosphorus deficiency stress. J. Plant Nuutr. 34: 665-699.

of wheat under P deficiency. Field Crops Res. 120: 311-319.

and Ioannis Therios

*Laboratory of Pomology, Aristotle University of Thessaloniki, Greece* 

efficiency.

**Author details** 

**10. References** 

121: 350-362.

Corresponding Author

 \*

Theocharis Chatzistathis\*

DOI 10.1002/jsfa.5672.

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 443

Mycorrhizal associations may be another factor increasing resistance to heavy metal toxicity, thus reducing the depression of biomass due to toxic conditions. Castillo et al. (2011) [80] found that when *Tagetes erecta* L. colonized by *Glomus intraradices* displayed a higher resistance to Cu toxicity. According to the same authors, *Glomus intraradices* possibly accumulated excess Cu in its vesicles, thereby enhanced Cu tolerance of *Tagetes erecta* L. [80].

Finally, other factors, such as the influence of *Bacillus* sp. on plant growth, in contaminated heavy metal soils, indicate that biomass may be stimulated under so adverse conditions. According to Brunetti et al. (2012) [81], the effect of the amendment with compost and *Bacillus licheniformis* on the growth of three species of *Brassicaceae* family was positive, since it significantly increased their dry matter. Furthermore, the strain of *Bacillus* SLS18 was found to increase the biomass of the species sweet sorghum (*Sorghum bicolor* L.), *Phytolacca acinosa Roxb*., and *Solanum nigrum* L. when grown under Mn and Cd toxicity conditions [82].

### **9. Conclusion and perspectives**

Biomass production is significantly influenced by many environmental, agronomic and other factors. The most important of them are air and soil temperature, soil humidity, photoperiod, light intensity, genotype, and soil nutrient availability. Soil fertility, i.e. the availability of nutrients in the optimum concentration range, greatly influences biomass production. If nutrient concentrations are out of the optimum limits, i.e. in the cases when nutrient deficiency or toxicity occurs, biomass production is depressed. Under nutrient deficient conditions, the farmers use chemical fertilizers in order to enhance yields and fruit production. However, since the prices of fertilizers have been significantly increased during the last two decades, a very good agronomic practice is the utilization of nutrient use efficient genotypes, i.e. the utilization of genotypes which are able to produce high yields under nutrient limited conditions. Although great scientific progress has been taken place during last years concerning nutrient use efficient genotypes, more research is still needed in order to clarify the physiological, genetic, and other mechanisms involved in each plant species.

On the other hand, in heavy metal contaminated soils, many plant species could be used (either as hyperaccumulators, or as fast growing-high biomass crops) in order to accumulate metals, thus to clean-up soils (phytoremediation). Particularly, the use of fast growing-high biomass species, such as Poplar, having also the ability to accumulate high amounts of heavy metals in their tissues, is highly recommended, as the efficiency of phytoremediation reaches its maximum. Particularly, since a given species typically remediates a very limited number of pollutants (i.e. in the cases when soil pollution caused by different heavy metals, or organic pollutants), it is absolutely necessary to investigate the choice of the best species for phytoremediation for each heavy metal. In addition to that, more research is needed in order to find out more strategies (apart from fertilization, the use of different *Bacillus* sp. strains, CO2 enrichment under controlled atmospheric conditions e.t.c.) to enhance biomass production under heavy metal toxicity conditions, thus to ameliorate the phytoremediation efficiency.

### **Author details**

442 Biomass Now – Cultivation and Utilization

Mn and Cd toxicity conditions [82].

species.

**9. Conclusion and perspectives** 

ability to absorb and accumulate great amounts of heavy metals in contaminated soils, i.e. the efficiency of phytoremediation. According to Li et al. (2012) [25], in order to achieve large biomass crops, heavy fertilization has been practiced by farmers. Application of fertilizers not only provides plant nutrients, but may also change the speciation and mobility of heavy metals, thus enhances their uptake. According to Li et al. (2012) [25], NPK fertilization of *Amaranthus hypochondriacus*, a fast growing species grown under Cd toxicity conditions, greatly increased dry biomass by a factor of 2.7-3.8, resulting in a large increment of Cd accumulation. High biomass plants may be beneficed and overcome limitations concerning metal phytoextraction from the application of chemical amendments, including chelators, soil acidifiers, organic acids, ammonium e.t.c. [21]. Mihucz et al. (2012) [79] found that Poplar trees, grown hydroponically under Cd, Ni and Pb stress, increased

Mycorrhizal associations may be another factor increasing resistance to heavy metal toxicity, thus reducing the depression of biomass due to toxic conditions. Castillo et al. (2011) [80] found that when *Tagetes erecta* L. colonized by *Glomus intraradices* displayed a higher resistance to Cu toxicity. According to the same authors, *Glomus intraradices* possibly accumulated excess Cu in its vesicles, thereby enhanced Cu tolerance of *Tagetes erecta* L. [80]. Finally, other factors, such as the influence of *Bacillus* sp. on plant growth, in contaminated heavy metal soils, indicate that biomass may be stimulated under so adverse conditions. According to Brunetti et al. (2012) [81], the effect of the amendment with compost and *Bacillus licheniformis* on the growth of three species of *Brassicaceae* family was positive, since it significantly increased their dry matter. Furthermore, the strain of *Bacillus* SLS18 was found to increase the biomass of the species sweet sorghum (*Sorghum bicolor* L.), *Phytolacca acinosa Roxb*., and *Solanum nigrum* L. when grown under

Biomass production is significantly influenced by many environmental, agronomic and other factors. The most important of them are air and soil temperature, soil humidity, photoperiod, light intensity, genotype, and soil nutrient availability. Soil fertility, i.e. the availability of nutrients in the optimum concentration range, greatly influences biomass production. If nutrient concentrations are out of the optimum limits, i.e. in the cases when nutrient deficiency or toxicity occurs, biomass production is depressed. Under nutrient deficient conditions, the farmers use chemical fertilizers in order to enhance yields and fruit production. However, since the prices of fertilizers have been significantly increased during the last two decades, a very good agronomic practice is the utilization of nutrient use efficient genotypes, i.e. the utilization of genotypes which are able to produce high yields under nutrient limited conditions. Although great scientific progress has been taken place during last years concerning nutrient use efficient genotypes, more research is still needed in order to clarify the physiological, genetic, and other mechanisms involved in each plant

their heavy metal accumulation by factor 1.6-3.3 when Fe (III) citrate was used.

Theocharis Chatzistathis\* and Ioannis Therios *Laboratory of Pomology, Aristotle University of Thessaloniki, Greece* 

### **10. References**


<sup>\*</sup> Corresponding Author

[8] oussadia O, Steppe K, Zgallai H, El Hadj SB, Braham M, Lemeur M, Van Labeke MC (2010) Effects of nitrogen efficiency on leaf photosynthesis, carbohydrate status and biomass production in two olive cultivars ('Meski' and 'Koroneiki'). Sci. Hortic. 123: 336-342.

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

[23] Frerot H, Petit C, Lefebvre C, Gruber W, Collin C, Escarre J (2003) Zinc and cadmium accumulation in controlled crosses between metallicolous and nonmetallicolous

[25] Li NY, Fu QL, Zhuang P, Guo B, Zou B, Li ZA (2012) Effect of fertilizers on Cd uptake of Amaranthus hypochondriacus, a High Biomass, Fast Growing and Easily Cultivated

[26] Liu Y, Wang K, Xu P, Wang Z (2012) Physiological responses and tolerance threshold to cadmium contamination in *Eremochloa ophiuroides*. Intern. J. Phytorem. 14: 467-480. [27] Jun R, Ling T (2012) Increase of Cd accumulation in five Poplar (*Populus* L.) with

[28] Wang R, Dai S, Tang S, Tian S, Song Z, Deng X, Ding Y, Zou X, Zhao Y, Smith DL (2012) Growth, gas exchange, root morphology and cadmium uptake responses of poplars and willows grown on cadmium-contaminated soil to elevated CO2. Environ Earth Sci DOI

[29] Ebbs SD, Kochian L (1997) Toxicity of zinc and copper to Brassica species: Implications

[30] Voyiatzis D, Petridou M (1997) Biology of horticultural plants. Publications of the

[31] Therios I (2009) Olives. Crop Production Science in Horticulture. CAB International.

[32] Pavel EW, Fereres E (1998) Low soil temperatures induce water deficits in olive (Olea

[33] Chaudhary MI, Adu-Gyamfi JJ, Saneoka H, Nguyen T, Suwa R, Kanai S, El-Shemy HA, Lightfoot DA, Fujita K (2008) The effect of phosphorus deficiency on nutrient uptake, nitrogen fixation and photosynthetic rate in mashbean, mungbean and soybean. Acta

[34] Pang J, Tibbett M, Denton MD, Lambers H, Siddique KHM, Bolland MDA, Revell CK, Ryan MH (2010) Variation in seedling growth of 11 perennial legumes in response to

[35] Hu J, Lin X, Wang J, Dai J, Cui X, Chen R, Zhang J (2009) Arbuscular mycorrhizal fungus enhances crop yield and P-uptake of maize (*Zea mays* L.): A field case study on a sandy loam soil as affected by long-term P-deficiency fertilization. Soil Biol Biochem 41:

[36] Blanke V, Wagner M, Renker C, Lippert H, Michulitz M, Kuhn AJ, Buscot F (2011) Arbuscular mycorrhizas in phosphate-polluted soil: interrelations between root

[37] Singh YP, Singh G, Sharma DK (2010) Biomass and bio-energy production of ten multipurpose tree species planted in sodic soils of indo-gangetic plains. J. Forestry Res.

Aristotle University of Thessaloniki, Thessaloniki, Greece. pp 51-65.

populations of *Thlaspi caerulescens* (Brassicaceae). New Phytologist 157: 643-648. [24] Wu L, Li Z, Han C, Liu L, Teng Y, Sun X, Pan C, Huang Y, Luo Y, Christie P (2012) Phytoremediation of soil contaminated with cadmium, copper and polychlorinated

biphenyls. Intern. J. Phytorem. 14: 570-584.

10.1007/s12665-011-1475-0.

Physiol. Plant. 30: 537-544.

UK. pp 179-209.

2460-2465.

21: 19-24.

Potential Cd Hyperaccumulator. Int. J. Phytorem. 14: 162-173.

different supply levels of Cd. Int. J. Phytorem. 14: 101-113.

for phytoremediation. J. Env. Qual. 26: 776-781.

europaea L.) trees. Physiol. Plant. 104: 525-534.

phosphorus supply. Plant Soil 328: 133-143.

colonization and nitrogen. Plant Soil 343: 379-392.

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 445


Plant. 31: 849-853.

York. pp 185-207.

804.

1672.

0956-4.

and 'Koroneiki'. HortSci. 44: 1994-1999.

Oct.-4 Nov. 2011, Crete, Greece. Article in Press.

*Melilotus officinalis*. Env. Exp. Bot. 78:18-24.

to nickel. Int. J. Phytorem. 14: 481-492.

heavy metals. J. Plant Nutr. 3: 643-654.

336-342.

[8] oussadia O, Steppe K, Zgallai H, El Hadj SB, Braham M, Lemeur M, Van Labeke MC (2010) Effects of nitrogen efficiency on leaf photosynthesis, carbohydrate status and biomass production in two olive cultivars ('Meski' and 'Koroneiki'). Sci. Hortic. 123:

[9] Reddy KR, Matcha SK (2010) Quantifying nitrogen effects on castor bean (*Ricinus communis* L.) development, growth and photosynthesis. Indus. Crops Prod. 31: 185-191. *[10]* Msilini N, Attia H, Bouraoui N, M'rah S, Ksouri R, Lachaal M, Ouerghi Z (2009) Responses of *Arabidopsis thaliana* to bicarbonate-induced iron deficiency. Acta Physiol.

[11] Chapin, FS, Van Cleve, K (1991) Approaches to studying nutrient uptake, use and loss in plants. In: Pearcy RW, Ehleringer JR, Mooney HA, Rundel PW, editors. Plant physiological ecology-Field methods and instrumentation. Chapman and Hall, New

[12] Chatzistathis T, Therios I, Alifragis D (2009) Differential uptake, distribution within tissues, and use efficiency of Manganese, Iron and Zinc by olive cultivars 'Kothreiki'

[13] Chatzistathis T, Papadakis I, Therios I, Patakas A, Giannakoula A, Menexes G (2012) Differential response of two olive cultivars to excess manganese. J. Plant Nutr. 35: 784-

[14] Chatzistathis T, Therios I (2011) Possible reasons for the differential micronutrient utilization efficiency of olive cultivars. Proc. 4th Intern. Olivebioteq seminar. Chania 31

[15] Fernandez R, Bertrand A, Garcia J, Tames RS, Gonzalez A (2012) Lead accumulation and synthesis of non- protein thiolic peptides in selected clones of *Melilotus alba* and

[16] Shen ZG, Li XD, Chen XM, Wang CC, Chua H (2002) Phytoextraction of Pb from a contaminated soil using high biomass species. J. Environ. Qual. 31: 1893-1900. [17] Dheri GS, Brar MS, Malhi SS (2007) Comparative phytoremediation of chromiumcontaminated soils by fenugreek, spinach and raya. Com. Soil Sci. Plant Anal. 38: 1655-

[18] Adhikari T, Kumar A (2012) Phytoaccumulation and tolerance of *Riccinus communis* L.

[19] Caldelas C, Araus JL, Febrero A, Bort J (2012) Accumulation and toxic effects of chromium and zinc in *Iris pseudacorus* L. Acta Physiol. Plant. DOI 10.1007/s11738-012-

[20] Santana KB, De Almeida AAF, Souza VL, Mangabeira PAO, Da Silva DC, Gomes FP, Dutruch L, Loguercio LL (2012) Physiological analyses of *Genipa americana* L. reveals a tree with ability as phytostabilizer and rhizofilterer of chromium ions for

[21] Chaney RL, Malik M, Li YM, Brown SL, Brewer EP, Angle JS, Baker AJM (1997)

[22] Baker AJM (1981) Accumulators and excluders-strategies in the response of plants to

phytoremediation of polluted watersheds. Env. Exp. Bot. 80: 35-42.

Phytoremediation of soil metals. Curr. Opin. Biotech. 8: 279-284.


[38] Therios I (1996) Mineral nutrition and fertilizers. Dedousis publications. Thessaloniki, Greece. pp 174-177. In Greek.

How Soil Nutrient Availability Influences Plant Biomass and How Biomass Stimulation Alleviates Heavy Metal

[56] Maruyama T, Higuchi Κ, Yoshida Μ, Tadano Τ (2005) Comparison of iron availability

[57] Citernesi AS, Vitagliano C, Giovannetti M (1998) Plant growth and root system morphology of *Olea europaea* L. rooted cuttings as influenced by arbuscular

[58] Rengel Z (2001) Genotypic differences in micronutrient use efficiency in crops. Com.

[59] Yang YY, Yung JY, Song WY, Suh HS, Lee Y (2000) Identification of rice varieties with high tolerance or sensitivity to lead and characterization of the mechanism of tolerance.

[60] Bah AM, Dai H, Zhao J, Sun H, Cao F, Zhang G, Wu F (2011) Effects of cadmium, chromium and lead on growth, metal uptake and antioxidative capacity in *Typha* 

[61] Ohki K (1985) Manganese deficiency and toxicity effects on photosynthesis, chlorophyll

[62] Foy CD, Farina MPW, Oakes AJ (1998) Iron-Manganese interactions among clones of

[63] Alam, S, Kamei S, Kawai S (2001) Amelioration of manganese toxicity in barley with

[64] Quartin VML, Antunes ML, Muralha MC, Sousa MM, Nunes MA (2001) Mineral

[65] Sarkar D, Pandey SK, Sud KC, Chanemougasoundharam A (2004) In vitro characterization of manganese toxicity in relation to phosphorus nutrition in potato

[66] Baccouch S, Chaoui A, Ferjani EEI (1998) Nickel toxicity: Effects on growth and

[67] Dai HP, Shan C, Wei Y, Liang JG, Yang T, Sa WQ, Wei AZ (2012) Subcellular localization of cadmium in hyperaccumulator *Populus* x *canescens*. Afr. J. Biotech. 11:

[68] Guney M, Zagury GJ, Dogan N, Onay TT (2010) Exposure assessment and risk characterization from trace elements following soil ingestion by children exposed to

[69] Wu G, Kang HB, Zhang XY, Shao HB, Chu LY, Ruan CJ (2010) A critical review on the bio-removal of hazardous heavy metals from contaminated soils: issues, progress, eco-

[70] Kumar NPBA, Dushenkov V, Motto H, Raskin I (1995) Phytoextraction: the use of plants to remove heavy metals from soils. Environ Sci Technol 29: 1232-1238. [71] Jadia CD, Fulekar MH (2009) Phytoremediation of hevy metals: recent techniques. Afr.

[72] Xu L, Luo YM, Teng Y, Zhang XL, Wang JJ, Zhang HB, Li ZG, Liu WX (2009) Soil environment quality and remediation in Yangtze River Delta IV. Soil acidification and heavy metal pollution in farmland soils around used electronic device disassembling

playgrounds, parks and picnic areas. J. Hazard. Mater. 182: 656-664.

environmental concerns and opportunities. J. Hazard. Mater. 174: 1-8.

imbalance due to manganese excess in triticales. J. Plant Nutr. 24: 175-189.

in leaves of barley and rice. Soil Sci. Plant Nutr. 51: 1037-1042.

mycorrhizas. J. Hortic. Sci. Biotech. 73: 647-654.

*angustifolia*. Biol. Trace Elem. Res. 142: 77-92.

nilegrass. J. Plant Nutr. 21: 978-1009.

iron. J. Plant Nutr. 24: 1421-1433.

3779-3787.

J. Biotechnol. 8: 921-928.

and transpiration in wheat. Crop Sci. 25: 187-191.

(*Solanum tuberosum* L.). Plant Sci. 167: 977-986.

metabolism of maize. J. Plant Nutr. 21: 577-588.

sites. Acta Pedol. Sin. 46: 833-839. In Chinese.

Soil Sci. Plant Anal. 32: 1163-1186.

Plant Physiol. 124: 1019-1026.

Toxicity in Soils: The Cases of Nutrient Use Efficient Genotypes and Phytoremediators, Respectively 447


[56] Maruyama T, Higuchi Κ, Yoshida Μ, Tadano Τ (2005) Comparison of iron availability in leaves of barley and rice. Soil Sci. Plant Nutr. 51: 1037-1042.

446 Biomass Now – Cultivation and Utilization

Greece. pp 174-177. In Greek.

Plant Anal. 32: 921-950.

var. Pekinensis) biomass. J. Plant Nutr. 35: 471-479.

agriculture. A review. Agron. Sustain. Dev. 30: 83-107.

nutrient use efficiency. Aus. J. Agric. Res. 54: 571-580.

remobilization. J. Plant Nutr. 34: 2006-2017.

and ionic parameters. J. Plant Nutr. 34: 522-533.

aminotransferase genes. Nat. Biotecnol. 19: 466-469.

Annu. Rev. Plant Physiol. Plant Mol. Biol. 46: 237–260.

sustainable ways. Plant Soil 247: 3-24.

J. 25: 641-649.

Biotechnol. 18: 450-453.

definitions and suggested improvement. Plant Soil 158: 29-37.

New Series, volume 15B. Springer Verlag, New York., pp 453-480.

[38] Therios I (1996) Mineral nutrition and fertilizers. Dedousis publications. Thessaloniki,

[39] Konkol NR, McNamara CJ, Bearce-Lee K, Kunoh H, Mitchell R (2012) Novel method of micronutrient application increases radish (*Raphanus sativus*) and shirona (*Brassica Rapa*

[40] Baligar VC, Fageria NK, He ZL (2001) Nutrient use efficiency in plants. Com. Soil Sci.

[41] Khoshgoftarmanesh AH, Schulin R, Chaney RL, Daneshbakhsh B, Afyuni M (2009) Micronutrient-efficient genotypes fro crop yield and nutritional quality in sustainable

[42] Gourley CJP, Allan DL, Rousselle MP (1994) Plant nutrient efficiency: A comparison of

[43] Gerloff GC, Gabelman WH (1983) Genetic basis of inorganic plant nutrition. In Lauchli A, Bieleski RL, editors. Inorganic Plant Nutrition. Encyclopedia and Plant Physiology

[44] Papadakis IE, Dimassi KN, Therios IN (2003) Response of 2 citrus genotypes to six boron concentrations: concentration and distribution of nutrients, total absorption and

[45] Jiang WZ, Ireland CR (2005) Characterization of manganese use efficiency in UK wheat cultivars grown in a solution culture system and in the field. J. Agric. Sci. 143: 151-160. [46] Jiang WZ (2006) Mn use efficiency in different wheat cultivars. Env. Exp. Bot. 57: 41-50. [47] Aziz T, Ahmed I, Farooq M, Maqsood MA, Sabir M (2011a) Variation in phosphorus efficiency among *Brassica* cultivars I: Internal utilization and phosphorus

[48] Yang F, Wang G, Zhang Z, Eneji AE, Duan L, Li Z, Tian X (2011) Genotypic variations

[49] Cakmak I (2002) Plant nutrition research: Priorities to meet human needs for food in

[50] Aziz T, Rahmatullah M, Maqsood MA, Sabir M, Kanwal S (2011b) Categorization of *Brassica* cultivars for phosphorus acquisition from phosphate rock on basis of growth

[51] Richardson AE, Hadobas PA, Hayes JE (2001) Extracellular secretion of Aspergillus phytase from *Arabidopsis* roots enables plants to obtain phosphorus from phytate. Plant

[52] Takahashi M, Nakanishi H, Kawasaki S, Nishizawa NK, Mori S (2001) Enhanced tolerance of rice to low iron availability in alkaline soils using barley nicotianamine

[53] Kochian LV (1995) Cellular mechanisms of aluminium toxicity and resistance in plants.

[55] Lopez-Bucio J, Martinez de la Vega O, Guevara-Garcia A, Errera-Estrella L (2000) Enhanced phosphorus uptake in transgenic tobacco that overproduce citrate. Nat.

[54] Marschner H (1995) Mineral nutrition of higher plants. Academic Press, London.

in potassium uptake and utilization in cotton. J. Plant Nutr. 34: 83-97.


106-113.

Conf. pp. 49-50.

Physiol. 169: 561-566.

Geoderma 170: 322-330.

Appl. Microbiol. Biotechnol. 93:1745–1753.

[73] Teofilo V, Marianna B, Giuliano M (2010) Field crops for phytoremediation of metal-

[74] Syc M, Pohorely M, Kamenikova P, Habart J, Svoboda K, Puncochar M (2012) Willow trees from heavy metals phytoextraction as energy crops. Biomass and Bioenergy 37:

[75] Wu LH, Li N, Luo YM (2008) Phytoextraction of heavy metal contaminated soil by *Sedum plumbizincicola* under different agronomic strategies. Proc. 5th Int. Phytotechnol.

[76] Ye ZH, Baker AJM, Wong MH, Willis AJ (2003) Copper tolerance, uptake, and

[77] Tang SR, Wilke BM, Brooks RR (2001) Heavy-metal uptake by metal-tolerant *Elsholtzia haichowensis* and *Commelina communis* from China. Com. Soil Sci. Plant. 32: 895-905. [78] Peng X, Yang B, Deng D, Dong J, Chen Z (2012) Lead tolerance and accumulation in three cultivars of *Eucalyptus urophylla X E. grandis*: implication for phytoremediation.

[79] Mihucz VG, Csog A, Fodor F, Tatar E, Szoboszlai N, Silaghi-Dumitrescu L, Zaray G (2012) Impact of two iron(III) chelators on the iron, cadmium, lead and nickel accumulation in poplar grown under heavy metal stress in hydroponics. J. Plant

[80] Castillo OS, Dasgupta-Schubert N, Alvarado CJ, Zaragoza EM, Villegas HJ (2011) The effect of the symbiosis between *Tagetes erecta* L. (marigold) and *Glomus intraradices* in the uptake of copper (II) and its implications for phytoremediation. New Biotech 29. [81] Brunetti G, Farrag K, Soler-Rovina P, Ferrara M, Nigro F, Senesi N (2012) The effect of compost and *Bacillus licheniformis* on the phytoextraction of Cr, Cu, Pb and Zn by three *Brassicaceae* species from contaminated soils in the Apulia region, Southern Italy.

[82] Luo S, Xu T, Chen L, Chen J, Rao C, Xiao X, Wang Y, Zeng G, Long F, Liu C, Liu Y (2012) Endophyte-assisted promotion of biomass production and metal-uptake of energy crop sweet sorghum by plant-growth-promoting endophyte Bacillus sp. SLS18.

contaminated land. A review. Environ Chem Lett. 8: 1-17.

accumulation by *Phragmites australis*. Chemosphere 50: 795-800.

Environ Earth Sci DOI 10.1007/s12665-012-1595-1.

*Edited by Miodrag Darko Matovic*

This two-volume book on biomass is a reflection of the increase in biomass related research and applications, driven by overall higher interest in sustainable energy and food sources, by increased awareness of potentials and pitfalls of using biomass for energy, by the concerns for food supply and by multitude of potential biomass uses as a source material in organic chemistry, bringing in the concept of bio-refinery. It reflects the trend in broadening of biomass related research and an increased focus on secondgeneration bio-fuels. Its total of 40 chapters spans over diverse areas of biomass research, grouped into 9 themes.

Photo by Patrick\_Lauzon / iStock

Biomass Now - Cultivation and Utilization

Biomass Now

Cultivation and Utilization