**3. Results**

## **3.1 Selected soil and plant properties**

Soil samples under all tree species were predominantly organic (65–85%). The mineral part of the soil is silty loam to silty clay loam (20–36% clay, 0.4–0.7% sand,

**103**

**Table 1.**

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme…*

63–80% silt) depending on the tree location in the study area (data not shown). Soil pH was neutral to alkaline for all tree species (**Table 1**). Two-way ANOVA did not demonstrate an interaction between tree species and sampling months (**Table 2**). Available N varied significantly with regard to monthly sampling and not to tree species (**Table 2**). Available Ca and P varied significantly with regard to tree species (**Table 2**). Total N and C:N ratio varied significantly with regard to tree species.

Six enzyme activities (EAs) were assessed representing C (β-glucosidase), N (β-glucosaminidase), P (phosphodiesterase, alkaline and acid phosphatase) and S (arylsulphatase) cycling. There was a clear and significant separation in the soil enzyme activity between the tree species in this forest, while there was no distinct trend due to the sampling time (**Figure 2**). Most of the separation on the metabolic capacity of the soil according to these six EAs was observed along axis 1, which explained 87.22% of the variability. *Tabebuia heterophylla* aligned along CA 1 separating this species from *Pisonia albida* and *Ficus citrifolia* which aligned along CA 2 (eigenvalue 12.3). The activities of β-glucosidase, alkaline and acid phosphatase were more closely associated with *Tabebuia heterophylla* than with the other tree species while β-glucosaminidase activity was more closely associated with *Ficus* 

Two-way analysis of variance was used to establish the effect of plant species and sampling period on soil enzyme activities. The only enzymes that were significantly affected by plant species were alkaline phosphatase (F = 8.18; P < 0.001), β-glucosidase (F = 8.86; P < 0.001) and β-glucosaminidase (F = 5.97; P = 0.007) (**Table 2**). Alkaline phosphatase was the only enzyme affected by monthly variations (F = 4.375; P = 0.007). Two-way ANOVA does not demonstrate significant

A total of 56 FAMEs were identified in the samples analysed and 13 FAMEs were used as indicators of different microbial groups as affected by tree species traits and sampling time (**Table 2**). According to CDA, significant differences (P < 0.0001) were detected in the FAME profiles of the soil microbial community structure as affected by tree species traits (**Figure 3A**). Canonical axis 1 explained 83% of the variation. The soil microbial communities under *Tabebuia heterophylla* were distinct from the microbial communities under *Ficus citrifolia* and *Pisonia albida*. FAMEs that contributed to the separation observed under *Tabebuia heterophylla* are the

interactions between tree species and monthly variations.

**3.3 Microbial community structure (EL-FAME)**

*Mean values of selected soil properties under three different tree species.*

*DOI: http://dx.doi.org/10.5772/intechopen.82579*

**3.2 Soil enzyme activities**

*citrifolia* trees.

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme… DOI: http://dx.doi.org/10.5772/intechopen.82579*

63–80% silt) depending on the tree location in the study area (data not shown). Soil pH was neutral to alkaline for all tree species (**Table 1**). Two-way ANOVA did not demonstrate an interaction between tree species and sampling months (**Table 2**). Available N varied significantly with regard to monthly sampling and not to tree species (**Table 2**). Available Ca and P varied significantly with regard to tree species (**Table 2**). Total N and C:N ratio varied significantly with regard to tree species.

### **3.2 Soil enzyme activities**

*Extremophilic Microbes and Metabolites - Diversity, Bioprospecting and Biotechnological...*

and evaporated under a N2

**2.4 Soil enzyme activity**

(PN) released in kg<sup>−</sup><sup>1</sup>

**2.5 Statistical analysis**

 soil h<sup>−</sup><sup>1</sup> .

(25 m × 0.2 mm) with ultra-high purity H2

incubation period and, after the incubation period was completed, 3 ml of 1.0 acetic acid was added to neutralize the mixture's pH; (b) partitioning of the FAMEs into an organic phase was achieved by adding 10 ml of hexane and centrifuging the preparations at 480×*g* for 10 min; (c) the hexane layer was transferred to a clean glass tube

hexane:methyl-tert butyl ether and transferred to GC vials for analysis. Extractions were analysed as described by [4]. A 6890 GC Series II (Hewlett Packard, Wilmington, DE, USA) equipped with a flame ionization detector and fused silica capillary column

extractions. The temperature program was ramped from 170 to 250°C at 5°C min<sup>−</sup><sup>1</sup>

method of the MIDI system). FAME concentrations (nmol g<sup>−</sup><sup>1</sup>

described. The fatty acids were identified and quantified by comparing the retention times and peak areas to MIDI standards. The MIDI software provides FAME relative peak areas (percentage) based on the total FAMEs in a sample (based on the Aerobe

comparing peak areas to an analytical standard (19:0, Sigma Chemical Co., St. Louis, MO) calibration curve. The FAMEs are described by the number of C atoms, followed by a colon, the number of double bonds and then by the position of the first double bond from the methyl (ω) end of the molecule. Cis isomers are indicated by c and branched fatty acids are indicated by the prefixes *i* and *a* for iso and anteiso, respectively. Other notations are Me for methyl, OH for hydroxy and cy for cyclopropane.

Study of enzyme activities was performed as described in [4, 6]. The activities of enzymes relevant in C cycling (β-glucosidase), C and N cycling (β-glucosaminidase), P cycling (alkaline phosphatase, acid phosphatase, phosphodiesterase) and in the S cycle (arylsulphatase) were assayed using 0.5 g of air-dried soil (<2 mm). Duplicate replicates and one control were used for all the soils that were tested; furthermore, the appropriate substrate was used for each assay and reactions were incubated for 1 h at 37°C at their optimal pH as described in [4]. For the controls, the substrate was added after the 1-h incubation period and subtracted from a sample control value. Enzyme activity is expressed in mg p-nitrophenol

All data analysed with JMP software were checked for normality and transformed to log10. To determine differences in soil FAMEs, enzyme activity and selected soil properties due to tree species and sampling periods, canonical discriminate analysis (CDA) was conducted with the JMP program. The first and second canonical discriminate functions were utilized to determine the distribution of enzyme activity and FAMEs as influenced by each tree species and by sampling period. SigmaPlot 10 software was used to conduct two-way analysis of variance (2-way ANOVA). Twoway analysis of variance was used to establish the effect of plant species and sampling period on soil community structure, enzyme activities and selected soil properties.

Soil samples under all tree species were predominantly organic (65–85%). The mineral part of the soil is silty loam to silty clay loam (20–36% clay, 0.4–0.7% sand,

stream and (d) FAMEs were suspended in 0.5 ml of 1:1

as the gas carrier was used to analyse the

as

soil) were calculated by

**102**

**3. Results**

**3.1 Selected soil and plant properties**

Six enzyme activities (EAs) were assessed representing C (β-glucosidase), N (β-glucosaminidase), P (phosphodiesterase, alkaline and acid phosphatase) and S (arylsulphatase) cycling. There was a clear and significant separation in the soil enzyme activity between the tree species in this forest, while there was no distinct trend due to the sampling time (**Figure 2**). Most of the separation on the metabolic capacity of the soil according to these six EAs was observed along axis 1, which explained 87.22% of the variability. *Tabebuia heterophylla* aligned along CA 1 separating this species from *Pisonia albida* and *Ficus citrifolia* which aligned along CA 2 (eigenvalue 12.3). The activities of β-glucosidase, alkaline and acid phosphatase were more closely associated with *Tabebuia heterophylla* than with the other tree species while β-glucosaminidase activity was more closely associated with *Ficus citrifolia* trees.

Two-way analysis of variance was used to establish the effect of plant species and sampling period on soil enzyme activities. The only enzymes that were significantly affected by plant species were alkaline phosphatase (F = 8.18; P < 0.001), β-glucosidase (F = 8.86; P < 0.001) and β-glucosaminidase (F = 5.97; P = 0.007) (**Table 2**). Alkaline phosphatase was the only enzyme affected by monthly variations (F = 4.375; P = 0.007). Two-way ANOVA does not demonstrate significant interactions between tree species and monthly variations.

### **3.3 Microbial community structure (EL-FAME)**

A total of 56 FAMEs were identified in the samples analysed and 13 FAMEs were used as indicators of different microbial groups as affected by tree species traits and sampling time (**Table 2**). According to CDA, significant differences (P < 0.0001) were detected in the FAME profiles of the soil microbial community structure as affected by tree species traits (**Figure 3A**). Canonical axis 1 explained 83% of the variation. The soil microbial communities under *Tabebuia heterophylla* were distinct from the microbial communities under *Ficus citrifolia* and *Pisonia albida*. FAMEs that contributed to the separation observed under *Tabebuia heterophylla* are the


#### **Table 1.**

*Mean values of selected soil properties under three different tree species.*


**Table 2.**

*Two-way analysis of variance for the effects of species (*Ficus citrifolia*,* Pisonia albida *and* Tabebuia heterophylla*) and sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011) on the soil properties, soil enzymes and EL-FAMEs in the humus soil substrate layer at the Guánica Dry Forest.*

Gram-positive (G+) markers *a*15:0 and *a*17:0, actinomycete 10Me18:0, protozoan marker 20:4ω6c and the fungal marker 18:3ω6c.

*Ficus citrifolia* and *Pisonia albida* grouped closer together but were still significantly different from each other as species, 95% confidence ellipse are clearly separated. FAMEs that contribute to differences in *Ficus citrifolia* vs. *Pisonia albida* are cy19:0 (Gram-negative, G−), 10Me16:0 or 10Me17:0 (actinomycete) and 18:2 ω6c (fungal marker). Differences in *Pisonia albida we*re due to *i*17:0 (G+), cy17:0 (G−), and 16:1 ω5c or 18:1ω9c (fungal markers).

**105**

**Figure 2.**

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme…*

*Canonical discriminate analysis of the soil enzyme activity at the Guánica Dry Forest. (A) canonical discriminate analysis of the soil enzyme activity as affected by tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla). Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (*Ficus citrifolia, Pisonia albida *and* Tabebuia heterophylla*), respectively. The multivariate mean for each tree species is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean. (B) canonical discriminate analysis of the soil enzyme activity as affected by sampling period during July to November 2011 at the Guánica Dry Forest of Puerto Rico. Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla), respectively. The multivariate mean for each month sampled is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean.*

*DOI: http://dx.doi.org/10.5772/intechopen.82579*

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme… DOI: http://dx.doi.org/10.5772/intechopen.82579*

#### **Figure 2.**

*Extremophilic Microbes and Metabolites - Diversity, Bioprospecting and Biotechnological...*

Gram-positive (G+) markers *a*15:0 and *a*17:0, actinomycete 10Me18:0, protozoan

heterophylla*) and sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011) on the soil properties, soil enzymes and EL-FAMEs in the humus soil substrate layer at the Guánica Dry Forest.*

*Two-way analysis of variance for the effects of species (*Ficus citrifolia*,* Pisonia albida *and* Tabebuia

*Ficus citrifolia* and *Pisonia albida* grouped closer together but were still significantly different from each other as species, 95% confidence ellipse are clearly separated. FAMEs that contribute to differences in *Ficus citrifolia* vs. *Pisonia albida* are cy19:0 (Gram-negative, G−), 10Me16:0 or 10Me17:0 (actinomycete) and 18:2 ω6c (fungal marker). Differences in *Pisonia albida we*re due to *i*17:0 (G+), cy17:0

marker 20:4ω6c and the fungal marker 18:3ω6c.

(G−), and 16:1 ω5c or 18:1ω9c (fungal markers).

**104**

**Table 2.**

*Canonical discriminate analysis of the soil enzyme activity at the Guánica Dry Forest. (A) canonical discriminate analysis of the soil enzyme activity as affected by tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla). Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (*Ficus citrifolia, Pisonia albida *and* Tabebuia heterophylla*), respectively. The multivariate mean for each tree species is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean. (B) canonical discriminate analysis of the soil enzyme activity as affected by sampling period during July to November 2011 at the Guánica Dry Forest of Puerto Rico. Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla), respectively. The multivariate mean for each month sampled is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean.*

#### **Figure 3.**

*Canonical discriminate analysis of the soil FAMEs at the Guánica Dry Forest. (A) canonical discriminate analysis of the soil FAME profiles as affected by tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla). Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (*Ficus citrifolia, Pisonia albida *and* Tabebuia heterophylla*), respectively. The multivariate mean for each tree species is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean. (3B) canonical discriminate analysis of the soil FAME profiles as affected by sampling period during July to November 2011 at the Guánica Dry Forest of Puerto Rico. Letters (A, B, C, D and E) represent sampling periods (July 2011, August 2011, September 2011, October 2011 and November 2011), respectively. The colours of the letters (pink, green and blue) represent tree species (Ficus citrifolia, Pisonia albida and Tabebuia heterophylla), respectively. The multivariate mean for each month sampled is a coloured and labelled circle. The size of the circles corresponds to a 95% confidence limit for the mean.*

In addition to the effects of tree species on FAME profiles of the microbial community structure, there were also significant monthly variations (p = 0.001) according to CDA along CA1, which explained 68.81% of the variation (**Figure 3B**). The CDA revealed that July, September and August clustered together along axis 1. August was separated from other months due to the predominance of *a*17:0 (G+) and cy17:0 (G− marker). This cluster of these 3 months was characterized by the

**107**

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme…*

presence of the fungal marker 18:1ω9c, the G− marker cy19:0 and the actinomycete indicator 10Me16:0. The second cluster formed was composed of October and November due to the predominance of two G+ markers (*a*15:0 and *i*17:0), two actinomycete markers (10Me17:0 and 10Me18:0), a protozoan marker (20:4ω6c)

Two-way analysis of variance was used to establish the effect of plant species and sampling period on soil community structure. The Gram-positive EL-FAMEs that were significantly affected due to tree species were iso15:0, iso17:0, antesio17:0, 10Me16:0 and 10Me17:0. Gram-negative markers affected by tree species were 17:0 cy and 19:0 cy and fungal markers were 18:1ω9c, 16:1ω5c, 18:2 ω6c and 18:3 ω6c. The Gram-positive EL-FAMEs that were significantly affected due to monthly variations were iso15:0, antesio15:0 and 10Me16:0. None of the Gram-negative markers were affected by monthly variations. Fungal markers 16:1ω5c, 18:2 ω6c and 18:3 ω6c were significantly affected by monthly variations. Two-way ANOVA did not demonstrate significant interactions between tree species and monthly variations (**Table 2**).

**4.1 Soil microbial communities in a dry forest as affected by tree species**

are known to contribute to the rate of decomposition of organic matter, which may drive the microbial community under each tree species [22–24]. Previous studies have determined that *Tabebuia heterophylla* leaves are tougher (350 N) and

to undergo decomposition. Soils under *Tabebuia heterophylla* presented a higher relative abundance of saprophytic fungal marker 18:3ω6c which may be indicative

Differences in the relative abundance of fungal markers at the Guánica Dry Forest could also be due to the microclimate of each tree species. The amplitude of temperature fluctuations encountered in this forest varies among plant species, a factor that has been documented to affect arthropods in this system [16]. The canopies of both *Pisonia albida* and *Ficus citrifolia* generally are taller than the tree canopy of *Tabebuia heterophylla*. This difference in height contributes to differences in temperature, water throughfall and soil moisture. Idiosyncratic tree species characteristics have the ability of modifying the amplitude of daily temperature at the Guánica Dry Forest [16] and these modifications will also affect the fungal

**4.2 Soil microbial communities in a dry forest as affected by monthly variations**

As found in our study, microbial community structure was influenced by monthly wet/dry events (**Figure 3B**). Previous studies have described significant responses of the soil microbial communities to wet/dry events [25]. Our results point towards differential responses between sporadic and continuous rainfall

have lower specific leaf area (SLA) (90 cm<sup>2</sup>

/g) and *Pisonia albida* (80 N; 100 cm2

of lower rates of litter fragmentation and decomposition.

community structure that is present under each tree species.

The evaluated tree species had the ability of modifying soil microbial communities. We found that the relative abundance of 8 out of 13 markers was higher under *Ficus citrifolia* than the rest of the tree species. *Ficus citrifolia* is a facultative deciduous tree, which means that it mostly exchanges leaves and has massive litter fall during very dry periods. Higher numbers of FAME markers under this species may be indicative of idiosyncratic effects that may be stabilizing the microbial community. Species traits such as leaf nutrients, leaf toughness and specific leaf area (SLA, cm2

/g) than those of *F. citrifolia* (175 N;

/g) [16] making them more difficult

/g)

*DOI: http://dx.doi.org/10.5772/intechopen.82579*

**4. Discussion**

110 cm<sup>2</sup>

and all the fungal markers (16:1ω5c, 18:3 ω6c and18:2ω6c).

*Tree Species and Precipitation Effect on the Soil Microbial Community Structure and Enzyme… DOI: http://dx.doi.org/10.5772/intechopen.82579*

presence of the fungal marker 18:1ω9c, the G− marker cy19:0 and the actinomycete indicator 10Me16:0. The second cluster formed was composed of October and November due to the predominance of two G+ markers (*a*15:0 and *i*17:0), two actinomycete markers (10Me17:0 and 10Me18:0), a protozoan marker (20:4ω6c) and all the fungal markers (16:1ω5c, 18:3 ω6c and18:2ω6c).

Two-way analysis of variance was used to establish the effect of plant species and sampling period on soil community structure. The Gram-positive EL-FAMEs that were significantly affected due to tree species were iso15:0, iso17:0, antesio17:0, 10Me16:0 and 10Me17:0. Gram-negative markers affected by tree species were 17:0 cy and 19:0 cy and fungal markers were 18:1ω9c, 16:1ω5c, 18:2 ω6c and 18:3 ω6c. The Gram-positive EL-FAMEs that were significantly affected due to monthly variations were iso15:0, antesio15:0 and 10Me16:0. None of the Gram-negative markers were affected by monthly variations. Fungal markers 16:1ω5c, 18:2 ω6c and 18:3 ω6c were significantly affected by monthly variations. Two-way ANOVA did not demonstrate significant interactions between tree species and monthly variations (**Table 2**).
