**4. Molecular markers for drought**

Accurate phenotypic markers must ultimately be associated with molecular markers to aid and accelerate current plant breeding efforts to select improved soybean varieties with better stress tolerance. The effectiveness of morphological and physiological markers can vary greatly according to the growth stage of the plant and the many variables in the environment, particularly under field conditions. The selection of molecular markers is based either on variations in genomic DNA or on variations in gene expression (transcriptome) patterns. Comparisons of the transcriptome and also proteome signatures of organs or tissues under stress conditions can provide a direct assessment of processes and/or components than can be developed into a useful marker for stress tolerance in breeding programs. In general, a molecular marker might identify variations in plant responses to stress at the gene level, or in certain regions of DNA, the composition of DNA or in the degree of DNA methylation. While the DNA content of a cell is independent of environmental conditions, organ specificity or growth stage, DNA methylation can be regulated by these endogenous and environmental conditions and so alter also the patterns of gene expression. Variable regions of the genome (transposons), which can include single copy and repetitive genome regions, can also have considerable impact on plant stress tolerance. Such genomic regions can be functional or 'silent' without any obvious function. More recently, next generation sequencing with the identification of single nucleotide polymorphisms, is increasingly being considered as a tool to identify useful molecular markers for stress tolerance.

Further, Quantitative trait loci (QTLs) are chromosomal regions (genomic loci) that might regulate the expression levels of proteins. QTLs can be associated with DNA segments that are considered to make a significant contribution to the expression of complex phenotypic traits

Identification and Application of Phenotypic

and Molecular Markers for Abiotic Stress Tolerance in Soybean 193

We have analysed the crown nodule transcriptome at two stages of development (i.e. on 7 and 11 week-old plants) in order to identify possible markers for nodule development/senescence that might underpin plant performance. A comparison of the crown nodule transcriptomes was performed on two cultivars (PAN809 and Highveld Top) after 7 and 11 weeks of growth. PAN 809 is generally recommended for use under long growing seasons and this variety is often used in moderate and hot regions of Southern Africa because it is well-suited to irrigated- and rain-fed cultivation conditions. Highveld Top was developed specifically to withstand the low night temperatures that are often experienced during the growing season at high altitude in Southern Africa. The first nodule harvest point (7 weeks) was chosen because the measured capacity of symbiotic nitrogen fixation was highest at this point of nodule development. The second harvest point (11 weeks) was chosen because symbiotic nitrogen fixation had decreased by 50% at this time point relative to week 7. Figures 8 and 9 illustrate the degree of genotypic variation in the nodule transcriptome signatures observed in the nodules of 7 and 11 week-nodules under optimal growth conditions. While 702 transcripts were changed in abundance in Highveld Top nodules harvested at week 11 compared to week 7, 1737 transcripts were differentially expressed in PAN809 nodules under the same conditions (Figure 8). Of these, 226 transcripts were identical and showed similar patterns of increase or decrease at week 11 relative to week 7 in both cultivars (Figure 8). Of the transcripts that were differentally expressed in PAN809 and Highveld Top nodules at week 7, the PAN809 nodule transcriptome had a much higher number of transcripts encoding proteins involved in cell wall and stress metabolism than Highveld Top (Figure 9). For example, an extensin-like protein and proteins involved in disease resistance were much more abundant in PAN 809 nodules than Highveld Top nodules at week 7. Moreover, transcripts encoding the extensinlike protein were enhanced to a much greater extent in PAN 809 than Highveld Top. Transcripts encoding disease resistance proteins were also much higher in PAN 809 nodules than Highveld Top nodules at 11 weeks. The biological relevance of these findings is currently

under investigation in relation to the different performance of these two cultivars.

Fig. 8. A comparison of differential gene expression in Highveld Top (HT) and PAN 809

nodules harvested from 7- and 11-week old plants.

such as stress tolerance. Quantitative traits are usually affected by more than one gene and by the environment. The association of morphological or physiological phenotypic markers with molecular markers in QTL analysis is considered to ultimately be the ideal approach to accelerate plant breeding programs in marker-assisted selection (MAS). MAS is the process whereby the identified markers of all types (morphological, physiological, biochemical or molecular) are used for indirect selection of required traits. MAS approaches however are far from trivial and they are often not cost effective in current breeding programs. Drought tolerance is a mutagenic trait, often influenced by large numbers of minor QTLs, rather than one or more major QTLs. This means that the development and effective large-scale application of MAS is still technically challenging, costly and time-consuming.

The identified molecular DNA markers should ideally be applicable using relatively simple methods that are amenable to automation and high throughput. However, the isolation DNA from plants is not technically demanding and can be carried out in any laboratory that has basic DNA isolation and characterisation equipment. Potential markers can be first evaluated in small sample subsets in order to confirm that the desired loci are present. Such methodologies normally require large sample collections or population sizes because the chosen markers have be applied to large segregating populations in order to determine whether the markers are linked to the required traits (Ribaut et al., 2002).

Relatively few markers for drought tolerance have been identified in soybean. However, progress in this regard will be accelerated once large collections of potential DNA sequences have been established for marker development (Shinozaki, 2007). Accurate genetic and physical maps of the soybean genome are also essential for the development of useful molecular markers for drought tolerance in soybean. Towards this goal, a total of 318 AFLP, 121 SSR, 108 RFLP, and 126 STS markers have been integrated into a linkage map composed of 509 RFLP, 318 SSR, 318 AFLP, 97 AFLP-derived STS, 29 BAC-end or EST-derived STS, 1 RAPD, and five morphological markers (Hisano et al., 2007, Xia et al., 2007). A further very useful advance has come from the sequencing of the soybean (Glycine max (L.) Merr. 'Williams 82') genome. Some 66,153 protein-coding loci are now available at: http://www.phytozome.net/soybean. Moreover, 3290 microsatellites (SSRs) identified from BAC end sequences of clones (comprising the 'Williams 82' physical map) were screened and two hundred and sixty-five SSRs were genetically mapped in at least one mapping population (Shoemaker et al., 2008).

### **4.1 Identification of markers using microarrays**

Microarrays involve the immobilisation of single-stranded DNA on a solid support that is hybridised with a single-stranded DNA or RNA population (Rockett and Dix, 1999). In microarray analysis gene expression and regulation patterns can be monitored on a large scale (Quackenbush, 2001). The technology facilitates screening for differently expressed genes in different plant varieties. Differently expressed genes can potentially be useful in MAS providing that they encode proteins involved in the traits of interest.

A spotted soybean cDNA microarray containing 36,000 elements derived from EST libraries is available that covers a wide range of tissues and organs at different developmental stages under optimal and stress conditions (Vodkin et al., 2004). Expressed sequence tags (ESTs) are coding regions within the DNA that can also be used for MAS or mapping purposes. The isolation of 6570 full-length sequences of soybean cDNAs derived from tissues exposed to different abiotic stresses will aid marker development (Umezawa et al., 2008) as will the isolation of ESTs from drought-stressed soybean root tips (Valliyodan & Nguyen, 2008).

such as stress tolerance. Quantitative traits are usually affected by more than one gene and by the environment. The association of morphological or physiological phenotypic markers with molecular markers in QTL analysis is considered to ultimately be the ideal approach to accelerate plant breeding programs in marker-assisted selection (MAS). MAS is the process whereby the identified markers of all types (morphological, physiological, biochemical or molecular) are used for indirect selection of required traits. MAS approaches however are far from trivial and they are often not cost effective in current breeding programs. Drought tolerance is a mutagenic trait, often influenced by large numbers of minor QTLs, rather than one or more major QTLs. This means that the development and effective large-scale

The identified molecular DNA markers should ideally be applicable using relatively simple methods that are amenable to automation and high throughput. However, the isolation DNA from plants is not technically demanding and can be carried out in any laboratory that has basic DNA isolation and characterisation equipment. Potential markers can be first evaluated in small sample subsets in order to confirm that the desired loci are present. Such methodologies normally require large sample collections or population sizes because the chosen markers have be applied to large segregating populations in order to determine

Relatively few markers for drought tolerance have been identified in soybean. However, progress in this regard will be accelerated once large collections of potential DNA sequences have been established for marker development (Shinozaki, 2007). Accurate genetic and physical maps of the soybean genome are also essential for the development of useful molecular markers for drought tolerance in soybean. Towards this goal, a total of 318 AFLP, 121 SSR, 108 RFLP, and 126 STS markers have been integrated into a linkage map composed of 509 RFLP, 318 SSR, 318 AFLP, 97 AFLP-derived STS, 29 BAC-end or EST-derived STS, 1 RAPD, and five morphological markers (Hisano et al., 2007, Xia et al., 2007). A further very useful advance has come from the sequencing of the soybean (Glycine max (L.) Merr. 'Williams 82') genome. Some 66,153 protein-coding loci are now available at: http://www.phytozome.net/soybean. Moreover, 3290 microsatellites (SSRs) identified from BAC end sequences of clones (comprising the 'Williams 82' physical map) were screened and two hundred and sixty-five SSRs were genetically mapped in at least one

Microarrays involve the immobilisation of single-stranded DNA on a solid support that is hybridised with a single-stranded DNA or RNA population (Rockett and Dix, 1999). In microarray analysis gene expression and regulation patterns can be monitored on a large scale (Quackenbush, 2001). The technology facilitates screening for differently expressed genes in different plant varieties. Differently expressed genes can potentially be useful in

A spotted soybean cDNA microarray containing 36,000 elements derived from EST libraries is available that covers a wide range of tissues and organs at different developmental stages under optimal and stress conditions (Vodkin et al., 2004). Expressed sequence tags (ESTs) are coding regions within the DNA that can also be used for MAS or mapping purposes. The isolation of 6570 full-length sequences of soybean cDNAs derived from tissues exposed to different abiotic stresses will aid marker development (Umezawa et al., 2008) as will the isolation of ESTs from drought-stressed soybean root tips (Valliyodan & Nguyen, 2008).

MAS providing that they encode proteins involved in the traits of interest.

application of MAS is still technically challenging, costly and time-consuming.

whether the markers are linked to the required traits (Ribaut et al., 2002).

mapping population (Shoemaker et al., 2008).

**4.1 Identification of markers using microarrays** 

We have analysed the crown nodule transcriptome at two stages of development (i.e. on 7 and 11 week-old plants) in order to identify possible markers for nodule development/senescence that might underpin plant performance. A comparison of the crown nodule transcriptomes was performed on two cultivars (PAN809 and Highveld Top) after 7 and 11 weeks of growth. PAN 809 is generally recommended for use under long growing seasons and this variety is often used in moderate and hot regions of Southern Africa because it is well-suited to irrigated- and rain-fed cultivation conditions. Highveld Top was developed specifically to withstand the low night temperatures that are often experienced during the growing season at high altitude in Southern Africa. The first nodule harvest point (7 weeks) was chosen because the measured capacity of symbiotic nitrogen fixation was highest at this point of nodule development. The second harvest point (11 weeks) was chosen because symbiotic nitrogen fixation had decreased by 50% at this time point relative to week 7. Figures 8 and 9 illustrate the degree of genotypic variation in the nodule transcriptome signatures observed in the nodules of 7 and 11 week-nodules under optimal growth conditions. While 702 transcripts were changed in abundance in Highveld Top nodules harvested at week 11 compared to week 7, 1737 transcripts were differentially expressed in PAN809 nodules under the same conditions (Figure 8). Of these, 226 transcripts were identical and showed similar patterns of increase or decrease at week 11 relative to week 7 in both cultivars (Figure 8). Of the transcripts that were differentally expressed in PAN809 and Highveld Top nodules at week 7, the PAN809 nodule transcriptome had a much higher number of transcripts encoding proteins involved in cell wall and stress metabolism than Highveld Top (Figure 9). For example, an extensin-like protein and proteins involved in disease resistance were much more abundant in PAN 809 nodules than Highveld Top nodules at week 7. Moreover, transcripts encoding the extensinlike protein were enhanced to a much greater extent in PAN 809 than Highveld Top. Transcripts encoding disease resistance proteins were also much higher in PAN 809 nodules than Highveld Top nodules at 11 weeks. The biological relevance of these findings is currently under investigation in relation to the different performance of these two cultivars.

Fig. 8. A comparison of differential gene expression in Highveld Top (HT) and PAN 809 nodules harvested from 7- and 11-week old plants.

Identification and Application of Phenotypic

PAN than HT.

**soybean nodule development** 

and Molecular Markers for Abiotic Stress Tolerance in Soybean 195

We compared nodule structure, carbon/nitrogen interactions and respiration in PAN, which is chilling-sensitive and in HT, which is more chilling-resistant (van Heerden et al., 2008). Under optimal growth conditions, SNF began to decline after 9 weeks in PAN nodules and after 11 weeks in HT. A transcriptome analysis was performed on PAN and HT nodules harvested from plants at 3 until 15 weeks after germination. Of the genes that showed the same developmental pattern in both varieties transcripts encoding a cysteine proteinase gene (Gma.8481.1.S1\_at) that belongs to a subgroup a vacuolar processing enzymes (legumains) was up-regulated in the senescent nodules. Nodule numbers were unaffected by dark chilling in both genotypes. The abundance of the nitrogenase and leghemoglobin proteins was not changed as a result of dark chilling but nodule respiration rates, nitrogenase activity and NifH and NifK mRNAs were decreased while nodule starch, sucrose and glucose were increased. Chilling-induced decreases in nodule respiration continued in PAN nodules after return to optimal temperatures but respiration recovered in HT by the end of the chilling period. This recovery was associated with a large decrease in the area of the intercellular spaces in the nodule cortex and infected zone in HT. This acclimatory response was not seen in PAN nodules. We conclude that the ability to regulate the oxygen diffusion barrier is an important component of ability of nodules to tolerate stress (van Heerden et al., 2008). The HT nodules were able to regulate both respiration and the area of the intercellular spaces during chilling and so control the oxygen diffusion barrier. We conclude that chilling-induced inhibition of SNF in PAN nodules was caused by the inhibition of respiration coupled to the failure to regulate the oxygen diffusion barrier effectively (van Heerden et al., 2008). Furthermore, the stress-induced limitations in SNF make an important contribution the greater chilling-induced inhibition of photosynthesis in

The characterisation of simple but accurate phenotypic markers for enhanced drought tolerance is important because drought is considered to be the most important factor limiting soybean productivity in the field. A comparison of shoot, root and nodule parameters in three genotypes: Prima 2000, glyphosate-resistant A5409RG and Jackson revealed a positive correlation between SNF and photosynthesis under optimal and drought conditions (Fenta et al., 2011). Considerable genotypic variation was observed in the responses of photosynthesis to drought. While Jackson and Prima performed better than A-5409RG in short-term drought, SNF in Jackson nodules was equally inhibited in all cultivars under long term drought conditions. Drought-induced decreases in shoot to root ratios occurred in all three cultivars, together with a reduction in whole plant biomass (Fenta et al., 2011). However, the shoot to root ratios under drought were significantly higher in A-5409RG than Jackson or in Prima, showing that there is considerable genotypic variation in the control of shoot to root ratios in soybean in response drought. We conclude that that the ability to sustain shoot biomass under the nitrogen limitation caused by impaired SNF could

**5.1 Characterization of chilling-induced tolerance traits in soybean** 

**5.2 Characterization of drought tolerance traits in soybean** 

used as a marker for drought tolerance in soybean (Fenta et al., 2011).

**5.3 A role for CLAVATA3/Embryo Surrounding Region (CLE) peptide signalling in** 

Nodule development is an energy demanding process and so it might be more economical for the plant, to be able to form few larger nodules compared to many small ones. Indeed,

Fig. 9. A comparison of the relative expression of different transcripts in the crown nodules from Pan 809 and Highveld Top. Positive values represent relative increases in transcript abundance while negative values represent decreases in the abundance of transcripts encoding proteins involved in RNA/signaling (RNA), cell wall metabolism (CW/Cell), defense or stress responses (Stress), regulation (Regul), hormone metabolism (Hormon), protein metabolism (Protein), primary metabolism (Carbo), secondary metabolism (Second) and transport (Transp).
