3.3.3. Identification of disease-responsive genes
SMD and FW are two serious diseases that adversely affect pigeonpea production. With an objective of identifying candidate genes for these diseases, Illumina 1G sequencing was used on the transcriptomes of FW-challenged root tissues and SMD-challenged leaf tissues of five each resistant (ICPL 87119, ICPB 2049, ICPL 20096, BSMR 736 and ICPL 7035) and susceptible genotypes (ICPL 87091, ICPL 99050, TAT 10, ICPL 332 and TTB 7). The number of Illumina tags (36 bp long) ranged from 18 644 113 (ICPL 87119) to 14 514 194 (TAT 10) for the 10 genotypes. The sequence data of these Illumina tags have been submitted to the National Center for Biotechnology Information (NCBI). The data can be accessed at (http://www.ncbi.nlm.nih.gov/
) and accession numbers are: SRA030523.1 to SRP005971.1. These tags were aligned to the CcTA (Supplementary Table S3
). As a result, ~35 million Illumina tags could be aligned to 54 426 TUSs. Numerical comparison of these tags between a pair of resistant and susceptible genotypes for a disease (usually the parents of a mapping population) was used to identify differentially expressed genes for a disease.
Since the numbers of Illumina tags mapped to the transcriptome assembly varied among genotypes, the data were normalized per million reads. For the SMD study, a numerical comparison of SMD-responsive reads generated from three resistant (ICPL 20096, BSMR 736 and ICPL 7035) and three susceptible (ICPL 332, TAT 10 and TTB 7) genotypes representing three mapping populations was conducted. The Log 2 threshold for this analysis was taken as −2 to +2. The number of TUSs showing expression differences at these cutoffs ranged from 7505 (BSMR 736 × TAT 10) to 10 497 (ICPL 20096 × ICPL 332). In the case of the TTB 7 × ICPL 7035 combination, the number of differentially expressed genes was 9402. Similarly, in the FW study, a comparison was made between the specific parental combinations used to develop two different mapping populations (with the same thresholds) to find TUSs with differential expression. The number of TUSs with significant differentially expressed genes ranged from 6673 (ICPB 2049 × ICPL 99050) to 11 518 (ICPL 87119 × ICPL 87091) (Fig. ).
Figure 3. Distribution of differentially expressed genes in SMD- and FW-responsive genotypes. Differential expression was calculated based on Log 2 value, with a threshold of less than −2 to greater than +2 number of differentially expressed gene was calculated (more ...)
Based on the expression values for differentially expressed genes in SMD- and FW-responsive genotypes, hierarchical clustering was done for SMD- and FW-responsive genes separately to compare the pattern of gene expression. These clusters show the pattern of co-regulated genes for the SMD-responsive genotypes (Fig. A) and for the FW-responsive genotypes (Fig. B).
Figure 4. Hierarchical clustering of differentially expressed TAs within SMD- and FW-responsive genotypes. Hierarchial clustering of the gene involved in SMD- and FW-stress responses was done using HCE version 2.0 beta web tool. These two dendrograms illustrate (more ...)
In order to study the gene expression pattern between two parental genotypes of a mapping population, the numbers of up-regulated and down-regulated TUSs were calculated with respect to the resistant parent. The numbers of up-regulated TUSs remained high in all the crosses studied, with an exception of ICPL 87119 × ICPL 87091, which had more down-regulated TUSs (11 364) when compared with up-regulated (154). Log 2 fold differences between the parental combinations are shown in Table .
Summary of differentially expressed TUSs across five parental combinations
Functional annotations of differentially expressed TUSs are described next, with the additional requirement of ≥5 fold differences across all the parental combinations. The annotation analysis was conducted in three ways: (i) TUSs differentially expressed across all the 10 genotypes, (ii) TUSs differentially expressed in 6 SMD-responsive genotypes separately and in 4 FW-responsive genotypes separately and (iii) the common set of TUSs that is differentially expressed in both FW- and SMD-responsive genotypes. Based on these analyses, in the first category, 6107 TUSs (with fold difference ≥5) were selected for functional classification. Considering an E
-value cutoff of ≤1E−08 and a bit-score value of ≥50, functional annotation for 3698 TUSs showed significant similarity with the UniRef non-redundant protein database. No significant matches were found for 2409 TUSs. For 3698 TUSs with functional classes, we found that in addition to basic housekeeping genes, these TUSs also showed homology to genes involved in stress response, such as proline-rich protein, Syringolide-induced protein, desiccation protective protein of soybean, ABA-responsive protein, and leucine zipper protein (Supplementary Table S4
Among these 3698 TUSs, 2106 could be assigned into three major categories: (i) molecular function, (ii) biological process and (iii) cellular component. These categories were further subcategorized, i.e. under molecular function category, the subcategory ‘binding’ accounted for highest percentage of TUSs (594), followed by ‘catalytic activity’ (513), ‘transporter’ (58), ‘structural molecule’ (52) and rest of the subcategories accounting for 75 TUSs. Similarly, under ‘biological process’ category, the highest number of TUSs were assigned to the subcategory ‘metabolic process’ (571) followed by ‘cellular process’ (542), response to ‘stimuli’ (152), ‘biological regulation’ (118), ‘establishment of localization’ (108) and 363 TUSs accounted for rest of the subcategories. Under ‘cellular component’ category, the highest percentage of TUSs was assigned to the subcategory ‘cell part’ (562) followed by ‘organelle’ (381), ‘organelle part’ (202), ‘macro molecule complex’ (158) and 75 TUSs were assigned to rest of the subcategories.
Differentially expressed genes included those encoding proline-rich proteins, zinc finger proteins, leucoanthocyanidin dioxygenase and RAS-related protein. There were seven TUSs that correspond to proline-rich protein. This protein forms a component of glutamate pathway and has multiple developmental and stress-related functions. High expression of this protein in leaves has been reported to play major role in the early stage of virus infection in soybean.28
The glutamate pathway assimilates nitrogen and produces glutamate, which then acts as a starting point for amino acid synthesis. A 5-fold up-regulation of this gene in resistant genotype (ICPL 2049) probably implicates its role in response to this stress. A gene for zinc finger protein showed an average of 5-fold differential expression in both SMD- and FW-responsive genotypes. This protein is a component of mitogen-activated protein kinase (MAPK) pathways which are demonstrated to play an important role in regulating the gene expression in response to various biotic as well as abiotic stress in species such as Arabidopsis.29,30
MAPK pathways transduce a large variety of external signals, leading to a wide range of cellular responses, including growth, differentiation, inflammation and apoptosis. A total of six TUSs had homologies to leucoanthocyanidin dioxygenase, and showed an average of 5-fold differential expression. The up-regulation of this gene (in the flavonoid pathway) is known to play an important role in defence against both biotic and abiotic stress by acting as a passive or inducible barrier against pathogens.31
A total of 21 TUSs showed homology to gene for RAS-related protein ARA-3. These TUSs were up-regulated 6-fold. This gene is involved in the ethylene-mediated signalling pathway, suggesting an important role in stress response.
With an objective of identifying candidate genes for FW and SMD, as mentioned above, the common set of TUSs with ≥5 fold expression difference was identified in SMD- and FW-responsive genotypes, separately. From this, 99 common TUSs were found for FW-responsive genotypes and 13 for SMD-responsive genotypes. Functional characterization of these genes showed function for 51 FW-responsive TUSs and 3 SMD-responsive TUSs (oxygen-evolving enhancer protein, NADH-ubiquinone oxidoreductase and sedoheptulose-1,7-biphosphate). FW responsive TUSs include genes such as mannose-1-phosphate guanyltransferase, prolinedehydrogenase, cellulose synthase, pectinesterase inhibitor, superoxide dismutase [Fe] and vacuolar protein sorting-associated protein.
Among FW-responsive genes were TUSs showing cellulose synthase homology. These genes are essential for secondary cell wall synthesis. Among the SMD-responsive TUSs, one showed homology to oxygen evolving enhancer protein, with an average 5-fold expression difference. These proteins are components of glycine-rich protein 3/wall-associated kinase. One TUS showed homology to a gene coding for NADH-ubiquinone oxidoreductase and was up-regulated in resistant with respect to susceptible genotypes for SMD. This is a common component for energy evolving pathways in the cell.
Considering biotic stress responsive genes in common for FW and SMD, no TUS was found common at the threshold of ≥5 fold difference. When this threshold was decreased to ≥2 fold, the number of common TUSs across FW- and SMD-responsive genotypes was found to be 192. Of this set, 99 TUSs were functionally annotated and 93 were uncharacterized proteins. These annotated TUSs showed sequence similarity to several stress responsive genes such as zinc protein, aminocyclopropane carboxylate oxidase, cysteine protease and hexokinase. For example, two TAs showed homology to a gene corresponding to zinc finger protein and expressed with an average of 3.4-fold difference. As mentioned, this gene is a component of MAPK. TUSs with sequence similarity to gene for 1-aminocyclopropane-1-carboxylate oxidase were expressed with an average 3.3 folds. This gene is a component of ethylene-biosynthesis pathway which plays an important role in ethylene biosynthesis at stress conditions.32
TUSs with sequence similarity to gene for synthesis of germination-specific cysteine protease also showed an average of 3-fold difference in expression value, this gene is responsible for cell death hence regulating response to stress. Sequence similarity for another gene encoding for hexokinase-2 was also discovered for one TUS which showed an average-fold difference of 2.9. This gene is known to play a major role in metabolic pathways, e.g. fructose and mannose metabolism, galactose metabolism and glycolysis.
Genes responding in the FW- and SMD-resistant lines will provide a rich basis for further explorations of the mechanisms of disease resistance for these important viral and fungal diseases, and may also be useful in identifying regulatory networks and targets for breeding efforts, As no controls (Illumina tags generated from non-challenged tissues) were used for identification for FW- and SMD-responsive genes, like recent studies in wheat (Triticum aestivum
and yam (Dioscorea alata
, it is, therefore suggested to use other techniques like qRT–PCR to validate and establish magnitudes of expression levels of identified genes before they are used for further studies.