A Study of the Genetics of Root Nodulation in Pigeonpea (Cajanus cajan) Using Indigenous Rhizobia

Abstract

Pigeonpea (Cajanus cajan) is an important grain legume, which is grown in many African countries largely for human and animal consumption. Leguminous crops such as pigeonpea fix atmospheric nitrogen (N) symbiotically in the root nodules thus elimination of the need for expensive chemical nitrogenous fertilizers. The determination of host plant x microsymbiont combinations that produce optimum crop productivity is important in the pigeonpea value chain. In addition, the characterization of both symbionts is necessary for exploiting the N fixation in tropical legumes including pigeonpea. There is a dearth of information regarding the agronomic performance of pigeonpea and other common legume species such as Bambara groundnut (Vigna subterranea), soybean (Glycine max) and tepary (Phaseolus acutifolius) that are cultivated in South Africa. Therefore, the aim of this study was to evaluate the symbiotic effectiveness of indigenous rhizobia on pigeonpea.. The specific objectives of the study were to: (i) collect rhizobial strains that are associated with root nodulation in pigeonpea from diverse locations across South Africa (ii) perform molecular characterization of the rhizobia that are associated with root nodulation in pigeonpea from diverse locations in South Africa (iii) to sequence the whole genome of a selected rhizobial strain derived from pigeonpea and determine its molecular characteristics and (iv) determine the effectiveness of the rhizobial strains with pigeonpea and other common tropical legume species. In the first objective of the study, forty soil samples were collected from diverse locations across the country and used for inoculating separately the seed of each of five randomly selected pigeonpea genotypes. The pigeonpea plants were raised in a N-depleted growth medium in the greenhouse. A split-plot experimental design with two replications was used in the study. After six weeks of growth, the plants were harvested to isolate rhizobia from the root nodules. Several morphological characteristics of the rhizobial colonies including shape and growth habit (type) were determined. In addition, a range of N fixation variables of the host pigeonpea plants was measured including the nodule dry weight (NDW) and shoot dry weight (SDW) per plant. A variety of the colony morphologies ranging from tiny to medium as well as cream white and large, watery oval colonies was observed. Two hundred and eighty putative pigeonpea rhizobial strains were obtained from the root nodules of the plants. Based on their morphological characteristics on YMA-CR, nutrient agar and peptone glucose agar, the isolates were deposited into the South African Rhizobium Culture Collection gene bank. There was ˃40.0 % difference in the number of nodules between ‘Genotype-5’ and ‘Genotype-4’ but the difference in NDW between the two genotypes was ˃80.0 %. In contrast, the heaviest dry shoots (0.4513 g) that were attained by ‘Genotype-3’, weighed 52.0% more than the lightest dry shoots that were observed for ‘Genotype-4’. The results indicated that the soil samples contained diverse rhizobial isolates with distinct morphological characteristics and significant differential N fixation ability of the pigeonpea genotypes suggesting that there was a potential to select for optimum host genotype x rhizobial strain combinations for N fixation in this legume species. In the second objective of the study aimed at the molecular characterization of the rhizobial strains derived from pigeonpea, two housekeeping bacterial genes (namely 16S rRNA and recA) were used to identify each rhizobial strain to the species level. In addition, the phylogenetic relationships among these rhizobial strains were determined. The results showed that 56 strains were confirmed as rhizobia and deposited into the national rhizobia collection bank. Two primers successfully amplified both the Rhizobium strain (30bp3) as well as several Bradyrhizobium strains (16a2p3, 15bp3, 11a2p3, 13bp3, 33ap4 and 19a1p3). Two novel genera of rhizobia (Phyllobacterium and Paraburkholderia), were associated with root nodulation in pigeonpea. There was a considerable variation in the size of the sequences of both the 16S rRNA and recA genes among the rhizobial isolates. The sequences of the 16S rRNA genes across the four genera averaged 1015.73 bp. The 16S rRNA Rhizobium phylogenetic tree showed that the rhizobial isolates obtained from pigeonpea were dispersed in six different clusters and grouped with several type species of the genus Rhizobium including R. tropici with a high (77.0%) similarity grouping. The 16S rRNA based phylogenetic tree showed that the novel Paraburkholderia rhizobial isolate ‘30a2p3’, grouped with several type strains including P. rhizosphaerae (with a similarity grouping >50%) but in the recA based phylogenetic tree, the same isolate was grouped in Cluster IV with 93.0% similarity grouping. The study concluded that the sequences of the two genes (16S rRNA and recA) of the isolates from pigeonpea could provide sufficient phylogenetic information about the isolates up to the species level and confirmed that this legume is promiscuous in diverse soils from South Africa. The objective focusing on the whole genome sequencing of a selected rhizobial strain derived from pigeonpea to determine its molecular characteristics selected the rhizobial strain 10ap3 (SARCC-755) (that was originally derived from pigeonpea in the trapping experiments). Upon DNA from the strain, the DNA libraries were prepared using the Nextera protocol (Illumina, USA) and paired-end (300bp x 2) sequenced on a MiSeq (Illumina) sequencer at the Biotechnology Platform, Agricultural Research Council-Onderstepoort Campus (Pretoria, South Africa). The genome was a large circular chromosome (6,297,373 bp) and containing the overall G + C content of 60.0%. The total number of genes in the genome of the strain was 6,013 of which 99.13% were coding sequences. However, only 5,833 of the genes were associated with proteins that could be assigned to specific functions. Several important genes that were found on the genome, included the genes for N metabolism, stress response, phosphorus metabolism and iron acquisition as well as adenosine monophosphate nucleoside for purine conversion. The nodulation gene (nolR), which functions as a DNA binding transcription factor was located on contig 12. Precursor genes for purine synthesis, for instance, inosine-5-monophosphate and adenylosuccinate, which are also responsible for nodule formation, were also present on the genome. The results showed that the genome of this strain (Rhizobium tropici SARCC-755) does not contain common nod and nif genes suggesting that an alternative pathway involving a purine derivative was involved in its symbiotic association with pigeonpea. The genome also possessed some genes that are associated with abiotic stresses and mineral nutrient acquisition thus making it a candidate for future formulation of commercial inoculants especially when considering its high symbiotic efficiency with pigeonpea. In the fourth study objective, which focused (i) on determining the relative performance of individual tropical legume species when inoculated separately with each of the specific rhizobial strains that were previously derived from pigeonpea, (ii) quantify the magnitude of the effects of interactions between the host tropical legume species x rhizobial strain on a range of N fixation variables and (iii) identify the winning (superior) rhizobial strains with the specific test legume species. Thirty-six rhizobial strains which were previously isolated from pigeonpea root nodules were used in the study. There was at least one strain representing each of four distinct rhizobial genera namely Bradyrhizobium, Paraburkholderia, Phyllobacterium and Rhizobium. The experiment was laid out as a split plot design with legume species as the main factor and rhizobial strain as the sub-factor. Each treatment was replicated twice. The data sets of several N fixation variables including NDW and SDW were measured and subjected to the analysis of variance and Pearson’s correlation analysis using SAS statistical software (version 9.3) followed by mean separation using LSD test at the 5.0% probability level. Further analysis using the GGE biplot model was carried out to understand better the relationship between the host plants and the microsymbionts. For objective (i) of the fourth study, three healthy seeds of each legume species were planted in a plastic pot filled with 1.65 kg sterile river sand saturated with Hoagland solution. At seven weeks after germination, each plant was harvested and gently washed with tap water before detaching the nodules carefully from the roots. Similarly, the shoot was separated from the roots for each plant prior to oven drying all the harvested plant parts at 70o C for 48 h followed by weighing to determine the dry weights. The results showed marked variability in the responses of the legume species to inoculations with individual rhizobial strains. Tepary bean showed poor nodulation as indicated by the chlorotic plants which contrasted sharply with those of Bambara groundnut. Pigeonpea responded differentially to each individual rhizobial strain resulting in marked differences in the nodule load per plant. Some rhizobial strains, (for instance, Rhizobium strain ‘26a-PP3’) induced profuse nodulation in Bambara groundnut but not in the other legume species. The principal component (PC) analysis showed that the first two principal components accounted for 78.74% of the total variation. Four N fixation variables, including the NDW and SDW, were moderately associated with PC1. The GGE biplot of the rhizobial strain x legume species interaction for NDW explained 82.44% of the total variation. For NDW, the environments represented by E3 (soybean) and E4 (pigeonpea) were positively correlated since their vectors were separated by an acute angle. However, E1 (Bambara groundnut) and E3 (soybean) were negatively correlated since they were characterized by an obtuse angle between them for the SDW. The ‘which-won where’ biplot for NDW explained 82.44% of the total variation of which PC1 and PC2 accounted for 50.40% and 32.04% of the total variation, respectively. Two rhizobial strains on the vertices of the polygon (Rhizobium sp. 36a-PP5) and (Rhizobium sp. 26a2-PP5) performed best with Bambara groundnut (E1) and soybean (E3), respectively. E3 (soybean) consisted of the longest vector line suggesting that it possessed a high discriminating ability. Two rhizobial strains, namely, (Rhizobium sp.; 33a-PP2) and (Rhizobium multihospitium; 37a-PP4) were identified as ideal for RDW and SDW, respectively. The biplot analysis also revealed that for SDW, E1 (Bambara groundnut) and E4 (pigeonpea), in that sequence, were plotted closet to the epicentre. The GGE biplot analysis also revealed that both pigeonpea and Bambara groundnut provided the most ideal symbiotic activity for NDW but tepary bean lacked the discriminatory ability for NDW. Further testing and validation of the symbiotic activities of the rhizobial strains identified in this study in field trials on diverse legume species and in multiple agro-ecological locations is recommended. It will also be desirable to identify new bio-inoculants for improving tepary bean productivity.