Differential response of kabuli and desi chickpea genotypes toward inoculation with PGPR in different soils

Pakistan is among top three chickpea producing countries but the crop is usually grown on marginal lands without irrigation and fertilizer application which significantly hampers its yield. Soil fertility and inoculation with beneficial rhizobacteria play a key role in nodulation and yield of legumes. Four kabuli and six desi chickpea genotypes were, therefore, evaluated for inoculation response with IAA-producing Ochrobactrum ciceri Ca-34T and nitrogen fixing Mesorhizobium ciceri TAL-1148 in single and co-inoculation in two soils. The soil type 1 was previously unplanted marginal soil having low organic matter, P and N contents compared to soil type 2 which was a fertile routinely legume-cultivated soil. The effect of soil fertility status was pronounced and fertile soil on average, produced 31% more nodules, 62% more biomass and 111% grain yield than marginal soil. Inoculation either with O. ciceri alone or its co-inoculation with M. ciceri produced on average higher nodules (42%), biomass (31%), grains yield (64%) and harvest index (72%) in both chickpea genotypes over non-inoculated controls in both soils. Soil 1 showed maximum relative effectiveness of Ca-34T inoculation for kabuli genotypes while soil 2 showed for desi genotypes except B8/02. Desi genotype B8/02 in soil type 1 and Pb-2008 in soil type 2 showed significant yield increase as compared to respective un-inoculated controls. Across bacterial inoculation treatments, grain yield was positively correlated to growth and yield contributing parameters (r = 0.294* to 0.838*** for desi and r = 0.388* to 0.857** for kabuli). PCA and CAT-PCA analyses clearly showed a site-specific response of genotype x bacterial inoculation. Furthermore, the inoculated bacterial strains were able to persist in the rhizosphere showing colonization on root and within nodules. Present study shows that plant growth promoting rhizobacteria (PGPR) inoculation should be integrated with national chickpea breading program in Pakistan especially for marginal soils. Furthermore, the study shows the potential of phytohormone producing strain Ca-34T as promising candidate for development of biofertilizer alongwith nodulating strains to get sustainable yield of kabuli and desi chickpea with minimum inputs at marginal land.


Introduction
Due to high nutritive value, chickpea (Cicer arietinum L.) is one of the earliest cultivated and third widely grown edible legume in tropical, sub-tropical and temperate regions of the world. Pakistan is the 2nd largest chickpea growing and third biggest chickpea producing country with an average yield of 561 kg/ha (FAO statistics, 2014 1 ) where the crop is grown as the principal winter pulse crop, fodder and green manure. Crop legumes, grown in rotation with cereal crops, can improve yields of the cereals and contribute to the total nitrogen (N) pool in soil. Legumes especially chickpea occupies special position regarding nutrition as well as soil fertility and improvement. It has the ability to grow well in poor soils as well as to improve them because of its efficient N fixation system. It can happily grow on marginal, poorly fertile sandy loam land, almost exclusively under rain-fed conditions in areas of low rainfall without any chemical or biological fertilizer. Soil factor exert greater influence than bacterial inoculation on plant growth, nitrogen fixation and nutrient uptake of plant (Neumann et al., 2011).
Sustainable production depends upon the manipulation of all genetic and environmental factors that influence crops by exploiting high yielding varieties and manipulation of its symbiotic system. The plant growth promoting rhizobacteria (PGPR) induce plant's nutrient acquisition, disease tolerance and play a vital role in crop yield (Lugtenberg and Kamilova, 2009). Growth and yield of the plant have been improved by repeated inoculation of highly effective rhizobia (Hynes et al., 2008) and/or co-inoculation with PGPR (Mirza et al., 2007;Mishra et al., 2009;Tsigie et al., 2011). Improving biological nitrogen fixation (BNF) in food crops may increase plant-based protein for human consumption and increase growth of subsequent crops with lesser chemical inputs.
Implementation of PGPR-based biofertilizer technology presents economic, environmental and agronomic benefits and could be used to a larger degree to partially replace the synthetic fertilizers (Silva and Uchida, 2000;) to improve economic yield under natural conditions. Repeated incorporation of rhizobia with more effective strains coupled with the addition of "helper bacteria" can add to the BNF of the crop. Strain x variety interaction is as important as that of strain or crop variety alone (Abi-Ghanem et al., 2011). Selection of best microbial strains and plant variety/cultivar is important because a strong effect of cultivar x microbe has been reported on BNF in soybean (Israel, 1981;Israel et al., 1986), beans (Valverde and Otabbong, 1997), peanuts (Wynne et al., 1980), lentils and peas (Hafeez et al., 2000;Abi-Ghanem et al., 2011). Comparison of N 2 fixation among various rhizobial strains demonstrated 42% variability among strains while 81% variability among different lentil cultivars (Hafeez et al., 2000). This further supports the fact that each host cultivar has a variable potential for nitrogen fixation and response toward rhizobial inoculation.
Of the two chickpea varieties, the desi type accounts for 85% and the kabuli type for 15% of the area. These two 1 FAO, 2014. www.fao.org. types differ in their yield potential and hence in nutrient requirements. Each ton of chickpea grain removes 121.9 kg of primary nutrients (67.3 N + 6.6 P + 48 K), 34.7 kg of secondary nutrients (18.7 Ca + 7.3 Mg + 8.7 S) alongwith ∼1000 g of four micronutrients (38 Zn + 868 Fe + 70 Mn + 11.3 Cu) (Aulakh et al., 1985;Prasad et al., 2002). Chickpea is although a hard crop and can grow well even in the marginal soil and soils of varying textures. In general, the soils of chickpea-growing areas have low organic carbon content, which is indicative of low soil fertility. Chickpea can be grown successfully in soils with a pH ranging from 6 to 9. Low pH (<4.6), besides limiting some micronutrient availability, causes some toxicity problems and poor nodulation. Depending on soil fertility, climate and plant factors, nutrient, e.g., N, P, Fe, B and S deficiency in the soil causes yield losses up to 10%, 29-45%, 22-90%, 100%, and 16-30%, respectively (Ali et al., 2002).
Members of genus Ochrobactrum are adapted to a wide range of habitats and usually considered as free-living and ubiquitous. Although frequently isolated from the soil and rhizosphere and forming a substantial population in the rhizosphere, it has considerable effect on plant growth (Faisal and Hasnain, 2006;Príncipe et al., 2007;Chakraborty et al., 2009) andbiocontrol agent (Cook et al., 1997). Furthermore, members of this genus have significant applications in the biodegradation of a range of toxic materials in soil (Yamada et al., 2008) and treatment of waste water (Ozdemir et al., 2003). O. ciceri was isolated from the nodules of chickpea plants in Pakistan (Imran et al., 2010) and contain IAA production ability in vitro. The response of inoculation and the rhizosphere colonization potential of this strain over other bacteria in the rhizosphere however, need to be determined in vivo.
Extensive breeding and management programs are going on for the development of high yielding chickpea genotypes with improved disease resistance but no data is available for the PGPR-inoculation response of these genotypes/varieties. This study was, therefore, planned to evaluate the response of both kabuli and desi genotypes/varieties to inoculation with PGPR to screen the best strain x genotype combination for further improvement of nodulation and yield of chickpea in local farming system. The ultimate objective was to develop a PGPR based biofertilizer for resource-poor farmers cultivating chickpea at marginal lands.

Plant Material
Six desi and four kabuli chickpea genotypes (Table 1), including both adapted varieties and advanced breeding lines (i.e., "genotypes") were obtained from Nuclear Institute of Agriculture and Biology (NIAB), Faisalabad. The seeds were multiplied under identical conditions at a single site to minimize the seed source effect. The desi and kabuli chickpea genotypes chosen in this study represent a wide range of variation in seed size and yield, disease resistance and origin. O. ciceri Ca-34 T and M. ciceri TAL-1148 were grown in LB and YEM, respectively, at 28 ± 2 • C over-night with constant shaking. The cells were harvested with centrifugation and suspended in saline to get 10 9 cells ml −1 . Bacterial cultures were mixed with sterilized finely ground carrier material (finally grinded filter mud). The seeds were mixed until they become coated with a thin film of bacterial inoculum. Seeds for the uninoculated control were coated in a similar manner with the sterilized carrier material prepared in saline. Coated seeds were air-dried in shade before sowing.  row spacing. Two rows at the beginning and two at the end of each block were kept for protection. Genotypes (10) were used as main plot while treatments (four) were used as subplots. Fields were prepared by pre-sowing irrigation (10 days before field preparation) and one bag DAP (N 18%, P 2 O 5 46%) fertilizer @ 125 Kg/Ha applied during the field preparation. No further irrigation or fertilizer was applied during the whole experiment.

Measurement and Data Analysis
Six plants were randomly selected from the middle rows of each replicate at 60, 120, and 180 DAS (days after sowing). Plant height and fresh weights were recorded immediately while dry weights were recorded after oven drying at 65 • C for 48 h. Number of seeds per pod was calculated by counting 25 randomly selected pods/plant from each treatment and taking their averages. 100-seed weight was calculated by taking the average of 300 seeds from each treatment. Grain and straw yield was calculated by taking the yield of 25 random plants from each treatment and then converted to Kg/ha. Total biomass was calculated from weight of sun-dried 25 plants (including grain and straw) from each treatment and converted to Kg/ha. Harvest index (HI) percentage was calculated using the formula: HI (%) = Seed yield/total biomass × 100. Relative effectiveness (RE) was determined by following equation and expressed in percent as described by Maâtallah et al. (2002).
Relative effectiveness (RE) = (DW ino /DW cont ) × 100 Where DW ino = Dry weight of inoculated plant; DW cont = dry weight of un-inoculated control plants. Split-plot ANOVA was performed using Statistix 8.1 and the means [±standard deviation (SD)] were compared using Tukey HSD all-pairwise comparison test. Mean separation was done using LSD at p = 0.05. Regression and Principal Component Analysis was performed using SPSS software package version 17.0 (SPSS, Inc., Chicago, IL).

Detection of Inoculants
Bacteria were quantified from chickpea roots at 30 DAS and nodules at 80 DAS using serial dilution plating technique. The nodules and roots were carefully detached from the plants. One gram soil tightly adhering to the roots was used to prepare serial dilution. Nodules were washed thoroughly with sterile distilled water, then flame sterilized by dipping in ethanol. Sterilized nodules were incubated overnight on LB agar plates to see the surface contamination. Only those nodules were processed for serial dilutions that did not show any surface contamination. The dilutions were spread onto YEM-congored to recover TAL-1148. Ca-34 T was recovered on YEM-congored plates containing kanamycin (30 µg), aztreonam (30 µg) and rifampicin (5 µg).
Colonies showing similar colony morphologies to those of the inoculant strains were subjected to immunoblotting. Polyclonal antibodies were raised in 6 months old female albino rabbits and immunoblotting was carried out using standard protocols (Somasegaran and Hoben, 1994). Both the antisera showed no reactivity with antigens of Pseudomonas spp. 96-51 (Rasul et al., 1998) and MST4.1 (Hafeez et al., 2006), Meso/rhizobium spp. IC-94 and IC-2002 (ICRISAT India; Provided by Solange A very weak signal was observed for the cross-reactivity of TAL-1148 with IC-94 and Ca-34 T with LAIII 106 but the signal intensity was very low compared to the positive reaction hence false detection was unlikely. Ochrobactrum-like colonies were further confirmed by colony PCR using O. intermedium specific forward primer F4  (Romero et al., 1995) and Ca-34 T specific reverse primer OcA-34 5 ′ -GCCCCCCTTTAAAATTTCAG-3 ′ . This primer was designed against the previously reported 46 nucleotides insertion in 16S rRNA of Ca-34 T (Imran et al., 2010). This 46bp insertion is present at E. coli 16S rRNA position 187 that folds into a stem loop structure and reported to prolong the helix H184 when placed on the 16S rRNA gene sequence secondary structure. The nodules were crushed on Fast Prep instrument in 200 µL distilled water and nodule lysates were used for PCR-detection of Ca-34 T using F4 and OcA-34. Moreover, the strain identity was confirmed using random primer OPC-13 that generates a specific fingerprint for Ca-34 T . Preparation of PCR reaction and  amplification conditions were same as described (Imran et al., 2010).

Effect of Inoculation on Chickpea Nodulation
Nodulation exhibited no consistent pattern with plant genotype, inoculation or soil type. Nodules were mostly observed along entire primary/tap root in most of the genotypes forming crown (whorls) around the root and very few nodules were observed on secondary roots (Figure 1). Nodule number and nodule biomass showed differential genotype response toward inoculation in both soils (P < 0.05) and generally nodule number and biomass was higher in soil 2 ( Table 3). Kabuli genotypes produced higher nodules and nodule biomass in unplanted soil than desi genotypes. Nodulation in kabuli genotypes Pb-Noor-2009, PKV-2 and desi genotypes Pb-2008, CH23/00 was significantly increased either by single inoculation with Ca-34 T (up to 42%) or co-inoculated along-with TAL-1148 (up to 29%) in both soils. Inoculation with TAL-1148 alone however, had nonsignificant effect on nodulation of most of the kabuli and desi genotypes in any soil. Comparison of treatment means (averaged over genotypes and replicates) showed that in soil 1, nodulation in Ca-34 T -inoculated plants was maximum (55.85 nodules per plant) followed by its co-inoculation (50.93 nodules per plant) while in soil 2, nodulation was maximum in co-inoculated plants

Effect of Inoculation on Plant Biomass, Relative Effectiveness and Crop Yield
Genotype (G) response was highly significant (P < 0.05) at all harvests of 60, 80, and 120 DAS. Inoculation (T) response was significant (P < 0.05) at 60 and 80 DAS only. Kabuli genotypes produced higher shoot and root lengths while desi genotypes produced higher root and shoot weights irrespective of soil condition (see Supplementary Tables 1, 2 for detail).
In soil 1, the plants inoculated either with Ca-34 T or coinoculated along-with the nodulating strain produced taller plants, higher root and plant biomass, more primary and secondary branches, straw weight and grain yield as compared to inoculation with TAL-1148 alone. In soil 2 (fertile soil), inoculation with TAL-1148 showed better results as compared to other treatments. When treatment means were compared (averaged over genotypes), effect of inoculation on chickpea plant height, primary/secondary branches was non-significant in soil 2 while statistically significant in soil 1 showing maximum inoculation response in co-inoculated plants. Yield data of individual plant showed a significant genotype-dependent inoculation response (P < 0.05) in both soils. Desi genotypes Pb-2008, CH23/00, and CH21/02 produced more primary and secondary branches, straw weight, grain yield, and seeds per plant in both soils. Generally, co-inoculation of Ca-34 T +TAL-1148 significantly improved yield contributing parameters in marginal soil (soil type 1) whereas the response of inoculation in fertile soil (type 2) was, although higher than soil 1, non-significant as compare to uninoculated control. Inoculation with Ca-34 T resulted in 63 and 22% increase while its co-inoculation with TAL-1148 resulted in 54 and 15% increase in grain yield (Kg/Ha) in marginal and fertile soils, respectively over control. RE of inoculation (Table 4) was highest for desi variety B8/02 (321.69) inoculated with Ca-34 T +TAL-1148 followed by kabuli genotype CM2008 (301.43) inoculated with Ca-34 T . Over all, kabuli genotypes showed higher RE as compared to desi genotypes in both soils while RE of soil 1 was higher as compared to soil 2.
Overall, harvest index (HI) was genotype-dependent and generally higher in fertile soil (type 2) for both genotypes ( Table 4). Most of the genotypes inoculated with Ca-34 T showed increase in HI in both soils (up to 148%).
Effect of inoculation on seed size (100 grain weight) was genotype-dependent and maximum in fertile soil (type 2; Table 4). Seed size was highest in desi genotypes PKV-2, Pb-2008 and CH23/00 in plants inoculated with Ca-34 T or co-inoculated with TAL-1148. The desi genotypes 93127 and CH21/02 showed improved seed size with TAL-1148 inoculation as compared to other treatments.
Grain yield was found to be significantly affected by inoculation in both soils ( Comparison of treatment means and genotype means ( Table 5) showed that biomass was lower in soil 1 (4331-5161 kg/Ha) as compared to soil 2 (6363-6929 kg/Ha). Maximum biomass was obtained in plants with mix inoculation in both soils. Of genotypes, desi types produced more biomass as compared to kabuli types ( Table 5).

Relationship among Parameters
The whole data was subjected for correlation analysis using SPSS and a positive linear relationship was found (r = 0.26-0.856 * * ) between grain yield and measured plant growth and yield parameters. Kabuli genotypes specifically showed higher correlation coefficient ratio (r-values) but no significant trait correlation was observed in any genotype after bacterial inoculation.
Linear regression effectively modeled the positive relationship of grain yield with plant dry weight, nodule biomass and straw yield, accounting for 70-82% of total variance. A positive FIGURE 3 | Categorical Principal component analysis (CATPCA) of plant traits measured across desi and kabuli chickpea genotypes in two different soils (Total variance explaiened: 99% for soil 1, 95% for soil 2). CATPCA is a non-linear PCA. Factor loadings in PC1 and PC2 are presented as vectors using external scale. quadric regression (R 2 = 0.916; 0.908) was observed for nodule parameters with plant dry weight and grain yield and for harvest index with plant biomass and nodule number against genotype, inoculation and soil type accounting for more than 80% of the total variance. When modeled separately, a positive regression was observed (Figure 2) for different parameters in both kabuli and desi genotypes. CAT-PCA (Figure 3) and PCA (Figure 4) captured more than 70-95% of the variance and clearly demonstrated the key genotype difference in both soils. The effect of soil was more pronounced showing all the genotypes in marginal soil loaded on negative while in fertile soil loaded on positive quadrant (Figure 3). PCA showed that soil effect was pronounced and most of the chickpea growth parameters were strongly positively correlated to each other (r 2 < 0.99) and positively loaded on PC1 (Figure 4).

Detection of Inoculants
Survival of inoculated bacteria within rhizosphere and nodule of chickpea and their persistence throughout the crop growth was tested. Based on plate-assay, it was observed that of total bacterial population attached to chickpea roots, 1.5% was similar to Ochrobactrum. In nodules, its population ranged from 1 × 10 3 to 8 × 10 5 per gram of nodule fresh weight (Figure 5) constituting about 0.04-5% of total  cultureable nodule population. These results were confirmed by specific PCR (Figure 6A) immunoblotting (Figure 6B), and randomly amplified polymorphic DNA (RAPD) analysis with primer OPC-13. Although, different genotypes showed variable response in both soils but kabuli genotypes showed more colonization in nodules as compared to desi genotypes. The presence of TAL-1148 was confirmed only in nodules and the population ranged from 5 × 10 4 to 8 × 10 7 in different genotypes.

Discussion
Pakistan ranks the third in terms of chickpea production globally, where, the Thal desert that cannot support/sustain major cash crops due to low fertility and lack of artificial irrigation is well known as home of chickpea. The cultivation in nutrient-deficient soils coupled with inadequate or no crop management (fertilizer, moisture application) results in low yield of crop (Kantar et al., 2007). High yield and disease resistance must be an essential traits for breeding (Micke, 1993;Richards, 2000) but needs to be combined with traits that improve crop economics such as increased N 2 fixation capacity (Herridge and Rose, 2000). The chickpea breeding program is also hampered by the low fertility status of soil on which the crop is usually grown. Rhizobial inoculation to legumes especially chickpea exert positive effects on growth and yield (Rodríguez-Navarro et al., 2000;Kyei-Boahen et al., 2005) in soils containing in-effective rhizobia (Sharma et al., 1983;Beck et al., 1991). The success of inoculation, however, depends on the environmental conditions, soil richness (Bottomley, 1992;Graham, 1992), number and application method of effective rhizobial cells , presence of high population of competing strains of rhizobia (Thies et al., 1991) and plant genotype (Hafeez et al., 1998).This study has evaluated and demonstrated the likely contribution of PGPR inoculation toward growth promotion of kabuli and desi chickpea under different soil conditions. Both desi and kabuli genotypes are botanically similar but genotypically different. They represent a wide range of variation in seed size and origin (parent source). Furthermore, the soil where these genotypes were evaluated, contrasts in nutrient and fertility status. One was an unplanted, marginal soil where no crop was grown for 8 years (soil type 1). Background total bacterial population of this soil was 3 × 10 5 .The other was fertile, regular legume growing soil where legumes (mungbean, chickpea) and castor were routinely grown (soil type 2) having bacterial population 7.5 × 10 9 . The experiments were conducted with a very little addition of external fertilizer (added before the experiment) and with zero irrigation. As chickpea can perform well under conditions of moisture stress in marginal soils, hence, drought tolerance in this crop is extremely desirable attribute for moisture deficient areas of the country. The inoculation response was more significant in soil 1 having poor indigenous rhizobial population and fertility (can be categorized as marginal land) which shows that soil fertility, nutrient status, and indigenous population of bacteria have a vital role in the development of plant microbe interaction. The pronounced effect of soil was very evident in PCA analysis where both soil types loaded differently on PC1 and PC2. The evidences suggest that this interaction is mainly controlled by indigenous microbial population, soil richness and by plant genotype hence, their co-selection under a given set of soil and environmental conditions may enhance the amount of fixed N. The inoculation effect was different in both soils which show that soil fertility, nutrient status and indigenous population of bacteria have a vital role in the development of plant microbe interaction. In fertile soil, the average genotype yield in un-inoculated control plants was higher than the reported yield ( Table 1). This high yield might be due to the Meso/brady/Rhizobium-rich soil and excellent fertility status of the soil. It is believed that general/universal inoculum for all systems is impossible to develop as effectiveness depends upon plant type, soil type, weather conditions and many unidentified factors . The data showed that soil type 1 was better for growing kabuli genotypes while soil type 2 was better for cultivation of both genotypes.
Inoculation of nodulating reference strain TAL-1148 alone showed non-significant nodulation response in both soils. High background rhizobial population usually hampers the growth, survival and colonization of inoculating rhizobial strains in the soil. Where indigenous population offers more competition, the effect of inoculation usually is non-significant. Desi genotype Pb-2008 showed increased nodulation in plants co-inoculated with Ca-34 T + TAL-1148. It is already established that inoculation with nitrogen-fixing/nodulating strain combined with PGPR exert more beneficial effect on plant as compared to single inoculation (Hameed et al., 2004). The synergistic effect of co-inoculating strains is well established in chickpea (Mirza et al., 2007;. Increased biomass and grain yield in Ca-34 T + TAL-1148 co-inoculated plants may be attributed to the nitrogen fixation potential of TAL-1148. Overall Ca-34 T -inoculation significantly increased nodule number, plant height, root proliferation and biomass, primary and secondary branching, plant biomass, straw weight and grain yield in single as well as multi-strain combination treatment. Most responsive genotypes toward Ca-34 T -inoculation were all four kabuli genotypes and two desi genotypes B8/02 and Pb-2008. These genotypes showed maximum increase in nodulation, biomass and grain yield by Ca-34 T -inoculation. Varietal difference for harvest index has been reported in chickpea, mungbean (Singh et al., 1980;Malik et al., 1981) and rice (Fida et al., 1993). The importance of changes in dry weight partitioning between organs have focused attention of scientists on harvest index as a specific selection criterion for plant breeders as the productivity of grain crops depends not only on dry matter accumulation (Kumar et al., 2006), but also on its effective partitioning to economically important plant parts. Improved harvest index has been responsible for the grain yield potential increase among major cereal species (Frey, 1981).
The root proliferation coupled with increased nutrient uptake and yield of plant might be the result of the production of IAA by O. ciceri strain Ca-34 T . IAA is a plant growth hormone that enhances the lateral root development in plants to facilitate uptake of more nutrients and water, consequently improving growth and yield (Barazani and Friedman, 1999;Shahid et al., 2012;Ali et al., 2014). Apart from IAA, Ochrobactrum spp. strains are involved in nutrient mobilization (phosphate and zinc), AHLproduction , production of siderophores (Chakraborty et al., 2009), antibiotic 2, 4-DAPG (Hassan et al., 2010) that directly or indirectly promote plant growth. Positive effects of inoculation with Ochrobactrum spp. have been reported in maize (Príncipe et al., 2007), mungbean (Faisal and Hasnain, 2006), wild Coffea Arabica L. (Muleta et al., 2008) under normal soils, and in chickpea in chromium contaminated soils (Riaz et al., 2010).
Ochrobactrum species, usually considered as free-living, are adapted to a wide range of habitats including soil, rhizosphere (Lebuhn et al., 2000), cotton root interior (McInroy and Kloepper, 1994), wheat roots (Sato and Jiang, 1996), deep-water rice endophyte (Verma et al., 2004;Tripathi et al., 2006) and nodules of Acacia mangium (Ngom et al., 2004), Lupinus albus (Trujillo et al., 2005), Cytisus scoparius (Zurdo-Pineiro et al., 2007) and chickpea (Imran et al., 2010). Although omnipresent, Ochrobactrum species are considered as weak colonizers of plants and less competent in the rhizosphere. Persistence of the both inoculating strains in the rhizosphere and nodules of both kabuli and desi genotypes, specifically O. ciceri Ca-34 T within the nodules, shows that they are competitive and can maintain substantial population level when inoculated to the seeds. As the population was obtained on antibiotic plates which were later confirmed by immunoblotting, specific PCR and RAPD, so the possibility of getting contaminants were ruled out. Our findings are in line with the earlier findings (Lebuhn et al., 2000 reporting that Ochrobactrum species constitute 2% of total rhizosphere and 0.3% of rhizoplane population.
As described in this study, several others reported a strong influence of plant variety on N 2 fixation. To enhance the amount of fixed N 2 under a given set of environmental conditions, the co-selection of all three main contributors, i.e., plant genotype, microbe and soil is indispensable. The inherent genetic variability of genotypes and high indigenous soil population is known to contribute significantly in nodulation and yield sometimes out-competing the inoculants strains. The chickpea plants exhibited substantial interaction between genotype/varieties, bacterial strains and soil type. Genotype and soil were found to be the most important factors contributing toward the plant response to bacterial inoculation.

Conclusions
Being a highly nutritive cheap food, chickpea occupies a special position among legumes but mostly grown on the marginal lands by resource-poor farmers. The field analysis of PGPR inoculation to some of the elite kabuli and desi cultivars developed in the country shows that both genotypes differentially respond to inoculation under nutrient-poor and nutrient-rich soils. The data suggests co-selecting the chickpea genotypes/cultivars along-with the highly efficient rhizobial and PGPR strains for improvement of BNF efficiency and yield in chickpea. This study has demonstrated the plant beneficial potential of O. ciceri strain Ca-34 T as single or multi-strain inoculum both in marginal and fertile soil. This quality makes it an excellent candidate for development of inoculum in combination with an efficient chickpea nodulating strain to get sustainable production of both genotypes of chickpea under nutrient-poor/rich soils.