Impact of Azospirillum sp. B510 Inoculation on Rice-Associated Bacterial Communities in a Paddy Field

Rice seedlings were inoculated with Azospirillum sp. B510 and transplanted into a paddy field. Growth in terms of tiller numbers and shoot length was significantly increased by inoculation. Principal-coordinates analysis of rice bacterial communities using the 16S rRNA gene showed no overall change from B510 inoculation. However, the abundance of Veillonellaceae and Aurantimonas significantly increased in the base and shoots, respectively, of B510-inoculated plants. The abundance of Azospirillum did not differ between B510-inoculated and uninoculated plants (0.02–0.50%). These results indicate that the application of Azospirillum sp. B510 not only enhanced rice growth, but also affected minor rice-associated bacteria.

The genus Azospirillum, which includes nitrogen-fixing Gram-negative Alphaproteobacteria, has been known for many years to be composed of plant-growth-promoting bacteria. Among the Alphaproteobacteria, Azospirillum is one of the most studied genera (8). This genus was isolated not only from the rice phyllosphere (7,17), but also from the rhizosphere of other plants (21). Rice growth was enhanced by inoculation with Azospirillum (15).
Azospirillum sp. strain B510 was isolated from surfacesterilized stems of field-grown rice (7). Rice tiller numbers and seed yield were significantly increased with B510 inoculation in paddy fields (13). Moreover, Yasuda et al. (26) showed that rice plants inoculated with B510 had greater resistance against two diseases caused by the virulent rice blast fungus Magnaporthe oryzae and by the virulent bacterial pathogen Xanthomonas oryzae. Thus, Azospirillum sp. B510 is a promising bacterial inoculant for plant growth promotion and agricultural practices (13).
The plant-growth-promoting effects of Azospirillum brasilense Sp245 were due to the production of indole-3acetic acid (24). Azospirillum cells inoculated into rice plants often spread from below-ground roots into above-ground tissues (4). However, Azosprillium sp. B510 inoculation gave rise to colonization in basal parts of rice plants in a laboratory experiment (26). Little is known about the effect of Azospirillum sp. B510 inoculation on rice-associated bacterial communities in field conditions. The objective of the present study was to determine whether Azospirillum sp. B510 inoculation changes rice-associated bacterial communities, including the B510 inoculant, when the effects of inoculation first appear in a paddy field.
Three sets of composite samples including at least three rice plants were independently prepared for the bases or shoots (approximately 100 g each) in each treatment, and then stored at −80°C until used. The composite samples of bases or shoots were homogenized without surface sterilization to prepare shoot-and base-associated bacterial cells containing both endophytes and epiphytes. DNA was extracted by the bacterial enrichment method (11). Bacterial communities were analyzed by triplicate 454 pyrosequence analyses targeting the V2-V3 region of the 16S rRNA gene using 454 barcode tags (20). Raw pyrosequencing reads were processed using the Quantitative Insights Into Microbial Ecology (QIIME) software package (2). The reads were assigned to each sample according to sample-specific barcodes. Low-quality reads with a length shorter than 300 bp, with an average quality score lower than 25, with mismatching primer sequences, or with ambiguous bases (denoted by N), were eliminated from downstream analyses. The forward and reverse primer sequences were removed from the qualityfiltered reads. The remaining sequences were clustered into operational taxonomic units (OTUs) at 97% similarity. Taxonomic assignment of representative OTUs was performed using the RDP Classifier (25) with default parameters. Principal-coordinates analysis (PCoA) was performed using weighted UniFrac distances (14). The project number for 454 pyrosequence reads in the National Center for Biotechnology Information (NCBI) database is ID DRA000965.
For bacterial communities, the total number of pyrosequence reads of 16S rRNA genes from base and shoot, with and without B510 inoculation, ranged from 16,156 to 33,906 (Table S1). Bacterial community structures were analyzed by PCoA (Fig. 1). PC1 (79.3%) showed that bacterial communities were separately clustered in the shoot and base. PC2 accounted for only 6.9% of the differences in microbial communities. There were no marked shifts in bacterial communities by B510 inoculation in either the shoot or base of the rice plants (Fig. 1).
The compositions of entire bacterial communities are shown at the class level (Table S2) and with more detailed taxa (Order, Family and Genus) ( Table 1). At the class level, inoculation significantly increased the relative (percent) abundance of the class Negativicutes (phylum Asterisks indicate a significant difference (P<0.05; t-test) between uninoculated controls (−) and inoculated plants (+). Data for bacterial taxa with low relative abundance (<1%) are not shown. Grey areas highlight the taxa with relative abundance, significantly different between inoculated and uninoculated rice plants. Firmicutes) in the rice-plant base (P<0.05; t-test), although Alphaproteobacteria (>30% of total sequences) were the most abundant in both shoot and base communities (Table 1). In the Alphaproteobacteria, Rhizobium (relative abundance, 9.5-20.7%) and Methylobacterium (15.2-22.0%) were dominant bacterial groups in the rice tissue. However, their relative abundances were not significantly influenced by inoculation. In contrast, relatively minor bacterial groups were significantly influenced by B510 inoculation: the relative abundance of Aurantimonas increased significantly in the shoot (from 4.8% in controls to 6.3% in inoculated plants) and Methylocystis decreased in the base (from 5.0% to 2.9%) with B510 inoculation. Our result was partially consistent with the previous observation that application of Azosprillium brasilense strains as inoculants did not influence the dominant members of the endophytic microbial communities in the rice phyllosphere using PCR-DGGE (22). DGGE profiles of the bacterial community might limit our understanding because its resolution is relatively low. However, in this study, minor bacterial groups, such as Auriontimonas (4.8% of relative abundance), Methylocystis (5%) and members of Veillonellaceae (>1%), were significantly influenced by inoculation. Aurantimonas has been found in above-ground parts of plants such as rice, soybean, Arabidopsis thaliana, and Jatropha curcas (1,5,16,18). Their population was likely controlled in a manner similar to that of legume-rhizobia symbiosis (10). These findings suggest that Aurantimonas spp. are ubiquitous in the phyllosphere and have close interactions with plants. Veillonellaceae is a single family of anaerobic Gram-negative bacteria within Firmicutes, unlike most other members of the Firmicutes (19). Members of Veillonellaceae have been detected from the human gut and from rice paddy soil (6,9). Hooda et al. (9) observed that consumption of soluble corn fiber led to increases in Veillonellaceae in the gut of healthy men. Thus, it is likely that the members of Aurantimonas and Veillonellaceae are associated with higher organisms, although their biology is largely unknown. Thus, the limited knowledge of bacterial functions makes it difficult to discuss the relationships between the bacterial community and rice growth promotion, although it is possible that growth enhancement was induced by minor bacterial changes.
In the present study, the B510 inoculant and natural Azospirillum populations were not differentiated because of the limitation of the pyrosequence read length (approximately 400-500 bp). However, it is possible to estimate the fate of the inoculant during rice growth in the field by observing the fluctuation of Azospirillum abundance. The relative abundance of Azospirillum was not significantly influenced by B510 inoculation, although it was low (0.02-0.5%) in both the base and shoot (Fig. 2), suggesting that the abundance of the B510 inoculant decreased at least to that of natural Azospirillum populations at 51 DAT (Fig. S1). Apart from the inoculation effect, the relative abundance of Azospirillum was higher in the base (0.26-0.29%) than in the shoot (0.02-0.04%) (Fig. 2). The initial bacterial community was not determined at 0 DAT in this field experiment. However, 5.9×10 4 CFU per plant of Azospirillum sp. B510 (0.6% of total inoculant of 10 6 ) was detected from surface-sterilized plant tissue in the present study, suggesting that Azospirillum sp. B510 cells colonized in rice tissue as endophytes just after 0 DAT.
In conclusion, the present study revealed that B510 inoculation significantly influenced minor bacterial groups of the shoot and base of rice plants, but not major bacterial groups. These results indicate that Azospillium sp. B510 inoculation not only stimulated rice growth, but also affected relatively minor rice-associated bacteria in the rice base and shoot.