Differential microbial diversity in blueberry rhizosphere microbial community
To profile the taxonomical compositions of rhizosphere microbial communities of blueberry and compare the taxonomical differences for these three blueberry varieties and bulk soil samples, we sequenced the V3–V4 region of 16S rDNA of bacteria and archaea from rhizosphere samples. In total, 997,713 high-quality 16S rRNA amplicons for 20 rhizosphere samples were obtained and analyzed. The number of sequences for these samples ranged from 31,591 to 73,918, with an average value 49,886 (Supplementary Table 1). After rarefying the final OTU table to 18,652 reads, we detected 6,280 OTUs for these rhizosphere soil samples, and the number of OTUs for blueberry rhizosphere microbial communities and bulk samples ranged from 1,495 OTUs to 2,548 OTUs (Supplementary Table 1).
The alpha diversities of rhizosphere microbial communities were compared between three blueberry varieties and bulk soil samples using the number of OTUs, Shannon index, and Simpson index (Figs. 1a–1c). We observed that the number of OTUs of microbial communities in bulk soil samples was significantly higher than that in three blueberry varieties, and the number of OTUs of rhizosphere microbial communities among three blueberry varieties was also significantly different (Kruskal–Wallis test, p < 0.05; Fig. 1a). As for species richness of rhizosphere microbial communities, we found that the Shannon and Simpson indexes of rhizosphere microbial communities of bulk soil samples were significantly higher than those of blueberry varieties, except Southern Highbush Blueberry (Figs. 1b–1c).
The similarities of rhizosphere microbial communities were also assessed among three blueberry varieties and bulk soil samples based on Bray–Curtis (Fig. 1d) and unweighted UniFrac distance metrics (Fig. 1e). The results of PCoA based on Bray–Curtis (Fig. 1d) and unweighted UniFrac distance metrics (Fig. 1e) revealed significant differences in taxonomical compositions between microbial communities of blueberry rhizosphere soil and bulk soil (p < 0.001, F = 6.815, One-way PERMANOVA, N = 9,999, Bray–Curtis dissimilarity index). The taxonomical compositions of rhizosphere microbial communities of three blueberry varieties also significantly differed (p < 0.001, F = 7.472, One-way PERMANOVA, N = 9,999, Bray–Curtis dissimilarity index).
Differential taxonomical composition in blueberry rhizosphere microbial community
To gain insights into the taxonomical compositions of blueberry rhizosphere microbial communities, we stratified the taxonomical structure of rhizosphere microbial communities at the phylum, order, and genus levels (Fig. 2). We compared the differences in taxonomical compositions between rhizosphere microbial communities of blueberry and bulk soil and among different blueberry varieties.
At the phylum level, we found that Proteobacteria, Actinobacteria, Acidobacteria, Firmicutes, Planctomycetes, and Verrucomicrobia constituted the six most enriched bacterial phyla among rhizosphere microbial community of three blueberry varieties and bulk soil (Fig. 2a). The predominant phylum is almost consistent with a previous study, which also reported that Proteobacteria, Actinobacteria, Acidobacteria, Bacteroidetes, Planctomycetes, Chloroflexi, and Verrucomicrobia were enriched in the rhizosphere microbiome of blueberry [19]. The proportion of Proteobacteria of each blueberry variety (Rabbiteye Blueberry: 40.81%±0.87%, Northern Highbush Blueberry: 36.79%±6.2%, Southern Highbush Blueberry: 36.2%±2.07%) was not different from that of bulk soil (39.42%±6.31%, t-test, all p > 0.05). The relative abundances of Actinobacteria of rhizosphere microbial communities of Rabbiteye Blueberry (24.72%±4.91%) and Northern Highbush Blueberry (22.93%±5.49%) varieties were significantly higher than those of bulk soil (14.57%±2.72%, t-test, p < 0.05). Although the relative abundance of Firmicutes increased in rhizosphere microbial communities of three blueberry varieties compared with bulk soil, the proportions in Northern Highbush Blueberry (6.24%±1.8%) and Southern Highbush Blueberry (6.02%±1.13%) were significantly different from that in bulk soil (2.97%±1.72%, t-test, p < 0.05). The relative abundances of Nitrospirae were significantly decreased in rhizosphere microbial communities of Rabbiteye Blueberry (0.26%±0.08%), Northern Highbush Blueberry (0.16%±0.03%), and Southern Highbush Blueberry varieties (0.25%±0.11%) compared with bulk soil (0.7%±0.29%, t-test, p < 0.05). Additionally, LDA was conducted to maximize the separation of rhizosphere microbial communities of three blueberry varieties and bulk soil based on the relative abundances of predominant phyla. We observed that rhizosphere microbial communities of three blueberry varieties and bulk soil could be distinctly differentiated by integrating a linear combination of phyla (Fig. 2b). Among the linear combination of phyla, we found that Planctomycetes, Gemmatinonadetes, Chloroflexi, and Verrucomicrobia were important for differentiating rhizosphere microbial communities of three blueberry varieties and bulk soil (Fig. 2b).
At the order level, we observed that Acidobacteriales, Actinomycetales, Xanthomonadales, Rhodospirillales, Rhizobiales, and Gaiellales were the six predominant bacterial orders in rhizosphere microbial communities of three blueberry varieties and bulk soil (Fig. 2c, Supplementary Figure 2a). Specifically, we found that the average relative abundances of Actinomycetales in rhizosphere microbial communities of Rabbiteye Blueberry (15.2%±3.37%) and Northern Highbush Blueberry (12.31%±4.41%) were increased compared with those of bulk soil (7.22%±2.77%) and Southern Highbush Blueberry (6.29%±3.64%). The average relative abundance of Xanthomonadales in rhizosphere microbial communities of Rabbiteye Blueberry (15.19%±2.71%) was significantly higher than those of bulk soil (5.81%±2.99%, t-test, p < 0.01), Northern Highbush Blueberry (9.13%±2.29%, t-test, p < 0.01), and Southern Highbush Blueberry (9.81%±0.59%, t-test, p < 0.05). We also profiled the taxonomical composition of rhizosphere microbial communities of blueberry varieties and bulk soil at the genus level, and we found that the specific distribution of genus contributed to the discrepancy of rhizosphere microbial communities (Fig. 2d, Supplementary Figure 2b).
Differential functional and phenotypic compositions in blueberry rhizosphere microbial community
The functional and phenotypic compositions in blueberry’s rhizosphere microbial community were profiled based on their taxonomical compositions (Fig. 3). As to the functional traits that collapsed to level 2 of the KEGG database, we found that the enrichment of enzyme families and environmental adaptation in rhizosphere microbial communities and the proportion of biosynthesis of other secondary metabolites was higher in Northern Highbush Blueberry (Supplementary Figure 3). The relative abundances of functional traits related to transporters, general function, ABC transporters, DNA repair and recombination proteins, two-component system, and urine metabolism were higher in the rhizosphere microbial community of blueberry varieties and bulk soil (Fig 3a). Moreover, we found that the functional compositions of the rhizosphere microbial communities of Rabbiteye Blueberry significantly differed from those of bulk soil (p < 0.05, F = 3.545 One–way PERMANOVA, N = 9,999, Bray–Curtis dissimilarity index) and Southern Highbush Blueberry (p < 0.05, F = 3.3, One–way PERMANOVA, N = 9,999, Bray–Curtis dissimilarity index). The rhizosphere microbial communities of three blueberry varieties and bulk soil could be distinctly distinguished by integrating a linear combination of functional components (Fig. 3b).
Additionally, we explored the phenotypic compositions of rhizosphere microbial communities between three blueberry varieties and bulk soil. We observed that the proportions of anaerobic microbiota, mobile elements, and stress tolerant significantly differed (Kruskal–Wallis test, p < 0.05, Fig. 3c). Specifically, the proportions of anaerobic microbiota of bulk soil (4.78%±1.63%) were higher than those of Rabbiteye Blueberry (2.14%±0.62%) and Northern Highbush Blueberry (3.08%±0.23%), except Southern Highbush Blueberry (4.06%±1.04%, Fig. 3c). The relative abundances of mobile elements in the rhizosphere microbial communities of three blueberry varieties (Rabbiteye Blueberry: 40.18%±4.51%, Northern Highbush Blueberry: 32.61%±4.25%, and Southern Highbush Blueberry: 25.95%±3.78%) were higher than those of bulk soil (23.67%±3.99%, Fig. 3c). The proportions of stress tolerant of rhizosphere microbial communities of Rabbiteye Blueberry (82.56%±3.59%) and Southern Highbush Blueberry (77.05%±2.46%), except Northern Highbush Blueberry (72.39%±4.53%), were higher than those of bulk soil (74.64%±4.38%, Fig. 3c).
Core blueberry rhizosphere microbiome
We extended our analysis to determine which OTUs are stable across in rhizosphere microbial communities of different blueberry varieties and bulk soil. We identified 728, 634, 777, and 712 OTUs as the core OTUs in rhizosphere microbial communities of Rabbiteye Blueberry, Northern Highbush Blueberry, Southern Highbush Blueberry and bulk soil (Fig. 4a), respectively. Eventually, 201 OTUs of 1,420 OTUs (14.2%) were identified as the core OTUs in rhizosphere microbial communities of blueberry varieties and bulk soil (Fig. 4a, Supplementary Table 2). Many OTU cases are mainly affiliated with Proteobacteria (78 OTUs), Actinobacteria (41 OTUs), Acidobacteria (34 OTUs), Firmicutes (16 OTUs), Chloroflexi (9 OTUs), and Planctomycetes (8 OTUs, Fig. 4b). The distribution of each core OTU in rhizosphere microbial communities of blueberry varieties was different (Fig. 4b), indicating that the relative abundance of core OTUs varied most among different blueberry varieties.
Identification of microbial biomarkers for classifying different blueberry varieties
To explore the taxonomical signatures among rhizosphere microbial communities of three blueberry varieties and bulk soil, we conducted LEfSe analysis to identify biomarkers for each blueberry variety based on the taxonomical compositions of rhizosphere microbial communities. Finally, we obtained 28 discriminative biomarkers with logarithmic LDA score > 3.5 (Fig. 5). At the phylum level, we found that Actinobacteria and Planctomycetes were identified as the biomarkers for Rabbiteye Blueberry and Southern Highbush Blueberry, respectively, whereas Verrucomincrobia and Chloroflexi were detected as the biomarkers for bulk soil (Fig. 5a). At the order level, we observed Clostridiales, Rhodospirillales, Rhizobiales, Gaiellales, Actinomycetals, Xanthomonadales, and Burkholderiales were identified as the biomarkers for three blueberry varieties (Fig. 5).
Patterns of co-occurrence network in blueberry rhizosphere microbial community
To gain more insights into the interactions among the microbial members of rhizosphere microbial communities of blueberry varieties, we extended our analysis to explore the patterns of OTUs co-occurrence network from an ecological perspective. The SparCC algorithm was applied to calculate the correlations between OTUs and the significant strong correlations (the value of absolute correlations > 0.8 and the p-value < 0.05) were chosen to construct the co-occurrence network. The co-occurrence network comprised of 198 nodes and 484 edges (Fig. 6). The density and average degree of the co-occurrence network were 0.025 and 4.89, respectively. The clustering coefficient of the co-occurrence network was 0.35 and the co-occurrence network could be clustered into seven clusters. Strong interactions existed between OTUs in the co-occurrence network. The members of co-occurrence network were mainly affiliated with Proteobacteria, Actinobacteria, Acidobacteria, Verrucomicrobia, and Firmicutes (Fig. 6). Among the 198 nodes, 74 nodes (37.4%) belonged to core OTUs and these OTUs were mainly affiliated with Proteobacteria, Actinobacteria, and Acidobacteria (Fig. 6). The OTUs with the highest average proportions of the co-occurrence network were members of core OTUs of rhizosphere microbial communities of blueberry varieties, which were affiliated with Xanthomonadaceae, Koribacteraceae, Gaiekkaceae, and Sinobacteraceae (Fig. 6).