Large Rhizosphere Bacterial Diversity Exits Among Wild Progenitor Species of Modern Sugarcane (Saccharum Spp. Inter-Specic Hybrids)

Rhizosphere is rich in highly diverse and complex microbial communities. Plant growth promoting rhizpbacteria and diazotrops are played crucial role in plant growth and development. In this study, rhizosphere soils were collected from ve wild Saccharum species- S. ocinarum L. cv Badila (BRS), S. barberi Jesw. cv Pansahi (PRS), S. robustum (RRS), S. spontaneum (SRS), and S. sinense Roxb. cv Uba (URS) for studied of rhizosphere and diazotroph bacterial diversity using 16S rRNA and nifH gene amplication and sequencing. We from species Out of the bacterial communities among we found a core microbiome of 31 rhizobacterial families spread across all the species analyzed. A total of 1099 OTUs were identied for diazotrophs with a core microbiome of 9 families distributed among all the sugarcane species. The core microbiomes were distributed across twenty genera-Bradyrhizobium, Dechloromonas, Desulfovibrio, Stenotrophomonas, Xanthobacter, Anaeromyxobacter, Azospirillum, Pseudoacidovorax, Methylobacterium, Azoarcus, Paenibacillus, Ideonella, Beijerinckia, Paraburkholderia, Burkholderia, Rucoccus, Geobacter, Sinorhizobium, Kosakonia, and Azotobacter. soil and Our present study aims to understand the role of rhizosphere bacterial communities and identify new species of nitrogen xing bacteria using high-throughput 16S and nifH gene sequencing by the Illumina platform. The current study reports interesting novel provides to study

molecular biology, especially the advent of high-throughput DNA sequencing technologies and the associated data analytics has helped understand the rhizosphere micro ora by culture-independent studies rapidly and at relatively low cost [10,9]. The next-generation sequencing approaches provide an e cient and comprehensive approach to identify microbial species in the rhizosphere irrespective of microbial abundance [11]. As a result, through the sequencing of 16S rRNA gene, the taxonomic characterization of highly diversi ed rhizospheric bacteria has increased remarkably [12,13]. Further, modern molecular techniques permit an in-depth analysis of soil bacterial communities' compositional and functional dynamics in changing soil environmental conditions, a recurring feature of agricultural soil [9,14].
Sugarcane is an important agricultural crop grown in nearly 110 countries world wide. China is the third largest sugarcane producer (a collective term for Saccharum species, but more commonly applied to Saccharum o cinarum L. and Saccharum spp. inter-speci c hybrids) and it is a major crop in southern China, accounting for ~90% of Chinese sugar production [15]. Over the years, sugarcane has been developed as a multi-purpose agro-industrial crop as it provides the raw material for different industries like food, thermal, energy/fuel, paper etc. [16,17]. Sugarcane is mostly grown as a monoculture for extended periods resulting in yield decline which is attributed to degraded soil, imbalanced soil biology, and build-up of pests and diseases. Restoration of soil biology and soil fertility is now emerging as a priority for improving soil health, reducing yield gap and sustaining pro table green agriculture. Hence, signi cant research is now being carried out on sugarcane rhizobacteria to understand their diversity and role in crop improvement. Several novel PGPRs from the sugarcane microbiome have been identi ed and used to improve crop production [18,19]. They are mostly involved in nitrogen xation and plant hormone production, thus positively affecting sugarcane growth [16,20,21]. Rhizobacteria play a signi cant role in nitrogen xation in sugarcane crops [22,23]. However, much remains to be learned about these diazotrophic rhizobacteria, a key driver of soil health and fertility. Our present study aims to understand the role of rhizosphere bacterial communities and identify new species of nitrogen xing bacteria using high-throughput 16S rRNA and nifH gene sequencing by the Illumina platform. The current study reports interesting and novel ndings on the diversity of bacterial communities in ve Saccharum species namely, S. o cinarum L. cv Badila (BRS), S. barberi Jesw. cv pansahi (PRS), S. robustum (RRS), S. spontaneum (SRS), and S. sinense Roxb. cv Uba (URS), and provides a knowledge base to study the in uence of sugarcane genotype on rhizosphere bacteria in this necessary sugar and energy crop.

Results
This study analyzed the rhizosphere soil microbiota of ve different critical ancestral sugarcane species using 16S rRNA and nifH gene sequencing to understand their bacterial community diversity, especially that ofdiazotrophs.
Data ltration, quality evaluation, and sequence optimization A total of 233,455 effective sequences number with an average length of 413 bp were obtained by sequencing 16S rRNA from different sugarcane species samples. nifH gene sequencing resulted in a total of 182,185 effective sequences with 357 bp average length from all the samples. Filtration of raw reads of 16S rRNA and nifH genes was done using QIIME quality lters, followed by OTU identi cation, clustering, and analysis, respectively (Additional le 2: Table S1 and S2).
Identi cation and analysis of Operational Taxonomic Units (OTUs)

Rarefaction curves
To assess microbial diversity among all the sugarcane samples, a Rarefaction curve was drawn. The presence of OTUs across all the 16S samples is shown in Additional le 1: Figure S1A. The highest number of OTUs was observed in S. sinense followed by S. robustum, S. barberi, S. o cinarum and S. spontaneum. Presence of the OTUs identi ed in nifH sequence data across all the samples is shown in Additional le 1: Figure S1B. The highest number of OTUs was observed in S. barberi and S. sinense followed by S. robustum and S. spontaneum. The clustering of OTUs was done based on 97% sequence similarity.

Rank abundance curves (RACs)
Comparison of rank abundance based on OTU ranks derived from 16S rRNA and nifH sequences (Additional le 1: Figure S2 A and B) was performed to visualize the relative species abundance across all the samples. RACs depict the species richness and species evenness, which help identify common and rare rhizobacterial species in the sugarcane species studied. Figure S2A, revealed that all the sugarcane samples showed high abundance at least OTU ranks i.e. 1, following this at 500 OTU rank equal species richness and evenness was displayed. At the highest OTU rank, i.e. 1000, all the samples showed least species abundance and high species evenness. S. sinense samples displayed the highest species richness among all the samples at highest OTU rank. But in (Additional le 1) Figure S2B, S. barberi samples displayed the highest species richness among all the samples at highest OTU rank. The S. spontaneum sample, compared to all the other samples, showed less species evenness at low OTU rank, i.e. approximately 180.

Venn diagrams
We analyzed common and unique OTUs based on 16S rRNA and nifH gene sequences for each sample in the Venn diagram ( Figure 1A & B). In the 16S rRNA sequence data, 6202 OTUs were identi ed collectively, of which 519 OTUs were common across all samples. According to the Venn diagram, the relative frequency of OTUs in the studied species is as follows: S. sinense>S.robustum>S.barberi>S. o cinarum>S. spontaneum. The highest number of OTUs was recorded in S. sinense. A total of 1099 OTUs were identi ed in all the samples of nifH gene sequence data. Among them, 14 were common OTUs found across all the samples. According to the Venn diagram, occurrence of OTUs in the plant samples was as follows: S.barberi>S. robustum>S. o cinarum>S. sinense>S. spontaneum. The highest numbers of OTUs were found in S. barberi and, S. spontaneum had the lowest number of OTUs for both 16S rRNA and nifH gene sequence data.

Principal component analysis (PCA)
To understand the rhizobacterial community composition, PCA plots were generated based on 16S rRNA and nifH gene samples OTUs data ( Figure S3 A and B). The OTUs present in the 16S samples shows that S. o cinarum and S. robustum are not identical, S. sinense and S. robustum are more similar, and S. spontaneum is distinct from all the other samples ( Figure S3 A). PCA of nifH gene samples showed a close identity between S. barberi and S. spontaneum whereas S. sinense, S. o cinarum and S. robustum remained distinct to one another.

Alpha diversity
Alpha diversity refers to the diversity within a particular sample individually, and it is usually represented by the species (i.e. species richness) enumerated in the test samples. Alpha diversity analysis was done using Shannon, Simpson and Chao Indices Rarefaction curves for both 16S and nifH sequence data. Additional le 1: Figure S4 A consists of two plots displaying Shanon, Simpson, and Chao Indices, built using 16S samples. The Shannon Index increased as both the species richness and the evenness in the community increase. Among the 16S rRNA data of all the samples analyzed, Shanon and Chao Index of S. sinense sample was the highest as compared to the other four samples with little variation among themselves. Whereas, Simpson Index increased as diversity increased. In the Simpson Index plot, S. o cinarum sample showed the highest value implying that it has more species diversity compared to the other four samples. Shannon and Chao Index of nifH gene data of all the samples showed more species diversity in S. barberi sample than others. According to the Simpson Index, of S. o cinarum sample was the highest compared to other (Additional le 1: Figure S4 B).

Community composition analysis
The relative abundance of the microbial communalities differed among ve sugarcane species analyzed.The abundant phyla identi ed in 16S samples in all sugarcane species are Proteobacteria, Firmicutes, Actinobacteria, Acidobacter, Bacteroidetes, Chloro exi, Gemmatimonadetes, Planctomycetes and Nitrospirae (Figure 2A). Firmicutes were the highest phyla present in S. o cinarum compare to other samples. Gemmatimonadetes was the highest in S. barberi, Acidobacter was highest in S. spontaneum, and Bacteroidetes was highest in S. sinense. Based on nifH gene samples, we observed the occurrence of Proteobacteria and Verrucomicrobia ( Figure 2B). Firmicutes were abundant in S. spontaneum samples compare to other samples. Many unclassi ed phyla were also represented abundantly in S. spontaneum followed by S. sinense and S. barberi samples.
Genus distribution using nifH gene data is presented in Figure 3B. Bradyrhizobium, Dechloromonas, Desulfovibrio, and Stenotrophomas were abundant in S. o cinarum, while Bradyrhizobium, Desulfovibrio, Xanthobacter and Anaeromyxobacter were the leading genera in S. barberi. Bradyrhizobium, Dechloromonas, Desulfovibrio, and Anaeromyxobacter were the dominant groups in S. robustum. S. spontaneum contained Bradyrhizobium, Azospirillum, Methanobacterium, and Paenibacilllus species. Bradyrhizobium, Dechloromonas, Xanthobacter and Anaeromyxobacter were present in S. sinense. Genus Burkolderia was found in S. barberi, S. spontaneum, and S. sinense while Beijerinckia was recorded in S. barberi, S. spontaneum, and S. sinense. Genus Idenella was commonly present in all the samples except S. spontaneum and Kosakonia was commonly present in S. barberi and S. spontaneum.

KRONA analysis
The KRONA software was used to visualize the species annotation results, in which the circles represent different points from the inside out class level, the size of the fan represented the relative proportions of the results of different OTU comments ( Figure 4). Species annotation results visualized with KRONA identi ed Bacillus, Pseudomonas, Pseudoarthobacter, Rhizobiales, Burkholderiales, Massilia and Streptomyces as the most dominant ones ( Figure 4A). Similar analyses carried out with nifH OTUs showed Bradyrhizobium, Dechloromonas, Desulfovibrio, Stenotrophomonas, Anaeromyxobacter etc. as the most common genera identi ed ( Figure 4B). The relative abundance of the top twenty diazotrophs at the genus level present in all sugarcane species is shown in Figure 5. Diazotrophs belonging to the genus Bradyrhizobium were present in all the samples tested.
Star analysis of all the sugarcane samples Star analysis was conducted using the top 10 genera of each sample. Star analysis for both 16S and nifH sequence data are presented in Additional le 1: Figure S5 A&B. Top 10 genera used for the 16S data based analysis were Pseudoarthobacter, Pseudomonas, Bacillus, Massilia, Gemmatimonas, Nitrospora, Haliangium, Ramibacter, Tumebacillus and Lysobacter. From this analysis, Bacillus genus was found high in S. o cinarum and S. robustum samples, while Pseudomonas was the dominant one in S. barberi, S. robustum, S. spontaneum and S. sinense. Tumebacillus genus was found only in S. spontaneum. Similarly, star analysis for nifH gene samples identi ed the presence of Bacillus in all the samples. Desulfovibrio was found in S. o cinarum, S. robustum and S. barberi in signi cant numbers. Xanthobacter was abundant in S. sinense. Anaeromyxo bacter was found in all the samples except that of S. spontaneum. Pseudoacidovorax genus was found only in S. robustum, S. barberi and S. sinense samples. Azospirillum and Methylobacterium were unique to S. spontaneum.

Beta diversity
Beta diversity measures species diversity among different samples collected from different or similar environments. We performed the principal coordinated analysis (PCoA) and UniFrac-based cluster analysis to understand the beta diversity of 16S rRNA and nifH gene in the sugarcane species studied (Figures 6 & 7). Based on 16S rRNA data, S. sinense, along with S. barberi, and S. o cinaruma long with S. robustum, formed independent clusters. And, S. spontaneum segregated away from all others, displaying higher beta diversity. The same was found true for nifH gene analysis ( Figure 7A). Hierarchical clustering based on the UniFrac cluster analysis showed similar results for both 16S rRNA and nifH gene data, containing similar identical sequences showing 0 (blue colour) distances. S. sinense to S. spontaneum showed more distance (red colour = 0.3) indicating dissimilarity in sequences of S. sinense to S. spontaneum samples ( Figure 6A and 7B).
The distribution of dominant genera based on their relative abundance performed with Bray-Curtis algorithm showed Bacillus being the dominant genus in all the samples except those from S. Spontaneum (Figure 8). Second dominant genus identi ed was Pseudomonas, while Pseudoarthobacter was the third leading one. S. spontaneum showed the presence of most unidenti ed genera in our analysis. Similarly, distribution of dominant genera in nifH samples showed the presence of Dechloromonas sp. abundant in all samples except those from S. barberi. Xanthobacter and Bradyrhizobiumg were found to be dominant in S. O cinarum and S. sinense, respectively ( Figure 9).

Discussion
There are several reports of identi cation and characterization of plant growth promoting rhizobacteria from sugarcane [34,35] and other crops [36][37][38] and they are also being used for crop productivity improvement. Most of the sugarcane studies are conducted with modern cultivated sugarcane hybrid varieties and there is no previous attempt to understand the diversity of rhizobacteria in its wild progenitors and closely related species such as S. o cinarum, S. spontaneum, S. robustum, S. barberi and S. sinensis as we did in this study. Here, we used high-throughput sequencing of 16S rRNA and nifH genes to study the rhizobacterial diversity and complexity in sugarcane species. Rhizospheric microorganisms are known to interact with plants for their survival and nutritional requirements [39,40]. Many of them are also bene cial bacteria to help plants with nutrient uptake and to cope with pathogens and abiotic stresses [18,41,42]. Considering the large microbial diversity in the rhizosphere, a large number of PGPRs are yet to be identi ed for various crops, including sugarcane. Bacteria are the most abundant of all the rhizospheric microbiota, and many are known to promote plant growth [2,3,42]. Our study made an effort to understand the diverse and complex diazotroph bacterial communities present in the rhizospheric soils of progenitors and closely related speciesof modern sugarcane hybrids as nitrogen xation plays a critical role in its growth and production [43,22].
16S rRNA sequence data revealed 6202 OTUs assigned to different bacterial species colonizing rhizosphere of different sugarcane species studied. Analysis of these OTUs showed that S. sinense rhizosphere has the largest number of rhizobacterial communities compared to other related species studied here. Besides this, alpha diversity analysis also predicted the highest bacterial diversity in S. sinense samples. However, other sugarcane species also showed considerable species diversity. Phylum level distribution studies identi ed the dominance of Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes, Chloro exi, Gemmatimonadetes, Planctomycetes and Nitrospirae. Many of them were reported to be present in sugarcane rhizosphere previously [44]. Out of the total OTUs identi ed, 1099 were from diazotrophs based on nifH gene data. Analysis of nifH gene sequence variation helped identify the top twenty genera (Bradyrhizobium, Dechloromonas, Desulfovibrio, Stenotrophomonas, anthobacter, Anaeromyxobacter, Azospirillum, Pseudoacidovorax, Methylobacterium, Azoarcus, Paenibacillus, Ideonella, Beijerinckia, Paraburkholderia, Burkholderia, Ru coccus, Geobacter, Sinorhizobium, Kosakonia and Azotobacter). Out of these, ten were found to x nitrogen in sugarcane and other plants.
Bradyrhizobium is known to x nitrogen in sorghum and sugarcane [45,46]. Members of Sinorhizobium genus are also known to x nitrogen symbiotically in leguminous alfalfa plants [47]. Azotobacter genus consists of seven different species and they are involved in atmospheric nitrogen xation in different crop plants [48]. Azotobacter and Beijerinckia were studied for diazotrophic attributes in early 1960s in rice and other cereal crops [49,50]. Contrary to Azotobacter, Beijerinckia genus is largely restricted to the tropics and its nitrogen xation ability has been reported in a variety of plants [51]. Kosakonia spp. x nitrogen on cucumber roots [52]. Roots of switchgrassare inhabited with nitrogen-xing bacteria belonging to Dechloromonas, Desulfovibrio, Azoarcus, Ideonella, Paraburkholderia and Burkholderia [53]. It is hard to identify and classify most of the bacteria in culture because of their morphological similarities. But, culture-independent methods like 16S rRNA sequencing and highly e cient, costeffective and provide accurate identi cation and classi cation of rhizobacteria. Overall, the dominant genera identi ed in this study are known to x atmospheric nitrogen for plant growth.
In the present study, we observed a few taxa such as Proteobacteria, Acidobacteria, Actinobacteria, Firmicute, Sphingomonas, Bradyrhizobium, and Gemmatimonas were dominant while genus such as Bacillus, Pseudomonas, Bradyrhizobium, Burkholderiaand rhizobium were dominant. In our previous studies, we observed that some taxa such as Proteobacteria, Acidobacteria, Actinobacteria, Firmicute, Bacteroidetes, Sphingomonas, Bradyrhizobium, Bryobacter, and Gemmatimonas were always present in the sugarcane rhizosphere [54,55]. Some genus such as Bacillus, Pseudomonas, Burkholderia, known for their plant growth promoting and nitrogen xating properties, were found to be enriched in the sugarcane rhizosphere [56,22,23]. Community composition analysis of 16S rRNA sequence data helped in tracking phylum and genus level distribution of rhizobacteria among different sugarcane species studied. Interestingly, we observed the presence of a few genera, namely Streptococcus, Rhodanbacter, Anaeromyxo bacter and Prevotella common among all the species studied here. The members of these genera were found commonly in soil and can colonize crop plants. From the previous reports it appears that Rhodanbacter, Anaeromyxobacter and Prevotella genera were isolated from different soil and plant sources and they were found to have nifH gene and nitrogen xation abilities [57][58][59]. Thus, we believe that more characterization of these bacteria colonizing the sugarcane rhizosphere will be bene cial for developing bio-based crop products to improve sugarcane crop productivity.
Diversity among these nitrogen xing bacteria was revealed by alpha and beta diversity analyses. The top ten genera with the highest abundance were found to be Bacillus, Desulfovibrio, Xanthobacter, Anaeromyxobacter, Pseudoacidovorax, Azospirillum and Methylobacterium. Among these Bacillus is a common bacterial diazotroph in sugarcane [60]. Desulfovibrio, Anaeromyxobacter, Azospirillum and Xanthobacter are nitrogen xers in rice [61][62][63][64]. Methylobacterium is a nitrogen xing bacterium, found in legumes [65], but such a function was not reported for Pseudoacidovorax.

Conclusions
All the sugarcane species studied here showed a signi cant number of nitrogen xing rhizobacteria, which strengthens the contention, that exploring rhizosphere bacteria may help develop a sustainable low sugarcane crop production system meeting its N requirement from atmospheric nitrogen xation requiring low nitrogen fertilizer input. Substantial genetic variation for rhizobacteria and diazotrophs communities exists among different progenitors and closely related species of modern cultivated sugarcane hybrids. However, considering the wide natural habitats of these wild species spanning both, tropics and sub-tropics, similar studies using accessions sourced from different locations and environmental conditions will greatly advance our understanding of sugarcane rhizo-microbiome. Future research should also focus on isolation and practical application of bene cial plant growth promoting rhizobacteria and diazotrophs. Filling the large knowledge gap on microbiota and sugarcane interactions is critical for exploiting these bene cial microbes for sustainable sugarcane agriculture.  Table S3. We sampled three soil samples randomly from each species and were used for DNA isolation.

Bacterial genome isolation and sequencing
Extraction of genomic DNA We used a culture-independent method [24] to study the bacterial composition of rhizospheric microbiome collected from the test species. Total microbial DNA was extracted using CTAB/SDS isolation method with minor modi cations [25]. The purity and concentration of DNA preparations were determined using 1% agarose gel electrophoresis. DNA samples diluted to 1ng DNA/μL using sterile water and used for further analysis. We mixed three DNA samples of each sugarcane species in the same concentration to get a mixed sample. Mixed samples were used for the sequencing analysis.
Identi cation of 16S rRNA and nifH genes Microbial 16S rRNA was ampli ed with the universal primers 341F (ACTCCTACGGGAGGCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT), which target the V3-V4 region. The nifH gene was ampli ed with primers Pol-F (TGCGAYCCSAARGCBGACTC) and Pol-R (ATSGCCATCATYTCRCCGGA), as previously reported [26]. Polymerase chain reaction (PCR) ampli cation of identi ed 16S rRNA and nifH genes was performed with Phusion® High-Fidelity PCR Master Mix (New England Biolabs). Thermal cycling consisted of initial denaturation at 98 ºC for 1 min, followed by 30 cycles of denaturation at 98 ºC for 10 s, annealing at 50 ºC for 30 s, and elongation at 72 ºC for 60 s, nally, 72 ºC for 5 min, following the protocols reported previously by Zhou et al. [27].

PCR product quanti cation and Library preparation and sequencing
Visualization and quanti cation of PCR products were conducted by mixing an equal volume of 1× loading buffer containing SYB green dye with the PCR products and electrophoresed on 2% agarose gel.
The 400-450bp DNA fragments were isolated and used for further experiments. Equimolar amounts of PCR products from all samples were pooled and the mixture was puri ed using Qiagen Gel Extraction Kit (Qiagen, Germany). Sequencing libraries of puri ed amplicons were generated using TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA) in accordance with manufacturer's protocol and index codes were added. The library quality and concentration were assessed on the Qubit@ 2.0 Fluorometer (Thermo Scienti c) and Agilent Bioanalyzer 2100 system. To perform sequencing the quali ed libraries were fed into the IlluminaHiSeq2500 platform.

Species annotation
OTUs were taxonomically annotated following a BLAST analysis against the Unite Database of each identi ed representative bacterial sequence done in QIIME software.
Phylogenetic relationship construction

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.