Amplification of genes by multiplex PCR
Three genes namely nifH, pqqC and accd-DR and the reference gene 16S rRNA were successfully amplified employing multiplex PCR from the metagenomic DNA extracted from the rhizospheric soil of rice plant (Fig. 1b). Four DNA bands of approx. 1500, 546, 390 and 113 bp corresponding to 16S rRNA, pqqC, nifH and accd-DR were repeatedly observed suggesting the efficacy of multiplex PCR in amplification of all the genes (Fig. 1b). That the amplified amplicons are indeed from bacteria is evident from the fact that the template DNA used also resulted in the amplification of 16S rRNA.
Aanlysis of nifH and pqqC diversity in rhizosphere of rice
A 390 bp amplicon of nifH amplified by multiplex PCR using metagenomic DNA was cloned in pGEM-T easy cloning vector to obtain large number of nifH clones. 96 clones were selected and insert was double digested with AluI and RsaI which generated 15 RFLP groups. One representative from each RFLP group of clones was sequenced and subjected to BLAST for nifH sequence diversity. It is evident from the results of Table 2 that the highest number of identical clones are present in SBT AK C3 (27) followed by SBT AK C1 (14). Furthermore, results of BLAST analysis showed that the sequences belonging to five clones (SBT AK C1, AK C3, AK C6, AK C45 and AK C52) are represented by different species of Bradyrhizobium/Mesorhizobium (out of 15 clones). The remaining ten clones contain sequences of different genera/order/classes (Table 2). It is also evident from the data that majority of the clones exhibit maximum similarity (91–100%) with unculturable bacteria. In fact, three clones showed 100% similarity with the uncultured bacterial sequences. Further analysis of sequences revealed the presence of highest percentage of sequences belonging to the phylum proteobacteria (10/15, 66.66%) comprising alpha (5/15), beta (2/15), gamma (1/15) and delta classes (2/15). Sequences of four clones with very low percentage (ranging from 6.66 to 13.33%) belonged to other group of bacteria. Sequence of the clone-SBT AK C2 did not show affiliation with nifH sequences of any cultured bacteria.
Table 2
Similarity of nifH sequences of various clones with uncultured and cultured bacterial sequences available in the database
nifH clones | Accession number of this study | Closest uncultured bacteria (accession number) | % Identity | Closest match (cultured bacteria) | Accession number | % Identity | Group affiliation |
SBT AK C1 (14) | MF680849 | AHN51144.1 | 100 | Bradyrhizobium japonicum | ACT67989.1 | 99 | α |
SBT AK C2 (3) | MF680850 | ACN23728.1 | 94 | - | - | - | Other |
SBT AK C3 (27) | MF680851 | CAG30111.1 | 100 | Bradyrhizobium sp. | BAF95631.1 | 98 | α |
SBT AK C6 (8) | MF680852 | AHN50517.1 | 96 | Bradyrhizobium japonicum | ACT67982.1 | 96 | α |
SBT AK C10 (6) | MF680853 | APA22235.1 | 95 | Methanocella conradii | WP014404752.1 | 94 | Archaea |
SBT AK C14 (4) | MF680854 | ADZ48369.1 | 100 | Azoarcus communis Rhodocyclales bacterium | AAB63033.2 OHC65356.1 | 98 98 | β |
SBT AK C24 (4) | MF680855 | ADX43226.1 | 93 | Methanoregula boonei | WP012106688.1 | 92 | Archaea |
SBT AK C26 (4) | MF680856 | APD29463.1 | 95 | Syntrophus gentianae | WP093883931.1 | 96 | α |
SBT AK C28 (3) | MF680857 | ADZ48348.1 | 98 | Desulfuromonadales sp. | OGR29997.1 | 99 | α |
SBT AK C45 (4) | MF680858 | AHN51275.1 | 99 | Bradyrhizobium sp. | ACT67983.1 | 99 | α |
SBT AK C50 (5) | MF680859 | AHN50775.1 | 98 | Opitutaceae bacterium | WP009512762.1 | 95 | Verrucomicrobia |
SBT AK C52 (6) | MF680860 | AHN51303.1 | 99 | Bradyrhizobium sp. Mesorhizobium loti | AML61104.1 BAF95636.1 | 99 98 | α |
SBT AK C53 (4) | MF680861 | BAP16836.1 | 91 | Gallionellales sp. | OGS90295.1 | 94 | β |
SBT AK C73 (2) | MF680862 | APA22124.1 | 97 | Nitrospirae bacteria | OGW27708.1 | 96 | Nitrospirae |
SBT AK C85 (2) | MF680863 | AJF14338.1 | 94 | Vibrio natriegens | AAD55588.1 | 93 | γ |
Query coverage- 100% for both uncultured and cultured bacteria. Number in bracket shows total number of clones present in each representative clone based on RFLP similarity. |
In the case of pqqC, 96 clones were selected following the protocol adopted for the nifH analysis. All the 96 clones clustered to twelve groups on the basis of RFLP pattern. Of the 12 representative clones, the highest number of identical clones were present in the SBT AK8 (28) followed by SBT AK1 (12) (Table 3). The remaining ten clones contained three to eight clones. Further analysis of representative sequences of six clones (SBT AK1, AK8, AK13, AK19, AK78 and AK96) revealed identical level of similarity with both the cultured and uncultured bacteria. However, sequences of the remaining six clones showed higher level of similarity with the cultured bacteria. Interestingly, of the 12 group of clones, 10 clones consisted members of cultured pseudomonads namely Pseudomonas putida, Pseudomonas sp. GM50, Pseudomonas lini, Pseudomonas sp. FSL W5-0299, Pseudomonas mandelii, Pseudomonas sp., Pseudomonas oryzae, Pseudomonas sp. PICF141, and Pseudomonas sihuiensis all belonging to γ-proteobacteria. Sequence of the clone SBT AK96 also belonged to a member of gamma class, Marinobacterium jannaschii (with 75% percent identity). Clone SBT AK40 did not show affiliation with any bacteria present in the NCBI database, nevertheless shared homology with the sequence of cultured bacterium, Candidatus Entotheonella with 100% query coverage and 74% identity (Table 3).
Table 3
Similarity of pqqC sequence of various clones with uncultured and cultured bacterial sequences available in the database
pqqC clones | Accession number of this study | Closest uncultured bacteria (accession number) | % Identity | Closest match (cultured bacteria) | Accession number | % Identity | Group affiliation |
SBT AK1(12) | MH453460 | ATI09183.1 | 99 | Pseudomonas putida | WP079226301.1 | 99 | γ |
SBT AK8 (28) | MH453461 | ATP13615.1 | 97 | Pseudomonas sp. GM50 | WP008008220.1 | 97 | γ |
SBT AK13 (6) | MH453462 | ATP13615.1 | 99 | Pseudomonas lini | WP048396408.1 | 99 | γ |
SBT AK19 (5) | MH453463 | ATI09183.1 | 100 | Pseudomonas putida | WP014755118.1 | 100 | γ |
SBT AK25(5) | MH453464 | ATP13613.1 | 97 | Pseudomonas sp. FSL W5-0299 | WP077749285.1 | 99 | γ |
SBT AK33 (6) | MH453465 | ATP13613.1 | 98 | Pseudomonas mandelii | WP083376520 | 99 | γ |
SBT AK36 (7) | MH453466 | ATP13613.1 | 97 | Pseudomonas sp. | WP018929700.1 | 99 | γ |
SBT AK40 (5) | MH453467 | - | - | Candidatus Entotheonella gemina | ETX06873.1 | 74 | Other |
SBT AK54 (7) | MH453468 | ATP13506.1 | 91 | Pseudomonas oryzae | WP090351895.1 | 94 | γ |
SBT AK78 (8) | MH453469 | ATP13615.1 | 99 | Pseudomonas sp. PICF141 | WP095630850.1 | 99 | γ |
SBT AK91 (4) | MH453470 | ATP13489.1 | 92 | Pseudomonas sihuiensis | WP092375741.1 | 96 | γ |
SBT AK96 (3) | MH453471 | ATP13506.1 | 75 | Marinobacterium jannaschii | WP027857990.1 | 75 | γ |
Query coverage- 100% for both uncultured and cultured bacteria. Number in bracket shows total number of clones present in each representative clone based on RFLP pattern similarity. |
NGS for nifH, pqqC and accd-DR diversity and bacterial community composition
With a view to gain better understanding of nifH sequence diversity, NGS approach was applied. Altogether, 90236 consensus sequences of nifH were retrieved and after removing 12324 (13.66%) chimeric sequences, 77912 (86.34%) pre-processed consensus sequences were obtained. With 77912 pre-processed consensus sequences, 21532 OTUs were obtained after clustering based on sequence similarity (similarity cut off = 0.97). Subsequently, 19487 OTUs having less than five reads were filtered and the remaining 2045 OTUs finally selected for the taxonomical abundance study. Altogether, fourteen types of genera containing varying number of OTUs were found. Halorhodospira was the dominant genus with 151 OTUs (7.38%) followed by Frankia (6.74%) and Bradyrhizobium (6.55%). Percent distribution of OTUs in other genera are presented in Fig. 2a. Sequences showing no alignment against taxonomic database comprised 65.86% (unknown) and those less than five in numbers included 1.22% (Fig. 2a).
Similarly, NGS of pqqC was done following the steps used for nifH analysis. Altogether, 15,805 OTUs were obtained after clustering based on the sequence similarity of total reads. Pseudomonas sp. was the dominant genus with 7,703 OTUs (48.73%) followed by Acinetobacter sp. (6.78%) and Azotobacter sp. (6.01%). Percent distribution of OTUs in other genera were; Klebsiella sp. (2.73%), Xanthomonas sp. (1.46%), Erwinia sp. (0.87%), Stenotrophomonas sp. (0.77%) and others (32.6%) (Fig. 2b).
NGS of accd-DR (113 bp) was done as per the steps used for nifH and pqqC analysis. Accordingly, of the 367529 paired-end reads, 361570 pre-processed consensuses sequences were obtained. Out of 361570 reads, a total of 69841 OTUs were identified and after removing OTUs with less than 5 reads, 28872 OTUs were finally selected for further analysis. Altogether, seven genera containing varying number of OTUs (above 1%) were found. Data showed the highest abundance of Acidovorax sp. (58.28%) followed by Paraburkholderia sp. (14.75%), Variovorax sp. (8.53%), Desmospora active (5.61%), Pseudomonas syringae (2.17%), Streptomyces sp. (1.80%), Kibdelosporangium aridum (1.01%) and others (7.85%) with less than 1% of OUTs (Fig. 2c)
Phylogenetic analysis of nifH, pqqC and accd-DR
Open reading frame (ORF) for amino acid sequence was checked and similar orientation for sequences of all the clones was made for multiple sequence alignment (MSA) (Fig. 3a, b, c). Phylogenetic tree of NifH sequences based on the deduced amino acid sequences was constructed by using sequences of fifteen clones of this study and twenty one similar sequences of NifH protein of cultured bacteria retrieved from NCBI database. It is evident from the tree that all the 15 clones could be placed in two clusters comprising NifH sequences of alpha-, beta-, gamma-, delta-proteobacteria and other bacterial nitrogenases (Fig. 4). Clone SBT C1 showed close relationship (98% amino acid similarity) to Geobacter sp.-OR 1 and Geobacter sp. M21 (94%). Clone SBT AK 45 is most closely related (99% amino acid similarity) to Geobacter pickeringii. Clone SBT AK28 shares 99% similarity with the sequence of Desulfuromonas sp. Clone SBT AK C14 shares 99% similarity with the nitrogenase amino acid sequence of the beta-proteobacteria Azoarcus sp. CC-YHH848 and Thauera sp. D20. Clone SBT AK85 is related to the sequence of Methylomonas (96%) and Vibrio natriegens NBRC (83%). Similarly, clone SBT AK73 shares 97% similarity with the amino acid sequence of the alpha-class bacterium Bradyrhizobium japonicum. Clone SBT AK53 shows similarity (93%) with the sequence of Gallionellales bacterium GWA2. Nitrogenase amino acid sequence of the clone SBT AK C52 shows close relationship with the sequence of Bradyrhizobium sp. (99%) followed by Mesorhizobium loti (98%). Clone SBT AK C3 shares 98% similarity (amino acid sequence) with the sequence of Xanthobacter tagetidis. NifH protein sequences of the clones SBT AK C50, C6 and C26 show 97, 95 and 96% similarity with the sequences of Bradyrhizobium japonicum, B. japonicum and Syntrophus gentianae respectively (Fig. 4). Clones SBT AK C2, C10 and C24 show relatedness (89 to 93% amino acid similarity) with the sequence of nitrogenase of bacteria namely Spirochaeta perfilievii, Methanocella conradii and Kiritimatiellales.
Similar to NifH, phylogenetic tree of PqqC was constructed using the deduced amino acid sequences of one representative from all the twelve clones and ten similar sequences of PqqC of cultured bacteria retrieved from NCBI data base. Evidently, all the twelve clones fall into two clusters with five groups in cluster I (10 clones) and one in cluster II (2 clones). Representative sequences from clones of cluster I show close relationship (99 to 100% amino acid similarity) with different species of Pseudomonas (Fig. 5). Of these, sequences of clones SBT AK33, AK36 and AK19 shared 100% similarity with the sequences of Pseudomonas species. Sequences of two clones (SBT AK8 and AK91) shared 96–97% similarity with different species of Pseudomonas but clone SBT AK54 had only 93% similarity with the sequence of P. oryzae. On the other hand, clones SBT AK40 and AK96 from cluster II grouped together and showed 100% homology between each other but did not show significant similarity (maximum 76% similarity) with the sequences of any species available in the database (Fig. 5). It is also evident from the phylogenetic analysis that there is a high degree of relatedness between majority of the Pseudomonas sp.
Construction of pylogenetic tree of accd-DR was possible by using top six OTUs nucleotide sequences assigned in NGS for bacterial species namely OTU1-Acidovorax sp., OUT13-Paraburkholderia sp., OTU26-Pseudomonas sp., OTU44-Streptomyces sp., OTU65- Kibdelosporangium phytohabitans and OTU78-Variovorax sp. accd-DR nucleotide sequences of above OTUs were converted into amino acid sequences with the help of translational tool. Resultant amino acid sequences of each OTU were queried for similarity (above 97%) search against NCBI database and phylogenetic tree was constructed. Among the six OTUs, sequences of OTU1, OTU13, OTU44 and OTU65 shared 100% similarity with Acidovorax citrulli, Paraburkholderia, Streptomyces sp. and Kibdelosporangium phytohabitans respectively. Sequences of OTUs78 and 26 showed 92 and 99% similarity with the sequences of Variovorax sp. and Pseudomonas sp. respectively (Fig. 6).