Comparative Evaluation of 16S rRNA and Housekeeping Gene-Specific Primer Pairs for Rhizobia and Agrobacteria Metagenomics ^†

Abstract

Of many housekeeping genes, gyrB and rpoB are used as alternative markers to 16S rDNA to analyze Rhizobia and Agrobacteria communities. However, the extent to which the targeted genes and their corresponding primers could be suitable in metagenomic studies within communities belonging to the two taxa remains elusive. This work evaluates in silico the taxonomic resolution of partial regions of two housekeeping and 16S rRNA genes in differentiating between Rhizobia and Agrobacteria. The study confirmed V5–V7 as the best 16S rDNA variable region for differentiating all the genera at a 100% threshold. However, rpoB and gyrB markers outcompeted the 16S rDNA in terms of taxonomic resolution regardless of the threshold, possibly replacing the use of 16S rDNA V-regions in metagenomics studies of Rhizobia and Agrobacteria.

1. Introduction

The 16S rRNA gene has been considered as the “gold standard” for microbial identification for many decades [1,2]. However, the selection of its most efficient variable (v) region(s) (V-region) and/or corresponding threshold(s) for microbiome analysis is still debated [3,4]. Several studies indicated that the efficiency of the 16S rDNA variable regions for high-throughput DNA sequencing analysis depends on multiple parameters such as the genetic background of the microorganisms of interest and the extent to which the evolution of their 16S rRNA genes occurred [5,6]. Rhizobia form a group of soil saprophytic bacteria capable of establishing a nitrogen-fixing symbiosis with most legume plants, playing a key role in soil fertility and sustainable agriculture [7]. Exploring Rhizobia communities is crucial for understanding biological nitrogen fixation in legume–Rhizobia symbiosis and its impact on soil health and agricultural productivity [8,9]. To date, studies have revealed that they are scattered across at least 18 genera of α- and β-proteobacteria [7], of which the majority are known to have conserved genetic structure across the rRNA operons [10]. Extended genetic conservation of the 16S rDNA has been reported for both alpha- and beta-Rhizobia [10,11]. Moreover, recent metagenomic studies have revealed that the 16S rRNA gene and corresponding variable regions may have insufficient taxonomic resolution for the accurate identification of Rhizobia and related plant-associated bacteria such as Agrobacteria (

Table A1. Overview of the discriminatory power of 16S rDNA V-regions for Rhizobia and Agrobacteria.

16S rDNA Region	Type of Sample	Main Findings	Reference
V4–V5	Soil samples and in silico analysis	V4–V5 marker failed to discriminate the Aminobacter–Mesorhizobium genera complex. It partially discriminated the Rhizobium–Allorhizobium–Neorhizobium–Pararhizobium complex. V5–V7 was found to be the most discriminant region	[17]
V3–V4 and V5–V7	Agrobacterium crown gall samples (gallobiome)	V3–V4 was outcompeted by V5–V7. V5–V7 yielded the highest number (4.3 fold) and percentage of bacterial reads after the HTAS analysis	[27]
V4, V4–V5	Soil samples	The V4-region failed to discriminate the Rhizobium complex, Burkholderia complex and Methylobacterium complex	[28]
V1–V3, V3–V4, V4, V4–V5, V6–V8, V6–V9	In silico analysis of 16 plant-related microbial genera	Insufficient resolution for several 16S V-regions. The V4 region failed to distinguish all the selected genera. Moreover, the widely used V3–V4 region was found to be less discriminant than V1–V3	[19]
full 16S rRNA gene	Bulk soil and soybean rhizosphere	Oxford Nanopore Technologies long-read sequencing was used to identify Bradyrhizobium populations at the species level	[29]

). The conserved protein-coding genes, also known as housekeeping genes, are required for the maintenance of basic cellular and metabolic functions and also are universally present in the bacterial kingdom [12]. Unlike the 16S rRNA gene, housekeeping genes exist in single-copy in bacterial genomes and their sequencing can avoid the overestimation of bacterial richness and abundance in ecological surveys [12,13]. Moreover, they were further proposed as robust markers for intra-species-level bacterial diversity analysis [14,15]. Therefore, some housekeeping genes, including gyrB (which encodes subunit B of the bacterial gyrase) and rpoB (which encodes subunit B of RNA polymerase), are used as alternative markers to 16S to analyze Rhizobia communities [14,16]. However, the extent to which the targeted genes and their corresponding primers could be suitable in metagenomic studies of all the genera of Rhizobia remains elusive. This work evaluates in silico the taxonomic resolution of partial regions of gyrB, rpoB and 16S rRNA genes in differentiating between Rhizobia and Agrobacteria genera. It proposes the more suitable DNA marker to use in rhizobial and agrobacterial community metagenomic studies.

2. Material and Methods

2.1. Data Collection

2.1.1. 16S rRNA Gene Sequences

The 16S rRNA gene sequences of the 18 Rhizobia genera reported by Gnangui et al. [17] were used in this study, with some modifications (Table 1). The Methylobacterium organophilum DSM 760^T 16S rDNA sequence was replaced due to quality concerns by the M. organophilum NBRC 15689^T (=DSM 18172^T) 16S sequence (accession number BPQV01000023.1) [18]. In addition, the newly sequenced Paraburkholderia graminis (P. graminis LMG 18924^T = C4D1M^T; CADIKA010000048.1) 16S rDNA data were also used. Thus, the analysis was based on 16S rDNA sequences of high quality. They were retrieved from the genomes of the type species of all the 18 genera of α- and β-proteobacteria harboring rhizobial species [17]. Moreover, the Agrobacterium tumefaciens type strain ATCC 4720^T 16S sequence of accession number JAAQPP01000028 was also included in the analysis, being a closely related genus of Rhizobia [7].

Table 1. List of alpha- and beta-proteobacteria genera harboring Rhizobia and Agrobacteria species used in this study (adapted from [17]).

N°	Rhizobia Genus 1	Genus Type Species 3	Genome Accession	Gene Full-Length Size (bp)
				16S rDNA	gyrB	rpoB
1	Allorhizobium	Allorhizobium undicola ORS 992T	NZ_JHXQ01000045	1482	2436	4140
2	Aminobacter	Aminobacter aminovorans DSM 7048T	NZ_SLZO01000023	1484	2448	4134
3	Azorhizobium	Azorhizobium caulinodans ORS 571T	AP009384	1482	2427	4131
4	Bradyrhizobium	Bradyrhizobium japonicum USDA 6T	NC_017249	1488	2436	4119
5	Cupriavidus 2	Cupriavidus necator N-1T	CNE_1c16970	1531	2526	4107
6	Devosia	Devosia riboflavina IFO13584T	NZ_JQGC01000043	1481	2454	4134
7	Ensifer	Ensifer adhaerens Casida AT	NZ_CP015880	1484	2436	4140
8	Mesorhizobium	Mesorhizobium loti DSM 2626T	NZ_QGGH01000001	1484	2472	4143
9	Methylobacterium	Methylobacterium organophilum NBRC 15689T	NZ_BPQV01000023	1482	2448	4131
10	Microvirga	Microvirga subterranea DSM 14364T	NZ_QQBB01000028	1486	2427	4131
11	Neorhizobium	Neorhizobium galegae HAMBI 540T	HG938353	1480	2436	4137
12	Ochrobactrum	Ochrobactrum anthropi ATCC 49188T	NC_009667	1482	2424	4134
13	Paraburkholderia 2	Paraburkholderia graminis  LMG 18924T	CADIK010000048	1532	2472	4107
14	Pararhizobium	Pararhizobium giardinii H152T	NZ_KB902704	1484	2436	4140
15	Phyllobacterium	Phyllobacterium myrsinacearum DSM 5892T	NZ_SHLH01000013	1484	2424	4155
16	Rhizobium	Rhizobium leguminosarum USDA 2370T	GCA_003058385	1480	2436	4140
17	Shinella	Shinella granuli DSM 18401T	NZ_SLVX01000061	1484	2433	4140
18	Trinickia 2	Trinickia symbiotica  JPY-345T	NZ_PTIR01000049	1530	2472	4107
19	Agrobacterium	Agrobacterium tumefaciens ATCC 4720T	JAAQPP010000028	1484	2436	4137

¹ Agrobacterium is not considered a rhizobium, but a closely related plant-associated bacterium [7]. ² The genera in bold belong to the class of beta-proteobacteria; they are known as beta-Rhizobia. ³ The nodulation capacity of each genus type species is accessible in the work of Gnangui et al. [17].

Table 1. List of alpha- and beta-proteobacteria genera harboring Rhizobia and Agrobacteria species used in this study (adapted from [17]).

N°	Rhizobia Genus 1	Genus Type Species 3	Genome Accession	Gene Full-Length Size (bp)
				16S rDNA	gyrB	rpoB
1	Allorhizobium	Allorhizobium undicola ORS 992T	NZ_JHXQ01000045	1482	2436	4140
2	Aminobacter	Aminobacter aminovorans DSM 7048T	NZ_SLZO01000023	1484	2448	4134
3	Azorhizobium	Azorhizobium caulinodans ORS 571T	AP009384	1482	2427	4131
4	Bradyrhizobium	Bradyrhizobium japonicum USDA 6T	NC_017249	1488	2436	4119
5	Cupriavidus 2	Cupriavidus necator N-1T	CNE_1c16970	1531	2526	4107
6	Devosia	Devosia riboflavina IFO13584T	NZ_JQGC01000043	1481	2454	4134
7	Ensifer	Ensifer adhaerens Casida AT	NZ_CP015880	1484	2436	4140
8	Mesorhizobium	Mesorhizobium loti DSM 2626T	NZ_QGGH01000001	1484	2472	4143
9	Methylobacterium	Methylobacterium organophilum NBRC 15689T	NZ_BPQV01000023	1482	2448	4131
10	Microvirga	Microvirga subterranea DSM 14364T	NZ_QQBB01000028	1486	2427	4131
11	Neorhizobium	Neorhizobium galegae HAMBI 540T	HG938353	1480	2436	4137
12	Ochrobactrum	Ochrobactrum anthropi ATCC 49188T	NC_009667	1482	2424	4134
13	Paraburkholderia 2	Paraburkholderia graminis  LMG 18924T	CADIK010000048	1532	2472	4107
14	Pararhizobium	Pararhizobium giardinii H152T	NZ_KB902704	1484	2436	4140
15	Phyllobacterium	Phyllobacterium myrsinacearum DSM 5892T	NZ_SHLH01000013	1484	2424	4155
16	Rhizobium	Rhizobium leguminosarum USDA 2370T	GCA_003058385	1480	2436	4140
17	Shinella	Shinella granuli DSM 18401T	NZ_SLVX01000061	1484	2433	4140
18	Trinickia 2	Trinickia symbiotica  JPY-345T	NZ_PTIR01000049	1530	2472	4107
19	Agrobacterium	Agrobacterium tumefaciens ATCC 4720T	JAAQPP010000028	1484	2436	4137

¹ Agrobacterium is not considered a rhizobium, but a closely related plant-associated bacterium [7]. ² The genera in bold belong to the class of beta-proteobacteria; they are known as beta-Rhizobia. ³ The nodulation capacity of each genus type species is accessible in the work of Gnangui et al. [17].

2.1.2. Housekeeping Gene Sequences

gyrB and rpoB full-length sequences belonging to the 19 selected Rhizobia and Agrobacteria type species were included in the analysis. Their relevant characteristics are reported in Table 1. They were also used to perform a comparative analysis of the taxonomic resolution of partial regions of both ribosomal and housekeeping genes.

2.2. Data Analysis

2.2.1. Re-Evaluation of the Discriminatory Power of 16S rRNA Gene V-Regions for Rhizobia and Agrobacteria

In silico evaluation of the discriminatory power of 16S rDNA V-regions for Rhizobia was previously performed [17]. It was refined in this study, together with data from Agrobacteria. Briefly, nine V-regions (V1 to V9) spanning the entire 16S rRNA gene of Rhizobia and Agrobacteria and their corresponding primers were selected (Table 2). The criteria for the selection included the extent to which these V-regions were targeted in plant-associated bacterial metagenomic analysis [19]. To evaluate their taxonomic resolution, (i) 19 Rhizobia and Agrobacteria full-length 16S rRNA gene sequences were aligned with MAFFT version 7 using MEGA v7 [20], (ii) a phylogenetic tree was reconstructed with a maximum likelihood (ML) method and (iii) pairwise similarity distances (P-distance) were calculated and used to identify the uniquely distinguishable taxa at 97% (OTU) or 100% (ASV) cut-offs [17,21]. Moreover, (iv) the full-length 16S-based sequence alignment was edited to match the total number of positions that correspond to those of each 16S variable region. Finally, for a given V-region, the total number of positions obtained after the editing was used to calculate similarity values that served to identify the uniquely distinguishable genera of Rhizobia and Agrobacteria [17,21].

2.2.2. Assessment of the Discriminatory Power of Housekeeping Genes for Rhizobia and Agrobacteria

gyrB and rpoB gene sequence analysis was performed following similar procedures to those described in Section 2.1.1 for 16S rDNA, with a few modifications. Briefly, the full-length size of the sequences of each housekeeping gene were aligned using MEGA Ver. 7. Subsequently, all the alignments were edited to partial sequences using the corresponding primers (Table 2). The selected sets of primers are the following: one pair of rpoB primers specific to Rhizobia [14], hereinafter “rpoB-1”, one pair of universal rpoB primers [13] (=“rpoB-2”) and one pair of universal gyrB primers [22] (=“gyrB-1”) (Table 2). The similarity values were calculated to identify the uniquely distinguishable taxa at 98% or 100% cut-offs, as recommended for housekeeping genes [14,22].

Table 2. Primers used to compare the taxonomic resolution of the 16S rRNA and housekeeping gene variable regions.

Target	Set of Primers	Forward Sequence (5′ to 3′)	Reverse Sequence (5′ to 3′)	Amplicon Size (bp) 3 (Rhizobia and Agrobacteria)
16S rRNA gene 1
V1–V2	27F/337R	AGAGTTTGATCMTGGCTCAG	CYIACTGCTGCCTCCCGTAG	320–350
V1–V3	27F/534R	AGAGTTTGATCMTGGCTCAG	ATTACCGCGGCTGCTGG	468–523
V3–V4	341F/805R	CCTACGGGNGGCWGCAG	GACTACHVGGGTATCTAATCC	440–465
V3–V5	341F/926Rb	CCTACGGGNGGCWGCAG	CCGTCAATTYMTTTRAGT	560–585
V4	515F/806R	GTGCCAGCMGCCGCGGTAA	GGACTACHVGGGTWTCTAAT	292
V4–V5	515F–Y/909-928R	GTGYCAGCMGCCGCGGTAA	CCCCGYCAATTCMTTTRAGT	413
V5–V7	799F/1193R	AACMGGATTAGATACCCKG	ACGTCATCCCCACCTTCC	409–417
V6–V9	928F/1492Rmod	TAAAACTYAAAKGAATTGACGGGG	TACGGYTACCTTGTTAYGACTT	605–612
V7–V9	1100F/1492Rmod	YAACGAGCGCAACCC	TACGGYTACCTTGTTAYGACTT	408–415
V1–V9	27F/1492Rmod	AGAGTTTGATCMTGGCTCAG	TACGGYTACCTTGTTAYGACTT	1445–1497
Housekeeping genes 2
“gyrB-1”	gyrB_aF64/gyrB_aR353	MGNCCNGSNATGTAYATHGG	ACNCCRTGNARDCCDCCNGA	287–302
“rpoB-1”	rpoB1479-F/rpoB1831-R	GATCGARACGCCGGAAGG	TGCATGTTCGARCCCAT	378–384
“rpoB-2”	Univ_rpoB_F_deg/ Univ_rpoB_R_deg	GGYTWYGAAGTNCGHGACGTDCA	TGACGYTGCATGTTBGMRCCCATMA	434–440

¹ For details about the selected 16S rDNA primers, please refer to Baker et al. [23], Johnson et al. [6] and Gnangui et al. [17]. ² “gyrB-1”, “rpoB-1” and “rpoB-2” refer to the pair of primers we selected from Barret et al. [22], Zhang et al. [14] and Ogier et al. [13], respectively. ³ Amplicon sizes were estimated in silico using all 19 proteobacterial genera harboring Rhizobia (n = 18 genera) and Agrobacteria (n = 1 genus).

Table 2. Primers used to compare the taxonomic resolution of the 16S rRNA and housekeeping gene variable regions.

Target	Set of Primers	Forward Sequence (5′ to 3′)	Reverse Sequence (5′ to 3′)	Amplicon Size (bp) 3 (Rhizobia and Agrobacteria)
16S rRNA gene 1
V1–V2	27F/337R	AGAGTTTGATCMTGGCTCAG	CYIACTGCTGCCTCCCGTAG	320–350
V1–V3	27F/534R	AGAGTTTGATCMTGGCTCAG	ATTACCGCGGCTGCTGG	468–523
V3–V4	341F/805R	CCTACGGGNGGCWGCAG	GACTACHVGGGTATCTAATCC	440–465
V3–V5	341F/926Rb	CCTACGGGNGGCWGCAG	CCGTCAATTYMTTTRAGT	560–585
V4	515F/806R	GTGCCAGCMGCCGCGGTAA	GGACTACHVGGGTWTCTAAT	292
V4–V5	515F–Y/909-928R	GTGYCAGCMGCCGCGGTAA	CCCCGYCAATTCMTTTRAGT	413
V5–V7	799F/1193R	AACMGGATTAGATACCCKG	ACGTCATCCCCACCTTCC	409–417
V6–V9	928F/1492Rmod	TAAAACTYAAAKGAATTGACGGGG	TACGGYTACCTTGTTAYGACTT	605–612
V7–V9	1100F/1492Rmod	YAACGAGCGCAACCC	TACGGYTACCTTGTTAYGACTT	408–415
V1–V9	27F/1492Rmod	AGAGTTTGATCMTGGCTCAG	TACGGYTACCTTGTTAYGACTT	1445–1497
Housekeeping genes 2
“gyrB-1”	gyrB_aF64/gyrB_aR353	MGNCCNGSNATGTAYATHGG	ACNCCRTGNARDCCDCCNGA	287–302
“rpoB-1”	rpoB1479-F/rpoB1831-R	GATCGARACGCCGGAAGG	TGCATGTTCGARCCCAT	378–384
“rpoB-2”	Univ_rpoB_F_deg/ Univ_rpoB_R_deg	GGYTWYGAAGTNCGHGACGTDCA	TGACGYTGCATGTTBGMRCCCATMA	434–440

¹ For details about the selected 16S rDNA primers, please refer to Baker et al. [23], Johnson et al. [6] and Gnangui et al. [17]. ² “gyrB-1”, “rpoB-1” and “rpoB-2” refer to the pair of primers we selected from Barret et al. [22], Zhang et al. [14] and Ogier et al. [13], respectively. ³ Amplicon sizes were estimated in silico using all 19 proteobacterial genera harboring Rhizobia (n = 18 genera) and Agrobacteria (n = 1 genus).

2.2.3. Comparative Analysis of the Discriminatory Power of 16S rRNA and Housekeeping Genes for Rhizobia and Agrobacteria

The ability to distinguish all known genera of Rhizobia and Agrobacteria at the genus level was applied as a metric in evaluating the taxonomic resolution of each genetic marker. The marker and corresponding pair of primers with the lowest uniquely distinguishable taxa was selected as the best target in the metagenomic analysis of Rhizobia and Agrobacteria (Figure A1).

3. Results

This study provides an evaluation of the discriminatory power of two housekeeping genes and nine commonly used 16S rRNA gene V-regions for differentiating Rhizobia and Agrobacteria. The main results are presented in the subsequent paragraphs.

3.1. Taxonomic Resolution of 16S rDNA V-Regions Re-Evaluated for Rhizobia and Agrobacteria

The analysis of nine variable regions of 16S rRNA revealed an insufficient taxonomic resolution for several V-regions, including the widely used V4, V3–V4, V3–V5, V4–V5 and V7–V9 regions (Figure 1). Moreover, V4 and V4–V5 were the only variable regions that could not discriminate all Rhizobia and Agrobacteria at one nucleotide polymorphism level (ASV level). In contrast, V5–V7 had the highest taxonomic resolution for differentiating Rhizobia and Agrobacteria at the genus level, regardless of the threshold. Although it appeared as the best target, the V5–V7 region has a limitation at lower cut-offs. For example, 17 out of 19 genera were discriminated at 97% (Figure 1).

3.2. The rpoB and gyrB Markers Outcompeted the 16S rDNA

The uniquely distinguishable taxa count showed that partial regions of rpoB and gyrB had sufficient resolution for differentiating all the bacteria genera (Figure 1). The partial sequences of these housekeeping genes were similar to those of the V5–V7 region in terms of sequence length (~400 bp). However, only rpoB and gyrB distinguished all the 19 type species of Rhizobia and Agrobacteria used in the study regardless of the threshold. In contrast, the analysis showed that the V5–V7 region could not distinguish between Aminobacter and Mesorhizobium at the 97% cut-off (Figure 1), suggesting that it was outcompeted by the housekeeping genes in terms of taxonomic resolution.

4. Discussion

We used the uniquely distinguishable taxa approach to evaluate the taxonomic resolution of different partial regions of gyrB, rpoB and 16S rRNA genes for distinguishing between Rhizobia and Agrobacteria in metagenomic analysis. The two groups of soil bacteria play important roles in agriculture in terms of nitrogen biochemical cycling and plant pathology, respectively [7]. Some metagenomic studies have reported the poor discriminatory power of the 16S rRNA gene for these plant-associated bacteria [17,19]. Insufficient resolution was confirmed for the nine 16S rRNA gene variable regions in this study. Our study further demonstrated that the usefulness of the 16S rRNA gene in metagenomic studies depends on the type of microorganism of interest and the extent to which their 16S rRNA genes have evolved [5,24]. Interestingly, rpoB and gyrB markers outcompeted the 16S rDNA in terms of taxonomic resolution regardless of the threshold, possibly replacing the use of 16S rDNA V-regions in the metagenomic studies of Rhizobia and Agrobacteria. Many studies reported the usefulness of housekeeping genes in microbiome analyses [25,26]. Pioneering studies first reported the use of single-copy housekeeping gene chaperonin-60 (cpn60) in microbial community analyses to improve resolution at the species and subspecies levels [25,26]. Subsequently, rpoB and gyrB genes have been successfully targeted to analyze bacterial communities from different environments, including legume plant nodules and rhizosphere soils [15,16]. However, to our knowledge, this is the first study on metagenomic analysis that aimed to demonstrate that partial sequences of rpoB and gyrB are highly appropriate markers for assessing the taxonomic structure of Rhizobia and Agrobacteria communities. Thus, rpoB and gyrB markers are recommended as they could provide a more comprehensive and unbiased picture of the composition of Rhizobia and Agrobacteria communities. A key limitation of our study is the lack of in vitro testing with environmental DNA. Thus, it will be interesting to address some potential important topics following this study, including (i) the validation of the potential of rpoB and gyrB markers for real-world application and (ii) the development of curated rpoB and gyrB DNA sequence databases [12].

5. Conclusions

Exploring plant-associated bacterial communities is crucial for understanding their impact on soil health and agricultural productivity. In this context, this work evaluated the taxonomic resolution of partial regions of two housekeeping and 16S rRNA genes for differentiating Rhizobia and Agrobacteria communities. It revealed that rpoB and gyrB gene markers have a good discriminatory power to differentiate between Rhizobia and Agrobacteria, and outcompeted the 16S rDNA in terms of taxonomic resolution. Thus, we recommend rpoB and gyrB markers for Rhizobia and Agrobacteria metagenomics. A limit to our performance appraisal is that the approach has not been tested yet with biological samples. Further analyses should confirm the potential of rpoB and gyrB markers using DNA samples. The findings could serve as indicators for future Rhizobia and Agrobacteria microbiome explorations.