A comparative transcriptome and fitness study found that the 191-kilobase conjugative plasmid pSfr64a behaves very differently after entering two bacterial species. In its native host, Sinorhizobium fredii GR64, the plasmid transfers at high frequency and its major transcriptional regions are highly active. After transfer into the bean symbiont Rhizobium etli, expression from those regions was strongly suppressed and conjugation became nearly undetectable.

In Agrobacterium tumefaciens, the same plasmid remained capable of high-frequency transfer, but weaker silencing coincided with extensive changes in host-gene expression, slower growth and a marked competitive disadvantage. The results support a model in which host-dependent transcriptional control determines the balance between plasmid propagation and host fitness. The researchers examined pSfr64a, a self-conjugative megaplasmid with a mosaic composition and a high transfer rate from its native S. fredii GR64 background.

RNA sequencing showed that most of its transcriptional activity in that host was concentrated in four sectors. The largest, extending from approximately 168 to 189 kilobases, contains genes associated with conjugation, including the Dtr system for DNA transfer, the Mpf system for mating-pair formation and traI, which participates in production of acyl-homoserine lactone quorum-sensing signals. A second sector near kilobase 5 contains a SAM-class I methyltransferase and a RHE_PE00001 paralogue.

The third, spanning roughly kilobases 13 to 19, includes genes associated with plasmid persistence and stability, while the fourth, around kilobases 23 to 25, contains IS21-associated genes. To determine whether the host effect extended beyond the conjugation machinery, the researchers inserted a green fluorescent protein reporter under control of the strong nptII promoter in an intergenic region of pSfr64a outside the Dtr and Mpf regions. Fluorescence was high when the reporter plasmid resided in S. fredii GR64 and A. tumefaciens but was at least threefold lower in R. etli CFN2001.

Because the reporter was driven by a constitutive promoter, the result was consistent with host-dependent silencing rather than the absence of plasmid-specific regulatory elements. Reporter fluorescence was also high when the same gene was carried on the natural R. etli conjugative plasmid pRet42a, further indicating that the suppression was associated with pSfr64a in the R. etli background. Laboratory mating assays quantified the effect on DNA transfer.

When pSfr64a::gfp was donated from S. fredii GR64, the mean conjugation frequency was 3,800 × 10^-4 transconjugants per output recipient cell for R. etli CFN2001, 210 × 10^-4 for A. tumefaciens GMI9023 and 430 × 10^-4 for A. tumefaciens UIA143. When the reporter plasmid was donated from R. etli CFN2001 to A. tumefaciens UIA143, the mean frequency was only 0.0001 × 10^-4 transconjugants per output recipient cell, approximately seven orders of magnitude below the value for transfer from S. fredii. In contrast, pSfr64a::gfp transferred from A. tumefaciens GMI9023 to UIA143 at a mean frequency of 380 × 10^-4.

The natural R. etli plasmid pRet42a transferred readily from R. etli to the same A. tumefaciens recipient, showing that the low pSfr64a transfer frequency was not a general incompatibility between the species. Comparative RNA sequencing of the plasmid showed that the same four active sectors remained recognizable in R. etli and A. tumefaciens, but expression was lower in both hosts than in S. fredii GR64, with stronger suppression in R. etli. Using the native host as the reference and thresholds of a Benjamini-Hochberg false-discovery rate below 0.05 and a log2 fold change of 1.5, the researchers classified 44 pSfr64a genes as differentially expressed.

Twenty-five were downregulated in both alternative hosts, including 19 in the conjugative region. In R. etli, genes in that region had log2 fold changes from −1.5 to −6.9, corresponding to reductions of approximately 2.8- to 119.4-fold. In A. tumefaciens, reductions ranged from 4.5- to 25.9-fold.

The minor changes observed for traM and traR suggested that stronger repression in R. etli was not simply caused by loss of these conjugation-associated regulators. The distinct transcriptional responses were accompanied by different fitness outcomes. In rich medium, acquisition of pSfr64a had little effect on the maximum specific growth rate of R. etli, which changed from 0.195 to 0.213 per hour, but reduced the corresponding rate of A. tumefaciens from 0.394 to 0.254 per hour.

A similar pattern occurred in minimal medium: the mean rate for R. etli changed from 0.175 to 0.148 per hour, whereas that for A. tumefaciens fell from 0.217 to 0.138 per hour. In eight-hour competition assays, plasmid-bearing R. etli had a mean competitive index of approximately 1.1, while plasmid-bearing A. tumefaciens had an index of approximately 0.25. At 48 hours, the respective indices were 1.9 and 0.09; at 72 hours, they were 1.2 and 0.04.

Thus, the R. etli result was compatible with no substantial disadvantage at eight hours, while the longer-term means were variable and remained above one. The A. tumefaciens results showed a sustained competitive disadvantage for cells carrying pSfr64a. Whole-transcriptome comparisons showed that the effects on host chromosomes were also species dependent.

Acquisition of pSfr64a produced no differentially expressed genes on the R. etli chromosome under the study’s criteria. Twelve differentially expressed genes were detected elsewhere in that host, all on its other plasmids: one was downregulated and 11 were upregulated. In A. tumefaciens, carriage of pSfr64a altered 49 chromosomal genes, comprising 18 downregulated and 31 upregulated genes.

These included lpxC, which participates in the first committed step of lipid A biosynthesis; two genes encoding nonessential tRNA-Ala products; rpsT, encoding the nonessential 30S ribosomal protein S20; and rpmF, encoding the nonessential 50S ribosomal protein L32. Reduced expression of lpxC, rpsT and rpmF was corroborated by quantitative reverse-transcription PCR. The authors considered it plausible that the cumulative effects of these changes contributed to the fitness deficit, but the experiments did not establish that they were its sole cause.

The A. tumefaciens response also displayed an unexplained regional pattern. Seventeen of the 18 downregulated genes were located on the circular chromosome, whereas 27 of the 31 upregulated genes were on the linear chromosome. The two replicons differ in replication organization and evolutionary origin: the circular chromosome depends on DnaA for replication, while the linear chromosome uses a RepABC system characteristic of megaplasmids.

Differences in topology or evolutionary history were proposed as possible explanations, but neither was directly tested. The observations establish that the transcriptional consequences of pSfr64a carriage in A. tumefaciens extended across both chromosome types rather than being confined to plasmid genes. Mutant analysis identified MucR, known as RosR in R. etli, as one component of the silencing response but not the complete mechanism.

Insertion disruption of mucR relieved some repression of the GFP reporter and increased pSfr64a transfer by approximately 100-fold. Adding spent S. fredii culture medium, which replenishes acyl-homoserine lactone signals, likewise increased transfer by about 100-fold. Combining the mucR mutation with spent medium produced a nonadditive increase of roughly 1,000-fold, yet conjugation remained far below the level observed in the native host.

These results separate the relief of reporter or selected-gene silencing from restoration of efficient transfer, indicating that additional host-dependent mechanisms constrain conjugation in R. etli. Overall, the study proposes that pSfr64a operates within a two-tier regulatory hierarchy. An upper layer of xenogeneic silencing, potentially supplemented by DNA methylation, restrains the plasmid in a new host; quorum sensing then acts at a subordinate level to promote transfer when the broader repression is sufficiently relieved.

The dominant silencing component in R. etli remains unknown. Common enterobacterial silencing proteins such as H-NS and HU/IHF are absent from rhizobial genomes, and preliminary tests with an inhibitor of 5-methylcytosine did not alter growth or conjugative-gene expression, but epigenetic modification was not comprehensively assessed. The comparison was limited to particular strains, plasmid backgrounds and laboratory conditions.

Within those limits, the experiments show that acquiring a megaplasmid creates an evolutionary conflict between plasmid propagation and host stability: efficient transfer in A. tumefaciens came with major transcriptional and fitness costs, whereas the stronger suppression observed in R. etli permitted coexistence with little measurable harm.

Subject of Research: Host-specific transcriptional regulation of a conjugative rhizobial megaplasmid after horizontal transfer

Article Title: Transcriptional silencing of a conjugative Sinorhizobium megaplasmid upon transfer to different rhizobial species

Article References: de la Cruz, C., Cervantes, L., Torres-Tejerizo, G., & Romero, D. (2026). Transcriptional silencing of a conjugative Sinorhizobium megaplasmid upon transfer to different rhizobial species. Frontiers in Microbiology, 17, Article 1977405. https://doi.org/10.3389/fmicb.2026.1977405

Image Credits: AI Generated

DOI: 10.3389/fmicb.2026.1977405

Keywords: rhizobial megaplasmid, horizontal gene transfer, xenogeneic silencing, bacterial conjugation, RNA sequencing, host fitness, quorum sensing, MucR, Rhizobium etli, Agrobacterium tumefaciens, Transcriptional, silencing