|Home | About | Journals | Submit | Contact Us | Français|
The type IV secretion system (T4SS) of the plant intracellular symbiont Sinorhizobium meliloti 1021 is required for conjugal transfer of DNA. However, it is not required for host invasion and persistence, unlike the T4SSs of closely related mammalian intracellular pathogens. A comparison of the requirement for a bacterial T4SS in plant versus animal host invasion suggests an important difference in the intracellular niches occupied by these bacteria.
Sinorhizobium meliloti is a soil bacterium that invades roots of host plants and enters cells of the plant cortex within a host-derived membrane compartment, differentiating into the symbiotic bacteroid form that supplies the host with fixed nitrogen compounds (19). The pSymA plasmid of the plant symbiotic bacterium S. meliloti 1021 contains virB1 to virB11 and virD4/traG genes encoding a putative type IV secretion system (T4SS) that has not yet been characterized (6). This report describes studies of the role of this secretion system in conjugation and symbiosis.
T4SSs play different roles in different bacteria. In many species, these systems function in conjugal transfer of DNA between bacteria (12). In animal pathogens that form a chronic, intracellular infection of their hosts, a functional T4SS is required for the bacteria to establish and maintain themselves in the proper vacuolar compartment (9, 40). T4SSs can also function as virulence determinants of nonintracellular plant and animal pathogenic bacteria, in some cases injecting effector proteins or DNA into the host cytosol (12, 28), and as determinants of the host range of plant symbiotic bacteria (23). The T4SS of S. meliloti 1021 is particularly interesting because the closely related pathogens (order Rhizobiales) Brucella and Bartonella spp. require their T4SSs in order to persist in the intracellular environments of their mammalian hosts (9, 30, 40).
Since successful host infection by Brucella spp. has been shown to require many of the same determinants that S. meliloti requires for symbiosis (e.g., bacA, the exoS/chvI two-component regulator system, and hfq), we tested whether the S. meliloti T4SS is involved in intracellular survival of this bacterium (24, 38, 43; L. Barra and G. C. Walker, unpublished observations). Both our own data (Fig. (Fig.1A)1A) and another report (6) have shown that, in an otherwise wild-type background, mutants with mutations in the S. meliloti virB operon encoding critical components of the T4SS remain capable of invading and forming a functional symbiosis on the plant host alfalfa (Medicago sativa). In contrast, Brucella spp. and Bartonella spp. require their T4SSs for persistence within their respective hosts (9, 30, 40).
Despite the fact that the S. meliloti T4SS is not required for nodulation of the plant host alfalfa, we set out to determine whether the T4SS of S. meliloti might have a function in this bacterium, given the diversity of functions that have been found for the T4SSs of other bacteria. We tested whether the T4SS is involved in (i) bacterial conjugation, (ii) determining the plant host range of S. meliloti, and (iii) allowing invasion to proceed in the absence of other invasion determinants.
A deletion of the virB genes virB6 to virB11, predicted to encode essential components of the T4SS apparatus, was generated by cloning 1.14 kb upstream of virB6 and 0.49 kb downstream of virB11, flanking the AB2001 lacZ-Gn cassette, into pK19mobsac (the lacZ-Gn cassette construct pAB2001 was a gift from A. Becker, University of Bielefeld) (7). This construct was mobilized into S. meliloti 1021 by triparental mating as previously described (20). Exconjugants were selected on M9 minimal medium containing 50 μg/ml gentamicin and 10% sucrose. virBΔ6-11 strains were checked for insertion of the lacZ-Gn cassette and for the absence of wild-type sequences by PCR (data not shown). Low-level expression of the virB::lacZ transcriptional fusion was observed both on X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) plates and in o-nitrophenyl-β-d-galactopyranoside assays of liquid culture (data not shown).
The T4SS-encoding virB genes of S. meliloti 1021 are very similar to the avh genes of the Agrobacterium tumefaciens pAtC58 plasmid, which are required for conjugal self-transmissibility of this plasmid (6, 10, 21). The T4SS proteins (VirB1 to VirB11 plus the VirD4/TraG substrate-coupling protein) and the TraA, TraC, and TraD DNA transfer and replication proteins are all required for a T4SS to function in DNA transfer (Table (Table1)1) (12). The pSymA plasmid of S. meliloti 1021 is predicted to encode a homolog of each of these proteins (6). Therefore, we tested the role of the S. meliloti T4SS in conjugation. We assayed conjugation through the T4SS by using a highly transmissible IncQ (RSF1010) plasmid, a method that has previously been used to quantitate conjugation through T4SSs of A. tumefaciens (10, 44). Selection for transfer of the pML122 IncQ plasmid to a naïve S. meliloti strain was performed on 200 μg/ml neomycin-LBMC plates (5, 44) (pML122 was a gift from A. N. Binns, University of Pennsylvania). Only 3.9 × 10−8 transconjugants per donor were detected using wild-type S. meliloti 1021 (Table (Table2).2). However, a >104-fold-higher level of transfer of pML122 (1.8 × 10−3 to 5.2 × 10−3 transconjugants per donor) was detected from an S. meliloti 1021 strain in which we had deleted the rctA open reading frame (Table (Table2).2). An S. meliloti mutant defective in rctA, a negative regulator of the vir genes, has previously been reported to be capable of transferring even the 1.4-Mb S. meliloti pSymA plasmid to a plasmidless A. tumefaciens strain (36). For these studies, we constructed an unmarked S. meliloti rctA deletion mutant by cloning 580 bp upstream of rctA and 579 bp downstream into pK19mobsac and generating the S. meliloti mutant strain by the same methods used to generate the virBΔ6-11 mutant. Deletion of rctA was confirmed by PCR (data not shown). In the rctA deletion strain, a much higher level of expression of the virB::lacZ transcriptional fusion was observed on X-Gal plates (data not shown).
The virBΔ6-11 mutants were not capable of conjugal transfer in either the wild-type S. meliloti 1021 or the rctA deletion mutant background, demonstrating a critical role for one or more of these T4SS components in S. meliloti conjugation (Table (Table22).
Symbiosis between Medicago truncatula (a diploid relative of alfalfa) and S. meliloti strain ABS7 (a native bacterial isolate from this plant) results in a more successful symbiosis than with S. meliloti strain 1021. The differences in molecular determinants between strains 1021 and ABS7 that are responsible for the difference in symbiotic efficiency are unknown. Based on studies of the symbiosis of Mesorhizobium loti strain R7A with the plant Leucaena leucocephala (23), we reasoned that the elimination of the T4SS of S. meliloti 1021 might similarly result in a more efficient symbiosis with the plant host M. truncatula. Wild-type M. loti R7A forms nonfunctional nodules on this plant host, while strains with mutations in the vir genes gain the ability to fully invade the plant and fix nitrogen (23). In principle, this could result from a protein injected by the T4SS of the wild-type bacterium triggering a plant defense response that aborts the symbiosis.
To test the effects of virB deletion on host range, 3-day-old M. truncatula cv. Jemalong A17 seedlings (46) were inoculated with 100 μl of a bacterial suspension (optical density at 600 nm = 0.05) of either S. meliloti ABS7, S. meliloti 1021, four separate isolates of the S. meliloti 1021 virBΔ6-11 mutant, or the S. meliloti 1021 expR::lacZ-Gn strain, which carries the lacZ-Gn cassette at a neutral site in the genome (as a control for the presence of the lacZ-Gn marker). Plants were grown on buffered nodulation medium (pH 6.5) plates (15). Plant height was measured at 7.5 weeks postinoculation. As shown in Fig. Fig.1B,1B, the deletion of the virB6 to -11 genes had neither a positive nor a negative effect on the efficiency of the symbiosis of S. meliloti 1021, indicating that the T4SS does not act as a critical determinant that prevents successful nodulation of M. truncatula by this strain.
Another possible role for the T4SS is that it might function as a secondary invasion determinant that acts synthetically with another bacterial effector. The S. meliloti 1021 exopolysaccharide effector molecule succinoglycan is required for efficient invasion of alfalfa and M. truncatula plants through their root hair cells (11). The succinoglycan-deficient S. meliloti 1021 exoY mutant is not able to colonize nodules efficiently. However, a small percentage of alfalfa plants inoculated with this strain do eventually show signs of a functional symbiosis after 6 to 7 weeks (49). In contrast, this succinoglycan-independent invasion has not been observed when the host plant M. truncatula is inoculated with the S. meliloti 1021 exoY mutant (data not shown). The fact that succinoglycan-independent invasion by S. meliloti 1021 occurs on one plant host but not another suggests that an additional host-specific, bacterial determinant might be involved in this process. The observation that a T4SS is involved in mammalian host-pathogen interactions suggested it as a potential candidate for such a determinant. Several host-pathogen interactions have been shown to require that protein effectors modulating host cell behavior be introduced into the host cytosol via a T4SS. Two notable cases are the Bartonella henselae infection of endothelial tissue, where the T4SS is required for translocation of Bep proteins into host cells, and the Helicobacter pylori infection of gastric tissue, where the T4SS translocates cytotoxin-associated antigen (CagA) into host cells (4, 31, 41). The activity of Bep and CagA proteins in the host cytosol produces dramatic changes in cell morphology, among other effects (41, 42). Similarly, infection thread formation in the S. meliloti/alfalfa symbiosis requires a complete reorganization of root hair cell morphology, an activity that in principle might be facilitated by T4SS-translocated factors.
To investigate the possibility that the T4SS of S. meliloti contributes to the low level of succinoglycan-independent nodule invasion of a succinoglycan-deficient strain, 3-day-old alfalfa cultivar Iroquois plants were inoculated with 100 μl of a bacterial suspension (optical density at 600 nm = 0.05) of either wild-type S. meliloti 1021, two independent virBΔ6-11 isolates, the succinoglycan-deficient exoY210::Tn5 mutant (51), or two independent isolates of the exoY210::Tn5 virBΔ6-11 double mutant. Plants were scored for successful invasion by the number of pink nodules that developed on the roots. A persistent pink color indicates that the plant nodule tissue is expressing leghemoglobin, which indicates that a functional nitrogen-fixing symbiosis has been established (26).
Plants inoculated with the S. meliloti 1021 exoY mutant formed 1.8 ± 2.6 pink nodules per plant, while plants inoculated with isolates of exoY210::Tn5 virBΔ6-11 double mutants formed 1.8 ± 2.9 or 1.7 ± 2.6 pink nodules per plant (Fig. (Fig.2)2) , indicating that the T4SS of S. meliloti 1021 is not responsible for effecting succinoglycan-independent invasion of alfalfa roots.
The recent phylogenetic analysis of bacterial T4SSs by similarity of VirB4 protein sequences does not provide a clear distribution of T4SSs into functional classes based on degree of similarity (18). Phylogenetic analysis of the S. meliloti VirB4 places it in the “A cluster” of the vir genes, along with the A. tumefaciens Avh proteins, which are involved in conjugation; the A. tumefaciens Vir proteins, which are involved in injecting bacterial DNA and proteins into plant hosts; the Bartonella Vir proteins, which are involved in establishing an intracellular infection of the mammalian host; the M. loti R7A symbiosis island-encoded Vir proteins, which are involved in determining bacterial competitiveness and host range; and the Rhizobium etli pRetCFN42d plasmid-encoded Vir proteins, whose function has not yet been determined (10, 23, 40). The percent protein sequence identity of each S. meliloti 1021 T4SS component to those of the other members of the “A cluster” and to Brucella suis 1330 are displayed in Table Table11 (1, 2, 21, 35, 37, 45). Although the protein sequences that are most similar to the S. meliloti Vir proteins are those of A. tumefaciens Avh and R. etli Vir, the defined functions of the “A cluster” of VirB4/AvhB4 proteins are very different, and thus very similar Vir proteins are involved in very different types of functions.
M. loti R7A and S. meliloti 1021 are the only two plant symbiotic bacteria in which the T4SS has been analyzed (23; this study). In neither case does a mutation in the T4SS (known to be present in only one copy in S. meliloti 1021) render these bacterial species incapable of establishing a functional symbiosis (Table (Table3)3) (6, 8, 16). Therefore, a T4SS may not be a requirement for an α-proteobacterial species to maintain an intracellular infection of a plant host. This is in contrast to the available data for α-proteobacterial, intracellular animal pathogens (Table (Table3).3). Of the two genera (Brucella spp. and Bartonella spp.) in which knockout mutations in a critical T4SS component have been generated, both lose their ability to maintain an intracellular infection of the mammalian host (30, 40). In Bartonella tribocorum and Brucella spp. the T4SS encoded by the virB and virD4 genes is absolutely critical for long-term survival of the bacteria within the host, establishing an intracellular replicative niche, and in Brucella abortus the T4SS has been shown to be required for fusion of the vacuole containing the endocytosed bacteria with the endoplasmic reticulums of infected cells and for preventing fusion of this vacuole with the lysosome (9, 40). All α-proteobacterial, intracellular mammalian pathogens that have been sequenced have at least some components of a T4SS (1, 3, 14, 22, 27, 29, 34, 35). This is also the case for all sequenced α-proteobacterial species that reside in intracellular compartments of insects and nematodes, including Wolbachia (wBm), a mutualist of the nematode Brugia malayi (13, 17, 27, 32, 33, 48, 50) These differences in the host cell invasion requirements of α-proteobacteria suggest that there may be critical differences in the determinants required to survive within a plant host versus an animal host.
We suggest as a possible explanation for the survival of T4SS-deficient rhizobial bacteria within plant cells that these bacteria exploit a peculiarity of the plant endocytic vacuolar sorting pathway. In contrast to mammalian and yeast cells, plants have two possible destinations for endocytosed material. One is the lytic vacuole, which is analogous to the lysosome of mammals, and the other is the nonlytic protein storage vacuole (39, 47). Though the infection droplet in which rhizobial bacteria are endocytosed by plant root cells has not yet been linked by protein markers to protein storage vacuoles or to nonlytic prevacuolar compartments, it is possible that bacteria residing in intracellular membrane-bound compartments evade lytic degradation by different mechanisms in plant cells than in mammalian cells (47). Thus, the fact that the T4SSs of S. meliloti 1021 and M. loti R7A are not required for the intracellular survival of these bacteria raises new questions about the vacuolar identity of the bacterium-containing compartments of plants versus animals.
We are extremely grateful to Brenda Minesinger for her assistance. We also thank Brenda Minesinger and other members of the Walker lab for helpful discussions and critical reading of the manuscript. We also thank Zhenying Liu and Andrew N. Binns (University of Pennsylvania) for providing plasmid pML122 and Anke Becker (University of Bielefeld) for providing the lacZ-Gn cassette construct pAB2001.
This work was supported by NIH grant GM31030 to G.C.W. and by a postdoctoral fellowship from the Fullbright/Spanish Ministry of Education and Science to J.L. G.C.W. is an American Cancer Society Research Professor.
Published ahead of print on 8 December 2006.