|Home | About | Journals | Submit | Contact Us | Français|
With an obligate intracellular lifestyle, Alphaproteobacteria of the order Rickettsiales have inextricably coevolved with their various eukaryotic hosts, resulting in small, reductive genomes and strict dependency on host resources. Unsurprisingly, large portions of Rickettsiales genomes encode proteins involved in transport and secretion. One particular transporter that has garnered recent attention from researchers is the type IV secretion system (T4SS). Homologous to the well-studied archetypal vir T4SS of Agrobacterium tumefaciens, the Rickettsiales vir homolog (rvh) T4SS is characterized primarily by duplication of several of its genes and scattered genomic distribution of all components in several conserved islets. Phylogeny estimation suggests a single event of ancestral acquirement of the rvh T4SS, likely from a nonalphaproteobacterial origin. Bioinformatics analysis of over 30 Rickettsiales genome sequences illustrates a conserved core rvh scaffold (lacking only a virB5 homolog), with lineage-specific diversification of several components (rvhB1, rvhB2, and rvhB9b), likely a result of modifications to cell envelope structure. This coevolution of the rvh T4SS and cell envelope morphology is probably driven by adaptations to various host cells, identifying the transporter as an important target for vaccine development. Despite the genetic intractability of Rickettsiales, recent advancements have been made in the characterization of several components of the rvh T4SS, as well as its putative regulators and substrates. While current data favor a role in effector translocation, functions in DNA uptake and release and/or conjugation cannot at present be ruled out, especially considering that a mechanism for plasmid transfer in Rickettsia spp. has yet to be proposed.
Type IV secretion systems (T4SSs) are macromolecular complexes that transport protein, DNA, and nucleoprotein across the bacterial cell envelope in both Gram-negative and Gram-positive species, as well as some wall-less bacteria and archaea (1, 32). Functioning in naked DNA uptake and release (60), conjugation (80), and the propagation of genomic islands (69), T4SSs are prominent factors in bacterial diversification and are responsible for the horizontal spread of antimicrobial resistance and virulence genes. T4SSs are also used by some species to deliver effector molecules (DNA and/or protein) into eukaryotic host cells (28), a process that facilitates infection and subsequent pathogenesis. It is assumed that all varieties of T4SSs form a channel that spans the cell envelope and culminates in a surface-exposed structure, such as a pilus (Fig. (Fig.1A).1A). Despite this conserved architecture, genetic diversity in a multitude of features, including gene composition and organization, underlies the hundreds of T4SSs identified through genome sequencing. Recently, T4SSs have been classified into four groups: F, P, I, and GI (70). F-T4SSs and P-T4SSs (previously known as type IVA) are widespread systems represented by the archetypes encoded by the F plasmid of Escherichia coli (tra and trb) and the pTi plasmid of Agrobacterium tumefaciens (vir), respectively. I-T4SSs (previously known as type IVB) are typified by the icm/dot system of IncI plasmids, and examples in Legionella spp. and Coxiella burnetii are the best characterized. GI-T4SSs, distinct systems that function in transferring the genomic islands with which they are associated (70, 71), are also widespread and can be further classified into sublineages based on gene content and arrangement (73). The growing diversity of T4SSs will undoubtedly continue to challenge attempts at their classification and the unraveling of their evolutionary origins.
Alphaproteobacteria of the order Rickettsiales are diverse obligate intracellular species with a wide range of eukaryotic hosts (22, 23, 105, 125). Many species within the two well-characterized families, Anaplasmataceae and Rickettsiaceae, pose severe threats to livestock and human health. The agricultural and medical ramifications have resulted in the rapid accumulation of over 30 complete or nearly complete genome sequences from a diverse array of Rickettsiales taxa. Despite the common ancestry (127) and strictly intracellular lifestyles of Rickettsiales, the manner of genome reduction and reliance on host resources vary greatly across lineages (36, 63, 95). While few syntenic regions are found across Rickettsiales genomes (63), a conserved P-T4SS is a particularly definitive feature of these bacteria. Since the completion of sequencing of the first Rickettsiales genome, that of Rickettsia prowazekii (5), a reduced P-T4SS (lacking homologs of virB1, virB2, virB5, and virB7) has been uncovered in all subsequently sequenced genomes, with anomalous duplication of genes homologous to virB4, virB6, virB8, and virB9 suggesting rich functionality and with genes split into multiple islets across the genomes. We recently performed a detailed informatics analysis of the P-T4SS of Rickettsia spp. and concluded that, relative to the canonical vir P-T4SS of A. tumefaciens, this transporter lacks only a homolog of virB5, the gene encoding the minor pilus subunit (55).
In this review, we expand our prior analysis of the Rickettsia T4SS, in which we named this transporter rvh (Rickettsiales vir homolog), to encompass T4SSs of all Rickettsiales (Fig. (Fig.1B).1B). An assumption is made that the acquisition of a P-T4SS was pivotal in the transition from an extracellular to an obligate intracellular lifestyle. We address the nature of duplication of rvh components (rvhB4, rvhB8, and rvhB9) and proliferation of another component (rvhB6) and draw special attention to the components that tend to elude automated genome annotation (rvhB1, rvhB2, and rvhB7). The translocated proteins encoded by the latter genes define the most plastic attributes of the rvh P-T4SS and, coupled with the deletion of a virB5 homolog, imply coevolution of the transporter and the bacterial cell envelope. Despite conservation of the rvh T4SS across the Rickettsiales, there is little information regarding rvh regulation and substrate transport. Learning more about the manner in which the various rvh T4SSs assemble and function in the bacterial outer membrane (OM) may present novel opportunities for vaccine development and drug targeting, and we discuss these possibilities in relation to rvh adaptations to host cell environments.
Previously, it was determined by phylogenetic analysis of P-T4SSs that lateral gene transfer (LGT) has spread these transporters across divergent bacterial lineages and that effector molecule translocation has evolved from the primitive function of conjugation (50). Consistent with this observation, it has been demonstrated previously that the rvh T4SS is related to P-T4SSs from certain Gammaproteobacteria (Legionella spp. and Photobacterium profundum) and Epsilonproteobacteria (Helicobacter pylori, Wolinella succinogenes, and Campylobacter jejuni) (26, 93). Our phylogeny estimation for 47 P-T4SSs based on half of the vir-like components (virB4, virB8 to virB11, and virD4) is in accord with this notion that the rvh P-T4SS was derived from a non-Alphaproteobacteria ancestor (Fig. (Fig.2).2). We recovered the trw P-T4SS, carried on plasmid pXcB of the Gammaproteobacteria species Xanthomonas citri, as the closest xenolog of rvh, as in a previous analysis based on 73 VirB4 and VirB4-like proteins (50). As demonstrated in other studies, the lvh P-T4SSs of Legionella spp., the trb P-T4SS of P. profundum, and the unnamed P-T4SS carried on plasmid pMLa of Mesorhizobium loti are the next closest xenologs of rvh, all having branched from the P-T4SSs of H. pylori, C. jejuni, and W. succinogenes that are involved in DNA uptake and release.
Our analysis, as well as the findings of prior studies (55, 106, 108), supports the vertical transmission of the rvh T4SS after the split of “Candidatus Pelagibacter” from the Rickettsiales ancestor. The most parsimonious recreation of an LGT event that equipped this ancestor with a P-T4SS involves a transporter containing 12 genes (rvhB1 to rvhB11 and rvhD4), with ancestral deletion, recombination, and duplication events creating a modified transporter comprising 18 genes (rvhB1 to rvhB3, rvhB4a and rvhB4b, rvhB6a to rvhB6e, rvhB7, rvhB8a and rvhB8b, rvhB9a and rvhB9b, rvhB10, rvhB11, and rvhD4). The modern genomic distribution of this ancestral operon is fragmented across all sequenced Rickettsiales genomes (Fig. (Fig.3;3; see also Fig. S1 in the supplemental material) and, coupled with lineage-specific gene deletion, proliferation, and truncation events, defines a variable rvh P-T4SS.
For Rickettsia spp., we reported previously that only one ortholog of each of the singly duplicated genes (rvhB4a, rvhB8b, and rvhB9a) adhered to the conserved properties of similar genes in other P-T4SSs (55). Both rvhB4b and rvhB8a encode proteins with similar defects in Rickettsia spp. and the other Rickettsiales, as observed in our prior informatics analysis (see Table S1 in the supplemental material). In contrast, the product of rvhB9b, which in Rickettsia spp. is missing the C-terminal domain, demonstrated previously via nuclear magnetic resonance (NMR) (16), cryo-electron microscopy (52), and crystallography (29) to bind VirB7 in other systems, is full-length in the Anaplasmataceae (see Table S1 in the supplemental material). We had hypothesized (55), given the mirrored arrangements of the rvhB9a-rvhB8a locus (antisense) and the rvhB8b-rvhB9b locus (sense), with rvhB7 at the center of these loci (Fig. (Fig.3),3), that an intrachromosomal recombination event truncated rvhB9b in Rickettsia spp.; however, this pattern does hold throughout the Rickettsiales order. rvhB9b is also truncated in the genomes of Orientia tsutsugamushi strains, yet the rvhB8b-rvhB9b locus is well separated from the rvhB9a-rvhB8a locus in O. tsutsugamushi genomes, as it is in the Anaplasmataceae genomes. The exact manner in which Rickettsiaceae rvhB9b became truncated is unclear; either the entire C islet (rvhB9a rvhB8a rvhB7 rvhB8b rvhB9b rvhB10 rvhB11 rvhD4) exhibits the ancestral gene order and the rvhB9a-rvhB8a-rvhB7 segment became separated outside Rickettsia spp. or a recombination event coupled the duplicate rvhB9 and rvhB8 genes in Rickettsia spp. to the remaining genes within islet C. Nonetheless, a major difference between the T4SSs of Anaplasmataceae and Rickettsiaceae is the presence of two full-length copies of rvhB9 in the former and a truncated rvhB9b in the latter.
The proliferated rvhB6 genes, encoding proteins RvhB6a to RvhB6d, are present in most sequenced Rickettsiales genomes and are arrayed contiguously within one predicted operon (Fig. (Fig.3).3). A fifth gene, rvhB6e, is found only in Rickettsiaceae genomes, and while arrayed with rvhB6a to rvhB6d in the Rickettsia genomes, the gene is well separated from the rvhB6 operon in O. tsutsugamushi genomes. The proteins encoded by rvhB6 are perhaps the most intriguing features of the rvh P-T4SS. While all the proteins contain a complete VirB6/TrbL (PF04610) domain that is comparable to full-length VirB6 and VirB6-like orthologs in other species harboring P-T4SSs, the Rickettsiales open reading frames (ORFs) encode additional flanking regions that lack similarity to any proteins in public databases (55). The nature of these “gangly arms” flanking the VirB6/TrbL domain, coupled with the unusual proliferation of the corresponding genes in the highly reduced Rickettsiales genomes, in which gene duplication is atypical, brings attention to these curious proteins. VirB6/TrbL proteins resemble ComEC channel proteins, which are involved in the uptake of environmental DNA (42). It is possible that RvhB6 proteins play a role in DNA import/export, with the various duplications forming multiple diverse channels that maximize the potential for LGT in environments (such as those in protozoa and macrophages) with high rates of congener contact. Given that VirB6 in A. tumefaciens has been demonstrated previously to facilitate substrate transfer (66) yet is not a component of the core P-T4SS channel (52), the multiple RvhB6 proteins may equip Rickettsiales with substrate selectivity in the various host environments encountered by these obligate intracellular bacteria. However, the strong selection keeping the rvhB6 loci contiguous, presumably within tightly regulated operons (55), suggests that posttranscriptional regulation of rvhB6 genes would be necessary for environment-specific utilization of the various duplicate proteins in the bacteria. This idea is supported by the rvhB6e gene of O. tsutsugamushi strain Ikeda, which is truncated and well separated from the remaining rvhB6 genes in islet A (Fig. (Fig.3).3). However, all four rvhB6 genes of Ehrlichia chaffeensis are coexpressed in tick and human cells, and various RvhB6-RvhB6 and RvhB6-RvhB9 interactions suggest more than one RvhB6 protein may assemble at the inner membrane (IM) channel region of the rvh T4SS (8). Nonetheless, given that two rickettsial RvhB6-encoding genes (rvhB6d in R. bellii strain OSU 85 389 and rvhB6e in R. massiliae) (see Table S2 in the supplemental material) have undergone pseudogenization, not all copies of these genes may be functional despite the near global conservation across the Rickettsiales.
VirB2 and related proteins constitute the major subunit of the Agrobacterium T-pilus and related structures (77, 79). The signal sequences of these proteins are processed in the IM, with subsequent species-specific processing of the C and/or N termini (44, 45, 72, 78, 80, 92, 112). This role in T-pilus formation is supported by findings from various studies demonstrating the presence of VirB2 in complexes with VirB5 (the minor subunit of the T-pilus) and VirB7 (17, 67, 76, 86, 115, 132). In particular, the VirB2-VirB5 pilus complex is dependent on the periplasmic interaction between VirB4 and VirB8, which promotes the formation of extracellular pili (132). Unlike pTi-encoded VirB5 (VirB5Ti), VirB2Ti is critical for substrate transfer through the P-T4SS scaffold (27, 64). Polymers (115) of these proteins probably span the entire periplasm (32), as made evident by the presence of at least two predicted transmembrane-spanning regions in nearly all available sequences (55). Recently, it was hypothesized that VirB2 polymers might “snake” along the VirB10 antenna domain, lining the outside of the entire P-T4SS channel from the IM to the OM (65) and binding VirB5 at the OM, with continual polymerization as a T-pilus in species that protract these extracellular appendages. However, it cannot be ruled out that VirB2 polymerizes within the chamber of the VirB7-VirB9-VirB10 core complex prior to binding VirB5 and extending extracellularly as the T-pilus (51), a hypothesis supported by the observation that the transmembrane helices within the TraF/VirB10 crystal appear to be caved in (in the absence of VirB2) (29).
Several prior reviews of T4SSs reported a lack of genes encoding either the major or the minor pilin subunit within Rickettsiales genomes. However, our recent bioinformatics analysis of 13 Rickettsia genomes identified a candidate VirB2-encoding ORF, rvhB2, predicted to be a single transcriptional unit well separated from other T4SS genes (55). In certain Anaplasmataceae genomes, a duplicated ORF previously named orfX (24, 94), associated with the major surface protein 2 (msp2) superfamily, has also recently been designated a putative virB2 homolog (23, 99). Here, we draw attention to recently identified ORFs in the O. tsutsugamushi and Anaplasmataceae genomes which suggest that VirB2 proteins are likely to be an essential component of the rvh P-T4SS (see references 35 and 114 and Fig. S2 in the supplemental material). Interestingly, unlike the genomes of Rickettsia spp., the remaining Rickettsiales genomes have rvhB2 present at least twice and the genomes of Wolbachia, Anaplasma, and Ehrlichia species show proliferation of this gene (yielding three or more paralogs) (Fig. (Fig.3).3). Together with a closely related ORF named orfY, a total of 22 candidate rvhB2 sequences in the Anaplasma marginale genome were identified (data not shown), exemplifying the operation of selection on the retention of numerous rvhB2 paralogs in the derived Anaplasmataceae. Like that of other duplicate rvh genes in Rickettsiales, the expression of rvhB2 paralogs may be specific to the host environment. Supporting this hypothesis, the rvhB2 paralogs of Anaplasma phagocytophilum are differentially expressed in tick and mammalian cell cultures (99). Alternatively, based on its probable secretion to the OM and likely exposure to the host immune system, the major pilin component may have become co-opted into a diverse antigen family that increases the chance of host immune avoidance, a role other duplicate genes and functional pseudogenes in the msp2 superfamily have (23). Supporting this possibility, the majority of rvhB2 genes in A. marginale are arrayed with msp2 and msp3 genes and their associated functional pseudogenes (25). Regardless of the provisional role that RvhB2 plays in rvh assembly and function, experimental evidence suggests surface exposure of the protein. For example, in A. marginale, RvhB2 induces a T-cell response in cattle as part of a protective bacterial membrane vaccine (89, 119), and the two orthologs in Neorickettsia risticii are coexpressed by an operon and localized predominantly at the poles, where they form focal complexes (85).
T4SSs typically encode small lipoproteins under 100 amino acids that are secreted to the periplasm and are essential for substrate transfer. In A. tumefaciens, the small lipoprotein VirB7 primarily binds and stabilizes VirB9 in the OM (13, 47, 118) via an essential disulfide bond (4, 11, 62, 132). Structural studies of this interaction in the P-T4SS encoded by the plasmid pKM101 of E. coli illustrate that a disulfide bond is not universal (16, 52), suggesting that other protein interactions in VirB7-VirB9 heterodimer formation can suffice (9). Along with VirB10 and VirB9, VirB7 is a component of the core T4SS complex, and it is inserted into the OM with the C-terminal domains of VirB9 and VirB10 (52). Prior comprehensive studies of type IV secretion have noted that VirB7 and related proteins are not always encoded within T4SSs (26, 93). However, given their small size, it is likely that many ORFs encoding these lipoproteins are not annotated by automated gene prediction methods, especially if the genomic positions of these ORFs are not arrayed with those of other T4SS genes.
We recently identified a putative VirB7 homolog, RvhB7, encoded within all sequenced genomes of Rickettsia spp. (55). rvhB7 is located upstream of the rvhB8b-rvhB9b-rvhB10-rvhB11-rvhD4 cluster and is flanked by the rvhB9a-rvhB8a locus, which is carried on the opposite strand (Fig. (Fig.3).3). Bioinformatics analysis revealed that RvhB7 contains at least one Cys residue in addition to the conserved lipoprocessing Cys that is characteristic of all VirB7-like proteins, suggesting that RvhB7 is similar to VirB7Ti and other T4SS lipoproteins that may bind the C-terminal domain of VirB9 via a disulfide bridge. Other than these conserved features, RvhB7 proteins contain a candidate P(ILV)NK motif in the C-terminal region that is typical of most VirB7-like sequences (16). Additionally, a conserved sequence, (KI)KSP, directly flanking the second conserved Cys on the N-terminal side was found to be a feature shared only by the RvhB7 proteins of the Rickettsia spp. and the ComB7 proteins of H. pylori and C. jejuni, illustrating the possibility that rvh T4SSs descended from DNA competence systems (55).
Herein, we expand our analysis of putative rickettsial T4SS lipoproteins and illustrate that these molecules are likely to be part of the conserved rvh core as in other systems (Fig. (Fig.4).4). RvhB7 sequences of Rickettsia spp. were used to identify hypothetical proteins encoded in the genomes of A. marginale (hypothetical protein AM306) and the Wolbachia symbiont of Culex quinquefasciatus (hypothetical protein WPa_0823) that share the conserved characteristics of RvhB7 proteins. More refined tools revealed putative ORFs in most of the remaining sequenced Anaplasmataceae genomes that are currently not annotated as genes. Each of the protein sequences analyzed contains a predicted lipoprocessing site (Cys), a second conserved Cys, and several candidate P(ILV)NK motifs. The latter feature continues to become less conserved with the addition of more diverse VirB7-like sequences, and even simplifying the motif to PhN+, where h and + represent nonpolar and positively charged residues, respectively, still does not encompass the diversity in the C-terminal regions of these sequences or facilitate approaches to multiple-sequence alignment. The addition of these putative Anaplasmataceae sequences refines the conserved region shared by RvhB7 and the ComB7 proteins of H. pylori and C. jejuni to (K/R)SP, further supporting the evolution of rvh from competence systems. Additionally, we identified the (K/R)SP motif in the TrwH protein of X. citri, which is consistent with the phylogenetic position of the trw T4SS of X. citri as the closest xenolog of rvh (Fig. (Fig.2).2). Importantly, a recent study reports that the putative RvhB7 protein in A. marginale OM vaccine preparations is immunogenic for cattle vaccinees expressing several common major histocompatibility complex class II haplotypes (119). Altogether, these data strongly imply that RvhB7 is part of the conserved core rvh T4SS (RvhB7, RvhB9, and RvhB10) and is possibly surface exposed at the OM.
Efficient T4SS transporter assembly across the cell envelope typically is associated with local disruption of peptidoglycan (PG) (Fig. (Fig.1).1). Descending from the free-living “Candidatus Pelagibacter,” Rickettsiales have not been found to encode a complete pathway for the synthesis of PG that is independent of host resources (see Fig. S3 in the supplemental material). PG from R. prowazekii similar to PGs from other Gram-negative species has been purified and demonstrated to incorporate d-alanine (103), and it is likely that a murein layer is synthesized in all Rickettsia spp., starting from host reserves of fructose-6-phosphate and/or glucosamine-6-phosphate. Prior studies have failed to detect PG in O. tsutsugamushi (3, 117), and a lack of enzymes able to convert fructose-6-phosphate to UDP-N-acetylglucosamine suggests the absence of PG in the cell envelope. Furthermore, alr, which encodes the enzyme that converts l-alanine to d-alanine, is deleted from O. tsutsugamushi genomes, as it is from all Anaplasmataceae genomes. Despite this finding, genes for most of the enzymes involved in amino sugar metabolism are present in the O. tsutsugamushi genomes (31, 98), as well as in A. marginale and Wolbachia genomes (49, 130). The same is true for genes encoding the enzymes responsible for the synthesis of lipid I and lipid II and the transport of anhydromuropeptides to the periplasm via the MurJ flippase (111). Genes encoding enzymes responsible for modification (i.e., transpeptidation and transglycosylation) of PG in the periplasm are highly conserved in Rickettsia spp. but found sporadically in the remaining Rickettsiales genomes. Outside the Rickettsiaceae, only A. marginale contains slt, the gene encoding the soluble lytic transglycosylase (LT) responsible for the excision of PG subunits from the murein layer. Rickettsia spp. encode the most complete pathway for recycling PG, with atypical proliferation of the IM AmpG permeases involved in PG subunit import to the cytoplasm (discussed below). Thus, from a genomics perspective, Rickettsia spp. and A. marginale are the only Rickettsiales that likely incorporate PG into their cell envelopes. The lack of rigid cell envelopes in Anaplasmataceae aside from A. marginale supports this viewpoint (33, 84, 109).
It is common for type II, III, and IV secretion systems, as well as DNA competence systems and bacteriophages, to encode LTs that hydrolyze PG (14, 40, 74, 81, 96). These specialized LTs (75) facilitate the local disruption of PG, allowing for efficient transporter assembly across the entire cell envelope or for host cell penetration in the case of bacteriophages. While typically much smaller than Slt and related PG autolysins, specialized LTs contain similar lysozyme-like folds and belong to an ancient glycohydrolase superfamily that includes plant chitinases, bacterial chitosanases, goose and hen g-type lysozymes, and phage T4 lysozymes (110). One of the best-characterized specialized LTs is VirB1Ti, which aside from its N-terminal lysozyme domain contains a processed C-terminal region that is secreted extracellularly and possibly involved in T-pilus formation (10, 133). While VirB1Ti is critical for T-pilus biogenesis, VirB1Ti mutants weaken but do not entirely abolish substrate transfer (15, 18, 53, 77), suggesting that Slt and/or other PG autolysins suffice in degrading PG for transporter assembly. VirB1-like proteins are also nonessential in other systems (14, 38, 39, 129); however, in the cag P-T4SS of H. pylori, the specialized LT is critical for both CagA translocation/phosphorylation and interleukin-8 induction in host cells (48).
We recently identified a virB1 homolog (rvhB1) that is conserved in all sequenced genomes of Rickettsia (55), supporting the likelihood that PG is a component of Rickettsia species cell envelopes. While rvhB1 is located independently of the other rvh genes in these genomes (Fig. (Fig.3),3), there is strong in silico evidence supporting RvhB1 as a specialized LT (55). Informatics data suggest that no RvhB1 homologs are encoded in the remaining Rickettsiales genomes (data not shown). Thus, if PG is synthesized in A. marginale and/or Wolbachia spp., the manner in which the T4SS scaffold assembles and spans the periplasm is likely to be different from that in Rickettsia spp.
The vir P-T4SS of A. tumefaciens and the icm/dot I-T4SS of Legionella pneumophila are regulated by two-component systems (54, 128, 134), suggesting tight correlation of T4SS gene expression and function. Other single genes have been demonstrated previously to control the expression of the vir loci of Brucella spp. (37, 41, 116). Regarding several intracellular species, it has been demonstrated that host conditions upregulate T4SS operons (21, 30, 113). Clustering of T4SS genes into one or a few operons undoubtedly enhances their coordinated regulation (32, 82). However, the conserved rvh T4SS is scattered in small islets throughout the genome in all lineages (Fig. (Fig.3),3), and in Rickettsia genomes, the 18 genes can be grouped into five islets (55). Counterbalancing this architectural anomaly, a single T4SS transcriptional regulator, ecxR (ECH_0795), has been identified in E. chaffeensis and induces coordinate expression of several of the rvh islets (30). Informatics data indicate that ecxR homologs are present in all genomes of the Anaplasmataceae but are not present in the Rickettsiaceae (see Table S3 in the supplemental material). Previously, we identified two conserved genes in Rickettsia spp. that are carried within predicted rvh operons. A RelA/SpoT-encoding gene adjacent to rvhD4 is of interest given that a stringent response protein similar to RelA/SpoT, Rsh, regulates the vir loci of Brucella spp. (41). A second gene encoding a conserved hypothetical protein is immediately downstream of the rvhB1 gene in all sequenced genomes, possibly having a role associated with type IV secretion. Undoubtedly, there will be further characterized rvh regulators given the need to coordinately express the scattered rvh islets.
Despite the rvh P-T4SS's being one of the features most conserved across Rickettsiales genomes, the function of the rvh P-T4SS is poorly characterized (55). The AnkA protein of A. phagocytophilum, which penetrates neutrophil nuclei and is essential for host infection (83, 104), has been described as the first rvh T4SS substrate and is able to be secreted via the vir P-T4SS of A. tumefaciens (83). However, no other protein or DNA molecules have been identified as substrates of the rvh T4SS, and the exact manner in which substrates are presented to and secreted by the transporter are unknown. For Rickettsia spp., two genes have garnered attention for possibly encoding rvh substrates. ralF, a gene encoding a Sec7 domain-containing protein, is known in prokaryotes only from Rickettsia spp. and Legionella spp. (34), and in L. pneumophila the protein is an I-T4SS effector that functions as a guanine nucleotide exchange factor in the recruitment of the ADP-ribosylation factor to occupied phagosomes (97). While the precise role of RalF in Legionella pathogenesis is unknown, an analogous function associated with phagosomal modification in Rickettsia spp. is unlikely given the immediate lysis of the phagosome upon host cell invasion. Furthermore, pseudogenization has eliminated nearly the entire ralF ORF in all sequenced genomes of spotted fever group rickettsiae. The second putative rvh effector is RickA, a protein considered to activate host Arp2/Arp3 complexes, resulting in actin nucleation (58, 68), which permits intercellular spread of some rickettsiae (61). Based on informatics data, which predict neither membrane nor periplasmic association of RickA (data not shown), and the fact that the protein is localized to the bacterial surface (59), it was hypothesized that RickA is secreted via the T4SS (58). Like ralF, the RickA gene is not present in all sequenced genomes of Rickettsia spp.; hence, if either or both of these candidates are true T4SS effectors, the rvh T4SS would secrete different substrates in various species/strains of Rickettsia, possibly contributing to lineage-specific pathogenicity.
It was speculated previously that the rvh T4SS must function in virulence factor secretion versus plasmid transfer because not all sequenced rickettsia genomes contain plasmids (102). Interestingly, the same investigation not only uncovered the first case of a plasmid system in a rickettsia (R. felis) but also provided electron microscopy images of pilus-like structures that were assumed to be associated with conjugation (102). As there is no complete set of genes in the R. felis genome encoding type II or type IV pili (data not shown) and the absence of a virB5 homolog would presumably prevent the formation of a T-pilus (as discussed above), the exact genetic architecture underlying these extracellular appendages in R. felis remains unknown. Similar pilus-like structures were observed in R. bellii strain RML 369-C (101) and R. massiliae (19), which both carry a nearly full set of genes related to the tra-trb operon of the F plasmid of E. coli, despite only the latter's harboring a plasmid system. Other rickettsiae harboring plasmids (e.g., R. africae, “R. monacensis,” and R. peacockii) do not carry full tra-trb-like operons, and no pilus-like structures have been reported. Thus, the significance of a tra-trb operon in relation to extracellular appendages and their possible role in the conjugation of rickettsial plasmids remains quite nebulous. What is clear is that a large mobile genetic element, which carries an F-T4SS highly similar to the F plasmid tra-trb operon, is detectable as products of pseudogenization across the sequenced Rickettsiales genomes (data not shown) and is highly proliferated in the genomes of O. tsutsugamushi strains (31, 98). Similar conjugation genes have also been detected in various rickettsiae associated with arthropods with no known association with vertebrates (124). Recent sequencing of the rickettsial endosymbiont of Ixodes scapularis (REIS) has revealed the largest number of conjugation genes within a Rickettsia genome, suggesting that ancient conjugative plasmids have propagated mobile elements across the Rickettsiales. Strong reductive evolution has eliminated the majority of these elements in most genomes, in particular a complete F-T4SS likely to be essential for conjugative transfer of plasmids. The ability of the rvh P-T4SS to function in conjugation cannot be ruled out, but the absence of any identified plasmids or pilus-like structures in the Anaplasmataceae argues strongly against a role in conjugation. However, a function in naked DNA uptake and release from host environments, particularly given the close phylogenetic relatedness of rvh to T4SSs that are involved in DNA competence (Fig. (Fig.2),2), cannot be overlooked (55). Rickettsiaceae may benefit from scavenging nucleotides upon DNA uptake since genes involved in de novo synthesis of nucleotides have been deleted relative to the genomes of Anaplasmataceae, which due to their vacuole-enclosed lifestyle do not have access to cytoplasmic nucleotides and have thus retained genes involved in purine and pyrimidine biosynthesis (23).
Hundreds of sequenced bacterial genomes provide a foundation for understanding the diversity of T4SSs that define the lifestyles of many pathogens, as well as contribute to the continuum of speciation through the transfer of components of the bacterial mobile gene pool. While informatics tools and laboratory studies are useful for unveiling T4SS architecture and function, evolution has done the crucial experiments, and many T4SSs can be seen as the end products of millions of years of bacterial coevolution with various host and vector cells. This information, combined with a wealth of experimental data that define the archetypal vir P-T4SS of A. tumefaciens (Fig. (Fig.5A),5A), allows for the observation of structural and functional diversification across a wide range of bacterial species. For example, in relation to the vir P-T4SS, the ptl P-T4SS of Bordetella pertussis has undergone three innovations that likely correlate with the secretion of its sole substrate, the pertussis holotoxin (PT) (Fig. (Fig.5B).5B). First, a virB1 homolog is absent in the ptl system, yet given the presence of PG in B. pertussis, a mechanism for local degradation of PG is required. This task is accomplished by the fusion of a glycohydrolase domain to the N-terminal region of PtlE (a VirB8 homolog) that has demonstrated peptidoglycanase activity in both B. pertussis and E. coli (107). Thus, because VirB8 proteins are bitopic in the IM, this N-terminal glycohydrolase domain is located in the periplasm, allowing for simultaneous degradation of PG and T4SS channel assembly. Second, a homolog of the type IV coupling protein (T4CP), VirD4, is deleted in the ptl system. This is explained by the sec-dependent secretion of PT subunits to the periplasm (88, 100), with holotoxin assembly likely driving the formation of the core Ptl scaffold around PT prior to secretion (121). Thus, a T4CP would not be needed for entry of substrates from the cytoplasm to the periplasm, with the energetics generated from the remaining two IM ATPases, VirB4 (PtlC) and VirB11 (PtlH), enough to drive translocation of PT out of the bacterial cell. Lastly, the lack of a virB5 homolog in the ptl system correlates with the lack of an observed T4SS pilus in B. pertussis (126), as host cell contact is not required for secretion of PT.
Similar comparisons between the T4SSs of members of the Rickettsiales and the vir system of A. tumefaciens can be made for inferring structural and functional diversification of the rvh transporter (Fig. 5C and D). Like the B. pertussis genome, all Rickettsiales genomes lack a virB5 homolog, suggesting that rvh-mediated secretion occurs in the absence of a T-pilus-like structure. As recent studies identify VirB5 as an adhesin involved in host cell recognition (2, 7, 131), a minor pilin homolog (and a T-pilus) would be unnecessary in Rickettsiales, as substrates would be directly secreted to and/or imported from the host environment (55). Thus, RvhB2 polymerization across the periplasm should terminate near the OM in all Rickettsiales species. As Rickettsia spp. are the only Rickettsiales lineage to synthesize lipopolysaccharide (LPS) (data not shown), RvhB2 may be highly surface exposed in the remaining Rickettsiales. This suggestion is supported by an RvhB2-induced T-cell response in cattle immunized with an A. marginale OM fraction (89). The lack of LPS may also expose a portion of the outer cap of the core T4SS, as antibodies specific for RvhB9 have been detected in dogs infected with Ehrlichia canis (46) and cattle infected with A. marginale (6, 122). Furthermore, both RvhB9 and RvhB10 elicit robust antibody and T-cell responses from cattle immunized with a protective A. marginale OM fraction (89-91). The loss of LPS synthesis in the Anaplasmataceae may have lead to greater exposure of the T4SS at the OM and may account for RvhB2 duplication and proliferation as a consequence of host cell immune system avoidance. In contrast to Rickettsiaceae genomes, all sequenced Anaplasmataceae genomes have full-length RvhB9 gene homologs, which may add to antigenic complexity in the OM portion of the T4SS. While the cell envelope composition of O. tsutsugamushi differs greatly from that of Rickettsia spp. and does not contain LPS (117), prediction of surface exposure of the T4SS is difficult since only one genome (that of strain Ikeda) contains the VirB2 paralog genes (Fig. (Fig.3).3). Nonetheless, as VirB9 is predicted to be surface exposed in other systems (16, 64), it is probable that some regions of the rvh transporter are surface exposed at the OM, thus making it worthy for exploration as a vaccine target.
It has been hypothesized previously that species of Rickettsiales that invade vertebrate immune cells would benefit from a lack of PG synthesis (63). This proposal is consistent with evidence from other systems that host receptors of PG, nucleotide-binding oligomerization domain 1 (Nod1) and Nod2 proteins, detect by-products of PG degradation via inefficient anhydromuropeptide recycling (20, 120). As all members of the Anaplasmataceae replicate within intracellular vacuoles (43), discarded PG fragments may still be undetected by host cells in species that potentially synthesize PG. For free-living Rickettsiaceae, either PG synthesis does not occur (as in the case of O. tsutsugamushi) or a strategy may exist for “hiding” PG fragment release from immune cells. As discussed above, Rickettsia spp. contain many genes involved in the degradation and recycling of PG (see Fig. S3 in the supplemental material). Exceptionally, all sequenced genomes of Rickettsia spp. contain three to four copies of ampG, which encodes a permease involved in the import of PG monomers from the periplasm to the cytoplasm. As such genes are typically present only once in bacterial genomes, ampG proliferation in Rickettsia spp. may hint at increased necessity to recycle PG subunits rather then shed them freely into the host cytoplasm. The presence of a virB1 homolog in only the genomes of Rickettsia spp. suggests that PG fragments are likely to be released during the assembly of the rvh T4SS and that this process of cell envelope rearrangement may make various components of the T4SS scaffold vulnerable to the host immune response. Furthermore, the processing of the C-terminal region of VirB1Ti (VirB1*) (10) and its subsequent extracellular secretion and role in T-pilus formation (87, 133) are likely not specific to the A. tumefaciens vir T4SS. Processed VirB1 products in the cell lysate from Brucella abortus were identified previously (39), and informatics data suggest that a conserved Ala is the likely cleavage site in many VirB1 homologs, with all predicted VirB1* sequences containing tracts of repeated residues, particularly Pro-rich tracts (55). Across 13 Rickettsia spp., rvhB1 has the highest average number of codons evolving under positive selection among the 18 rvh genes (55), suggesting possible coevolution with host cell components. Nonetheless, RvhB1, as well as a putative RvhB1* form, may pose novel vaccine targets specific for species of Rickettsia.
Despite the genetic intractability of Rickettsiales as obligate intracellular bacteria, advances in understanding the mechanisms involved in the pathogenicity of these species are being made (12, 36, 123). This work outlines major achievements in the identification and characterization of components of the rvh T4SS, as well as its putative regulators and substrates. Past and present bioinformatics approaches have greatly facilitated our understanding of the genetic architecture of the rvh scaffold, and experimental evidence has identified several promising vaccine targets. Our synopsis here suggests a single event of inheritance of the rvh T4SS in the Rickettsiales progenitor, with lineage-specific diversification of rvh components likely a result of modifications to cell envelope structure. This coevolution of the rvh T4SS and cell envelope structure is likely driven by adaptations to various host cells and thus identifies the transporter as an important target for vaccine development. While current data favor a role in effector translocation, functions in DNA uptake and release and/or conjugation cannot at present be ruled out, especially considering that no mechanism for plasmid transfer in Rickettsia spp. has yet been proposed. Furthermore, the genomes of several attenuated strains of Rickettsia (R. prowazekii Madrid E and R. rickettsii Iowa) and species with no known pathogenicity in arthropod or vertebrate cells (REIS and R. peacockii Rustic) carry most or all of the rvh genes. Taking this finding into consideration, we expect major advances in the near future regarding knowledge about the rvh T4SS and its potential involvement in host disease.
Studies by Y. Rikihisa and M. Lin (Curr. Opin. Microbiol. 13:59-66, 2010) and H. Niu, V. Kozjak-Pavlovic, T. Rudel, and Y. Rikihisa (PLoS Pathog. 19:e1000774, 2010) were published during the production of this paper. They add substantial information regarding rvh substrates of the Anaplasmataceae.
The project described herein was supported by award numbers R01AI017828 and R01AI59118 (to A.F.A.) and R01AI053692 (to W.C.B.) from the National Institute of Allergy and Infectious Diseases (NIAID) and by funding through NIAID contract HHSN266200400035C to B.W.S.
The content of this paper is the responsibility solely of the authors and does not necessarily represent the official views of the NIAID or the National Institutes of Health.
We are grateful to Jim Kaper (University of Maryland) for the invitation to contribute this review. We thank Nick Carbonetti (University of Maryland) for helpful discussion on B. pertussis type IV secretion, Julie Hotopp (University of Maryland) for insight on Anaplasmataceae genomics, and Guy Palmer (Washington State University) and Steve Dumler (Johns Hopkins) for discussions on Anaplasmataceae biology. We are appreciative to members of the Azad lab and the Cyberinfrastructure Group (VBI) for critical advice during the completion of this work.
Joseph J. Gillespie was born and raised in suburban Philadelphia, PA. In 1998 he obtained his B.S. degree from Widener University, studying the effects of controlled fire on arthropod populations in the New Jersey Pine Barrens. He continued to study arthropods in his M.S. (University of Delaware) and Ph.D. (Texas A&M University) programs, specializing in molecular evolution and phylogenetics. In 2006 he was hired by the Virginia Bioinformatics Institute (VBI) at Virginia Tech to study the bioinformatics of arthropod-borne bacteria. Since then he has developed an interest in the biology of obligate intracellular bacteria, especially rickettsiae. Of particular interest is the manner in which these bacteria coevolve with their eukaryotic hosts. He has remained with VBI as a senior research scientist and is also a visiting scientist in the Department of Microbiology and Immunology at the University of Maryland School of Medicine. He resides with his family in Maryland.
Kelly A. Brayton was born in Akron, OH, and grew up in India, Malaysia, Turkey, Greece, Lebanon, and Texas. She holds a B.A. in biology from Texas A&M and a Ph.D. in biochemistry from Purdue University. She started studying hemoparasitic diseases during her postdoctoral stint at the Onderstepoort Veterinary Institute in South Africa. She is currently an associate professor of microbial genomics in the Department of Veterinary Microbiology and Pathology, School for Global Animal Health, Washington State University, where she has developed a genomics program for veterinary pathogens. She led the sequencing efforts that resulted in genome sequences for the cattle pathogens A. marginale and Babesia bovis, among others. The availability of these sequences has catalyzed research on these organisms—allowing Kelly and her colleagues to pursue research on mechanisms of immune evasion, persistence, virulence, and transmission and vaccine development.
Kelly P. Williams (Virginia Bioinformatics Institute) has broad interest in bacterial molecular biology and evolution, with special interests in RNAs, genomic islands, and phylogeny. His Ph.D. work (at the University of California, San Diego) was on the biochemistry of bacteriophage transcription. Postdoctoral research during appointments at the Salk Institute, the Consiglio Nazionale delle Ricerche (Rome, Italy), and the Whitehead Institute employed in vitro selection to explore diverse RNA functions. His laboratory at Indiana University focused on the mechanism and evolution of bacterial tmRNA.
Marco A. Quevedo Diaz was born in Lima, Peru. He gained his M.S. degree in microbiology from the Faculty of Natural Sciences, Comenius University, in Bratislava, Slovakia, studying pdr genes in Saccharomyces cerevisiae. From 1998 to 2003, he was enrolled in the Ph.D. program in the Department of Rickettsiology at the Institute of Virology, Slovak Academy of Science, studying Coxiella burnetii phase variation and genetic transformation. Since that time, he has been fascinated with the biology of obligate intracellular bacteria, particularly rickettsiae and their interaction with eukaryotic hosts. Currently, he is at the Department of Microbiology and Immunology, School of Medicine, University of Maryland.
Wendy C. Brown was born and raised in upstate New York. She obtained her B.A. degree in microbiology from Smith College and M.P.H. and Ph.D. degrees from Yale University, studying infectious disease epidemiology and T-cell immunology. She transitioned to research on tick-borne pathogens at the International Laboratory for Research on Animal Diseases in Nairobi, Kenya, where she studied vaccine development for theileriosis, and then went to Texas A&M University (as an associate professor), where she studied T-cell responses to B. bovis, and finally to Washington State University, where she has also worked on T-cell immunity to the rickettsial pathogen A. marginale. There, she is a Regents Professor in the Department of Veterinary Microbiology and Pathology and the School for Global Animal Health. Of particular interest is targeting outer membrane protein complexes, specifically the bacterial type IV secretion system, for vaccine development. She resides with her family in the Palouse.
Abdu F. Azad is a professor of microbiology and immunology at the University of Maryland School of Medicine, Baltimore. He obtained his Ph.D. from the Johns Hopkins University School of Public Health, investigating the biology of a liver nematode, Capillaria hepatica, in Norway rats. His diverse scientific interests started with the epidemiology of human intestinal parasites, flea taxonomy (he described five new species), and the ecology and natural history of mammal-borne pathogens in Asia and Africa. His current research is focused primarily on the biology of arthropod-borne rickettsial pathogens, particularly understanding how these bacterial agents with reduced genomes cause infection and disease. Additionally, he has continued his long-term interest in investigating the genetic and molecular bases of preerythrocytic stages of malaria parasites in protective immunity. Aside from his research, he considers training as his major professional endeavor and is very proud of the scientific accomplishments of his former students and postdoctoral fellows.
Bruno W. Sobral was born in Brazil. His undergraduate education was in agricultural engineering and his Ph.D. was in genetics at Iowa State University, with postdoctoral work in molecular evolution. After starting his own research group at the California Institute of Biological Research, La Jolla, in 1991, he went on to be the vice president of scientific programs at the National Center for Genome Resources in Santa Fe, NM. In 2000 he started the Virginia Bioinformatics Institute (VBI) at Virginia Tech as the founding executive and scientific director and as a professor in the Department of Plant Pathology, Physiology, and Weed Science. His Cyberinfrastructure Section at VBI has focused on bioinformatics, computational biology, and informatics-based approaches to infectious disease research. He is particularly interested in and focused on transdisciplinary approaches to research and development and currently is a professor and the director of the Cyberinfrastructure Section.
Editor: H. L. Andrews-Polymenis
Published ahead of print on 22 February 2010.
†Supplemental material for this article may be found at http://iai.asm.org/.
‡The authors have paid a fee to allow immediate free access to this article.