The genome sequence of Rickettsia felis revealed a number of rickettsial genetic anomalies that likely contribute not only to a large genome size relative to other rickettsiae, but also to phenotypic oddities that have confounded the categorization of R. felis as either typhus group (TG) or spotted fever group (SFG) rickettsiae. Most intriguing was the first report from rickettsiae of a conjugative plasmid (pRF) that contains 68 putative open reading frames, several of which are predicted to encode proteins with high similarity to conjugative machinery in other plasmid-containing bacteria.
Using phylogeny estimation, we determined the mode of inheritance of pRF genes relative to conserved rickettsial chromosomal genes. Phylogenies of chromosomal genes were in agreement with other published rickettsial trees. However, phylogenies including pRF genes yielded different topologies and suggest a close relationship between pRF and ancestral group (AG) rickettsiae, including the recently completed genome of R. bellii str. RML369-C. This relatedness is further supported by the distribution of pRF genes across other rickettsiae, as 10 pRF genes (or inactive derivatives) also occur in AG (but not SFG) rickettsiae, with five of these genes characteristic of typical plasmids. Detailed characterization of pRF genes resulted in two novel findings: the identification of oriV and replication termination regions, and the likelihood that a second proposed plasmid, pRFδ, is an artifact of the original genome assembly.
Altogether, we propose a new rickettsial classification scheme with the addition of a fourth lineage, transitional group (TRG) rickettsiae, that is unique from TG and SFG rickettsiae and harbors genes from possible exchanges with AG rickettsiae via conjugation. We offer insight into the evolution of a plastic plasmid system in rickettsiae, including the role plasmids may have played in the acquirement of virulence traits in pathogenic strains, and the likely origin of plasmids within the rickettsial tree.
Citation: Gillespie JJ, Beier MS, Rahman MS, Ammerman NC, Shallom JM, et al. (2007) Plasmids and Rickettsial Evolution: Insight from Rickettsia felis. PLoS ONE 2(3): e266. doi:10.1371/journal.pone.0000266
Academic Editor: Berend Snel, Utrecht University, Netherlands
Received: December 29, 2006; Accepted: February 8, 2007; Published: March 7, 2007
Copyright: © 2007 Gillespie et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work is funded through NIH grants AI59118 and AI17828 to AFA and NIAID contract HHSN266200400035C to BSS.
Competing interests: The authors have declared that no competing interests exist.
All members of the genus Rickettsia (Rickettsiales; Rickettsiaceae) are obligate intracellular parasites of eukaryotes , with some species that are pathogenic and are known to cause harmful diseases in humans, e.g., R. prowazekii, the causative agent of epidemic typhus and R. rickettsii, the agent of Rocky Mountain spotted fever . Some rickettsiae are important both as emerging pathogens  and as selected agents for the development of bioweapons . Given the diversity of Rickettsia spp. ,  and their associated pathologies, a sound understanding of the evolutionary relationships of these diverse bacteria is essential for species-level diagnostics and potential vaccine targeting. The system of classification of rickettsiae has traditionally grouped Rickettsia spp. into the spotted fever (SFG) and typhus group (TG) rickettsiae  with R. bellii considered ancestral to both groups , . R. canadensis was subsequently added to this ancestral group (AG) rickettsiae . Recent phylogenies based on various molecular markers have resulted in conflicting tree topologies [e.g., 10,11,12,13]; however, when AG rickettsiae are included they are nearly always basal to the remaining TG and SFG rickettsiae clades , , . Moreover, only with the inclusion of AG rickettsiae (for rooting the analyses) is a unique lineage recovered consisting of R. akari and R. felis , , , as well as R. australis and some other unidentified rickettsiae from booklouse (Liposcelis sp.) and parasitic wasp (Neochrysocharis sp.) hosts . This clade, referred to hereafter as transitional group (TRG) rickettsiae, is interesting from a host perspective because, while both R. akari and R. felis have been classified as SFG rickettsiae, neither species purportedly parasitizes a tick host, with R. akari found in mites  and R. felis found in fleas , , , , .
R. felis has been difficult to place phylogenetically because it displays some genotypic and phenotypic attributes of both SFG and TG rickettsiae, e.g., association with insect, hemolytic activity, actin-based motility, transovarial maintenance in the vector hosts, and serological cross-reactivity. In addition, the genome sequence of R. felis revealed morphological surprises, such as the presence of plasmids (pRF) and conjugative pili, as well as genetic traits atypical of most rickettsial genomes, such as elevated copies of several transposase families, spoT genes and genes coding for uncharacterized proteins with ankyrin (ANK) and tetratricopeptide (TPR) motifs , . Despite tremendous laboratory efforts these features had not been identified in Rickettsia in the pre-genomic era .
The discovery of a plasmid system in R. felis is rather peculiar since no other bacteria in the Rickettsiales (Ehrlichia, Anaplasma, Neorickettsia, and Wolbachia) are known to harbor plasmids based on their available completed genome sequences. Plasmids are known only from a few other obligate intracellular bacteria, including the γ-proteobacterium Q fever agent Coxiella burnetti , , str. MCS of an unclassified species of Mycobacterium , and the distantly related Chlamydiaceae species Chlamydia trachomatis , , , Chlamydia muridarum , Chlamydophila pneumoniae , , Chlamydophila psittaci , , , Chlamydophila caviae , and Chlamydophila felis . The existence of a putative conjugative plasmid presents a historical riddle when R. felis is considered within a phylogenetic context. R. felis is neither ancestral nor derived within the rickettsial evolutionary tree , , , yet it is the only published rickettsial genome to date that contains autonomous and seemingly functional plasmids , . This entails one of two evolutionary scenarios for rickettsiae. First, all ancestral rickettsiae once contained functional plasmids that have been lost in all lineages for which a genome has been sequenced (except R. felis). Given the phylogenetic position of R. felis, this implies multiple losses of the plasmid and raises the question of why, among all ten sequenced rickettsial genomes, a single maintenance of a plasmid system would remain in R. felis. Under this hypothesis, if some plasmid genes are essential for rickettsial fitness, then the lineages without plasmids may have had the plasmid genes incorporated into their chromosomes where they have become a permanent fixture, as is the case for the plasmidless C. burnetti isolate Scurry Q217 ,  and some plasmidless Chlamydia spp. , , . Thus, pRF genes that are not present in other rickettsiae would likely be involved in the specific biology of R. felis, and may involve critical proteins for its survival and/or virulence. Otherwise they could be pseudogenes in the early stages of decay.
Alternatively, perhaps R. felis acquired a plasmid from another organism and has retained this plasmid through the incorporation of certain genes on the plasmid into its life cycle such that loss of these genes would be deleterious for the survival and/or virulence of R. felis. This hypothesis implies that pRF genes would be less related to other rickettsial orthologous genes than are genes on the R. felis chromosome. Garnering support for either evolutionary model requires the evaluation of each plasmid gene through phylogenetic estimation and comparative analysis with other organisms that may likely have contributed to the structure and composition of pRF via vertical (phylogenetic) and horizontal (e.g., conjugative) gene transfer (HGT).
Ogata et al.  proposed that the genes found on the plasmid and chromosome of R. felis are homologous (sharing common ancestry) and likely undergo exchanges between the replicons. To test this claim, as well as to rule out one of the two hypotheses stated above, we present a phylogenetic analysis that discerns the mode of inheritance of the genes on the pRF plasmid that are also found on the chromosome of R. felis and other rickettsiae. Furthermore, we characterize the composition of the pRF plasmid in light of other plasmid-containing obligate intracellular bacteria, adding novel information that strengthens the hypothesis that pRF is conjugative and self-replicating. This comparison of the evolution of independent replicons in R. felis will ultimately lend resolution to the ambiguity that has long plagued the systematic placement of this perplexing rickettsial taxon.
Results and Discussion
An estimated phylogeny from 15 chromosomal-encoded proteins present in nine rickettsial genomes and two strains of Wolbachia endosymbionts, is presented in Figure 1 and used as a reference for the robust relationships of the four main lineages of rickettsiae (AG, TG, TRG, SFG). This single most parsimonious tree, 8061 steps in length, was the result of an exhaustive search that analyzed all possible trees from treespace. From a total of the 12263 amino acid characters in the concatenated alignment, 3039 were parsimony-informative. This phylogeny is congruent with other recently published rickettsial molecular phylogenies , , , thus supporting our establishment of the TRG rickettsiae as a lineage distinct from other previously established taxonomic categories for rickettsiae.
Figure 1. Phylogeny estimation from analysis of fifteen R. felis proteins.
Phylogeny estimation under parsimony of fifteen R. felis proteins (hypothetical protein RF_0005, threonyl-tRNA synthetase, preprotein translocase SecA subunit, uncharacterized low-complexity protein RF_0864, pyruvate phosphate dikinase precursor, leucyl-tRNA synthetase, hypothetical protein RF_0556, NAD-specific glutamate dehydrogenase, DNA polymerase III alpha chain, O-antigen export system permease protein RfbA, thioredoxin, NADPH-dependent glutamate synthase beta chain and related oxidoreductases, putative TIM-barrel protein in nifR3 family, UDP-3-O-[3-hydroxymyristoyl] glucosamine, and zinc/manganese ABC transporter substrate binding protein TroA_c) from nine rickettsial species (Rickettsia bellii, R. canadensis, R. prowazekii, R. typhi, R. akari, R. felis, R. conorii, R. rickettsii, and R. sibirica) and two strains of Wolbachia. Branch support is from one million bootstrap replicates. Genome information was compiled from the PATRIC Website. * Total R. felis genome size: 1,485,148 bp = chromosome; 62,829 bp = pRF and 39,263 bp = pRFδ.doi:10.1371/journal.pone.0000266.g001
Summary information of nine published rickettsial genomes illustrates the increased genome size of R. felis relative to other non-AG rickettsiae (Figure 1). While R. felis genome size is most similar to R. bellii str. RML369-C, it is interesting that the presence of pRF does not result in a remarkably larger genome size, given the lack of plasmids in R. bellii str. RML369-C . In light of the recent discovery of pili in R. bellii str. RML369-C , it may be possible that genes once present on a plasmid have been incorporated into the chromosome, accounting for the larger size of the R. bellii genome relative to other rickettsiae (save R. felis). Other R. felis genome summary statistics, including gene number, percent GC and percent coding, are not significantly different than other rickettsial genomes (Figure 1).
Phylogenies of Chromosomal and pRF genes
A phylogeny estimated from 21 conserved hypothetical proteins found only on the chromosomes of 10 Rickettsia spp. (Figure 2A) is similar to our phylogeny based on six fewer genes (Figure 1) as well as other recently published rickettsial trees , , , suggesting that the conserved rickettsial hypothetical proteins are phylogenetically informative markers. This suggests that these uncharacterized ORFs code for functional proteins, as we might expect less conserved genes or pseudogenes to elicit a phylogenetic signal different from the organismal phylogeny , , . Even though these 21 proteins are hypothetical, the bootstrap values increased for several of the branches compared to the tree based on 15 proteins (Figure 1), suggesting that adding more data for phylogeny estimation is better for recovering the apparent relatedness amongst these 10 Rickettsia spp., an approach recently verified using concatenated nucleotide alignments . Thus, the failure for this phylogenetic position of R. felis to be recovered in many previous studies is likely due to the fewer number of genes analyzed. For instance, when we analyzed nine genes, we recovered the same tree topology but with weaker bootstrap support (data not shown). Furthermore, analyses of fewer than nine genes did not consistently recover the R. akari/R. felis clade. In light of these findings, we strongly caution against the use of single or few genes to estimate historical divergence within rickettsiae. This problem can be seen in the comparison of recent trees estimated from 16S rDNA sequences. While the tree from Perlman et al.  is in general agreement with our estimates (Figure 1 & 2A) and that of Abergel et al. , the 16S rDNA trees from Lawson et al. , Kikuchi et al. , Fournier et al. , and our unpublished analyses do not recover the AG rickettsiae as basal to the remaining Rickettsia spp. that are pathogenic in vertebrates. These conflicting trees are likely due less to differences in analytical methods than variance in the sampled taxa, thus highlighting that taxon sampling is just as important as concatenating multiple datasets for estimating a robust phylogeny of rickettsiae.
Figure 2. Comparison of phylogeny estimations from exclusively chromosomal proteins and proteins present on the chromosome and plasmids of R. felis.
(A) Estimated phylogeny of 21 exclusively chromosomal proteins from 10 rickettsial strains. (B) Estimated phylogeny of 10 proteins present on the chromosome and plasmids of R. felis. “Ancestral” (red) refers to primitive rickettsiae with no known potential for host virulence. TG (aquamarine) = typhus group, TRG (light blue) = transitional group and SFG (brown) = spotted fever group. TG and TRG boxes depict the major differences in tree topologies. The pRF genes are boxed and shaded. Results from both analyses of amino acids are from an exhaustive search under parsimony with branch support from one million bootstrap replications.doi:10.1371/journal.pone.0000266.g002
While our generated phylogenies described above (Figure 1, 2) are robust and in agreement with other well-supported studies, the consensus phylogeny of the seven ORFs found on both pRF and the R. felis chromosome, as well the chromosomes of the other sampled rickettsiae, tells a different story (Figure 2B). Most unexpected is the non-monophyly of the pRF and chromosomal genes of R. felis, which strongly suggests that the analyzed pRF genes were not vertically passed over time in the lineage leading to R. felis, but rather were likely inherited horizontally from other bacterial plasmids and non-bacterial DNA. Furthermore, none of the individual plasmid gene trees are in agreement with the chromosomal phylogeny (Figure 3). The majority of the single trees place the plasmid genes basal to the non-R. bellii taxa (Figure 3A, C, D, E, G). One tree groups the plasmid genes closely with R. canadensis (Figure 3B), while one tree is mostly discordant with any reasonable rickettsial phylogeny (Figure 3F). Thus, both consensus and individual pRF phylogenies suggest an affinity of pRF with AG rickettsiae over any other rickettsial group. The inclusion of the pRF genes in rickettsial phylogeny estimation has a profound impact on tree topology, as TG and SFG rickettsiae are grouped as monophyletic to the exclusion of TRG rickettsiae, which groups with R. canadensis and pRF (Figure 2B, 3H). This is perhaps our most compelling evidence for the separation of TRG rickettsiae from SFG rickettsiae.
Figure 3. Individual phylogeny estimations for the seven pRF proteins used in the combined analysis of pRF.
(A, B) Majority rule consensus trees. (C–H) Strict consensus trees. All analyses were of amino acids from an exhaustive search under parsimony with branch support from one million bootstrap replications. Bootstrap values are placed above branches. Percentages of nodes recovered in majority rule consensus trees are shown below branches. Scores are tree lengths, with total characters and number of parsimony informative characters provided.doi:10.1371/journal.pone.0000266.g003
Our observation that pRF genes are closely related to AG rickettsial genes is further supported by analyzing the distribution of the R. felis genes found exclusively on pRF (Table 1). Thirty-eight of the 68 pRF genes are not found on the R. felis chromosome, and none of these 38 genes are present in any of the SFG. Furthermore, only three exclusively pRF genes occur in TG and one truncated gene is found in R. akari. However, 10 exclusively pRF genes (or inactive derivatives) occur in the ancestral rickettsiae, with five of these genes characteristic of typical plasmids. While our phylogenetic analysis suggests that the pRF plasmid of R. felis is likely composed of many products of HGT, possibly through conjugation with other bacteria (especially AG rickettsiae), the question still remains: of all sequenced rickettsial genomes, why do plasmids occur exclusively in R. felis? We attempt to answer this in light of the characterization of the plasmid genes and their distribution in R. felis chromosome, other rickettsiae, other bacterial taxa, and other non-bacterial organisms.
Table 1. Annotation of R. felis large (pRF) and small (pRFδ) plasmid-encoded genes that are absent on the R. felis chromosome, and their distribution in other rickettsiae, other bacterial sequences, and other non-bacterial taxa.doi:10.1371/journal.pone.0000266.t001
Characterization and Distribution of pRF Genes
The pRF plasmid in R. felis could be maintained for a variety of reasons, all of which are not mutually exclusive. First, since it contains proteins with high homology to bacterial conjugation proteins (tra genes), pRF could function in fertility as a (F)plasmid, exchanging genetic material with other bacterial congeners via a pilus. Support for this role comes from the presence of putative DNA transfer proteins TraA Ti (pRF38/39), TraD Ti (pRF37) and TraD (pRF43/44), and two F-pilus assembly/aggregate stabilization homologs of E. coli TraGF (pRF46 and pRF47). Other important conjugative proteins are found on the R. felis chromosome, including two competence protein ComE3 orthologs (RF0020 and RF0021), a DNA primase (RF0786) similar to the E. coli TraC protein that replicates transferred DNA in recipient cells, a protein (RF0705/RF0706) similar to the P-pilus assembly protein FimD, and a protein (RF0964) similar to E. coli F-pilin acetylation protein TraX . Thus, the majority of conjugative transfer genes and other structural proteins involved in conjugation are present in the R. felis genome, suggesting at the very least that R. felis once had the capacity to transfer plasmid genes. This is further supported by the presence of some of these proteins in other rickettsiae (Table 1 and 2), particularly the ancestral taxa. As previously noted , the presence of conserved and fully intact type IV secretion system (T4SS) components (found in all rickettsiae) could allow for efficient transfer of plasmid DNA, much like that found in the vir and dot/icm T4SSs of Agrobacterium tumefaciens and Legionella pneumophila, respectively, that have dual functions as DNA-transfer machines and effector protein translocators .
Table 2. Annotation of R. felis large (pRF) and small (pRFδ) plasmid-encoded genes that are present on the R. felis chromosome, and their distribution in other rickettsiae, other bacterial sequences, and other non-bacterial taxa.doi:10.1371/journal.pone.0000266.t002
Second, pRF may exist as a resistance, or (R)plasmid, that can allow a tolerance for antibiotics or poisons present in the niche of R. felis. While there is an elevated level of drug resistance genes on the R. felis chromosome , including six R. felis-specific proteins, there are no genes on pRF that suggest the plasmid plays any role in antibiotic or poison evasion. Third, pRF could function as a Col-plasmid, killing other bacteria with colicine gene products (bacteriocins). So far, no colicine gene orthologs have been identified in any rickettsiae genomes. Fourth, because 17 of the genes on the pRF plasmid have not been characterized (or annotated), it cannot be ruled out that pRF isn't a degrative plasmid, enabling the digestion of unusual substances like toluene or salicylic acid, which may be encountered in primary and secondary host environments.
Finally, pRF could be a virulence plasmid, allowing the bacterium to be pathogenic to its primary and secondary hosts. As a prerequisite to virulence, pathogenic bacteria must have proteins involved in recognition of and adaptation to host cells. pRF contains both ANK (pRF14, pRF42) and TPR (pRF12, pRF15, pRF16, pRF17, pRF18, pRF24, pRF26) motif-containing genes that are typically involved in protein-protein interactions in eukaryotic cells , , , , , , . ANK-containing genes are known from other intracellular pathogens such as Wolbachia pipientis , Ehrlichia phagocytophila , C. burnetii  and Legionella pneumophila  and likely play an important role in the manipulation of host cell physiology. In addition, one copy of a surface antigen, sca12 (pRF25), is also present on pRF. Previously, two pRF genes were implicated as likely virulence factors : a patatin homolog (pRF11), pat2, which is a patatin-like phospholipase, and a hyaluronidase gene (pRF56), which codes for an enzyme, or “spreading factor”, that increases host tissue permeability and enables antigenic disguise. BLASTP results suggest that pat2 is found in other α-proteobacteria as well as plants (Table S1), while the putative hyaluronidase is similar to distantly related bacteria as well as proteins from various arthropod genomes (Table S2). These proteins are strong candidates for testing the role of pRF in virulence. Furthermore, they may provide the best explanation for why R. felis is the only known Rickettsia species that has retained plasmids. Pat2 of pRF has been shown to be unrelated to other patatins present on the R. felis chromosome and the chromosomes of other rickettsiae , , and, although these patatins have been ruled out as factors for phagosomal escape , an ascribed phospholipase activity and conservation of A2 active sites suggests that Pat2 likely has a role in R. felis virulence that is distinct from other rickettsiae. Interestingly, the pFra plasmid of Yersinia pestis, the causative agent of plague, encodes a phospholipase D gene, Ymt, that is essential for colonization of the flea midgut and eventual transmission to a secondary mammalian host , . This virulence plasmid is absent in the species Yersinia pseudotuberculosis, which is the less virulent immediate ancestor to Y. pestis , . Additionally, the presence of a hyaluronidase gene on pRF, which is not found in other rickettsiae, suggests an important role in the life cycle of R. felis. These two genes may be critical to the survival and virulence of R. felis such that their failure to integrate into the chromosome has driven the retention of plasmids in R. felis (and other as yet unidentified plasmid-containing Rickettsia spp.). Thus, it is probable that some genes on pRF confer R. felis with the ability to recognize, invade and cause virulence in host cells. This is consistent with another bacterial pathogen, Bacillus anthracis, the causative agent of anthrax, that seemingly acquired virulence plasmids from a source other than its most common ancestor, which lacks plasmids . Like B. anthracis and Y. pestis, R. felis could have acquired its pRF plasmid in a serendipitous event that allowed for colonization of a new primary host (flea) as well as a range of secondary vertebrate hosts, a mechanism recently proposed to explain “host jumping” in pathogenic bacteria . This hypothesis is certainly consistent with the unique combination of SFG and TG rickettsiae characteristics that define R. felis.
Further support for a functional plasmid that is probably essential for the life cycle of R. felis comes from analysis of other interesting proteins on pRF that have plasmid-like characteristics. For instance, there are 12 ORFs on pRF that are similar to transposase (tnp) genes from 10 different transposase (or inactive derivative) families (Tables 1 and 2). This high occurrence of transposases suggests that pRF genes have been frequently rearranged through recombination mediated by tnp elements. This is further supported by the presence of two putative integrase genes (pRF14, pRF42), which typically integrate viral DNA into host chromosomes, a putative TnpR resolvase gene that codes for a site specific recombinase that is typical of plasmids (pRF32), a putative lon gene coding for an ATP-dependent serine protease La (bacterial type) (pRF26) likely involved in DNA-binding and cytokinesis, three copies of a DnaA-like protein (pRF05, pRF19, pRF20) that could function in initiation of plasmid replication, and a site-specific recombinase with similarity to DNA invertase Pin homologs (pRF66). Two parA genes (pRF03, pRF23) on the plasmid could be involved in cytokinesis, and one parB gene (pRF35) is likely responsible for cleaving ssDNA and particularly super-coiled plasmid DNA. Finally, pRF22 is similar to the plasmid Ppsr1 ORF 12 from Pseudomonas syringae .
Some plasmid-containing bacteria have incorporated toxin-antitoxin (TAT) systems for keeping the partitioning and inheritance of plasmids stable , , , . Under constitutive expression with the antitoxin component on the plasmid and the toxin on the chromosome, the lethality of the more stable toxin is mediated by the unstable antitoxin. Upon faulty segregation of plasmids after cell division, plasmidless daughter cells are killed by elevated toxin levels due to the breakdown of the unstable antitoxin , , . Aside from the 16 toxin and 14 antitoxin genes identified in the R. felis chromosomal genome, we suggest that another ORF (RF1343) may encode a putative toxin with mild homology to the mazF toxin gene of several plasmid-containing bacteria. The mazEF TAT module, first characterized in E. coli , , is found on the chromosomes of many bacteria and functions in programmed cell death . Interestingly, a BLAST search using RF1343 as the query resulted in two other Rickettsia spp. containing this putative toxin: R. bellii str. RML369-C and R. akari, suggesting that all other rickettsiae have lost this gene. Alternatively, this putative mazF toxin gene could have been transferred from the AG rickettsiae to TRG rickettsiae. No other α-proteobacteria seem to contain a mazF toxin gene, and the most similar homologs are from three species of Firmicutes and one β-proteobacterium, Ralstonia eutropha str. H16 (Table S3). Searches for a complementary antitoxin mazE-like protein in other rickettsiae were unsuccessful. While TAT systems are found in many free-living bacteria, they are rare among obligate intracellular pathogens , . The presence of RF1343 in the R. felis genome may hint at an evolutionary relic of a once functional system for regulating programmed cell death in the absence of stable plasmid inheritance . Furthermore, the retention of only one component of this TAT module could allude to a switch in function of the mazF toxin homolog for adaptation to eukaryotic hosts, as has been suggested for other R. felis toxin and antitoxin genes . Such “neofunctionalization”  could likely be true for many of the duplicate genes acquired laterally in R. felis that have not been subjected to decay in the midst of selection favoring the original ortholog. Interestingly, another component of the maz system, mazG, whose gene product functions as a nucleoside triphosphate pyrophosphohydrolase, is present and extremely variable across rickettsiae (data not shown).
Identification of pRF oriV and replication termination regions
We identified the putative origin of replication (oriV) of pRF (Figure 4) based on results from GenSkew (http://mips.gsf.de/services/analysis/genskew), an application for computing and plotting nucleotide skew data. It has previously been reported for Borrelia burgdorferi linear and circular plasmids that the oriV maps to regions of the plasmids wherein a significant and pronounced switch in DNA strand compositional asymmetry (AT and CG skew) occurs . We have identified the minimum cumulative skew of pRF at positions 20523K (AT-skew) and 21453 (CG-skew) which occurs within pRF23 (parA homolog) and pRF24 (hypothetical protein) (Figure 4). ParA functions in cytokinesis and replication initiation , and a BLASTP search using pRF23 retrieved with high homology other identified ParA proteins and orthologs with predicted functions in DNA replication and plasmid partitioning (Table 3). Additionally, within a ten-gene range from the pRF23 gene we found two other genes coding for putative Lon (pRF23) and DnaA (pRF23) proteins whose functions in replication are described above. The putative region of replication termination is marked by the minimum CG- and maximum AT-skews (Figure 4) and, rather interestingly, is flanked on both the 5′- and 3′- sides by five of the seven genes associated with the conjugation apparatus (the other two genes are close on the 3′-end). Thus it is highly likely that we have determined the oriV and replication termination regions, providing more evidence that pRF is indeed a functionally-replicating plasmid.
Figure 4. Characteristics and summary information of predicted origin of replication (oriV) of the pRF plasmid of Rickettsia felis.
(A) Schematic map of the pRF with shaded regions containing the putative oriV (right) and replication termination region (left). The region outlined in the dark dashed line depicts the portion of the plasmid missing in pRFδ (pRF15-pRF38). Grey boxes depict genes, with gene names described in Tables 1 and 2. Red lines depict coding strands, and yellow blocks depict areas of gene overlap. (B) AT-skew of pRF, with AT-skew (blue), cumulative AT-skew (red) and minimum AT-skew (orange). (C) CG-skew of pRF, with CG-skew (blue), cumulative CG-skew (red) and maximum CG-skew (orange). Plots generated and values computed with GenSkew (http://mips.gsf.de/services/analysis/genskew).doi:10.1371/journal.pone.0000266.g004
Table 3. Results of a BlastP search using pRF23 (parA) as a query.doi:10.1371/journal.pone.0000266.t003
Second R. felis Plasmid, pRFδ?
Ogata et al. ,  identified a second plasmid of 39,263 bp in R. felis, pRFδ, which is identical to pRF except for the deletion of 24 contiguous genes, pRF15 to pRF38 (Figure 4A). Of these deleted genes, 12 are exclusive to pRF (Table 1), while the other 12 are present on both the R. felis chromosome and pRF (Table 2). We call into question the existence of pRFδ for five reasons. First, several of the important genes predicted to be involved in plasmid maintenance and replication are absent in pRFδ, including genes coding for two putative Dna-like chromosomal replication initiator proteins (pRF19 and pRF20), the putative cytokinesis regulatory protein ParA (pRF23), a putative structural maintenance of chromosomes protein (ABC_SMC_euk) (pRF27), and ParB, a protein implicated in the cleavage of ssDNA and supercoiled plasmid DNA (pRF35). Second, six pRF proteins that have homology to proteins from other plasmid-containing bacteria are absent in pRFδ, including the ORF similar to P. syringae plasmid Ppsr1 ORF12 (pRF22), rickettsial hypothetical protein pRF29, the plasmid-encoded site specific recombinase TnpR (pRF32), a DNA polymerase III epsilon subunit-like protein with WGR domain (pRF34), and the putative conjugative transfer proteins TraD Ti (pRF37) and TraA Ti (pRF38). Third, our recent attempts (unpublished data) and those of Pornwiroon et al.  to amplify pRFδ in a different strain of R. felis (str. LSU) were unsuccessful. Fourth, our predicted oriV of pRF that is substantiated by gene composition, sharp change in coding strand and nucleotide compositional skew, is deleted in pRFδ (Figure 4A), suggesting that another means of plasmid replication would be responsible for its continual inheritance. Finally, the fifth reservation we have with the existence of pRFδ in R. felis deals with plasmid incompatibility. Plasmid incompatibility is the failure of two co-resident plasmids to be stably inherited without external selection . Incompatibility arises either by conflict in common replication or maintenance elements found in each unique plasmid, or by interference with the ability to correct stochastic fluctuations in copy number of the co-resident plasmids . Even though pRFδ is lacking several of the important genes suspected in plasmid replication and maintenance, the presence of other genes identical to pRF would likely result in plasmid incompatibility, either by conflict in maintenance, replication or regulation of copy number. Collectively, these five points suggest that pRFδ may be an artifact of the genome sequencing and assembly procedures, and that only one plasmid, pRF, probably occurs in strains of R. felis.
Horizontal Transfer of pRF Genes
Likely due to their highly reductive genomes , , , there is relatively little HGT within α-proteobacteria , and particularly rickettsiae , , . This is consistent with other parasitic bacteria with small genomes, such as Mycoplasma genitalium and Chlamydia spp. ,  and the actinobacterium Tropheryma whipplei . Some examples of HGT between α-proteobacteria and their hosts are well known. For instance, segments of the genome of a Wolbachia endosymbiont (wOvo) have been inserted into the nuclear genome of its host, a human-parasitic nematode, Onchocerca volvulus . Similarly, a large segment of DNA from a Wolbachia endosymbiont of the aduki bean beetle, Callosobruchus chinensis, has been detected on the X chromosome in the nuclear genome of the beetle . Still, plasmids are thus far unknown from other Rickettsiales, suggesting other means for genetic exchange between Wolbachia spp. and their hosts. HGT between chlamydiae and rickettsiae has been suggested for the transfer of ADP/ATP-translocase genes (tlc) from the former to the latter roughly 1.5 billion years ago while both organisms likely inhabited single-celled primitive eukaryotes , . Regarding the hypothetical immediate ancestor to TG and SFG rickettsiae, HGT has been proposed for the acquirement of S-adenosylmethionine synthetase (metK) from an unknown γ-proteobacterium . Recently, the genome sequence of R. bellii str. RML369-C revealed the presence of many genes with high homology to several obligate intracellular symbionts of amoebae, suggesting that at one time R. bellii likely exchanged genetic information, particularly with the ancestors of Legionella and Protochlamydia, while residing in an amoeba-like ancestral protozoan . While the R. bellii str. RML369-C genome encodes a complete set of putative conjugal DNA transfer genes (Tables 1 & 2), the lack of plasmids suggests that either R. bellii once had plasmids and has lost them or integrated them into its genome, or that R. bellii possesses the ability to naturally uptake DNA from its environment. Regardless, it is certainly plausible to suggest that R. bellii and R. felis have had the capability to exchange DNA either by conjugation or some other as of yet uncharacterized mechanism.
While the extent to which the putative conjugation proteins of pRF function in the lateral exchange of genes is still a mystery, it is probable that some degree of HGT has occurred between pRF and other bacteria, particularly AG rickettsiae and other distantly related bacteria. It is well known that plasmids function as vehicles for DNA transfer, and the characterization of the composition of pRF described above implicates as many as 32 of the total 68 pRF genes as putative candidates for HGT (Tables 1 & 2). These 32 genes are not found in other rickettsiae, yet 18 of them are present with high homology in other bacterial genomes, and 14 share limited similarity with non-bacterial sequences. Interestingly, 13 of the total 68 pRF genes are similar to proteins present on plasmids in other bacteria, suggesting that their origin may be from other distantly-related plasmid-containing bacteria. We calculated the %GC for each pRF gene to determine if a significant deviation from the average base composition of the plasmid or R. felis chromosome exists (Table 1 & 2). Within a range (±5) of the base composition of the R. felis genome (32.5% GC), only seven pRF genes deviate from the average base composition (pRF02, pRF04, pRF14, pRF25, pRF51, pRF65, pRF67). None of these genes are known from plasmids of other bacteria, and it is likely that they are just extremes to the average base composition of the R. felis genome. However, absence of base compositional bias alone should not be used to rule out HGT, as the base composition of transferred DNA segments will likely become nearly identical to the host genome over time , , , especially in intracellular symbionts wherein elevated mutation rates are typical .
Many bacterial genes passed horizontally likely fail at becoming an integral component of the new genome, with their eventual loss of function and subsequent decay into pseudogenes , . Only two pRF proteins are unique to R. felis (pRF09 and pRF59), having no putative function or orthology with other known proteins. This could, again, be a consequence of the highly reductive genomes of rickettsiae, or it could reflect two genes that have recently been laterally acquired and have yet to undergo decay. BLAST searches against the public virus protein database recovered no similarity between these ORFs and viruses, supporting the recent proposal that proteins with no recognizable homology to any other life are less likely to have been acquired from viruses than are other characterized proteins .
Chromosomal Virulence Factors Derived from Primitive Plasmids
The number of intracellular pathogens that have acquired the conjugal transfer systems of plasmids and modified them for the export of toxins is growing , . The presence of putative type IV pili in R. felis could imply that this machinery is responsible for the secretion of virulence factors and other proteins involved in host tropism, as was recently determined for Francisella tularensis virulence . Alternatively, a modification of a second “adapted” T4SS could be involved in the shunt of effectors into eukaryotic host cells. This is at least the more likely scenario as other plasmidless rickettsiae seem to have the ability to secrete effector molecules directly into host cells . The machinery for bacterial T4SS has been suggested to have undergone a great amount of HGT . Phylogenetic analysis implies that the ancestors to all chromosomally encoded T4SSs are plasmid-derived, with the lateral transfer of plasmid-encoded genes to the chromosome much more likely than vice versa . This supports the hypothesis that all rickettsiae may have once harbored plasmids, with T4SS components derived from these plasmids, either through independent acquirement events or one event ancestral to all rickettsiae. Indeed, the putative TraDF gene (pRF43) shares close similarity with the virD4 genes of other rickettsiae (Table 2). Interestingly, AG rickettsiae contain additional copies of virD4 that are more similar to pRF43 than to the other non-AG rickettsiae virD4 genes, further strengthening the argument for HGT between R. felis and AG rickettsiae.
In theory, the presence of duplicated T4SS genes would allow for bacteria to continue the use of pili-mediated conjugation with a separate adapted conjugation system for the secretion of virulence factors. The only two Rickettsia spp. known to have conjugative pili are R. felis str. URRWXCal2 and R. bellii str. RML369-C , , . Not surprising is that these two genomes also contain elevated copies of transposable elements, insertion sequences and repetitive elements relative to other sequenced rickettsial genomes . They also contain more HGT candidates than other rickettsiae, particulalry genes likely inherited from Legionellaceae and Parachlamydiaceae . While R. felis seemingly has the ability to transfer plasmids via its pilus, the genome of R. bellii str. RML369-C does not harbor plasmids by which to exchange DNA. However, conjugative transposons are also efficient at transferring DNA through the pilus . Indeed, another intracellular pathogen, Legionella pneumophila, performs T4SS-mediated conjugative transfer of both chromosomal  and plasmid  DNA. Thus, even in the absence of plasmids, rickettsiae with pili are likely able to exchange DNA via conjugation with other bacteria, as well as uptake genetic material from non-bacterial organisms.
In rickettsiae, no proteins potentially involved in virulence have been demonstrated to be secreted via the T4SS. However, we predict that some rickettsiae T4SS effectors characterized in the future will be genes inherited horizontally from distantly-related organisms, as products of HGT in prokaryotic genomes are often associated with pathogenicity . For example, HGT has been implicated in the acquirement of bacterial α2-macroglobulins, which function as colonization factors, in many diverse bacteria exploiting higher eukayotes as hosts . The recently sequenced genome of Chlamydophila abortus revealed several highly variable genomic islands, some of which contain genes involved in host invasion and virulence . Waterfield et al.  have demonstrated that a high level of similarity exists between pathogenicity islands from insect and mammalian pathogens. While no pathogenicity islands have been described from rickettsiae, it is likely that virulence factors could have resulted from previously acquired pathogenicity islands that are difficult to detect in highly reductive genomes. For instance, the presence of gene clusters involved in virulence or immunity on plasmids are often flanked by repetitive and direct repeats , as is known for microcin C51 plasmid genes of E. coli that are believed to be laterally inherited . Some of these genes are present in various rickettsiae in truncated and highly variable copies (data not shown). One microcin, mccE, is similar to a rickettsial gene that codes for rimJ, a ribosomal-protein-alanine N-acetyltransferase. While conserved across all rickettsiae, rimJ is unknown from other Rickettsiales, with the closest putative orthologs found in various genomes of Coxiella and Bacillus (Table S4). Even if the similarity between rimJ and mccE is the result of convergence in the C-terminal catalytic domain, the close affinity of rimJ to other bacteria well diverged from rickettsiae, many of which harbor plasmids (Table S4), may hint at the horizontal inheritance of this gene in rickettsiae. Also, one of the above-mentioned pRF genes possibly involved in R. felis virulence, the hyaluronidase (pRF56), has close similarity with several proteins from insects, suggesting it could be derived from an insect host gene and is involved in host manipulation (Table S2).
The search for virulence factors involved in rickettsial pathogenicity has not been easy and will entail intense scrutiny of the predicted ORFs with unknown function across the growing number of rickettsial genomes. Due to their streamlined evolution with constraints on genome size relative to many other bacteria, the detection of HGT in rickettsiae will be difficult for the following reasons: 1) most of the introduced genes are likely rapidly degraded and pruned from the genome, 2) accelerated rates of nucleotide evolution quickly disguise horizontally-inherited genes by homogenizing overall genomic base composition, and 3) neofunctionalization of co-opted genes only retains those motifs important in structure and function, making homology assignment virtually undetectable for similarity algorithms. Nevertheless, through an understanding of these properties of rickettsial evolution, coupled with rigorous phylogeny estimation and in vitro characterization, will a repertoire of horizontally inherited virulence factors begin to emerge, further illustrating the means by which these parasitic bacteria manipulate hosts cells throughout their life stages.
Plasmid Plasticity and Associated Virulence Traits
Of the 53 sequenced α-proteobacterial genomes currently available on the NCBI microbial genome database (http://www.ncbi.nlm.nih.gov/genomes/lproks.cgi), 16 are known to contain plasmids. Aside from Rickettsia, two other genera have species both with and without plasmids, Nitrobacter and Rhodopseudomonas, suggesting that plasmids are not essential for all species within α-proteobacterial genera that contain them. This plasticity for harboring plasmids is further intricate when analyzing intraspecific plasmid number. Three α-proteobacterial species differ in the number of plasmid types within their genomes: Mesorhizobium (2–3), Rhizobium (5–6) and Silicibacter (1–2). Thus it appears typical for at least α-proteobacterial plasmid-containing species that plasmids can be present or absent, and that the number of plasmid types per genome is plastic. This pattern extends to include four other genera of obligate intracellular pathogens (Table 4). It seems that plasmids are plastic across all five genera of the obligate intracellular pathogens listed, and in the instances where virulence factors are associated with the plasmid, plasmidless strains have incorporated the essential genes into their chromosomes (35,37,38,39,40). Thus, even though laboratory methods for detecting plasmids, such as pulse-field gel electrophoresis or PCR, may rule out the presence of plasmids within genomes, plasmids may reside within chromosomes either permanently or temporarily, not unlike the behavior of the bacteriophage lambda. Still, integration of plasmids into chromosomes is considered rare, and transfer of plasmid genes to the chromosome is likely the most efficient means for avoiding deleterious mutations or elimination by unequal segregation of important plasmid genes .
Table 4. Comparative genomic analysis of five plasmid-containing genera of obligate intracellular pathogens.doi:10.1371/journal.pone.0000266.t004
Plasmids are also plastic when analyzed across the genome sequences of five other selected pathogenic bacteria, namely Bacillus, Legionella, Neisseria, Pseudomonas and Yersinia (Table S5). Except for Yersinia, all five genera contain both strains with and without plasmids, and aside from Neisseria and Legionella, the number of plasmid types per species is variable. Although three sequenced genomes of Neisseria revealed no presence of plasmids, a larger sampling of several pathogenic and commensal neisseriae uncovered six plasmid types, with some containing genes with moderate homology to the recently identified meningococcal disease-associated phage . The virulence traits associated with the plasmids of Bacillus and Yersinia were discussed above and raise questions as to whether or not plasmidless strains have the genes necessary for virulence incorporated into their chromosomes. Analysis of two close strains of Legionella pneumophila, Paris and Lens, revealed three variable plasmids, with only the Paris strain harboring a T4SS encoded on a multicopy plasmid . Interestingly, robust sampling of L. pneumophila populations exposed that some genomes have this T4SS-plasmid integrated into the chromosome . Similarly, of three strains of Pseudomonas, pv. tomato DC3000, pv. phaseolicola 1448A and pv. syringae B728a, only the first two contain plasmids, with two plasmid types per genome , , . Analysis of the plasmidless str. pv. syringae B728a with other pseudomonad genomes revealed a genomic island with high similarity to a conjugative plasmid from P. aeruginosa str. PAO1, pKLC102, which carries a pil operon encoding type IV sex pili, as well as recombination and replication machinery , . Regarding Pseudomonas pv. tomato DC3000, even though copies of virulence genes are present on pDC3000A, the plasmid is not involved in virulence, as copies of the virulence genes have functional paralogs on the chromosome . Thus plasmids may initially be involved in virulence, but once the virulence genes are inserted into the chromosome, the plasmid genes are no longer essential for fitness. However, mutational analysis of T4SS genes on the plasmid pVir of Campylobacter jejuni str. 81–176 identified five genes that affect in vitro invasion of intestinal epithelial cells, suggesting the plasmid is essential for host colonization . Similarly, O'Connell and Nicks  demonstrated that plasmid-cured strains of Chlamydia muridarum grew smaller plaques than the wildtype and were unable to accumulate glycogen within intercytoplasmic inclusions. Thus, while plasmid plasticity may hint at a reduced role for plasmids in virulence, particularly when virulence genes are copied on the chromosome, it is likely that many plasmid systems are still essential for survival and/or virulence.
The number of gram-negative pathogenic bacteria that use adapted T4SSs for the export of virulence factors to hosts is growing, e.g. Agrobacterium tumefaciens, Bartonella tribocorum, Brucella spp., Helicobacter pylori, Bordetella pertussis, L. pneumophila, R. prowazekii, and Yersinia enterocolitica , , . Given that plasmid conjugation machineries are likely the ancestors to many of these modified T4SSs, the role of plasmids in the acquirement of virulence is an exciting area for research, particularly for systems wherein plasmids are plastic, such as rickettsiae. Future analysis of plasmid plasticity and characterization of plasmid-associated virulence traits will hopefully answer an important question: why do the sequenced genomes of other pathogenic rickettsiae not harbor plasmids? And more interestingly, why would R. felis need a plasmid with virulence traits when it has the same core set of genes (e.g., pld, tlyc, rompB, pat, etc.) shared by the other nine sequenced rickettsial genomes, some of which have already been implicated in host adherence and pathogenicity? Studies demonstrating the differential expression of pRF genes during the life cycle of R. felis may lead to the identification of genes involved in host invasion and virulence. Indeed, plasmid gene expression has been shown to be highly variable throughout the life cycle of another intracellular pathogen, C. trachomatis . Future studies should also determine if R. felis is capable of integrating pRF into its chromosome, and how plasmid-curing affects overall fitness.
We have demonstrated above through a phylogenomic/bioinformatic evaluation that there is strong support for the presence of a single plasmid in R. felis, pRF, and that many of the plasmid genes have probably been horizontally inherited from exchanges with other organisms either through a rudimentary conjugation apparatus or an adaptation of the T4SS that allows for dual transfer of plasmid DNA as well as translocation of effector molecules. Thus, we doubt the recent proposal that all pRF genes were inherited in one HGT event (Figure 5A), and that this event occurred either in R. felis  or an ancestor to TRG rickettsiae . Rather, our characterization of the R. felis plasmid within a phylogenomic context suggests that the primitive rickettsial ancestor likely harbored plasmids (Figure 5B). In the Rickettsia ancestor plasmids would have 1) eased the exchange of DNA with other intracellular pathogens, 2) fostered the duplication of certain genes (especially transposases), creating gene families and new genes through neofunctionalization, 3) facilitated the integration of plasmid genes into the chromosome, and 4) provided for a means to acquire a conjugation operon that would eventually give rise to the T4SS. Under this evolutionary model the plasticity of plasmids can be explained in relation to life history and overall fitness costs associated with plasmids: those lineages with strict host specialization, particularly lineages undergoing extreme gene loss and accelerated rates of nucleotide evolution (i.e., TG rickettsiae), would no longer need to harbor plasmids. Those lineages with a large genome size, elevated numbers of mobile elements and a less strict host range might retain plasmids.
Figure 5. Comparison of two hypotheses for the evolution of plasmids in rickettsiae.
(A) The appearance of a plasmid system in R. felis (as a member of SFG rickettsiae) as recently suggested (Ogata et al., 2005b; Blanc et al. 2007). (B) Our hypothesis centered on the notion that the ancestor to all rickettsiae harbored a plasmid system with subsequent losses in the ancestors to the TG and SFG rickettsiae, and in R. canadensis and R. bellii str. RML369-C. Red = ancestral rickettsiae, Aquamarine = typhus group, light blue = transitional group, brown = true spotted fever group. Trees are from Figure 2A.doi:10.1371/journal.pone.0000266.g005
We stress that the discovery of a plasmid system in R. felis, as well as the presence of conjugative pili in R. bellii and close genomic similarities between the two species, should be considered the opening of Pandora's box, as subsequent completed rickettsial genomes will likely yield more plasmid systems and other means for non-vertical exchange of genetic material within rickettsiae and between rickettsiae and other organisms. For instance, it was recently presented at the annual meetings of The American Society of Rickettsiology  that R. bellii str. OSU85-389 contains a putative conjugative plasmid (sequence as yet unpublished). Based on this finding, R. akari str. Hartford may also contain a conjugative plasmid, as it shares a unique clade (TRG rickettsiae) in the rickettsial tree with R. felis that is unique in that it colonizes mite and insect hosts respectively. The presence of a plasmid perhaps allowed for the horizontal acquisition of virulence factors from other mite- and insect-associated endosymbionts that occupy similar niches of TRG rickettsiae. The supposed finding of a plasmid in R. bellii str. OSU85-389 is compatible with our analysis of the pRF gene distribution throughout other rickettsiae, as we detected many of the plasmid-specific pRF genes in the genomes of two R. bellii strains. Plasmids would undoubtedly be beneficial for R. bellii as it has one of the largest host ranges of any Rickettsia species.
We predict that as more genomic sequences become available for other Rickettsia spp., the four clades defined herein using phylogenetic estimation (AG, TG, TRG, SFG) will remain strongly supported, and that R. australis and other rickettsiae with either recent host switches or the presence of plasmids, will likely fall within the AG and TRG rickettsiae. However, because the recently sequenced genome of another member of the SFG rickettsiae, R. massiliae, revealed a large genome size and the presence of a tra gene cluster similar to that found in R. bellii , plasmids may be uncovered in other as yet unsequenced SFG rickettsiae. Given the genomic similarities shared between the R. felis and R. bellii (and likely R. massiliae) genomes, it is likely that genetic mobility boosts the versatility and plasticity of these microorganisms. However, the role plasmids play in host colonization and virulence is not well understood, and will likely only become more apparent with both the discovery of plasmids in other rickettsiae, as well as the in vitro characterization of the contribution of plasmids to host recognition, invasion and pathogenicity. This aspect of rickettsiology is promising for the future development of vaccines for virulent rickettsiae, as well as for the construction of shuttle vectors for which plasmids may provide the first means for in vivo experimental models , .
Materials and Methods
We analyzed only those Rickettsia spp. for which a genome sequence was available: Rickettsia bellii str. RML369-C (NC_007940), R. bellii str. OSU85 389 (NZ_AARC00000000), R. canadensis str. McKiel (NZ_AAFF01000001), R. prowazekii str. Madrid E (NC_000963), R. typhi str. Wilmington (NC_006142), R. akari str. Hartford (NZ_AAFE01000001), R. felis str. URRWXCal2 (NC_007109), R. conorii str. Malish 7 (NC_003103), R. rickettsii (NZ_AADJ01000001), and R. sibirica str. 246 (NZ_AABW01000001). Additionally, to root the baseline phylogeny we included orthologous sequences from two outgroup taxa, one from the Wolbachia endosymbiont of Drosophila melanogaster (NC_002978.6), and one from the Wolbachia endosymbiont strain TRS of Brugia malayi (NC_006833.1). Methods of phylogeny estimation were the same for generating the baseline phylogeny (Figure 1), the phylogenies of exclusively chromosomal genes (Figure 2A) and those present on chromosomes and the pRF plasmid of R. felis (Figure 2B), and the single pRF genes (Figure 3). Initially, BLASTP  searches against the NCBI protein database were done using R. felis amino acid sequences as queries (for BLASTP specifics and threshold see below). For the baseline phylogeny, we analyzed 15 conserved rickettsial proteins (hypothetical protein RF_0005, threonyl-tRNA synthetase, preprotein translocase SecA subunit, uncharacterized low-complexity protein RF_0864, pyruvate phosphate dikinase precursor, leucyl-tRNA synthetase, hypothetical protein RF_0556, NAD-specific glutamate dehydrogenase, DNA polymerase III alpha chain, O-antigen export system permease protein RfbA, thioredoxin, NADPH-dependent glutamate synthase beta chain and related oxidoreductases, putative TIM-barrel protein in nifR3 family, and UDP-3-O-[3-hydroxymyristoyl] glucosamine). For the exclusively chromosomal dataset, we used the Rickettsia Orthologous Groups database at the PATRIC Website  to compile 21 hypothetical proteins that are present in all ten published rickettsial genomes. Our rationale for using hypothetical proteins was that these proteins have never been analyzed in prior studies and thus would test previous phylogeny estimates, as well as evaluate the phylogenetic utility of uncharacterized ORFs. The seven proteins comprising the third dataset consist of genes present of both the R. felis chromosome and pRF and distributed across all 10 rickettsial chromosomal genomes. Initially, 10 proteins qualified by this criterion; however, gene duplications, truncations and split genes did not permit the inclusion of pRF02, pRF19 and pRF25 (Table 2). pRF43 was also excluded because of its uncertain homology with chromosomal virD4 genes.
Retrieved rickettsial (and Wolbachia outgroup) protein sequences were exported in Fasta format and aligned locally using default parameters in the command-line version of the program MUSCLE , . Aligned datasets were converted to Nexus format using the program seqConverter.pl, version 1.1 . Each Nexus file was concatenated manually into a combined executable Nexus file and analyzed under parsimony in an exhaustive search in the program PAUP* version 4.10 (Altivec) . Branch support was assessed using the bootstrap  with default settings in PAUP. We performed one million bootstrap replications. Tree images were exported from PAUP* and manually adjusted in Adobe® Illustrator® CS2 v.12.0.1. Single gene phylogeny estimations were performed similarly.
Characterization and distribution of pRF genes
We performed BLASTP searches against the NCBI protein database using the 68 genes present on the pRF plasmid (NC_007110) as queries. The nr (All GenBank+RefSeq Nucleotides+EMBL+DDBJ+PDB) database was used, coupled with a search against the Conserved Domains Database. Searches were performed across all organisms with composition-based statistics. No filter was used. Default matrix parameters (BLOSUM62) and gap costs (Existence:11 Extension: 1) were implemented, with an inclusion threshold of 0.005. We then compiled the 68 genes into two tables: one based on pRF genes that were present only on pRF (Table 1), and another with genes also found on the R. felis chromosome (Table 2). When strong support for a function was given for closely related sequences, we modified the existing annotation for some genes. The presence or absence of the genes across 1) other rickettsiae, 2) other bacteria and 3) other viruses, archaea and eukaryotes was then assigned. For presence/absence in other viruses, archaea and eukaryotes, we did not reject sequences with low homology, as recent studies suggest that regions of divergent proteins sharing limited homology can actually be the result of molecular mimicry , . We also determined whether or not pRF genes were present in the plasmids of other bacteria.
Identification of pRF oriV
We predicted the putative origin of replication (oriV) of pRF (Figure 4) based on results from GenSkew (http://mips.gsf.de/services/analysis/genskew), an application for computing and plotting nucleotide skew data. We then plotted a ten-gene range from the center of the cumulative AT and GT skews to determine putative oriV and replication termination regions through identification of genes known to typically occur in these regions.
All relevant materials, including updates to pRF annotation and sequence alignments used to generate phylogenetic trees, will be available in a future update of the PATRIC rickettsial database .
Results of a BlastP search using pRF11
(0.04 MB DOC)
Results of a BlastP search using pRF56
(0.04 MB DOC)
Results of a BlastP search using RF1343
(0.04 MB DOC)
Results of a BlastP search using RP693
(0.04 MB DOC)
Comparative genomic analysis of five pathogenetic bacteria harboring plasmids with associated virulence factors.
(0.07 MB DOC)
We thank Micah Worley (UMB) for helpful discussion, and Matt Yoder (Texas A&M University) for assistance with regular expressions.
Conceived and designed the experiments: AA JG MB MR. Performed the experiments: JG JS. Analyzed the data: JG JS. Contributed reagents/materials/analysis tools: AA BS JG AP. Wrote the paper: AA BS JG MB MR NA AP.
- 1. Perlman SJ, Hunter MS, Zchori-Fein E (2006) The emerging diversity of Rickettsia. Proc Biol Sci 273: 2097–2106.
- 2. Hackstadt T (1996) The biology of rickettsiae. Infect Agents & Disease 5: 127–143.
- 3. Azad AF, Beard CB (1998) Rickettsial pathogens and their arthropod vectors. Emerg Infect Dis 4: 179–86.
- 4. Azad AF, Radulovic S (2003) Pathogenic rickettsiae as bioterrorism agents. Ann N Y Acad Sci 990: 734–738.
- 5. Weisburg WG, Dobson ME, Samuel JE, Dasch GA, Mallavia LP, et al. (1989) Phylogenetic diversity of the Rickettsiae. J Bacteriol. 171: 4202–4206.
- 6. Fournier P-E, Dumler JS, Greub G, Zhang J, Wu Y, et al. (2003) Gene sequence-based criteria for the identification of new Rickettsia isolates and description of Rickettsia heilongjiangensis sp. nov. J Clin Microbiol 41: 5456–5465.
- 7. Philip RN, Casper EA, Anacker RL, Cory J, Hayes SF, et al. (1983) Rickettsia bellii sp. nov.: A tick-borne rickettsia, widely distributed in the United States, that is distinct from the spotted fever and typhus biogroups. Int J Syst Bacteriol 33: 94–106.
- 8. Stothard DR, Clark JB, Fuerst PA (1994) Ancestral divergence of Rickettsia bellii from the spotted fever and typhus groups of Rickettsia and antiquity of the genus Rickettsia. Int J Syst Bacteriol 44: 798–804.
- 9. Stothard DR, Fuerst PA (1995) Evolutionary analysis of the spotted fever and typhus groups of Rickettsia using 16S rRNA gene sequences. Syst. Appl. Microbiol. 18: 52–61.
- 10. Stenos J, Walker DH (2000) The rickettsial outer-membrane protein A and B genes of Rickettsia australis, the most divergent of the spotted fever group. Int J Syst Evol Microbiol 50: 1775–1779.
- 11. Ishikura M, Fujita H, Ando S, Matsuura K, Watanabe M (2002) Phylogenetic analysis of spotted fever group Rickettsiae isolated from ticks in Japan. Microbiol Immunol ; 46: 241–247.
- 12. Kikuchi Y, Sameshima S, Kitade O, Kojima J, Fukatsu T (2002) Novel clade of Rickettsia spp. from leeches. Appl Environ Microbiol 68: 999–1004.
- 13. Eremeeva ME, Madan A, Shaw CD, Tang K, Dasch GA (2005) New perspectives on rickettsial evolution from new genome sequences of rickettsia, particularly R. canadensis, and Orientia tsutsugamushi. Ann N Y Acad Sci 1063: 47–63.
- 14. Abergel C, Blanc G, Monchois V, Renesto-Audiffren P, Sigoillot C, et al. (2006) Impact of the excision of an ancient repeat insertion on Rickettsia conorii guanylate kinase activity. Mol Biol Evol 23: 2112–2122.
- 15. Ogata H, LaScola B, Audic S, Renesto P, Blanc G, et al. (2006) Genome sequence of Rickettsia bellii illuminates the role of amoebae in gene exchanges between intracellular pathogens. PLoS Genet 2: e76.
- 16. Huebner RJ, Jellison WL, Pmerantz C (1948) Rickettsialpox-a newly recognized rickettsial disease. IV. Isolation of a rickettsia apparently identical with the causative agent of rickettsialpox from Allodermanyssus sanguineus, a rodent mite. Public Health Rep 61: 1677–1682.
- 17. Adams JR, Schmidtmann ET, Azad AF (1990) Infection of colonized cat fleas, Ctenocephalides felis (Bouche), with a rickettsia-like microorganism. Am J Trop Med Hyg 43: 400–409.
- 18. Azad AF, Sacci JB Jr, Nelson WM, Dasch GA, Schmidtmann ET, et al. (1992) Genetic characterization and transovarial transmission of a typhus-like rickettsia found in cat fleas. Proc Natl Acad Sci USA 89: 43–46.
- 19. Bouyer DH, Stenos J, Crocquet-Valdes P, Moron CG, Popov VL, et al. (2001) Rickettsia felis: Molecular characterization of a new member of the spotted fever group. Int J Syst Evol Microbiol 51: 339–347.
- 20. Zavala-Velazquez JE, Zavala-Castro JE, Vado-Solis I, Ruiz-Sosa JA, Moron CG, et al. (2002) Identification of Ctenocephalides felis fleas as a host of Rickettsia felis, the agent of a spotted fever rickettsiosis in Yucatan, Mexico. Vector Borne Zoonotic Dis 2: 69–75.
- 21. Ogata H, Robert C, Audic S, Robineau G, Blanc G, et al. (2005) Rickettsia felis, from culture to genome sequencing. Ann NY Acad Sci 1063: 26–34.
- 22. Ogata H, Renesto P, Audic S, Robert C, Blanc G, et al. (2005) The genome sequence of Rickettsia felis identifies the first putative conjugative plasmid in an obligate intracellular parasite. PLoS Biol 3: e248.
- 23. Kill K, Binnewies TT, Sicheritz-Pontén T, Willenbrock H, Hallin PF, et al. (2005) Genome update: Sigma factors in 240 bacterial genomes. Microbiology 151: 3147–3150.
- 24. Samuel JE, Frazier ME, Kahn ML, Thomashow LS, Mallavia LP (1983) Isolation and characterization of a plasmid from phase I Coxiella burnetii. Infect Immun 41: 488–493.
- 25. Samuel JE, Frazier ME, Mallavia LP (1985) Correlation of plasmid type and disease caused by Coxiella burnetii. Infect Immun 49: 775–779.
- 26. Copeland A, Lucas S, Lapidus A, Barry K, Detter JC, et al. (2006) Complete sequence of chromosome and plasmid of Mycobacterium sp. MCS. US DOE Joint Genome Institute.
- 27. Lovett M, Kuo C-C, Holmes K, Falkow S (1980) Plasmids of the genus Chlamydia. In “Current Chemotherapy and Infectious Disease. Proceedings of the 11th ICC and the 19th ICAAC,”. Amer Soc Microbiol. Washington, DC.pp. 1250–1252.
- 28. Hyypia T, Larsen SH, Stahlberg T, Terho P (1984) Analysis and detection of chlamydial DNA. J Gen Microbiol 130: 3159–3164.
- 29. Palmer L, Falkow S (1986) A common plasmid of Chlamydia trachomatis. Plasmid 16: 52–62.
- 30. Everett KD, Bush RM, Andersen AA (1999) Emended description of the order Chlamydiales, proposal of Parachlamydiaceae fam. nov. and Simkaniaceae fam. nov., each containing one monotypic genus, revised taxonomy of the family Chlamydiaceae, including a new genus and five new species, and standards for the identification of organisms. Int J Syst Bacteriol 49 Pt 2: 415–440.
- 31. Pickett MA, Everson JS, Pead PJ, Clarke IN (2005) The plasmids of Chlamydia trachomatis and Chlamydophila pneumoniae (N16): Accurate determination of copy number and the paradoxical effect of plasmid-curing agents. Microbiology 151: 893–903.
- 32. Joseph T, Nano FE, Garon CF, Caldwell HD (1986) Molecular characterization of Chlamydia trachomatis and Chlamydia psittaci plasmids. Infect Immun 51: 699–703.
- 33. Campbell LA, Kuo CC, Grayston JT (1987) Characterization of the new Chlamydia agent, TWAR, as a unique organism by restriction endonuclease analysis and DNA-DNA hybridization. J Clin Microbiol 25: 1911–1916.
- 34. Peterson EM, de la Maza LM (1988) Restriction endonuclease analysis of DNA from Chlamydia trachomatis biovars. J Clin Microbiol 26: 625–629.
- 35. Read TD, Myers GS, Brunham RC, Nelson WC, Paulsen IT, et al. (2003) Genome sequence of Chlamydophila caviae (Chlamydia psittaci GPIC): Examining the role of niche-specific genes in the evolution of the Chlamydiaceae. Nucleic Acids Res 31: 2134–2147.
- 36. Azuma Y, Hirakawa H, Yamashita A, Cai Y, Rahman MA, et al. (2006) Genome sequence of the cat pathogen, Chlamydophila felis. DNA Res 13: 15–23.
- 37. Savinelli EA, Mallavia LP (1990) Comparison of Coxiella burnetii plasmids to homologous chromosomal sequences present in a plasmidless endocarditis-causing isolate. Ann N Y Acad Sci 590: 523–533.
- 38. Willems H, Ritter M, Jager C, Thiele D (1997) Plasmid-homologous sequences in the chromosome of plasmidless Coxiella burnetii Scurry Q217. J Bacteriol. 179: 3293–3297.
- 39. Kalman S, Mitchell W, Marathe R, Lammel C, Fan J, et al. (1999) Comparative genomes of Chlamydia pneumoniae and C. trachomatis. Nat Genet 21: 385–389.
- 40. Read TD, Brunham RC, Shen C, Gill SR, Heidelberg JF, et al. (2000) Genome sequences of Chlamydia trachomatis MoPn and Chlamydia pneumoniae AR39. Nucleic Acids Res 28: 1397–1406.
- 41. Lawrence JG, Ochman H (1997) Amelioration of bacterial genomes: Rates of change and exchange. J Mol Evol 44: 383–397.
- 42. Lawrence JG, Ochman H (1998) Molecular archaeology of the Escherichia coli genome. Proc Natl Acad Sci USA 95: 9413–9417.
- 43. Liu L, Harrison PM, Kunin V, Gerstein M (2004) Comprehensive analysis of pseudogenes in prokaryotes: Widespread gene decay and failure of putative horizontally transferred genes. Genome Biol 5: R64.
- 44. Lawson ET, Mousseau TA, Klaper R, Hunter MD, Werren JH (2001) Rickettsia associated with male-killing in a buprestid beetle. Heredity 86: 497–506.
- 45. Cascales E, Christie PJ (2003) The versatile bacterial type IV secretion systems. Nat Rev Microbiol 1: 137–149.
- 46. Blatch GL, Lassle M (1999) The tetratricopeptide repeat: A structural motif mediating protein–protein interactions. Bioessays 21: 932–939.
- 47. Caturegli P, Asanovich KM, Walls JJ, Bakken JS, Madigan JE, et al. (2000) AnkA: An Ehrlichia phagocytophila group gene encoding a cytoplasmic protein antigen with ankyrin repeats. Infect Immun 68: 5277–5283.
- 48. Gough J, Karplus K, Hughey R, Chothia C (2001) Assignment of homology to genome sequences using a library of hidden Markov models that represent all proteins of known structure. J Mol Biol 313: 903–919.
- 49. Mosavi LK, Minor DL Jr, Peng ZY (2002) Consensus-derived structural determinants of the ankyrin repeat motif. Proc Natl Acad Sci USA 99: 16029–16034.
- 50. Chung S, Mondon P, Chang YC, Kwon-Chung KJ (2003) Cryptococcus neoformans with a mutation in the tetratricopeptide repeat-containing gene, CCN1, causes subcutaneous lesions but fails to cause systemic infection. Infect Immun 71: 1988–1994.
- 51. Cazalet C, Rusniok C, Bruggemann H, Zidane N, Magnier A, et al. (2004) Evidence in the Legionella pneumophila genome for exploitation of host cell functions and high genome plasticity. Nat Genet 36: 1165–1173.
- 52. Raoult D, Audic S, Robert C, Abergel C, Renesto P, et al. (2004) The 1.2-megabase genome sequence of Mimivirus. Science 306: 1344–1350.
- 53. Wu M, Sun LV, Vamathevan J, Riegler M, Deboy R, et al. (2004) Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: A streamlined genome overrun by mobile genetic elements. PLoS Biol 2: E69.
- 54. Seshadri R, Paulsen IT, Eisen JA, Read TD, Nelson KE, et al. (2003) Complete genome sequence of the Q-fever pathogen Coxiella burnetii. Proc Natl Acad Sci USA 100: 5455–5460.
- 55. Blanc G, Renesto P, Raoult D (2005) Phylogenetic analysis of rickettsial patatin-like protein with conserved phospholipase A2 active sites. Ann N Y Acad Sci 1063: 83.
- 56. Whitworth T, Popov VL, Y X-J, Walker DH, Bouyer DH (2005) Expression of the Rickettsia prowazekii pld or tlyC gene in Salmonella enterica serovar typhimurium mediates phagosomal escape. Infect Immun 73: 6668–6673.
- 57. Hinnebusch J, Cherepanov P, Du Y, Rudolph A, Dixon JD, et al. (2000) Murine toxin of Yersinia pestis shows phospholipase D activity but is not required for virulence in mice. Int J Med Microbiol. 290: 483–487.
- 58. Hinnebusch J, Rudolph A, Cherepanov P, Dixon JD, Schwan T, et al. (2002) Role of Yersinia murine toxin in survival of Yersinia pestis in the midgut of the flea vector. Science 296: 733–735.
- 59. Sodeinde OA, Subrahmanyam YV, Stark K, Quan T, Bao Y, et al. (1992) A surface protease and the invasive character of plague. Science 258: 1004–1007.
- 60. Achtman M, Zurth K, Morelli C, Torrea G, Guiyoule A, et al. (1999) Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis. Proc Natl Acad Sci USA 96: 14043–14048.
- 61. Waterfield NR, Wren BW, Ffrench-Constant RH (2004) Invertebrates as a source of emerging human pathogens. Nat Rev Microbiol 2: 833–841.
- 62. Scully LR, Bidochka MJ (2006) Developing insect models for the study of current and emerging human pathogens. FEMS Microbiol Lett 263: 1–9.
- 63. Sundin GW, Mayfield CT, Zhao Y, Gunasekera TS, Foster GL, et al. (2004) Complete nucleotide sequence and analysis of pPSR1 (72,601 bp), a pPT23A-family plasmid from Pseudomonas syringae pv. syringae A2. Mol Genet Genomics. 270: 462–476.
- 64. Jensen RB, Gerdes K (1995) Programmed cell death in bacteria: Proteic plasmid stabilization systems. Mol Microbiol 17: 205–210.
- 65. Yarmolinsky MB (1995) Programmed cell death in bacterial population. Science 267: 836–837.
- 66. Couturier M, Bahassi EM, Van Melderen L (1998) Bacterial death by DNA gyrase poisoning. Trends Microbiol 6: 269–275.
- 67. Engelberg-Kulka H, Glaser G (1999) Addiction modules and programmed cell death and anti-death in bacterial cultures. Annu Rev Microbiol 53: 43–70.
- 68. Hayes F (2003) Toxins–antitoxins: Plasmid maintenance, programmed cell death, and cell cycle arrest. Science 301: 1496–1499.
- 69. Gerdes K, Christensen SK, Lobner-Olesen A (2005) Prokaryotic toxin–antitoxin stress response loci. Nature Rev Microbiol 3: 371–382.
- 70. Buts B, Lah J, Dao-Thi MH, Wyns L, Loris R (2006) Toxin-antitoxin modules as bacterial metabolic stress managers. Trends Biochem Sci 30: 673–679.
- 71. Metzger S, Dror IB, Aizenman E, Schreiber G, Toone M, et al. (1988) The nucleotide sequence and characterization of the relA gene of Escherichia coli. J Biol Chem 263: 15699–15704.
- 72. Masuda Y, Miyakawa K, Nishimura Y, Ohtsubo E (1993) chpA and chpB, Escherichia coli chromosomal homologs of the pem locus responsible for stable maintenance of plasmid R100. J Bacteriol 175: 6850–6856.
- 73. Aizenman E, Engelberg-Kulka H, Glaser G (1996) An Escherichia coli chromosomal “addiction module” regulated by guanosine-3′5′-bispyrophosphate: A model for programmed bacterial cell death. Proc Natl Acad Sci U S A 93: 6059–6063.
- 74. Zhang YX, Li J, Guo XK, Wu C, Bi B (2004) Characterization of a novel toxin-antitoxin module, VapBC, encoded by Leptospira interrogans chromosome. Cell Res 14: 208–216.
- 75. Pandey DP, Gerdes K (2005) Toxin-antitoxin loci are highly abundant in free-living but lost from host-associated prokaryotes. Nucleic Acids Res 33: 966–976.
- 76. Engelberg-Kulka H, Amitai S, Kolodkin-Gal I, Hazan R (2006) Bacterial programmed cell death and multicellular behavior in bacteria. PLoS Genetics 2: e135.
- 77. Frank AC, Alsmark CM, Thollesson M, Andersson SG (2005) Functional divergence and horizontal transfer of type IV secretion systems. Mol Biol Evol 22: 1325–1336.
- 78. Picardeau M, Lobry JR, Hinnebusch BJ (2000) Analyzing DNA strand compositional asymmetry to identify candidate replication origins of Borrelia burgdorferi linear and circular plasmids. Genome Res 10: 1594–1604.
- 79. Fogel MA, Waldor MK (2006) A dynamic, mitotic-like mechanism for bacterial chromosome segregation. Genes Dev 20: 3269–3282.
- 80. Pornwiroon W, Pourciau SS, Foil LD, Macaluso KR (2006) Rickettsia felis from cat fleas: Isolation and culture in a tick-derived cell line. Appl Environ Microbiol 72: 5589–5595.
- 81. Novick RP, Clowes RC, Cohen SN, Curtiss R III, Datta N, et al. (1976) Uniform nomenclature for bacterial plasmids: A proposal. Bacteriol Rev 40: 168–189.
- 82. Novick RP (1987) Plasmid incompatibility. Microbiol Rev 51: 381–395.
- 83. Andersson SG, Eriksson AS, Naslund AK, Andersen MS, Kurland CG (1996) The Rickettsia prowazekii genome: A random sequence analysis. Microb Comp Genomics 1: 293–315.
- 84. Andersson SGE, Kurland CG (1998) Reductive evolution of resident genomes. Trends Microbiol 6: 263–268.
- 85. Andersson JO, Andersson SGE (1999) Insights into the evolutionary process of genome degradation. Curr Opin Genet Dev 9: 664–671.
- 86. Fitzpatrick DA, Creevey CJ, McInerney JO (2006) Genome phylogenies indicate a meaningful alpha-proteobacterial phylogeny and support a grouping of the mitochondria with the Rickettsiales. Mol Biol Evol 23: 74–85.
- 87. Ogata H, Audic S, Renesto-Audiffren P, Fournier PE, Barbe V, et al. (2001) Mechanisms of evolution in Rickettsia conorii and R. prowazekii. Science 293: 2093–2098.
- 88. Hsiao WWL, Ung K, Aeschliman D, Bryan J, Finlay BB, Brinkman FSL (2005) Evidence of a large novel gene pool associated with prokaryotic genomic islands. PLoS Genetics 1: e62.
- 89. Ochman H, Lawrence JG, Groisman EA (2000) Lateral gene transfer and the nature of bacterial innovation. Nature 405: 299–304.
- 90. Brinkman FS, Blanchard JL, Cherkasov A, Av-Gay Y, Brunham RC, et al. (2002) Evidence that plant-like genes in Chlamydia species reflect an ancestral relationship between Chlamydiaceae, cyanobacteria, and the chloroplast. Genome Res. 12: 1159–1167.
- 91. Raoult D, Ogata H, Audic S, Robert C, Suhre K, et al. (2003) Tropheryma whipplei Twist: A human pathogenic actinobacteria with a reduced genome. Genome Res 13: 1800–1809.
- 92. Fenn K, Conlon C, Jones M, Quail MA, Holroyd NE, et al. (2006) Phylogenetic relationships of the Wolbachia of nematodes and arthropods. PLoS Pathog 2:
- 93. Kondo N, Nikoh N, Ijichi N, Shimada M, Fukatsu T (2002) Genome fragment of Wolbachia endosymbiont transferred to X chromosome of host insect. PNAS 99: 14280–14285.
- 94. Wolf YI, Aravind L, Koonin EV (1999) Rickettsiae and Chlamydiae: Evidence of horizontal gene transfer and gene exchange. Trends Genet 15: 173–175.
- 95. Greub G, Raoult D (2004) Microorganisms resistant to free-living Amoebae. Clin Microbiol Rev 17: 413–433.
- 96. Blanc G, Ogata H, Robert C, Audic S, Suhre K, et al. (2007) Reductive genome evolution from the mother of Rickettsia. PLoS Genetics 3: e14.
- 97. Brown JR (2003) Ancient horizontal gene transfer. Nat Rev Genet 4: 121–132.
- 98. van Passel MWJ, Bart A, Luyf AC, van Kampen AH, van der Ende A (2006a) Compositional discordance between prokaryotic plasmids and host chromosomes. BMC Genomics 7: 26.
- 99. van Passel MWJ, van der Ende A, Bart A (2006b) Plasmid diversity in Neisseriae. Infect Immun 74: 4892–4899.
- 100. Itoh T, Martin W, Nei M (2002) Acceleration of genomic evolution caused by enhanced mutation rate in endocellular symbionts. PNAS 99: 12944–12948.
- 101. Lerat E, Daubin V, Ochman H, Moran NA (2005) Evolutionary origins of genomic repertoires in bacteria. PLoS Biol. 3: e130.
- 102. Yin Y, Fischer D (2006) On the origin of microbial ORFans: Quantifying the strength of the evidence for viral lateral transfer. BMC Evol Biol 6: 63.
- 103. Cao TB, Saier MH Jr (2001) Conjugal type IV macromolecular transfer systems of Gram-negative bacteria: Organismal distribution, structural constraints and evolutionary conclusions. Microbiology 147: 3201–3214.
- 104. Sexton JA, Vogel JP (2002) Type IVB secretion by intracellular pathogens. Traffic 3: 178–185.
- 105. Hager AJ, Bolton DL, Pelletier MR, Brittnacher MJ, Gallagher LA, et al. (2006) Type IV pili-mediated secretion modulates Francisella virulence. Mol Microbiol 62: 227–237.
- 106. Baron C, O'Callaghan D, Lanka E (2002) Bacterial secrets of secretion: EuroConference on the biology of type IV secretion processes. Mol Microbiol 43: 1359–1365.
- 107. Snyder EE, Kampanya N, Lu J, Nordberg E, Rajasimha H, et al. (2007) The VBI PathoSystems Resource Integration Center (PATRIC). Nucleic Acids Res. (Database Issue). In press:
- 108. Licht TR, Wilcks R (2006) Conjugative gene transfer in the gastrointestinal environment. Adv Appl Microbiol 58: 77–95.
- 109. Miyamoto H, Yoshida S, Taniguchi H, Shuman HA (2003) Virulence conversion of Legionella pneumophila by conjugal transfer of chromosomal DNA. J Bacteriol 185: 6712–6718.
- 110. Vogel JP, Andrews HL, Wong SK, Isberg RR (1998) Conjugative transfer by the virulence system of Legionella pneumophila. Science 279: 873–876.
- 111. Nakamura Y, Itoh T, Matsuda H, Gojobori T (2004) Biased biological functions of horizontally transferred genes in prokaryotic genomes. Nat Genet 36: 760–766.
- 112. Budd A, Blandin S, Levashina EA, Gibson TJ (2004) Bacterial α2-macroglobulins: Colonization factors acquired by horizontal gene transfer from the metazoan genome? Genome Biol 5: R38.
- 113. Thomson NR, Yeats C, Bell K, Holden MT, Bentley SD, et al. (2005) The Chlamydophila abortus genome sequence reveals an array of variable proteins that contribute to interspecies variation. Genome Res. 15: 629–640.
- 114. Hacker J, Kaper JB (2000) Pathogenicity islands and the evolution of microbes. Annu Rev Microbiol 54: 641–679.
- 115. Fomenko DE, Metlitskaya AZ, Peduzzi J, Goulard C, Katrukha GS, et al. (2003) Microcin C51 plasmid genes: Possible source of horizontal gene transfer. Antimicrob Agents Chemother 47: 2868–2874.
- 116. Lusher M, Storey CC, Richmond SJ (1989) Plasmid diversity within the genus Chlamydia. J Gen Microbiol 135: 1145–1151.
- 117. Thomas NS, Lusher M, Storey CC, Clarke IN (1997) Plasmid diversity in Chlamydia. Microbiology 143: 1847–1854.
- 118. McClenaghan M, Honeycombe JR, Bevan BJ, Herring AJ (1988) Distribution of plasmid sequences in avian and mammalian strains of Chlamydia psittaci. J Gen Microbiol 134: 559–565.
- 119. Buell CR, Joardar V, Lindeberg M, Selengut J, Paulsen IT, et al. (2003) The complete genome sequence of the Arabidopsis and tomato pathogen Pseudomonas syringae pv. tomato DC3000. PNAS 100: 10181–10186.
- 120. Feil H, Feil WS, Chain P, Larimer F, DiBartolo G, et al. (2005) Comparison of the complete genome sequences of Pseudomonas syringae pv. syringae B728a and pv. tomato DC3000. PNAS USA 102: 11064–11069.
- 121. Joardar V, Lindeberg M, Jackson RW, Selengut J, Dodson R, et al. (2005) Whole-genome sequence analysis of Pseudomonas syringae pv. phaseolicola 1448A reveals divergence among pathovars in genes involved in virulence and transposition. J Bacteriol 187: 6488–6498.
- 122. Klockgether J, Reva O, Larbig K, Tümmler B (2004) Sequence analysis of the mobile genome Island pKLC102 of Pseudomonas aeruginosa C. J Bacteriol 186: 518–534.
- 123. Bacon DJ, Alm RA, Hu L, Hickey TE, Ewing CP, et al. (2002) DNA sequence and mutational analyses of the pVir plasmid of Campylobacter jejuni 81-176. Infect Immun 70: 6242–6250.
- 124. O'Connell CM, Nicks KM (2006) A plasmid-cured Chlamydia muridarum strain displays altered plaque morphology and reduced infectivity in cell culture. Microbiology 152: 1601–1607.
- 125. Christie PJ (2001) Type IV secretion: Intercellular transfer of macromolecules by systems ancestrally related to conjugation machines. Mol Microbiol 40: 294–305.
- 126. Strauch E, Goelz G, Knabner D, Konietzny A, Lanka E, et al. (2003) A cryptic plasmid of Yersinia enterocolitica encodes a conjugative transfer system related to the regions of CloDF13 Mob and IncX Pil. Microbiology 149: 2829–2845.
- 127. Pearce BJ, Fahr MJ, Hatch TP, Sriprakash KS (1991) A chlamydial plasmid is differentially transcribed during the life cycle of Chlamydia trachomatis. Plasmid 26: 116–122.
- 128. Eremeeva ME, Madan A, Dasch GA (2006) Genome sequence of Rickettsia bellii OSU 85-389. 20th Meeting of The American Society for Rickettsiology in conjunction with the 5th International Conference on Bartonella as Emerging Pathogens. September 2-7, 2006. Asilomar Conference Grounds, Pacific Grove, California, USA. Abstract #11.
- 129. Holden M, Rajandream MA, Bentley S (2005) Food for thought. Nature Rev Microbiol 3: 912–913.
- 130. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, et al. (1997) Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res 25: 3389–3402.
- 131. Edgar RC (2004) MUSCLE: A multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5: 113.
- 132. Edgar RC (2004) MUSCLE: Multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32: 1792–1797.
- 133. Bininda-Emonds O (2006) seqConverter. pl, version 1.1. ed, Institut fur Spezeille Zoologie und Evolutionsbiologie mit Phyletischem Museum, Friedrich-Schiller-Universitat Jena.
- 134. Swofford D (1999) PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4 ed. Sunderland, , MA: Sinauer.
- 135. Felsenstein J (1985) Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791.
- 136. Stebbins CE, Galan JE (2001) Structural mimicry in bacterial virulence. Nature 412: 701–705.
- 137. Sikora S, Strongin A, Godzik A (2005) Convergent evolution as a mechanism for pathogenic adaptation. Trends Microbiol 13: 522–527.