|Home | About | Journals | Submit | Contact Us | Français|
The recent discoveries of the pRF and pRM plasmids of Rickettsia felis and R. monacensis have contravened the long-held dogma that plasmids are not present in the bacterial genus Rickettsia (Rickettsiales; Rickettsiaceae). We report the existence of plasmids in R. helvetica, R. peacockii, R. amblyommii, and R. massiliae isolates from ixodid ticks and in an R. hoogstraalii isolate from an argasid tick. R. peacockii and four isolates of R. amblyommii from widely separated geographic locations contained plasmids that comigrated with pRM during pulsed-field gel electrophoresis and larger plasmids with mobilities similar to that of pRF. The R. peacockii plasmids were lost during long-term serial passage in cultured cells. R. montanensis did not contain a plasmid. Southern blots showed that sequences similar to those of a DnaA-like replication initiator protein, a small heat shock protein 2, and the Sca12 cell surface antigen genes on pRM and pRF were present on all of the plasmids except for that of R. massiliae, which lacked the heat shock gene and was the smallest of the plasmids. The R. hoogstraalii plasmid was most similar to pRM and contained apparent homologs of proline/betaine transporter and SpoT stringent response genes on pRM and pRF that were absent from the other plasmids. The R. hoogstraalii, R. helvetica, and R. amblyommii plasmids contained homologs of a pRM-carried gene similar to a Nitrobacter sp. helicase RecD/TraA gene, but none of the plasmids hybridized with a probe derived from a pRM-encoded gene similar to a Burkholderia sp. transposon resolvase gene.
The genus Rickettsia (Rickettsiales; Rickettsiaceae) consists of small gram-negative obligately intracellular alphaproteobacteria associated with eukaryotic hosts. They occur in a wide range of arthropods (38, 39) but are best known as vertebrate pathogens transmitted by blood-feeding arthropods. Rickettsia prowazekii and Rickettsia typhi, transmitted by lice and fleas, respectively, are the etiologic agents of epidemic and endemic typhus. Various species associated with ticks are the agents of spotted fevers occurring around the world, but others found in ticks have not been associated conclusively with human disease (5, 37). Rickettsia species are the closest known microbial relatives of mitochondria (3, 58) and have undergone the reductive genome evolution characteristic of strict endosymbiotic bacteria (2, 4, 12). Rickettsial genomes are typically 1.1 × 106 to 1.5 × 106 base pairs in length, with a GC content of approximately 30%, and have high levels of both synteny and noncoding content (12). Many biosynthetic pathways that are present in free-living bacteria have been replaced by transport systems in rickettsiae, which renders them dependent on eukaryotic host cells for essential metabolites (42).
The recent discoveries of pRF in Rickettsia felis by whole-genome sequencing (35) and of pRM in Rickettsia monacensis (7) contravened the long-held dogma that plasmids are not present in rickettsiae. The pRM plasmid was identified by pulsed-field gel electrophoresis (PFGE) and Southern blot analyses of DNAs from R. monacensis populations containing a transposon encoding green fluorescent protein and chloramphenicol acetyltransferase markers integrated into pRM (7). Sequence analysis of the 23.5-kbp pRM plasmid revealed similarities in coding capacity to that of the 63-kbp pRF plasmid. However, the plasmids do not have syntenic gene organization, and both contain genes not found on the other that were in some cases apparently derived from other bacteria through transposon activity (7, 23, 35). The pRF plasmid was suggested to be unique to R. felis based on the failure of PCR assays to provide evidence of pRF-like DNA sequences in other rickettsiae (35). A subsequent phylogenetic analysis suggested an origin of pRF in the ancestral-group rickettsiae (23), an inference supported by discovery of pRM in R. monacensis (7).
The existence of plasmids in R. felis and R. monacensis raises questions concerning their potential presence in other Rickettsia spp., the degree of gene conservation among such plasmids, and their role in rickettsial biology. In this report, we demonstrate the existence of plasmids in Rickettsia helvetica, Rickettsia peacockii, Rickettsia massiliae, and four “Rickettsia amblyommii” isolates from Ixodid ticks. A newly described isolate, “Rickettsia hoogstraalii,” from the Argasid tick Carios capensis (27), contained a plasmid with probable homologs of six genes found on pRM, but the gene complements of plasmids from rickettsiae isolated from Ixodid ticks were less similar to that of pRM. Although the “R. hoogstraalii” and “R. amblyommii” isolates have not been described formally as Rickettsia spp., DNA sequence analyses have shown that they meet the gene sequence standards for designation as Rickettsia spp. (21), and they are referred to as R. hoogstraalii and R. amblyommii throughout this report. The R. amblyommii isolates and R. peacockii contained at least two plasmids of different sizes, and those of R. peacockii were lost during long-term serial passage of the organism in cultured tick cells. Of the Rickettsia spp. analyzed, only Rickettsia montanensis did not contain a plasmid. The occurrence of plasmids in a diverse range of Rickettsia spp. suggests that they may be widespread in the genus, and their existence in the face of pressure exerted by reductive genome evolution suggests an important role in rickettsial biology.
All rickettsiae were grown in the Ixodes scapularis ISE6 cell line, maintained in L-15B300 medium (28) as described previously (31). They included Rickettsia monacensis, passages 61 to 64 (48); the R. monacensis transformant Rmona658B, passages 13 to 15 (7); R. montanensis isolate M5/6, passages 17 to 27 since receipt from Robert Heinzen at NIH, Rocky Mountain Laboratories (10); R. felis isolate LSU, passages 24 to 27 (40); R. peacockii isolate DAE100R, passages 2 to 4 and 114 to 123 (47); R. helvetica isolate C9P9, passages 10 to 14 (9, 44); R. hoogstraalii isolate RCCE3, passages 43 to 46 (27); R. massiliae isolate Rs1 from Rhipicephalis sanguineus, passages 3 to 5 (U. G. Munderloh, unpublished data); R. amblyommii isolates from Amblyomma americanum, including WB-8-2 (passages 46 to 49) (14, 26, 51), AaR/SC (passages 4 to 6) (T. J. Kurtti, unpublished), and MOAa (passages 3 to 5) (57); and a fourth R. amblyommii isolate, Ac/Pa from Amblyomma cajennense from Panama, passages 2 to 4 (Munderloh, unpublished). Rickettsiae were purified from infected host cells suspended in L-15B300 medium by passage through a 25-gauge syringe needle six times to lyse the cells, passage of the lysates through 2-μm syringe filters to remove cellular debris, and centrifugation of the filtrates at 18,400 × g at 4°C for 5 min to pellet rickettsiae. Genomic DNAs were prepared from purified rickettsiae as described previously (7), and the identities of all isolates were confirmed by PCR amplification and sequencing (ABI 377 automated sequencer at Advanced Genetic Analysis Center, University of Minnesota) of the ompA gene (ompB for R. helvetica), using oligonucleotide primers (see Table S1 in the supplemental material) and reaction conditions described in references 41, 43, 44, 46, 49, 56, and 59. The sequences were compared to reference sequences deposited in GenBank. The identity of the R. massiliae isolate Rs1 was confirmed by PCR amplification and sequencing of all or a portion of the ompA, ompB, sca4, gltA, 16S rRNA, and 17-kDa antigen genes.
Purified rickettsiae and uninfected control cell lysates prepared as described above were embedded in low-melting-point agarose, digested with proteinase K in the presence of sodium lauryl sarcosine and 0.5 M EDTA, and subjected to PFGE as described previously (7). Following electrophoresis, the gels were stained in TAFE buffer (20 mM Tris-free base, 5 mM free acid EDTA, 0.00025% glacial acetic acid) with SYBR green 1 (Cambrex, Rockland, ME) to visualize DNA.
Rickettsial DNAs and uninfected ISE6 tick cell controls (mitochondrial DNA) were electrophoresed in PFGE gels, depurinated, and transferred to a Zeta Probe GT genomic membrane (Bio-Rad, Hercules, CA) as described previously (6). The blots were hybridized with digoxigenin-labeled probes (see below) at 55°C (16S mitochondrial, pRM6, pRM16, and pRM21 probes) or 50°C (pRM8, pRM12, pRM13, and pRM23 probes), washed, and exposed to Kodak X-OMAT AR film as described previously (6, 7). All probes were prepared by PCR amplification, using primers described below, with a DIG probe synthesis kit (Roche, Indianapolis, IN). Tick 16S mitochondrial gene probe reaction mixtures (11) included 50 ng of uninfected control ISE6 cell DNA as a template and the T1F and T1R primers (Table (Table1).1). The R. monacensis pRM gene probe reaction mixtures included 5 ng of pRM plasmid as a template and primer pairs (Table (Table1)1) derived from sequences of genes and pseudogenes on pRM (7), including pRM6, pRM8, pRM12, pRM13, pRM16, pRM21, and pRM23. PCR cycling conditions were as follows: 1 cycle at 95°C for 2 min; 40 cycles at 95°C for 30 s, 48°C for 30 s, and 72°C for 1 min; and 1 cycle at 72°C for 7 min. The exceptions were pRM16 and pRM21 reactions, which employed 50 and 52°C annealing temperatures, respectively, rather than 48°C. The pRM6, -8, and -16 primers were also used to amplify the cognate genes from rickettsial genomic DNA preparations for DNA sequence analysis. All primers were synthesized by Integrated DNA Technologies (Coralville, IA).
Because of the likely presence of mitochondria in rickettsial preparations purified by filtration from tick cell lysates, we investigated the potential presence in those PFGE gels of mitochondrial DNA bands that could be interpreted falsely as plasmid DNA bands. Uninfected tick ISE6 cell lysates contained DNAs that migrated as two SYBR green-stained bands, at approximately 30 and 40 kbp, relative to linear DNA markers (Fig. (Fig.1,1, left panel, center versus left lanes). Lysates of ISE6 cells infected with the R. monacensis transformant Rmona658B (7) contained DNAs that migrated at the same 30- and 40-kbp positions relative to the marker DNAs (right versus left lanes) as well as DNA bands corresponding to the circular, supercoiled, and linear pRM plasmid isomers (7), which migrated at approximately 50, 25, and 20 kbp relative to marker DNAs, respectively. Southern blots of the same gel hybridized with an I. scapularis 16S mitochondrial DNA probe showed that only the 30- and 40-kbp bands were recognized (center panel), while the rickettsial DNA bands hybridized with a probe specific for the GFPuv gene on the pMOD658 transposons integrated into the chromosome and plasmid (right panel). The results showed that host cell mitochondrial DNA was present in the rickettsia-infected cell lysates and migrated in PFGE gels as approximately 30- and 40-kbp bands relative to DNA markers.
We used PFGE and Southern blots to screen Rickettsia spp., including R. monacensis and R. felis, for the presence of plasmids. The ISE6 cell mitochondrial DNA bands were observed in all eight rickettsia preparations (Fig. (Fig.2A).2A). The R. monacensis and R. felis plasmid DNA bands migrated at approximately 25 and 50 kbp (pRM) and 60 to 70 kbp (pRF) relative to marker DNAs. The R. amblyommii isolate WB-8-2 lane contained prominent plasmid DNA bands with similar mobilities to those of the pRM bands, while a fainter plasmid band was present at approximately 50 kbp in the R. helvetica lane. In the R. hoogstraalii lane, plasmid DNA bands were present at approximately 50 to 70 kbp. In the R. massiliae lane, a plasmid DNA band was present at 25 kbp and a probable plasmid DNA band comigrated with the 40-kbp mitochondrial band, as judged by the intense staining of that band relative to the bands of the same size in the other lanes. The results indicated the presence of relatively large plasmids in R. felis and R. hoogstraalii, plasmids of intermediate size in R. monacensis, R. amblyommii, and R. helvetica, a smaller plasmid in R. massiliae, and the absence of plasmids in R. montanensis and R. peacockii (but see below).
We used Southern blots to confirm the above results and to gain a preliminary assessment of gene conservation among the plasmids. Replicate PFGE gels were Southern blotted and hybridized with probes for the R. monacensis pRM6, pRM16, and pRM21 plasmid genes, which are similar to the R. felis pRF genes and encode a small heat shock protein (Hsp), a DnaA-like replication initiator protein, and a cell surface antigen protein, respectively (7). We also used probes for the pRM12 and pRM13 genes, which are similar to chromosomal genes of other Rickettsia species and encode proline/betaine membrane transporter and SpoT stringent response proteins, respectively. Lastly, we used probes for the pRM8 and pRM23 genes, similar to genes of noncongener bacteria and encoding helicase RecD/TraA and transposon resolvase proteins, respectively.
The pRM16 probe hybridized to the R. monacensis, R. felis, R. amblyommii, R. helvetica, R. hoogstraalii, and R. massiliae plasmid DNA bands but did not hybridize to DNA of R. peacockii or R. montanensis (Fig. (Fig.2B).2B). The pRM6 probe hybridized to the plasmid DNA bands of all the rickettsiae except for that of R. massiliae and did not hybridize to DNAs of R. peacockii and R. montanensis (Fig. (Fig.2C).2C). The pRM21 probe hybridized more weakly than the pRM6 and pRM16 probes to all of the rickettsial plasmid DNA bands and to all chromosomal DNAs migrating above the position of the 100-kbp marker (Fig. (Fig.2D).2D). A strong hybridization signal in the R. felis lane above the position of the 100-kbp marker is also present in Fig. 2B and C. The R. felis chromosome does not contain a pRM16 homolog (23). Hybridization of the pRM16 probe to R. felis DNA migrating above the position of the 100-kbp marker may be due to the presence of a circular plasmid isomer unresolved from sheared chromosomal DNA, while the signal at 60 to 70 kbp probably represents hybridization to supercoiled and linear plasmid isomers.
The pRM8 probe hybridized to the plasmid DNA bands of R. monacensis, R. amblyommii, R. hoogstraalii, and R. helvetica, while the pRM12 and pRM13 probes hybridized only to the R. hoogstraalii plasmid and the pRM23 probe hybridized only to the R. monacensis plasmid (data not shown). PCR amplification of pRM6, pRM8, and pRM16 sequences was successful for the majority of the rickettsiae, and sequence similarities of the PCR amplicons to the cognate pRM sequences ranged from 82 to 95%, providing evidence that they were amplified from homologous genes.
PFGE and Southern blot analyses of R. peacockii maintained in cultured ISE6 cells for approximately 120 passages had indicated the absence of a plasmid (see above). However, PFGE analysis of early-passage (third and fourth passages) R. peacockii lysates revealed plasmid DNA bands at the approximately 25- and 50-kbp migration positions of the pRM plasmid and at the 60- to 70-kbp position of pRF (Fig. (Fig.3A).3A). Southern blot analyses showed that the pRM16, pRM6, and pRM21 probes hybridized to both the 25- to 50-kbp and 60- to 70-kbp plasmid bands (Fig. 3B, C, and D), while the pRM8, pRM12, pRM13, and pRM23 probes did not hybridize (data not shown). Further analyses of R. peacockii at passages 3, 39, and 66 showed that the plasmid bands were present in SYBR green-stained PFGE gels at approximately the same relative intensities (data not shown). The results suggested the presence of two plasmids in early-passage R. peacockii that were no longer present after over 100 serial passages in cultured tick cells.
PFGE analyses of R. amblyommii isolates WB-8-2, MOAa, AaR\Sc, and Ac\Pa, from Amblyomma ticks collected in Tennessee, Missouri, South Carolina, and Panama, respectively, showed that all possessed plasmids migrating as approximately 25- and 45- to 50-kbp bands relative to marker DNAs (Fig. (Fig.4A),4A), while additional bands that migrated at approximately 60 and 100 kbp were visible in lanes containing Ac\Pa and AaR\Sc lysates. Southern blots showed that the pRM16, pRM6, and pRM21 probes hybridized to the 25- and 45- to 50-kbp bands (Fig. 4B, C, and D). Interestingly, the pRM16 probe produced a signal at the 60-kbp band position in the WB-8-2, MOAa, and AaR\Sc lanes but not in the Ac\Pa lane, while the pRM6 probe produced 60- and 100-kbp band hybridization signals in the Ac\Pa lane but not in the other lanes. The results were consistent with the presence of a plasmid similar in size to pRM in all four R. amblyommii isolates, the presence of a larger plasmid carrying a pRM16 homolog but no pRM6 homolog in the WB-8-2, MOAa, and AaR\Sc isolates, and the presence of a larger plasmid in the Ac\Pa isolate, with a pRM6 homologue but no pRM16 homolog. The larger plasmid species may be multimer isomer forms of the smaller plasmids, but differential hybridization of the pRM6 and pRM16 probes suggests that they have at least limited sequence differences.
The PFGE gel and Southern blot results, summarized in Table Table2,2, indicated that homologs of DnaA-like replication initiator protein, small Hsp2, and Sca12 cell surface antigen genes were on all the rickettsial plasmids, with the exception of the R. massiliae plasmid. This plasmid lacked a small Hsp2 gene, was the smallest of the plasmids, and was the least well conserved relative to pRM. The plasmid of R. hoogstraalii that was isolated from an argasid tick was better conserved relative to pRM than were the plasmids from the other rickettsiae, all of which were isolated from ixodid ticks.
In this report, we have shown that plasmids are present in spotted fever group rickettsiae, including R. amblyommii, R. massiliae, and R. peacockii, and in “transitional group” members (23, 27), including R. felis (in which Ogata et al. first demonstrated plasmids ), R. helvetica, R. hoogstraalii, and R. monacensis. Plasmids thus exist in phylogenetically diverse members of the Rickettsia, and preliminary analysis of their coding capacities shows considerable variation (Table (Table2).2). The large proportions of transposon sequences and genes apparently transferred from noncongener bacteria on the fully sequenced pRM and pRF plasmids (7, 23, 35) indicate that rickettsial plasmids may be highly plastic vehicles providing access to a “horizontal gene pool” (53, 54). They might thereby mitigate the constrictive genetic effects exerted by population bottlenecks and reductive genome evolution in arthropod endosymbionts (2, 4, 12).
Of the rickettsiae analyzed, only R. montanensis isolate M5/6 did not contain plasmids. The M5/6 strain was isolated nearly 50 years ago (10) and has a long and uncertain passage history. It may once have contained a plasmid that was lost during serial passage in culture, similar to the loss of plasmids by serially cultured R. peacockii. We also obtained evidence for the presence of multiple plasmid species in some rickettsial isolates. The early-passage R. peacockii isolate and the four R. amblyommii isolates contained a smaller plasmid, similar in size to pRM, and larger plasmids that were probably 20 to 30 kbp greater in size, based on their relative mobilities in PFGE gels. The presence of more than one plasmid species in five rickettsia isolates resembled the original description of pRF and an apparent deletion form, pRFδ, in R. felis isolate California 2 (35). However, our present and previous analyses of R. felis isolate LSU have revealed the presence of only one plasmid species (7, 40).
We have shown plasmid loss in serially passaged R. peacockii and previously used the same method to demonstrate plasmid loss and rearrangements in Borrelia burgdorferi strain JMNT, which was serially passaged with cultured tick cells following its isolation from a hamster (29). The Borrelia plasmid changes were accompanied by changes in infectivity, consistent with similar observations made by others (45). Many functions essential for B. burgdorferi host interactions reside on its 21 linear and circular plasmids, and vertebrate-infectious strains isolated from nature usually have the full complement of plasmids (15). Plasmid loss reduces infectivity for tick hosts (52) but usually does not affect propagation in culture (16, 50). Loss of plasmids during serial culture of R. peacockii, which is maintained by transovarial transmission in the tick host and is not a vertebrate pathogen (34, 47), was suggestive of a possible role for plasmids in adaptation to the tick host. These results highlight the importance of using new isolates for host interaction and pathogenesis studies and may explain the apparent absence of plasmids from the genomes of sequenced pathogenic rickettsial isolates, such as Rickettsia rickettsii and Rickettsia conorii (12). R. felis (35) and R. massiliae (13) remain the two examples where published genome sequencing revealed the presence of plasmids.
Clues to the potential role of plasmids in rickettsial biology and host interactions may be drawn from comparisons of their coding capacities and the known interactions of other arthropod-borne microbes and hosts. We observed that all of the rickettsial plasmids, except for that of R. massiliae, encoded a small Hsp (Table (Table2).2). Both nonpathogenic arthropod endosymbionts and arthropod-borne pathogens of vertebrates have been demonstrated in numerous studies to express various Hsps in response to developmental or environmental changes in the host (20). The Hsps are molecular chaperones that are expressed in response to a wide range of stress conditions and have been defined classically as playing diverse roles in folding, assembly, stabilization, intracellular localization, secretion, function, and degradation of other proteins (22). However, the small α-crystallin-type Hsps (α-Hsps) are a particularly interesting group because they respond to temperature as well as pH, osmotic, and chemical stresses and may function in stabilization of membranes and nucleic acids in addition to proteins (32). In some plant symbiotic and arthropod endosymbiotic bacteria, α-Hsp genes are carried on plasmids (32, 33, 55), as they are on pRM, which contains two α-Hsp genes, pRM6 and -7, in an operon (7). A proteome analysis of R. felis showed that pRM6 and -7 homologs were the only plasmid-carried genes that were expressed detectably in Xenopus laevis XTC-2 cells (36).
In B. burgdorferi, two α-Hsps were expressed at much higher levels in low-passage than in high-passage strains (17), suggesting a possible host-adaptive function of the α-Hsps that was not necessary in serial culture. The rickettsial plasmid-carried α-Hsp genes might also play a role in host adaptation. Rickettsiae face significant changes in pH, osmotic pressure, CO2 and O2 levels, metabolite concentrations, and temperature during alternating periods of starvation and blood feeding in the tick host and upon transfer between arthropod and vertebrate hosts (30). The phenomenon of reactivation (24), in which rickettsiae in a quiescent noninfective state in unfed flat ticks resume cell division and regain infectivity during blood feeding, implies alterations in rickettsial physiology and gene expression. The location of genes encoding α-Hsps, in addition to membrane transport proteins, cell surface antigens, and unique rickettsial proteins of unknown function, on a plasmid that may be present in multiple copies per cell might facilitate enhanced transcription and expression of genes involved in adaptation to changes in host physiology. The example of B. burgdorferi again offers an illustration of how that might be achieved through a molecular mechanism other than plasmid copy number. The ospA, -B, and -C major outer surface protein genes of B. burgdorferi are carried on plasmids whose supercoiling state is influenced by temperature and are differentially expressed in response to environmental changes. The transcriptional activity of the osp promoters is regulated by DNA supercoiling, suggesting that the conformational state of the plasmids serves as a regulatory transducer of altered outer surface protein expression in response to environmental change (1). That regulatory effect may extend to other environmental changes typical of the tick life cycle, because plasmid supercoiling states in bacteria are regulated by DNA gyrase and topoisomerase activities. The activity of gyrase, in particular, is dependent on the intracellular [ATP]/[ADP] ratio, which in turn is influenced by such factors as altered osmolality and O2 tension (19, 25), which vary during the tick life cycle. We note that DNA gyrase genes as well as multiple copies of genes encoding ATP/ADP membrane translocases for importation of host cell ATP are present in rickettsial genomes that are distinguished by their otherwise spare coding capacity (42). Rickettsiae thus appear to be theoretically well equipped to follow the B. burgdorferi paradigm of transcriptional regulation of host-adaptive genes carried on plasmids through a mechanism that could simultaneously sense the energy state of the host cell and alter plasmid conformation. That possibility is lent support by the extensive evidence for convergent evolution of genetic mechanisms for antigenic variation among bacterial and protozoan vector-borne pathogens (8).
The existence of plasmids and mobile genetic elements in the rickettsiae suggests that they have significantly greater potential for genetic diversity and host adaptation than was previously believed (18). It is now necessary to determine the full extent of plasmid distribution and evolution in the genus Rickettsia, the mechanisms of plasmid maintenance, whether the plasmids are mobile, and their role in host adaptation and virulence.
This research was supported by NIH grant RO1 AI49424 to U.G.M.
Published ahead of print on 7 December 2007.
†Supplemental material for this article may be found at http://aem.asm.org/.