|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: YP AvdE. Performed the experiments: WK. Analyzed the data: AvdE. Contributed reagents/materials/analysis tools: YP VD DSAB KAJ EV MCJM DV. Wrote the paper: YP AvdE.
Chlamydia comprises a group of obligate intracellular bacterial parasites responsible for a variety of diseases in humans and animals, including several zoonoses. Chlamydia trachomatis causes diseases such as trachoma, urogenital infection and lymphogranuloma venereum with severe morbidity. Chlamydia pneumoniae is a common cause of community-acquired respiratory tract infections. Chlamydia psittaci, causing zoonotic pneumonia in humans, is usually hosted by birds, while Chlamydia abortus, causing abortion and fetal death in mammals, including humans, is mainly hosted by goats and sheep. We used multi-locus sequence typing to asses the population structure of Chlamydia. In total, 132 Chlamydia isolates were analyzed, including 60 C. trachomatis, 18 C. pneumoniae, 16 C. abortus, 34 C. psittaci and one of each of C. pecorum, C. caviae, C. muridarum and C. felis. Cluster analyses utilizing the Neighbour-Joining algorithm with the maximum composite likelihood model of concatenated sequences of 7 housekeeping fragments showed that C. psittaci 84/2334 isolated from a parrot grouped together with the C. abortus isolates from goats and sheep. Cluster analyses of the individual alleles showed that in all instances C. psittaci 84/2334 formed one group with C. abortus. Moving 84/2334 from the C. psittaci group to the C. abortus group resulted in a significant increase in the number of fixed differences and elimination of the number of shared mutations between C. psittaci and C. abortus. C. psittaci M56 from a muskrat branched separately from the main group of C. psittaci isolates. C. psittaci genotypes appeared to be associated with host species. The phylogentic tree of C. psittaci did not follow that of its host bird species, suggesting host species jumps. In conclusion, we report for the first time an association between C. psittaci genotypes with host species.
Chlamydia comprises a group of obligate intracellular bacterial parasites responsible for a variety of diseases in humans and animals, including several zoonoses. It was proposed in 1999 that the single genus of Chlamydia should be reassigned into two genera, Chlamydia and Chlamydophila, based on clustering analyses of the 16S rRNA and 23S rRNA genes , which has not been widely accepted by the chlamydial research community. Recently however, reversion to the single genus Chlamydia was recommended , with the Chlamydia nomenclature used here.
Chlamydia trachomatis can cause diseases with severe morbidity, such as trachoma, urogenital infection and lymphogranuloma venereum –. Several serovars and genotypes have been identified, but which have not been linked to disease or clinical outcome , . Chlamydia pneumoniae is a common cause of community-acquired pneumonia, bronchitis, pharyngitis and sinusitis . Although C. pneumoniae often causes mild or subclinical infections, its persistence in the host can lead to the establishment of chronic pathologies and has been implicated with arteriosclerosis  and coronary heart diseases , . Chlamydia psittaci which can cause zoonotic pneumonia in humans are usually hosted by birds –. Transmission of C. psittaci from birds to humans is frequently reported and veterinarians, poultry farmers, bird breeders and pet shopkeepers are in particular at high risk –.
Chlamydia abortus has been associated with abortion and fetal death in mammals, including humans, and is hosted by goats, sheep and less frequently by cattle, horses and pigs . The microorganism has also been associated with pneumonia, conjunctivitis, arthritis as well as epididymitis and has been isolated from the faeces of healthy sheep and goats –. Pregnant women are at risk when exposed to animals infected with C. abortus and may suffer severe infections, including spontaneous abortion –.
Multi-locus sequence typing (MLST) based on the partial sequences of seven housekeeping genes was first used to evaluate the population structure of Neisseria meningitidis . Previously, we developed an MLST scheme to understand the population genetic structure of C. trachomatis and C. pneumoniae and the diversity of these species and to evaluate the association between genotype and disease . In the present study, the MLST scheme was used to evaluate the population structure of C. psittaci and C. abortus. Results of cluster analyses of concatenated sequences of the 7 housekeeping fragments expanded and validated the proposed typing system for all chlamydial species. The results indicated that the 7 housekeeping fragments used in our study were likely representative for the whole genome sequence. C. psittaci genotypes were associated with their host species; the C. psittaci phylogenetic tree however, did not follow that of its host bird species. Furthermore, C. psittaci 84/2334, formerly considered as the missing link between C. abortus and C. psittaci, was clearly typed as C. abortus.
Among 132 Chlamydia isolates 44 sequence types (STs) were identified; 19 STs among 60 C. trachomatis isolates, 4 STs among 18 C. pneumoniae isolates, 12 STs among 34 C. psittaci isolates and 4 among 16 C. abortus isolates (Table S1). Previously, it was shown that in C. trachomatis the homologous recombination rate is low . We determined the index of association for the different Chlamydia species according to Haubold . Significant linkage disequilibrium was detected for C. psittaci, C. abortus, C. trachomatis and C. pneumoniae (Table S2). In addition, we tested the sequences for evidence of recombination using the maximum chi square . Recombination events were not detected in C. pneumoniae and C. abortus sequences. In C. trachomatis, 5 putative recombination events were detected between a pair of hflX alleles (putative recombination site at position 422) and 4 oppA alleles (putative recombination sites at positions 500 and 506), respectively. In C. psittaci, we detected one putative recombination event in hemN (at position 426) (Table S3). Together these data suggest that the role of recombination in diversity of Chlamydia species is low and that Chlamydia species are clonal. In phylogenetic analyses with clonal bacterial species, where mutation is more important than recombination, it is preferable to use concatenated sequences of the MLST loci rather than allelic profiles, because the magnitude of changes between alleles is lost in allelic profiles . This method was subsequently used to study relatedness of closely related species , , . Phylogenetic analysis of all 132 strains of Chlamydia using the Neighbour-Joining algorithm with the maximum composite likelihood model of the single 3120 base pairs (bp) sequence of the aligned concatenated loci resulted in a tree comparable to that obtained with 16S rRNA gene and 23S rRNA gene sequences  and to the previous reported tree based on concatenated sequences of 6 MLST loci  (Fig. 1). In the present tree, three main groups are identified among 60 C. trachomatis isolates, consistent with earlier reported results of analyses of 26 C. trachomatis isolates (not shown) . In addition, C. pneumoniae LPCoLN isolated from koala branched separately from C. pneumoniae from patients, consistent with the results of Myers and colleagues . C. psittaci M56 from a muskrat branched separately from the main cluster of C. psittaci isolates, while C. psittaci 84/2334 grouped together with C. abortus, suggesting that 84/2334 belongs to C. abortus.
Concatenated sequences of the MLST loci were used to estimate the diversity among and divergence between Chlamydia species. The diversity among C. trachomatis and C. pneumoniae is limited (Table 1). Nucleotide substitutions in protein encoding genes can be either synonymous or non-synonymous (resulting in a changed amino acid). Darwinian selection may lead to the retention of non-synonymous substitutions. The number of synonymous and non-synonymous substitutions may indicate the degree of selection operating in the population. The average number synonymous substitutions per synonymous site dS in C. trachomatis and C. pneumoniae were comparable, while the average number of nonsynonymous substitutions per nonsynonymous site dN was three times higher among C. trachomatis (Table 1). However, considering C. pneumoniae from humans only, than dS in C. pneumonia was much lower than that in C. trachomatis and dN of C. pneumonia even 7 fold lower than that of C. trachomatis, indicating that C. pneumoniae from human is much more clonal than C. trachomatis in accordance with previous observations .
The dS and dN among the 34 C. psittaci isolates was much higher than among any other Chlamydia included in this study (Table 1). However, the dS of C. psittaci without M56 and 84/2334 was considerably lower, but still higher than among C. pneumoniae, C. abortus or C. trachomatis. The dS and dN among the 16 C. abortus isolates were comparable to those among the C. pneumoniae isolates, reconfirming the homogeneity observed in this species, in accord with previous results . Inclusion of C. psittaci 84/2334 to C. abortus increased the diversity among the species, which remained lower than the diversity among C. psittaci without M56 and 84/2334 (Table 1). The degree of selection can be expressed by the dN/dS ratio; a ratio dN/dS higher than 1 indicates a positive selection (altered amino acid substitutions are common), while a rate lower than 1 indicates negative selection (silent substitutions). The C. trachomatis and C. pneumoniae have the highest dN/dS, but in both species synonymous substitutions are in excess (Table 1).
The divergences between Chlamydia species adjacent in the tree can be expressed as the average number of nucleotide substitutions per site between two species populations (DXY) and the number of fixed differences, i.e. mutations uniform within the species populations. Since most fixed differences are neutral and accumulate at a rate proportional to the mutation rate (molecular clock), the number of fixed differences are indicative for the elapsed time since two populations evolved from a common ancestor. The DXY between Chlamydia species closely positioned in the N-J tree were similar with two exceptions (Table 2). First, the divergence between C. pneumoniae and C. pecorum was twofold higher than that between other neighbours in the tree. Of note, the DXY and the number of fixed differences between C. pneumoniae from human and LPCoLN was small and consistent with the results obtained by comparisons of available C. pneumoniae genome sequences by Myers and colleagues . Second, the DXY between C. psittaci and C. abortus was remarkably low and the number of fixed differences was even lower when compared to that between C. pneumoniae isolated from human and LPCoLN (Table 2). In addition, only C. abortus and C. psittaci shared polymorphisms. However, while the DXY between C. psittaci without 84/2334 and C. abortus with 84/2334 was similar to that between the whole C. psittaci and C. abortus populations, the number of fixed differences was considerably higher, but still much lower than between other neighbouring species in the tree. In addition, shared polymorphism were absent between C. psittaci without 84/2334 and C. abortus with 84/2334. The divergence between 84/2334 and C. abortus was limited and of the same magnitude as that between human C. pneumoniae and LPCoLN, supporting the notion that 84/2334 and C. abortus are one species.
The N-J tree clearly showed that C. psittaci M56 branched separately from the main group of C. psittaci strains isolated mainly from birds, while C. psittaci 84/2334 grouped together with C. abortus isolates (Fig. 1 and and2).2). Cluster analyses of the 7 individual housekeeping gene fragments showed that with four gene fragments M56 grouped close to the remaining (except 84/2334) C. psittaci isolates. With gene fragments gatA, hflX and fumC M56 grouped neither with C. psittaci nor with C. abortus. Cluster analyses of the 7 individual housekeeping gene fragments showed that in all instances 84/2334 grouped together with C. abortus (Fig. S1). In addition, the hemN sequence of 84/2334 was identical to that of all but one C. abortus isolates. Together, with the analyses of the divergence between C. abortus and C. psittaci (Table 2) these results suggest that 84/2334 belongs to C. abortus.
Using BURST ST24 and ST28 were defined as singletons, although 13 isolates from parrots/parakeets were ST24 (group I) and 8 isolates (7 from ducks and one from human) were ST28 (group II). Two clonal complexes or groups were defined: group III comprises 5 isolates from pigeons and one isolate from human and group IV was formed by the two isolates from turkey (data not shown).
Results of cluster analyses performed with concatenated sequences of MLST loci of C. psittaci excluding 84/2334 and M56 showed association between host species and C. psittaci genotype with the aforementioned four groups being recognized (Fig. 3). One isolate, C. psittaci VS225 from a parakeet, was found to lie between the group I isolates from parrots/parakeets and the group II isolates from duck. The C. psittaci WC bovine isolate did not group with any of the four main groups. Clustering was not associated with the geographic origin of the isolates or of their corresponding host species as parrots/parakeets which have their origin in three different continents, grouped together (Table S1).
In a previous study we used the present MLST scheme to analyze clonal groupings among C. trachomatis and C. pneumoniae strains . A database hosted at http://pubmlst.org/chlamydiales/was developed and sited at the University of Oxford. The website offers a large number of ways to query the database and to further break down and export the results . In the present study we used the MLST scheme to explore the population genetics of C. psittaci and C. abortus isolates. The main findings are that C. psittaci genotypes are associated with their host species and that isolate 84/2334, formerly classified as C. psittaci, is most likely typed as C. abortus.
We have shown in our previous study that an UPGMA tree produced from the allelic profiles and from concatenated allele sequences of 28 C. trachomatis isolates resulted in three groups of sequence types . The urogenital strains were distributed over two separated groups; one consisted solely of strains with the frequent occurring serovars E, D and F. Strains isolated from patients with lymphogranuloma venereum (LGV strains) grouped in a single cluster, which also included C. trachomatis B/TW5, although not being identical. The close relatedness of B/TW5 was supported by results of comparisons of IncA sequences, showing that B/TW5 shares IncA polymorphisms with LGV strains, which were not found among other serovars . The present study, with C. trachomatis strains supplemented with isolates from a study by Ikryannikova and colleagues  to a total of 60, yielded the same three groups in cluster analyses. Recently, another MLST scheme was described showing clonal groupings among C. trachomatis strains with a group consisting exclusively of LGV strains . These results differ somewhat from those reported in our previous study .
Recently, a phylogenetic tree was reported that was based on concatenated sequences of 110 conserved protein sequences extracted from the available whole genome sequences of C. trachomatis, C. pneumoniae, C. pecorum, C. felis, C. psittaci, C. abortus and C. caviae , . Our tree that is based on the concatenated sequences of 7 housekeeping fragments is very similar to the original tree. In addition, we showed that in our tree C. pneumoniae LPCoLN isolated from koala branched separately from the remaining human C. pneumoniae strains consistent with the results reported by Myers and colleagues who compared whole genome sequences of LPCoLN and C. pneumoniae AR39, TW-183, CWL029 and J138 isolated from human providing evidence that humans were originally infected zoonotically . Together, these results indicate that the 7 housekeeping fragments used in our study are representative for the whole genome sequence.
Phylogenetic analyses of the concatenated allele sequences of C. psittaci revealed an association between C. psittaci genotype and host species. Genotype was not associated with geographic origin, as the host birds had their natural habitat in different countries and covering four different continents, yet the C. psittaci isolated from them were of a single genotype. This identified association between C. psittaci genotype and host species has not been previously observed, which may due to the difference in sequences studied. Cluster analyses of C. psittaci has previously been performed using ompA , , ompA based PCR-RFLP  or multi loci variable number of tandem repeats (MLVA) . ompA encodes the major outer membrane protein and is subjected to host immuno-pressure and its sequence may therefore not reflect the genetic make up of Chlamydia. MLVA based on genetic variation of tandem repeats may be also less suitable to assess evolutionary history by phylogenetic analyses, because it lacks sequence information.
Recently Hackett and colleagues examined ~32 kilobases of aligned nuclear DNA sequences from 19 independent loci for 169 bird species, representing all major extant groups, and recovered a robust phylogeny from a genome-wide signal supported by multiple analytical methods . In their phylogenetic tree based on Maximum Likelhood analyses ducks and turkeys group relatively close together, while pigeons and parrots are more distantly related. In contrast, our results show that C. psittaci from ducks grouped closer to those from parrots and more distantly from those isolated from turkeys, indicating that C. psittaci phylogeny did not follow that of their host species birds. Possibly, C. psittaci may have been transferred from Psittaciformes (order of birds to which parrots and parakeets belong) to Ansiformes (order of birds to which ducks belong) or vice versa. Alternatively, C. psittaci has been transferred to different bird species more than once from a source other than birds. Both scenarios imply host species jump. Incidental transmission is not likely because C. psittaci isolates from these birds and from ducks were isolated in different countries and as far as known in different years. In addition, C. psittaci M56 from a muskrat has it own branch in the phylogenetic tree obtained by using concatenated sequences of MLST loci similar as in a split network graph based on a sequence similarity matrix of ompA sequences reported by Sachse and colleagues . This and the observation, that C. psittaci WC from a cow also does not group with C. psittaci from birds, might indicate that C. psittaci may have spread among mammals, allowing for the development of mammal species specific C. psittaci genotypes as well. Analyses of more C. psittaci isolates from different host species will shed more light on this interesting question. Recently Mitchell and colleagues showed that C. pneumoniae found among animals is more divers than those found among humans . In addition, their results support two separate animal-to-human cross species transfer events in the evolutionary history of this pathogen. Although C. psittaci has a broad host range and numerous cases of transmissions from birds to humans have been described , reports of human to human transmission of C. psittaci are rare , , indicating that C. psittaci has not yet adapted to the human host.
Of note, two C. psittaci strains were isolated from human, suggesting that these were incidental zoonotic infections, without establishing C. psittaci in humans. C. psittaci CPMN isolated from human and thereafter repeatedly passaged in ferrets  and C. psittaci humaan E strain grouped together with strains isolated from pigeons and from ducks, respectively. Interestingly, the latter case was a turkey farmer, of whom pharyngeal and nasal swab were taken before the arrival of a new flock of turkeys arrived at the farm and repeatedly thereafter . Initially, C. psittaci with ompA genotype E was cultured, but after three to six weeks ompA genotype A was found, the same ompA genotype found among the turkeys. Our data suggest that the initial infection may be acquired from waterfowl.
The 16 C. abortus isolates comprised 4 STs (5 when 84/2334 is included). Of note, strains LLG and POS have the same genotype and grouped separately from the remaining C. abortus isolates, consistent with the results of Laroucau and colleagues using MLVA  and immunological analysis . Furthermore, MLST analyses in this study differentiated the vaccine strain 1B from its parental strain AB7, both classified as the same genotype by MLVA . The 4 C. abortus genotypes were not grouped according to their host species (sheep or goat).
The much lower divergence between C. psittaci and C. abortus compared to that between the other species, suggests C. psittaci and C. abortus have diverged from a common ancestor and much more recently than the others in the tree have from their common ancestor. Alternatively, C. psittaci might have been transmitted from birds to ruminants and adapted to the new hosts as already suggested by Pudjiatmoko and colleagues . In our study, host species jump by C. abortus may be indicated by C. psittaci 84/2334, which grouped close to all C. abortus strains, showing limited divergence with C. abortus with only 7 fixed nucleotide substitutions and should therefore be classified as C. abortus. Originally, C. psittaci 84/2334 was isolated from a yellow-crown amazon (parrot) and classified as C. psittaci based on its reaction with specific sera against the major outer membrane protein (MOMP). However, among 60 C. psittaci isolates from birds, strain 84/2334 had a unique ompA AluI restriction pattern, indicating that this strain differs from the majority of C. psittaci strains . It was suggested that C.psittaci 84/2334 was an intermediate between C. psittaci and C. abortus based on analysis of ompA, rnpB and the rrn (part of the region between the 16S rRNA 23SrRNA genes) sequences . In addition, C.psittaci 84/2334 was found to have DNA sequences that were identical to an extrachromosomal plasmid in duck C. psittaci strain N352, while extrachromosomal plasmids are not found in C. abortus strains . However, trees based on full length 16S rRNA and 23S rRNA gene sequences showed that C. abortus strains grouped together with C. psittaci strains . Similar results were obtained with cluster analyses based on 390 bp rnp sequences. In an N-J tree based on these sequences C. psittaci 6BC and C. abortus B577 were indistinguishable . Cluster analyses of rnp sequences extracted from GenBank showed C. psittaci 84/2334 grouped closer to C. abortus isolates than to C. psittaci isolates (data not shown). Also, the difference in the 222 base pairs rrn sequence between C. psittaci strains and C. abortus strains is limited. C. psittaci 84/2334 differs at only one position with 5 C. abortus isolates of which their rrn sequences are publicly available. C. psittaci 84/2334 and the 5 C. abortus isolates differ at 3 to 4 positions with C. psittaci strains. Again, in a tree based on rrn sequences extracted from GenBank, C. psittaci 84/2334 grouped closer to C. abortus than to C. psittaci (data not shown). In addition, an extended rrn sequence of 315 bp of C. psittaci 84/2334 also differed at only one position from the C. abortus sequences . In our cluster analyses we used up to 10-fold more sequences information than the single 16S RNA, 26S rRNA, rnn and rnp sequences and showed clear separation between C. psittaci isolates and C. abortus isolates with one exception: C. psittaci 84/2334 grouped together with C. abortus. Cluster analyses of the individual housekeeping fragments showed that in all instances strain 84/2334 grouped together with C. abortus. Nevertheless, for six of the seven housekeeping fragments strain 84/2334 has a unique allele. This could indicate that strain 84/2334 is unique within the group of C. abortus isolates. In addition, only C. abortus from a limited set of host (goats and sheep) have been included in the analyses. Inclusion of more C. abortus isolates from different hosts will most likely result in more C. abortus genotypes some of these maybe more related or identical to that of C. psittaci 84/2334. Ultimately, whole genome sequences of more C. psittaci strains isolated from different host including C. psittaci 84/2334 will be ideal to assess these questions.
In conclusion, C. psittaci genotypes are associated with host species and C. psittaci 84/2334 should be reclassified as C. abortus.
All 132 Chlamydia isolates (currently present in the MLST data base for Chlamydiales at http://pubmlst.org/chlamydiales/) were included in the study (Supplementary Table S1). This includes 26 C. trachomatis and 16 C. pneumoniae strains from a earlier study  and 30 C. trachomatis strains submitted to the MLST database by Dr. Ikryannikova, who kindly gave her permission to use the data . Detailed analyses were performed with 34 C. psittaci strains isolated from different bird species and mammals from different geographic locations and 16 C. abortus strains. For comparison, sequences of C. caviae, C. felis, C. pecorum and C. muridarum were also included . C. psittaci strains were isolated form dead animals brought to the veterinary clinic for autopsy, or from samples sent to the laboratory by veterinary practitioners for chlamydial diagnosis. C. abortus B577 (VR656) was purchased from ATCC. Greek C. abortus strains were isolated from infected placentae or aborted fetuses. C. abortus MA/231184, MB/312, ME/4004, MF/337, MD/3920, FAS, FAG, VPIG and the variant strains LLG and POS have been previously described, as well as the reference strains AB7, A22, and S26/3 , . The vaccine strain 1B (Enzovax) was purchased from Intervet (Intervet-Hellas) and strain Krauss-15 was kindly provided by Dr Jones (Moredun Research Institute, UK). C. psittaci was cultured in Buffalo green monkey cells (ATCC CCL-26) identifying the organisms by the IMAGEN™ Chlamydia immunofluorescence staining (Dakocytomation, Denmark), as previously described  and C. abortus stocks were produced in McCoy cell monolayers in the continuous presence of cycloheximide (1 µg/ml) and subsequently typed with monoclonal antibodies against the major outer membrane protein (MOMP) and the polymorphic membrane proteins (Pmps) as previously described .
DNA extraction, PCRs and DNA sequencing were performed as previously described . All alleles of the partial sequences of the 7 housekeeping genes are accessible via the MLST website for Chlamydiales (http://pubmlst.org/chlamydiales/). New allele sequences have been deposited in GenBank (accession no.: HM776459-HM776511).
Sequences of fragments from seven housekeeping genes (enoA, fumC, gatA, gidA, hemN, hlfX, oppA) were analyzed as previously described . Allele numbers and genotypes were identified at http://pubmlst.org/chlamydiales/. Clonal complexes were identified using BURST in the START2.0 software package at http://pubmlst.org/software/analysis/start2/. Clonal complexes consisted of sequence types that shared 6 of 7 alleles with at least 1 other sequence type in the complex. The index of association for the different Chlamydia species according to Haubold  and sequences were tested for recombination using maximum chi square  with the START2.0 software package.The number of synonymous and non-synonymous substitutions per site was determined using DnaSP 4.0 . Phylogenetic and molecular evolutionary analyses of the seven housekeeping fragments (individually or concatenated) were conducted using MEGA version 4  or SplitsTree version 4.0 , generating a Neighbor-Joining (N–J) tree using the Maximum Composite Likelihood model.
All isolates used in this study according to Chlamydia species, host and ST.
(0.12 MB XLS)
Detecting linkage disequilibrium in MLST data of Chlamydia species.
(0.03 MB DOC)
Detection of putative recombination events in Chlamydia species.
(0.04 MB DOC)
The authors like to acknowledge stimulating discussion with Dr. A. Bart. Melanie Nguyen is acknowledged for technical assistance. Dr. W.A. Paxton is acknowledged for critical reading of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
Funding: The development of the Chlamydiales MLST web site has been funded by the Wellcome Trust. Veerle Dickx is funded by the Federal Public Service of Health, Food Chain Safety and Environment (convention RF-6177). This research was partly funded by the Sixth Framework Programme of the European Commission, Proposal/Contract no.: 512061 (Network of Excellence ‘European Virtual Institute for Functional Genomics of Bacterial Pathogens’ [http://www.noe-epg.uni-wuerzburg.de]) and partially by the Belgian Federal Public Service of Health, Food Chain Safety and Environment (convention RF-6177). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.