Chlamydia trachomatis virulence factor CT135 is stable in vivo but highly polymorphic in vitro

Abstract

Chlamydia trachomatis is an important human pathogen causing both ocular and sexually transmitted disease. Recently, we identified CT135 as an important virulence determinant in a mouse infection model. Results from CEL 1 digestion assays and sequencing analyses indicated that CT135 was much more polymorphic in high in vitro passage reference serovars than it was in clinical strains that had undergone limited passaging. Herein, we used targeted next-generation sequencing of the CT134–135 locus, from reference strains and clinical isolates, enabling accurate discovery of single nucleotide polymorphisms and other population genetic variations. Our results indicate that CT134 is stable in all C. trachomatis serovars examined. In contrast, CT135 is highly polymorphic in high-passaged reference ocular and non-LGV genital serovars, with the majority of the mutations resulting in gene disruption. In low-passaged ocular clinical isolates, CT135 was frequently disrupted, whereas in genital clinical isolates CT135 was intact in almost all instances. When a serovar K isolate, with an intact CT134 and CT135, was subjected to serial passage in vitro CT134 remained invariable, while numerous gene interrupting mutations rapidly accumulated in CT135. Collectively, our data indicate that, for genital serovars, CT135 is under strong positive selection in vivo, and negative selection in vitro.

CT135 is a critical Chlamydia trachomatis virulence factor that undergoes rapid inactivation upon in vitro passage.

Graphical Abstract Figure.

CT135 is a critical Chlamydia trachomatis virulence factor that undergoes rapid inactivation upon in vitro passage.

Chlamydia trachomatis is an obligate intracellular bacterium that is a leading cause of preventable blindness and sexually transmitted diseases worldwide. Chlamydia trachomatis natural infection tropism is limited to humans. As such, non-human primate and murine models have been extensively used to define the biology of infection, disease pathogenesis and correlates of protective immunity (Morrison and Caldwell 2002; Kari et al. 2011). Historically, a significant limitation to the identification of chlamydial virulence factors was the lack of a system for genetic manipulation. Recent advances arising from whole genome sequence data for comparative genomics (Bachmann, Polkinghorne and Timms 2014) the development of a plasmid-based transformation system (Wang et al. 2011), and an insertional mutagenesis procedure (Johnson and Fisher 2013) has enhanced opportunity for identifying chlamydia virulence factors. Two such determinants are the chlamydial plasmid (Song et al. 2013) and CT135 (Sturdevant et al. 2010). CT135 encodes a hypothetical protein (MW 38.3 kDa) with four membrane-spanning domains, indicative of it being an inclusion membrane protein (Inc) (Lutter et al. 2012). CT135 is encoded in an operon together with CT134 (MW 14.9 kDa), which also contains membrane-spanning domains and is a predicted Inc (Dehoux et al. 2011; Lutter et al. 2012).

Recently, we showed that mice infected with an in vitro highly passaged population of C. trachomatis serovar D UW-3 resulted in early (D-EC) and late (D-LC) infection clearance phenotypes (Sturdevant et al. 2010). Strains of each phenotype were isolated and plaque cloned for whole genome sequencing (WGS). The WGS data indicated that the strains were genetically identical, except for single nucleotide indels in CT135. Thus, mutations in CT135 are sufficient to attenuate the in vivo virulence of C. trachomatis serovar D in the female mouse genital tract. The purpose of the current study was to undertake an in-depth genetic analysis of the CT134-135 locus in C. trachomatis reference stocks and clinical isolates by employing next-generation sequencing to elucidate single nucleotide polymorphisms (SNPs) and other population genetic variations.

Primers were designed to span CT133 (CT133F 5^′-TGAGGACCGTTTGGTGTTCCG) and CT136 (CT136R 5^′-TCTATAAATTCTAT CCCAGCAAGC) and used with all C. trachomatis genomic DNA templates, in PCR reactions employing KAPA Hi Fi HotStart (KAPA Biosystems) polymerase. The resulting amplicons were subjected to deep sequencing as random NextTera XT shotgun libraries applying the Illumina MiSeq platform, according to the manufacturer's protocols. Depth of read coverage varied from 2000 to 25 000. Reads were referenced mapped to C. trachomatis serovar D UW-3 chromosome sequence (Stephens et al. 1998; NCBI NC_000117.1), wherein both CT134 and CT135 are annotated as intact. Gapped-read alignments were achieved with Bowtie 2 (Langmead and Salzberg 2012) and SNPs were called using Freebayes (Blankenberg et al. 2014) and TRAMS version 1.0.0. In all cases, high-quality SNPs indicated were supported by a Phred quality score of ≥ 3. To assess for non-specific PCR-induced mutations, the serovar L2 PCR amplicon was cloned into pCR-XL-TOPO (Life Technologies), transformed into E. coli TOP10 and DNA from a single colony was used as template for PCR, using the CT133F and CT136R primers. The resultant amplicon was deep sequenced and mapped to the L2 chromosome sequence (Thomson et al. 2008; NCBI NC_010287.1); no mutations were detected in either CT134 or CT135, suggesting low frequency of PCR-induced background mutations. Thus, the mutations seen in the deep-sequencing data set are fundamentally true mutations existing in a non-clonal C. trachomatis serovar population.

Initially, the amplicon deep-sequencing procedure was applied to 15 reference stock C. trachomatis elementary body (EB) populations from our culture collection. The majority of these strains were isolated 40 plus years ago and have been maintained in our lab since the 1980s. All stocks have been passed multiple times, initially in eggs and more recently in HeLa or McCoy cells.

The deep-sequencing data indicate that CT134 is highly conserved across all C. trachomatis serovars, and predicted as intact in all cases. The majority of SNPs (14 of19) are synonymous (Fig. 1). Ocular serovars revealed only three SNPs in total, all of which are synonymous. Genital serovars revealed six SNPs in total, with five synonymous and one non-synonymous. LGV serovars revealed 10 SNPs, with six synonymous and four non-synonymous.

Figure 1.

SNPs in CT134 and CT135 from C. trachomatis reference serovars as determined by deep sequencing. Primers designed to span CT133 and CT136 were used with all C. trachomatis genomic DNA templates, in PCR reactions. The resulting amplicons were subjected to deep sequencing as random NextTera XT shotgun libraries applying the Illumina MiSeq platform. Reads were referenced mapped to C. trachomatis serovar D UW-3 chromosome sequence (Stephens et al. 1998, NCBI NC_000117.1). SNP position are shown using the C. trachomatis D UW-3 genomic sequence numbering as the reference. CT134: 151671–152084 and CT135: 152143–153225. Not shown is a single intragenic region LGV serovar-specific SNP, a TA>T deletion, which occurs at position 152119. In all cases, high-quality SNPs indicated were supported by a Phred quality score of ≥3.

CT135 is more divergent, revealing considerable polymorphism (Fig. 1, Table S1, Supporting Information). The gene is predicted as intact in all LGV serovars with 26 non-frameshifting SNPs in total, 16 being non-synonymous. In contrast, ocular serovars were consistently detected with no intact CT135 coding sequence (CDS) (24 SNPs overall: four synonymous, eight missense, five non-sense and seven indel frameshifts; two insertions and five deletions). Strikingly, and in contrast to CT134 with no CDS-interrupting mutations, 12 of 24 CT135 mutations are CDS disrupting. From this collective evidence, we infer strong negative selection to disrupt the CT135 locus in all ocular serovars when cultured in vitro. Within the genital biovar, only two serovars (G and K) encode intact CT135 devoid of SNPs, when compared to the annotated chromosome reported by Stephens et al. (1998). The remaining six genital serovars (D, E, F, H, I and J) all have CDS-disrupting SNPs (42 SNPs total; 6 synonymous, 14 missense, 6 nonsense and 16 indel frameshifts; 7 insertions and 9 deletions). Serovar I is heavily mutated harboring 27 SNPs: only 3 synonymous and 14 CDS disrupting. Nonsense mutations included all three of the stop codon variants; ochre, amber and opal. Resembling the ocular serovars, there is a clear trend for genital serovars towards CT135 interrupting mutations (22 of 42 total SNPs). As limited knowledge exists regarding CT135 function, we cannot predict which substitutions may impact protein function. Thus, our determinations of CT135 CDS-inactivating mutations are likely minimum estimates.

Previously, we reported on two plaque clones, D-EC and D-LC, that were isolated from mice infected with our parental seed stock, containing a mixture of organisms varying in virulence (Sturdevant et al. 2010). D-EC and D-LC carried unique mutations in CT135. These same two mutations are seen in our data set for the serovar D population dataset, specifically for D-LC an AT>A frameshift deletion at base 152275 and for D-EC a C>CT insertion at base 152 685 (Fig. 1). In addition to these two indels, we detected two other CDS-disrupting mutations, confirming that our serovar D seed stock is indeed polymorphic at the CT135 locus. We generated amplicons from our current seed stocks of D-EC and D-LC and despite > 10 additional passages in vitro, both remain clonal retaining their original indel identified in CT135. This strongly suggests that once the CT135 locus is functionally compromised in a particular clone there is limited additional variation.

We leveraged the fact that CT135 in our serovar K stock lacked mutations and asked whether further passage in vitro would result in the accumulation of CT135 mutations. The serovar K stock was cultured in McCoy cells grown in D-MEM high glucose medium plus 10%FBS and 1 μg ml⁻¹ cycloheximide. Chlamydia were harvested at 48 h p.i. and immediately passed onto a fresh McCoy cell monolayer. This process of culture, harvest and reinfection was continued for nine additional passages in total. DNA was extracted from EBs after each passage, used as PCR template and resultant amplicons subjected to deep sequencing. No mutations were detected in the CT134 locus at any of the nine passages (Fig. 2). Whereas no mutations in CT135 were detected in any of the first four passages, by passage five (p5) a nonsense mutant, an insertion mutant as well as a deletion mutant were already detected (Fig. 2, Table S2, Supporting Information). Thereafter, detected SNPs incremented at each subsequent passage, reaching 16 by p9, but without any synonymous mutations (three missense, four nonsense and nine indel; three insertions and six deletions). Notably the mutations that occur at each passage were retained following further passage (i.e. three SNPs detected in p5 were present in each subsequent pass). This SNP pattern, heavily biased toward CT135 CDS disruption, is strikingly similar to that seen in other reference serovars. Curiously, few instances of repeat mutations at a single chromosomal base were detected; this suggests that despite being particularly targeted for insertional/disruptive mutation, no mutational hot spot exists within the gene.

Figure 2.

SNPs in CT135 in serially passaged serovar K UW-31 as determined by deep sequencing. Serovar K, initial start passage number 40, was serially passed every 48 h in McCoy cells with DMEM-10 high glucose 10% FBS + 1 ug ml⁻¹ cycloheximide for nine additional passes. Genomic DNA was prepared from EBs from each passage and used for PCR and the resulting amplicons were subjected to deep sequencing and SNP analysis as described in the legend to Fig. 1. In all cases, high-quality SNPs indicated were supported by a Phred quality score of ≥3.

Taken together, these data support a unique genetic mechanism wherein CT135 targeted mutation is occurring in individual chlamydial organisms within the collective microbial population. In part, this could explain why most of the reported C. trachomatis genome sequences, which are generated from non-clonal populations, all report intact CT135 loci. (i.e. the consensus readout of a mixed base population from individual independent mutations would appear as an intact CDS). Interestingly, sequencing data indicate that this phenomenon also occurs with high-passaged reference stocks of C. muridarum. The initial C. muridarum strain Nigg genome sequence was determined from a high-passage non-clonal population, and the CT135 homolog (TC0412) was reported as intact (Read et al. 2000). Subsequent studies have clearly shown that plaque-cloned organisms from this population have a variety of TC0412 CDS-disrupting SNPs (Ramsey et al. 2009; Chen et al. 2015; Jasper et al. 2015). Unfortunately, no recent field isolates of C. muridarum are available for sequence analysis; thus, we do not know if TC0412 is intact in vivo.

To extend our findings obtained with C. trachomatis reference serovars, we next looked for polymorphisms within CT134/135 in low-passaged naturally occurring ocular and genital clinical strains. For this we chose a random sampling of 42 genomic DNA templates from our large collection isolated from low-passaged cultures, obtained from distinct geographic areas and previously applied to map mutations in the tryptophan synthase locus (Caldwell et al. 2003). We generated amplicons from 10 ocular, 30 genital and 2 LGV biovar isolates for deep sequencing.

The two L2b isolates (LGV biovar) had intact CT134 and CT135 sequences. Of the 10 ocular serovar A and B isolates analyzed, both CT134 and CT135 were intact in only 2 isolates. CT134 was intact in all but one of the remaining eight isolates. That isolate had an 1106-bp deletion (serovar D UW-3 bases 150968–152074) that removed the N-terminus of CT133, a predicted methyl transferase, and all but the final eight bases of CT134. The deletion boundary had an eight base (AAAGATCG) direct repeat at each end, suggesting a recombination-mediated deletion event. Although the CT135 CDS is predicted intact in this isolate, given that CT134 and CT135 are cotranscribed as part of an operon (Albrecht et al. 2010; Dehoux et al. 2011), the upstream deletion removes the promoter region ablating CT135 transcription. For the remaining 7 (of 10) ocular isolates, CT135 is interrupted by a variety of indels and nonsense mutations.

Thirty genital isolates were analyzed, a set which had representation from all genital non-LGV serovars (D, E, F, G, H, I, J, K and genital serovar B isolates). The CT134 locus was intact in all isolates examined. In contrast to what was observed with the reference stocks, of these genital isolates the CT135 locus was intact in all but two (a serovar H and serovar E). Not only was CT135 intact in the majority of the tested genital isolates, there was also a notable absence of any SNPs. We conclude that CT135 is highly stable in genital clinical isolates, and that the locus is under positive selection in vivo. Interestingly, our CT135 results are similar to our previously reported distinct ocular vs genital serovar functional differences at the cytotoxin (Carlson et al. 2004) and tryptophan synthase loci (Caldwell et al. 2003).

In summary, the deep-sequencing results presented herein support the conclusion that CT135 is subject to selective pressure, and thus, likely an important C. trachomatis virulence or host-adapted gene. The findings indisputably show that, with genital chlamydial serovars, there is positive selection to maintain the CT135 locus in vivo and a surprisingly powerful negative selection to disrupt the gene in vitro. At the present time, it is not known what the in vivo selection signals are, but a plausible explanation may be for chlamydiae to evade host innate or adaptive defenses. Why there is strong selection to inactivate this function in vitro is unclear.

SUPPLEMENTARY DATA

Supplementary data are available at FEMSPD online.

FUNDING

This work was supported by the Public Health Agency of Canada and the Intramural Research Program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

Conflict of interest. None declared.