A study of C. trachomatis concordance in heterosexual partnerships (dyads) was conducted in Indianapolis during April 2000–October 2003; participants were sexually active heterosexual men and women 15–25 years of age who visited an urban STI clinic (12). Written informed consent was obtained, and the study was approved by the Indiana University–Purdue University Institutional Review Board. Eligibility was defined as self-reported sexual activity between the partners during the previous 30 days. A total of 210 heterosexual dyads were established by research disease intervention specialists and enrolled. C. trachomatis infection was identified by Amplicor CT/NG (Roche Diagnostics, Indianapolis, IN, USA) nucleic acid amplification test and cell culture, as previously described (12). Of the 210 dyads, 130 contained >1 C. trachomatis–infected partner; for 45 dyads, both partners were infected and had identical ompA genotypes.

For the MLST study, we used remainder samples from 56 members of 28 dyads who were concordant for C. trachomatis infection and ompA genotype. These samples were provided to investigators at Children’s Hospital Oakland Research Institute (CHORI) in a de-identified and blinded fashion. Thus, CHORI research was considered not to involve human subjects, and informed consent was not required.

We used 56 samples (from cervix in women and urethra in men) from 28 dyads in which persons within each dyad were concordant for C. trachomatis infection and ompA genotype. Additionally, we used for analysis MLST data for 20 C. trachomatis reference strains (A/Sa1, A/HAR13, B/TW5/OT, Ba/Apache2, C/TW3/OT, D/UW3/Cx, Da/TW448, E/Bour, F/ICCal3, G/UW57/Cx, H/UW4/Cx, I/UW12/Ur, Ia/UW202, J/UW36/Cx, Ja/UW92, K/UW36/Cx, L₁/440, L₂/434, L₂a/UW396, L₃/404) and 172 clinical samples in the MLST database (http://www.mlst.net).

ompA genotyping of the samples had been previously performed at Indiana University (13,14) as part of the earlier C. trachomatis concordance study. Cultured and noncultured clinical samples were sent to CHORI for analysis (11). DNA was extracted, and MLST for 7 housekeeping genes was performed by using primers as described (11; http://www.mlst.net; Technical Appendix Table 1). A consensus sequence was created from forward and reverse sequences, and the genes were concatenated and queried against all 202 MLST sequences in the database (15). Sequence output was used to identify each unique allelic profile to assign an ST, and all STs were deposited in the C. trachomatis database (http://chlamydia.mlst.net). The concatenated sequences of the 7 MLST loci and the allelic profiles for each sample were used to identify sample relatedness.

ompA genotypes were defined on the basis of homology with reference strains of C. trachomatis. If >1 SNP was identified when sequences were compared to those of the closest hit reference strain, a number was used to denote the presence of the SNP(s) (e.g., Ia4) (15).

Strain clustering and SNP analyses were performed as described (11). Briefly, clusters of related and singleton STs as well as evolutionary patterns among the isolates and for the entire dataset were determined by using eBURST (http://eburst.mlst.net). Neighbor-joining and minimum evolution methods in MEGA4 (http://www.megasoftware.net) were used to construct the trees along with multiple substitution models, including p-distance and Jukes-Cantor; all methods gave similar results. To test support for each node in the tree, we performed 1,000 bootstrap replicates.

All SNPs were identified for each ST by using the PROC FREQ tool in SAS software (SAS Institute, Cary, NC, USA). The probability of association of a SNP with an ST was determined by using a classification index (16). Variance across the dataset was determined using the Levene test (17). A p value <0.05 was considered significant.

The Table shows the distribution of MLST STs and ompA genotypes for each dyad with SNP location(s), if present, for each of the 7 MLST housekeeping genes. We noted that in some cases, the DNA extracted directly from the patient sample was sufficient for MLST, whereas in other cases, the cultured sample was required because the patient sample did not yield sufficient DNA for MLST. We found no differences in MLST results when we compared DNA directly extracted from the patient sample with DNA extracted from the culture of the same sample.

For each of the 28 dyads, both partners had the same MLST ST. A total of 9 STs were found in the 56 samples from the 28 dyads: ST15, ST19, ST23, ST34, ST39, ST45, ST46, ST47, and ST55. Twenty-one dyads harbored MLST STs that matched STs of samples obtained from geographically diverse areas currently included in the MLST database, whereas 7 dyads harbored newly identified STs not present in the database (Technical Appendix Table 2). Among these 7 dyads, 4 MLST STs were unique to Indianapolis: ST45 (dyad 22), ST46 (dyads 23–26), ST47 (dyad 27), and ST55 (dyad 28); these results reflect the SNPs in various alleles (Table).

Eleven ompA genotypes were represented in the study and, by study design, were identical within dyads. The ompA genotypes included E, F, H, Ia, J, and K sequences that were identical to those of reference strains and 5 variant ompA genotypes of D1, D2, E6, F4, and Ia4 that had SNPs compared with the reference strains.

By combining MLST ST and ompA genotype data, we identified 13 unique strains (9 by MLST and 4 by ompA genotype) among the samples from our study group. Among these 13 strains, 8 were unique to Indianapolis and were found in 12 dyads (dyads 1, 6, 9, 12, 18, 22, 23, 24, 25, 26, 27, and 28) (Table; Technical Appendix Table 2). Moreover, of the 13 strains identified among the dyad samples, 9 (69%) contained gene sequences that suggested recombination within the genome, meaning that the ompA genotype was different from the ompA genotype that should be associated with the MLST ST if the genome were just 1 strain. For example, for strains from dyads 10 and 12, the ompA was D2, but the sequences of the 7 housekeeping genes (MLST ST34) matched the 7 housekeeping genes of strain F/ICCal3 from the MLST database. Putative recombinants (boldface in Table) represented a rate of 32% (9 of 28 samples).

MLST STs and ompA genotypes for each sample in the MLST dataset are shown in Technical Appendix Table 2. We found substantial variability of ompA genotypes associated with MLST STs, as shown previously (11). For the STs for the 56 Indianapolis samples, ST15 was associated with ompA genotypes J and K; ST34 was associated with ompA genotypes D2, E, and F; and ST19 was associated with ompA genotypes D, G, H, I, and J. online Technical Appendix Table 3 shows the characteristics of the alleles for each MLST locus based on the inclusion of the Indianapolis dataset in the MLST database.

The association of disease phenotype with 3 clonal complexes (CCs) was identified by eBURST (Figure 1), similar to those we reported previously: C. trachomatis strains that cause trachoma A, B, Ba, and C (CC-A); noninvasive STIs with low population prevalence (CC-B); and noninvasive, globally prevalent D/Da, E, and F STIs (CC-C). The Indianapolis strains were confined to noninvasive STI CCs (B and C), as expected. The strains associated with 3 of the 4 unique STs in the Indianapolis samples are seen in CC-C (Figure 1).

The minimum-evolution tree also displayed 3 disease clusters (Figure 2); each of the Indianapolis strains is denoted next to the corresponding ST. Cluster I grouped noninvasive, low-prevalence STIs (eBURST CC-B), including a subcluster of strains that cause trachoma (eBURST CC-A). Cluster II grouped only invasive LGV strains. Cluster III grouped noninvasive, prevalent D/Da, E, and F STIs (eBURST CC-C). The tree constructed based on amino acid analysis showed similar clustering (data not shown).

The 9 putative Indianapolis recombinants were localized on the MLST tree with strains of the same ompA genotype and were recombinants of strains within the same cluster. Most recombinants were in cluster III. Four dyads with ST46 (unique to Indianapolis) had ompA genotype E and homology of the 7 housekeeping genes to reference strain Da/TW448 but with SNPs in leuS. D1/90i and D1/91i (ST45) and D2/96i, D2/97i, D2/189i, and D2/190i (ST34) were recombinants with homology to F/ICCal3 with SNPs in glyA and F/ICCal3, respectively. In cluster I, J/112i and J/113i shared the same ST as K/186i and K/187i and were recombinants with K/42nl and K/49nl. H/114i and H/115i were recombinants with G/SotonG1.