Materials and methods

Clinical specimens

One hundred consecutive C trachomatis tissue culture-positive samples from routine diagnostics from 5 October to December 2006 were genotyped using ompA sequencing and MLST. These samples also included nvCT-positive samples, which were found using Cobas Amplicor (negative results) and mutant-specific real-time PCR.15 In addition, the results of the ompA sequencing of 237 C trachomatis-positive clinical specimens, collected during a one-year period in 1999–2000,10 were used for comparison, and a subset of 53 genotype E samples were also examined by MLST.

The age of the individuals examined was recorded from referral sheets to the laboratory.

ompA typing

Sequence determination was according to our previously described method.10

Multilocus sequence typing

The system comprises five target regions (figure 1) and was performed as previously described13 with a minor modification. This consisted of the pbpB gene being amplified as two separate fragments, using the primer pairs pbpB1F/pbpB823R (5′–TATATGAAAAGAAAACGACGCACC–3′ and 5′–CAGCATAGATCGCTTGCCTAT–3′) and pbpB1455F/pbpB2366R (5′–GGTCTCGTTTTTGATGTTCTATTC–3′ and 5′–TGGTCAGAAAGATGCTGCACA–3′), respectively. The pbpB gene was amplified using the same cycling parameters as for the other regions.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Schematic overview of the five multilocus sequence typing (MLST) regions. Each region is named after the target gene. The hctB has repetitive elements (grey) and the pbpB was analysed in two reactions (pbpB1F/pbpB823R and pbpB1455F/pbpB2366R).

Discriminatory power

The discriminatory power (D) was determined as previously described by Hunter and Gaston,16 for the ompA sequencing, the MLST system and the combination of both methods. D is the index of the discriminatory power expressed in a percentage. As D must be based on epidemiologically unrelated isolates, only 97 isolates from 2006 were used in this calculation, because three isolates were from sexual contacts.

Statistics

To test the difference of proportion with genotype E in 1999–2000 and 2006 we used the Pearson χ² test. The significance level was set at 5% (two-sided). Statistical analyses were performed using SPSS software V.15.0.

Results

ompA genotyping

Among the 100 consecutive Chlamydia tissue culture-positive samples in the 2006 material, there were 69 (69.0%) genotype E isolates, with an identical ompA sequence. Genotype E was thus significantly more prevalent (p<0.001) than in 1999–2000 (47.3%, figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

ompA genotype distribution of 237 urogenital Chlamydia trachomatis isolates from a one-year study in Örebro, Sweden, in 1999–200010 compared with the distribution of genotypes in consecutive urogenital isolates (n=100) in the same county from October to December 2006. All 41 new variant (nvCT) cases were excluded before comparison, but are denoted in the figure as genotype E.

Among the genotype E isolates cultured in 2006, all 41 of the nvCT displayed identical ompA sequence compared with the reference strain E/Bour (GenBank no X52557). However, three of these nvCT-positive samples were also Cobas Amplicor positive, indicating multiple infections also including wild-type C trachomatis (wtCT). Excluding the 41 nvCT isolates, the distribution of ompA genotypes (n=59) was as follows: genotype B (3.4%), D (11.9%), E (47.5%), F (13.6%), G (10.2%), H (1.7%), Ia (1.7%), J (3.4%) and K (6.8%), which is similar to the genotyping results from 1999–2000 (see figure 2). The mean age in 2006 was lower (21.1 years, range 15.8–35.4, median 20.1 years) for the 41 patients with nvCT compared with the 59 patients with wtCT (24.3 years, range 15.4–56.3, median 21.4 years). In the one-year period in 1999–2000 the mean age was 27.2 years (range 13.5–63.5).

Multilocus sequence typing

The distribution of the MLST profiles from 2006 is shown in tables 1 and 2. Briefly, complete MLST profiles were obtained from 26 out of 28 wtCT genotype E isolates, and among these 14 (58%) unique sequence types were identified. In contrast, all nvCT isolates displayed an identical and unique MLST profile: 21 (target region hctB), 19 (CT058), 1 (CT144), 2 (CT172) and 1 (pbpB). However, one isolate had a too short sequence product and an inconclusive result in pbpB, but was typeable to be 1, 2 or 38.

The MLST profiles of the 31 isolates that displayed other ompA genotypes than genotype E are shown in table 2. In short, among the genotype D isolates (n=7), the MLST method and the ompA genotyping both identified three different genetic variants. For the genotype F isolates (n=8) the ompA sequences were identical, but the MLST method showed four different sequence types and one additional isolate was not typeable in hctB. All genotype G isolates (n=6) showed different sequence types, while two ompA sequences were identified. Two genotype J isolates were identified, displaying identical ompA sequences but different sequence types. All genotype K isolates (n=4) displayed an identical ompA sequence, but three different sequence types. It is notable that in four of the ompA genotype D and three of the ompA genotype F isolates the MLST profile was identical (5, 19, 7, 1, and 4).

Out of 53 genotype E isolates from the 1999–2000 material, with an identical ompA sequence, we only managed to identify four complete MLST profiles (data not shown). However, we were able to type the target region hctB in 16 of the 53 isolates, but none displayed the allele variant number 21, which was represented in all nvCT in 2006 (see table 1).

Discriminatory power

The D was calculated using 97 unrelated isolates. D for ompA sequencing in this material was 49.5% and for the MLST method it was 83.5%. With a combination of both methods, D=83.7%.

Discussion

Our genotyping study of 100 consecutive C trachomatis isolates shows that the outbreak of nvCT is of clonal nature and that MLST provides a substantially higher discriminatory power than ompA sequencing in genotype E isolates, resulting in more than a threefold higher number of variants compared with ompA sequencing when nvCT cases are excluded.

We have previously shown that, in principle, our MLST system provides high resolution for genotyping of selected C trachomatis cases.13 In the present study MLST was further challenged by typing 100 consecutive samples and compared with ompA genotyping. This investigation confirmed that MLST is also superior to ompA sequencing in discrimination capacity for unselected specimen collections.

For genotypes E, F, G and K, comprising 87% of all C trachomatis cases, the discriminatory capacity of MLST was substantially higher than when utilising ompA sequencing. For the seven genotype D isolates both typing systems resulted in three genetic variants, which is in contrast to a previous study in which genotype D samples showed five different ompA sequences, but 11 divergent MLST sequence types.17 However, this was in a study of men who have sex with men and the high number of partners they had may explain the different outcome. For the other genotypes (B, H, I and J) no difference between the systems could be seen, but the number of cases was also too low to enable a valid comparison.

It is also noteworthy that using the five MLST targets and ompA sequencing in combination was beneficial, as four of the ompA genotype D and three of the ompA genotype F isolates displayed a single identical sequence type. This would also be useful for comparisons with many previous and presumably future studies using ompA genotyping. Nevertheless, overall ompA sequencing only slightly increased the number of genetic variants identified.

We found that all nvCT isolates were of genotype E, with an identical ompA sequence, and displayed an identical and unique MLST profile, which strongly indicates a clonal spread of the new variant. In contrast, among 26 of 28 wtCT genotype E isolates, 14 (58%) unique MLST types were obtained. The mean age of patients infected with nvCT was lower than that of patients harbouring wild-type strains. This was not surprising, considering that young persons in Sweden are more sexually active and have more partners, and thus nvCT is more rapidly spread in younger age groups. The nvCT has been detected in high proportions all over Sweden,18–20 but surprisingly few cases have been found in neighbouring countries or other countries.2 21–23 Furthermore, in a recent study the MLST analysis of 48 arbitrarily selected nvCT isolates from four countries had one single sequence type, which indicates a clonal spread.2 The present investigation of a local study population confirmed these findings. It also indicates that although the MLST system has high resolution, there is no obvious risk that the system is too variable and that it would give misleading results when used for molecular epidemiology of sexual networks.

At present it is not known when or where this nvCT appeared and started to be transmitted in Sweden. However, in our study in 1999–2000,10 one suspected mutant (of 237 positive samples derived from a total of 2195 patients), which was positive in conventional ompA gene PCR but repeatedly negative in Cobas Amplicor, was identified. This specimen was also determined as genotype E, that is, as the present nvCT, but 47% of the samples in that previous study were genotype E. At that time, this Cobas Amplicor-negative isolate was considered a plasmid-free C trachomatis strain.24

Efforts have been made to apply the present MLST method to C trachomatis specimens from the previous study in 1999–2000.10 However, complete MLST profiles were obtained in only a limited number of cases. This is explained by the storage of the samples at –20°C and previous repeated freeze thawing, resulting in DNA degradation.18 25 Nevertheless, we managed to show that in one target region (hctB) the allele variant for genotype E in 1999–2000 differed when compared with that in 2006. It is not surprising that a fluctuation of specific C trachomatis strains predominates in a population, but as conventional ompA typing has lower differentiating capacity this is not easily seen.

There are limitations in our study. First, we apparently only succeeded in analysing a few samples from 1999–2000 in the MLST system. This reduces the possibilities for comparing variation of sequence types over time. In addition, the number of isolates analysed in 2006 (n=100) also means that statistical changes are difficult to obtain. Second, clinical and epidemiological data for the Chlamydia-infected individuals in the study were mainly lacking. Analysis of the spread of specific C trachomatis strains in sexual networks was thus not possible and was also beyond the aims of the study.

Although for two decades genotyping of C trachomatis has been restricted almost entirely to ompA sequencing, new methods with other targets have recently been reported. One system combines ompA sequencing with analysis of tandem repeats.14 An MLST system intended for tracking evolutionary changes has also been published.26 The advent of new multitarget typing systems provides more information and highlights the need for standardised nomenclature systems.

In conclusion, the identical MLST profile of all nvCT isolates strongly indicates a clonal spread in the population, as shown by this study, in which a significant (p<0.001) increase in genotype E from 47% in 1999–2000 to 69% in 2006 was detected. Excluding the nvCT isolates, the distribution of ompA genotypes is similar to the genotyping results from 1999–2000. The overall discrimination of C trachomatis isolates was high using the present MLST system. Accordingly, the MLST system enables improved molecular epidemiological characterisation of sexually transmitted Chlamydia infections.