Analysis of Genetic Heterogeneity in Chlamydia trachomatis Clinical Isolates of Serovars D, E, and F by Amplified Fragment Length Polymorphism

Abstract

Amplified fragment length polymorphism (AFLP) fingerprinting of clinical isolates of Chlamydia trachomatis serovars D, E, and F showed a low percentage of genetic heterogeneity, but clear differences were found. Isolates from index patients and partners had identical AFLP patterns and AFLP markers. Characterization of these AFLP markers could give more insight into the differences in virulence and clinical course of C. trachomatis infections.

It is still unknown why infections with particular Chlamydia trachomatis serovars run either a symptomatic or an asymptomatic clinical course. Several studies have shown relationships between specific serovars and clinical disease, but the results are still contradictory (6, 11, 20, 29). It is not unlikely that variation at the genomic level could give more insight into the differences in virulence, tissue tropism, disease pathogenesis, and epidemiology. Furthermore, it could lead to the identification of strains within one serovar with different pathogenic features that could explain the differences in disease manifestations found in the literature for isolates of the same serovar. The serotyping (18, 21) and genotyping (7, 14, 18, 19) techniques for C. trachomatis are based on variations in one gene or protein: the major outer membrane protein (MOMP) or its gene, omp1. However, Chlamydia species and C. trachomatis isolates have also been differentiated at the genomic level by genomic restriction fragment length polymorphism (RFLP) analysis (23, 24), random amplification of polymorphic DNA (26), arbitrary primer PCR (13), genomic hybridization (5), and rRNA spacer analysis (15). Recently, a new fingerprinting technique has been developed: amplified fragment length polymorphism (AFLP) analysis (30). This PCR-based fingerprinting technique has shown to be a reproducible and powerful tool for taxonomic, diagnostic, and epidemiological applications (9, 10, 25, 30). The specific advantages of AFLP analysis over the conventionally used techniques are its use of small amounts of DNA (10 to 50 ng) and its reproducibility. This study aimed to optimize AFLP for investigation of genomic variation in C. trachomatis. Variation between and within isolates of the most prevalent urogenital C. trachomatis serovars, serovars D, E, and F, of the trachoma biovar was analyzed.

C. trachomatis isolates obtained from 29 heterosexual patients and their C. trachomatis-infected partners were included in this study (13). C. trachomatis serovars were determined by a PCR-based RFLP analysis of the omp1 gene and by serotyping (14, 18): 5 pairs were infected with serovar D, 13 pairs were infected with serovar E, and 11 pairs were infected with serovar F (the pairs are numbered from 1 to n for each serovar, the index patient is indicated by the letter A, and the partner is indicated by the letter B).

Cell culture was performed by routine procedures (18, 21). C. trachomatis elementary body (EB) DNA was isolated by treating a sonicated cell culture suspension with DNase I, proteinase K, and RNase A, followed by EB DNA isolation with the High Pure PCR Template Preparation kit (Boehringer, Mannheim, Germany). Each specimen was subjected to amplification by a 206-bp β-globin PCR to check for the presence of human DNA (12). C. trachomatis plasmid-specific primers were used to detect C. trachomatis DNA (17). Mycoplasma DNA was detected by PCR as described previously (22) to ensure that there was no contaminating mycoplasmal DNA. For DNA fingerprinting by AFLP analysis (10, 16, 30), C. trachomatis DNA was digested with EcoRI (New England Biolabs Inc., Beverly, Mass.) and MseI (New England Biolabs Inc.). AFLP images were analyzed with GelCompar, version 4.0, software. The similarity was expressed by the Pearson product moment correlation coefficient (r), and groupings were obtained by the unweighted pair-group matrix analysis (UPGMA) clustering algorithm as described previously (16).

The isolated C. trachomatis DNA was free of human and mycoplasma DNA, as assessed by PCR, ensuring C. trachomatis-specific fingerprinting patterns in the AFLP analysis. Different primer combinations with 0 or 1 selection base were tested (Eco0 and Mse-C, Eco-A and Mse-0, Eco0 and Mse-0) by using the five serovar D pairs. The most discriminatory primer combination for AFLP analysis appeared to be the pair Eco0–Mse-C. Figure 1 shows representative AFLP patterns for 15 strains of serovars D, E, and F (five strains of each serovar). Clustering was found at r equal to 91.0% ± 2.3% for serovars D, E, and F. Both unique AFLP markers and markers which were absent from specific strains were found. In addition, different AFLP markers were observed between 425 and 525 bp. However, while three different marker patterns were found for serovars E and D (Table 1, patterns A, B and C), only pattern B was found for serovar F. Of the most aberrant strains (from index patients), the corresponding strains from the partners have also been analyzed by AFLP analysis, as shown in Fig. 2 (eight different AFLP markers are indicated; r = 89.0% ± 3.6%). Interestingly, these specific AFLP markers are identical for both index patients and their partners. To confirm the existence of different strains within C. trachomatis serovars D, E, and F identifiable by the presence or absence of AFLP markers, five strains were further analyzed by AFLP analysis with Eco-A–Mse0. As shown in Fig. 3 (F1, D1, D3, E11, and E12, strains from both the index patient and the partner; r = 87.8% ± 2.1%), this independent AFLP analysis confirmed the existence of different strains of the same serovar, since in this AFLP analysis identical unique AFLP markers were also found for strains from both index patients and their partners.

FIG. 1.

Digitized AFLP patterns and dendrograms for the three most prevalent urogenital C. trachomatis serovars, serovars D, E, and F. The dendrograms were constructed by the UPGMA clustering method. AFLP analysis was performed with primers MseC and Eco-0. The scale represents the product-moment correlation coefficient (r; in percent).

TABLE 1.

Patterns of C. trachomatis AFLP markers between 425 and 525 bp

Pattern	AFLP marker no.a	Strain example
A	1, 2, 3, and 4	D4, E12
B	1, 2, and 4	D3, E6, F3
C	1 and 3	D1, E11

^aAFLP markers 1 to 4 are indicated in Fig. 2. 

FIG. 2.

Digitized AFLP patterns and dendrograms for C. trachomatis serovar D, E, and F isolates: pairs from index patients (A) and the corresponding partners (B). The dendrograms were constructed by the UPGMA clustering method. AFLP analysis was performed with primers Mse-C and Eco-0. The scale represents the product-moment correlation coefficient (r; in percent). Arrows indicate AFLP markers.

FIG. 3.

Digitized AFLP patterns and dendrograms for pairs of C. trachomatis serovar D, E, and F isolates. The dendrograms were constructed by the UPGMA clustering method. AFLP analysis was performed with primers Mse-0 and Eco-A. The scale represents the product-moment correlation coefficient (r; in percent). Arrows indicate AFLP markers.

This study showed that although genomic heterogeneity was small (Fig. 1 and 2), within C. trachomatis serovars D, E, and F unique AFLP markers were observed. These AFLP markers were confirmed by the identical AFLP patterns found for isolates from the corresponding partners. When a different primer combination was applied, the existence of strain variants was confirmed by showing unique AFLP markers, as shown in Fig. 3, emphasizing both the reliability and usefulness of this fingerprinting approach. However, since the differences are small but reproducible, it is advisable that strains be analyzed in the same PCR run and the same polyacrylamide gel. Phylogenetic analysis of the C. trachomatis MOMP and the omp1 gene showed no evolutionary relationships among serovars that corresponded to biological or pathological phenotypes (tissue tropism, disease manifestation, and epidemiological success) (28). Since AFLP analysis showed genomic variation between and within C. trachomatis serovars and AFLP analysis of C. pneumoniae (2) and C. psittaci (1) yielded clusters that correlated with clinical syndromes, it could be a valuable tool for answering the intriguing question of whether specific C. trachomatis strains that cause symptomatic and asymptomatic infections exist within a particular serovar. For this, the use of different enzymes and primer pairs could potentially enlarge the genomic variations found. Characterization of the AFLP markers found at the gene level by cloning and sequencing (3, 4) will be greatly facilitated by the recently published C. trachomatis genome sequence (27). Other techniques have been used to differentiate C. trachomatis strains at the genome level (8, 23, 24) and also found intraserovar heterogeneities. However, one of the important advantages of AFLP analysis over non-PCR-based genomic analysis is that instead of the 1 μg of C. trachomatis DNA needed for RFLP analysis, only 10 to 50 ng of C. trachomatis DNA is needed for AFLP analysis, significantly reducing the labor needed to propagate chlamydial strains (instead of the usual four to six six-well plates, only four shell vials were needed).

In conclusion, although the genetic heterogeneity by AFLP fingerprinting of C. trachomatis clinical urogenital isolates of serovars D, E, and F was low, clear AFLP markers were found. By this approach the unique AFLP markers found could be characterized and the observed differences in pathogenicity of C. trachomatis might be related to particular strains and/or specific genes.