TEXT

Top Abstract Text References

Chlamydia trachomatis is an obligate intracellular pathogen that replicates inside human epithelial cells in a membrane-bound vacuole termed an inclusion (4, 16). The C. trachomatis developmental cycle in the infected cells is characterized by the fusion of various small homotypic inclusions containing reticulate bodies to one large single inclusion which then contains the replicating bacteria (5, 7). Key features for the survival of C. trachomatis in the cell and the pathogenesis of the C. trachomatis infection are the bacterial ability to prevent a fusion between this single inclusion and the cellular lysosomal compartments and the bacterial ability to acquire nutrients from the host cell cytosol (4, 16). Chlamydial proteins that are translocated into the inclusion membrane, so-called Inc proteins, might be involved in both processes (11). To date, at least 11 distinct Inc proteins in the inclusion membrane of C. trachomatis-infected cells have been identified (2, 3, 13). Although the various Inc proteins of C. trachomatis share a very low primary sequence homology, a typical bilobed hydrophobic domain of approximately 60 amino acids is present in all of them (2). Recently, C. trachomatis strains causing unusual nonfusogenic inclusions were identified (15). The growth cycle of this phenotype in epithelial cells was characterized by the occurrence of multiple small nonfusogenic inclusions in epithelial cells. This phenotype was present in only 176 (1.5%) of 11,440 clinical isolates (15). Immunofluorescence microscopy of cells infected with C. trachomatis of the nonfusogenic phenotype revealed that IncA was undetectable in the inclusion membrane. In addition, IncA could not be detected by immunoblot analysis in whole-cell lysates of cells infected with nonfusogenic C. trachomatis (15). These data strongly suggested that one of the functions of IncA is to facilitate homotypic fusion of inclusions. Sequence analysis of incA of two strains exhibiting the nonfusogenic phenotype and lacking detectable IncA protein identified two nonsynonomous mutations in the gene compared to the DNA sequence of the prototypic C. trachomatis strain (serovar D/UW-3/Cx) (14). One mutation leads to the replacement of isoleucine at codon 47 with threonine (I47T), modifying the characteristic hydrophobic domain of the protein. The other substitution (E116K) is localized outside this domain (15). These results suggested that the incA I47T mutations might be associated with the nonfusogenic phenotype.

The present study aimed to investigate the correlation between the incA I47T mutation and the nonfusogenic phenotype in C. trachomatis. In order to assess the occurrence of the I47T polymorphism as well as possible additional polymorphisms in incA, the incA genes of 25 C. trachomatis isolates of different serovars were sequenced (Table 1). Selection of the isolates was at random, with the only condition being that all major serovars should be present. Isolates were propagated and DNA was extracted from infected cells as described previously (8). incA was then amplified by thermocycling using the primer pair 5'-AGCCATAGGATCTGGTTTCAGCGA-3' and 5'-GCGCGGATCCTAGGAGCTTTTTGTAGAGGGTGA-3 (15). From each isolate, three independently generated PCR products, one of which was cloned into the PCR2.1 vector (Invitrogen, Groningen, The Netherlands), were sequenced using the above-mentioned primers, a second primer set (5'-CGCTCCACAAATCACTATTGTTC-3' and 5'-TGAGGCAACAGAGCCTTTAAGA-3'), and -21M13 and M13Reverse primers (Perkin-Elmer, Nieuwekerk aan de IJssel, The Netherlands). Sequence analysis was performed on a automatic sequencer (Applied Biosystems Incorporated, Foster City, Calif.) as described previously (10). The accession numbers of incA sequences are listed in Table 1.

Database similarity searches (1) revealed that 19 out of 25 incA genes contained one or more point mutations compared to the DNA sequence of the prototypic incA-positive and fusogenic strain (D/UW-3/Cx) (14). No deletions were found, and none of the mutations introduced a stop codon.

On the basis of amino acid substitution patterns in the hydrophobic domain of IncA, there were four major different incA sequence types among the isolates. Nine (36%) of the strains studied contained incA genes encoding putative IncA proteins which were identical to the prototype sequence, indicated as sequence type a (Table 1). Four (16%) strains had incA sequences in which only substitutions outside the hydrophobic domain were identified (sequence types i to l) (Table 1). incA sequences from five (20%) other strains (sequence types d to h) were characterized by the presence of a mutation at codon 62 or at codon 75. Hydropathy plot analysis, using a window size of seven (6), predicted that none of these substitutions would result in a significant change in the secondary structure of the hydrophobic domain of IncA (not shown). In addition, it has been reported that such strains express IncA in the inclusion membrane and are fusogenic (3), indicating that the substitutions do not interfere with IncA expression and fusogenicity.

Seven (28%) strains with the mutation I47T (sequence types b and c) were identified. The frequency of this mutation was significantly higher than that described for the nonfusogenic phenotype (15). Therefore, it is questionable whether I47T predicts nonfusogenicity of strains carrying this mutation; thus, we studied IncA expression and fusogenicity of nine Chlamydia isolates representing the four incA sequence types (indicated in Table 1) using immunofluorescence microscopy. HeLa 229 cells (ATCC CCL2.1) cultured on coverslips were infected with each of the isolates a multiplicity of infection of 1 to 2. Infections were carried out as described previously (8), and cells were ethanol fixed 20 to 30 h postinfection and indirectly immunostained with an antiserum directed against IncA (3) essentially as described previously (12). Confocal immunofluorescence microscopy of infected cells immunostained with anti-IncA demonstrated clear bright fluorescent staining of one large single inclusion in each infected cell, irrespective of the sequence type of the strain used for infection (Fig. 1A to E). In addition, staining of reticulate bodies circumscribing the luminal surface of the single inclusions was also observed (Fig. 1A to E). In contrast, mock-infected cells showed no reactivity (Fig. 1F). The typical rim-like staining pattern of the inclusion membrane strongly indicates that the strains expressed IncA which was localized in the inclusion membrane. In addition, only one large inclusion per infected cell was present, indicating that the neither the I47T mutation nor one of the other incA polymorphisms interfered with the characteristically homotypic fusion of the inclusions. In conclusion, the alteration in the hydrophobic domain of IncA due to the mutation I47T is not associated with a deficiency in IncA expression and does not interfere with the translocation of IncA to the inclusion membrane and the fusogenicity of the inclusion. Therefore, other mutations in incA or its putative regulatory regions, alone or in conjunction with mutations and lesions in other genes, should be considered as alternative explanations for nonfusogenicity of inclusions.

View larger version (112K): [in this window] [in a new window]   FIG. 1.   Fluorescence microscopic analyses of HeLa cells infected with serovar D', D, F, E, and L2 (panels A to E, respectively) and noninfected cells (F) fixed for microscopy 20 to 30 h postinfection. Immunostaining with anti-IncA indicates incA expression and the occurrence of fusogenicity in all infected cells, despite the presence of the mutations I47T and E116K (A and B), the mutation I47T (C and D), or any other polymorphisms in incA (E). The bar in each panel indicates 10 µm.

Nucleotide sequence accession numbers. The nucleotide sequences of incA are deposited in the GenBank database under the accession numbers indicated in Table 1.