Topological Analysis of Chlamydia trachomatis L2 Outer Membrane Protein 2

Abstract

Using monospecific polyclonal antisera to different parts of Chlamydia trachomatis L2 outer membrane protein 2 (Omp2), we show that the protein is localized at the inner surface of the outer membrane. Omp2 becomes immunoaccessible when Chlamydia elementary bodies are treated with dithiothreitol, and protease digestions indicate that Omp2 has a possible two-domain structure.

Chlamydiae are obligate intracellular pathogenic bacteria with a unique biphasic life cycle. In the extracellular environment, Chlamydia exists as the infectious, metabolically inactive elementary body (EB). Upon host cell entry, EBs convert to the intracellular replicating form, the reticulate body (RB). Both EBs and RBs have an outer membrane with a gram-negative-like structure, but no peptidoglycan layer has been detected (13). As in gram-negative bacteria, the Chlamydia outer membrane complex (COMC) can be extracted with 2% Sarkosyl. Chlamydia trachomatis OMC contains genus-specific lipopolysaccharides (LPS) and three dominant proteins, the major outer membrane protein (MOMP; 39.5 kDa) (7) and the cysteine-rich outer membrane proteins Omp2 (60/62 kDa) and Omp3 (12 kDa) (15). In the EB, MOMP is present as a trimer in association with LPS (4), in agreement with its porin function (2). Specific variable domains of MOMP and the LPS are the only known surface-exposed components of C. trachomatis EBs (5, 19).

The transition of RB to EB coincides with the synthesis and incorporation of Omp2 and Omp3 in the Chlamydia outer membrane (16, 25). Extended disulfide cross-linking of these proteins is thought to contribute to cell wall rigidity and osmotic stability of the EB. Omp2 and Omp3 are encoded by a bicistronic operon (1, 21). Chlamydia psittaci Omp3 is shown to be a lipoprotein (11), indicating that it is situated at the inner surface of the outer membrane. The topology and specific localization of Omp2 have thus been questioned (12, 22).

Omp2 is a major immunogen in chlamydial infections (24), but the use of immunoelectron microscopy has previously failed to detect Omp2 on the surface of chlamydiae (8, 28). Studies by Everett and Hatch (12) indicate that Omp2 is situated in the periplasmic space and is therefore not a true membrane constituent. These studies have shown that C. psittaci 6BC Omp2 is water soluble when COMC is reduced by dithiothreitol (DTT) and is not detected by hydrophobic affinity labelling. Both observations are contrary to results for both MOMP and Omp3. Protease digestions raise questions. Everett and Hatch find that C. psittaci EB Omp2 is not cleaved by trypsin, whereas C. trachomatis L2 Omp2 is cleaved by trypsin treatment of EBs (12). Furthermore, Ting et al. (26) found that a small part of Omp2 was cleaved off by trypsin treatment of EBs from C. psittaci guinea pig inclusion conjunctivitis strain and that the trypsin treatment reduced the infectivity of C. psittaci guinea pig inclusion conjunctivitis. Additionally, these researchers reported that Omp2 prepared from COMC binds preferentially to formaldehyde-fixed HeLa cells (26). Consequently, Ting et al. (26) concluded that Omp2 was partly exposed on the EB surface.

The posttranslational modification of the Omp2 protein is unique. A traditional leader peptide (amino acids [aa] 1 to 22) is cleaved off upon maturation of the protein (1), but approximately 50% of C. trachomatis L2 Omp2 is cleaved additionally at a secondary cleavage site (aa 41/aa 42) where there is no sign of a traditional signal peptidase recognition site (27). This two-step posttranslational modification results in a 60/62-kDa doublet of the protein (1).

Analysis of the C. trachomatis L2 Omp2 amino acid sequence is shown in Fig. 1. The N-terminal part of the mature protein (aa 23/42-84) is hydrophilic, with a high surface probability (Fig. 1). This region of the protein is highly variable between species of Chlamydia (17), but the hydrophilicity and high surface probability are conserved (data not shown). The region contains many trypsin cleavage sites (Fig. 1). The remaining part of the protein (aa 85 to 547) is highly conserved within the Chlamydia genus and contains all the structurally important cysteine residues (Fig. 1).

FIG. 1.

Analysis of the C. trachomatis L2 Omp2 amino acid sequence with the Wisconsin Genetics Computer Group sequence analysis software package (9). Kyte and Doolittle (KD) hydrophilicity calculations are based on observations concerning individual amino acid side chains (20), whereas surface probability predictions are based on structural data (10, 18). Trypsin sites indicate potential protease recognition sites. Antibodies were generated against the regions of the protein marked at the bottom.

To study the topology of Omp2 in C. trachomatis L2, we produced polyclonal monospecific antibodies to different parts of Omp2. Histidine-tagged fusion proteins of Omp2 have previously been generated and affinity purified (24). This includes aa 23 to 84 (Ctr-Omp2_aa23–84), Ctr-Omp2_aa23–182, Ctr-Omp2_aa167–434, and Ctr-Omp2_aa420–547 (Fig. 1). Monospecific polyclonal antibodies were raised against these fusion proteins (Ctr-Omp2_aa23–182, Ctr-Omp2_aa167–434, and Ctr-Omp2_aa420–547) by immunizing New Zealand White rabbits as previously described (21). The fusion protein Ctr-Omp2_aa23–84 was used to affinity purify antibodies against this fragment from the serum produced by immunization with Ctr-Omp2_aa23–182 as described previously (14). The specificity of the purified polyclonal antibody was tested by enzyme-linked immunosorbent assay (24). The serum reacted strongly with peptides covering the variable region of C. trachomatis L2 Omp2 (aa 23 to 84, aa 35 to 52, aa 46 to 64, and aa 59 to 73) (data not shown).

To evaluate the resistance of Omp2 to proteases, freshly purified C. trachomatis L2 EBs were digested with soluble trypsin (Sigma, St. Louis, Mo.). Reactions were terminated by adding sodium dodecyl sulfate (SDS) sample buffer followed by boiling the preparation (100°C, 5 min). The protein samples were separated by SDS–polyacrylamide gel electrophoresis (PAGE; 10% polyacrylamide) and electrotransferred to polyvinylidene difluoride membranes (Millipore, Bedford, Mass.). Immunodetection was done as described previously (3, 6). Similar digestion profiles were obtained with sera generated against Ctr-Omp2_aa23–182, Ctr-Omp2_aa167–434, and Ctr-Omp2_aa420–547 (Fig. 2A through C). The Omp2 double band was detected at 62/60 kDa (double arrow, Fig. 2A). With increasing amounts of trypsin, a 55-kDa band became pronounced (single arrow, Fig. 2A). As indicated by the trypsin map (Fig. 1), this could result from either N-terminal or C-terminal digestion. However, the digestion results in a single band (55 kDa), and it is therefore likely that the N-terminal part of Omp2 is cleaved off, since trypsin cleavage sites are present at aa 89 and 93 (Fig. 1). This possibility was verified by using the affinity-purified antibody against Ctr-Omp2_aa23–84. This antibody recognized only full-length Omp2 (Fig. 2D). The different protease susceptibility of the amino-terminal fragment versus that of the remaining part of the protein produces important structural evidence of a highly rigid, inert C-terminal domain that contains all structurally important cysteine residues (Fig. 1). Everett and Hatch (12) proposed that the trypsin sensitivity of Omp2 was due not to surface exposure of Omp2 but solely to the ability of the enzyme to penetrate the outer membrane of C. trachomatis L2. To determine whether this was the case, we probed the electrotransferred digestions with a monoclonal antibody against ribosomal protein S1 (23). Ribosomal protein S1 was totally degraded by trypsin (Fig. 2E), demonstrating that trypsin had access not only to the surface of the EBs but also to the periplasmic space and to the cytosol, even in freshly purified EBs. We repeated the study, using immobilized proteases as described previously (United States Biochemical Corporation, Cleveland, Ohio). Protease digestions were terminated after 5 min of reaction time. No digestion of Omp2 was seen with high protease concentrations (1 μg/μl) (Fig. 2F, lane 2). A positive control indicated significant digestion of the surface-exposed Mycoplasma hominis PG21 membrane protein (70 kDa) under similar conditions (Fig. 2F, lane 4) (21).

FIG. 2.

Immunoblotting of SDS-PAGE-separated trypsin digests. Lanes labelled EB are nontreated EBs. Lanes labelled with trypsin are EBs digested with 0.025 (+), 0.25 (++), or 2.5 (+++) μg of trypsin per ml for 30 min. Membranes were probed with anti-Ctr-Omp2_aa23–182 (A), anti-Ctr-Omp2_aa167–434 (B), anti-Ctr-Omp2_aa420–547 (C), or anti-Ctr-Omp2_aa23–84 (D), and a monoclonal antibody against Chlamydia ribosomal protein S1 (E). The double arrow in panel A indicates the localization of the Omp2 double band 60/62 kDa, and the single arrow indicates the major product of digestion (55 kDa). (F) Immunodetection upon digestion with immobilized trypsin. Lanes 1 and 2, immunodetection of EBs with anti-Ctr-Omp2_aa420–547; lanes 3 and 4, digestion of the surface of M. hominis with a monoclonal antibody against a 70-kDa surface-exposed membrane protein (21). Lanes 1 and 3 were not treated. Lanes 2 and 4 were treated with trypsin.

Immunoelectron microscopy was used to investigate the localization of Omp2 epitopes. C. trachomatis L2 OMC was adsorbed to carbon-coated grids and reacted with each of the four antibodies as previously described (5). A secondary goat anti-rabbit antibody, conjugated with 10-nm colloidal gold particles, was used (Biocell, Cardiff, United Kingdom). All antibodies reacted in a similar manner. Membranes were not labelled with the antibodies (Fig. 3A). To test the immunodetection of Omp2 upon reduction, membranes were treated with 5 or 50 mM DTT prior to immunodetection. This agent reduces the extensive disulfide bridging in COMC, thereby enabling the antibodies to react with Omp2 (Fig. 3B and C).

FIG. 3.

Immunoelectron microscopy with the polyclonal monospecific anti-Ctr-Omp2_aa167–434 with a colloidal gold-conjugated goat anti-rabbit secondary antibody. (A) COMC. (B) COMC treated with 5 mM DTT. (C) COMC treated with 50 mM DTT. (D) EBs. (E) EBs treated with 5 mM DTT. (F) EBs treated with 50 mM DTT. Bar, 0.1 μm.

The four monospecific polyclonal antisera prepared against Omp2 should be able to detect surface-exposed epitopes of Omp2 on whole Chlamydia EBs. None of the sera reacted with the untreated EB surface by immunoelectron microscopy (Fig. 3D). Treatment of EBs with 5 mM DTT did not change the lack of immunodetection (Fig. 3E). However, when the EBs were treated with 50 mM DTT, a strong reaction was seen (Fig. 3F), similar to that observed with COMC (Fig. 3C).

In order to elucidate the chemical changing in the Chlamydia outer membrane, we separated the proteins from C. trachomatis OMC and EBs upon treatment with DTT by SDS-PAGE. Our results are in agreement with observations previously reported for C. psittaci 6BC (12). Minor extraction (∼1%) of Omp2 was seen when EBs were treated with DTT (50 mM), whereas a larger proportion of Omp2 (∼10%) was liberated from COMC upon extraction with DTT (50 mM) (data not shown). MOMP was not extracted from the membranes or from whole EB particles upon treatment with DTT. These results demonstrate that a reduction of the outer membrane facilitates the immunodetection of Omp2, indicating that Omp2 is leaking through the pores created by MOMP upon reduction or that the MOMP porin renders Omp2 accessible to antibodies.

In conclusion, protease digestions and immunoelectron microscopy indicate that C. trachomatis L2 Omp2 is localized at the inner surface of OMC with a small protease-susceptible amino-terminal region and a large, rigid C-terminal domain. The structural and functional roles of the highly variable amino-terminal region remain to be elucidated.