INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The cell envelope structure of Chlamydia is similar to that of other gram-negative bacteria, with an outer membrane (OM) containing lipopolysaccharide, a periplasm, and an inner membrane. However, two envelope features are unique to chlamydiae: an apparent lack of or deficiency in peptidoglycan and the presence of disulfide-bond-cross-linked proteins in the OM and the periplasm (reviewed in reference 15). Nevertheless, the infectious elementary body (EB) form of chlamydiae, but not the dividing reticulate body (RB) form, is osmotically stable, and chlamydiae are sensitive to -lactams and D-cycloserine. The sensitivity of chlamydiae to peptidoglycan synthesis-inhibiting drugs in the possible absence of peptidoglycan has been termed the chlamydial anomaly by Moulder (29). The scope of the anomaly has been expanded by the recent sequencing of the genomes of several chlamydial strains (18, 35, 38), revealing the presence of what has been thought to be all of the genes required for peptidoglycan synthesis (7). Ghuysen and Goffin (11) proposed a solution to the anomaly, suggesting that chlamydiae use their peptidoglycan genes to synthesize a glycanless wall polymer whose synthesis is penicillin sensitive. Three key points of their proposal are as follows: (i) the predicted amino acid sequence of the three chlamydial penicillin-binding proteins suggests that they are capable of carrying out cross-linking transpeptidase reactions but are incapable of transglycosylating N-acetylglucosamine-N-acetylmuramic acid disaccharides into a glycan polymer; (ii) potential chlamydial N-acetylmuramoyl-L-alanine amidases cleave cross-linked peptidyl polymers from the disaccharide subunits; and (iii) the disaccharide subunits, as part of lipid II, serve strictly as carriers and therefore do not accumulate in stoichiometric amounts. Ghuysen and Goffin (11) further speculated that the glycanless polymer may be covalently linked to lipoproteins in the inner membrane or OM or to a highly disulfide-bond-cross-linked protein structure in the periplasm. A disulfide-cross-linked periplasmic structure was proposed by Everett and Hatch (10) to consist of cysteine-rich proteins (CRPs) that are encoded by a bicistronic operon and that are made only late in the chlamydial developmental cycle (1, 17, 22, 31). The CRPs consist of a 60-kDa doublet, which is the product of posttranslational processing of the omcB gene product (2), and a 12- to 15-kDa lipoprotein (9), which is a product of omcA and is speculated to be anchored to the OM by its lipid moiety (10).

The CRPs are located in the sodium lauryl sarcosine (Sarkosyl)-insoluble fraction, called the Sarkosyl chlamydial OM complex (COMC), which consists of integral OM proteins and the highly disulfide-cross-linked periplasmic CRPs (10, 15). The predominant protein in COMCs is the major OM protein (MOMP). The MOMP lacks a homolog in other bacteria but functions, like many other gram-negative OM proteins, as a porin (4, 40, 41). It is present throughout the chlamydial developmental cycle and is disulfide cross-linked in EBs but not in logarithmically dividing RBs (16, 17, 33). A third class of proteins, variously referred to as polymorphic OM proteins (POMPs) and polymorphic membrane proteins, was identified in the Sarkosyl COMC of ovine abortion strains of Chlamydia psittaci by Cevenini et al. (6) and others (12, 13, 27, 37). Genes encoding 9 and 21 potential POMPs are present in the genomes of Chlamydia trachomatis and Chlamydia pneumoniae, respectively (18, 35, 38). The similarity of predicted POMP amino acid sequences across chlamydial species is low, ranging from about 10 to 60%. Similarities among POMPs within a chlamydial strain are almost equally low, with the notable exception of the ovine enzootic abortion (OEA) strain C. psittaci S26/3 POMPs 90A and 90B, which are identical, and POMPs 91A and 91B, which are 89% similar to each other and 86 and 87% similar, respectively, to POMPs 90A and 90B (26).

The POMPs share no homology with other bacterial proteins but do share common features among themselves, including multiple GGA (I, L, and V) and FXXN repeats, large size (90 to 187 kDa), and carboxy-terminal phenylalanine residues (14, 26, 27, 38). The last property and predicted secondary structures suggest that POMPs are located in the OM, and immunoelectron and immunofluorescence microscopy studies indicate that one or more of the ovine POMPs and at least one C. pneumoniae POMP are surface exposed (19, 24, 25). However, it is not clear that all POMP genes are expressed as proteins and that all POMPs are located in the OM. For example, 5 of the 21 C. pneumoniae POMP genes contain frameshift mutations and 1 is truncated to encode a 56-kDa protein, and 1 of the predicted POMPs of OEA C. psittaci (POMP 98A; GenBank accession number U722499), C. trachomatis D, and C. pneumoniae lacks a signal sequence (14, 18, 38). The function of the POMPs is unknown. The paralogous nature of the proteins suggests that the more distantly related POMPs may have distinct functions (14), and animals that were immunized with insoluble detergent extracts of C. psittaci stains, which likely contained POMPs, were protected against challenge with infectious organisms (3, 39).

The purpose of the present study was to determine the cellular location, disulfide-cross-linked nature, and developmental stage of synthesis of POMPs in C. psittaci avian strain 6BC.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Purification of EBs and RBs. L cells were infected with C. psittaci 6BC. RBs or EBs were harvested at 15, 24, 28, and 48 h postinfection and purified by density centrifugation in a Beckman SW28 rotor for 30 min at 80,000 × g with a three-step gradient of 29, 34, and 40% Hypaque-76 (Nycomed Inc., Princeton, N.J.) in Dulbecco phosphate-buffered saline (GIBCO, Grand Island, N.Y.) containing 0.5 mM MgCl₂ and 1.0 mM CaCl₂ (PBS). The 48-h harvest was treated with the nonionic detergent Nonidet P-40 (0.25%; Sigma Chemical Co., St. Louis, Mo.) in PBS for 5 min at room temperature before purification to eliminate osmotically fragile RBs (17). RBs were collected at the 29%-34% interface; EBs were collected at the 34%-40% interface.

Trypsin digestion. To determine the surface exposure of proteins, purified EBs were incubated with trypsin (60 µg/ml; type III from bovine pancreas; Sigma) in PBS for 30 min at 37°C. Trypsin was inactivated by incubation with trypsin inhibitor (120 µg/ml; type II-O from chicken egg white; Sigma) for 5 min at room temperature. The samples were pelleted by centrifugation in a microcentrifuge at 9,000 × g for 10 min and washed once with PBS containing trypsin inhibitor.

Labeling of EBs. To label OM proteins, 25 µCi of 3-(trifluoromethyl)-3-(m-[¹²⁵I]iodophenyl)diazarine ([¹²⁵I]TID; Amersham, Arlington Heights, Ill.) was added to gradient-purified EBs harvested from 2 × 10⁸ infected cells in 200 µl of PBS. The mixture was incubated for 1 h on ice in a dimly lit room before the [¹²⁵I]TID was activated by exposure to a long-wave UV lamp for 1 h (10). Proteins in EBs and detergent-insoluble complexes were fractionated by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), and labeled proteins were detected by phosphorimaging of the dried gel or an immunoblot of the gel.

Preparation of detergent-insoluble complexes. Purified RBs and EBs were extracted at 37°C for 30 min with either 2.0% SDS or 2.0% Sarkosyl (Sigma) in PBS. The samples were subjected to sonication for 30 s, pelleted by centrifugation at 14,500 × g for 20 min at 4°C, and washed once with the detergents to produce SDS-insoluble complexes and Sarkosyl COMCs. In one experiment, EBs were incubated for 30 min at 37°C in PBS containing 20 mM dithiothreitol (DTT) and 5% -mercaptoethanol (ME) and pelleted by centrifugation at 9,000 × g for 10 min before extraction with detergents.

SDS-PAGE and immunoblotting. All samples were suspended in Laemmli solubilization buffer, with or without 5% ME and 10 mM DTT, heated to 90°C, and fractionated by SDS-PAGE on 7.5 to 15% gradient gels (21). Prestained protein standards were purchased from Bio-Rad, Hercules, Calif. (low range, catalog no. 161-0305; high range, catalog no. 161-0309; and broad range, catalog no. 161-0318).

Mass spectrometry. Sarkosyl COMCs were fractionated by SDS-PAGE, and proteins were detected by zinc sulfate staining (5). Proteins in gel slices were treated with trypsin, and peptides were extracted and analyzed by matrix-assisted laser desorption ionization (MALDI)-time of flight mass spectrometry as described by Shevchenko et al. (36).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of OM proteins in C. psittaci 6BC. The protein profiles of whole EBs and the Sarkosyl COMC fraction of EBs of C. psittaci 6BC were examined by SDS-PAGE and immunoblotting (Fig. 1). Following harvesting and purification, EBs were treated with [¹²⁵I]TID, a photoactivatable lipophilic reagent that specifically labels the portions of OM proteins that are embedded within the membrane. TID does not label periplasmic proteins or proteins in the inner membrane because it cannot enter or pass through the hydrophilic periplasm (10). Following treatment with [¹²⁵I]TID and before the preparation of Sarkosyl COMCs, half of the EB preparations were treated with trypsin to test whether proteins were surface exposed.

View larger version (72K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE and immunoblot analysis of Sarkosyl COMCs of EBs labeled with [125I]TID. EBs were harvested and purified at 48 h postinfection and treated with [125I]TID. Where indicated, EBs were treated with trypsin before the preparation of Sarkosyl COMCs. All samples were heated in the presence of 10 mM DTT and 5% ME before electrophoresis. (A) Coomassie brilliant blue-stained gel, with molecular weights (MW, in thousands) of prestained protein standards at each side. (B) Phosphorimage of the dried gel. (C) Immunoblot of a duplicate gel reacted with monospecific sheep antiserum raised against C. psittaci ovine strain POMP 90A. (D) Phosphorimage of the immunoblot. Lanes: 1, EBs; 2, EBs treated with trypsin; 3, Sarkosyl (Sark) COMCs prepared from EBs; 4, Sarkosyl COMCs prepared from trypsin-treated EBs. The positions of putative POMPs (positions 1 to 6), an unknown protein (position 7), the CRPs, and the MOMP are indicated at the right side of each panel. T1, T2, and T3 indicate tryptic fragments. The assignment of TID-labeled tryptic fragments to specific immunoreactive proteins was determined by superimposition of the phosphorimage in panel D on the immunoblot in panel C.

Disulfide bonds in POMPs. All POMP genes in C. trachomatis and C. pneumoniae and the six ovine C. psittaci POMP genes identified in DNA sequence databases are predicted to encode proteins that contain cysteine residues. To investigate whether the putative POMPs of 6BC are interpeptide cross-linked by disulfide bonds, the effects of reducing agents on the ability of POMPs in detergent-insoluble complexes to enter polyacrylamide gels were examined. The protein profiles of the Sarkosyl COMCs and the SDS-insoluble complexes electrophoresed in the presence of reducing agents were identical, consisting of the MOMP and likely degradation products of the MOMP, protein 7, the CRPs, and the putative high-molecular-weight POMPs (Fig. 2A, lanes 2 and 4). Because the strong anionic detergent SDS dissolves both the inner membrane and the OM of bacteria, the presence of these proteins in the SDS-insoluble fraction suggests that they are part of one or more insoluble supramolecular complexes. When the insoluble complexes were electrophoresed in the absence of DTT and ME, CRPs, the MOMP, and protein 7 failed to enter the gel (Fig. 2B, lanes 2 and 4), confirming that these proteins are highly interpeptide disulfide cross-linked. In contrast, diffusely Coomassie-stained bands were noted in the 98-kDa putative POMP region of the gel, in the absence of reducing agents (Fig. 2B, lanes 2 and 4). Immunoblot analysis confirmed that at least some of the diffusely stained bands cross-reacted with antibody raised against ovine POMP 90A (Fig. 2D, lanes 2 and 4). These results suggest that at least some of the C. psittaci 6BC POMPs, although part of a supramolecular complex, are not extensively interpeptide disulfide cross-linked.

View larger version (70K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE and immunoblot analysis of detergent-insoluble fractions of EBs. EBs were harvested and purified at 48 h postinfection. Where indicated, EBs were treated with 20 mM DTT and 5% ME before the preparation of Sarkosyl COMCs and SDS-insoluble complexes. All preparations were suspended in Laemmli buffer (21) with (A and C) and without (B and D) the addition of reducing agents (10 mM DTT and 5% ME), heated to 90°C, fractionated by SDS-PAGE, and analyzed by Coomassie staining (A and B) or immunoblotting with anti-POMP 90A (C and D). Lanes: 1, EBs; 2, Sarkosyl (Sark) COMC; 3, Sarkosyl COMC prepared from reduced EBs; 4, SDS-insoluble complex; 5, SDS-insoluble complex prepared from reduced EBs. The intensively stained band below the MOMP at approximately 40 kDa (lanes 1 to 3) is a degradation product of the MOMP. MW, molecular weight standards, in thousands.

Stage-specific expression of POMPs. To determine the developmental stage specificity of the putative POMPs, whole chlamydiae and Sarkosyl COMCs were analyzed by SDS-PAGE and immunoblotting at 15, 24, 28, and 48 h postinfection (Fig. 3). The 15- to 28-h harvests consisted of density-gradient-purified RBs, and the 48-h preparations were purified, Nonidet P-40-treated EBs (to eliminate osmotically fragile RBs in the EB fraction). The material loaded on the gels was adjusted so that the amount of the MOMP, which is present throughout the cycle, was approximately the same in all preparations. Under these conditions of analysis, the putative POMPs and the late-stage-specific CRPs were not detected in either whole RBs or COMC preparations at 15 h but were detected in increasing amounts as the infection progressed to completion between 28 and 48 h postinfection. These observations indicate that the six putative POMPs of C. psittaci 6BC are late-stage specific.

View larger version (90K): [in this window] [in a new window]   FIG. 3.   SDS-PAGE (A) and immunoblot analysis (B) of EB and RB proteins during the developmental cycle. Electrophoresis was carried out under reducing conditions, and the blot was reacted with anti-POMP 90A. Lanes: 1, 15-h RBs; 2, 24-h RBs; 3, 28-h RBs; 4, 48-h EBs; 5 to 8, Sarkosyl COMCs of the samples shown in lanes 1-4. Prestained protein standards (MW, in thousands) were run at the sides of the panels. The positions of putative POMPs (positions 1 to 6), the CRPs, and the MOMP are indicated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We identified three POMPs in C. psittaci 6BC that cross-react with antiserum prepared against the POMP 90 family of OEA C. psittaci S26/3. It is likely that three additional proteins (proteins 1 to 3) in C. psittaci 6BC are POMPs on the basis of their large sizes, location in the OM, and the lack of predicted non-POMP OM proteins of similar sizes in the C. trachomatis and C. pneumoniae genome databases (18, 35, 38). Two of the six putative POMPs have a surface-exposed, trypsin-sensitive site. The precise role of disulfide bonds in maintaining C. psittaci POMPs in the OM is not clear. In contrast to the MOMP, protein 7, and the CRPs, some proportion if not all of the POMPs were capable of migrating in gels at approximately the same rates in the presence and absence of reducing agents, suggesting that they are not interpeptide cross-linked. On the other hand, the POMPs were released from the SDS-insoluble complexes and partially solubilized from the Sarkosyl COMCs when the complexes were prepared from EBs treated with reducing agents. It is possible that intrapeptide disulfide bonds play some role in maintaining POMPs in detergent-insoluble complexes, perhaps by allowing a noncovalent association of POMPs with themselves or with a cross-linked MOMP in the OM. Alternatively, they may associate with supramolecular structures in the periplasm, such as the glycanless wall polymer proposed by Ghuysen and Goffin (11) or the disulfide-cross-linked CRP complex proposed by Everett and Hatch (10). The ability of the MOMP to form oligomeric complexes in SDS has been shown by Wyllie et al. (40) and McCafferty et al. (28)

Immunoelectron microscopic studies by Longbottom et al. (24, 25) indicated that one or more members of the POMP 90 family are exposed on the surfaces of 24-h RBs and 48-h EBs of ovine C. psittaci S26/3 24 and 48 h postinfection. This observation is consistent with our finding that C. psittaci 6BC POMPs were found in gradient-purified RBs at 24 and 28 h postinfection and in EBs. However, we failed to detect POMPs and the CRP doublet proteins when we examined middle-stage, logarithmically dividing RBs, suggesting that C. psittaci 6BC POMPs, like the CRPs, are late-stage specific and therefore are not required for RB growth and cell division. In a preliminary study, Lindquist and Stephens (23) reported that transcripts orthologous to the nine C. trachomatis D POMP genes could be detected in C. trachomatis L2 by reverse transcription-PCR between 10 and 48 h postinfection, suggesting that C. trachomatis POMPs are made early in the growth cycle. However, the POMP transcripts were not quantified, nor were their levels compared with the levels of other chlamydial gene transcripts; thus, the possibility that C. trachomatis POMP genes are relatively more highly expressed during the late phase of the developmental cycle cannot be excluded.

The numbers of POMP genes in the genomes of C. psittaci 6BC and other C. psittaci strains are not known. Six POMP genes have been identified in the genome of C. psittaci S26/3; however, by analogy to C. trachomatis and C. pneumoniae, additional POMP genes are likely to be present in the genomes of C. psittaci strains. Our finding of six POMPs in the OM of C. psittaci 6BC is higher than the numbers found thus far in C. trachomatis (two) and C. pneumoniae (four) (19, 30). It is possible that many more POMP genes are expressed as proteins in chlamydiae growing in tissue cultures but are present in amounts not readily detected by conventional staining and labeling techniques. Alternatively, some POMP genes may be expressed only under specific in vivo conditions, as speculated by Birkelund et al. (8).

The C. pneumoniae POMP genes are particularly interesting in that six genes do not appear to encode full-length proteins and thus may represent a reservoir for recombination or mutation (14, 18). At least one example of mutation has already been identified: the open reading frame (Cpn 449/450) that encodes Pmp10, also referred to as OMP5, in strain CDC/CWL-029/VR1310 contains a frameshift, as sequenced by Kalman et al. (18), but is expressed as a full-length POMP, as reported by Knudsen et al. (19) and Pedersen et al. (34). Pedersen et al. (34) speculated that the expression of the open reading frame was dependent on the addition or deletion of nucleotides from a poly(G) tract within the coding region, most likely at the genomic level.

An interesting result of our study was the finding that sulfo-NHS-biotin, which is useful for labeling proteins exposed on the surfaces of eukaryotic cells, appeared to penetrate the OM of chlamydial EBs, labeling a wide range of proteins that were not labeled with [¹²⁵I]TID. The EB form of chlamydiae is impermeable to ATP, GTP, and amino acids that are taken up by specific transport mechanisms in RBs (17). Bavoil et al. (4) suggested that the impermeability of EBs may be a function of the MOMP porin, which they found in a liposome swelling assay to require the reduction of disulfide bonds and the blockage of sulfhydryl groups with iodoacetamide for activity. However, Wyllie et al. (40) found that an MOMP reconstituted into planar lipid bilayers maintained porin activity without the blockage of sulfhydryl residues and that treatment of the bilayers with oxidative reagents did not inhibit activity. It is possible, therefore, that the hydrophilic cross-linking reagent gained entry to the periplasm through the MOMP porin and that the impermeability of EBs is related to the inactivity of specific transport systems in the inner membrane rather than to exclusion by the OM.

Another interesting result of our study was the discovery of 48-kDa protein 7, which is one of the interpeptide-disulfide-cross-linked OM proteins found in EBs but which does not appear to be late-stage specific (Fig. 3). Mass spectroscopic analysis of a gel slice containing protein 7 revealed the presence of an MOMP in the band. It is possible that protein 7 is a posttranslationally modified form of MOMP, as has been proposed by Kuo et al. (20) for the MOMP of C. trachomatis L2. However, Kuo et al. (20) presented evidence that a "standard" MOMP, rather than a protein of higher molecular weight, is glycosylated, as is the case for protein 7. Alternatively, protein 7 may be an unknown OM protein which is contaminated by the highly abundant MOMP, the result of smearing during electrophoresis. In either case, it is interesting that Newhall et al. (32) found in multiple serovars of C. trachomatis a similar band that reacted on immunoblots with the sera of some patients infected with C. trachomatis. We are currently attempting to identify the homolog of protein 7 in a DNA-sequenced strain of C. trachomatis.