Role of a Highly Conserved Bacterial Protein in Outer Membrane Protein Assembly

See allHide authors and affiliations

Science  10 Jan 2003:
Vol. 299, Issue 5604, pp. 262-265
DOI: 10.1126/science.1078973


After transport across the cytoplasmic membrane, bacterial outer membrane proteins are assembled into the outer membrane. Meningococcal Omp85 is a highly conserved protein in Gram-negative bacteria, and its homolog Toc75 is a component of the chloroplast protein-import machinery. Omp85 appeared to be essential for viability, and unassembled forms of various outer membrane proteins accumulated upon Omp85 depletion. Immunofluorescence microscopy revealed decreased surface exposure of outer membrane proteins, which was particularly apparent at the cell-division planes. Thus, Omp85 is likely to play a role in outer membrane protein assembly.

The cell envelope of Gram-negative bacteria consists of an outer membrane (OM) and an inner membrane (IM), separated by the periplasm. The IM is a phospholipid bilayer, whereas the OM is asymmetric with phospholipids and lipopolysaccharides (LPS) in the inner and outer leaflet, respectively. Whereas integral IM proteins are mostly α-helical, OM proteins (OMPs) present a different structure, the β barrel (1). OMPs are synthesized in the cytoplasm with an NH2-terminal signal sequence required for IM translocation by the Sec machinery (2). After translocation, the signal sequence is cleaved off, releasing the mature protein in the periplasm. Little is known about the subsequent steps in OMP biogenesis. OMPs fold in the periplasm before their insertion into the OM (3), and LPS stimulates the folding of OMPs in vitro (4). Also, the periplasmic chaperone SurA stimulates OMP folding (5), whereas another periplasmic protein, Skp, plays an unidentified role in OMP biogenesis (6). The insertion of proteins into membranes generally requires a proteinaceous machinery. However, no components of the putative OMP insertion machinery have been identified. Such components are likely to be conserved among Gram-negative bacteria and essential for their viability. A protein possibly fulfilling these criteria is the surface antigen designated Omp85 in Neisseria spp. and D15 inHaemophilus spp. Genes encoding Omp85/D15 homologs are present in all Gram-negative bacteria examined (7) [fig. S2 (8)]. Attempts to delete the structural gene for this protein in Haemophilus ducreyi andSynechocystis sp. were unsuccessful, suggesting it is an essential protein (7, 9). Furthermore, because theomp85 gene is located in close proximity to other genes involved in the biogenesis of OM components (8), we investigated the possible function of the neisserial Omp85 in OMP assembly.

To verify whether Omp85 was essential for viability, we inactivated the chromosomal omp85 gene in a strain carrying an intact omp85 gene under an isopropyl-β-d-thiogalactopyranoside (IPTG)–inducible promoter on a plasmid, designated pRV2100 [fig. S1 (8)]. Wild-type and omp85-mutant bacteria, both containing pRV2100, were grown on a plate containing 0.1 mM IPTG and then streaked onto plates containing IPTG or glucose (Fig. 1A). The wild-type strain grew normally overnight on all plates, whereas the mutant strain failed to form colonies outside the initial streak on plates containing ≤10 μM IPTG. Thus, omp85 expression was essential for growth. To examine the growth characteristics in further detail, we grew the strains for 2 hours in liquid medium containing IPTG and then washed and grew them further in medium without IPTG. After 8 hours, both cultures had reached the same optical density (Fig. 1B). The amount of Omp85 had reached the chromosomally encoded level in the wild type, whereas it was drastically decreased but still detectable in the mutant (Fig. 1C). After dilution of the cultures into fresh medium, growth of the mutant clearly lagged behind that of the wild type (Fig. 1B), and Omp85 was undetectable by immunoblotting after another 11 hours of incubation (Fig. 1C). Thus, Omp85 was essential for viability.

Figure 1

Omp85 is essential in N. meningitidis.(A) Wild-type (WT) and the omp85mutant (Omp85), both containing plasmid pRV2100, were grown overnight on plates containing either IPTG at the concentration indicated or glucose. (B) The same strains were grown in liquid medium containing 0.1 mM IPTG. Att 0, cells were collected by centrifugation, washed, and resuspended in medium without IPTG containing 0.4% glucose. At t 1, the cultures were diluted 40-fold in medium with glucose and grown further. (C) Cells collected at t 0,t 1, and t 2 were analyzed by immunoblotting with a mouse polyclonal antiserum raised against nonnative Omp85 (left). Additionally, samples of the wild-type strain with and without pRV2100 were analyzed to compare omp85expression levels (right).

To evaluate the putative role of Omp85 in OMP biogenesis, we analyzed the cell-envelope protein patterns of wild-type and omp85-mutant cells collected at time t 2(Fig. 1B). When the cell envelopes were analyzed under denaturing conditions, no major effect of Omp85 depletion on the OMP profiles was observed (Fig. 2A). The major OMPs of Neisseria meningitidis are the trimeric porins PorA and PorB. These trimers are stable and do not dissociate into monomers in SDS–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer with a low concentration of SDS (0.2%), if not heated before electrophoresis (seminative conditions) (10). When analyzed under such conditions (Fig. 2A), PorA and PorB could be seen to have accumulated mainly as monomers in Omp85-depleted cells. Immunoblotting confirmed the accumulation of PorA and PorB monomers in the mutant (Fig. 2C). The misassembly of the porins in the mutant was underscored by their increased sensitivity to trypsin (Fig. 2A), and their extractability from the cell envelope preparation with urea (Fig. 2B) demonstrated that they were not integrally inserted in the membrane.

Figure 2

Omp85 depletion affects OMP biogenesis. The wild-type (C) and the omp85 mutant (M), each containing pRV2100, were grown in glucose-containing medium as in Fig. 1B. At timet 2, cells were collected and cell envelope protein patterns were analyzed by SDS-PAGE with or without denaturation at 95°C as indicated. (A) Coomassie blue–stained gel. Where indicated, cell envelopes were treated with trypsin. (B) Coomassie blue–stained gel of cell envelope samples extracted with urea; s, urea-soluble fraction; ns, nonsoluble fraction. (C to G) Immunoblot analysis of cell envelopes with antibodies against PorB and PorA (C), PilQ (D), FrpB (E), OMPLA (F), or IgA1 protease (G). The antibodies to FrpB do not recognize the native oligomeric form of the protein. Molecular size markers (in kD) are indicated.

We also investigated the effect of Omp85 depletion on the assembly of other OMPs. The secretin PilQ, involved in type IV pili formation, forms very stable high molecular weight oligomers (11). Upon Omp85 depletion, the amount of oligomers was severely reduced, and monomeric PilQ accumulated (Fig. 2D). The heterooligomeric complex formed by the siderophore receptor FrpB and RmpM (12) was not assembled in the absence of Omp85, because FrpB was found to be monomeric under seminative SDS-PAGE conditions (Fig. 2E). OM phospholipase A (OMPLA) is normally present in a folded monomeric form (13), which displays heat modifiability—i.e., the correctly folded monomer migrates faster in seminative SDS-PAGE than the heat-denatured form (14). In the Omp85-depleted strain, OMPLA accumulated mainly in its unfolded form (Fig. 2F). Autotransporters, such as the neisserial immunoglobulin A1 (IgA1) protease (15), contain a COOH-terminal β-barrel domain, which forms a pore in the OM allowing transport of the NH2-terminal passenger domain across the OM (16). After translocation, the passenger domain is usually cleaved and either released into the extracellular medium or bound noncovalently to the cell surface. In the Omp85-depleted strain, the processed passenger domain of IgA1 protease was barely detectable, and the full-length form was found to accumulate (Fig. 2G). Thus, Omp85 appeared to possess a general function in the assembly of OMPs.

We investigated the cell-surface exposure of OMPs by immunofluorescence microscopy on formaldehyde-fixed bacteria (8) with antibodies directed against PorB, PorA, PilQ, and an OMP of unknown function, NspA (17). In all cases, labeling was weaker after Omp85 depletion (as shown for PorB labeling in Fig. 3, A and B). Because the amount of porins was not reduced after Omp85 depletion (Fig. 2A), most of the proteins produced apparently were not accessible to the antibodies in the intact cells, i.e., they were not properly inserted in the OM. Moreover, instead of the wild-type diplococcal phenotype, we found many bacteria in groups of four (tetrapacs) under Omp85 depletion (Fig. 3, A and B), indicating a cell-division defect (18). Such tetrapacs were only rarely observed in the wild-type culture (an example is shown in Fig. 3C). Whereas the wild-type cells were equally labeled over the entire surface with all four antibodies used, the mutant cells were irregularly labeled (Fig. 3D). Labeling was absent in division septa, where new membranes are expected to be synthesized. This phenotype reinforces the role of Omp85 in OMP biogenesis.

Figure 3

Immunofluorescence microscopy analysis. Wild-type (WT) (A and C) and mutant (Omp85) (B and D) cells collected at time t 2 (Fig. 1B) were fixed with formaldehyde and stained with antibodies directed against PorB (A and B) or PilQ (C and D) and Alexa fluorochrome-conjugated secondary antibody. Bar: 1.5 mm (A and B) and 0.5 mm (C and D). Neither wild-type nor mutant cells bound antibody Mn23G2.38, which reacts exclusively with denatured PorA, indicating that formaldehyde treatment does not denature PorA.

Because LPS stimulates OMP folding in vitro (4), the accumulation of nonnative OMPs in theomp85 mutant could be a consequence of an LPS biogenesis defect. Indeed, after Omp85 depletion, the amount of LPS in the cell envelope was reduced by 60%. However, this reduction might be a consequence of defective assembly of the putative LPS transport machinery. We disfavor the possibility that Omp85 is directly involved in LPS biogenesis and only indirectly in OMP biogenesis for the following reasons: (i) in contrast to Omp85, LPS is not essential inN. meningitidis (19), and OMPs were still correctly assembled into the OM of an LPS-deficient strain (20). (ii) An omp85 homolog is also present in bacteria lacking LPS, such as Treponema pallidum andBorrelia burgdorferi (fig. S2) (8). (iii) Omp85 is homologous to the chloroplast protein-import machinery component Toc75 (7) [fig. S3 (8)], suggesting a role for Omp85 in protein rather than in lipid transport. Consistently, the Synechocystis Omp85 homolog SynToc75 has been shown to form channels in planar lipid bilayers with a high affinity for peptides (21).

We obtained evidence for a direct role of Omp85 in OMP assembly by demonstrating an interaction between a native Omp85 complex and denatured PorA. Cross-linking experiments suggested that Omp85 belongs to a multisubunit complex (22). Consistently, we detected a high molecular weight complex containing Omp85 in cell envelopes of wild-type strain H44/76 analyzed under seminative conditions (Fig. 4A). Proteins from OM vesicles of a porA mutant overexpressing Omp85 were separated by seminative SDS-PAGE and blotted onto a nitrocellulose membrane. The membrane was then incubated with purified denatured PorA, and the binding of PorA to the membrane was detected with antibodies to PorA. PorA bound specifically to the native Omp85 complex and not to the denatured Omp85 or to any other protein in the OM vesicle preparation (Fig. 4B). The binding of PorA to the Omp85 complex, possibly mediated by the NH2-terminal domain of Omp85, which is periplasmically exposed in the proposed topology model (fig. S4) (8), underscores a direct role of this complex in OMP assembly.

Figure 4

Direct interaction of denatured PorA with a high molecular weight Omp85 complex. (A) Cell envelopes of strain H44/76 were analyzed by SDS-PAGE under denaturing and seminative conditions followed by immunoblot analysis with antiserum to Omp85. (B) OM vesicles of a porA mutant of H44/76 containing pRV2100 cultivated with IPTG were separated by seminative SDS-PAGE, and the proteins were blotted onto a nitrocellulose membrane. Where indicated, lanes from the membrane were incubated with purified monomeric PorA. The blots were further incubated with primary antibodies against Omp85 or PorA and developed. The high molecular weight form of Omp85 (HMW) and the denatured Omp85 monomer are indicated. Molecular size marker proteins (in kD) are indicated.

The presence of Toc75 homologs in Gram-negative bacteria, includingSynechocystis, suggests a common evolutionary origin (23). Because Toc75 is involved in protein import into chloroplasts, the bacterial homologs have been suggested to be involved in protein export (23). However, protein transport through the OM is not an essential process in Gram-negative bacteria. The observation that Omp85 is essential for viability and is ubiquitous in Gram-negative bacteria supports a more general essential function, such as the assembly of OMPs as demonstrated in this study. After transfer of the ancestral gene to the plant nucleus early in endosymbiosis, the transporter might have reversed orientation to mediate the import of proteins into the stroma of chloroplasts (7).

Supporting Online Material

Materials and Methods


Figs. S1 to S4

Table S1


  • * To whom correspondence should be addressed. E-mail: j.p.m.tommassen{at}


View Abstract

Navigate This Article