INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The proper functioning of proteins requires their correct localization to predetermined cellular locations. In the gram-negative bacterium Escherichia coli K-12, a special class of proteins known as porins (24) must be properly targeted and assembled in the outer membrane to form channels that facilitate the diffusion of hydrophilic solutes. The atomic structure showed that porins primarily consist of antiparallel  strands that are arranged in a pseudocyclic -barrel structure, which encloses a water-filled channel (6, 29). On the periplasmic side,  strands are adjoined by short turns, whereas long loops provide connections on the medium-exposed side. Although the majority of loops are surface exposed, one or more loops fold inward thus restricting the channel. Surface-exposed loops are often utilized by bacteriophages as their receptors (35, 36, 37).

Outer membrane proteins (OMPs) are synthesized in the cytoplasm as precursors with an amino-terminal signal sequence that assists in exiting the cytoplasm (26). Cleavage of the signal sequence on the periplasmic side yields mature molecules that contain the necessary information for outer-membrane targeting (2, 8). It is increasingly recognized that the targeting of mature porin molecules to the outer membrane is associated with folding events that allow monomers to gain structures crucial for assembly into protease- and heat-resistant trimers (11, 19, 22, 27, 38). Thus, targeting and folding are likely to be overlapping events. The issue of whether or not trimerization occurs prior to outer membrane insertion has not been fully resolved.

Besides intragenic information, proper assembly of OMPs relies on several extragenic factors. The periplasmic proteins that catalyze correct disulfide bond formation (DsbA) and peptidyl prolyl cis-trans isomerization (SurA and FkpA) have profound effects on OMP assembly (for reviews see references 8 and 23). Periplasmic protein Skp shows affinity to unfolded soluble OMPs, thus suggesting its role as a general chaperone (5, 28). Besides proteins, the lipid components of the outer membrane, phospholipids and lipopolysaccharide (LPS), also play an important role in OMP assembly (1, 13, 14, 30). As LPS is exclusively localized in the outer membrane, it is possible that OMPs reach the outer membrane through their interactions with LPS.

Alterations within the primary sequence or extragenic factors can lead to defective OMP assembly. Assembly-incompetent intermediates tend to misfold and are diverted to aggregation and degradation pathways. Major protease DegP, whose activity resides in the periplasmic space, is responsible for the degradation of misfolded envelope proteins (16, 33). DegP's activity is essential when bacterial cultures are grown in rich media at 42°C. The reason for this essentiality is not entirely clear, but presumably a greater number of misfolded proteins accumulate under these growth conditions. In an exciting development concerning DegP, a recent study has shown that it contains both proteolytic and general chaperone activities (32). Aberrant OMP assembly in the envelope leads to elevated DegP levels (8, 9, 17, 23). Presumably a rise in proteolytic or chaperone activity assists cells in coping with the physiological stress by minimizing the population of misfolded polypeptides. Elevated DegP levels rely on the activation of an alternative sigma factor, ^E, which controls degP transcription (17).

In this study, we carried out an in vivo analysis of OmpF mutants with altered carboxyl termini, which are occupied by phenylalanine residues in most OMPs. Tommassen's group first reported the significance of this conserved residue in PhoE biogenesis; its replacement or deletion affected PhoE's outer membrane incorporation and assembly into trypsin- and sodium dodecyl sulfate (SDS)-resistant trimers (7, 34). Replacement of OmpF's terminal phenylalanine with unrelated residues affected one or more steps of assembly. Expression of assembly-defective OmpF proteins compelled DegP's presence, thus pointing to a critical role for DegP in mutant OMP assembly.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, media, and chemicals. PLB3260 (lamB106 ompF-lacZ fusion) and RAM831 (lamB106 ompF80 cog192 [OmpG⁺] ompC-lacZ fusion) were used as host strains for pTrc99A (Pharmacia) and its recombinant derivatives. A degP::Km^r null allele was obtained from KS474 (33). PND2000 (MC4100 degP-lacZ fusion) (9) was used to assess degP's regulation. Luria broth and M63 salt-based minimal medium were prepared as described by Silhavy et al. (31). When cultures were grown on minimal medium, 0.4% glycerol was the supplemented carbon source. Maltodextrins were purchased from Pfanstiehl Laboratories and were further purified as described previously (21). When required, ampicillin (50 µg/ml), chloramphenicol (25 µg/ml), and kanamycin (25 µg/ml) were added to the growth medium. Presoaked antibiotic discs were purchased from Difco. [³⁵S]methionine Expre protein labeling mixture was purchased from DuPont-New England Nuclear. Heat-killed Staphylococcus aureus cells (Pansorbin) were purchased from Calbiochem. DNA-modifying enzymes were purchased from Promega. All other chemicals were of analytical grade.

Cloning, mutagenesis, and other genetic manipulations. The ompF205 gene (19) without its native promoter was amplified from the chromosome using primers complementary to the start (5'-GAGGGTAATAAACCATGGTGAAGCGCAATATTCTGGCAGTG-3') and end (5'-GACGAGGATCCATTATGGTTACAGAAGG-3') of the gene. Restriction sites for NcoI (start of the gene; underlined) and BamHI (end of the gene, underlined) were incorporated for cloning into a pTrc99A vector plasmid where an isopropyl--D-thiogalactopyranoside (IPTG)-inducible promoter controlled the expression of ompF205. The first two residues of the terminal phenylalanine codon of ompF (UUC) were randomly altered in PCRs by utilizing a mutagenic reverse primer (5'-GAGGGATCCTATTAGNNCTGGTAAACGATACCCAC-3') and the forward NcoI primer listed above. The reverse primer incorporated a BamHI site (underlined) and randomized nucleotides (NN) at positions corresponding to the first two residues of the terminal UUC codon of ompF. PCR fragments were restricted with NcoI and BamHI and cloned into pTrc99A. DNA sequence analysis of several independent clones was carried out to reveal alterations at ompF's terminal codon. The ompF gene and its IPTG-regulated promoter from plasmid clones were recombined first onto a lambda vector and then to the chromosome by a method described by Boyd et al. (4). Plasmid isolation, transformation, and P1 transduction were carried out utilizing standard laboratory protocols.

Protein methods. Whole-cell envelopes from cultures grown overnight were obtained by the French press lysis procedure (19). Envelope samples were analyzed by SDS-11% polyacrylamide gel electrophoresis (PAGE). For better resolution of OmpF from nearby protein bands, urea (4 M) was added to the running gel. Protein bands were visualized after staining gels with Coomassie blue. When required, envelope samples were treated with sodium sarcoysl (1% final concentration) at room temperature for 30 min. Sarcosyl-insoluble material was pelleted by centrifuging samples at 50,000 rpm for 30 min at 4°C in a tabletop ultracentrifuge (TLA120.2 rotor; Beckman).

Pulse-chase labeling and trimer assays. Protein-labeling experiments were carried out with cells that also produced OmpG (10). Constitutive expression of the OmpG porin provided identical growth patterns of strains expressing the various forms of OmpF. Pulse-chase labeling and trimer extractions were carried out as described previously (19, 22). Equal amounts of trichloroacetic acid (TCA)-precipitable radioactive counts were used for immunoprecipitating OmpF with polyclonal antibodies raised against trimers or denatured OmpF (19). Immunocomplexes were removed by heat-killed Staphylococcus aureus cells, solubilized in sample buffer, and analyzed by SDS-PAGE after heating at 62.5 or 100°C for 5 min. Gels were fluorographed, dried, and exposed to X-ray films at 70°C. Radioactive protein bands were quantified by a Bio-Rad gel analyzer.

-Galactosidase assays. Assays were carried out as described by Miller (18).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of OmpF mutants with substitutions at F340. A variant ompF allele, ompF205, was utilized to obtain mutants with substitutions at the carboxy-terminal phenylalanine residue (F340). The single alteration of R82 to S in OmpF205 increases its channel size, thus permitting the transport of larger sugars such as maltodextrins (3). The Dex⁺ phenotype (ability to grow on maltodextrins) of OmpF205 thus provides a powerful tool for mutant isolation and characterization (19). The ompF205 gene was cloned into plasmid vector pTrc99A, where an IPTG-inducible promoter controlled its expression. A comparison of envelope protein profiles from strains expressing OmpF205 either from the chromosome or from the plasmid under noninducing (without IPTG) conditions showed that they produced similar levels of the OmpF205 protein (Fig. 1; compare lanes 3 and 5).

View larger version (23K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE analysis of whole-cell envelopes obtained from strains expressing chromosomally encoded OmpF (lanes 1 and 2), chromosomally encoded OmpF205 (lanes 3 and 4), and plasmid-encoded OmpF205 under noninducing (without IPTG) growth conditions (lanes 5 and 6). Some strains also expressed OmpC (lanes 1, 3, and 5) or OmpG (lane 6).

Effects of F340 substitutions on OmpF levels and channel activities. Examination of envelopes obtained from cultures grown overnight revealed the extent of the defects in various mutants (Table 1). The F340Y and F340L substitutions produced relatively little effect on OmpF levels. Substitutions involving histidine and proline produced an intermediate effect, while those involving arginine, serine, and threonine greatly reduced the protein level. A substitution involving alanine produced the severest effect, as no OmpF could be detected.

Substitutions of F340 affect trimerization and not synthesis. To determine the levels of newly synthesized mature OmpF molecules and trimers, cells were labeled with [³⁵S]methionine for 20 s and chased for 30 s in the absence of new protein synthesis. This period was sufficient to complete the processing of any residual OmpF precursor molecules into mature forms. OmpF was solubilized from labeled cells by utilizing a method that efficiently extracts soluble and insoluble forms of the protein including SDS-resistant metastable and stable trimers (19, 22). Prior to immunoprecipitation, labeled protein extracts were divided into two halves, one of which was boiled to convert the various species of OmpF into a denatured form. Aliquots of denatured (boiled) or native (unboiled) cell extracts corresponding to an equal number of TCA-precipitable counts were then used for immunoprecipitation with antibodies raised against denatured OmpF or native trimers, respectively. The data presented in Fig. 2 showed that levels of OmpF synthesis in all mutants were very similar to that from the parental strain. The slightly reduced levels of OmpF detected in the case of the proline, histidine, threonine, and alanine substitutions may reflect degradation immediately after synthesis. These results showed that the various substitutions at the terminal OmpF residue did not substantially interfere with the protein's synthesis and precursor processing.

View larger version (58K): [in this window] [in a new window]   FIG. 2.   Relative levels of newly synthesized total and trimeric OmpF. Cells expressing the various OmpF proteins were labeled for 20 s with [35S]methionine and chased for 30 s with 20 mM nonradioactive methionine and chloramphenicol (200 µg/ml). Immunoprecipitation conditions are described in the text. Immunocomplexes were solubilized in SDS sample buffer, heated for 5 min at 100°C, and analyzed by SDS-PAGE. OmpF-specific bands were quantified from two independent experiments, and values relative to the parental protein bands were graphed.

Substitutions of F340 affect trimerization kinetics. Trimerization kinetics were examined by labeling cells with [³⁵S]methionine for 20 s and chasing for 1, 2, 4, 6, and 8 min in the absence of new protein synthesis. Trimer-specific antibodies were used to immunoprecipitate trimers from aliquots of labeled cell extracts corresponding to an equal number of TCA-precipitable counts. After immunoprecipitation, precipitates were boiled to convert various trimeric species into denatured forms, which were then analyzed by SDS-PAGE. The results presented in Fig. 3 showed that, for the parental protein, trimers accumulated in two distinct steps: almost two-thirds of the total SDS-resistant trimers rapidly assembled within 1 min of chase, while the remaining one-third assembled with a slower kinetic pathway lasting 5 to 7 min. Mutants with an F340Y or F340L substitution had levels of trimers similar to that of the parental protein after 1 min of chase and displayed kinetics of trimerization very similar to that observed for the parental protein (Fig. 3).

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Kinetics of OmpF trimerization. Cells expressing the various OmpF proteins were labeled with [35S]methionine for 20 s and chased with nonradioactive methionine in the presence of chloramphenicol. OmpF trimers extracted from samples withdrawn after various chase periods were immunoprecipitated with OmpF trimer-specific antibodies. Immunocomplexes were heated at 100°C for 5 min and analyzed by SDS-PAGE. OmpF-specific bands were quantified and plotted.

Up-regulation of degP and destabilization of the outer-membrane permeability barrier by mutant OmpF proteins. The expression of degP is known to be up-regulated in response to either the presence of misfolded proteins in the envelope or overexpression of OMPs (8, 9, 17). Plasmids expressing various OmpF proteins were introduced into a strain containing degP::lacZ⁺ so that the effect on degP expression could be analyzed by measuring -galactosidase activity. The presence of a vector plasmid or a plasmid expressing the parental OmpF205 protein did not alter basal -galactosidase activity and hence degP expression (Table 2). In contrast to what was found for the parental protein, the presence of plasmids expressing mutant proteins elevated degP expression (Table 2). As the mutant proteins were not overproduced, the elevated degP expression observed was likely due to the transient accumulation of misfolded assembly intermediates. It is interesting to note that the OmpF mutants with an F340Y or F340L substitution, which appeared normal in trimer assays (Fig. 3), also up-regulated degP transcription.

DegP's involvement in the biogenesis of mutant OmpF proteins. The DegP protease has been shown to degrade misfolded proteins in the envelope (22, 33). We examined the role of DegP in the biogenesis of mutant OmpF proteins produced under noninducing (without IPTG) growth conditions. It was reasoned that if DegP is responsible for degrading mutant OmpFs, their levels might rise in cells devoid of DegP. On the other hand, if the accumulation of mutant OmpF proteins is toxic, the absence of DegP may cause lethality. Plasmids encoding OmpF205 and three variants with an F340Y, F340L, or F340A substitution were successfully introduced into a degP null strain at 30°C but not those expressing an OmpF with an F340P, F340H, F340R, F340S, or F340T substitution (Table 3). We surmise from these observations that DegP mediates the degradation of certain mutant proteins that become toxic if left to accumulate. Curiously, the expression of OmpF with an F340A substitution in a DegP strain conferred temperature sensitivity at 37°C, suggesting an exaggerated assembly defect at the elevated growth temperature.

Overproduction of protease-deficient DegP reverses the lethal phenotype of OmpF mutants. Recently it has been shown that DegP contains both protease and chaperone activities (32). Since the degP null allele utilized in this study disrupts both activities, it could not be concluded which of the two activities is essential in strains expressing mutant OmpF proteins. To address this, we utilized a variant DegP protein in which an S210A substitution abolished the protease activity without affecting its proposed chaperone function (32). Overproduction of DegP_S210A in a degP null background permitted the growth of all strains expressing mutant OmpF proteins at 37°C (Table 3), thus showing a protective role for protease-deficient DegP in overcoming the lethal effects of mutant OmpFs.

View larger version (58K): [in this window] [in a new window]   FIG. 4.   Effects of DegPS210A on OmpF levels (A), detergent solubility (B), and membrane localization (C). OmpF was examined by Western blot analysis utilizing antibodies that recognized denatured OmpF and OmpA and an unknown protein migrating between OmpF and OmpA. All strains expressed wild-type DegP. Samples from strains overproducing DegPS210A are indicated by a plus sign, an upward arrow, or the protein designation. Inner-membrane (IM) and outer-membrane (OM) fractions corresponding to buoyant densities of 1.15 and 1.22, respectively, were analyzed.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The fact that phenylalanine occupies the carboxyl-terminal position of nearly all OMPs points to a role for this residue in biogenesis (34). Support for this notion initially came from mutant analysis in which the deletion or substitution of phenylalanine impaired PhoE's assembly (34). Curiously the assembly defect was observed only when the mutant PhoE proteins were overproduced (7). The in vivo data presented in this study provided many useful insights concerning the assembly and physiology of mutant OmpF proteins. Unlike what was found for PhoE, defects in OmpF assembly were observed even when OmpF was synthesized at a level similar to that produced from the chromosomal allele.

Assembly defects of OmpF F340 mutants. Substitutions of F340 with dissimilar amino acids drastically affected the assembly and stability of SDS-resistant metastable and stable trimers. The extent of assembly defect was dependent on the substituting amino acid residue. As the synthesis of mature monomers was virtually unaffected by the various substitutions, the lack of trimers was primarily due to a defect in an assembly step involving monomer trimerization. This defect was distinct from the one displayed by previously described OmpF assembly mutants, where a late assembly step involving the conversion of metastable trimers to stable trimers was conditionally blocked (19). The reduced ability to produce trimers suggested that substitutions at the terminal position might have affected events such as the correct folding of monomers or their proper interactions with assembly-promoting factors LPS and Skp.

Physiological effects of mutant OmpF proteins. The data presented in this study showed that the expression of all mutant OmpF proteins elevated degP transcription, which is known to be up-regulated via ^E in the event of aberrant envelope protein assembly (9, 17). Interestingly, although OmpF_Y340 and OmpF_L340 showed parental trimerization kinetics, they too induced degP expression to about the same extent as did the assembly-defective mutant OmpF proteins. So the amplitude of degP up-regulation did not truly reflect the degree of assembly defect observed. It is conceivable that, even though a greater pool of misfolded assembly intermediates accumulates in severely defective mutants than in those with no defect or a less severe defect, the rapid turnover of the signal for degP up-regulation causes the signal to be about the same in all cases. The hypersensitivity phenotypes of assembly mutants were more reflective of their assembly defects than was degP up-regulation. It could be that molecules which are not degraded as rapidly as those that induce degP upregulation confer the hypersensitive phenotype. Additionally, different threshold levels of molecules may be needed to produce diverse phenotypes.

Critical role for DegP in mutant OmpF assembly. As stated above, several mutant OmpF proteins were synthesized normally yet their levels in the envelope were significantly lower than that of the parental protein. This reflected their degradation prior to assembling into stable trimers. The degradation of mutant OmpF proteins appeared to be mediated by the DegP protease whose activity resides in the periplasm (33). DegP has been previously shown to play an important role during the biogenesis of envelope proteins in E. coli (22, 32, 33). Here we have shown that the requirement for DegP for the viability of cells expressing OmpF bearing an F340P, F340H, F340R, F340T, or F340S substitution was absolute. Initially, this conclusion was reached due to our inability to construct strains lacking DegP and simultaneously expressing the aforementioned mutant OmpF proteins from plasmids under noninducing conditions. However, when ompF genes from plasmids were recombined into the chromosome, the further lowering of protein expression allowed DegP to become dispensable. Interestingly, the requirement for DegP under these conditions became conditional, i.e., cellular toxicity was observed when mutant OmpF expression was induced by IPTG. It should be emphasized that induction with IPTG did not lead to overexpression of OmpF proteins from the recombinant chromosomal alleles; rather, their levels simply reached a point where the presence of DegP became mandatory. It is conceivable that in a DegP background, the accumulated mutant polypeptides acted as a "sink" for folding factors whose depletion resulted in disarrayed protein assembly and ultimately cell death. The results presented here thus provided genetic evidence for the crucial role of DegP in the event of aberrant OMP assembly.

Role of protease-deficient DegP in mutant OmpF assembly. Elevated expression of a variant DegP_S210A protein which is deficient in protease activity reversed the lethal effects of several OmpF mutants in a degP null background. This rescue from toxicity was not accomplished by promoting degradation of mutant proteins or correcting their assembly. Rather, it appeared that a significant population of mutant polypeptides was sequestered from the normal assembly pathway. While the level of correctly assembled (detergent-insoluble, thermostable, and outer-membrane-localized) molecules did not increase, a detergent-soluble and highly thermolabile species accumulated when DegP_S210A was overproduced. Curiously, this overproduction of DegP_S210A also preferentially promoted the localization of mutant polypeptides to the inner membrane, where DegP_S210A itself is localized. This raises the possibility that the inner-membrane localization of mutant polypeptides occurs through their binding to DegP_S210A. Recent electron microscopic and cross-linking studies showed that DegP consists of a hexameric double-layered ring structure with a central cavity that has a higher affinity for unfolded than for folded polypeptides (12). In light of this, it is tempting to suggest that misfolded OmpF polypeptides are captured within the double-ring structure of DegP_S210A. This "molecular capturing" of misfolded polypeptides would clear the assembly pathway for OMP molecules that fold correctly and assemble into detergent-insoluble thermostable forms that are localized to the outer membrane.