Selective pressure for rapid membrane integration constrains the sequence of bacterial outer membrane proteins

Abstract

Almost all bacterial outer membrane proteins (OMPs) contain a β barrel domain that serves as a membrane anchor, but the assembly and quality control of these proteins are poorly understood. Here, we show that the introduction of a single lipid-facing arginine residue near the middle of the β barrel of the Escherichia coli OMPs OmpLA and EspP creates an energy barrier that impedes membrane insertion. Although several unintegrated OmpLA mutants remained insertion-competent, they were slowly degraded by the periplasmic protease DegP. Two EspP mutants were also gradually degraded by DegP but were toxic because they first bound to the Bam complex, an essential heteroligomer that catalyzes the membrane insertion of OMPs. Interestingly, another EspP mutant likewise formed a prolonged, deleterious interaction with the Bam complex but was protected from degradation and eventually inserted into the membrane in a native conformation. The different types of interactions between the EspP mutants and the Bam complex that we observed may correspond to distinct stages in OMP assembly. Our results show that sequences that significantly delay assembly are disfavored not only because unintegrated OMPs are subjected to degradation, but also because OMPs that assemble slowly can form dominant-negative interactions with the Bam complex.

Fig. 1

Location of arginine mutations in OmpLA and EspP. A.–B. The crystal structures of OmpLA (PDB:1QD5; Snijder et al., 1999) and the EspP β barrel (PDB: 2QOM; Barnard et al., 2007) and the position of the arginine substitutions that were analyzed in this study are shown. Red: arginine substitutions that blocked insertion of the protein into the OM in wild-type cells; yellow: arginine substitutions that delayed membrane insertion; green: arginine substitutions that did not affect protein assembly.

Fig. 2

Introduction of lipid-facing arginine residues near the middle of the OmpLA β barrel destabilizes the protein in wild-type E. coli. A. MC4100 transformed with pMDG2 (P_trc-ompLA) or a pMDG2 derivative encoding the indicated OmpLA mutant were subjected to pulse-chase labeling after the addition of IPTG. Immunoprecipitations were then conducted using an anti-OmpLA antiserum. B. The percentage of the pulse-labeled protein that remained at each time point is shown.

Fig. 3

OmpLA arginine variants are stable and gradually acquire PK resistance in the absence of DegP. JMR352 transformed with pMDG2 or a pMDG2 derivative encoding the indicated OmpLA mutant were subjected to pulse-chase labeling after the addition of IPTG. The OM was permeabilized and half of each sample was treated with PK. Immunoprecipitations were then conducted using an anti-OmpLA antiserum. The percentage of the protein that remained and that was resistant to PK digestion at each time point is shown. Because it was not possible to distinguish proOmpLA from mature OmpLA in some of the pulse-labeled samples, the protein observed in untreated cells after a 2 min chase was defined as 100%.

Fig. 4

The PK resistance of OmpLA arginine mutants in a degP- strain is due to membrane integration. JMR352 transformed with pMDG2 or a pMDG2 derivative encoding the indicated OmpLA mutant were pulse labeled and subjected to a 20 min chase after the addition of IPTG. In (A) the OM was then permeabilized and equal portions of each sample were left untreated, treated with PK or treated with PK after the addition of DDM. In (B) cells were disrupted by sonication. After a portion of the total cell extract (T) was set aside, membranes (Memb) were separated from soluble proteins (Sol) by centrifugation. Equal portions of the membrane fractions were then left untreated, treated with PK or treated with PK after the addition of DDM. Immunoprecipitations were conducted using an anti-OmpLA antiserum.

Fig. 5

OmpLA arginine mutants do not acquire PK resistance in wild-type E. coli at low temperature. MC4100 (A) or JMR352 (B) transformed with pMDG2 or a pMDG2 derivative encoding the indicated OmpLA mutant were shifted to 28° C and subjected to pulse-chase labeling after the addition of IPTG. The OM was permeabilized and half of each sample was treated with PK. Immunoprecipitations were then conducted using an anti-OmpLA antiserum.

Fig. 6

The introduction of a lipid-facing arginine residue near the extracellular or periplasmic side of the EspP β barrel does not significantly affect assembly. AD202 transformed with pRLS5 (P_trc-espP) or a pRLS5 derivative encoding the indicated EspP mutant were subjected to pulse-chase labeling after the addition of IPTG. Immunoprecipitations were then conducted using an antiserum generated against an EspP C-terminal peptide. The percentage of the passenger domain that was released from the β domain by proteolytic cleavage at each time point is shown.

Fig. 7

The introduction of a lipid facing arginine residue near the middle of the EspP β barrel delays or blocks membrane integration. A. AD202 transformed with pRLS5 (P_trc-espP) or a pRLS5 derivative encoding the indicated EspP mutant were subjected to pulse-chase labeling after the addition of IPTG. Half of the cells were treated with PK, and immunoprecipitations were conducted using an antiserum generated against an EspP C-terminal peptide. The percentage of the passenger domain that was released from the β domain by proteolytic cleavage or surface exposed at each time point is plotted in (B) and (C). The percentage of the radiolabeled protein that remained at each time point is shown in (D). E. Cell membranes isolated from AD202 that produced wild-type EspP or the G1123R mutant were heated at the indicated temperature in SDS-PAGE sample buffer and the free β barrel domain was detected by Western blot.

Fig. 8

EspP mutants that contain a lipid facing arginine interact stably with the Bam complex. A and B. AD202 transformed with pDULE-Bpa and a derivative of pRI22 encoding wild-type EspP or EspP (G1123R), EspP (L1121R) or EspP (I1119R) with an amber mutation at position 1214 were subjected to pulse-chase labeling after the addition of IPTG. One aliquot of cells was UV-irradiated while an equal aliquot was left untreated. In part (A), PK was added to half of each sample. Immunoprecipitations were then conducted using the indicated antisera. The non-irradiated samples from part (A) are shown in Fig. S9.

Fig. 9

EspP mutants that contain a lipid facing arginine near the middle of the β barrel inhibit Bam complex function. A and B. DPR959 transformed with pRLS5 or a pRLS5 derivative encoding the indicated EspP mutant were incubated at 37° C on LB agar plates containing ampicillin (100 μg/ml) and either no IPTG or 100 μM IPTG. C. AD202 transformed with pSCrhaB2 (vector), pJH207 (P_rhaB2-espPΔ5) or a pJH207 derivative encoding the indicated EspPΔ5 mutant were inoculated into LB containing trimethoprim (50 μg/ml) at OD₅₅₀=0.0001. The growth of each culture at 37° C was monitored at OD₅₅₀. D. AD202 transformed with pRLS5 and a derivative of pJH207 encoding wild-type EspPΔ5 or the indicated EspPΔ5 mutant were subject to pulse-chase labeling after the addition rhamnose and IPTG. Half of the cells were treated with PK, and immunoprecipitations were conducted using an antiserum generated against an EspP N-terminal peptide.