The targeting pathway of Escherichia coli presecretory and integral membrane proteins is specified by the hydrophobicity of the targeting signal

Abstract

Previous studies have demonstrated that presecretory proteins such as maltose binding protein (MBP) and outer membrane protein A (OmpA) are targeted to the Escherichia coli inner membrane by the molecular chaperone SecB, but that integral membrane proteins are targeted by the signal recognition particle (SRP). In vitro studies have suggested that trigger factor binds to a sequence near the N terminus of the mature region of OmpA and shunts the protein into the SecB pathway by blocking an interaction between SRP and the signal peptide. By contrast, we have found that the targeting pathway of a protein under physiological conditions is dictated by the composition of its targeting signal. Replacement of the MBP or OmpA signal peptide with the first transmembrane segment of AcrB abolished the dependence on SecB for transport and rerouted both proteins into the SRP targeting pathway. More modest alterations of the MBP signal peptide that simply increase its hydrophobicity also promoted SRP binding. Furthermore, we obtained evidence that SRP has a low affinity for typical signal peptides in vivo. These results imply that different classes of E. coli proteins are targeted by distinct pathways because bacterial SRP binds to a more restricted range of targeting signals than its eukaryotic counterpart.

Figure 1

PreMBP and ProOmpA variants used in this study. The sequence of the N terminus of preMBP and proOmpA and the leader peptidase cleavage site (arrow) is shown. The point mutations, substitutions, and insertions in each variant are underlined. TM-MBP and TM-OmpA do not have a cleaved signal peptide. The proOmpA analyzed herein contains a proline in place of the naturally occurring glutamine at position −2 (italics). This change was introduced to create a restriction site that could be used to make the proOmpA variants.

Figure 2

Protease-sensitivity of TM-MBP and TM-OmpA. TST1 transformed with pJH28, pJH29, pHL11, or pHL12 were pulse-labeled, converted to spheroplasts, and treated with proteinase K. MBP- and OmpA-containing polypeptides were immunoprecipitated with anti-MBP and anti-HA antibodies, respectively. Lanes: 1 and 2, cells producing MBP or OmpA; 3 and 4, cells producing TM-MBP or TM-OmpA. Proteinase K was added to the samples in lanes 1 and 3.

Figure 3

Replacement of the MBP and OmpA signal peptides with a TMS reroutes the proteins into the SRP targeting pathway. (A) HDB51 (P_BAD-ffh secB⁺) and HDB52 (P_BAD-ffh secB⁻) transformed with pJH28, pJH29, pHL11, or pHL12 were grown in M9 medium containing arabinose (lanes 1–4) or glucose (lanes 5–8), radiolabeled, converted to spheroplasts, and treated with proteinase K. MBP- and OmpA-containing polypeptides and CAT were immunoprecipitated. Lanes: 1, 2, 5, and 6, HDB51 cells; 3, 4, 7, and 8, HDB52 cells. Proteinase K was added to the samples shown in lanes 1, 3, 5, and 7. (B) HDB50 and HDB51 transformed with pJH29 were grown in M9 medium containing glucose. The experiment shown in A was repeated except that 1% Triton X-100 was added to the samples in lanes 1 and 4 immediately before proteinase K treatment. Lanes: 1–3, HDB51 cells; 4–6, HDB52 cells. Proteinase K was added to the samples in lanes 1, 2, 4, and 5.

Figure 4

Increasing the distance between the signal peptide and the mature region of proOmpA does not change its targeting pathway. TST1 and HDB50 (secB⁻) transformed with pHL11, pHL13, or pHL14 were radiolabeled, and OmpA-containing polypeptides were immunoprecipitated with anti-HA antibodies. The precursor encoded on each plasmid is indicated at the top. Lanes: 1, 3, and 5, TST1 cells; 2, 4, and 6, HDB50 cells.

Figure 5

Increasing the hydrophobicity of the MBP signal peptide reroutes the protein into the SRP pathway. HDB51 and HDB52 transformed with a plasmid encoding the indicated MBP* mutant were grown in M9 medium containing arabinose (lanes 3 and 4) or glucose (lanes 5 and 6) and radiolabeled. TST1 and HDB50 transformed with pJH28 were grown in M9 medium containing glucose and radiolabeled to provide molecular weight markers (lanes 1 and 2). MBP-containing polypeptides were immunoprecipitated and resolved on a 10% NuPage gel with a standard Mops buffer (NOVEX). Lanes: 3 and 5, HDB51 cells; 4 and 6, HDB52 cells.

Figure 6

MBP remains SecB-dependent in cells that overproduce SRP and FtsY. (A) CK1953 (secB⁻) transformed with the indicated plasmid were pulse-labeled and endogenous MBP-containing polypeptides were immunoprecipitated. PreMBP and MBP markers were generated as described in Fig. 5. (B) cK1953 transformed with the indicated plasmid were grown as in A but not radiolabeled. Ffh and FtsY were detected by Western blot.