Targeting, Insertion, and Localization of Escherichia coli YidC
2002; Elsevier BV; Volume: 277; Issue: 15 Linguagem: Inglês
10.1074/jbc.m200311200
ISSN1083-351X
AutoresMalene L. Urbanus, Linda Fröderberg, David Drew, Petra Björk, Jan‐Willem L. de Gier, Josef Brunner, Bauke Oudega, Joen Luirink,
Tópico(s)ATP Synthase and ATPases Research
ResumoYidC was recently shown to play an important role in the assembly of inner membrane proteins (IMPs) both in conjunction with and separate from the Sec-translocon. Little is known about the biogenesis and structural and functional properties of YidC, itself a polytopic IMP. Here we analyze the targeting and membrane integration of YidC using in vivo and in vitro approaches. The combined data indicate that YidC is targeted by the signal recognition particle and inserts at the SecAYEG-YidC translocon early during biogenesis, unlike its mitochondrial homologue Oxa1p. In addition, YidC is shown to be relatively abundant compared with other components involved in IMP assembly and is predominantly localized at the poles of the cell. YidC was recently shown to play an important role in the assembly of inner membrane proteins (IMPs) both in conjunction with and separate from the Sec-translocon. Little is known about the biogenesis and structural and functional properties of YidC, itself a polytopic IMP. Here we analyze the targeting and membrane integration of YidC using in vivo and in vitro approaches. The combined data indicate that YidC is targeted by the signal recognition particle and inserts at the SecAYEG-YidC translocon early during biogenesis, unlike its mitochondrial homologue Oxa1p. In addition, YidC is shown to be relatively abundant compared with other components involved in IMP assembly and is predominantly localized at the poles of the cell. inner membrane protein signal recognition particle fifty-four homologue transmembrane segments green fluorescent protein Trigger Factor Luria Bertani isopropyl-1-thio-β-d-galactopyranoside inverted inner membrane vesicles In Escherichia coli, inner membrane protein (IMP)1 integration can occur via Sec-dependent or Sec-independent mechanisms (1.de Gier J.-W. Luirink J. Mol. Microbiol. 2001; 40: 314-322Crossref PubMed Scopus (82) Google Scholar). The majority of IMPs are targeted to the membrane by the signal recognition particle (SRP) and its receptor FtsY that mediates co-translational targeting to the Sec-translocon (2.Herskovits A.A. Bochkareva E.S. Bibi E. Mol. Microbiol. 2000; 38: 927-939Crossref PubMed Scopus (97) Google Scholar). The SRP, which consists of the 4.5 S RNA and Ffh (for fifty-four homologue) is homologous to the eukaryotic SRP but less complex in composition. The core Sec-translocon consists of the integral membrane components SecY, SecE and SecG, which form a heterotrimer, and the peripheral subunit SecA (3.Manting E.H. Driessen A.J.M. Mol. Microbiol. 2000; 37: 226-238Crossref PubMed Scopus (210) Google Scholar). The translocon serves as a translocating pore for both secretory proteins and IMPs. SecA is an ATPase that functions as a molecular motor that drives the translocation of secretory proteins and large periplasmic domains of IMPs through the SecYEG pore. Recent evidence shows that a novel component, YidC, is specifically involved in Sec-dependent IMP integration (4.Luirink J. Samuelsson T. de Gier J.-W. FEBS Lett. 2001; 1: 1-5Crossref Scopus (122) Google Scholar). Using a site-specific photocross-linking procedure, YidC was shown to interact with the nascent IMPs, FtsQ, Lep, and mannitol permease, as they move laterally from the Sec-translocon into the lipid bilayer (5.Beck K. Eisner G. Trescher D. Dalbey R.E. Brunner J. Müller M. EMBO Rep. 2001; 2: 709-714Crossref PubMed Scopus (138) Google Scholar, 6.Houben E.N.G. Scotti P.A. Valent Q.A. Brunner J. de Gier J.-W. Oudega B. Luirink J. FEBS Lett. 2000; 476: 229-233Crossref PubMed Scopus (73) Google Scholar, 7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar, 8.Urbanus M.L. Scotti P.A. Fröderberg L. Sääf A. de Gier J.-W. Brunner J. Samuelson J.C. Dalbey R.E. Oudega B. Luirink J. EMBO Rep. 2001; 6: 524-529Crossref Scopus (154) Google Scholar). This interaction appeared to be specific for the transmembrane segments (TMs) in the nascent polypeptide. Moreover, YidC could be co-purified with the Sec-translocon suggesting a physical connection (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar). Upon depletion of YidC, the assembly of Sec-dependent IMPs such as Lep and FtsQ is hampered, although the effect is relatively mild (8.Urbanus M.L. Scotti P.A. Fröderberg L. Sääf A. de Gier J.-W. Brunner J. Samuelson J.C. Dalbey R.E. Oudega B. Luirink J. EMBO Rep. 2001; 6: 524-529Crossref Scopus (154) Google Scholar,9.Samuelson J.C. Chen M.Y. Jiang F.L. Möller I. Wiedmann M. Kuhn A. Phillips G.J. Dalbey R.E. Nature. 2000; 406: 637-641Crossref PubMed Scopus (423) Google Scholar). In contrast to most IMPs, some small phage coat proteins (like Pf3 coat and M13 procoat proteins) insert into the inner membrane independent of the Sec-translocon (10.Dalbey R.E. Kuhn A. Ann. Rev. Cell Dev. Biol. 2000; 16: 51-87Crossref PubMed Scopus (84) Google Scholar). These proteins were considered to partition spontaneously into the lipid bilayer (i.e. without the requirement of any proteinaceous factors and only depending on the proton motive force to energize the process). However, it was recently demonstrated that membrane assembly of M13 coat protein is almost completely blocked upon depletion of YidC suggesting a crucial role for YidC in the integration of Sec-independent proteins (9.Samuelson J.C. Chen M.Y. Jiang F.L. Möller I. Wiedmann M. Kuhn A. Phillips G.J. Dalbey R.E. Nature. 2000; 406: 637-641Crossref PubMed Scopus (423) Google Scholar, 11.Samuelson J.C. Jiang F. Yi L. Chen M. de Gier J.-W. Kuhn A. Dalbey R.E. J. Biol. Chem. 2001; 16: 16Google Scholar). YidC is homologous to the mitochondrial IMP Oxa1p and to the thylakoid membrane protein Alb3, which both have been implicated in membrane protein integration (4.Luirink J. Samuelsson T. de Gier J.-W. FEBS Lett. 2001; 1: 1-5Crossref Scopus (122) Google Scholar). Oxa1p is essential for the correct insertion of a subset of both mitochondrial-encoded IMPs (like pCoxII) and nuclear-encoded IMPs (like Oxa1p itself). Interestingly, mitochondria do not have an SRP-like targeting pathway or a Sec-like translocon suggesting that Oxa1p might function in a fashion similar to YidC in the Sec-independent route. YidC is a polytopic IMP that spans the membrane six times with an N-in, C-in topology and a large, poorly conserved periplasmic domain between TM1 (signal anchor sequence) and TM2 (Fig. 2A) (12.Sääf A. Monne M. de Gier J.-W. von Heijne G. J. Biol. Chem. 1998; 273: 30415-30418Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). As is true for most polytopic IMPs, hardly anything is known about the targeting, assembly, localization, and cellular abundance of YidC. Here, we present evidence that YidC is targeted co-translationally by the SRP to the Sec-translocon, which appears to be required for proper assembly. Nascent YidC is shown to contact SecA, SecY, and pre-existing YidC very early during biosynthesis. YidC is found to be present in excess over SecYEG. Finally, using GFP fusion technology we found that YidC accumulates at the poles of the cell. Restriction enzymes, Expand long template PCR system, and Lumi-LightPLUS Western blotting substrate were from Roche Molecular Biochemicals. Megashort script T7 transcription kit was from Ambion Inc. [35S]methionine and protein A-Sepharose were from Amersham Biosciences. T4 Ligase and T4 DNA polymerase were from Epicentre Technologies. All other chemicals were supplied by Sigma. Antiserum against a C-terminal peptide of YidC has been described previously (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar). In addition, an antiserum against purified hisYidC (see below) was raised in rabbit by Agrisera (Umeå, Sweden). The antisera directed against Trigger Factor (TF) and SecA were gifts from W. Wickner, and anti-SecY was a gift from A. Driessen. E. coli strain MRE600 was used to prepare a lysate for translation of in vitrosynthesized mRNA and suppression of UAG stop codons in the presence of (Tmd)Phe-tRNAsup. Strain MC4100 grown in Luria Bertani (LB) medium was used to isolate inverted membrane vesicles (IMVs) and as an expression host for plasmid pEH1hisYidC (overexpression of YidC carrying a hexahistidine tag)(13). The E. coli strain JS7131 in which yidC is under the control of the araBADoperator/promoter (9.Samuelson J.C. Chen M.Y. Jiang F.L. Möller I. Wiedmann M. Kuhn A. Phillips G.J. Dalbey R.E. Nature. 2000; 406: 637-641Crossref PubMed Scopus (423) Google Scholar) was used to study the complementation of YidC depletion by expression of YidC-GFP. The 4.5 S RNA conditional strain FF283 and the SecE depletion strain CM124 were cultured and depleted as described previously (14.Ribes V. Römisch K. Giner A. Dobberstein B. Tollervey D. Cell. 1990; 63: 591-600Abstract Full Text PDF PubMed Scopus (163) Google Scholar, 15.Traxler B. Murphy C. J. Biol. Chem. 1996; 271: 12394-12400Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar, 16.de Gier J.-W. Scotti P.A. Sääf A. Valent Q.A. Kuhn A. Luirink J. von Heijne G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14646-14651Crossref PubMed Scopus (112) Google Scholar). The conditional SecA strain BA13 and its control strain DO251 were cultured and depleted for SecA essentially as described (17.Oliver D.B. Beckwith J. Cell. 1982; 30: 311-319Abstract Full Text PDF PubMed Scopus (214) Google Scholar, 18.Qi H.Y. Bernstein H.D. J. Biol. Chem. 1999; 274: 8993-8997Abstract Full Text Full Text PDF PubMed Scopus (70) Google Scholar). The SecG deletion strain KN370 (19.Nishiyama K. Hanada M. Tokuda H. EMBO J. 1994; 13: 3272-3277Crossref PubMed Scopus (157) Google Scholar) was cultured in M9 minimal medium to the mid-exponential growth phase as described previously (16.de Gier J.-W. Scotti P.A. Sääf A. Valent Q.A. Kuhn A. Luirink J. von Heijne G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14646-14651Crossref PubMed Scopus (112) Google Scholar). Plasmid p H+ (20.Bost S. Belin D. J. Biol. Chem. 1997; 272: 4087-4093Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar) was used to complement the SecG deletion strain KN370 with SecG. Plasmid pEH1YidC, harboring yidC under control of a lac promoter, was constructed in a 3-step procedure as follows. First, the 5′-region of yidC was cloned NcoI-KpnI from pAra14-YidC (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar) in the pEH1 vector. Second, using pAra14-YidC as a template a SmaI site was introduced downstream of yidC by PCR. Subsequently, the 3′ yidC fragment was cloned KpnI-SmaI in pEH1YidC NcoI-KpnI to generate full-length yidC. For YidC-GFP expression, YidC was fused at its C terminus to the N terminus of a GFP variant that has been selected to fold well in E. coli (21.Drew D.E. von Heijne G. Nordlund P. de Gier J.-W. FEBS Lett. 2001; 2: 220-224Crossref Scopus (190) Google Scholar). The YidC-GFP fusion was cut out with XbaI-HindIII from plasmid pWaldoYidC-GFP and cloned XbaI-SmaI (the HindIII site of the DNA fragment harboring the YidC-GFP fusion was made blunt using T4 DNA polymerase) into the pEH1 expression vector yielding pEH1YidC-GFP. Plasmids pC4meth52YidCTAG17 and pC4Meth96YidCTAG17 were obtained by PCR using pEH1YidC as a template. They encode truncated YidC fused to a C-terminal 4× methionine tag to improve the labeling efficiency and contain an amber mutation at position 17 to enable sup-tRNA photocross-linking (see below). The nucleotide sequences of all cloned mutant genes were confirmed by DNA sequencing. Truncated mRNA was prepared as described previously (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar) from HindIII-linearized pC4Meth52YidCTAG17 or pCMeth96YidCTAG17. For photocross-linking, (Tmd)Phe was site-specifically incorporated into YidC nascent chains by suppression of a UAG stop codon using (Tmd)Phe-tRNAsup in an E. coli in vitro translation system containing [35S]methionine to label the nascent chains. This procedure has been described previously (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar, 8.Urbanus M.L. Scotti P.A. Fröderberg L. Sääf A. de Gier J.-W. Brunner J. Samuelson J.C. Dalbey R.E. Oudega B. Luirink J. EMBO Rep. 2001; 6: 524-529Crossref Scopus (154) Google Scholar). Targeting to IMVs, photocross-linking and carbonate extraction (to separate soluble and pheriperal membrane proteins from integral membrane proteins) were carried out as described previously (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar). Carbonate soluble and insoluble fractions were either trichloracetic acid-precipitated or immunoprecipitated. The material used for immunoprecipitation was 4-fold the amount used for trichloracetic acid precipitation. Sample analysis (SDS-PAGE and phosphorimaging) was as described previously (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar). Strains FF285, BA13, DO251, CM124, and KN370 were grown to the mid-exponential phase of growth. Expression of YidC was induced for 3 min with either 1 mm IPTG (from pEH1YidC in strain CM124) or 0.2%l-arabinose (from pAra14-YidC in the strains FF283, BA13, DO251, and KN370). Aliquots were pulse-labeled for 30 s with 150 μCi of [35S]methionine and converted to spheroplasts as described previously (16.de Gier J.-W. Scotti P.A. Sääf A. Valent Q.A. Kuhn A. Luirink J. von Heijne G. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 14646-14651Crossref PubMed Scopus (112) Google Scholar). Aliquots of the spheroplast suspensions were incubated on ice for 1 h in the presence or absence of proteinase K (0.3 mg/ml). Subsequently, phenylmethylsulfonyl fluoride was added to the spheroplast suspensions (0.33 mg/ml) to inhibit the protease. Finally, the samples were precipitated in 10% trichloracetic acid and immunoprecipitated with antisera against YidC, OmpA, or AraB/BandX and analyzed using SDS-PAGE. The cellular abundance of YidC was determined essentially as described (22.van Klompenburg W. Whitley P. Diemel R. von Heijne G. de Kruijff B. Mol. Membr. Biol. 1995; 12: 349-353Crossref PubMed Scopus (15) Google Scholar). An overnight culture of MC4100 was diluted 1:100 in LB medium and grown at 37 °C to the mid-exponential phase of growth. Cells were harvested and diluted for life cell counts by plating on LB-agar. From the remainder of the samples a range of 0.05–0.0025 OD660units of cells was analyzed using SDS-PAGE and Western blotting next to a 1–20-ng standard of purified hisYidC. The blots were incubated with antibodies directed against hisYidC and developed using Lumi-LightPLUS Western blotting substrate. The chemiluminescent signal was detected using the Fluor-S MultiImager (Bio-Rad) and quantified using Multi-Analyst 1.0.2 software (Bio-Rad). HisYidC was purified from Top10F′ containing pEH1hisYidC as described previously (13.van der Laan M. Houben E.N.G. Nouwen N. Luirink J. Driessen A.J.M. EMBO Rep. 2001; 21: 519-523Crossref Scopus (95) Google Scholar). Protein concentration was determined using the DC protein assay (Bio-Rad) and bovine serum albumin as a standard. YidC-GFP was expressed in the conditional YidC depletion strain JS7131. Cells were grown for 6 h at 37 °C in LB medium supplemented with 0.2%l-arabinose (to generate non-depleted control cells) or 10 μm IPTG (to deplete for YidC and express YidC-GFP). Cells were then harvested and used for analysis of YidC/YidC-GFP expression and fluorescence microscopy. To monitor YidC and YidC-GFP expression, cells were washed with M9 minimal medium and resuspended in M9 minimal medium supplemented with 0.2% l-arabinose or 10 mm IPTG. Subsequently, the cells were labeled for 1 h with [35S]methionine (150 μCi/ml), trichloracetic acid-precipitated, washed with cold acetone, and resuspended in 10 mm Tris-HCl pH 8.0/2% SDS. YidC and YidC-GFP were immunoprecipitated with anti-YidC serum and analyzed by SDS-PAGE. For localization of YidC-GFP, cells expressing YidC-GFP were harvested and resuspended in 50 mm Tris-HCl pH 8.0, 200 mmNaCl, 15 mm EDTA. GFP fluorescence was viewed with an Axioplan 2 microscope (Zeiss) equipped with a filter set for fluorescein isothiocyanate. Images were captured using a CCD camera (Color Cool View, Photonic Sciences) and Image Pro-Plus software. We analyzed the pathway of targeting and membrane insertion of nascent YidC (a polytopic IMP) using an in vitrotranslation/targeting/photocross-linking assay that we initially developed for the analysis of molecular interactions during biosynthesis of the less complex IMPs, FtsQ and Lep (6.Houben E.N.G. Scotti P.A. Valent Q.A. Brunner J. de Gier J.-W. Oudega B. Luirink J. FEBS Lett. 2000; 476: 229-233Crossref PubMed Scopus (73) Google Scholar, 7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar). Radiolabeled nascent chains of YidC of 52 and 96 amino acids were synthesized by translation of truncated mRNA in a membrane-free E. coli extract in the presence of [35S]methionine. Purified inverted IMVs were added from the start of the translation reaction to allow co-translational membrane targeting and insertion of the translation intermediates. To specifically probe the molecular environment of the first TM, a stop codon (TAG) was introduced at position 17 in the center of the TM region and suppressed during in vitro synthesis by the addition of (Tmd)Phe-tRNAsup that carries a photoreactive probe. After the translation/insertion reaction, one-half of each sample was irradiated with UV light to induce cross-linking whereas the other half was kept in the dark to serve as a control. The samples were extracted with carbonate to separate untargeted material from the membrane-integrated material, and cross-linking partners were identified by immunoprecipitation. In both constructs, the TAG17 mutation was efficiently suppressed by (Tmd)Phe-tRNAsup (data not shown), resulting in nascent YidC of the expected apparent molecular weight. Assuming that the ribosome covers ∼35 amino acids, the 52-mer only partially exposes the TM outside the ribosome, and the photocross-linking group is just exposed outside the ribosome (Fig. 1A). Using this construct, a major cross-linking adduct of ∼60 kDa was observed in the untargeted (carbonate soluble) nascent chains (Fig. 1B, lanes 1 and 2). Immunoprecipitation identified TF as the cross-linking partner (Fig. 1B, lane 3). Furthermore, two less prominent cross-linking adducts of ∼55 and ∼100 kDa were identified as Ffh, the protein component of the SRP, and SecA (Fig. 1B, lanes 4 and 5). Anti-Ffh also precipitated smaller adducts that might represent degradation products. The 52-mer was less efficiently integrated in the membrane than the longer YidC construct as judged by the criterion of carbonate resistance (∼25% versus∼45%, not shown). Membrane-integrated 52-mer gave rise to distinct cross-linking adducts of ∼40, ∼60, and ∼100 kDa (Fig. 1B, lanes 6 and 7), which were immunoprecipitated with anti-SecY, anti-YidC, and anti-SecA, respectively (Fig. 1B, lanes 8–10). For identification of nascent YidC-YidC adducts, antiserum raised against the C terminus of YidC was used to distinguish full-length YidC from nascent YidC. A fourth cross-linking adduct of ∼55 kDa was not immunoprecipitated with any of the antisera used and remains to be identified. In the longer 96-mer, the TM and ∼38 residues of the first periplasmic domain are exposed outside the ribosome (Fig. 1A). Untargeted nascent chains still cross-linked to TF, giving rise to several distinct adducts possibly caused by the binding of TF at several positions in this longer nascent chain (23.Valent Q.A. de Gier J.-W. van Heijne G. Kendall D.A. ten Hagen-Jongman C.M. Oudega B. Luirink J. Mol. Microbiol. 1997; 25: 53-64Crossref PubMed Scopus (149) Google Scholar). Ffh and SecA were cross-linked but to a lesser extent than with the shorter nascent chains (Fig. 1C, lanes 1–5). The membrane-integrated nascent chains showed predominantly cross-linking to YidC and SecA. SecY was also cross-linked, albeit much less efficiently than with the 52-mer (Fig. 1C, lanes 6–10). Taken together, these data show that the first TM in short nascent YidC interacts with TF and SRP and inserts into the inner membrane in a carbonate resistant configuration in the vicinity of pre-existing YidC, SecY, and SecA. The difference in cross-linking efficiencies of the 52-mer and 96-mer to SecY suggests that the first TM of YidC is close to SecY only early during the membrane insertion process. We have shown that nascent YidC interacts with Ffh, TF, SecA, SecY, and YidC. To investigate whether these interactions reflect an in vivo dependence on the SRP targeting pathway or Sec-translocon for correct membrane assembly of the full-length protein, we monitored the requirements for in vivo assembly of YidC in a proteinase K accessibility assay. When YidC is correctly targeted and assembled in the inner membrane, the large periplasmic loop between TM1 and TM2 (Fig. 2A) is proteinase K-resistant except for a small region near TM1 (Fig. 2A) (12.Sääf A. Monne M. de Gier J.-W. von Heijne G. J. Biol. Chem. 1998; 273: 30415-30418Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). Consequently, proteinase K treatment of spheroplasts derived from wild-type cells results in a small band shift of YidC in SDS-PAGE (Fig. 2B, lanes 1 and 2). OmpA and band X are periplasmic and cytoplasmic control proteins, respectively, used to monitor spheroplast formation (24.de Gier J.-W. Mansournia P. Valent Q.A. Phillips G.J. Luirink J. von Heijne G. FEBS Lett. 1996; 399: 307-309Crossref PubMed Scopus (135) Google Scholar). In addition, proOmpA processing was monitored to check SecA and SecE depletion. Depletion of 4.5 S RNA, the RNA component of SRP, affected the assembly of YidC into the inner membrane as is evident from the appearance of full-length YidC in proteinase K-treated spheroplasts (Fig. 2B). Depletion of SecA had a similar effect whereas depletion of SecE had an even stronger effect (Fig. 2, C and D). The SecE depletion strain was used because it enables the most efficient inactivation of the SecYEG core translocon thus far (15.Traxler B. Murphy C. J. Biol. Chem. 1996; 271: 12394-12400Abstract Full Text Full Text PDF PubMed Scopus (90) Google Scholar). Upon depletion of SecE, SecY is rapidly degraded by the protease FtsH (25.Kihara A. Akiyama Y. Ito K. Proc. Natl. Acad. Sci. U. S. A. 1995; 92: 4532-4536Crossref PubMed Scopus (210) Google Scholar). We also monitored the assembly of YidC in a SecG deletion background (Fig. 2E). The Sec-translocon component SecG also seems to be required for the efficient assembly of YidC. Unfortunately, we could not monitor the assembly of YidC in a YidC depletion background because the YidC background expression levels from the plasmid-borne copy of YidC were sufficient to alleviate chromosomal YidC depletion. Taken together, these data confirm the involvement of the SRP, SecA, and SecYEG in the proper targeting and assembly of YidC into the inner membrane. To determine the abundance of YidC in E. coli cells a semiquantitative Western blot procedure was employed. MC4100 was grown to the mid-exponential phase of growth and harvested for life cell counts and analysis by SDS-PAGE and Western blotting. A range of 0.05–0.0025 OD660 units of cells was analyzed (Fig. 3). To calculate the amount of YidC from Western blot, the chemiluminescent signal of a standard of purified hisYidC (1–20 ng) was quantified and fitted. A linear relationship between the amount of hisYidC and the chemiluminscent signal was observed throughout this range. The amount of YidC in the cell samples was calculated and correlated to the life cell counts. A MC4100 cell contains ∼0.27 fg of YidC corresponding to 2700 copies of YidC per cell. A duplicate experiment resulted in a similar amount of 2500 copies of YidC per cell (data not shown). Where the inner membrane assembly of IMPs occurs has not yet been studied. Because YidC appears to be exclusively involved in IMP insertion and assembly (9.Samuelson J.C. Chen M.Y. Jiang F.L. Möller I. Wiedmann M. Kuhn A. Phillips G.J. Dalbey R.E. Nature. 2000; 406: 637-641Crossref PubMed Scopus (423) Google Scholar), we have studied the spatial localization of YidC in the inner membrane using GFP-tagging technology. GFP was fused to the C terminus of YidC, and cloned into the IPTG inducible expression vector pEH1. The YidC depletion strain JS7131, in which yidC is under the control of the araBAD operator/promoter, was transformed with pEH1 (empty control vector) and pEH1YidC-GFP. To test if the fusion construct was functional, JS7131 (pEH1YidC-GFP) and its control strain JS7131 (pEH1) were plated onto LB-agar plates containing 0.2% l-arabinose (to generate non-depleted control cells) or 0.2% glucose (to generate YidC-depleted cells). Because YidC is essential for growth, the control strain JS7131(pEH1) could not grow in the absence of l-arabinose. In contrast, JS7131 (pEH1YidC-GFP) grew in the presence and absence ofl-arabinose without any discernable differences (data not shown). These observations demonstrate that even at uninduced expression levels, leaky expression of the plasmid-encoded YidC-GFP is able to complement depletion of YidC. However, in liquid culture uninduced JS7131 (pEH1YidC-GFP) grows slightly slower than JS7131 (pEH1) grown in the presence of 0.2% l-arabinose. An additional induction with 10 μm IPTG was required for optimal growth (Fig. 4A). The expression of YidC-GFP was monitored by labeling JS7131 (pEH1YidC-GFP) cells grown either in the presence of 0.2% l-arabinose or 10 μm IPTG with [35S]methionine and subsequent immunoprecipitation using antiserum to YidC (Fig. 4B). Cells in which the chromosomal yidC but not the plasmid yidC-GFP copy was expressed contained only a trace amount of YidC-GFP and a normal amount of YidC, whereas in YidC-depleted/YidC-GFP-induced cells only YidC-GFP could be detected at a level comparable with chromosome encoded YidC. Taken together, these data unambiguously show that the YidC-GFP fusion is functional. In addition, cell fractionation studies indicated that both YidC and YidC-GFP are membrane-associated (data not shown). To visualize YidC-GFP in living cells, cells of strain JS7131 (pEH1YidC-GFP) were grown in LB medium in the presence of 10 μm IPTG to express the fusion protein. Cells from different growth stages were inspected by fluorescence microscopy. Surprisingly, at all growth stages fluorescence was concentrated at the poles of the JS7131(pEH1YidC-GFP) cells, indicating that YidC-GFP is predominantly localized at the poles of the cell (Fig. 4C). As an additional control, other polytopic IMPs were fused to GFP and expressed at physiological levels. In contrast to YidC-GFP, fluorescence of all tested IMPs was distributed uniformly over the inner membrane (data not shown), indicating that the localization of YidC-GFP is specific for YidC and not an artifact caused by the fusion to GFP. So far, the biogenesis of E. coli IMPs has been studied using only a very limited set of model IMPs (1.de Gier J.-W. Luirink J. Mol. Microbiol. 2001; 40: 314-322Crossref PubMed Scopus (82) Google Scholar). Among these model IMPs there are hardly any complex polytopic IMPs. Here we analyze the biogenesis of YidC, a polytopic IMP that has recently attracted attention as a factor that might play a key role in the membrane integration and assembly of E. coli IMPs (7.Scotti P.A. Urbanus M.L. Brunner J. de Gier J.W. von Heijne G. van der Does C. Driessen A.J. Oudega B. Luirink J. EMBO J. 2000; 19: 542-549Crossref PubMed Scopus (301) Google Scholar, 9.Samuelson J.C. Chen M.Y. Jiang F.L. Möller I. Wiedmann M. Kuhn A. Phillips G.J. Dalbey R.E. Nature. 2000; 406: 637-641Crossref PubMed Scopus (423) Google Scholar). Using a combined in vitro and in vivo approach we provide evidence that the polytopic IMP YidC follows the SRP/Sec-translocon/YidC pathway for its membrane integration. In addition, we show that YidC is predominantly localized at the cell poles with an abundance of ∼2500–3000 copies per cell. Recently, we have analyzed the pathway of targeting and membrane insertion of the bitopic model IMP FtsQ by studying the sequential interactions of nascent FtsQ in the cytosol and membrane in an in vitro photocross-linking assay. This revealed a sequential interaction of the TM, first with SecY and then with YidC (8.Urbanus M.L. Scotti P.A. Fröderberg L. Sääf A. de Gier J.-W. Brunner J. Samuelson J.C. Dalbey R.E. Oudega B. Luirink J. EMBO Rep. 2001; 6: 524-529Crossref Scopus (154) Google Scholar). The most prominent SecY interaction was found when the TM was only partially exposed outside of the ribosome, whereas the strongest YidC interaction was found when the TM and ∼25–30 residues of the periplasmic domain were exposed outside the ribosome. In this study, two YidC constructs were used (52YidC and 96YidC) that resemble these two FtsQ constructs with respect to the distance between the ribosome and the TM and carry a photocross-linking probe in the exposed TM1 region (Fig. 1A). In contrast to FtsQ, the cross-linking to YidC appears rather complex. The early interaction of nascent YidC with SecY that is almost absent in the longer construct is comparable with the transient interaction of nascent FtsQ with SecY. Different from FbQ are the early interactions with pre-existing YidC and SecA that both persist and even increase in the longer construct. This might indicate that YidC inserts close to SecY/SecA/YidC in a flexible environment that is shielded from water given the sensitivity of the cross-linking probe to quenching by water. The mechanism of membrane insertion of polytopic IMPs may not be generic. The sugar co-transporter melibiose permease has been claimed (26.Bassilana M. Gwizdek C. EMBO J. 1996; 15: 5202-5208Crossref PubMed Scopus (29) Google Scholar) to insert completely independently from the Sec machinery, whereas mannitol permease inserts at SecYE but does not require SecA nor SecG (27.Koch H.-G. Hengelage T. Neumann-Haefelin C. MacFarlane J. Hoffschulte H.K. Schimz K.-L. Mechler B. Müller M. Mol. Biol. Cell. 1999; 10: 2163-2173Crossref PubMed Scopus (129) Google Scholar, 28.Koch H.G. Müller M. J. Cell Biol. 2000; 150: 689-694Crossref PubMed Scopus (57) Google Scholar). It has been suggested that SecA only plays a role in the translocation of large periplasmic loops that are not present in mannitol permease (29.Neumann-Haefelin C. Schäfer U. Müller M. Koch H.G. EMBO J. 2000; 19: 6419-6426Crossref PubMed Scopus (109) Google Scholar). YidC does contain a large periplasmic loop between TM1 and TM2, which might explain its dependence on SecA (Ref.12.Sääf A. Monne M. de Gier J.-W. von Heijne G. J. Biol. Chem. 1998; 273: 30415-30418Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar and this study). Remarkably, SecA is cross-linked to TM1 in 52YidC, which does not expose any periplasmic sequence. It remains to be determined how the future requirement for SecA is sensed in such short nascent chains. Interestingly, the absence of SecG strongly affected YidC assembly, whereas the deletion of SecG only affects protein secretion at low temperatures (19.Nishiyama K. Hanada M. Tokuda H. EMBO J. 1994; 13: 3272-3277Crossref PubMed Scopus (157) Google Scholar), and SecG has not previously been implicated in membrane protein assembly. The effect on YidC assembly may be related to the proposed function of SecG in the modulation of the SecA cycle of membrane insertion and deinsertion (31.Nishiyama K.I. Suzuki T. Tokuda H. Cell. 1996; 85: 71-81Abstract Full Text Full Text PDF PubMed Scopus (187) Google Scholar). Alternatively, SecG may be specifically involved in connecting YidC to the core SecYE translocon thus indirectly influencing YidC assembly. Apparently, YidC inserts via the SRP/Sec-translocon/YidC pathway. In contrast, the mitochondrial YidC homologue Oxa1p assembles solely at Oxa1p itself in the absence of SRP and Sec homologues (32.Hell K. Herrmann J.M. Pratje E. Neupert W. Stuart R.A. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 2250-2255Crossref PubMed Scopus (182) Google Scholar). This may be related to the different topology of Oxa1p, which lacks the first TM of YidC (4.Luirink J. Samuelsson T. de Gier J.-W. FEBS Lett. 2001; 1: 1-5Crossref Scopus (122) Google Scholar). In the cytosolic fraction 52YidC is cross-linked to both SRP (consistent with the in vivo dependence on SRP) and TF. This confirms the very early interaction of TF with nascent chains as observed previously for PhoE (23.Valent Q.A. de Gier J.-W. van Heijne G. Kendall D.A. ten Hagen-Jongman C.M. Oudega B. Luirink J. Mol. Microbiol. 1997; 25: 53-64Crossref PubMed Scopus (149) Google Scholar). It has been suggested that TF specifically interacts with the early mature region of secretory proteins to prevent interaction with the SRP and funnel the proteins into the SecB/SecA targeting route (33.Beck K. Wu L.F. Brunner J. Müller M. EMBO J. 2000; 19: 134-143Crossref PubMed Scopus (146) Google Scholar). Our observations are not consistent with a restricted role for TF in secretion. Apparently, TF is well able to interact with the TM sequence of a nascent IMP. We propose that TF is juxtaposed to all nascent sequences near the ribosome exit site, whereas SRP displaces TF when the exposed region is of sufficient hydrophobicity (Ref. 34.Lee H.C. Bernstein H.D. Proc. Natl. Acad. Sci. U. S. A. 1998; 6: 3471-3476Google Scholar and this study). Surprisingly, the E. coli strain MC4100 contains ∼2500–3000 copies of YidC per cell. The core translocon components SecY and SecE occur with ∼300–600 and 200–400 copies per cell, respectively (30.Driessen A.J.M. J. Membr. Biol. 1994; 142: 145-159Crossref PubMed Scopus (63) Google Scholar). Components of the accessory complex SecDFYajC are even less abundant with ∼30–100 copies per cell (30.Driessen A.J.M. J. Membr. Biol. 1994; 142: 145-159Crossref PubMed Scopus (63) Google Scholar). Assuming that SecYEG tetramers form functional translocation units, a cell contains ∼100–200 functional translocons. Hence, YidC appears to be in excess, arguing against an exclusive association with the SecYEG translocon even when YidC would be dimeric (13.van der Laan M. Houben E.N.G. Nouwen N. Luirink J. Driessen A.J.M. EMBO Rep. 2001; 21: 519-523Crossref Scopus (95) Google Scholar). This may be related to the dual function of YidC: in conjunction with the Sec-translocon and as a separate entity. YidC is a bona fide IMP spanning the membrane six times with short cytoplasmic N and C termini (12.Sääf A. Monne M. de Gier J.-W. von Heijne G. J. Biol. Chem. 1998; 273: 30415-30418Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar). Intriguingly, fusion of the C terminus to GFP did not affect the functioning of YidC and revealed a predominant localization at the poles of the E. coli cell. At present it is too early to speculate whether the cell poles function as dedicated IMP insertion sites perhaps related to the high membrane curvature in this region. We will investigate whether other Sec-translocon components colocalize with YidC at the cell poles using the GFP-tagging and Fluorescence Resonance Energy Transfer technology. We thank C. M. ten Hagen-Jongman for technical support.
Referência(s)