Flexibility in targeting and insertion during bacterial membrane protein biogenesis

www.elsevier.com/locate/ybbrc

Biochemical and Biophysical Research Communications 362 (2007) 727–733

Flexibility in targeting and insertion during bacterial membraneprotein biogenesis

Edwin van Bloois, Corinne M. ten Hagen-Jongman, Joen Luirink *

Department of Molecular Microbiology, Institute of Molecular Cell Biology, Vrije Universiteit, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands

Received 30 July 2007Available online 20 August 2007

Abstract

The biogenesis of Escherichia coli inner membrane proteins (IMPs) is assisted by targeting and insertion factors such as the signalrecognition particle (SRP), the Sec-translocon and YidC with translocation of (large) periplasmic domains energized by SecA and theproton motive force (pmf). The use of these factors and forces is probably primarily determined by specific structural features of anIMP. To analyze these features we have engineered a set of model IMPs based on endogenous E. coli IMPs known to follow distincttargeting and insertion pathways. The modified model IMPs were analyzed for altered routing using an in vivo protease mappingapproach. The data suggest a facultative use of different combinations of factors.� 2007 Elsevier Inc. All rights reserved.

Keywords: Membrane protein assembly; Protein targeting; Ribosome; SRP; YidC Sec-translocon; Escherichia coli

The biogenesis of Escherichia coli IMPs is accomplishedin three consecutive steps: (i) membrane targeting, (ii)insertion into the lipid bilayer, and (iii) folding and assem-bly into the final functional structure. The biogenesis ofonly a few IMPs has been studied in detail pointing at mul-tiple versatile pathways involving assistance by differentproteinaceous factors [1,2]. Targeting of most IMPs ana-lyzed thus far requires a conserved and essential systemcomprising the signal recognition particle (SRP) and itsreceptor FtsY [3]. The E. coli SRP consists of the signalbinding protein Ffh and the 4.5S RNA. The SRP bindsto hydrophobic targeting signals present in nascent IMPs.The ribosome nascent chain–SRP complex is then trans-ferred to the Sec-translocon via FtsY.

The Sec-translocon is a conserved heterotrimeric com-plex, consisting of the IMPs SecY, SecE, and SecG andfunctions as a protein conducting channel for both secre-tory proteins and IMPs [1]. The ATPase SecA is peripher-ally associated with the SecYEG complex and drives the

0006-291X/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.bbrc.2007.08.053

* Corresponding author. Fax: +31 20 5987155.E-mail address: [email protected] (J. Luirink).

translocation of secretory proteins and larger periplasmicdomains of IMPs. Additionally, the proton motive force(pmf) is required for protein translocation and many IMPsrequire the pmf for insertion and assembly [2].

YidC was recently identified as an essential IMP that isin part associated with the Sec-translocon [1]. It plays apivotal but poorly defined role downstream of the Sec-tran-slocon in later steps of IMP biogenesis such as the recogni-tion and lateral transfer of TMs from the Sec-translocon(FtsQ, Lep), the assembly of TMs (MtlA) and folding ofthe IMPs into their native structure (LacY). YidC alsooperates as a separate Sec-independent insertase. Thisalternative integration pathway is used by a few relativelysimple IMPs, including the small phage coat proteinsM13 and Pf3, and the endogenous IMPs Foc and MscL.

The different targeting and integrating factors such asSRP, Sec(A)YEG, and YidC can be envisioned as modules.Different combinations of these modules result in the differ-ent membrane biogenesis pathways for specific IMPs [2].The use of the different targeting and integration modulesis probably primarily determined by specific structural fea-tures of an IMP. Only a limited collection of IMPs hasbeen analyzed in detail with respect to the requirements

728 E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733

for membrane assembly. Hence our understanding of thefeatures of IMPs that funnel them into specific pathwaysis limited. Here, we have engineered model IMPs to iden-tify specific features that determine targeting, insertionand translocation in vivo.

Materials and methods

Reagents and sera. Restriction enzymes, Expand long template PCRsystem and Lumi-LightPLUS Western blotting substrate were from RocheMolecular Biochemicals. [35S]methionine and Protein A Sepharose werefrom Amersham Biosciences. T4 ligase and alkaline phosphatase werefrom Invitrogen. Antisera used were from our own collection. Antiserumagainst influenza haemagglutinin (HA) was from Sigma.

Strains and plasmids. Strain Top10F’ (Invitrogen) was used as routinehost for all plasmid constructs. The 4.5S RNA depletion strain FF283, theSecE depletion strain CM124 and the YidC depletion strain FTL10 weregrown as described [4,5].

The model IMPs were constructed by PCR. For Foc-P2, the P2domain (residues 78–324) of Lep was amplified from plasmid pRD8 astemplate [6]. Next, Foc (residues 1–73) was amplified using plasmidpC4Meth-Foc as template [7] and the P2 domain was appended onto Foc.The PCR product was cloned into pEH1, pEH3, and pBAD18 [8,9]. Pf3-H1Lep-P2 was constructed as follows. Plasmid pC4Meth-48Pf3LepTAG7[10] was used as template to construct 18Pf3-Lep by nested PCR and thePCR product was cloned into pC4Meth, yielding pC4Meth-18Pf3-Lep. H2(residues 62–77) was deleted by PCR using pC4Meth-18Pf3-Lep as tem-plate. Pf3-H1Lep-P2 coding sequences were introduced into pEH1, pEH3,and pBAD18. Plasmid pC4Meth-Foc-Foc-HA was obtained by twoindependent PCRs, using pEH1-Foc-HA as template [7]. Primers weredesigned such that during the first PCR both an NcoI and BamHI sitewere introduced at the 5 0 or 3 0 end of the PCR product. For the secondPCR, primers were designed such that both an EcoRI and NcoI site wereintroduced at the 5 0 or 3 0 end of the PCR product. The PCR products weredigested with NcoI and BamHI or EcoRI and NcoI and cloned into theEcoRI and BamHI sites of pC4Meth. An HA-tag (Tyr-Pro-Tyr-Asp-Val-Pro-Asp-Tyr-Ala), preceded by a flexible linker peptide (Pro-Gly-Gly) wasattached to the C-terminus of Foc-Foc by exchanging the NcoI/HindIIIfragment of pC4Meth-Foc-Foc with the corresponding fragment of pEH3-Foc-HA [7]. Foc-Foc-HA coding sequences were introduced into pEH1,pEH3, and pBAD18. Nucleotide sequences were verified by DNAsequencing. Primer sequences are available upon request.

Assay for in vivo membrane assembly. Strains MC4100, FF283,CM124, and FTL10 were grown to early-log phase. Cells harboringderivatives of pEH1 or pEH3 (MC4100, CM124, and FTL10) wereinduced for 1 min by adding IPTG (1 mM) and cells harboring derivativesof pBAD18 (FF283) were induced for 3 min by adding L-arabinose (0.2%).Cells were labeled with [35S]methionine (30 lCi/ml), converted to spher-oplasts and processed as described [4]. When indicated, the protonophoreCCCP (50 nM) was added 1 min prior to induction to dissipate the pmf orsodium azide (2 mM) was added 3 min prior to induction to block SecAfunction.

Results and discussion

Expression and detection of model IMPs

The F1Fo-ATPase subunit Foc only requires YidC formembrane insertion [12–14]. In contrast, Lep, the principalsignal peptidase of E. coli, was shown to use a combinedSec-YidC translocon and assembles in a strongly Sec-,but hardly YidC-dependent fashion [11]. Together, thesefindings have led to the concept of two distinct insertionsites, SecYEG-YidC and ‘‘YidC-only’’ [1]. Membrane tar-

geting of Lep requires the SRP/FtsY system, whereas themechanism of Foc is debated [7,12–14].

To study the features of an IMP that determine thetargeting and insertion pathway, we have engineeredmodel IMPs based on Foc and Lep (Fig. 1A). BothLep and Foc span the membrane twice with an N-out,C-out topology. In contrast to Foc, Lep has a largetranslocated C-terminal catalytic (P2) domain of �240residues. To determine the influence of this domain onthe selection of assembly pathways, we fused the P2domain to the C-terminus of Foc. Furthermore, we havereduced the ‘‘complexity’’ of Lep by deleting its secondTM. To assay translocation of the N-terminus, the first18 residues of Pf3 were fused to the N-terminus, result-ing in Pf3-LepH1-P2. Most ‘‘complex’’ (>2 TMs) poly-topic IMPs analyzed (e.g. LacY and MalF) require theSec-translocon for insertion [15,16]. To establish theimpact of the number of TMs in an IMP on the inser-tion pathway we made a tandem fusion of Foc. To per-mit immunodetection and determination of topology, anHA-tag was added to the C-terminus resulting in theFoc-Foc-HA construct.

To examine membrane assembly and the topology ofthe model IMPs, wild-type cells expressing the constructswere pulse-labeled, converted to spheroplasts and ana-lyzed by protease accessibility. In this assay, the P2domain, the HA-tag and the Pf3 N-terminal extensionserve as topological markers as they are sensitive to pro-teinase K when translocated across the inner membrane[7,11]. The constructs were immunoprecipitated usingantiserum directed against the P2 domain of Lep (Foc-P2 and Pf3-LepH1-P2) or antiserum against the HA-epi-tope (Foc-Foc-HA).

As shown in Fig. 1B, Foc-P2 and Foc-Foc-HA are acces-sible to proteinase K, indicating that the P2 domain andHA-tag are translocated. Translocation of the N-terminalregion of Pf3-LepH1-P2 is indicated (arrow) by a slightshift in molecular weight in protease treated spheroplasts.This protease protected form is degraded in detergent-sol-ubilized cells, confirming that the construct is intrinsicallyproteinase K sensitive but protected in spheroplasts bythe IM. Trigger factor (TF) and OmpA are cytoplasmicand outer membrane control proteins used to monitorspheroplast formation and proteinase K treatment, respec-tively. Combined, these results suggest proper membraneassembly and confirm the expected topology of the threenovel constructs.

Foc-P2 is assembled via the SRP/Sec(A)YEG/YidC

pathway

To examine whether the SRP is required for targeting ofFoc-P2, we used the 4.5S RNA depletion strain FF283 inprotease mapping. The data presented in Fig. 2A show thatFoc-P2 is to a large degree protected against proteolysis byproteinase K upon depletion of 4.5S RNA, suggesting thatthe SRP is required for membrane targeting. As a control,

Fig. 1. Membrane assembly of model IMPs. (A) Membrane topology of wild-type Leader peptidase (Lep), Foc and derivative hybrid constructs. (B)Analysis of membrane assembly and topology by protease mapping. MC4100 wild-type cells expressing the model IMPs were grown, pulse-labeled,converted to spheroplasts and treated with proteinase K (PK) as described in Materials and Methods. The samples were immunoprecipitated usingantibodies directed against Lep and the HA-tag (bottom) or antibodies against OmpA and trigger factor (TF) (top). See the text for further details. Adistinct proteolytic fragment of Pf3-H1Lep-P2 is indicated (^).

Fig. 2. Foc-P2 is assembled via the SRP/Sec(A)YEG/YidC pathway. Foc-P2 was analyzed for its SRP, Sec-translocon, YidC and pmf dependence bytransforming the strains FF283 (A), CM124 (B), FTL10 (C) with a plasmid encoding Foc-P2. The strains expressing this construct were grown under non-depleting or depleting conditions and subjected to protease mapping as described in Fig. 1. The YidC content in 0.1 OD660 unit of cells used in (C) wasanalyzed by immunoblotting. MC4100 cells expressing Foc-P2 were treated with CCCP to dissipate the pmf and subjected to protease mapping (D).

E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733 729

the processing of the SRP-independent protein proOmpAwas monitored and appeared unaffected, suggesting thatinactivation of the Sec-translocon had not occurred.

Next, we studied involvement of the Sec-translocon inthe membrane assembly of Foc-P2, using the SecE deple-tion strain CM124 in which the essential secE gene is under

730 E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733

control of an arabinose-inducible promoter. Upon deple-tion of SecE, Foc-P2 is fully protected against proteolysis,indicating that translocation of the P2 domain is almostcompletely blocked under these conditions (Fig. 2B). Thissuggests that proper topogenesis of Foc-P2 requires theSec-translocon. Efficient depletion of SecE is indicated bythe accumulation of proOmpA (Fig. 2B) that depends onthe Sec-translocon for translocation and processing. Toinvestigate a possible role of SecA in the translocation ofthe P2 domain in Foc-P2, azide was used to inhibit theATPase activity of SecA. Translocation of the P2 domainof Foc-P2 was almost completely blocked in the presenceof azide (data not shown), showing that SecA is requiredfor the translocation of the P2 domain.

To investigate the role of YidC in the membrane assem-bly of Foc-P2, we used the YidC depletion strain FTL10 inwhich the yidC gene is under control of an arabinose-inducible promotor. Depletion of YidC only slightlyaffected the accessibility of Foc-P2 towards proteinase K,suggesting that YidC is to a large extent dispensable forinsertion and translocation of the P2 domain (Fig. 2C).Processing and translocation of proOmpA that is indepen-dent of YidC, appeared unaffected under these conditions,suggesting that inactivation of the Sec-translocon had notoccurred. YidC depletion was verified by immunoblotting,using YidC specific antiserum (Fig. 2C, lower panel).

Fig. 3. Pf3-H1Lep-P2 only requires the SRP. Pf3-H1Lep-P2 was analyzed for i(A), CM124 (B), FTL10 (C), and MC4100 treated with CCCP (D) as describeprotease protected form (d) and a proteinase K-sensitive cross-reacting proteinwas analyzed by immunoblotting.

To study whether the pmf is involved in translocation ofthe P2 domain of Foc-P2, we used CCCP, a protonophorethat dissipates the proton gradient across the IM. CCCPtreatment severely affected the proteinase K accessibilityof Foc-P2, showing that the pmf is essential for the translo-cation of the large P2 domain in Foc-P2 (Fig. 2D). Addi-tion of CCCP also resulted in the accumulation ofproOmpA, confirming a collapse of the pmf.

Combined, the data suggest that Foc-P2 is targeted bythe SRP and assembles in a Sec-dependent but YidC-inde-pendent mechanism resembling Lep. Not surprisingly, theSec-translocon (including SecA) and the pmf are requiredfor translocation of the large P2 domain at the C-terminusof this construct [2]. Surprising is the YidC-independentinsertion of this construct. Apparently, P2 located at theC-terminus of the hybrid, is able to dictate insertion ofthe upstream region in the Sec-translocon possibly by a(partly) post-translational mechanism. In contrast, unfusedFoc approaches the YidC insertase co-translationally [7].Remarkably, the endogenous E. coli IMP CyoA wasrecently shown to insert by yet another vectorial two-stepprocess. YidC-dependent insertion of the N-terminaldomain was shown to be followed by Sec-dependent trans-location of the large C-terminal periplasmic domain mak-ing the complete process very dependent on YidC [17–19]. In vitro studies will be required to establish whether

ts SRP, Sec-translocon, YidC and pmf dependence using the strains FF283d in Fig. 2. For clarity, the proteolytic fragment of Pf3-H1Lep-P2 (^), the

(*) are indicated. The YidC content in 0.1 OD660 unit of cells used in (C)

E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733 731

Foc-P2 uses a novel, partly post-translational insertionmechanism.

Pf3-H1Lep-P2 only requires the SRP

Fig. 3A shows that the SRP pathway is needed for mem-brane targeting of Pf3-LepH1-P2. In the absence of 4.5SRNA, significant protease protection of Pf3-LepH1-P2 isobserved (Fig. 3A; indicated by a closed circle). Depletionof 4.5 S RNA did not affect processing of proOmpA. Theoccasional protease insensitivity of a small fraction ofOmpA under SRP depletion conditions has been observedbefore [19]. The requirement of SRP is consistent within vitro recruitment of SRP at Lep H1 in nascent Lep [20].

Pf3-LepH1-P2 is accessible to proteinase K regardless ofwhether SecE is depleted or not (Fig. 3B). Likewise, inhibi-tion of SecA by azide had no discernible effect on the acces-sibility of Pf3-LepH1-P2 towards proteinase K (data notshown). This suggests that proper topogenesis of thisconstruct does not require the Sec-translocon and SecA.Strikingly, there is also no measurable protection of Pf3-LepH1-P2 from added proteinase K when YidC is depleted(Fig. 3C) suggesting that membrane assembly of this con-struct is not dependent on YidC either. YidC depletionwas verified by immunoblotting, using YidC specific antise-rum (Fig. 3C, bottom panel). Expression of Pf3-LepH1-P2in FTL10 reproducibly generates a faster migrating prod-uct (asterisk) even in the presence of YidC, which might

Fig. 4. Foc-Foc-HA uses the SRP/YidC pathway for membrane assembly.dependence using the strains FF283 (A), CM124 (B), FTL10 (C), and MC410OD660 unit of cells used in (C) was analyzed by immunoblotting.

represent a proteinase K sensitive protein, cross-reactingwith the Lep antiserum. The same effect of YidC depletionon the membrane assembly of Pf3-LepH1-P2 was observedusing YidC depletion strain JS7131 [11] (data not shown),confirming that YidC is not required for membrane assem-bly of this construct. Dissipation of the pmf by CCCP hasvery little effect on the accessibility of Pf3-LepH1-P2towards proteinase K (Fig. 3D), suggesting that the pmfis not required for the translocation of the small periplas-mic domain of this construct.

The unexpected observation that insertion of Pf3-LepH1-P2 occurs completely independent of the pmf,Sec-translocon and YidC suggests that this construct usesa ‘‘spontaneous’’ insertion mechanism analogous to theendogenous IMP KdpD [21]. Alternatively, the apparentindependencies may reflect a facultative use of either theSec-machinery or YidC for insertion. Consistently, LepH1 is cross-linked to both SecY and YidC in nascent Lepat a very early stage during membrane insertion whenLep is only 50 amino acids in length suggesting affinityfor both factors [10]. Furthermore, the N-terminus ofLep is translocated independent of the Sec-transloconand appears only slightly affected by YidC depletionin vivo [11,22]. Finally, reconstitution experiments haveshown that nascent Lep that exposes only H1 can insertin proteoliposomes that contain either YidC or SecYEG[23]. The observed independence of the pmf is consistentwith previous findings using a similar construct [22].

Foc-Foc-HA was analyzed for its SRP, Sec-translocon, YidC and pmf0 treated with CCCP (D) as described in Fig. 2. The YidC content in 0.1

732 E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733

In contrast, insertion of wild-type Pf3 coat does require thepmf [24]. This apparent discrepancy might be due to therelatively hydrophobic character of H1 as compared tothe TM of Pf3 coat [25].

Foc-Foc-HA uses the SRP/YidC pathway for membrane

assembly

Membrane targeting of Foc-Foc-HA is SRP dependentas evidenced by the relatively strong effect on the accessibil-ity of Foc-Foc-HA towards proteinase K in the absence4.5S RNA (Fig. 4A). Depletion of SecE has very little effecton the insertion of Foc-Foc-HA (Fig. 4B). Similarly, inser-tion of Foc-Foc-HA was not affected upon inactivation ofSecA by azide (data not shown). This suggests that mem-brane assembly of this construct does not require the Sec-translocon and SecA. Depletion of YidC had a pronouncedeffect on the accessibility of the HA-epitope in Foc-Foc-HAto added proteinase K, suggesting that YidC is critical formembrane integration (Fig. 4C). YidC depletion was veri-fied by immunoblotting, using YidC specific antiserum(Fig. 4C, bottom panel). Interestingly, dissipation of thepmf by CCCP severely affected the proteinase K accessibil-ity of Foc-Foc-HA (Fig. 4D), suggesting that the pmf isrequired for the translocation of small polar domains ofthis construct.

The Foc-Foc-HA construct appears to be targeted by theSRP to the ‘‘YidC-only’’ insertase. YidC is required andsufficient for membrane insertion of Foc both in vivo andin vitro, whereas the requirement for SRP is debated[12,13]. Like Foc, all other known substrates of the YidC-only insertase are small (1 or 2 TMs) suggesting that YidCcan only catalyze insertion of small IMPs in the absence ofa Sec-translocon. However, the data reported here arguethat size per se (the number of TMs) is not an importantconstraint for YidC requirement. Extrapolating the data,YidC might be sufficient for insertion of more complexendogenous E. coli IMPs provided they have small translo-cated loops. In this respect, YidC would have a similarcapacity as its homologue Oxa1 that operates in the mito-chondrial IM in the absence of a Sec-translocon [26]. Inter-estingly, the membrane assembly of E. coli MelB, a highlyhydrophobic transporter with 12 TMs, has been shown tooccur independent of SecA, SecY and SecE [27]. This maypoint to a critical role for YidC in the membrane assemblyof MelB.

In contrast to Foc, Foc-Foc-HA clearly requires the pmffor assembly which may be related to the presence of theHA-tag [13,28]. Notably, the pmf is required for the trans-location of negatively charged residues present in translo-cated domains of IMPs [2] and therefore the presence oftwo negatively charged residues in the HA-tag mightexplain the pmf dependency of Foc-Foc-HA. In conclusion,our data underline the important role of the SRP in IMPtargeting but also indicate the very flexible use of existinginsertion ‘‘modules’’ (SecYEG and YidC) to mediate themembrane assembly of endogenous and engineered IMPs.

Furthermore, the translocation of polar domains is ener-gized by distinct and cooperating ‘‘modules’’ (SecA andthe pmf). Notably, assembly is operationally defined hereas the translocation of periplasmic domains. It remains tobe investigated whether an altered mode of ‘‘assembly’’influences folding, complex formation, stability and func-tioning in general.

Acknowledgments

We thank Dirk-Jan Scheffers, Wouter Jong and Jan-Willem de Gier for critical reading of the manuscript.E.B. is supported by the Council for Chemical Sciencesof the Netherlands Society for Scientific Research.

References

[1] J. Luirink, G. von Heijne, E. Houben, J.W. de Gier, Biogenesis ofinner membrane proteins in Escherichia coli, Annu. Rev. Microbiol.59 (2005) 329–355.

[2] S.J. Facey, A. Kuhn, Membrane integration of E. coli modelmembrane proteins, Biochim. Biophys. Acta. 1694 (2004)55–66.

[3] J. Luirink, I. Sinning, SRP-mediated protein targeting: structure andfunction revisited, Biochim. Biophys. Acta. 1694 (2004) 17–35.

[4] L. Froderberg, E. Houben, J.C. Samuelson, M. Chen, S.K. Park, G.J.Phillips, R. Dalbey, J. Luirink, J.W. De Gier, Versatility of innermembrane protein biogenesis in Escherichia coli, Mol. Microbiol. 47(2003) 1015–1027.

[5] K. Hatzixanthis, T. Palmer, F. Sargent, A subset of bacterial innermembrane proteins integrated by the twin-arginine translocase, Mol.Microbiol. 49 (2003) 1377–1390.

[6] R.E. Dalbey, W. Wickner, Leader peptidase catalyzes the release ofexported proteins from the outer surface of the Escherichia coli

plasma membrane, J. Biol. Chem. 260 (1985) 15925–15931.[7] E. van Bloois, G. Jan Haan, J.W. de Gier, B. Oudega, J. Luirink,

F(1)F(0) ATP synthase subunit c is targeted by the SRP to YidC inthe E. coli inner membrane, FEBS Lett. 576 (2004) 97–100.

[8] L. Hashemzadeh-Bonehi, F. Mehraein-Ghomi, C. Mitsopoulos, J.P.Jacob, E.S. Hennessey, J.K. Broome-Smith, Importance of usinglac rather than ara promoter vectors for modulating the levels oftoxic gene products in Escherichia coli, Mol. Microbiol. 30 (1998)676–678.

[9] L.M. Guzman, D. Belin, M.J. Carson, J. Beckwith, Tightregulation, modulation, and high-level expression by vectorscontaining the arabinose PBAD promoter, J. Bacteriol. 177(1995) 4121–4130.

[10] E.N. Houben, R. Zarivach, B. Oudega, J. Luirink, Early encountersof a nascent membrane protein: specificity and timing of contactsinside and outside the ribosome, J. Cell Biol. 170 (2005) 27–35.

[11] J.C. Samuelson, M.Y. Chen, F.L. Jiang, I. Moller, M. Wiedmann, A.Kuhn, G.J. Phillips, R.E. Dalbey, YidC mediates membrane proteininsertion in bacteria, Nature 406 (2000) 637–641.

[12] M. van der Laan, P. Bechtluft, S. Kol, N. Nouwen, A.J. Driessen,F1F0 ATP synthase subunit c is a substrate of the novel YidCpathway for membrane protein biogenesis, J. Cell Biol. 165 (2004)213–222.

[13] L. Yi, N. Celebi, M. Chen, R.E. Dalbey, Sec/SRP requirements andenergetics of membrane insertion of subunits a, b, and c of theEscherichia coli F1F0 ATP synthase, J. Biol. Chem. 279 (2004)39260–39267.

[14] J.W.L. de Gier, P. Mansournia, Q.A. Valent, G.J. Phillips, J. Luirink,G. von Heijne, Assembly of a cytoplasmic membrane protein inEscherichia coli is dependent on the signal recognition particle, FEBSLett. 399 (1996) 307–309.

E. van Bloois et al. / Biochemical and Biophysical Research Communications 362 (2007) 727–733 733

[15] K. Ito, Y. Akiyama, In vivo analysis of integration of membraneproteins in Escherichia coli, Mol. Microbiol. 5 (1991) 2243–2253.

[16] B. Traxler, C. Murphy, Insertion of the polytopic membrane proteinMalF is dependent on the bacterial secretion machinery, J. Biol.Chem. 271 (1996) 12394–12400.

[17] D.J. du Plessis, N. Nouwen, A.J. Driessen, Subunit a of cytochrome ooxidase requires both YidC and SecYEG for membrane insertion, J.Biol. Chem. 281 (2006) 12248–12252.

[18] N. Celebi, L. Yi, S.J. Facey, A. Kuhn, R.E. Dalbey, Membranebiogenesis of subunit II of cytochrome bo oxidase: contrastingrequirements for insertion of N-terminal and C-terminal domains, J.Mol. Biol. 357 (2006) 1428–1436.

[19] E. van Bloois, G.J. Haan, J.W. de Gier, B. Oudega, J. Luirink,Distinct requirements for translocation of the N-tail and C-tail of theEscherichia coli inner membrane protein CyoA, J. Biol. Chem. 281(2006) 10002–10009.

[20] E.N.G. Houben, P.A. Scotti, Q.A. Valent, J. Brunner, J.W.L. de Gier,B. Oudega, J. Luirink, Nascent Lep inserts into the Escherichia coliinner membrane in the vicinity of YidC, SecY and SecA, FEBS Lett.476 (2000) 229–233.

[21] S.J. Facey, A. Kuhn, The sensor protein KdpD inserts into theEscherichia coli membrane independent of the Sec translocase andYidC, Eur. J. Biochem. 270 (2003) 1724–1734.

[22] J.-I. Lee, A. Kuhn, R.E. Dalbey, Distinct domains of an oligotopicmembrane protein are Sec-dependent and Sec-independent formembrane insertion, J. Biol. Chem. 267 (1992) 938–943.

[23] E.N.G. Houben, M.L. Urbanus, M. van der Laan, C.M. ten Hagen-Jongman, A.J.M. Driessen, J. Brunner, B. Oudega, J. Luirink, YidCand SecY mediate membrane insertion of a type I transmembranedomain, J. Biol. Chem. 277 (2002) 35880–35886.

[24] J. Rohrer, A. Kuhn, The function of a leader peptide in translocatingcharged amino acyl residues across a membrane, Science 250 (1990)1418–1421.

[25] V.M. Delgado-Partin, R.E. Dalbey, The proton motive force, actingon acidic residues, promotes translocation of amino-terminaldomains of membrane proteins when the hydrophobicity of thetranslocation signal is low, J. Biol. Chem. 273 (1998) 9927–9934.

[26] B.S. Glick, G. von Heijne, Saccharomyces cerevisiae mitochondrialack a bacterial-type Sec machinery, Protein Sci. 5 (1996) 2651–2652.

[27] M. Bassilana, C. Gwizdek, In vivo membrane assembly of the E.colipolytopic protein, melibiose permease, occurs via a Sec-independentprocess which requires the protonmotive force, EMBO J. 15 (1996)5202–5208.

[28] M. van der Laan, N. Nouwen, A.J. Driessen, SecYEG proteolipo-somes catalyze the Deltaphi-dependent membrane insertion of FtsQ,J. Biol. Chem. 279 (2004) 1659–1664.