Title: Cotranslational Membrane Protein Biogenesis at the Endoplasmic Reticulum
Abstract: In eukaryotic cells, most polypeptides destined to become membrane proteins are initially integrated into the membrane of the endoplasmic reticulum (ER) 1The abbreviations used are: ER, endoplasmic reticulum; TMS, transmembrane sequence; cryo-EM, cryo-electron microscopy; FRET, fluorescence resonance energy transfer; SRP, signal recognition particle; SR, SRP receptor; RNC, ribosome-nascent chain complex; SA, signal anchor; CFTR, cystic fibrosis transmembrane conductance regulator.1The abbreviations used are: ER, endoplasmic reticulum; TMS, transmembrane sequence; cryo-EM, cryo-electron microscopy; FRET, fluorescence resonance energy transfer; SRP, signal recognition particle; SR, SRP receptor; RNC, ribosome-nascent chain complex; SA, signal anchor; CFTR, cystic fibrosis transmembrane conductance regulator. before being sorted to the location at which they function. Integration occurs at sites in the ER membrane termed translocons that are comprised of a specific set of membrane proteins (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar, 2Schnell D.J. Hebert D.N. Cell. 2003; 112: 491-505Abstract Full Text Full Text PDF PubMed Scopus (191) Google Scholar). In most cases, proteins are integrated into the bilayer cotranslationally, i.e. at the same time that they are being synthesized by ribosomes. During this process, the biosynthetic machinery mediates the integration of transmembrane sequences (TMSs) into the nonpolar core of the bilayer and delivers aqueous cytoplasmic and luminal domains to the appropriate compartments. Simultaneously, a nascent protein may undergo covalent modification (e.g. signal sequence cleavage, disulfide bond formation, and N-glycosylation), folding, and interactions with other proteins (e.g. chaperones) that ultimately lead to the assembly of the polypeptide into a functional monomeric or multimeric complex (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar, 2Schnell D.J. Hebert D.N. Cell. 2003; 112: 491-505Abstract Full Text Full Text PDF PubMed Scopus (191) Google Scholar, 3Deutsch C. Neuron. 2003; 40: 265-276Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar). Membrane protein biogenesis is therefore exceedingly complex, especially because the mechanisms involved are further constrained by the need to maintain the permeability barrier of the membrane. Here we highlight the most recent advances in our understanding of cotranslational integration at the ER membrane, focusing on four overlapping areas: translocon structural and functional states; nascent chain topogenesis; insertion of TMSs into the bilayer; and nascent chain regulation of integration. Other processes coupled with integration (e.g. covalent modification, folding, assembly, and quality control) and protein integration at other membranes are beyond the scope of this minireview. Structure—The mammalian ER translocon consists of the core heterotrimeric Sec61 complex (Sec61αβγ) and TRAM along with several associated protein complexes (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar). The Sec61p complex in the yeast ER membrane contains homologous core components in addition to the Sec62/63/71/72 complex, whereas the prokaryotic translocase consists of a heterotrimeric SecYEG complex that is homologous to the Sec61 complex (2Schnell D.J. Hebert D.N. Cell. 2003; 112: 491-505Abstract Full Text Full Text PDF PubMed Scopus (191) Google Scholar). Membrane protein integration in bacteria also involves YidC, a protein that functions in both a Sec-dependent and Sec-independent manner (4Kuhn A. Stuart R. Henry R. Dalbey R.E. Trends Cell Biol. 2003; 13: 510-516Abstract Full Text Full Text PDF PubMed Scopus (107) Google Scholar). The mammalian translocon forms an aqueous pore that spans the ER membrane, and the walls of the pore are formed primarily by Sec61α (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar). Cryo-EM images indicate that 2–4 copies of the Sec61 complex oligomerize to form rings with pores of ∼20 Å (or indentations that may result from pore sizes below the resolution limit) that align coaxially with the nascent chain exit tunnel in the ribosome (5Hanein D. Matlack K.E.S. Jungnickel B. Plath K. Kalies K.-U. Miller K.R. Rapoport T.A. Akey C.W. Cell. 1996; 87: 721-732Abstract Full Text Full Text PDF PubMed Scopus (279) Google Scholar, 6Beckmann R. Bubeck D. Grassucci R. Penczek P. Verschoor A. Blobel G. Frank J. Science. 1997; 278: 2123-2126Crossref PubMed Scopus (292) Google Scholar, 7Ménétret J.-F. Neuhof A. Morgan D.G. Plath K. Radermacher M. Rapoport T.A. Akey C.W. Mol. Cell. 2000; 6: 1219-1232Abstract Full Text Full Text PDF PubMed Scopus (182) Google Scholar, 8Beckmann R. Spahn C.M.T. Penczek P.A. Sali A. Frank J. Blobel G. Cell. 2001; 107: 361-372Abstract Full Text Full Text PDF PubMed Scopus (319) Google Scholar, 9Morgan D.G. Ménétret J.-F. Neuhof A. Rapoport T.A. Akey C.W. J. Mol. Biol. 2002; 324: 871-886Crossref PubMed Scopus (89) Google Scholar). The first atomic level view of translocon proteins has been provided by the crystal structure of an archeal SecYEβ (10van den Berg B. Clemons Jr., W.M.J. Collinson I. Modis Y. Hartmann E. Harrison S.C. Rapoport T.A. Nature. 2004; 427: 36-44Crossref PubMed Scopus (983) Google Scholar). Because the SecYEβ complex was crystallized in the absence of substrate and hence is in a translocation-inactive and sealed or "stand-by" conformation, one can only conjecture how the complex might rearrange during substrate translocation and integration. Interestingly, this SecYEβ structure is proposed to form a functional translocation pore with a single heterotrimer, which is at odds with the multimeric Sec61 (5Hanein D. Matlack K.E.S. Jungnickel B. Plath K. Kalies K.-U. Miller K.R. Rapoport T.A. Akey C.W. Cell. 1996; 87: 721-732Abstract Full Text Full Text PDF PubMed Scopus (279) Google Scholar, 6Beckmann R. Bubeck D. Grassucci R. Penczek P. Verschoor A. Blobel G. Frank J. Science. 1997; 278: 2123-2126Crossref PubMed Scopus (292) Google Scholar, 7Ménétret J.-F. Neuhof A. Morgan D.G. Plath K. Radermacher M. Rapoport T.A. Akey C.W. Mol. Cell. 2000; 6: 1219-1232Abstract Full Text Full Text PDF PubMed Scopus (182) Google Scholar, 8Beckmann R. Spahn C.M.T. Penczek P.A. Sali A. Frank J. Blobel G. Cell. 2001; 107: 361-372Abstract Full Text Full Text PDF PubMed Scopus (319) Google Scholar, 9Morgan D.G. Ménétret J.-F. Neuhof A. Rapoport T.A. Akey C.W. J. Mol. Biol. 2002; 324: 871-886Crossref PubMed Scopus (89) Google Scholar) and SecYEG (11Manting E.H. van der Does C. Remigy H. Engel A. Driessen A.J.M. EMBO J. 2000; 19: 852-861Crossref PubMed Scopus (168) Google Scholar) complexes observed in cryo-EM studies as well as the detection of SecY oligomerization using FRET (12Mori H. Tsukazaki T. Masui R. Kuramitsu S. Yokoyama S. Johnson A.E. Kimura Y. Akiyama Y. Ito K. J. Biol. Chem. 2003; 278: 14257-14264Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). Further experiments are necessary to ascertain the structure and oligomeric state of a functioning mammalian translocon (e.g. FRET studies to directly determine translocon component stoichiometry and arrangement as well as the magnitude and dynamics of translocon conformational changes). Function—The mammalian ER translocon is a dynamic multicomponent and multifunctional complex that facilitates protein translocation across and integration into the ER membrane (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar). Ribosomes synthesizing membrane or secretory proteins are identified by a signal sequence in the nascent chain that binds to SRP (Fig. 1, i) (13Keenan R.J. Freymann D.M. Stroud R.M. Walter P. Annu. Rev. Biochem. 2001; 70: 755-775Crossref PubMed Scopus (479) Google Scholar). A GTP-dependent interaction between SRP and SR then elicits the binding of the RNC to the translocon and the transfer of the nascent chain to the translocon (Fig. 1, ii). Although some signal sequences are cleaved from the nascent chain by signal peptidase (Fig. 1, iii), others contain additional residues and are not cleaved; instead, such sequences move laterally into the bilayer to form a TMS that is termed a "signal anchor" (SA) sequence. As translation proceeds, different regions of the nascent chain must be directed to the cytosol, lumen, or bilayer without disrupting the permeability barrier of the membrane. To this end, the ribosome and translocon undergo a series of conformational changes that are designed to alternately gate the cytosolic or luminal end of the pore and thereby direct the nascent chain into the appropriate aqueous compartment. Following RNC targeting to the translocon, an ion-tight RNC-translocon junction seals off the nascent chain from the cytosol, thereby directing the nascent chain into the ER lumen (Fig. 1, iv) (14Crowley K.S. Liao S. Worrell V.E. Reinhart G.D. Johnson A.E. Cell. 1994; 78: 461-471Abstract Full Text PDF PubMed Scopus (297) Google Scholar, 15Liao S. Lin J. Do H. Johnson A.E. Cell. 1997; 90: 31-41Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Upon detection of a TMS (see below), the luminal end of the pore is closed by the action of the luminal Hsp70 chaperone BiP (Fig. 1, v), and shortly thereafter the RNC-translocon seal is opened to allow the cytoplasmic domain of the membrane protein to move into the cytosol while the luminal end of the pore remains closed (Fig. 1, vi and vii) (15Liao S. Lin J. Do H. Johnson A.E. Cell. 1997; 90: 31-41Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 16Haigh N.G. Johnson A.E. J. Cell Biol. 2002; 156: 261-270Crossref PubMed Scopus (99) Google Scholar). For a polytopic membrane protein with multiple TMSs, we presume that the translocon pore is alternately sealed at its cytosolic and luminal ends by the RNC and BiP, respectively, though this has yet to demonstrated experimentally. Upon termination of translation and release of the membrane protein into the bilayer, the ribosome-free translocon remains assembled, and its pore is sealed on the luminal side by the BiP-mediated gate (Fig. 1, viii) (17Hamman B.D. Hendershot L.M. Johnson A.E. Cell. 1998; 92: 747-758Abstract Full Text Full Text PDF PubMed Scopus (349) Google Scholar). The translocon also undergoes dramatic conformational changes during its functional cycle. Not only do TMSs of the translocon proteins part to allow nascent chain TMSs to move laterally into the bilayer, but the inner diameter of the pore also increases from 9 to 15 Å in the ribosome-free closed state (e.g. Fig. 1, i) to 40–60 Å in the RNC-bound state (e.g. Fig. 1, iv) to a smaller diameter when the pore is sealed by both BiP and the RNC (Fig. 1, vii) (16Haigh N.G. Johnson A.E. J. Cell Biol. 2002; 156: 261-270Crossref PubMed Scopus (99) Google Scholar, 18Hamman B.D. Chen J.-C. Johnson E.E. Johnson A.E. Cell. 1997; 89: 535-544Abstract Full Text Full Text PDF PubMed Scopus (239) Google Scholar). One reason for this dynamic flexibility may be the need to accommodate multiple TMSs that may leave the translocon in pairs or groups (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar). The above mechanisms that maintain the permeability barrier of the ER membrane were determined by measuring the accessibility of fluorescent-labeled nascent membrane proteins to hydrophilic collisional quenching agents located on either the cytosolic side or on both sides of the ER membrane. These studies revealed that the pore is never in an aqueous continuum with both the cytosol and lumen; during translocation of luminal domains, the pore is sealed by the RNC-translocon junction, and during synthesis of cytosolic domains, the luminal end of the pore is sealed by the BiP-mediated gate (15Liao S. Lin J. Do H. Johnson A.E. Cell. 1997; 90: 31-41Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 16Haigh N.G. Johnson A.E. J. Cell Biol. 2002; 156: 261-270Crossref PubMed Scopus (99) Google Scholar). Moreover, large collisional quenching agents have been observed to move completely through a RNC-bound mammalian translocon pore occupied by a translocating nascent chain, thereby indicating that there is no constriction in the pore during translocation (18Hamman B.D. Chen J.-C. Johnson E.E. Johnson A.E. Cell. 1997; 89: 535-544Abstract Full Text Full Text PDF PubMed Scopus (239) Google Scholar). The fluorescence-detected reduction in pore size observed with a translocon that is sealed by the action of BiP may be accomplished by a conformational change in the translocon core that creates a constriction, but whatever the structural basis of the smaller pore, it does not completely block ion movement through the pore (17Hamman B.D. Hendershot L.M. Johnson A.E. Cell. 1998; 92: 747-758Abstract Full Text Full Text PDF PubMed Scopus (349) Google Scholar). It is important to emphasize that this fluorescence-based experimental approach directly measures ion movement and accessibility and therefore reveals how an ion-tight seal is maintained at the translocon throughout integration to maintain ion gradients such as the Ca2+ stores in the ER lumen. In contrast, the existence of a gap between the ribosome and translocon in cryo-EM images (7Ménétret J.-F. Neuhof A. Morgan D.G. Plath K. Radermacher M. Rapoport T.A. Akey C.W. Mol. Cell. 2000; 6: 1219-1232Abstract Full Text Full Text PDF PubMed Scopus (182) Google Scholar, 8Beckmann R. Spahn C.M.T. Penczek P.A. Sali A. Frank J. Blobel G. Cell. 2001; 107: 361-372Abstract Full Text Full Text PDF PubMed Scopus (319) Google Scholar, 9Morgan D.G. Ménétret J.-F. Neuhof A. Rapoport T.A. Akey C.W. J. Mol. Biol. 2002; 324: 871-886Crossref PubMed Scopus (89) Google Scholar) led to the suggestion that the permeability barrier is maintained by a constriction in the translocon pore, a view supported by the crystal structure of the monomeric SecYEβ (10van den Berg B. Clemons Jr., W.M.J. Collinson I. Modis Y. Hartmann E. Harrison S.C. Rapoport T.A. Nature. 2004; 427: 36-44Crossref PubMed Scopus (983) Google Scholar). The differing interpretations are most likely explained by the differences in the samples being analyzed. Whereas the fluorescence experiments examine integration or translocation intermediates that are functional and fully assembled in the membrane, the translocons used for high resolution cryo-EM images are detergent-solubilized and hence devoid of lipids, the translocon core protein TRAM, and other translocon-associated proteins, any of which may be important in maintaining a functional ion-tight RNC-translocon junction. Also, because an archeal organism would have to seal an unoccupied SecYEβ pore without the assistance of a BiP-like molecule on the other side of the membrane, a constriction that completely closes the pore may be required for such organisms. TMS Orientation—The orientation of a TMS in the bilayer, Nexo/Ccyt (type I, N terminus toward the lumen) or Ncyt/Cexo (type II, N terminus toward the cytoplasm), is determined by multiple features of the topogenic sequence (19Goder V. Spiess M. FEBS Lett. 2001; 504: 87-93Crossref PubMed Scopus (134) Google Scholar). Charged residues that flank the hydrophobic core of TMSs are the primary topogenic determinants, with the more positive end typically located in the cytoplasm in accordance with the "positive-inside rule" (20von Heijne G. J. Mol. Biol. 1992; 225: 487-494Crossref PubMed Scopus (1402) Google Scholar)(although not all eukaryotic membrane proteins strictly adhere to this rule (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar)). In addition, the hydrophobic core of the TMS influences orientation because greater total hydrophobicity and length of the nonpolar region of the TMS both promote Nexo/Ccyt orientation (19Goder V. Spiess M. FEBS Lett. 2001; 504: 87-93Crossref PubMed Scopus (134) Google Scholar) as does the hydrophobicity gradient within the TMS (21Harley C.A. Holt J.A. Turner R. Tipper D.J. J. Biol. Chem. 1998; 273: 24963-24971Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar). Folding of the nascent chain prior to arrival of a SA TMS at the translocon can also influence orientation (19Goder V. Spiess M. FEBS Lett. 2001; 504: 87-93Crossref PubMed Scopus (134) Google Scholar). Although these properties of a TMS have been shown to be important in topogenesis, the components of the biosynthetic machinery that decipher this information have yet to be identified. The fate of a TMS (its orientation and whether it is integrated or not) could be decided in part by soluble and membrane-bound proteins associated with the core components of the translocon (22Hegde R.S. Lingappa V.R. Trends Cell Biol. 1999; 9: 132-137Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar); the existence of such factors has been demonstrated by both biochemical (e.g. Ref. 23Hegde R.S. Voigt S. Lingappa V.R. Mol. Cell. 1998; 2: 85-91Abstract Full Text Full Text PDF PubMed Scopus (87) Google Scholar) and genetic (24Tipper D.J. Harley C.A. Mol. Biol. Cell. 2002; 13: 1158-1174Crossref PubMed Scopus (51) Google Scholar) approaches. In addition, phospholipids may affect topography (25van Voorst F. de Kruijff B. Biochem. J. 2000; 347: 601-612Crossref PubMed Scopus (63) Google Scholar) and influence substrate folding (26Bogdanov M. Dowhan W. J. Biol. Chem. 1999; 274: 36827-36830Abstract Full Text Full Text PDF PubMed Scopus (180) Google Scholar). Thus, the unique topography adopted by most native membrane proteins cannot be simplified or generalized to a single or very few interactions but instead results from the summation of several physicochemical properties of the TMSs and the combined influences of nascent chain interactions with the translocon and associated components. The existence of alternate topographical isoforms for some proteins (see Ref. 1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar for examples) presumably arises from a delicate balancing of different regulatory factors (22Hegde R.S. Lingappa V.R. Trends Cell Biol. 1999; 9: 132-137Abstract Full Text Full Text PDF PubMed Scopus (69) Google Scholar). The flanking regions can also influence the recognition of a TMS during biogenesis. The inversion of the charges that flank the N-terminal TMS of the Glut1 glucose transporter (27Sato M. Hresko R. Mueckler M. J. Biol. Chem. 1998; 273: 25203-25208Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar), the removal of positive charges from an internal loop (28Sato M. Mueckler M. J. Biol. Chem. 1999; 274: 24721-24725Abstract Full Text Full Text PDF PubMed Scopus (58) Google Scholar), or the introduction of basic residues into an exofacial loop (29Gafvelin G. Sakaguchi M. Andersson H. von Heijne G. J. Biol. Chem. 1997; 272: 6119-6127Abstract Full Text Full Text PDF PubMed Scopus (124) Google Scholar) results in the exclusion of hydrophobic TMSs from the membrane. Topogenesis of Polytopic Membrane Proteins—In the simplest model of polytopic protein topogenesis, the orientation of the first TMS inserted in the bilayer dictates the orientation of succeeding TMSs because of the requirement to alternate TMS orientations. However, the integration of many multispanning proteins appears to be more complex; polytopic proteins have multiple topogenic determinants, and proper integration and orientation often require the coordinated interaction of multiple topogenic segments of the nascent chain. Furthermore, an increasing number of studies has detected instances of delayed TMS insertion into the bilayer and of TMS insertion and reorientation that is nascent chain context-dependent. The topogenesis of many polytopic proteins has been studied by analyzing the topogenic activity of individual TMSs and by examining the coordinate topogenic activity of multiple TMSs using various reporter domains (30van Geest M. Lolkema J.S. Microbiol. Mol. Biol. Rev. 2000; 64: 13-33Crossref PubMed Scopus (165) Google Scholar). Many topogenic sequences, when examined separately, adopt an orientation that is opposite to their orientation in the native protein; hence, these TMSs require the presence of adjacent TMSs for proper topogenesis. In other cases, individual TMSs can assume the proper orientation in the translocon but must be paired with another TMS for efficient integration into the bilayer (30van Geest M. Lolkema J.S. Microbiol. Mol. Biol. Rev. 2000; 64: 13-33Crossref PubMed Scopus (165) Google Scholar). In several cases, a downstream TMS is required for the integration of a preceding TMS with weak topogenic activity (Fig. 2A), as has been observed for Sec61α (31Wilkinson B.M. Critchley A.J. Stirling C.J. J. Biol. Chem. 1996; 271: 25590-25597Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar), CFTR (32Lu Y. Xiong X. Helm A. Kimani K. Bragin A. Skach W.R. J. Biol. Chem. 1998; 273: 568-576Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar), and band 3 protein (33Ota K. Sakaguchi M. Hamasaki N. Mihara K. J. Biol. Chem. 1998; 273: 28286-28291Abstract Full Text Full Text PDF PubMed Scopus (65) Google Scholar). Such downstream determinants can even impart a transmembrane orientation onto a preceding hydrophilic segment (34Ota K. Sakaguchi M. von Heijne G. Hamasaki N. Mihara K. Mol. Cell. 1998; 2: 495-503Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar). In other cases, TMSs with weak topogenic determinants require a strongly hydrophobic upstream TMS for integration (35Ota K. Sakaguchi M. Hamasaki N. Mihara K. J. Biol. Chem. 2000; 275: 29743-29748Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar) either because an interaction between the two TMSs is required for integration and/or because the first TMS alters the translocon so as to promote its interaction with the second TMS (Fig. 2B). Interestingly, in one case increasing the loop length between the two TMSs resulted in translocation of the weak TM domain, suggesting that the strong TMS may have integrated prior to the arrival of the second TMS at the translocon (35Ota K. Sakaguchi M. Hamasaki N. Mihara K. J. Biol. Chem. 2000; 275: 29743-29748Abstract Full Text Full Text PDF PubMed Scopus (33) Google Scholar). Also, TMSs bearing charged residues have been shown to require specific electrostatic interactions (charge pairing) with neighboring TMSs for stable integration (3Deutsch C. Neuron. 2003; 40: 265-276Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar, 36Sato Y. Sakaguchi M. Goshima S. Nakamura T. Uozumi N. J. Biol. Chem. 2003; 278: 13227-13234Abstract Full Text Full Text PDF PubMed Scopus (28) Google Scholar). Time-dependent Reorientation—Goder and Spiess (37Goder V. Spiess M. EMBO J. 2003; 22: 3645-3653Crossref PubMed Scopus (112) Google Scholar) recently showed that each SA TMS initially orients Nexo/Ccyt in the translocon and that any inversion to the opposite orientation must occur within seconds. Consistent with previous studies, they also found that the final distribution of Nexo/Ccyt and Ncyt/Cexo orientations could be altered by adding or deleting a charge residue adjacent to a particular SA TMS or by altering its hydrophobicity. Moreover, TMS reorientation was hindered by glycosylation and by an increased length of intervening loop, suggesting that such inversion can be sterically prohibited and/or that the translocation of a certain length of polypeptide commits the substrate to a particular topography (38Goder V. Bieri C. Spiess M. J. Cell Biol. 1999; 147: 257-265Crossref PubMed Scopus (80) Google Scholar). Whether each TMS in a polytopic protein experiences a similar charge-, hydrophobicity-, and time-dependent evaluation and possible reorientation in the translocon has yet to be determined. Some membrane proteins undergo delayed TMS insertion or rearrangement. For example, internal TMSs (here defined as any TMS synthesized after the initial signal or SA sequence) of nascent CFTR (32Lu Y. Xiong X. Helm A. Kimani K. Bragin A. Skach W.R. J. Biol. Chem. 1998; 273: 568-576Abstract Full Text Full Text PDF PubMed Scopus (93) Google Scholar) and K+ channel (39Tu L. Wang J. Helm A. Skach W.R. Deutsch C. Biochemistry. 2000; 39: 824-836Crossref PubMed Scopus (57) Google Scholar) are required for efficient targeting. Moreover, aquaporin 1 (AQP1) is integrated as a loosely folded, four-TMS protein with extramembranous hydrophobic regions and during maturation undergoes topographical reorientation of multiple internal TMSs and peptide loops to assume its native six-TMS structure (40Lu Y. Turnbull I.R. Bragin A. Carveth K. Verkman A.S. Skach W.R. Mol. Biol. Cell. 2000; 11: 2973-2985Crossref PubMed Scopus (104) Google Scholar). The related AQP4 protein assumes the mature six-spanning topography cotranslationally, and the relevant differences in the topogenic determinants of AQP1 and AQP4 that dictate these disparate biosynthetic pathways have been identified (41Foster W. Helm A. Turnbull I. Gulati H. Yang B. Verkman A.S. Skach W.R. J. Biol. Chem. 2000; 275: 34157-34165Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar). Similarly, TMS10 of band 3 protein is translocated to the lumen and does not fully integrate into the membrane until both TMS12 and TMS1–3 have been assembled in the bilayer (42Kanki T. Sakaguchi M. Kitamura A. Sato T. Mihara K. Hamasaki N. Biochemistry. 2002; 41: 13973-13981Crossref PubMed Scopus (41) Google Scholar). The molecular environment of the lateral TMS pathway from the translocon pore into the bilayer has been experimentally assessed using photoreactive probes positioned within the TMS of intermediates at defined stages of integration. Upon insertion into the translocon pore, internal and SA TMSs are proximal to both translocon proteins and phospholipids (1Johnson A.E. van Waes M.A. Annu. Rev. Cell Dev. Biol. 1999; 15: 799-842Crossref PubMed Scopus (501) Google Scholar, 43Heinrich S.H. Mothes W. Brunner J. Rapoport T.A. Cell. 2000; 102: 233-244Abstract Full Text Full Text PDF PubMed Scopus (215) Google Scholar, 44Meacock S.L. Lecomte F.J.L. Crawshaw S.G. High S. Mol. Biol. Cell. 2002; 13: 4114-4129Crossref PubMed Scopus (68) Google Scholar, 45Heinrich S.U. Rapoport T.A. EMBO J. 2003; 22: 3654-3663Crossref PubMed Scopus (82) Google Scholar, 46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). These studies showed that TMSs have immediate access to phospholipid, but the length of time a TMS remains adjacent to a translocon protein(s) during integration is controversial. One proposed mechanism for TMS integration is a partitioning of the nonpolar TMSs into the hydrophobic lipid bilayer from the aqueous pore. This mechanism is supported by photocross-linking experiments that showed TMSs were only transiently proximal to translocon proteins (43Heinrich S.H. Mothes W. Brunner J. Rapoport T.A. Cell. 2000; 102: 233-244Abstract Full Text Full Text PDF PubMed Scopus (215) Google Scholar, 45Heinrich S.U. Rapoport T.A. EMBO J. 2003; 22: 3654-3663Crossref PubMed Scopus (82) Google Scholar, 47Mothes W. Heinrich S.U. Graf R. Nilsson I. von Heijne G. Brunner J. Rapoport T.A. Cell. 1997; 89: 523-533Abstract Full Text Full Text PDF PubMed Scopus (168) Google Scholar). In such a scenario, translocon proteins gate TMS access to the lipid phase without specifically recognizing or interacting with the TMS, thereby creating a one-step integration process in which a SA TMS moves directly from the pore into the bilayer. An alternative mechanism for TMS integration is based on other photocross-linking studies that showed internal and SA TMSs from a wide range of substrates were adjacent to translocon proteins for prolonged times (44Meacock S.L. Lecomte F.J.L. Crawshaw S.G. High S. Mol. Biol. Cell. 2002; 13: 4114-4129Crossref PubMed Scopus (68) Google Scholar, 46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar, 48Do H. Falcone D. Lin J. Andrews D.W. Johnson A.E. Cell. 1996; 85: 369-378Abstract Full Text Full Text PDF PubMed Scopus (207) Google Scholar, 49Knight B.C. High S. Biochem. J. 1998; 331: 161-167Crossref PubMed Scopus (19) Google Scholar). Some TMSs remain adjacent to Sec61α and/or TRAM (46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar, 48Do H. Falcone D. Lin J. Andrews D.W. Johnson A.E. Cell. 1996; 85: 369-378Abstract Full Text Full Text PDF PubMed Scopus (207) Google Scholar) as well as the recently identified PAT-10 (44Meacock S.L. Lecomte F.J.L. Crawshaw S.G. High S. Mol. Biol. Cell. 2002; 13: 4114-4129Crossref PubMed Scopus (68) Google Scholar) until translation terminates. Because in some cases more than 500 Å of fully extended polypeptide separates the TMS from the peptidyl-tRNA (46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar, 48Do H. Falcone D. Lin J. Andrews D.W. Johnson A.E. Cell. 1996; 85: 369-378Abstract Full Text Full Text PDF PubMed Scopus (207) Google Scholar), the nascent chain tether is sufficiently long to allow a TMS to diffuse away from the translocon, and the fact that it does not do so indicates that the translocon is actively regulating TMS release into the bilayer. Furthermore, because the TMSs of several different substrates form α-helices that occupy a fixed position in the translocon relative to Sec61α and TRAM, it appears that nascent chain TMSs are not able to rotate or move freely within the translocon and hence are bound to a translocon protein(s) during integration (46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). The non-random positioning of each TMS within the translocon differed for different TMSs examined in this study, which indicates that individual TMSs bind differently (most likely by van der Waals contacts along complementary surfaces) to Sec61α, TRAM, and/or other proteins prior to being released into the bilayer (46McCormick P.J. Miao Y. Shao Y. Lin J. Johnson A.E. Mol. Cell. 2003; 12: 329-341Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar). Moreover, TMSs of polytopic proteins interact with distinct translocon proteins during integration in a manner dependent on their relative order of insertion (44Meacock S.L. Lecomte F.J.L. Crawshaw S.G. High S. Mol. Biol. Cell. 2002; 13: 4114-4129Crossref PubMed Scopus (68) Google Scholar). Taken together, these results indicate that TMS traffic through the translocon is controlled by the binding of TMSs to translocon proteins and hence that the translocon plays an active regulatory role in membrane protein assembly in part by controlling the lateral movement of TMSs into the bilayer. A recent FRET study has revealed that inside the ribosomal exit tunnel close to the peptidyltransferase center the ribosome induces a nascent chain TMS to fold into a conformation compatible with an α-helix (50Woolhead C. McCormick P.J. Johnson A.E. Cell. 2004; 116: 725-736Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar). In contrast, a nascent secretory protein passes through the ribosome in an extended conformation. Furthermore, the folded TMS of a nascent membrane protein is sequentially proximal to two ribosomal proteins, L17 and L39, that are not exposed to a nascent secretory protein in the tunnel (50Woolhead C. McCormick P.J. Johnson A.E. Cell. 2004; 116: 725-736Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar). These interactions coincide with changes in the gating of the aqueous translocon pore (15Liao S. Lin J. Do H. Johnson A.E. Cell. 1997; 90: 31-41Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar). Therefore, it appears that the recognition of TMS folding by L17 initiates a transmembrane communication pathway that elicits the BiP-mediated closure of the luminal end of the pore (Fig. 1, v) and that subsequent TMS recognition by L39 mediates the opening of the RNC-translocon junction (Fig. 1, vi) (50Woolhead C. McCormick P.J. Johnson A.E. Cell. 2004; 116: 725-736Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar). Thus, maintenance of the ER membrane permeability barrier and the TMS-dependent conversion of the operational mode of the translocon from translocation to integration are apparently regulated by the nascent chain via a sophisticated series of interactions that involve components of both the ribosome and translocon as well as at least one luminal protein (15Liao S. Lin J. Do H. Johnson A.E. Cell. 1997; 90: 31-41Abstract Full Text Full Text PDF PubMed Scopus (213) Google Scholar, 16Haigh N.G. Johnson A.E. J. Cell Biol. 2002; 156: 261-270Crossref PubMed Scopus (99) Google Scholar, 50Woolhead C. McCormick P.J. Johnson A.E. Cell. 2004; 116: 725-736Abstract Full Text Full Text PDF PubMed Scopus (304) Google Scholar). Moreover, the ability of the ribosome to recognize TMSs and effect long range conformational changes may provide a basis for some aspects of topogenesis such as the mechanism by which a C-terminal TMS promotes integration of a weakly hydrophobic upstream TMS (Fig. 2A). The ribosome, translocon, and associated proteins therefore constitute a functional unit in which nascent chain-dependent structural features are recognized and communicated from one component to another, thereby effecting a coupled and ordered response that dictates membrane protein topogenesis, insertion, and assembly. The biosynthesis of a fully folded, functional membrane protein at the ER involves complex interactions of the nascent chain with the ribosome, translocon, membrane, and several other proteins. This complexity is then magnified by the differing requirements of a multitude of individual proteins for proper integration and assembly. Because we have information about the integration of only a few substrates, it is fair to say that our understanding of membrane protein biogenesis is still fragmentary and that many structural and mechanistic aspects of integration have yet to be discovered. Among the many important and fundamental questions that remain are the following. What is the oligomeric state of the intact mammalian translocon when bound to a RNC engaged in translocation or integration? How and where is an internal TMS rotated into an Ncyt/Cexo orientation? By what mechanism does a TMS promote the insertion of another TMS into the bilayer? Does this involve direct interactions between two TMSs? How and when do major rearrangements such as the transition from four to six TMSs in AQP1 occur? Is this structural change translocon-mediated? Do the TMSs of a polytopic protein leave the translocon singly, in pairs, or in a group? When do the TMSs of a polytopic protein assemble into their native structure? Do the ribosome and BiP alternate closing the pore during polytopic membrane protein integration? If so, does the ribosome induce the folding of each TMS to regulate changes at the translocon while the TMS is far inside the ribosome? What happens when the TMSs are separated by short loops (only a few residues)? Finally, the kinetics of cotranslational integration remain largely unexplored because of the difficulty in obtaining a synchronized and homogeneous sample to examine the time dependence of TMS translation, integration, and assembly. Yet by regulating the kinetics of nascent chain-translocon interactions at different points along the integration pathway, the biosynthetic machinery could, among other things, avoid misfolded proteins and allow TMSs to dynamically reorient within the translocon and to facilitate the insertion of TMSs of intermediate hydrophobicity and thereby profoundly influence the final folded state of the substrate protein. In conclusion, recent progress has demonstrated that the functional stages of membrane protein integration involve dynamic conformational changes of the ribosome-translocon machinery and that substrate integration is subject to regulation by the translocon, the ribosome, and the nascent chain itself. A major challenge for the future will entail determining how well the paradigms identified to date apply to every protein. Addressing the above questions will require the use of complex biochemical and biophysical approaches applied to an increasing array of substrates as well as structural studies on different functional states of the translocon and biophysical studies of the dynamics of those states. We thank David W. Andrews and William R. Skach for valuable suggestions.