Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2009 Jul 14; 106(28): 11588–11593.
Published online 2009 Jul 6. doi:  10.1073/pnas.0900638106
PMCID: PMC2710650
Biophysics and Computational Biology

Insertion of short transmembrane helices by the Sec61 translocon


The insertion efficiency of transmembrane (TM) helices by the Sec61 translocon depends on helix amino acid composition, the positions of the amino acids within the helix, and helix length. We have used an in vitro expression system to examine systematically the insertion efficiency of short polyleucine segments (Ln, n = 4 … 12) flanked at either end by 4-residue sequences of the form XXPX-Ln-XPXX with X = G, N, D, or K. Except for X = K, insertion efficiency (p) is <10% for n < 8, but rises steeply to 100% for n = 12. For X = K, p is already close to 100% for n = 10. A similar pattern is observed for synthetic peptides incorporated into oriented phospholipid bilayer arrays, consistent with the idea that recognition of TM segments by the translocon critically involves physical partitioning of nascent peptide chains into the lipid bilayer. Molecular dynamics simulations suggest that insertion efficiency is determined primarily by the energetic cost of distorting the bilayer in the vicinity of the TM helix. Very short lysine-flanked leucine segments can reduce the energetic cost by extensive hydrogen bonding with water and lipid phosphate groups (snorkeling) and by partial unfolding.

Keywords: hydrophobic mismatch, membrane protein synthesis, membrane proteins, molecular dynamics simulation, lipid bilayer

The thickness of the hydrocarbon core of bilayer membranes is ≈30 Å, which leads to the expectation that transmembrane (TM) helices of helix-bundle membrane proteins (MPs) should be ≈20-aa long to span the hydrocarbon core. Consistent with this expectation, the average helix length of α-helical MPs of known 3D structure is 26 ± 5 residues (http://blanco.biomol.uci.edu/mptopo). A few membrane proteins, most notably the aquaporins (1) and H+/Cl exchangers (2) have membrane-buried helices as short as n = 10 residues, corresponding to a helix length of (n − 1) × 1.5 = 13.5 Å. Although there is ample evidence that Sec61/SecY translocons can insert short helices, the biophysical basis of short-helix insertion has not been explored systematically.

Kuroiwa et al. (3) found that the Sec61 translocon could efficiently insert artificial stop-transfer polyleucine sequences as short as 9 residues across rough microsomal membranes. Chen and Kendall (4) subsequently found that the Escherichia coli SecY translocon is also able to insert 10-residue polyleucine segments across the inner membrane. In a detailed examination of the molecular code that the Sec61 translocon uses for identifying TM segments, Hessa et al. (5) found that efficient stop-transfer segments comprised of short polyleucine helices could also be engineered into a multispanning membrane protein, specifically the E. coli leader peptidase (Lep). In a recent model-peptide study, Krishnakumar and London (6) determined the minimum hydrophobic length required for the formation of poly(Leu) or poly(LeuAla) TM helices in small unilamellar vesicles formed from diacyl phosphatidylcholines with chain acyl lengths of 14 to 24 carbons. Their results were consistent with the TM insertion of hydrophobic sequences as short as 11–12 leucine residues in vesicle membranes formed from dioleoylphosphatidylcholine (DOPC).

A limitation of the biochemical and model-peptide studies to date is that they can only be inferentially compared with each other. To overcome this limitation, we have used the same set of TM helices (H-segments) for biochemical, biophysical, and computational studies. By using a microsomal, in vitro expression system to examine systematically the Sec61 insertion efficiency of short polyleucine segments flanked at either end by 4-residue sequences of the form XXPX-Ln-XPXX (n = 4–12; X = G, N, D, or K), we found that helices with n = 10–12 were readily inserted into the membrane. We also synthesized selected sequences from the XXPX-Ln-XPXX family and examined by oriented circular dichroism (OCD) the incorporation of the peptides into oriented palmitoyloleoyl phosphocholine (POPC) and POPC:POPG (palmitoyloleoyl phosphatidylglycerol) multibilayers at various concentrations and hydrations. We found that short helices of the XXPX family were indeed stable in a transmembrane configuration in oriented multibilayers. Finally, to explore the physical principles underlying the stability of short helices under conditions of severe hydrophobic mismatch, we carried out molecular dynamics (MD) simulations of some of the peptides in transmembrane orientations. Taken together, the results from this multifaceted approach suggest that the biophysics of hydrophobic mismatch controls the insertion of short helices across the endoplasmic reticulum (ER) membrane.


Integration of Polyleucine H-Segments Across ER Membranes by the Sec61 Translocon.

We first determined insertion efficiencies of systematically designed H-segments of the form XXPX-Ln-XPXX (X = G, N, D, or K) across the endoplasmic membrane by using the approach of Hessa et al. (5, 7), summarized in Fig. 1A. In brief, the H-segments were introduced near the middle of the large luminal P2 domain of the model protein, Lep. The protein was then expressed in vitro in the presence of ER-derived dog pancreas rough microsomes (RMs). The insertion probability (efficiency) p was calculated on the basis of the fractions of singly (f1g) and doubly (f2g) glycosylated proteins by using p = f1g/(f1g + f2g) determined from quantitative phosphoimager scans of SDS/PAGE gels. Hessa et al. (5, 7) showed that p is proportional to the Boltzmann function [1 + exp(ΔGapp/RT)]−1, which we consequently used to characterize the probability-of-insertion data.

Fig. 1.
Integration of polyleucine H-segments into microsomal membranes. (A) Leader peptidase (Lep) model protein construct. Wild-type Lep, consisting of 2 TM helices (TM1, TM2) and a large luminal domain (P2), inserts into rough microsomes in an Nlum-Clum orientation. ...

The central Ln stretch was flanked by the tetrapeptides XXPX- and –XPXX to “insulate” it from the surrounding sequence in the Lep model protein. For the present work, we took H-segments with X = G flanks as the reference segments. The immediate sequence environment outside the H-segment is quite polar (ENGIRLSETS/H-segment/VPGQQNATWI), hence there is no additional hydrophobic flanking segment that could be “recruited” into the membrane-spanning H-segment. Examples of SDS/PAGE gels showing the translation products of GGPG-Ln-GPGG constructs with n = 8, 10, and 12 are shown in Fig. 1B. Control experiments show that the identities of the TM1 and TM2 helices of the Lep constructs have little influence on the insertion efficiencies of the H-segment (8).

Fig. 2A shows plots of the probability of insertion as a function of n for X = G, N, D, and K. The probabilities of insertion for the reference set of H-segments with X = G are plotted as solid black squares. These data accurately conform to a Boltzmann distribution, shown as a solid black curve. This distribution means that the insertion process has the appearance of equilibrium between the inserted and translocated states measured by the relative amounts of singly and doubly glycosylated translation products. The data show that the probability of insertion is ≈0 for n = 8 or fewer leucines, but rises steeply to ≈90% for n = 12 or greater. These data confirm the earlier observations of Kuroiwa et al. (3) and Chen and Kendall (4). Assuming an α-helix, 12 leucines corresponds to a length of 16.5 Å, which cannot span the typical 30-Å-thick hydrocarbon core of lipid bilayers without considerable negative hydrophobic mismatch.

Fig. 2.
Length dependence of various parameters. (A) Efficiency of insertion of XXPX-Ln-XPXX constructs in dog pancreas rough microsomes. Data for X = G (black squares) were collected for n = 5, 6, 8, 10, 12, and 19. The solid curve through the points is a best-fit ...

Replacement of the flanking Gly residues with Asn (Fig. 2, green triangles) had little effect on the probability of insertion, whereas replacement with Asp (Fig. 2, red circles) caused a small shift of the insertion probabilities toward longer runs of leucines. In contrast, replacement of the flanking glycines with Lys residues (Fig. 2, inverted blue triangles) caused a shift of the probability curve to shorter lengths so that 90% efficiency was achieved for n = 10 leucines. If these results are due to physical interactions of the peptides with the membrane, then it should be possible to construct model membrane systems in which short synthetic peptides are stable across lipid bilayer membranes.

Integration of Synthetic Polyleucine H-Segments Across Oriented Phospholipid Membranes.

We synthesized 2 families of H-segments, each containing n = 6, 8, 10, or 12 leucines. The glycine family (designated GLn) had X = Gly flanks; the lysine family (designated KLn) had X = Lys flanks. In methanol and trifluoroethanol solutions, the GLn and KLn peptide families were α-helical for all values of n as determined by solution CD spectra (Fig. S1). The conformations of the GLn family in phosphate buffer depended on n (Fig. S2A). For n = 6 or 8, the CD spectra were consistent with β-sheet structure, probably due to aggregation, whereas for n = 10 or 12, the conformations were predominantly α-helical. In contrast, the CD spectra of the KLn family in phosphate buffer were consistent with random coil for n = 6 or 8 and predominately α-helical for n = 10 or 12 (Fig. S2B). We conclude from these measurements that a conformational transition to α-helical occurs for both peptide families between n = 8 and n = 10.

The ability of the peptides to insert across a lipid bilayer was examined by means of OCD spectroscopy of peptides in oriented multibilayer arrays that were formed by depositing methanol solutions of lipid and peptide on quartz substrates. Highly oriented arrays form spontaneously after slow evaporation of the methanol followed by hydration via vapor-phase equilibration with saturated salt solutions at 66% or 93% relative humidity (RH) (9, 10). Although the bilayers are not fully hydrated at these RHs, the disposition of incorporated peptides is generally the same as in excess water (9, 11). Bilayers were composed of either POPC or 1:1 POPC:POPG; peptide:lipid concentrations ranged from 1:12 to 1:100. Bilayer/peptide multilayers (consisting of stacks of ≈1,000 bilayers) formed in this way are highly aligned parallel to the quartz substrate as determined by X-ray diffraction.

OCD spectra were obtained for all peptides in both POPC and POPC:POPG at all peptide:lipid ratios (Fig. S3). Representative OCD data for the GLn and KLn peptide families (n = 6 or 12) are shown in Fig. 3. Shown as solid curves in each panel are the theoretically expected spectra for TM helices aligned perpendicular and parallel to the bilayer. The nearly identical spectra for GL12 and KL12 (Fig. 3 A and B, respectively) show that the H-segments are predominately helical and aligned normal to the bilayer plane. The GL12 spectra adhere slightly less accurately to the ideal helix-normal spectra than the KL12 peptides, possibly indicating improved helicity due to the lysine residues. The spectra for GL6 and KL6, on the other hand, are not readily interpretable; OCD spectra such as these have not been previously reported as far as we can establish and their interpretation is not obvious. The solution spectra (Fig. S2) of the n = 6 and 8 peptides indicate β-sheets for GLn and random coil for KLn, suggesting that the conformations may be more extended than in oriented bilayers. It may be significant that the OCD spectra have minima at about the same wavelength as for the solution CD spectra of GL6 and GL8. One possibility is that the n = 6 and n = 8 peptides have a distorted, partially helical conformation. Although the OCD spectra provide no obvious information about the orientation of the peptides relative to the bilayer normal, one can say with confidence that the conformations and/or orientations of the peptides change as the lengths increase from n = 6 to n = 12 leucines.

Fig. 3.
Oriented circular dichroism (OCD) spectra of synthetic XXPX-Ln-XPXX peptides in oriented multilamellar POPC bilayer arrays for X = G and K. Oriented multilamellar bilayer arrays were formed on 2.5-cm-diameter quartz plates, as described in detail elsewhere ...

We characterized the changes in the OCD spectra empirically by plotting the positions of wavelength minima as a function of n (Fig. 2B). The curves are remarkably similar to the probability curves of Fig. 2A in that they reveal a sharp break between n = 8 and n = 10. We cannot interpret this change structurally, but it is clear that for n ≥ 10, the synthetic H-segments form TM α-helices, as predicted by the translocon experiments. The broad agreement between the synthetic-helix and translocon data are consistent with a common underlying physical principle.

Molecular Dynamics Simulations of Transmembrane Polyleucine H-Segments.

What is the conformation of the synthetic peptides in lipid bilayers, and how does the lipid bilayer adapt to them? To examine these questions, we carried out all-atom molecular dynamics simulations of the GLn and KLn peptides in POPC lipid bilayers for n = 6, 8, 10, 12, and 20 (Figs. S4 and S5). The transbilayer phosphate-to-phosphate thickness of ER lipid membranes (12) is ≈38 Å, similar to the thicknesses of DOPC and POPC at moderate hydrations (13). The number of atoms, the simulation box size (a × b × c) at equilibrium, the simulation time te (ns) required to reach equilibrium, and the lateral area per lipid (Alipid) for each system are reported in Table 1. If it were technically feasible to run all-atom simulations for milliseconds or seconds, one might observe transitions between membrane-parallel interfacial helices and transmembrane helices. However, all-atom simulations on millisecond or longer time scales are not yet feasible. Nonetheless, our simulations provide critical insights into the behavior of bilayers in the vicinity of our H-segment transmembrane helices.

Table 1.
Summary of the molecular dynamics simulations for various properties of GLn and KLn simulations in POPC bilayers

Our simulations (Table 1) indicate that Alipid is proportional to the helical length: GL6–12 and KL6–12 all have much lower values than GL20 and KL20. In the GLn systems, Alipid is directly proportional to the number of leucines, with a slope of 0.4 Å2 per residue, an intercept of 46.5 Å2 (R = 0.996). In the case of KLn systems, on the other hand, it seems that KL8, KL10, and KL12 have reached a plateau of 51.4 Å2. KL6 does not respect this trend, but it also failed to remain helical. Moreover, Alipid is 1.4 Å2 lower in the KL20 than in the GL20 system, which is likely explained by the fact that the KL20 system is 2 Å thicker than the GL20 system as measured by the parameter c (Table 1). Overall, our results indicate that the systems thicken as the length of the inserted helices decrease. This observation seems contrary to any intuitive expectation that a bilayer (and consequently system thickness) should shrink with decreases in helix length to ensure hydrophobic matching between the two. There are 2 reasons for thickening. First, water plays an important stabilizing role. Second, the lipid bilayer adopts a distinct shape around each type of peptide. The roles of water and bilayer shape changes become apparent by comparing the GL6, KL6, GL12, and KL12 systems (Fig. 4).

Fig. 4.
Summary of all-atom molecular dynamics simulations of GL6, GL12, GK6, and GK12 peptides in POPC bilayers. (Left) Cut-away snapshots of the peptides spanning the lipid bilayer. The representative snapshots were taken during the simulation production runs ...

The importance of water is revealed by snapshots of the 4 systems at the end of the 10-ns production run (Fig. 4 Left). Water (orange) penetrates deeply into the bilayer for GL6 and even through the bilayer for KL6, because of hydration of the peptide backbone. Water penetration is much less for the longer GL12 and KL12 peptides. This water penetration, shown quantitatively in Fig. 4 Right, results from the formation of cavities around each peptide due to the bilayer collapsing around the peptides (see below) through hydrophobic mismatch. Lysine groups are much longer and more polar than glycine, and these properties must necessarily have an important effect on a peptide's ability to stabilize itself through hydrogen-bonding interactions, as suggested by Fig. 4 and quantitated by the number of hydrogen bonds (nHB) in Table 1. There are nearly twice as many hydrogen-bonding interactions for the KLn peptides, which means that the lysine residues add additional stability that backbone interactions alone cannot provide in the case of the GLn peptides. These lysine interactions are a complex form of snorkeling (14, 15) between lysine sidechains and lipid phosphate groups. A complete analysis of the hydrogen bonding of the peptides is included in Fig. S7.

The shape of the bilayer surrounding the GLn and KLn peptides reveals the adaptation of bilayer and peptide-to-hydrophobic mismatch (Fig. 5). The mismatch is made possible by changes in the lipid order parameters in the peptide neighborhood (Fig. S6). The shape of the bilayer around each peptide was analyzed by using a model based on the Landau-de Gennes theory, in which the free energy of lipid–protein interactions is determined by minimizing the functional integral of the local free energy density (16) (see SI Text). The variation of membrane thickness (z) as a function of the radial distance (r) from the peptide obtained from the shape analysis (Fig. 5) provides the entrée for parameterizing and computing the free energy change ΔG associated with mismatch deformation relative to GL20 and KL20 (see SI Text). In simple terms, the energetic cost of deformation increases as the square of the difference of the hydrophobic mismatch between the peptide and the unperturbed lipid bilayer. Qualitatively, Fig. 5 shows that the GL12 and GL6 mismatch is higher than the mismatch between KL12 and KL6, meaning that it is energetically more costly to deform the bilayer around the GLn peptides than that around the KLn peptides. The computed relative free energies of deformation for GLn and KLn are compared in Fig. 2C. For values of n between 20 and 12, the mismatch energetic cost is about the same for the 2 peptide families. But for n < 12, the free energy cost for GLn is persistently greater than that for KLn. This result implies that it should be energetically easier for the translocon to insert short KLn peptides than to insert GLn peptides, which is in fact the case (Fig. 2A).

Fig. 5.
Radial time-averaged profiles of the bilayers in the simulations. The average thickness z(r) of the bilayer is shown as a function of the radial distance r from the average peptide axis. (Left) The time-averaged positions of the lipid carbonyl and peptide ...


We have used a consistent set of model TM segments (H-segments) for molecular biological, biophysical, and computational studies to gain insights into the mechanism of insertion of short polyleucine segments across the ER membrane by the Sec61 translocon. We used a microsomal in vitro expression system to examine systematically the Sec61 insertion efficiency of short polyleucine segments (Ln, n = 4 … 12) flanked at either end by 4-residue sequences of the form XXPX-Ln-XPXX with X = G, N, D, or K (Fig. 1). With the exception of X = K, insertion efficiency rises steeply between n = 8 to 12. For X = K, the range of steep rise shifts to shorter lengths (n = 8 to 10).

To explore the question of whether or not simple physical principles can explain the in vitro Sec61 results, we synthesized members of the GLn and KLn families of sequences, and examined by OCD their incorporation into POPC- and POPC:POPG-oriented multibilayers at various protein:lipid ratios and hydrations. The results were largely independent of lipid and hydration (Fig. S3). A comparison of the OCD data for GL6 and GL12 with KL6 and KL12 was revealing (Figs. 3 and and4).4). For n ≥ 10, peptides incorporated into oriented lipid bilayers had OCD spectra unambiguously characteristic of transmembrane helices. This result shows that insertion of TM helices as short as 10 leucines by the Sec61 translocon are consistent with a physical partitioning of TM segments between translocon and lipid bilayer (5, 7).

OCD spectra for n ≤ 8 were ambiguous with respect to orientation and secondary structure, but solution CD spectra suggested that that the peptides are likely rich in β or extended conformations. A plot of the OCD wavelength minima against n (Fig. 4) revealed a sharp break between n = 8 and n = 10, mirroring the length-dependent Sec61 probability of insertion (Fig. 2). The solution spectra (Fig. S2) and the MD simulations (see below) raise the possibility that the shorter sequences cross the membrane with imperfect α-helical or extended conformations. But whatever their conformations and orientations may be, one can conclude that the shorter sequences interact differently with bilayers than do the longer ones.

Molecular dynamics simulations (Fig. 5) show that the α-helical conformation is stable for all of the peptides examined, except for the KL6 peptide, which partially unfolds. The unfolding is stabilized in the simulation by waters entering the bilayer along the peptide backbone. We do not know, however, whether a similar phenomenon can occur in real membranes. The amount of hydrogen bonding with water and bilayer headgroups distinguishes the KLn from the GLn peptides (Table 1); the KLn peptides have about twice as many stabilizing hydrogen bonds due to the strong interactions of the positively charged lysine sidechains. This hydrogen bonding apparently allows the bilayer to adapt more readily to the KLn peptides, i.e., without so much hydrophobic mismatch, than to the GLn peptides. The resulting energetic cost of exposing the peptide bonds to the bilayer hydrocarbon interior (17) is apparently compensated for by improved interactions of the lysine sidechains with lipid phosphates and water (18) for KL6. Although one might expect a similar effect for KL8, none was observed, presumably because the cost of deforming the bilayer was smaller than the cost of exposing the peptide bonds to the bilayer interior.

An analysis of the shape of the bilayer in the vicinity of the peptides revealed consistent hydrophobic mismatch that, as expected, increased in proportion to the difference between the length of the leucine helices and the unperturbed bilayer thickness (Fig. 5). The shape analysis allowed the positive free energy cost of the hydrophobic mismatch for each peptide to be computed relative to the GL20 and KL20 peptides, which had essentially no mismatch (Fig. 2C). The energetic cost arises from the energy required to distort the bilayer to minimize the exposure of the nonpolar leucines to the aqueous phase. The most significant observation may be that the energetic costs for the GLn and KLn peptides rise roughly in parallel as n decreases to 10 or 12; but for n < 10, the energetic cost of mismatch becomes significantly smaller for the KLn peptides. This smaller energetic cost is most easily explained by the increased hydrogen bonding of the of the KLn peptides, which reduces, in effect, the amount of mismatch.

The consistent picture that emerges from these studies is that lipid bilayers with acyl chains of 16–18 carbons easily adapt to TM helices as short as 10–12 leucines without too severe an energy penalty. The energetic cost of inserting helices shorter than ≈10 residues rises dramatically and prohibitively unless the leucines are flanked by a sufficient number of lysine residues, whose additional hydrogen bonding interactions with the phospholipid headgroups (snorkeling) reduce hydrophobic mismatch and consequently the energetic cost of short-helix insertion. What do these findings say about the mechanism of insertion of TM segments by the Sec61 translocon? First, no new insertion mechanism need be invoked to explain the insertion of short TM segments. The ability of the translocon to guide the insertion of short helices is completely consistent with the physical principles elucidated by our synthetic peptide studies and our molecular dynamics simulations. Second, despite the complicated inner workings of the ribosome–translocon complex (19, 20), the final “selection rules” for transmembrane insertion of nascent chains by the translocon are broadly consistent with physical partitioning of single or possibly multimers of TM helices between the translocon and membrane bilayer (2123).


Expression in Vitro.

Constructs in pGEM1 were transcribed and translated in the TNT SP6 Quick Coupled Transcription/Translation System. One microgram DNA template, 1 μL [35S]-Met (5 μCi), and 1 μL microsomes (a gift from Dr. M. Sakaguchi, Hyogo University, Hyogo, Japan) were added at the start of the reaction, and samples were incubated for 90 min at 30 °C. Translation products were analyzed by SDS/PAGE and gels were quantitated on a Fuji FLA-3000 phosphoimager by using the Reader V1.8J/Image Gauge V 3.45 software. The membrane-insertion efficiency of a given mutant was calculated as the quotient between the intensity of the singly glycosylated band divided by the summed intensities of the singly glycosylated and doubly glycosylated bands. A small correction for the amount of doubly glycosylated protein was used as described by Hessa et al. (7). On average, the glycosylation levels vary by no more than ±2% between repeat experiments.

Peptide Synthesis.

The sequences of the peptides were 6, 8, 10, and 12 Leu flanked by the groups GGPG or KKPK. In addition, each peptide carried a Trp residue on the C terminal to facilitate measurements of concentration. The peptides were synthesized on a 433A Applied Biosystems automatic synthesizer with stepwise solid-phase procedures (24) by using fluorenylmethoxycarbonyl chemistry and trifluoroacetic cleavage.

Solution Circular Dichroism.

CD measurements were performed by using an upgraded Jasco-720 spectropolarimeter (Japan Spectroscopic). Normally, 10 to 30 scans were recorded between 190 and 260 nm at ambient temperature (≈25 °C), by using a 1 mm optical path. All spectra were corrected for background scattering by subtracting a vesicle-only spectrum measured with an appropriate concentration of vesicles in buffer, without the peptide. See SI Text for more details.

Oriented Circular Dichroism.

Oriented CD measurements were performed by using the Jasco-720 spectropolarimeter. Oriented CD spectra were measured by using oriented multibilayers deposited on a quartz slide according to the procedures of Huang and colleagues (2527).

Molecular Dynamics Simulations.

All-atom MD simulations of the GLn and KLn peptides with n = 6, 8, 10, 12, and 20 were carried out in POPC bilayers hydrated with 30 water molecules per lipid. Each system consisted of a single peptide in a TM configuration, 280 POPC lipids, and 8,400 water molecules. The simulations were run at a constant temperature of 300 K and a constant pressure of 1 bar until the dimensions of the simulation cell remained stationary (fluctuated around a stable average) for 10 ns. The details of the system set-up, equilibration, and simulation protocols are provided in the SI Text.

Supplementary Material

Supporting Information:


This work was supported by the National Institute of General Medical Sciences Grants GM46823, GM68002, GM74637 (to S.H.W.) and the National Science Foundation Grants CHE-0417158 and CHE-0750175 (to D.J.T.); as well as grants from the Swedish Cancer Foundation, the Swedish Foundation for Strategic Research, the Marianne and Marcus Wallenberg Foundation, and the Swedish Research Council (to G.v.H.), and grants from the Magnus Bergvall Foundation (to I.N.). We thank Le Fond Québecois de la Recherche sur la Nature et les Technologies for a graduate fellowship to S.J.


The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/cgi/content/full/0900638106/DCSupplemental.


1. Sui HX, Han B-G, Lee JK, Walian P, Jap BK. Structural basis of water-specific transport through the AQP1 water channel. Nature. 2001;414:872–878. [PubMed]
2. Dutzler R, Campbell EB, Cadene M, Chait BT, MacKinnon R. X-ray structure of a CIC chloride channel at 3.0 Å reveals the molecular basis of anion selectivity. Nature. 2002;415:287–294. [PubMed]
3. Kuroiwa T, Sakaguchi M, Mihara K, Omura T. Systematic analysis of stop-transfer sequence for microsomal membrane. J Biol Chem. 1991;266:9251–9255. [PubMed]
4. Chen H, Kendall DA. Artificial transmembrane segments: Requirements for stop transfer and polypeptide orientation. J Biol Chem. 1995;270:14115–14122. [PubMed]
5. Hessa T, et al. The molecular code for transmembrane-helix recognition by the Sec61 translocon. Nature. 2007;450:1026–1030. [PubMed]
6. Krishnakumar SS, London E. Effect of sequence hydrophobicity and bilayer width upon the minimum length required for the formation of transmembrane helices in membranes. J Mol Biol. 2007;374:671–687. [PMC free article] [PubMed]
7. Hessa T, et al. Recognition of transmembrane helices by the endoplasmic reticulum translocon. Nature. 2005;433:377–381. [PubMed]
8. Meindl-Beinker NM, Lundin C, Nilsson I, White SH, Von Heijne G. Asn- and Asp-mediated interactions between transmembrane helices during translocon-mediated membrane protein assembly. EMBO Rep. 2006;7:1111–1116. [PMC free article] [PubMed]
9. Hristova K, et al. An amphipathic α-helix at a membrane interface: A structural study using a novel X-ray diffraction method. J Mol Biol. 1999;290:99–117. [PubMed]
10. Wimley WC, White SH. Designing transmembrane α-helices that insert spontaneously. Biochemistry. 2000;39:4432–4442. [PubMed]
11. Hristova K, Dempsey CE, White SH. Structure, location, and lipid perturbations of melittin at the membrane interface. Biophys J. 2001;80:801–811. [PMC free article] [PubMed]
12. Mitra K, Ubarretxena-Belandia I, Taguchi T, Warren G, Engelman DM. Modulation of the bilayer thickness of exocytic pathway membranes by membrane proteins rather than cholesterol. Proc Natl Acad Sci USA. 2004;101:4083–4088. [PMC free article] [PubMed]
13. Hristova K, White SH. Determination of the hydrocarbon core structure of fluid dioleoylphosphocholine (DOPC) bilayers by X-ray diffraction using specific bromination of the double-bonds: Effect of hydration. Biophys J. 1998;74:2419–2433. [PMC free article] [PubMed]
14. Segrest JP, Feldman RJ. Membrane proteins: Amino acid sequence and membrane penetration. J Mol Biol. 1974;87:853–858. [PubMed]
15. Killian JA, von Heijne G. How proteins adapt to a membrane-water interface. Trends Biochem Sci. 2000;25:429–434. [PubMed]
16. Jensen MØ, Mouritsen OG. Lipids do influence protein function: The hydrophobic matching hypothesis revisited. Biochim Biophys Acta. 2004;1666:205–226. [PubMed]
17. Jayasinghe S, Hristova K, White SH. Energetics, stability, and prediction of transmembrane helices. J Mol Biol. 2001;312:927–934. [PubMed]
18. Freites JA, Tobias DJ, von Heijne G, White SH. Interface connections of a transmembrane voltage sensor. Proc Natl Acad Sci USA. 2005;102:15059–15064. [PMC free article] [PubMed]
19. Pitonzo D, Yang Z, Matsumura Y, Johnson AE, Skach WR. Sequence-specific retention and regulated integration of a nascent membrane protein by the ER Sec61 translocon. Mol Biol Cell. 2009;20:685–698. [PMC free article] [PubMed]
20. Skach WR. The expanding role of the ER translocon in membrane protein folding. J Cell Biol. 2007;179:1333–1335. [PMC free article] [PubMed]
21. Rapoport TA, Goder V, Heinrich SU, Matlack KES. Membrane-protein integration and the role of the translocation channel. Trends Cell Biol. 2004;14:568–575. [PubMed]
22. Von Heijne G. Formation of transmembrane helices in vivo-–Is hydrophobicity all that matters? J Gen Physiol. 2007;129:353–356. [PMC free article] [PubMed]
23. White SH. Membrane protein insertion: The biology–physics nexus. J Gen Physiol. 2007;129:363–369. [PMC free article] [PubMed]
24. Stewart JM, Young JD. Solid Phase Peptide Synthesis. Rockford, IL: Pierce; 1984.
25. Olah GA, Huang HW. Circular dichroism of oriented α helices. I. Proof of the exciton theory. J Chem Phys. 1988;89:2531–2538.
26. Olah GA, Huang HW. Circular dichroism of oriented α helices. II. Electric field oriented polypeptides. J Chem Phys. 1988;89:6956–6962.
27. Wu Y, Huang HW, Olah GA. Method of oriented circular dichroism. Biophys J. 1990;57:797–806. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...