QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Data Supplement All Versions of this Article: mol.107.043042v1 73/4/1092    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Riek, R. P. Articles by Graham, R. M. Search for Related Content PubMed PubMed Citation Articles by Riek, R. P. Articles by Graham, R. M.

Wide Turn Diversity in Protein Transmembrane Helices Implications for G-Protein-Coupled Receptor and Other Polytopic Membrane Protein Structure and Function

Computational and Structural Biology Division (R.P.R.) and Molecular Cardiology (A.A.F., G.B., R.M.G.) Divisions, Victor Chang Cardiac Research Institute, Darlinghurst, New South Wales, Australia; and Department of Pharmacology, School of Medical Sciences, University of New South Wales, Kensington, New South Wales, Australia (A.A.F.)

Received for publication October 30, 2007.

Accepted for publication January 17, 2008.

	Abstract

Top Abstract Materials and Methods Results and Discussion Conclusion References

 
Previously, we showed that perturbations of protein transmembrane helices are manifested as one of three types of noncanonical structures (wide turns, tight turns, and kinks), which, compared with -helices, are evident by distinctive C_iC_x distances. In this study, we report the analysis of more than 3000 transmembrane helices in 244 crystal structures from which we identified 70 wide turns (29 proline- and 41 nonproline-induced). Based on differences in the C_iC_i-4 and C_iC_i-5 profiles, we show that wide turns can be subclassified into three distinct subclasses (W₁, W₂, and W₃) that differ with regard to the number and position of backbone i i-5 H-bonds formed N-terminal to the perturbing or signature proline or nonproline residue. Although wide turns generally produce changes in helical direction of 20° to 30° and a lateral shift in the helical axis, some of the W₃ subclass are associated with changes of <5°. We also show that the distinct architectural features of wide turns allow the carbonyl bond of the i-4th residue, which is located on the widened loop of a wide turn, to be directed away from the helical axis. This provides regions of flexibility within helical regions allowing, for example, unique opportunities for interhelical H-bonding, including interactions with glycine zipper motifs, and for ion and cofactor binding. Furthermore, differences in wide-turn subtype usage by related protein family members, such as the G-protein-coupled receptors rhodopsin and the β2-adrenergic receptor, can significantly affect the orientation and position of residues critical for ligand binding and receptor activation.

These C_iC_x distance-difference plots reveal alterations in regular -helical geometry because, although peptide bond lengths are essentially invariant and the distance between C atoms of adjacent residues remains constant, rotations about the bonds on either side of the C atom of each residue produce changes in the and dihedral angles, leading to altered distances between C atoms two or more residues apart. As a result, non--helical regions are evident as positive or negative deviations in the distances between C atoms two to five residues apart in sequence, compared with the equivalent residues in an -helix (Riek et al., 2001; Rigoutsos et al., 2003) (Supplemental Fig. 1).

Although C_iC_x distance-difference plots allow wide turns, tight turns, and kinks to be identified, neither such plots nor other analyses of helical architecture provide sufficient detail to fully characterize noncanonical regions. For example, Sugeta and Miyazawa (1967) and others (Aqvist, 1986; Christopher et al., 1996; Bansal et al., 2000; Visiers et al., 2000; Cordes et al., 2002; Bright and Sansom, 2003; Mohapatra et al., 2004; Lopera et al., 2005) determined the helical axis, the bend angle of the helix, the rise per residue, and the inter-residue angle, by "best-fitting" the helical regions to a canonical four residue -helical template with the use of a centroid calculation (involving four consecutive C atoms) or a least-squares procedure, which involves one to two turns of helix either side of the bend created by a noncanonical segment, to calculate the bend angle. These approaches suffer from an inability to evaluate helical geometry on a residue-to-residue basis. Moreover, noncanonical segments, such as wide turns, involve an eccentric enlargement of the helical diameter, resulting in lateral displacement of the helical axis. This behavior is not accurately modeled using available methods, which rather, artifactually skew and twist the helical axis. Other approaches to identify helical perturbations that result in wide turns, also called -bulges, require the presence of one or more (i i-5) hydrogen-bonds (H-bonds) (Kabasch and Sander, 1983; Richards and Kundrot, 1988; Frishman and Argos, 1995; Labesse et al., 1997; Weaver, 2000; Fodje and Al-Karadaghi, 2002) and a turn of -helix before and after the wide turn (Cartailler and Luecke, 2004). As a result, wide turns not bounded by complete -helical turns may not be identified.

View larger version (15K): [in this window] [in a new window]   Fig. 1. Ribbon traces showing paths of wide turns and an -helix. A, an -helix and W3 wide turn induced by prolines viewed along the helical axis from C to N terminus. This wide turn example has <5° change in direction. a, position at which the widened turn of the helix starts and then resumes -helical geometry one full turn further N-terminal. The position of the pyrrolidine ring is indicated by an asterisk. Green ribbon: -helix; orange ribbon: wide turn. B, the yellow ribbon shows the path of a W1 wide turn looping out and around the pyrrolidine ring. The gray lines show the path of an -helix. The path of a wide turn formed by a partial unwinding of the helix is indicated by the cyan ribbon. C, view of B rotated by 90° showing the divergent paths of these wide turns. (Gray ribbon not shown.)

	Materials and Methods

Top Abstract Materials and Methods Results and Discussion Conclusion References

 
Database and Identification of Helices. Transmembrane helical regions were identified using Insight II (Accelrys, San Diego, CA). To determine potential H-bonding interactions typical of a 3₁₀-helix (i i-3), -helix (i i-4), or -helix (i i-5), hydrogen atoms were added either at pH 7.0 or at the pH specified in the structure file.

We treated the same proteins from different species as separate examples, if they display sequence differences in the noncanonical structures. Some proteins are highly represented with multiple entries in the database, particularly structures for photosynthetic reaction centers and bacteriorhodopsin. Initially, we considered all such entries if the resolution criteria were met (resolution, 3.0 Å; R-value, 0.3). However, if the structures of the transmembrane helices were identical in the various entries, only one representative member was evaluated. In instances in which structural variations were evident between entries for the same protein as a result of differences in crystallization conditions, to the different mutant forms, or different functional states (e.g., oxidized or reduced) or to crystallization in the presence or absence of different inhibitors or different metal ions, all such entries were considered in evaluating the detailed geometry of wide turns.

Identification of Noncanonical Regions. Noncanonical segments were identified based on C_iC_x distance-difference plots (Supplemental Table 1), as described previously (Riek et al., 2001). Here we focus on the wide turn or -bulge category of noncanonical structures, which display C_i-2 and C_i-3 distance differences that are positive and C_i-4 and C_i-5 distance differences that are negative.

In noncanonical helical regions with proline as the helix-perturbing residue, the position containing the proline is designated as the signature residue position (S₀). Nonprolyl residues that disrupt -helical geometry result in C_iC_x distance-difference profiles similar to those produced by prolyl residues (Supplemental Fig. 1, D-F). In such instances, these nonprolyl residues are the signature residues. Thus, such nonprolyl residues are defined as the signature residue if their substitution by a proline would produce a similar C_IC_x profile.

Helical Parameters. The effects of the residues within a noncanonical region on helical structure can be seen from changes in the position and direction of the helical axis. To accurately determine such changes, we characterized transmembrane protein helices on a residue-to-residue basis. This was done by evaluating helical parameters that are based on the coordinates of the atoms forming each C_iC_i+1 segment. From these coordinates, nine parameters could be determined: the three backbone dihedral angles, , , and ; the three bond lengths; and the three bond angles of the backbone atoms. These parameters alone allow a helical segment of as little as four residues to be constructed. The regularity of such a four-residue helix enables other helical parameters to be calculated, including the helical spoke angle, the radius of the helical segment, and the direction and length (rise per residue) of the helical axis (Riek et al., 2001). Based on these considerations, to determine the helical axis for any residue-to-residue pair, we constructed a de novo computed four-residue helix using the parameters determined from the particular C_iC_i+1 segment being examined (Supplemental Fig. 2). The regularity of such de novo constructed helices enabled precise trigo-nometrical determination of the positions of two helical axis points for each four-residue helix. The helical axis vector, which is equivalent to the rise per residue for each C_iC_i+1 segment within the helical region of interest (Supplemental Data, Fig. 2C), was determined from the helical axis points. The vector of the helical axis indicates the helical path of a segment and is defined in terms of its direction, position, and length.

View larger version (39K): [in this window] [in a new window]   Fig. 2. Effect of an insertion or loss of a residue on the helical architecture of various transmembrane segments. A, insertion of a residue in an -helical region is associated with a wide-turn noncanonical structure. This is shown with the transmembrane region of the H chain of B. viridis (B. v.). Alignment of this region with the corresponding transmembrane regions (both -helices) of R. sphaeroides (R.s.) and T. tepidum (T.t.) shows the region containing Leu30 of B. viridis looping out as a widened turn (orange ribbon) before resuming as an -helical structure. B, loss of a residue within a wide turn is associated with an -helical structure. This is illustrated in the case of helix 2 in Escherichia coli ubiquinol oxidase (E.c.), which is homologous to the A chain of bovine (Bov.) cytochrome c oxidase (BCCO). In helix 2 in BCCO, there are conserved prolines at two positions (Pro72 and Pro84) and a conserved asparagine (Asn80). Aligning the ubiquinol oxidase with the Paracoccus dentrificans (P.d.) or R. sphaeroides (R.s.) BCCOs places the proline (Pro72) and asparagine (Asn80) residues (both inducing wide turns in the bovine) at comparable positions, with the N-terminal regions of the proline-induced wide turns all following a similar backbone path. However, the region between Pro117 and Asn124 of ubiquinol oxidase (equivalent to Pro72 and Asn80, respectively, of BCCO) is -helical (orange ribbon), lacking the second wide turn as seen in the other three structures. Key conserved residues are indicated by dots. Arrows indicate positions of the two conserved prolines.

To assess the potential for backbone H-bond formation, the orientation of the C=O bond with respect to the helical axis was determined. For each segment in a wide turn region, the angle between the C=O bond and the helical axis was calculated. A backbone H-bond could form if the distance between the amide hydrogen atom donor and the carbonyl oxygen atom acceptor was 3.0Å and the C-H···O and H···O=C angles were 110° and 90° (Desiraju and Steiner, 1999), respectively.

Detailed Characterization of Wide Turns. With a regular -helix, the Cβ atom of the side chain is directed toward the N terminus. To examine how and where the effects of the pyrrolidine ring of proline or the side chain of other signature residues start to affect helical structure in a noncanonical region, we chose to examine regions of the transmembrane helix extending from three positions C-terminal to eight positions N-terminal of the signature residue. This length of helix completely spans all noncanonical segments and, in addition, allows the steric effects of the pyrrolidine ring or other side chains that influence residues N-terminal to a noncanonical element to be characterized. Including residues beyond either end of wide-turn regions enabled parameters arising from coordinates spanning three or four residues; for example, inter-residue dihedral angles to be shown for the entire noncanonical region. Plots of the different parameters of these extended segments, therefore, include for comparison the capping helical structures at both ends of the perturbed region; such plots reveal whether an -helical segment or another noncanonical segment is part of a region being investigated.

To orient noncanonical regions for comparison of structural features, the helical axis point for the signature residue at position i was placed at coordinates 0,0,0, and then the helical axis vector was rotated to lie along the positive x-axis (Supplemental Fig. 3). Having fixed the orientation of the helical axis vector (defined by the helical axis points for residues i and i + 1), the entire helical region was then rotated to allow the C_i atom to lie on the z-axis (that is, at coordinate 0,0,z, where z is equal to the radius of the segment from i to i + 1); this orientation was designated the standard orientation.

	Results and Discussion

Top Abstract Materials and Methods Results and Discussion Conclusion References

 
More than 3000 individual transmembrane helical regions from 244 crystal structures deposited in the Protein Data Base (Berman et al., 2000) through August 2007 were examined. The dataset was restricted to crystal structures with resolution 3.0 Å and an R value 0.3. This set comprised 31 different proteins, including multichain protein complexes, with 350 transmembrane helices containing 70 wide turns. Although there is some uncertainty in the atom positions of structures with a resolution > 2.2 Å but < 3.0 Å, we included such structures, because, as noted by Senes et al. (2001), extended helical structures have an inbuilt degree of self-correction because the backbone atom positions have a greater degree of accuracy. NMR structures were considered, but were not included, with the exception detailed below, because the structures determined do not necessarily represent the lowest energy state of all segments of transmembrane helices. As a result, deviations in C=O and N-H bond orientations may limit the formation of potential backbone H-bonds.

Identification and Classification of Wide Turn Subclasses. Wide turns are characterized by the widening of a turn of a helix N-terminal to the signature residue or by the unwinding and lateral displacement of a two- to three-residue section of a helix (Fig. 1). As a result, the widened helix contains 4.4 residues per turn compared with 3.6 for an -helix. This is evident, for example, in the H chain of the photosynthetic reaction centers of Rhodobacter sphaeroides and Thermophilus tepidum, each with a single -helical transmembrane-spanning region. Alignment of these regions with the analogous region in the evolutionarily related organism, Blastochloris viridis, reveals the insertion of a single leucine residue (Leu30) in this transmembrane region, which, in contrast to the R. sphaeroides and T. tepidum structures, is a helix that now contains a wide turn (Fig. 2A). The opposite may also be the case (that is, deletion of a residue in an evolutionarily conserved transmembrane segment results in the switch from a wide turn to an -helix) (Fig. 2B).

View larger version (43K): [in this window] [in a new window]   Fig. 3. CiCx distance-difference profiles and H-bonding patterns for W1, W2, and W3 wide turns. A-C, CiCx distance-difference plots. Color code, difference between the distance between the C atom of any residue, n, in the noncanonical regions indicated and that of residues at 2 (n-2), 3 (n-3), 4 (n-4), 5 (n-5) positions N-terminal, and the comparable distances in an -helix, are indicated by the blue, green, red, and yellow traces, respectively. A, W1, E. coli Acrb multidrug efflux pump, helix 6, signature residue: Pro490. B, W2, BCCO, G chain, helix 1, signature residue: Pro26. C, W3, BCCO, A chain, helix 2: signature residue: Pro72. D, plots in A-C combined to allow comparison of differences in profiles. W1 wide turn (——); W2 wide turn (- -); W3 wide turn (- - -). The colors of the flat rectangles on the x-axis correspond to the colors of the helical axis vector for each segment (see Fig. 5). E-G, H-bonding interactions in the proteins indicated in A-C, respectively, determined as described in Riek et al. (2001). Blue rectangles represent the amide hydrogen of the residue indicated, red rectangles represent the carbonyl oxygen, and pink rectangles denote prolyl residues. The amino acid sequences are indicated in the single letter code and are offset (from top to bottom) by four residues in keeping with the formation of n to n-4 H-bonds in an -helix. Numbers above and below indicate sequence numbers. H-bond type: solid line (vertical): i i-4; dashed line (sloping from upper right to lower left): i i-5. Bond length (indicated by the numbers on the connecting lines) color code: brown, <2.0 Å; red, <2.2 Å; orange, <2.4 Å; orange-yellow, <2.5 Å; mustard yellow, <2.6 Å; green, < 2.7 Å; blue, < 2.8 Å; cyan, < 3.0 Å. H, different helical paths of the above W1, W2, and W3 wide turns. The segments have been aligned such that the signature residues at position S0 are coincident.

View larger version (36K): [in this window] [in a new window]   Fig. 5. Position and direction of NiNi+1, CiCi+1, and CiCi+1 helical axis vectors over the noncanonical region for proline-induced W1 (yellow), W2 (pink), and W3 (blue) wide turns. (Note: the axis vectors of an -helix would form a straight line.) An example from each wide turn subclass was placed in the "standard" orientation and combined to form a single display. The CS0CS+1 reference segment vector and the preceding eight N-terminal CiCi+1 vectors are shown along with the backbone N, C, and C atoms. Examples: W1 subclass induced by Pro490 in helix 6 of E. coli Acrb multidrug efflux pump (yellow); W2 subclass induced by Pro26 in helix 1 of the G chain of BCCO (pink); W3 subclass induced by Pro72 in helix 2 of the A chain of BCCO (blue). A, view along the z-axis. B, view along the x-axis. C, view along the y-axis. The N, C, and C backbone atoms indicate the path of the helix and show the backbone at the S-4 position looping out around the position (arrow) of the pyrrolidine ring (not shown). D, a view similar to B, with the backbone atoms removed and the S0 position highlighted by a white circle. The extent of the change in the vectors is indicated by the lighter coloring of the S-8 vectors. The W1 wide turn shows the greatest change in helical direction, the W2 wide turn a smaller divergence, and the W3 minimal change in direction. E and F, helical axis vectors of the same noncanonical segments as in A-D but determined based on the N and C atoms. The orientations are the same as shown in A and C, above, with the N and C vectors in E labeled showing position N-terminal to S0. The orientations of the NiNi+1 and CiCi+1 axis vectors are similar to the CiCi+1 vectors with the exception of the CS-4CS-3 and NS-3NS-2 axis vectors, which show a reversed direction and indicate segments with a left-handed orientation. Color code: S0 or reference vector, light (C series) and dark red (N series); S-1, orange; S-2, yellow; S-3, green; S-4, cyan; S-5, blue; S-6, purple; S-7, dark pink; S-8, gray.

As shown in Table 3, the range of and values at the S_-5 to S_-3 positions for the different wide-turn subclasses over-lap and cannot be used to distinguish among them. Likewise, as noted by Fodje and Al-Karadaghi (2002; see Figure 2a therein), although a number of points lie outside the "preferred" region in the Ramachandran - plot (Lovell et al., 2003) for the residues at the 5 position, (equivalent to the S_-3 position in our study), a comparable number lie within the "preferred" region. Indeed, analysis of the / dihedral angles lead us to use the term "wide turn" instead of "-bulge," given that if the individual pairs of averaged and dihedral angles determined for the S_-5 to S_-2 positions (Table 3) are used to construct regular helices, none form a -helix (that is, a helix with 4.4 residues per turn and i i-5 H-bonding). In fact, the different and combinations for the S_-3 position all produce either a left-handed helix or a circular helical path with a rise per residue of 0 Å—a structural impossibility. Likewise, the and values for position 5 (S_-3) (Table 3 of Fodje and Al-Karadaghi, 2002) also produce a circular structure. Most of these structures cannot form because of severe steric clashes between atoms on the adjacent turns. In addition, no and angle combinations yield a helix with H-bonding as the major H-bond type. Although a -bulge appears to mimic a segment of -helix, the backbone dihedral angles indicate that it forms a structure with less than perfect -helical character. Thus, and dihedral angles, per se, can be used neither to characterize wide turns nor to discriminate between different noncanonical types.

View this table: [in this window] [in a new window]   TABLE 3 and dihedral angles ± 1 S.D. for the different subclasses of wide turns 0 marks the position of the signature residue. Negative values indicate positions N-terminal to the signature residue. Position 3 in the helix-C region in the results of Cartaillier and Luecke (2004) is equivalent to the position of the signature residue. In the Fodje and Al-Karadaghi study (2002), position + 1 corresponds to the signature residue position.

In several instances of wide turns, the number of H-bonds indicated by H-bonding diagrams is fewer than the number expected based on the subclass indicated by the C_iC_xdistance-difference profile. One such example is the wide turn induced by Pro215 in helix 5 of bovine rhodopsin (Protein Data Base code 1U19). This has a C_iC_x distance-difference profile typical of the W₃ subclass (Fig. 4A), but the H-bonding diagram indicates only two rather than three H-bonds (Fig. 4B). This is due to the combination of a angle at position 211 that is less negative (53° versus -40°) than for an -helix and a angle that is more negative (-180° versus -65°) at position 212. This forces the His211 C=O bond to deviate from its normal direction by 75° (Fig. 4C). This marked deviation of the 211 C=O bond is associated with a coincident reorientation of the H-N bond at position 212, which positions its hydrogen atom below the 110° cutoff angle (H-N···O) required for the formation of an S_-3S_-8 H-bond. It is clear from a consideration of such aberrant wide turns that they could well be missed if their identification was based merely on the number of H-bonds. In contrast, our method of detecting wide turns based on C positions, which are the least likely to be perturbed even with wide turns that show the most marked structural deviation from that of an -helix, is much more robust.

View larger version (33K): [in this window] [in a new window]   Fig. 4. AW3 wide turn with fewer than predicted H-bonds. A, the CiCx distance-difference plot of a noncanonical region spanning Met207 to Leu216 in helix 5 of bovine rhodopsin indicates a W3 wide turn based on the width of the negative deflections (red and yellow traces). B, H-bonding diagram indicating that the H-bond between Met207:O and Phe212:HN does not form. The absent H-bond between Met207:O and Phe212:HN is indicated by the gray dotted line. C, helical structure of the noncanonical region containing the wide turn induced by Pro215 shows the skewed orientation of the His211 C=O bond as a result of its dihedral angle being markedly less negative (+53°) than that of an -helix (-40°). This also produces a coincident reorientation of the H-N bond at position 212, leading to the distance between Met207:O and Phe212:HN (4.1Å) being too great to allow an H-bond to form. In addition, the orientation of the Phe212 HN bond, directed toward the helical axis, results in an N-H···O angle of 60°—much less than the minimum 110° required for an H-bond.

Influence of Wide-Turn Subclasses on Helical Architecture. Residues in the vicinity of wide-turn signature residues can also have a modifying effect on helical structure. To evaluate such effects, we examined noncanonical regions on a residue-to-residue basis. To this end, we determined the changes in direction of the helical axis vector at each position (Fig. 5) and examined the effects on the different wide turns arising from sequence and rotamer differences. By comparing the differences in direction (helical bend angle) and position between the reference axis vector and the vector for the S_-7-S_-6 segment (the last -helical segment before the beginning of a noncanonical region), it is possible to measure the different wide turns (Fig. 5). When the helical axis vectors determined between adjacent N, C, or C groups of atoms of the peptide backbone are displayed concurrently, the effect of each backbone dihedral angle on helical structure and direction is revealed (Fig. 5, A, E, and F).

Analysis of the helical bend angles shows differences between wide-turn subclasses but also within subclasses. Except for the W₃ subclass, proline-induced wide turns show greater bend angles, and a narrower range of bend angles, than do the wide turns induced by nonproline amino acids (Table 4). With the diversity of side-chain structure for the nonproline amino acids, the wider range in bend angles is not unexpected. Although the signature residue, particularly if proline, is the major disruptive influence on the integrity of -helical structure, residues within the wide-turn region also have an effect on the direction of the helix. This is seen with the shifts in position and direction of the axis vectors for the intermediate segments compared with the reference segment vector (Fig. 5, E and F). These shifts indicate that the change in helical direction is not limited to a single point on the helix but is the result of cumulative changes occurring over a number of segments. Figure 5, A-C, shows the paths of the N-terminal helical segments of a proline-induced wide turn of each subclass and the extent of the shift in helical axis. It is evident from Fig. 5 that the W₁ subclass produces the largest shift in the helical axis in the region immediately N-terminal to the helical perturbation. The path traced by the axis vectors shows that for proline-induced wide turns, the backbone loops around the pyrrolidine ring and then bends away from the proline.

With proline-induced wide turns, the pyrrolidine ring and the formation of a single H-bond act as the equivalent of inserting a "wedge" into the helix, forcing it to bend away from the side containing the proline. With each additional H-bond formed in the W₂ and W₃ wide turns, another wedge is inserted into the helix but at an orientation equivalent to the position of the additional H-bond. The effect or "shape" of the wedge will be dependent upon the effects of the residues within the wide turn. Given that W₃ wide turns have three or more H-bonds, the combined effect can offset the effect of each individual H-bond, resulting in negligible change in helical direction.

Aromatic, β-branched, and large-volume residues predominate in the eight positions N-terminal to the signature residue in proline-induced wide turns, whereas glycine, alanine, serine, and threonine are usually found at two or more positions within this region in wide turns with a nonproline residue at the signature position. W₃ wide turns have a greater proportion of large volume residues compared with the W₁ and W₂ subclasses. Examples from a wider range of transmembrane protein families are needed before residue patterns characterizing each subgroup can be determined. Although wide turns are found throughout the transmembrane region, the majority of available examples lie toward the C-terminal region. However, there were a number of N-terminal region wide turns that were not included in the survey because they were either incomplete, in that they occurred in the last turn of the helix where the partial wide turn transitioned into coil-like or other nonhelical geometries before the wide turn was completed, or the remaining N-terminal portion of the helix was one turn or less in length.

Although structures solved by NMR do not provide consistent H-bonding patterns to allow confident delineation of wide-turn subclasses, the C_iC_x distance-difference profiles can be used to study the structure of the helical transmembrane segments. Analysis of the NMR-derived structures of PufX (Tunnicliffe et al., 2006), a single transmembrane-spanning protein organizing the photosynthetic reaction center (RC)-light harvesting complex of R. sphaeroides revealed up to four regions displaying profiles with wide turn characteristics (Supplemental Fig. 4). Calculation of the mean and standard deviations of the C_iC_x distance differences determined for these various NMR structures, at each position, revealed some segments that are relatively fixed and others where the helical architecture is flexible, varying between wide turn and almost -helical geometries [e.g., the segment encompassing residues 34 to 37] (Fig. 6). In a number of purple bacteria, the photosynthetic RC is surrounded by a ring of light-harvesting β polypeptide chains (LH1). The RC reduces a quinone (Q_B) successively over a number of cycles, and when fully reduced, it moves from the RC through the membrane to an adjacent cytochrome bc₁ complex (Cogdell et al., 2006). In R. sphaeroides, PufX is proposed to fulfill two functions: to enable the formation of a LH1-RC dimer and to provide a gap in the LH1 ring to allow the delivery of Q_B to the bc₁ complex. The glycine zipper motifs present in PufX would stabilize dimer formation. Thus, it seems reasonable to suppose that flexibility in the transmembrane region of PufX is required to provide an opening within the bilayer for Q_B to move within the membrane, while at the same time maintaining its contacts with the LH1-RC dimer complex.

View larger version (36K): [in this window] [in a new window]   Fig. 6. Mean CiCx distance-differences (top) and their standard deviations (histogram, bottom) for the helical region of PufX. The mean CiCx distance-difference profiles show the presence of four regions with wide-turn characteristics at Gly30, Gly35, Gly40, and Phe47 in the transmembrane segment (Trp22 Gly52). As is evident from the standard deviation histograms, the wide turn at Phe47 shows very little variation in CiCx distance-differences compared with the other three wide turns. Consistency of the Phe47 wide-turn structure is probably the result of hydrophobic/aromatic interactions between Trp33 and Val37 and between Phe38 and Leu42. Color codes as indicated in Fig. 3.

An example of such a switch in the position of a wide turn is seen in helix 10, subunit I of bovine cytochrome c oxidase (BCCO) (Tsukihara et al., 2003; Muramoto et al., 2007). In the reduced state, the wide turn has a W₂ profile with Gly384 as the signature residue. In the oxidized form, however, a W₃ wide turn is present with the signature residue position now at Phe387 (Fig. 7). To allow this isomerization to a different signature residue, two residues within this region (Leu381 and Ser382) have had to change their rotamer conformations (Fig. 7). Moreover, the wide turn has introduced a region of flexibility that is not available with an -helical geometry, or with a tight turn or kink.

View larger version (31K): [in this window] [in a new window]   Fig. 7. Redox-dependent alteration in a wide turn from a W2 to W3 subclass. A, helical path of the oxidized (green) and reduced (orange) helix 10 (A chain of BCCO) showing W3 and W2 configurations, respectively, and associated with different rotameric conformations of Ser382. The different rotamers allow Ser382:H to act as an H-bond donor for binding to a pyrrole ring on heme A or to an OH moiety on the tail of heme A. B and C, H-bonding diagrams (see Fig. 3 for details) of the helix in A with Phe387 as the signature residue in the oxidized form (B), and Gly384 as the signature residue in the reduced form (C), which has changed to -helical structure in the region N-terminal to Ala385 and Val386.

View larger version (45K): [in this window] [in a new window]   Fig. 8. Modifying effects of nonsignature residues on helical direction. A, differences in the helical axis vectors for W1 wide turn-containing helices with proline or leucine at the S-2 position and threonine or proline, respectively, at the signature residue position. Green, sensory rhodopsin II (SR2), helix 5; orange, E. coli Acrb multidrug efflux pump, helix 6. The effective size of the pyrrolidine rings is indicated by the green and orange balls. With proline at the S0 position (orange), the N-terminal helical path moves out and around the ring with the helical path bending away. Proline at the S-2 position (green) has the N-terminal path unwinding to place the ring beyond the helical path. B, the extent of the displacement in A, is shown by the 2.0-Å lateral shift (SR2) in the axis vector (blue) of the S-5 segment N-terminal to the reference segment (red). By comparison, the lateral displacement of a proline-induced W1 wide turn lacking a prolyl residue at the S-2 position is only 1.1 Å. C, comparison of the effects of a proline at the S0 and S-2 positions on helical structure. The SR2 helix above is shown in green. The W1 wide turn in helix 6 of the E. coli Acrb multidrug efflux pump is shown in orange, and the path of an -helix is shown in gray. The structures are aligned with the prolines superimposed and indicated by the blue ball.

Structural Effects of Wide-Turn Subclasses. Wide turns alter the register of a face of a helical segment C-terminal to the turn, compared with the orientation of the face before the turn, and the extent of this shift differs between the subclasses. It is smallest for the W₁ subclass (45°-75°), intermediate for the W₂ subclass (60°-90°) and largest for the W₃ subclass (80°-120°). This provides flexibility in helical architecture, allowing interactions between residues on the transmembrane segment in question with those on an adjacent helix, and in particular, allows backbone-to-backbone interactions that would otherwise not be possible with a rigid -helical structure.

An example of two wide turns underlying such interhelical interactions is shown in Fig. 9. Here, a number of CH···O type H-bonds (predominantly CH···O) stabilize interactions between transmembrane helix 1 of the A chain of BCCO, involving a glycine zipper motif (GxxxGxxG) (MacKenzie et al., 1997; Russ and Engelman, 2000; Senes et al., 2000, 2004; Kim et al., 2005), G₂₃AWAGMVG₃₀, and helix 2, which is in an antiparallel orientation to helix 1 and contains two wide turns. Both wide turns belong to the W₃ subclass and produce <5° changes in helical direction. As is evident, the first wide turn of helix 2 orients the C=O bond of residue Phe68 such that it interacts via a CH···O H-bond with the H2 atom of Gly23 in helix 1. Likewise, the second wide turn orients the C=O bond of Gly76 in helix 2 such that it can interact via a CH···O H-bond with H2 of Gly16 in helix 1. These interactions would not be possible if helix 2 lacked these wide turns, because the C=O bonds of Phe68 and Gly76 would not be directed toward their respective H-bonding partner residues. Moreover, the widening of the helix at the positions of the two wide turns brings these residues to within H-bonding distance of the adjacent helix. Note that because Phe68 and Gly76 bond to glycine donors (Gly23 and Gly16), the CH···O bonds thus formed are stronger than usual CH···O bonds (Desiraju and Steiner, 1999; Vargas et al., 2000; Scheiner et al., 2001). The lack of side chains on these glycines also facilitates interhelical packing that sterically would not be possible if residues containing side chains occupied these positions (Fig. 9). Finally, because wide turns permit backbone-to-backbone interactions, the interacting helices pack more closely than would be possible with interactions involving side chain-to-backbone or side chain-to-side chain bonding, because the number of contacts, electrostatic or van der Waals, that can form with these latter interactions, is likely to be less. Other examples of a glycine zipper motif stabilizing interhelical contacts have been reported (Senes et al., 2000), but to our knowledge this is the first example of a glycine zipper interacting with a wide turn.

View larger version (47K): [in this window] [in a new window]   Fig. 9. Multiple interhelical interactions made possible by the presence of wide turns and glycyl residues in adjacent transmembrane helices. Multiple backbone-to-backbone C-H···O H-bonds (dashed lines) between helices 1 (orange) and 2 (green) of the A chain of BCCO are shown. Helix 1 contains a glycine (Gly23) that interacts with Phe68 located on a W3 wide turn in helix 2 induced by Pro72. Helix 1 also contains another glycine (Gly16) that interacts with both Gly76 and Gly77 located on another helix 2W3 wide turn induced by Asn80. In addition, backbone-to-side chain interhelical H-bonds are evident between His12:O and Asn80:Hβ1 (2.3Å), Gly23:O and Met69:H2 (2.5Å), Pro72:O and Tyr19:Hβ1 (2.7Å), and Asn80:O and His12:H1 (1.8Å). An additional side chain-to-side chain H-bond occurs between Thr17:O and Trp81:H1 (2.2Å). The intrahelical H-bond, Gly76:OAsn80:Hβ2 orients the Asn80 side chain, enabling the interhelical H-bond Asn80:OHis12:H1 to form.

With the W₃ turns involving glycyl donors, although both the H1 and H2 atoms can form H-bonds, their H-bonding roles are distinct. The H1 atom is involved in intrahelical H-bonding or interhelical H-bonding to a carbonyl oxygen on an adjacent helix that runs in a parallel orientation. In contrast, the H2 atom can only form interhelical H-bonds with a backbone carbonyl on an adjacent helix if it runs in antiparallel orientation. This orientation is required; a parallel orientation precludes H-bonding because both the C-H2 bond and the C=O bond are aligned in the same direction.

Binding of Cofactors, Ions, and Ligands. Wide turns have previously been noted to provide binding site for cofactors and ions (Fodje and Al-Karadaghi, 2002; Cartailler and Luecke, 2004). Several interesting examples are also evident from our analyses. For example, a W₂ non-proline-induced wide turn is evident in helix 10 of the A chain of BCCO that binds the cofactor, HEME-A. A W₂ wide turn (signature residue Gly801) is also seen in the absence of Ca²⁺ in helix 7 of rabbit muscle Ca²⁺ ATPase. However, in the presence of Ca²⁺, the wide turn structure is disrupted because residues previously involved in forming it now use their carbonyls to form a Ca²⁺ binding pocket. Finally, a proline-induced W₂ wide turn in helix 5 of bovine rhodopsin helps to correctly position residues that line the binding pocket for the β-ionone ring of retinal. Upon photoactivation and isomerization of the chromophore from 11-cis to all-trans retinal to form the LUMI rhodopsin photo-intermediate, the β-ionone ring is now oriented so that it can interact with the S_-4 carbonyl oxygen of the wide turn (Nakamichi and Okada, 2006).

With the recent release of the structure for another member of the G-protein-coupled superfamily, the β₂-adrenergic receptor (AR) (Cherezov et al., 2007), we can now identify regions within the transmembrane helices that differ from those of rhodopsin. Although the transmembrane domains (TM) in the two structures have similar orientations, there are both similarities and differences in their wide turns, the latter potentially being important for receptor activation by their distinctive ligands—biogenic amines in the case of the β₂-AR and 11-cis retinal in the case of rhodopsin. These include a wide turn in TM II of both receptors that in both instances is of the W₂ subtype, and one in TM V that is of the W₃ subtype in rhodopsin but is a W₂ wide turn in the β₂-AR. The β₂-AR also has a W₁ wide turn induced by Pro168 in TM IV, with the carbonyl bond of Thr164 at the S_-4 position of this wide turn being directed toward Ser203 on TM V. Although Thr164 and Ser203 are too widely separated to form an H-bond, it is of interest that Ser203 is critically involved in ligand binding, via an H-bond interaction with the m-hydroxyl of the catechol ring, as well as in receptor activation (Strader et al., 1989; Liapakis et al., 2000). Thus, Thr164 in the β₂-AR may be important for positioning the ligand to allow interaction between Ser203 and its m-hydroxyl. Rhodopsin has a proline-induced kink rather than a wide turn at the equivalent region in TM IV that, in contrast to the β₂-AR, orients the carbonyl bond of Ala168 (the residue equivalent to Thr164 in the β₂-AR) away from TM V (Supplemental Fig. 6).

In terms of the W₂ versus W₃ wide-turn difference in TM V of rhodopsin and the β₂-AR, this changes the shape of the transmembrane domain containing these noncanonical regions to alter the positioning and orientation of residues N-terminal to the signature residue, including not only Ser 203 at the S_-8 position, but Ser207 at the S_-4 position. Like Ser203, Ser207 is also involved in ligand binding (in this instance via an H-bond interaction with the catechol ring p-hydroxyl) and in receptor activation (Strader et al., 1989, Liapakis et al., 2000). Thus, in the absence of a crystal structure for the β₂-AR, it is evident that modeling its binding pocket based on the rhodopsins crystal structure would malposition this critical pair of serine residues, even if the same rotamer positions were maintained as those for Met207 and His211—the rhodopsin TM V residues equivalent to Ser203 and Ser207 in the β₂-AR, respectively.

	Conclusion

Top Abstract Materials and Methods Results and Discussion Conclusion References

 
Using methods to characterize noncanonical regions in transmembrane helices on a reside-to-residue basis, we have identified three different classes of wide turns based on C_iC_x distance difference profiles and the number and position of potential H-bonds. We also show that wide turns result from a series of discrete changes over a number of residues, equivalent to either the widening of a turn or an unwinding of the helix. The path of the helical axis vectors indicates that for proline-induced wide turns, the helix bends around the proline instead of away from the proline, as occurs with tight turns and kinks. The structure of the wide turn introduces flexibility in the helix, which affords unique opportunities for interactions with adjacent helices. With the tools developed, we can more accurately define helical structure and determine the effect of side-chain rotamer conformations, thus bringing us closer to being able to predict with greater accuracy, secondary and, to some degree, even tertiary structure, from primary sequence.

	Acknowledgements

 
We are most grateful to Drs. Merridee Wouters, Siiri Iismaa and Daniela Stock for reading the manuscript and for critical input.

	Footnotes

 
Supported in part by grant 354400 from the National Health and Medical Research Council, Australia.

ABBREVIATIONS: RC, reaction center; BCCO, bovine cytochrome c oxidase; AR, adrenergic receptor; TM, transmembrane domain.

The online version of this article (available at http://molpharm.aspetjournals.org) contains supplemental material.

Address correspondence to: Dr. Robert M. Graham, 384 Victoria Street, Darlinghurst, NSW 2010, Australia. E-mail: b.graham{at}victorchang.edu.au

	References

Top Abstract Materials and Methods Results and Discussion Conclusion References

 
Aqvist J (1986) A simple way to calculate the axis of an -helix. Comput Chem 10: 97-99.[CrossRef]

Bansal M, Kumar S, and Velavan R (2000) HELANAL: a program to characterize helix geometry in proteins. J Biomol Struct Dyn 17: 811-819.[Medline]

Barlow DJ and Thornton JM (1988) Helix geometry in proteins. J Mol Biol 201: 601-619.[CrossRef][Medline]

Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, and Bourne PE (2000) The Protein Data Bank. Nucleic Acids Res 28: 235-242.[Abstract/Free Full Text]

Bright JN and Sansom MSP (2003) The flexing/twirling helix: exploring the flexibility about molecular hinges formed by proline and glycine motifs in transmembrane helices. J Phys Chem B 107: 627-636.

Cartailler JP and Luecke H (2004) Structural and functional characterization of -bulges and other short intrahelical deformations. Structure 12: 133-144.[Medline]

Cherezov V, Rosenbaum DM, Hanson MA, Rasmussen SG, Thian FS, Kobilka TS, Choi HJ, Kuhn P, Weis WI, Kobilka BK, et al. (2007) High-resolution crystal structure of an engineered human beta-2 G protein-coupled receptor. Science 318: 1258-1265.[Abstract/Free Full Text]

Christopher JA, Swanson R, and Baldwin TO (1996) Algorithms for finding the axis of a helix: fast rotational and parametric least-squares methods. Comput Chem 20: 339-345.[CrossRef][Medline]

Cogdell RJ, Gall A, and Kohler J (2006) The architecture and function of the light-harvesting apparatus of purple bacteria: from single molecules to in vivo membranes. Q Rev Biophys 39: 227-234.[CrossRef][Medline]

Cordes FS, Bright JN, and Sansom MSP (2002) Proline-induced distortions of transmembrane helices. J Mol Biol 323: 951-960.[CrossRef][Medline]

Desiraju G and Steiner T (1999) Archetypes of the weak hydrogen bond—C-H···O and C-H···N interactions in organic and organometallic systems, in The Weak Hydrogen Bond in Structural Chemistry and Biology, pp 29-121, Oxford University Press, Oxford, UK.

Dunbrack RL Jr and Cohen FE (1997) Bayesian statistical analysis of protein sidechain rotamer preferences. Protein Sci 6: 1661-1681.[Medline]

Fodje MN and Al-Karadaghi S (2002) Occurrence, conformational features and amino acid propensities for the pi-helix. Protein Eng 15: 353-358.[Abstract/Free Full Text]

Frishman D and Argos P (1995) Knowledge-based protein secondary structure assignment. Proteins 23: 566-579.[CrossRef][Medline]

Kabasch W and Sander C (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22: 2577-2637.[CrossRef][Medline]

Kim S, Jeon T-J, Oberai A, Yang D, Schmidt JJ, and Bowie JU (2005) Transmembrane glycine zippers: Physiological and pathological roles in membrane proteins. Proc Natl Acad Sci U S A 102: 14278-14283.[Abstract/Free Full Text]

Labesse G, Colloc'h N, Pothier J, and Mornon JP (1997) P-SEA: a new efficient assignment of secondary structure from C alpha trace of proteins. Comput Appl Biosci 13: 291-295.[Abstract/Free Full Text]

Liapakis G, Ballesteros JA, Papachristou S, Chan WC, Chen X, Javitch JA (2000) The forgotten serines. A critical role for Ser-203^5.42 in ligand binding to and activation of the β2-adrenergic receptor. J Biol Chem 275: 37779-37788.[Abstract/Free Full Text]

Lopera JA, Sturgis JN, and Duneau JP (2005) Ptuba: a tool for the visualization of helix surfaces in proteins. J Mol Graph Model 23: 305-315.[CrossRef][Medline]

Lovell SC, Davis IW, Arendall WB 3rd, de Bakker PI, Word JM, Prisant MG, Richardson JS, and Richardson DC (2003) Structure validation by C geometry: , and Cβ deviation. Proteins 50: 437-450.[CrossRef][Medline]

Lovell SC, Word JM, Richardson JS, and Richardson DC (2000) The penultimate rotamer library. Proteins 40: 389-408.[CrossRef][Medline]

MacKenzie KR, Prestegard JH, and Engelman DM (1997) A transmembrane helix dimer: structure and implications. Science 276: 131-133.[Abstract/Free Full Text]

Mohapatra PK, Khamari A, and Raval MK (2004) A method for structural analysis of -helices of membrane proteins. J Mol Model 10: 393-398.[CrossRef][Medline]

Muramoto K, Hirata K, Shinzawa-Itoh K, Yoko-O S, Yamashita E, Aoyama H, Tsukihara T, and Yoshikawa S (2007) A histidine residue acting as a controlling site for dioxygen reduction and proton pumping by cytochrome c oxidase. Proc Natl Acad Sci U S A 104: 7881-7886.[Abstract/Free Full Text]

Nakamichi H and Okada T (2006) Local peptide movement in the photoreaction intermediate of rhodopsin. Proc Natl Acad Sci U S A 103: 12729-12734.[Abstract/Free Full Text]

Richards FM and Kundrot CE (1988) Identification of structural motifs from protein coordinate data: secondary structure and first-level supersecondary structure. Proteins 3: 71-84.[CrossRef][Medline]

Riek RP, Rigoutsos I, Novotny J, and Graham RM (2001) Non--helical elements modulate polytopic membrane protein architecture. J Mol Biol 306: 349-362.[CrossRef][Medline]

Rigoutsos I, Riek P, Graham RM, and Novotny J (2003) Structural details (kinks and non-alpha conformations) in transmembrane helices are intrahelically determined and can be predicted by sequence pattern descriptors. Nucleic Acids Res 31: 4625-4631.[Abstract/Free Full Text]

Russ WP and Engelman DM (2000) The GxxxG motif: a framework for transmembrane helix-helix association. J Mol Biol 296: 911-919.[CrossRef][Medline]

Scheiner S, Kat T, and Gu Y (2001) Strength of the C H·O hydrogen bond of amino acid residues. J Biol Chem 276: 9832-9837.[Abstract/Free Full Text]

Senes A, Engel DE, and DeGrado WF (2004) Folding of helical membrane proteins: the role of polar, GxxxG-like and proline motifs. Curr Opin Struct Biol 14: 465-479.[CrossRef][Medline]

Senes A, Gerstein M, and Engelman DM (2000) Statistical analysis of amino acid patterns in transmembrane helices: the GxxxG motif occurs frequently and in association with beta-branched residues at neighboring positions. J Mol Biol 296: 921-936.[CrossRef][Medline]

Senes A, Ubarretxena-Belandia I, and Engelman DM (2001) The C-H.O hydrogen bond: A determinant of stability and specificity in transmembrane helix interactions. Proc Natl Acad Sci U S A 98: 9056-9061.[Abstract/Free Full Text]

Strader CD, Candelore MR, Hill WS, Sigal IS, and Dixon RA (1989) Identification of two serine residues involved in agonist activation of the β-adrenergic receptor. J Biol Chem 264: 13572-13578.[Abstract/Free Full Text]

Sugeta H and Miyazawa T (1967) General method for calculating helical parameters of polymer chains from bond lengths, bond angles and internal-rotation angles. Biopolymers 5: 673-679.[CrossRef]

Tsukihara T, Shimokata K, Katayama Y, Shimada H, Muramoto K, Aoyama H, Mochizuki M, Shinzawa-Itoh K, Yamashita E, Yao M, et al. (2003) The low-spin heme of cytochrome c oxidase as the driving element of the proton-pumping process. Proc Natl Acad Sci U S A 100: 15304-15309.[Abstract/Free Full Text]

Tunnicliffe RB, Ratcliffe EC, Hunter CN, and Williamson MP (2006) The solution structure of the PufX polypeptide from Rhodobacter sphaeroides. FEBS Lett 580: 6967-6971.[CrossRef]

Vargas R, Garza J, Dixon DA, and Hay BP (2000) How strong is the C-H·O=C hydrogen bond? J Am Chem Soc 122: 4750-4755.[CrossRef]

Visiers I, Braunheim BB, and Weinstein H (2000) Prokink: a protocol for numerical evaluation of helix distortions by proline. Protein Engineering 13: 603-606.[Abstract/Free Full Text]

Weaver TM (2000) The pi-helix translates structure into function. Protein Sci 9: 201-206.[Medline]

This Article Abstract Full Text (PDF) Data Supplement All Versions of this Article: mol.107.043042v1 73/4/1092    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Riek, R. P. Articles by Graham, R. M. Search for Related Content PubMed PubMed Citation Articles by Riek, R. P. Articles by Graham, R. M.