Figure 5. (a) Change in relative orientation of the A sheet and F helix, shown by a superposition of the F-helix (fragment 2c) in native and cleaved α₁-antitrypsin. In the native structure the F-helix is coloured black and the A-sheet red. In the cleaved structure the F-helix coloured blue and the A-sheet green. (b) and (c) Packing of residues of the F-helix (fragment 2c). The interface between the F-helix an the A-sheet (dark blue) in (b) native and (c) cleaved α₁-antitrypsin. Phe147, Ile157, Asn158 (ball-and-stick representation), Val161, Thr165, Ile169 and Thr180 are coloured magenta; and Val185 green. In the native structure the backbones of Leu172 and Val173 make hydrogen bonds to Asn158. In the cleaved structure the Leu172 hydrogen bond is not present. Other, non-conserved, residues that interact with the A-sheet are coloured cyan. Fragment 2c contains several conserved residues: Phe147, Ile157, Asn158, Val161, Thr165, Gly167, Ile169 and Thr180. (b) and (c) show the packing of these residues in native and cleaved α₁-antitrypsin: Phe147 is sandwiched between the F-helix and the A-sheet and is in contact with the C^γ of Thr180. Ile157 and Val161 are in contact with the A-sheet below (e.g. Val185). Asn158 makes hydrogen bonds to main-chain atoms of the strand that links the F-helix to s3A. It is mutated in antithrombin Rouen VI Asn187→Asp [Bruce et al 1994]. Thr165, Gly167 and Ile169 are involved in making a turn between the F-helix and the strand connecting it to s3A. Thr180 is located at the bottom of this strand in contact with Phe147, and forms hydrogen bonds to the side-chain of Asp177, also located on the strand linking the F-helix and s3A.

Figure 2. Rigid fragments in the comparison between (a) native and (b) cleaved α₁-antitrypsin (see also Table 1). The largest rigid fragment, which acts as the “scaffold” in the S→R conformational change, is fragment 1a (light blue). It contains 144 residues, comprising helices A, G, and H, the B-sheet, the C-sheet and the top of s5A. Hinge regions common to fragments 1a and 1b are shown in magenta. The main-chain atoms of fragment 1a in native and cleaved α₁-antitrypsin can be superposed with a root-mean-square (r.m.s.) deviation of 0.58 Å/atom, over 144 residues. This is an extremely low value, comparable to the experimental error. Fragment 1b contains helices hA, hB, hC and M plus the loop C-terminal to helix I. It is shown in yellow except for hinge regions that it shares with other fragments. Hinge regions common to fragments 1b and 2d are shown in brown, the hinge region common to fragments 1b and 1c is shown in orange; and part of the A-helix, a hinge region common to fragments 1a and 1b — is shown in magenta. Fragment 1c contains s6A (dark green) which links fragments 1a and 1b and s5A (orange) which is a hinge region common to fragment 1c and 1b. Fragment 2a (red) contains only the top of s3A and the region C-terminal to it. In the comparison described here, between native and cleaved α₁-antitrypsin, it is the smallest of the fragments. However, in comparisons between other pairs of serpin structures, fragment 2a is larger, including the top of s2A. The top of s2A is poorly defined in the known crystal structures of native α₁-antitrypsin. Therefore we did not include it in our calculations. Fragment 2a links fragment 1a and fragment 2b. Fragment 2b (pink) consists of most of strands s1A, s2A and s3A. [Stein and Chothia 1991] reported that these three strands move along with the F-helix as a rigid unit. However in our comparison of native/cleaved α₁-antitrypsin, it is clear that the F-helix does change its orientation with respect to the A-sheet. Figure 5 shows the change in packing between the F-helix and the A-sheet in the native and cleaved structures. Fragment 2c (dark brown) contains the F-helix and part of the strand connecting the F-helix to s5A. The F-helix packs against, s1A, s2A and s3A and its movement is probably coupled to shifts in fragment 2b. However it does alter its orientation with respect to the A-sheet with small changes in packing, possibly as a result of the inserted reactive centre loop (s4A). Fragment 2d contains the D-helix (dark blue). The C-helix (in brown) and the base of the B-helix (brown) act as hinges common to this fragment and fragment 1b. Fragment 2e (light green) contains the E-helix, and parts of the loop flanking this helix. It contains Tyr138, which is hydrogen bonded to His93 on the D-helix. Though these are not conserved residues they are also present in antithrombin (where the mutations His120→Tyr and Tyr166→Cys: both result in dysfunction) and are usually Tyr and His in other serpins. Those portions of the structure not assigned to rigid fragments (e.g. the reactive centre loop) are in grey, as they are either regions that undergo plastic deformation during the conformational change, or are insufficiently well resolved to be included in the comparison (e.g. the loop at the top of the D-helix).

Figure 3. Sequence and secondary structure of the rigid fragments. The colours of the fragments are as in Figure 2. Regions of plastic deformation are shown as inverted (white on black) text.

Figure 4. Schematic analysis of relationships between fragment movements during the S→R transition in α₁-antitrypsin. Each fragment is coloured as in Figure 2(a) to (c) and the logical flow of conformational change is indicated by arrows. This Figure should be viewed in conjunction with Figure 2. Cleavage of the reactive centre produces a movement in fragment 2a as the new strand inserts, possibly triggered by a conformational change in Arg196 which is salt-bridged to Glu354 (P5). The initial steric clashes made by an inserting reactive centre loop involve the top of strand s3A (fragment 2a), and the top of strand s5A (which is the hinge between fragments 1b and 1c). Movement of these two fragments is transmitted to the rest of the serpin: s3A and s5A (fragments 2a and 1c) move apart from each other, to allow the insertion of the reactive centre loop. However there is also movement in fragment 1b, as s5A is a region of common substructure (or hinge) between fragments 1b and 1c. Further down the A-sheet, the bottom half of s1A, s2A and s3A (fragment 2b) moves away from s5A in order to allow the reactive centre loop to insert. Fragment 2a contains the top half of s3A, and fragment 2b the bottom half. It is logical to suppose that the initial contacts made by the inserting reactive centre are with the top half (fragment 2a). Therefore it is likely that movement in fragment 2a will cause movement in fragment 2b as they are physically connected through s3A. Furthermore fragment 2b is packed onto fragment 1b, which is also shifted in response to the inserting reactive centre loop via s5A. These results are further supported by the comparison between native α₁-antitrypsin and native antithrombin [Whisstock 1996], where insertion of the reactive loop to P14 (just two residues) causes a dramatic shift at the opposite end of the molecule with s1A, s2A and s3A shifting in response to the movement at the top of s3A, but not as a result of actual physical contact with the inserting reactive centre loop. The F-helix (fragment 2c) packs against fragment 2b, and their movements are coupled, though there is some shift in the F-helix relative to fragment 2b, and this is reflected by subtle changes in the packing at the interface between these two regions (Figure 5(b). Other movements arise by propagation of conformational change. They are caused not by direct contact with the reactive centre loop itself, but by contacts with other regions that are in contact with the reactive centre loop. The D and E-helices (fragments 2d and 2e) are in contact with many mobile elements and their movements are complex. The Figure shows fragments packed against these two regions which could induce or affect their movement. Movements of the D and E-helices must be critical to the overall mechanism of serpin action, because mutations in these regions can produce serpin dysfunction. It seems the movements of these two helices may be coordinated, because the D and E-helices remain packed against each other in both S and R states. Indeed, the D and E-helices must move to allow the A-sheet to split apart and the reactive centre loop to insert. The E-helix is packed against the loop between hI and s5A. A comparison between latent antithrombin and latent PAI-1 reveals that this loop determines the extent of movement of the E-helix, which in turns affects movement in s1A, s2A, s3A, hF and hD. Thus although latent antithrombin and latent PAI-1 have undergone the full S→R, transition, there are subtle differences in the positions of some secondary structural elements, which appear to be controlled by the extent to which hE can move. In antithrombin the D-helix can itself mediate conformational change. Binding of heparin pentasaccharide to the D-helix causes expulsion of the partially inserted reactive centre loop from the A-sheet, the reverse of the S→R transition.

Figure 6. Side-chain interactions in the breach and shutter. (a) The breach region in native α₁-antitrypsin. Trp194, Phe190 and Ile188 are coloured green; Tyr244 (at the top of the Figure), Met374, Leu383 and Phe384 cyan; and Leu51 magenta. Asp341 and Lys191 are shown in “ball-and-stick” representation, in red. Broken lines represent hydrogen bonds. (b) The breach region of cleaved α₁-antitrypsin. Trp194, Phe190 and Ile188 are green; Tyr244 (at the top of the Figure), Met374, Leu383 and Phe384 cyan; Phe51 magenta; and Thr345 red. Asp341, Glu346 and Lys191 are shown in ball-and-stick representation, in red. (c) The shutter region in native α₁-antitrypsin. Ser53, Pro54 and Ser56 are cyan; His334 and Asn186 green; and the side-chain of Ser116 red. (d) The shutter region in cleaved α₁-antitrypsin. Ser53, Pro54 and Ser56 are cyan; Ala350 and Met351 magenta; His334 and Asn 186 green; and the side-chain of Thr59 dark blue.

Figure 7. Native (black) and cleaved (red) α₁-antitrypsin superposed on fragment 1a (broken lines). The movements in the A-sheet caused by the insertion of the reactive centre loop (in cyan in native and dark blue in cleaved) can be clearly seen, as well as the shift in the B-helix at the bottom of the Figure. The reactive centre loop swings down around a point roughly at the top of the centre of the A sheet. (a) Front view, (b) side view.

Figure 8. Comparison of selected side-chain interactions in native (black) and cleaved (red) α₁-antitrypsin. (a) Conformational change in the loop at the top of s3A around Trp194 and rotation of side-chain of Phe190, to remain in the same position relative to fragment 1a during the S→R transition. The reactive centre loop of the native molecule is shown in magenta, and that of the cleaved in green. (b) “Substitution” of similar side-chains in packing of the A-sheet against fragment 1b. The native and cleaved structures were superposed on the B-helix (part of fragment 1b), at the left of the Figure, and two strands from the A-sheet that pack against it are shown at the right. s2A from the native structure (black) and s3A from the cleaved structure (red) occupy the same position in space relative to the B-helix, as do s1A from the native structure and s2A from the cleaved. In this Figure the superposed strands s2A (native)/s3A (cleaved) are nearer the viewer and s1A (native)/s2A (cleaved) are farther from the viewer. There is some conservation of the buried surface of the A-sheet achieved by deployment in the same region of space of structurally similar but non-homologous side-chains. Thus Leu118 (native) and Leu184 (cleaved) occupy similar positions in space; as do Asn116 (native)/Asn186 (cleaved) and Thr114 (native)/Ile188 (cleaved).

Table 1. Rigid substructures of native and cleaved α₁-antitrypsin, and the changes in their relative geometry