PPS 97' - Helix-Sheet Packing

Helix-Sheet Packing

Because the periodicity of an alpha-helix is different to that of a beta-sheet, a regular intercalated interface, such as that proposed between a pair of alpha-helices (refer to helix-helix packing), would not seem to be possible.

Earlier studies showed a discrepancy as to whether there is regular local intercalation between a single beta-sheet residue and its four surrounding helix residues i+1, i+4, i+5, i+8 as in Figs. 1 and 2 (Cohen et al., 1982) or no regular intercalation at all (Janin and Chothia, 1980). The latter point of view is of a largely flat surface with a few irregular intercalations.

The Complementary Twist Model

(Refer to Chothia, 1984)

The Interface Residues on the Alpha-Helix

The helix residues which pack against the beta-sheet tend to belong to two neighbouring ridges of the ±4n form (see previous section on helix-helix packing). This is illustrated in the diagrams (Figs. 1 and 2) below.

Fig. 1. The cylindrical surface of an alpha-helix unrolled and flattened; each circle represents a side chain. The N-terminal end is at the top. Two neigbouring ±4n ridges are highlighted in red and blue. After Chothia (1984)

Fig. 2. The same grooves highlighted in an alpha-helix.

To view these highlighted residues in an alpha helix, download this model helix(19Kb), then apply this RasMol script

Only three residues in each ridge are indicated, because this is approximately the extent of the interface with a typical beta-sheet. This is due to the twist of the sheet: the strands twist away from the helix residues situated further along the grooves.

The Interface Residues on the Beta-Sheet

The typical interface between an alpha-helix and a beta sheet involves three neighbouring strands of the sheet. This is indicated in Fig. 3 below. The central strand has three residues in contact with the helix, while the outer strands twist away from the helix in one direction. Thus, due to the overall twist of the beta sheet, there is a 'plateau' of residues diagonally spanning two corners.

Fig. 3 Three neighbouring strands of a beta-sheet forming an interface with an alpha-helix. The small black circles represent the side chains of residues on the lower surface, while the large circles are side chains on the upper surface. The shaded circles are those residues which, due to the twist of the sheet, form a raised surface extending from one upper (U) corner to the other; the remaining two residues are in a down (D) position. After Chothia (1984)

Fig. 4 The same residues in a beta-sheet (the positions highlighted are of alpha carbons). Three orthogonal views are shown, indicating the raised residues, resulting from the twist. This sheet is in the PDB structure 1ofv (flavodoxin; Luschinsky et al., 1991).

This may be viewed using RasMol by downloading the flavodoxin PDB file: 1ofv (86Kb) [Bbk|BNL|ExP|Waw|Hal] and then using this script to highlight beta-strands 1, 2 and 3. Then keep this window open (see below).

The Complementary Twist in a Real Protein

We will now examine a helix-sheet interface in the flavodoxin structure (1ofv) that you have just looked at (see above).

Three scripts are provided:

script to show the same sheet with the entire sidechains of the upper-surface residues. Notice how the sidechain conformations comprise a flat surface. Residue 53 is a glycine; the alpha carbon is shown as a spacefilling atom.
script to show the helix (helix 1) which packs against the three strands of sheet which you have previously viewed.
script to show the helix and sheet. Which alpha-helix side chains are involved in the interface in this case? Are any other ±4n ridges involved? Try illustrating this with RasMol.

Neighbouring alpha helices packing against beta-sheets

Consider two helices packing against each other and a beta-sheet. The typical twist of a beta-sheet, which means that the angle between neighbouring strands approximately 4.5Å apart is -19°, means that the two helix axes could not be parallel. The twist of the sheet implies that the expected angle between the two axes is about -40°. In the ridges-and-grooves model, the closest helix-helix angle to this value is the 52° between two helices with intercalating ±4n ridges. 80% of alpha-helices which pack onto neighbouring helices and a beta-sheet were found to exhibit this particular ridge-and-groove interaction with each other (Chothia, 1984). Note that when a ±4n ridge fits into the ±3n grooves of a neighbouring helix, the expected angle between the axes is 23° (refer to the page on helix-helix packing).

Beta-Strand Packing

Index to Section 9

References

Last updated 7th April '97