Supplementary Information for: A de novo peptide hexamer with a mutable channel. Walk, Bristol BS8 1TD, UK. UK.

Transcription

1 SI.1 Supplementary Information for: A de novo peptide hexamer with a mutable channel Nathan R. Zaccai, 1 Bertie Chi, 1,2 Andrew R. Thomson, 2 Aimee L. Boyle, 2 Gail J. Bartlett, 2 Marc Bruning, 2 Noah Linden, 3 Richard B. Sessions, 1 Paula J. Booth, 1 R. Leo Brady 1 and Derek N. Woolfson 1,2 1 School of Biochemistry, University of Bristol, Medical Sciences Building, University Walk, Bristol BS8 1TD, UK. 2 School of Chemistry, University of Bristol, Cantock s Close, Bristol BS8 1TS, UK. 3 Department of Mathematics, University of Bristol, University Walk, Bristol, BS8 1TW, UK. These authors contributed equally to this work.

2 SI.2 Supplementary Methods Bioinformatics Co-ordinates of the CC-Hex parent structure were submitted to the Secondary Structure Matching service (SSM) at EBI ( (Krissinel, E.; Kenrick, K. (2004) Secondary structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Cryst. D60, ) using standard options, with three exceptions: that individual chains were not allowed to be matched (to exclude single helix hits); that chain connectivity could be ignored (to identify any putative parallel/antiparallel matches); and that the lowest precision setting was used to identify as many close hits as possible. Superimpositions of the hits identified were calculated using the McLachlan algorithm (McLachlan, A.D. (1982) Rapid Comparison of Protein Structures, Acta Cryst A38, ) as implemented in the program ProFit (Martin, A.C.R.; Porter, C.T., using just C α co-ordinates. Modelling The molecular mechanics (MM) potential energy surfaces of the 13 models were mapped by varying the side-chain angles of residue 24 over a regular grid and calculating the Coulombic and van der Waals energies of each conformation. Firstly, the χ 1 and χ 2 side chain torsion angles of position 24 were analysed in the three mutant structures. The χ 2 angles differed significantly between structures (29 ± 46, defined as the angle between C α -C β vector and C β -O δ1 -O δ2 or C γ -N δ1 -C δ2 plane)) while the χ 1 angles were all similar (-80 ± 11). Inspection also showed that the only conformation available to residues in position 24 in the hexamer is g+ for steric reasons. The 13 models were constructed by superimposing copies of CC-Hex- D 24 Y 15 and CC-Hex-H 24 chains (chosen from the mixed crystal structure with residue 24 χ 1 angles closest to the average) onto the CC-Hex-D 24 structure. The χ 2 angles of position 24 were stepped through 24 increments giving 15 6 conformations per model. The lowest energy conformation of each model is plotted in Figure S5. The approach is validated by the fact that the experimental conformation of the mixed CC-Hex-D 24 Y 15 :CC-Hex-H 24 hexamer is located as the lowest energy out of the ~148 million searched. The molecular mechanics calculations were performed using the CVFF force field in Discover 2.98 (Accelrys).

3 SI.3 Analytical HPLC traces showing purified peptides as laballed. All chromatograms were collected on a Jasco J2100 series instrument, using a phenomenex prodigy column (100 x 4.6 mm), eluting with a linear gradient of water and acetonitrile (20 % to 80 % (CC-Tet and CC-Tet-Φ22), or 20 % to 60 % (rest)), each containing 0.1 % TFA at a flow rate of 1 ml/min. Black traces are for absorbance at 220 nm, and the red line for absorbance at 280 nm. CC-Tet CC-Tet-Φ22

4 SI.4 CC-Hex CC-Hex-Φ22

5 SI.5 CC-Hex-H24 CC-Hex-D24

6 SI.6 CC-Hex-D24Y15

7 SI.7 Supplementary Results Supplementary Table 1 X-ray crystal data collection, phasing and refinement statistics (SAD). *Highest-resolution shell is shown in parentheses. # The structures of CC-Tet-Φ22 and CC-Hex-Φ22 were refined against unmerged data ( F(+) and F(-) ). CC-Tet-Φ22 (3R4H) CC-Hex-Φ22 (3R3K) Data collection Space group P I Cell dimensions a, b, c (Å) 84.8, 84.8, , 54.5, α, β, γ ( ) 90.0, 90.0, , 90.0, 90.0 Resolution (Å) ( ) * ( ) R merge (0.871) (0.339) I / σi 17.0 (4.9) 9.3 (2.8) Completeness (%) 100 (100) 95.6 (72.1) Redundancy 43.9 (45.4) 5.7 (2.2) Refinement Resolution (Å) No. reflections # 11,164 10,181 R work / R free / / No. atoms Protein 1, Ligand/ion - / - 8 / 6 Water B-factors Protein (main chain / side chains) 53.2 / / 25.1 Ligand/ion - / / 37.4 Water R.M.S. deviations Bond lengths (Å) Bond angles ( )

8 SI.8 Supplementary Table 1 contd. X-ray crystal data collection, phasing and refinement statistics (molecular replacement). *Highest-resolution shell is shown in parentheses. CC-Tet (3R4A) CC-Hex-H24 (3R47) Data collection Space group I P 4 2 Cell dimensions a, b, c (Å) 44.7, 50.9, , 55.8, α, β, γ ( ) 90.0, 90.0, , 90.0, 90.0 Resolution (Å) ( )* ( ) R merge (0.624) (0.575) I / σi 12.8 (2.9) 23.6 (1.6) Completeness (%) 100 (99.9) 99.9 (100) Redundancy 6.7 (6.7) 7.2 (5.3) Refinement Resolution (Å) No. reflections 7,271 14,714 R work / R free / / No. atoms Protein 883 2,577 Solvent B-factors Protein (main chain / side chains) 27.8 / / 41.8 Solvent R.m.s. deviations Bond lengths (Å) Bond angles ( )

9 SI.9 Supplementary Table 1 contd. X-ray crystal data collection, phasing and refinement statistics (molecular replacement). *Highest-resolution shell is shown in parentheses. CC-Hex-D24 (3R46) CC-Hex-D24Y15: CC-Hex-H24 mix (3R48) Data collection Space group P C 2 Cell dimensions a, b, c (Å) 55.2, 55.2, , 50.9, 31.7 α, β, γ ( ) 90.0, 90.0, , 91.1, 90.0 Resolution (Å) ( )* ( ) R merge (0.576) (0.398) I / σi 37.9 (2.4) 16.0 (1.8) Completeness (%) 99 (90.3) 88.9 (46.2) Redundancy 23.7 (8.4) 6.3 (3.7) Refinement Resolution (Å) No. reflections 23,338 14,303 R work / R free / / No. atoms Protein 1,403 1,396 Ligand/ion 12 / 5 6 / - Water B-factors Protein (main chain / side chains) 19.2 / / 35.4 Ligand/ion 47.0 / / - Water R.m.s. deviations Bond lengths (Å) Bond angles ( )

10 SI.10 Supplementary Table 2: Coiled-coil parameters for the CC-Tet and CC-Hex peptides. CC-Tet-Φ 22 CC-Tet CC-Hex-Φ 22 CC-Hex-D 24 CC-Hex-H 24 CC-Hex-D 24 Y 15 : CC-Hex-H 24 mix Coiled-coil parameters 1 Radius (Å) 7.09 (0.05) 6.84 (0.06) 9.11 (0.08) 9.18 (0.05) 9.11 (0.04) 9.08 (0.09) Pitch (Å) (15.8) (12) (38.9) (26.9) (31.4) (31.9) 1 Calculated for the central two heptads using TWISTER (Strelkov & Burkhard, J. Struct. Biol. 137, 54 (2002)); numbers in parentheses are standard deviations.

11 SI.11 a b c Structures Helix-contact angle ( ) Discrepancy, ( ) n Idealised Observed Ideal geometry (2β) Observed (example) 1 (2β - 51) 2 (2β - 103) 3 (2β - 154) Pore diameter (Å) (2zta) N/A N/A N/A (1gcm) 90 (CC-Tet) 107 (1mz9) ± ± (CC-Hex) ± (2hy6) ±

12 SI (1ek9) ± Supplementary Figure 1: Theoretical and observed helix contacts as a function of oligomer state. (a) The heptameric GCN4-p1 mutant (Liu, J. et al. Proc. Natl. Acad. Sci. U. S. A. 103, (2006)), illustrating the staggered helix-helix interfaces. (b) Helical-wheel representation of a hexamer showing the angles and interactions described in the main text. (c) Geometrical analysis of helix-helix contacts in theoretical and observed coiled-coil and coiled-coil-like rings with 2 15 helices. In panel b, the angle α (coloured blue) = 360 /n, where n is the number of helices in the assembly; in this case, n = 6, and α = 60. This gives 2β (coloured red), by the relationship α, which is the ideal helix-packing angle in a regular geometrical arrangement, i.e. an equilateral triangle for trimer and so on. 2β can be compared with various helix-contact angles (coloured green) that would be predicted/preferred on the basis of offset double heptad repeats and helix geometry in higher-order coiled coils (J. Walshaw and D. N. Woolfson J. Struct. Biol 144, (2003)). From this, there are three possible hydrophobic seam offset angles of 51, 103 and 154. The example given is for two, offset hydrophobic (Φ) repeats, ΦxxxΦxx and xxxφxxφ within a conventional heptad assignment abcdefg, as observed in CC-Hex, which gives the 103 angle. In turn, this has a minimum discrepancy angle ( ) from an ideal geometry for a hexamer, see Panel c; i.e., in this case, 2β = 120, and = = 17. Panel c tabulates the angles defined above for ideal assemblies containing 2 15 helices. In addition, experimental contact angles were calculated from the named X-ray crystal structures using TWISTER (S. V. Strelkov and P. Burkhard, J. Struct. Biol. 137, 54-64, (2002)). The discrepancy parameters, 1, 2 and 3, represent the angular discrepancy between the geometrically ideal helix-contact angles and the three hydrophobic-seam angles. The lowest value of 1, 2 or 3 for a given 2β is highlighted in red. The final column gives measured pore diameters determined using PoreWalker (M. Pellegrini- Calace et al., PLOS Comp. Biol., 5, 1-16 (2009)). Structures shown in green are classical coiled coils with known examples; those coloured orange are complex coiled coils with known examples; those shaded grey have yet to be observed.

13 SI.13 Supplementary Figure 2: Small α-helical pores most similar to CC-Hex. Top-left, M2 pentamer (PDB 1eq8). Top-middle, Cartilage oligomeric matrix protein (COMP, PDB 1vdf). Top-right, tryptophan-zipper pentamer (PDB 1t8z). Bottom-left, CC-Hex. Bottom-middle, HTHP hexamer (PDB 2oyy). Bottom-right, mutant GCN4-p1 heptamer (PDB 2hy6). In each case the coiled-coil helices identified by SOCKET are coloured; and the side chains classified as knobs are shown as sticks and coloured by the colours of the rainbow, i.e., those assigned as a and d sites are red and green, respectively. The non-coiled-coil regions are coloured grey. All coiled-coil regions were identified using the default SOCKET packing cutoff of 7 Å, except M2, which required an 8.5 Å value before any KIH interactions were observed. Images created with PyMol (

14 SI.14 Supplementary Figure 3: Design, synthesis and characterisation of a peptide fragment from HTHP corresponding to coiled-coil hexamer region. (A & B) PyMol generated image of hexameric tyrosine haem protein (HTHP, PDB 2oyy) showing regions with knobs-into-holes interactions identified by SOCKET. Heptad positions a,b,c,d,e,f,g are colored red, orange, yellow, green, cyan, blue, and purple, respectively. (C) One chain of the hexamer showing the heme group bound to Tyr 43. (D)

15 SI.15 Region of the HTHP structure corresponding to the designed peptide. The residues highlighted are buried against the angled helices surrounding the outside of the hexamer-forming helices. (E) Sequences for the wild-type helical fragment (the SOCKET-assigned register is shown in coloured text, but is extended in grey for illustration). The designed model sequence was derived by modification of the hydrophobic residues at f positions to lysine to improve solubility and to isolate the putative hexameric coiled-coil interface. The model peptide was synthesised and purified using standard protocols as described above. (F) The CD spectrum of the model peptide (500 µm, PBS buffer, 5 C), which shows no significant α-helical content, consistent with lack of assembly.

16 SI.16 Supplementary Figure 4: ph-dependent folding of L24D, L24H and the 1:1 mixture of L24D:L24H in solution. Helicity as judged by the CD signal at 222 nm ([θ] 222 ) as a function of ph at 20 C. Key: L24D (diamonds), L24H (squares), experimental mixture (circles), and the calculated averages of the L24D and L24H data (broken black lines). The solid lines for L24D and the mixture are fits assuming a single pk a, which return the same pk a value of 5.2 in both cases. N.b., the parent CC- Hex did not show any appreciable changes in helicity or stability over this range of ph.

17 SI.17 Supplementary Figure 5: Thermal unfolding curves for the CC-Hex-D 24 and CC-Hex-H 24 mutants and the equimolar CC-Hex-D 24 :CC-Hex-H 24 mixture as a function of ph. Helicity ([θ] 222 ) versus temperature plots. Key: CC-Hex-D 24 (red), CC-Hex-H 24 (blue), experimental mixture (purple), and the calculated averages of the CC-Hex-D 24 and CC-Hex-H 24 data (broken black lines). In all cases, except for at ph 3.4, the experimental mixture is more thermally stable than that predicted by the simple average of data for the individual mutants.

18 SI.18 Supplementary Figure 5 continued.

19 SI.19 Supplementary Figure 6: Analytical ultracentrifugation of the CC-Hex mutants at ph 4.4 and 30,000 rpm. Representative sedimentation equilibrium curves for CC-Hex-D 24 (red diamonds), CC- Hex-H 24 (blue squares) and a 1:1 mixture of the two peptides (purple circles). Data recorded at ph 7.4 for the parent, CC-Hex (black crosses), are given as a reference. The CC-Hex data fitted to a singleideal species of 20,319 Da (95% confidence limits +119 and -111 Da), the monomer mass being 3,375 Da; those for the 1:1 mixture also fitted well to a single-ideal species of 19, 630 Da (95% confidence limits +260 and -269), the averaged monomer mass being 3,313 Da. Neither the data for the CC-Hex-D 24 nor the CC-Hex-H 24 peptides fitted well using single-ideal species model, both gave non-ideal residuals as shown in the upper panels; various monomer to n-mer models also failed to return good fits to the experimental data. Conditions and parameters are given in the Materials and Methods.

20 SI.20 Supplementary Figure 7: Modelled structures and relative internal energies of the 13 combinations of Asp and His at position 24 of CC-Hex structure. Top: Observed (green) and minumum energy models (mauve) for CC-Hex-D 24 (left), the,cc-hex-d 24 Y 15 :CC-Hex-H 24 mixture (center) and CC-Hex-H 24 (right). Bottom: relative internal energies (electrostatic plus van der Waals components) for the minimum-energy structures of the 13 combinations. Note that the energies are not free energies but represent the potential energies of the structures calculated in the gas phase with rigid geometry, hence the extremities of the graph indicate the unfavourable nature of packing too many like-charges together and the right hand side shows the extra steric repulsion of the larger histidine side chains which is relieved in reality by splaying the bundle. Key: red and blue discs represent the Asp and His-containing chains, respectively; thus, from left (Asp 6 ) to right (His 6 ) this plot shows variants with increasing numbers of His-containing chains.