De novo design and characterization of an apolar helical hairpin peptide at atomic resolution: Compaction mediated by weak interactions -- Data Supplement - Supplemental Data -- Proceedings of the National Academy of Sciences

Supplementary material for Ramagopal et al. (2001) Proc. Natl. Acad. Sci. USA 98 (3), 870–874. (10.1073/pnas.031442998)

a,b-Dehydro Residues
. a,b-Dehydro residues, which generally are represented by the symbol D are characterized by a double bond between a- and b-carbon atoms. This simple modification is seen to impart dramatic change to the conformational and biochemical properties of peptides containing them. These residues frequently are found in naturally occurring peptides from fungal and microbial metabolite origin. They are also constituents of polycyclic peptide antibiotics such as nisin, epidermin, subtilin, etc., which form a class of antibiotics known as lantibiotics.

a,b-Dehydrophenylalanine (DPhe) can occur in two isomeric forms, Z-isomer (D^ZPhe) and E-isomer (D^EPhe). In the former one the C=O group is in trans position with respect to the phenyl group, while in the later case it is in cis position (Fig. 6). Most of the conformational studies on dehydrophenylalanine residues are on D^ZPhe perhaps because all chemical synthetic procedures result in exclusively the D^ZPhe, the Z-isomer. In peptide HH21 all the DPhe residues are Z-isomers. DPhe has been demonstrated to be useful in modulating the biochemical properties of bioactive peptides as well as in the design of protein secondary structural motifs [for more details see the review article by Jain and Chauhan (1)].

Because of the modified hybridization sates of Ca and Cb atoms, DPhe gains unusual conformational characteristics compared to normal Phe; it becomes an achiral residue and all of its atoms including hydrogen atoms will be restricted to an approximate plane. Fig. 7 represents the slight deviation from planarity due to steric clash. This modification restricts the orientation of the DPhe side chain with respect to any secondary structure in which it is incorporated. Technically, DPhe with single rotamer and added planarity simplifies the three-dimensional complexity of the side-chain atomic positions to two dimensions. Implying, in a given backbone architecture, if one knows the coordinates for backbone atoms of DPhe, approximate coordinates for the side-chain atoms can be generated. This simplified geometry of DPhe was exploited in introducing long-range interactions between the two secondary structures, separated in sequence.

Observations, Its Exploitation, and the Result.
Some times small peptide structures reveal very important clues for de novo design. Structure of Boc-Ala-DPhe-DPhe-NHMe assumes both left-handed and right-handed 3₁₀-helical conformation in the solid state (2), where L-Ala assumes both positive and negative f, y values to accommodate itself in left-handed as well as right-handed 3₁₀-helical structures. DPhe residues from these shape-complement helices stack one above the other in a similar way as shown in Fig. 1. Similar stacking also was observed in a deca-peptide crystal structure, where the molecule assumes both left-handed and right-handed 3₁₀-helical conformation in solid state (3). Further, in proteins it has been found that 60% of the aromatic residues are involved in aromatic-aromatic pairing (4), suggesting preferred interactions among the aromatic residues. Based on these ideas we asked whether the concerted action of aromatic pairing is capable of driving the association of peptides containing both aromatic and aliphatic residues resulting in aromatic/aliphatic patterning? If so, can we exploit this aromatic/aliphatic patterning in design of super secondary structure without maintaining the binary patterning of polar/apolar amino acids? The structure of HH21 highlights such a possibility. Figs. 8 and 9 are given to support and to give better explanation for the design strategy. In Fig. 10, schematic representation of the structure of HH21 is given with all the atoms involved in weak interaction(s) at the helix-helix interior and at the turn region are labeled. Hydrogen bonds are represented by dotted lines. Parameters for hydrogen bonds and torsion angles in HH21 are given in Tables 3 and 4, respectively. Hydrophobic Stretches in Proteins, Similarity Between the Turn in HH21 and Proteins.
The present work assumes added significance in the view of numerous stretches of completely apolar residues, which are being increasingly noticed in proteins of fully sequenced genomes of different organisms. A striking example is a 239-residue stretch found in Epstein-Barr virus of length 641 residues (SWISSPROT code p03211), considering Gly as apolar residue, and a 30-aa hydrophobic stretch can be seen without residues Ala, Gly, and Pro (SWISSPROT code p04540). We also have searched for the turns in proteins, which are similar to that present in HH21 to find out the possibility of replacing the turn region with other natural residues. Backbone coordinates for five residues from DPhe⁸ to Gly¹² are used for the search. The program SPASM first generates five pseudo atoms corresponding to five residues and searches for similar motifs in proteins (5). Total number of hits with rms deviation 0.4, 0.3, 0.2 and 0.1 are 73, 41, 11, and 0, respectively (number of hits for a four-residue stretch from Ala⁹ to Gly¹² are 394, 175, 83, and 10, respectively, Protein DataBank release NOV-99, less than 25% homology). In Fig. 11, two examples of such turns are given and are superposed on the turn region in HH21 using the program InsightII, Molecular Simulations, Waltham, MA. rms deviations were recalculated for all the backbone atoms from DPhe⁸ to Gly¹² except for N, Ca of DPhe⁸ and C and O atoms of Gly¹². Because DPhe⁸ has trigonal geometry around Ca atom, only C=O moiety was used in superposition considering the fact that carbonyl oxygen of the DPhe⁸ accept 4 ® 1 type I b-turn hydrogen bond at the turn region of HH21. Further C=O moiety of the Gly¹² is not considered for the superposition because C-terminal secondary structure is different in different cases. In the two cases the deviations were 0.41 Å and 0.34 Å for the turns in proteins with Protein DataBank codes 1nxb and 1tca, respectively. The turn occurs at Asp³¹ to Thr³⁵ and Thr³¹⁰ to Ile³¹⁴ in proteins with their Protein DataBank codes 1nxb and 1tca, respectively (1nxb is a neurotoxin from a sea snake and 1tca is Triacylglycerol hydrolase). It is interesting to note that in 72 cases of 73 similar turns found in proteins with rms deviation less than 0.4 Å, Gly occupies the fourth position, which is expected considering the ability of Gly to take positive f, y values. It may be seen that the turn observed in the HH21 is compatible with the turns seen in proteins.

References:

1. Jain, R. M. & Chauhan, V.S. (1996) Biopolymers (Peptide Science) 40, 105-119.

2. Tuzi, A., Ciajolo, M. R., Guarino, G., Temussi, P. A., Fissi, A. & Pironi, O. (1993) Biopolymers 33, 1111-1121.

3 Ramagopal, U. A., Ramakumar, S. & Chauhan, V. S. (1999) Acta Crystallogr. A 55, Supplement, 376.

4. Burley, S. K. & Petsko, G. A. (1985) Science 229, 23-28.

5. Kleywegt, G. J. (1999) J. Mol. Biol. 285, 1887-1897.

Some of the important works on C-HL O and other weak interactions:

1. Taylor, R. & Kennard, O. (1982) J. Am. Chem. Soc. 104, 5063-5070.

2. Taylor, R. & Kennard, O. (1984) Acc. Chem. Res. 17, 320-326.

3. Wahl, M. & Sundaralingam, M. (1997) Trends Biochem. Sci. 22, 97-102.

4. Derewanda, Z. S., Lee, L. & Derewenda, U. (1995) J. Mol. Biol. 252, 248-262.

5. Fabiola, G. F., Krishnaswamy, S., Nagarajan, V. & Pattabi, B. (1997) Acta Crystallogr. D 53, 316-320.

6. Burley, S. K. & Petsko, G. A. (1988) Adv. Protein. Chem. 39, 125-189.