Designing amino acid residues with single-conformations
Protein Engineering, Design & Selection vol. 19 no. 9 pp. 401–408, 2006
Published online June 24, 2006 doi:10.1093/protein/gzl024
Designing amino acid residues with
single-conformations
Tran T.Tran1, Herbert Treutlein1 and
Antony W.Burgess1,2
1
Ludwig Institute for Cancer Research and Cooperative Research Centre for
Cellular and Growth Factors, Melbourne, Victoria 3050, Australia
2
To whom correspondence should be addressed.
E-mail:
Introduction
Although a large number of chemically and conformationally
diverse peptides and proteins can be produced through different combinations of the 20 natural amino acids encoded
by DNA, the introduction of non-coded amino acids, such as
a-amino isobutyric acid (Aib) (Aubury et al., 1978; Paterson
et al., 1981; Prasad et al., 1984) or sarcosine (N-methylglycine) can enhance the diversity of synthetic peptides for
drug discovery. Apart from the increased diversity, non-coded
amino acids, or terminal blocking groups (Vasquez et al.,
1983), have been utilized in many ways: (i) to confer
resistance to enzymatic degradation (Mock et al., 1981;
Spatola, 1983; Michel et al., 1989), (ii) to probe the function
of particular residues in protein structures (Turk et al., 1975;
Juvvadi et al., 1992; Rutledge et al., 1996), (iii) for the
process of de novo peptide design (Marshall and Bosshard,
1972; Burgess et al., 1973; Prasad et al., 1984; DeGrado
et al., 1991; Karle et al., 1991; Ramnarayan et al., 1995),
(iv) for the design of structures with novel folds or secondary
structures (Appella et al., 1996; Seebach et al., 1996;
Che et al., 2006; Salaun et al., 2006), (v) to create new
combinatorial chemistry libraries (Simon et al., 1992; Kessler,
1993; Zuckermann et al., 1992, 1994; Miller et al., 1995;
Allinger et al., 1996) and (vi) to be used in de novo drug
Materials and methods
The conformational energy (f, c) maps for the N-acetyl-N0 methyl amino acid amides were obtained by minimization
with torsion forcing at every 10 or 15 increment of the f
and c dihedral angles using the discover program of MSI, the
CFF91 (Hwang et al., 1994; Maple et al., 1994) force field
(Maple et al., 1994) and our previously derived thioamide
force field parameters (Tran et al., 2001a,b,c,d). Energy
minimizations were performed using the artificial torsion
forcing function 4444[1 cos(V V0) Discover Force-Field
Manual (1995), www.accelrys.com], convergence criteria of
0.01 kcal/mol, no non-bonded cut-off, and a dielectric
constant of 80 (Tran et al., 2001c). For dipeptides containing
the penicillamine modification, the w torsional angles were
searched systematically to ensure the appropriate positioning
of the large side chain for each f, c conformer.
To identify amino acid/peptide combinations which favor
a single conformer, we determined all of the energy minima
on the (f, c) conformational energy map and identified
conformational minima within 2.0 kcal/mol of the global
minimum, which were separated from the other minima by an
energy barrier greater than kT (0.6 kcal/mol).
Almost 40 years ago, Ramakrishnan and Ramachandran
(1965) predicted the favorable areas of the (f, c) surface for
glycine and alanine dipeptides using a classical hard sphere
model and Mandel et al. (1977) used the ‘Derivation diagram’
to explain the steric contacts at each point on the f, c surface.
We have used similar diagrams to describe the pairwise
interactions influencing the conformational stability in particular regions of the (f, c) energy maps. As the f dihedral
angle of Ac-Ala-NHMe is rotated, the On1 and the Hn atoms
interact with three major atoms or groups of atoms, Cb, Hb
and Pn, where Pn is defined as the four atoms in the peptide,
Cn, On, Nn+1 and Hn+1. When the c dihedral angle is rotated,
the Hn+1 and the On atoms interact with Cb, Hb and Pn1,
The Author 2006. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail:
401
Drug design can benefit from the use of non-coded amino
acids, such as a-amino isobutyric acids (Aib) or sarcosine
(N-methyl-glycine). Non-coded amino acids can confer
resistance to enzymatic degradation and increase the conformational stability of the peptides. We have simulated the
conformational effects of combining N-methylation, bulky
groups on the Ca atom and/or thioamides using the class II
CFF91 force field and our thioamide force field parameters.
Although single amino acid substitutions (e.g. Aib) can
restrict the available conformations, they do not necessarily
lead to unique conformers, however, we predict that some
of the amino acids described in this report will fold to a
single f, w conformation (e.g. N-methylated and thioamide
penicillamine). Several other amino acid/thiopeptide
combinations were designed, which are predicted to prefer
only two conformations. Novel amino acids of this type
should prove useful for designing peptides with defined
conformations.
Keywords: conformationally restricted amino acids/
N-methylation/penicillamine/protein engineering/thioamide
design where a specific fit to a target surface is required
(Veber et al., 1979, 1981).
Novel amino acids often introduce extra substituents on the
peptide backbone: the amide nitrogen, the a-carbon or the C0 .
The effects of single additions or substitutions at these three
positions have been reported by others (Manavalan and
Momany, 1980; Spatola, 1983; Burgess, 1994; Mohle et al.,
1995; Sewald et al., 1995; Koksch et al., 1997; Valle et al.,
1988, 1989, 1991). We have previously derived force field
parameters for thioamides (Tran et al., 2001a,b,c,d, 2002),
which enable studies on the conformational and hydrogen
bonding effects of substituting S for O at the peptide carbonyl
group. The aim of this paper is to discover/design non-coded
amino acids that have single conformational states at
particular positions. This can be achieved by combining
thioamide substitution, N-methylation and addition of bulky
group onto the Ca atom.
T.T.Tran et al.
where Pn1 is defined as the peptide group containing the four
atoms, Nn, Hn, Cn1 and On1. These interactions are
displayed on the conventional (f, c) energy maps in Figure 2.
remain in the b-sheets regions: (75 , 135 ) and (135 , 75 )
(see Act-NMeAlat-NMe2 in Figure 2).
Results
The (f, c) conformational energy map for Act-Aib-NHMe is
similar to that of Ac-Aib-NHMe with six conformational
energy minima at: (60 , 45 ), (60 , 45 ), (180 , 75 ),
(180 , 75 ), (60 , 165 ) and (60 , 165 ) (Figure 3 and
Table 1). A search in the Cambridge Structural Database
(Allen et al., 1993) resulted in the Boc-Gly-Alat-Aib-OMe
(Jensen et al., 1985) as the only X-ray crystallographic
structure containing thio-peptide at the N-terminal of Aib.
The (f, c) dihedral angle for the Aib residue in BocGly-Alat-Aib-OMe is (53 , 42 ), which is close to the
global minimum (60 , 45 ) of the calculated (f, c) map for
Act-Aib-NHMe.
The conformations available to Ac-Aibt-NHMe and
Act-A (...truncated)
