Av. Filikov et Tl. James, STRUCTURE-BASED DESIGN OF LIGANDS FOR PROTEIN BASIC DOMAINS - APPLICATION TO THE HIV-1 TAT PROTEIN, Journal of computer-aided molecular design, 12(3), 1998, pp. 229-240
A methodology has been developed for designing ligands to bind a flexi
ble basic protein domain where the structure of the domain is essentia
lly known. It is based on an empirical binding free energy function de
veloped for highly charged complexes and on Monte Carlo simulations in
internal coordinates with both the ligand and the receptor being flex
ible. HIV-1 encodes a transactivating regulatory protein called Tat. f
inding of the basic domain of Tat to TAR RNA is required for efficient
transcription of the viral genome. The structure of a biologically ac
tive peptide containing the Tat basic RNA-binding domain is available
from NMR studies. The goal of the current project is to design a ligan
d which will bind to that basic domain and potentially inhibit the TAR
-Tat interaction. The basic domain contains six arginine and two lysin
e residues. Our strategy was to design a ligand for arginine first and
then a superligand for the basic domain by joining arginine ligands w
ith a linker. Several possible arginine Ligands were obtained by searc
hing the Available Chemicals Directory with DOCK 3.5 software. Phytic
acid, which can potentially bind multiple arginines, was chosen as a b
uilding block for the superligand. Calorimetric binding studies of sev
eral compounds to methylguanidine and Arg-/Lys-containing peptides wer
e performed. The data were used to develop an empirical binding free e
nergy function for prediction of affinity of the ligands for the Tar b
asic domain. Modeling of the conformations of the complexes with both
the superligand and the basic domain being flexible has been carried o
ut via Biased Probability Monte Carlo (BPMC) simulations in internal c
oordinates (ICM 2.6 suite of programs). The simulations used parameter
s to ensure correct folding, i.e., consistent with the experimental NM
R structure of a 25-residue Tat peptide, from a random starting confor
mation. Superligands for the basic domain were designed by joining tog
ether two molecules of phytic acid with peptidic and peptidomimetic li
nkers. The Linkers were refined by varying the length and side chains
of the linking residues, carrying out BPMC simulations, and evaluation
of the binding foe energy for the best energy conformation. The disso
ciation constant of the best ligand designed is estimated to be in the
low- to mid-nanomolar range.