ANTI-RETROVIRAL DRUG DESIGN BASED ON STRUCTURE ANALYSIS OF PROTEASES

¹Jindřich Hašek , ¹Jan Dohnálek , ¹Jarmila Dušková, ²Jan Konvalinka, ²Martin Hradilek, ²Milan Souček, ³Juraj Sedláček, ³Jiří Brynda, ⁴Eva Buchtelová

¹Institute of Macromolecular Chemistry Acad.Sci.CR, Heyrovského nám.2, 162 06 Praha 6
²Institute of Organic Chemistry and Biochemistry of Acad.Sci.CR, Flemingovo nám.2., 166 37 Praha 6
³Institute of Molecular Genetics Acad.Sci.CR, Flemingovo nám.2, 166 37 Praha 6
⁴Sincrotone Elettra, Padriciano 99, 34012 Trieste, Italy

Rational drug design based on experimental determination of enzyme-inhibitor bonding using x-ray diffraction is used in parallel with random screening of large numbers of potential inhibitors using combinatorial chemistry. In spite of more than 70 determined structures of HIV protease-inhibitor complexes of different types no satisfactory rules for the best inhibitor design were found. In addition, it seems that virtually all residues forming binding site of the HIV protease (8,23,81,82,84,150), (50,108,123,181,182,184), (28,29,30,32,47,84), (128,129,130,132,147,184) except of the catalytic aspartates 25 and 125 can be naturally mutated forming thus a new generations of viruses resistant against inhibitors originally applied [1].

Our project deals with experimental determination of HIV proteases complexed with ligands of different affinity resulting from screening. Monitoring the changes in interactions in sequences of inhibitors with protease we can follow some rules according to which new screening collections should be proposed. The series of 20 wide-spectral (with good mutational resistance) inhibitors denoted according to the scheme shown in the following table as SE, SQ, SI, RE, RQ, RI, OE OI, SRE, SRQ, SRI, SSE, SSQ, SSI, RSE, RSQ, RSI, RRE, RRQ, RRI is designed to explain [2,3]:

1. the influence of hydroxyl group originally supposed to replace the activated water molecule in the cleft site,
2. the influence of the (S/R) configuration of CH(OH) group, and
3. the influence of the charge, polarization and hydrophobicity of the residue in the P2' site.

Table 1. K_i values [picomol.l^-1] and shorthand notation for native HIV-1 protease complexed with 22 hydroxyethylamine, ethylamine and hydroxyethylene inhibitors of the general type Boc-Phe-[X]-Phe-P2'-Phe-NH₂ and C₉H₆NCO-Asn-Phe-[X]-NC₉H₁₄CO-NHC(CH₃)₃. Bold face numbers denote an area of the reasonably efficient inhibitors. Compound code (bold letters) consists of two parts: code for P3-P2-P1-[X]-P1’ and code for P2’-P3’. For example, the complex HIV-1 protease with Boc-Phe-[(R) CH(OH) CH₂ NH]-Phe-Glu-Phe-NH₂ is denoted as RE.

Table 1 summarizes K_i values [in picomol.l^-1] of native HIV-1 protease with 22 different inhibitors of the general type P3-P2-P1-[X]-P1'-P2'-P3’ belonging to the classes of hydroxyethylamine, ethylamine and hydroxyethylene inhibitors [4]. The table shows data for complexes native HIV-1 protease with 20 inhibitors synthetised in IOCB [2] and saquinavir - Ro-31-8959 ( C₉H₆NCO-Asn-P1-[X]-C₉NH₁₄CO-NH₂ C(CH₃)₃ ), here denoted as RoB. The groups related to S3, S2, S1, S1’,S2’ ,S3’ are:

P3 none in all cases except for saquinavir where two condensed rings C₉H₆N form the end of the chain.
P2 is Boc in all cases, except saquinavir where Asn is the P2 position.
P1 The function group at the P1 site is Phe in all cases.
[X] Hydroxyethylene (R or S), hydroxyethylamine (SR, SS, RS, RR) and ethylamine (Ea) spacers are used as isoster spacers.
P1' Hydrophobic groups Phe or DIQ (NC₉H₁₄) are available in both possible configurations R/S.
P2' The influence of four different residues (Glu, Gln, t-But, Ile) in the P2' position make difference in K_i from one to three orders in magnitude. The data for saquinavir (Ro-31-8959 ) are given in the eigth column.
P3´ is Phe-NH₂ in all cases except saquinavir, where the chain is terminated at P2’ by t-But.

The data for Saquinavir are given in the eigth line. Dash means that the experimental value of K_i is not available. The symbol* denotes structures that have been measured in the IMC, the symbol^a denotes structures under investigation, the symbol ⁺ denotes structures determined by a group of prof.Hingenfeld (to be published) and the symbol ^# refers to molecular structures determined by Krohn et al [5].

Crystal structure of the complex SQ clearly shows that the NH of P1’(Phe) forms hydrogen bridges with the catalytic aspartates D25, D125 and thus the hydroxy group does not play the role designed by hydroxyethylamine-inhibitor designers. It explains better inhibition properties of EoE complex (where CH(OH) group is replaced by CH₂) in comparison with SE and RE complexes. It seems that the interaction of hydroxy group need not be decisive for a design of the best inhibitor (notice that EoE belongs to the few best known inhibitors K_best= 50 pM ).

New structure determinations show that some rules for good inhibitors derived by different authors from structure and function correlation on limited number of structures are of limited validity. It seems that:

1. The best inhibitor need not replace the activated water molecule by hydroxyl group.
2. The configuration at the hydroxyl group need not be S (resembling L-amino acids) as it has already been confirmed with many inhibitors designed later.
3. The inhibitor need not possess two-fold symmetry.
4. Varying biochemical properties in P2’ need not lead to the monotoneous variation of the binding affinity because of possible conformational changes in other parts if inhibitor or protease.

The work is supported by A4050811 and 203/97/PO31.

1. Babime R.E., Bender S.L.: Molecular Recognition of Protein-Ligand Complexes: Applications to Drug Design. Chem.Rev.97, 1359-1472 (1997).
2. Konvalinka J., Litera J., Weber J., Vondrášek J., Hradilek M., Souček M., Pichová I., Majer P., Štrop P., Sedláček J., Heuser A.-M., Kottler H., Kräusslich H.-G.: Configurations of Diastereomeric Hydroxyethylene Isosteres Strongly Affect Biological Activities of Series of Specific Inhibitors of Human-Immunodeficiency -Virus Proteinase. Eur.J.Biochem.,250,559-566(1997).
3. Sedláček J., Fábry M., Hořejší M., Brynda J., Luftig R.B., Majer P.: A Rapid Screening Method for Biological Activity of Human Immunodeficiency Virus Proteinase Inhibitors by Using a Recombinant DNA-Derived Bacterial System. Anal.Biochem. 215, 306-309 (1993).
4. Dohnálek J., Hašek J., Brynda J., Fábry M., Sedláček J., Hradilek M., Konvalinka J., Souček M., Adams M.J.,
5. Naylor C.E.: Complex of HIV-1 Protease with (S)-Boc-Phe-y[CH(OH)CH2NH]-Phe-Glu-Phe-NH2. Protein Sci., 5, 167 (1996).
6. Krohn A., Redshaw S., Ritchie J.C., Graves B.J., Hatada M.H.: J.Med.Chem., 3340-3342 (1991).