Potent and Highly Selective Human Immunodeficiency Virus Type 1

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 1711-1715, March 1993 Medical Sciences Potent and highly selective human immunodeficiency virus type 1 (HIV-1) inhibition by a series of a-anilinophenylacetamide derivatives targeted at HIV-1 reverse transcriptase RUDI PAUWELS*t, KOEN ANDRIES*, ZEGER DEBYSER*, PAUL VAN DAELE*, DOMINIQUE SCHOLS*, PAUL STOFFELS*, KAREN DE VREESE*, ROBERT WOESTENBORGHSI, ANNE-MIEKE VANDAMME*, CORNELUS G. M. JANSSENt, JEF ANNE*, GEERT CAUWENBERGHI, JAN DESMYTER*, JOZEF HEYKANTS*, MARCEL A. C. JANSSENt, ERIK DE CLERCQ*, AND PAUL A. J. JANSSEN* *Rega Institute for Medical Research, Katholieke Universiteit Leuven, B-3000 Leuven, Belgium; and *Janssen Research Foundation,. Turnhoutseweg 30, B-2340 Beerse, Belgium Communicated by Baruch S. Blumberg, September 8, 1992

ABSTRACT In vitro evaluation ofa large chemical library hydroxyethoxy)methyl]-6-(phenylthio)thymine (HEPT) de- of pharmacologically acceptable prototpe compounds in a rivatives (4), the dipyridodiazepinones (5), the pyridinones high-capacity, cellular-based screening system has led to the (6), and the bis(heteroaryl)piperazines (7). All of these com- discovery of another famlly of human m virus pounds interact allosterically with the RT-template/primer type 1 (HIV-1) inhibitors. Through optimiation of a lead complex (1, 2, 5-9). compound, several a-anilMnphenylacetamide (a-APA) deriv- Here we describe another chemical family of HIV-1- atives have been identified that inhibit the repUcation ofseveral specific RT inhibitors, the a-anilinophenylacetamide (a- HIV-1 strains (1B/LAI, RF, NDK, MN, HE) in a variety of APA) derivatives, which inhibit HIV-1 replication in vitro at host cell ypes at concentrations that are 10,000- to 100,000- nanomolar concentrations 10,000- to 100,000-fold lower than fold lower than their cytotoxic concentrations. The IC5. of the their cytotoxic concentrations. One of them, R 89439, has a-APA derivative R 89439 for HIV-1 cytopathicity in MT-4 good oral availability and favorable pharmacokinetics and cells was 13 nM. The median 90% inhibitory concentration can be easily synthesized on a large scale. (IC5o) in a variety of host cells was 50-100 uM. Although these a-APA derivatves are active against a tetrahydrohmkao [4,5,1-jk][1,4]benzodiazepin-2(1H)-thione-(TIBO)-reslstant MATERIALS AND METHODS HIV-1 strain, they do not inhibit replication of HIV-2 (strains Compounds. The a-APA derivatives were prepared from ROD and EHO) or simian immunodeficency virus (strins 2,6-dichlorobenzaldehyde, a substituted aniline, and sodium Mac251, mndGB1, and agm3). An HIV-1 strain containing the cyanide in acetic acid (10), followed by hydrolysis of the Tyrl1 -) Cys mutation in the reverse transcriptase region intermediate aminonitrile in sulfuric acid (11). The radiola- displayed reduced sensitivity. a-APA derivative R 89439 in- beled a-APA derivative R 18893 and its two enantiomers R hibited virion and recombiant reverse transciptase of HIV-1 87231 and R 87232 were synthesized by tritiation of3-bromo- but did not inhibit that of HIV-2. Reverse transcriptase inhi- 2,6-dichlorobenzylalcohol to 2,6-dichloro-3-tritiobenzylalco- bition depended upon the template/primer used. The rela- hol. Oxidation to the benzaldehyde, aminonitrile synthesis tively uncomplicated synthesis of R 89439, its potent anti- with 2-nitroaniline, and hydrolysis gave the 3H-labeled R HIV-1 activity, and its favorable pharmacokinetic profie make 18893 in a 40% radiochemical yield. Resolution by means of R 89439 a good candidate for clinical studies. HPLC over a chiral column (ChiralCel O.J.) afforded both enantiomers with enantiomeric excess >99%o (261 GBq/ As of 1992, the World Health Organisation (WHO) estimates nmol, radiochemical purity >99.0%6). that worldwide -10 million people are infected with the Viruses. The origin of the virus strains was as follows: IIIB human immunodeficiency virus (HIV). In particular, Africa; (LAI) and RF (R. C. Gallo, National Institutes of Health); Southeast Asia, and South America are now faced with a NDK (12); MN (Medical Research Council AIDS Directed rapidly growing number of HIV infections. In addition to Programme Reagent Project and the National Institute of preventive campaigns and treatment of opportunistic dis- Allergy and Infectious Diseases AIDS Research and Refer- eases, managing and counteracting the HIV/AIDS pandemic ence Reagent Program); HE, isolated in our laboratory from also requires effective antiviral therapy. Ideally, anti-HIV a Belgian AIDS patient; ROD (L. Montagnier, Pasteur In- drugs should be highly selective, have good oral availability stitute, Paris); EHO (13); simian immunodeficiency virus and favorable pharmacokinetics, and allow large-scale pro- (SIV)M=,251 (14); SIV., (15); SIV,mndGBl (16). The HIV-1 duction at reasonable costs to bring them within reach of mutant strains 13MB1 (Leu00 -) Ile) and 13CN1 (Tyrl81-* more of the world population. Cys) were isolated in our laboratory after serial passage ofthe We described previously a family of highly potent and IIIB/LAI strain (MT-4 cells) and the NDK strain (CEM) cells, selective reverse transcriptase (RT) inhibitors, the tetrahy- respectively, in the presence of TIBO R 82913 at 0.5 ,tg/ml droimidazo[4,5,1-jkl[1,4]benzodiazepin-2(1H)-one and (unpublished data). -thione (TIBO) derivatives that exhibit distinctive specificity Antiviral Assays. These assays have been described in for HIV-1 (1, 2) and are devoid of serious side effects upon more detail elsewhere (1, 17). Exponentially growing MT-4 administration in vivo (3). Other compounds have been reported, which, despite their different chemical structures, Abbreviations: HIV-1, human immunodeficiency virus type 1; exhibit a TIBO-like antiviral profile. These include the a-APA, a-anilinophenylacetamide; ddI, dideoxyinosine; TIBO, tet- [1-(2- rahydroimidazo[4,5,1-jkI[1,4]benzodiazepin-2(1H)-one and -thione; HEPT, [142-hydroxyethoxy)methyl]-6(phenylthio)thymine; RT, re- The publication costs of this article were defrayed in part by page charge verse transcriptase; p.i., postinfection; MTT, 3-(4,5-dimethylthia- payment. This article must therefore be hereby marked "advertisement" zol-2-yl)-2,5-oliphenyltetrazolium bromide. in accordance with 18 U.S.C. §1734 solely to indicate this fact. tTo whom reprint requests should be addressed. 1711 1712 Medical Sciences: Pauwels et al. Proc. Natl. Acad Sci. USA 90 (1993) cells were either infected with 100-500 CCID50 (1 CCID50 to parallel cultures in microtiter plates at different time points being the viral dose infective for 50% of the cell cultures) or post infection (p.i.) at concentrations -100-fold higher than mock-infected. Five (MT-4 cells) to 6 (MOLT-4) days after their IC50 values in the MT-4/MTT assay. HIV-1 p24 core infection the viability of mock- and HIV-infected cells was protein (p24) was determined 29 hr p.i. by a sandwich ELISA examined either microscopically by trypan blue exclusion (DuPont). (MOLT-4) or spectrophotometrically (MT-4, MOLT-4) by a RT Assays. In the exogenous RT assay (8), the reaction tetrazolium, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl- mixture (50 ,ld) contained 50 mM Tris HCl (pH 8.4), 10 mM tetrazolium bromide (MTT)-based method (17). Phytohe- MgCl2, 100 mM KCI, 2.2 mM dithiothreitol, 2.5 ,uM dGTP, magglutinin-stimulated peripheral blood lymphocytes were and 0.05% (wt/vol) Triton X-100. The templates poly(C) or infected with a concentrated HIV-1 [strain IIIB(LAI)J stock poly(dC) and the primer (dG)12 18 (Pharmacia) were used at 40 solution. Fresh medium without compound was added at day and 6 ,ug/ml, respectively. Recombinant HIV-1 RT p66/p66 4 after infection. HIV-1 p24 core protein in the cell culture (P. J. Barr, Chiron) and HIV-2 p68/p68 RT (S. Hughes, medium was quantified at 8 days after infection by an Frederick Cancer Research Facility) were used at concen- antigen-capture assay (DuPont). The inhibitory activity of the trations of 1.1 nM. Recombinant RT Tyr'81 -- Cys, Tyr181 -_ compounds on expression of HIV-1 proteins in CEM cells Ile, and Tyr188 Leu mutants were constructed by site- was determined at day 4 after infection by an indirect directed mutagenesis as described (19) and were used at 20 immunofluorescence-laser cytofluorometric (FACS) method nM. The vector pKRT2 containing the HIV-1 RT-encoding (18). All experiments were conducted at least twice, and final fragment of HIV-1 under control of trc promoter was ob- data are expressed as the median of the IC50 values. tained through the AIDS Research and Reference Reagent Time of Addition Experiment. Delineation of the drug- Program, Division of AIDS (National Institute of Allergy and sensitive phase was determined in MT-4 cells infected at a Infectious Diseases) (20). For kinetic studies, recombinant high multiplicity of infection (0.1-1) with HIV-1 (IIIB/LAI). p66/pSi (R. Bhikhabhai and B. Strandberg, University of After 60-min incubation at 37°C, unadsorbed virus was re- Uppsala) was used. In the endogenous RT assay, the reaction moved by three washing steps. Compounds were then added mixture (50 ,ul) consisted of 50 mM Tris-HCl (pH 8.4), 2.5 mM

H R2

0 NH

H R

Compound Mr Isomer R1 R2 IC50,t nM CC50,t AM SOi RT IC50,1 p&M Test R 15345 325.2 (+) OCH3 H 610 15 25 9 R 18893 340.2 (±) NO2 H 88 80 909 1 R 87231 340.2 (+) NO2 H 1,700 82 48 67 R 87232 340.2 (-) NO2 H 33 66 2,000 0.4 R 88703 337.2 (+) C(O)CH3 H 26 130 5,000 R 88976 337.2 (+) C(O)CH3 H 6,500 120 18 R 88977 337.2 (-) C(O)CH3 H 19 52 2,737 R 89439 351.2 (+) C(O)CH3 CH3 13 710 54,615 0.2 R 90384 351.2 (+) C(O)CH3 CH3 2,100 270 129 R 90385 351.2 (-) C(O)CH3 CH3 5 420 84,000 Reference AZT 267.2 0.5 3.5 7,000 ddC 211.2 190 86 453 ddI 236.2 4,800 1,000 208 PMEA 273.2 15,300 380 25 R 82150 287.4 44 550 12,500 R 82913 321.9 33 34 1,030 R 86183 321.9 5 140 28,000 E-EPU 306.4 48 200 4,167 L-697,661 352.2 22 320 14,545 Ro 318959 670.9 14 17 1,214 FIG. 1 Structure-activity relationship for inhibition of HIV-1 cytopathicity and cytotoxicity in MT-4 cells by a-APA derivatives. Reference compounds include the following: PMEA, 9-(2-phosphonylmethoxyethyl)adenine; R 82150, TIBO derivative (+)-(S)-4,5,6,7-tetrahydro-5- methyl-6-(3-methyl-2-butenyl)imidazo[4,5,1-jkl[l,4]benzodiazepin-2(1H)-thione; R 82913, TIBO derivative (+)-(S)-4,5,6,7-tetrahydro-9-chloro- 5-methyl-6-(3-methyl-2-butenyl)imidazo[4,5,1-jk][1,4]benzodiazepin-2(1H)-thione; R 86183, TIBO derivative (+)-(S)-4,5,6,7-tetrahydro-8- chloro-5-methyl-6-(3-methyl-2-butenyl)imidazo[4,5,1-jk][1,4]benzodiazepin-2(1H)-thione; E-EPU, HEPT derivative 5-ethyl-1-ethoxymethyl-6- (phenylthio)uracil; L-697,661, pyridinone derivative 3-1[(4,7-dichloro-1,3-benzoxazol-2-yl)methyl]amino)-5-ethyl-6-methylpyridin-2(1H)-one (Merck Sharp & Dohme), provided by M. Goldman; Ro 318959, HIV protease inhibitor QC-AsnPhe[CH(OH)CH2N]DIQ-NHtBu [where QC, quinoline-2-carbonyl; DIQ, (4as, 8as)-decahydro-3(S)-isoquinolinecarbonyl], provided by N. Roberts (Roche Products, Welwyn Garden City, U.K.) tIC50 in MT-4 cells 5 days p.i. (MTT procedure). tFifty percent cytotoxic concentration in MT-4 cells (MTT procedure). §Selectivity index or ratio of CC50 to IC50. lIC50 of HIV-1 virion RT activity with poly(C)-(dG)j2j8 and dGTP as the template/primer and substrate, respectively. Medical Sciences: Pauwels et al. Proc. Natl. Acad. Sci. USA 90 (1993) 1713

MgCl2, 100 mM KCI, 4 mM dithiothreitol, 30 mg of bovine assay conditions, dideoxyinosine (ddI) and 3'-azido-3'- serum albumin per ml, 0.5 mM EGTA, and 0.01% (wt/vol) deoxythymidine (AZT) had an IC50 of 4800 nM and 0.5 nM, Triton X-100. Of the four deoxynucleotides (dNTPs), three respectively (Fig. 1). were used at a saturating concentration of 100 AM, whereas Antiretroviral Activity Profile of a-APA Derivative R 89439. the tritium-labeled dGTP (Amersham) was used at a concen- Next we examined the possible role of different host cell tration of 2.5 ,uM. systems and the susceptibility of different retroviruses to R 89439, ddI, and AZT (Table 1). In phytohemagglutinin- RESULTS stimulated peripheral blood lymphocyte cultures, HIV-1 infection was monitored by measuring the viral p24 core Discovery of the Lead Compound. From a chemical library production in the cell supernatant at 8 days p.i. Under these of 90,000 entities we selected 2000 compounds that belonged conditions R 89439 achieved a median 50%1o inhibition at 14 to different chemical classes and that were without effect in nM (Table 1). R 89439 and R 18893 proved >90%o protective standard toxicological and pharmacological assays. These at "'100 and 600 nM, respectively. In the non-human T-lym- compounds were evaluated in a high-flux screening system photropic virus type I-transformed CD4+ T-cell line CEM, R for their cellular cytotoxicity and capacity to inhibit HIV-1 89439 inhibited HIV-1 antigen expression, as measured by [IIIg(LAI)] replication in MT-4 cells. This screening program FACS analysis at an IC50 of 4 nM (Table 1). led to the discovery of the a-APA derivative R 15345, which We subsequently investigated the sensitivity of a number inhibited HIV-1 replication by 50%o at a concentration of 0.6 ofHIV-1 strains to R 89439 inhibition in MT-4 cells (Table 1). ,uM (Fig. 1). Cytotoxicity was seen at a 25-fold higher The MN strain, which has a V3 loop typical for strains concentration. Through evaluation of structurally related circulating in North America and Europe, proved the most compounds we identified R 18893 as a more potent inhibitor sensitive; its IC50 was 6 nM. The Haitian strain RF, the HE of HIV-1 replication with an IC50 of 88 nM (Fig. 1). This strain (Belgium), and the Zairian strain NDK exhibited IC50 increased potency was not paralleled by higher cytotoxicity, values of 15, 30, and 22 nM, respectively. An HIV-1 strain as the selectivity index [ratio of the 50% cytotoxic concen- that was totally resistant to the TIBO derivative R 82913 tration (CCso) to IC50] was 909. A program of synthesis was (IC5o, 18 pM; strain 13MB1) was still sensitive to R 89439 subsequently initiated, aimed at enhancing the anti-HIV (IC5o, 44 nM). On the other hand, an HIV-1 strain carrying a potency. The HIV-1 iihibition of R 18893 appeared to be Tyr-181 -- Cys mutation (strain 13CN1) was nearly 500-fold stereospecific because the newly synthesized (-)-isomer R less sensitive than the prototype IIIB/LAI strain. No activity 87232 was 50-fold more potent (IC50, 33 nM) as compared was found against the HIV-2 strains ROD and EHO and the with the (+)-isomer R 87231 (IC50, 1700 nM) (Fig. 1). When SIV strains MAC251, mndGB1, and agm3. The latter strain is the nitro-substituent in the aniline moiety of the a-APA inhibited by TIBO compound R 86183 at an IC5o of2,M (21). structure was replaced by an acetyl group (R 88703), the The combined anti-HIV-1 activity of a-APA derivative R anti-HIV-1 activity increased (IC50, 26 nM). The potency was 89439 with either AZT, ddI, and the protease inhibitor Ro further optimized by introducing an additional methyl group 318959 proved to be synergistic (data not shown). in the 5-position of the aniline moiety. This a-APA deriva- Mechanism of Action Studies. The drug-sensitive phase in tive, R 89439, had an IC5o of 13 nM and a selectivity index of the HIV-1 replicative cycle was determined by a time of 54,614. The stereospecificity seen for the enantiomers of R addition experiment in which equipotent drug concentrations 18893 extended to these derivatives. The (-)-isomer of R (i.e., 100-fold IC50 concentrations) were added at different 89439, R 90385, was the most potent of the series; it had an times p.i. to MT4 cells (Fig. 2). Analysis ofp24 antigen levels IC50 of S nM and a selectivity index of 84,000. Under similar shortly after the end of the first replicative cycle indicated Table 1. Antiretroviral spectrum of a-APA derivative R 89439, ddI, and AZT R 89439 ddI AZT Cell IC50, CC50, ICso, CCso, ICso, CCso, Virus Strain line nM ,M SI nM AM SI nM ,M SI HIV-1 IIIB(LAI) MT-4t 13 710 54,615 4,800 1000 208 0.5 3.5 7,000 IIIB(LAI) PBL* 14 >300 >21,429 210 >250 >1,190 2.1 52 24,762 IIIB(LAI) CEM§ 4 >500 >125,000 1,400 760 543 1.1 38 34,545 RF MT-4 15 710 47,333 4,000 1000 250 5.2 3.5 673 NDK MT-4 22 710 32,273 14,000 1000 71 1.7 3.5 2,059 MN MT-4 6 710 118,333 1,300 1000 769 0.4 3.5 8,750 HE MT-4 30 710 23,667 9,400 1000 106 0.9 3.5 3,889 13MB1 MT-4 44 710 16,136 1,800 1000 556 0.6 3.5 5,833 13CN1 MT-4 5,900 710 120 6,800 1000 147 1.1 3.5 3,182 HIV-2 ROD MT-4 >710,000 710 <1 18,000 1000 56 2.6 3.5 1,346 EHO MT-4 >710,000 710 <1 14,000 1000 71 2.2 3.5 1,591 SIV mac2si MT-4 >710,000 710 <1 5,300 1000 189 1.8 3.5 1,944 mndGB MOLT41 >250,000 >250 <1 12,000 >250 >21 1.1 >4 >3,636 agm3 MOLT411 >250,000 >250 <1 11,000 >250 >23 1.1 >4 >3,636 Data for IC5o represent median values of at least two separate experiments. CC50, cytotoxic concentration; SI, selectivity index or ratio of CC5o to IC5o. Strains were as follows: RF, Haitian strain; NDK, Zairian strain; MN, HIV-1 strain with V3-loop typical for Europe/USA; HE, Belgian strain; 13MB1, HIV-1 strain resistant to TIBO R 82913, obtained after serial passage of IIIB/LAI in MT-4 cells and contains a Leul(X Ile mutation in RT; 13CN1, HIV-1 strain resistant to TIBO R 82913, obtained after serial passage ofNDK in CEM cells and contains a Tyr181 Cys mutation in RT; mndGB, derived from mandrills; agm3, derived from African green monkeys. tDetermined 5 days p.i. by the MTT procedure. tDetermined by measuring HIV-1 p24 core protein production 7 days p.i. Cytotoxicity was determined by the MTT procedure. §HIV-1 antigen expression was determined by FACS analysis. lData determined by the MTT procedure. 'iData determined by the trypan blue dye-exclusion method. 1714 Medical Sciences: Pauwels et al. Proc. Natl. Acad Sci. USA 90 (1993)

rium dialysis experiments, the same stereospecificity as well as the template-dependent pattern became apparent (data not shown). Also within a series of a-APA derivatives, a correlation was found when the IC50 values for inhibition ofHIV-1 04 replication were compared with the IC50 values for RT

C., inhibition (Fig. 1). Enzyme kinetic analysis of RT inhibition by R 89439 indicated that the effects were noncompetitive with respect to the substrate dGTP (Ki, 0.34 uM) and the template/primer poly(C) (dG)i2_j8 (data not shown). HIV-1 ._ RT mutants obtained by site-directed mutagenesis in which Cts Tyr'81 was replaced by either cysteine or isoleucine, displayed reduced sensitivity by, at least, a factor of 250 and 1250, respectively. Similarly, the Tyr88 Leu mutation also led to >1250-fold reduction in sensitivity (Table 2). 0 5 10 15 20 25 Pharmacokinetic Profile of a-APA Derivatives. To select a Time of compound addition, hr suitable candidate for clinical evaluation, four a-APA deriv- FIG. 2. Variation of p24 antigen levels in acutely infected MT-4 atives were tested in healthy male volunteers. The plasma cells (29 hr) with time of addition of compounds dextran sulfate (Mr levels of the active (-) enantiomers were determined by 5000) at 50 ,g/ml (-*-), aurintricarboxylic acid at 50 ,ug/ml enantioselective HPLC. The ratio of plasma levels, deter- (- -*- -), AZT at 0.05 ,g/ml (o), dideoxycytidine ddC at 5 ,ug/ml (o), mined at 8 hr after oral intake of 100 mg in solution, over the ddl at 100 &g/ml (A), TIBO R 82913 at 0.05 ,ug/ml (o), a-APA R 87303 IC50 values was the highest for R 89439 (Cp, 50 ng/ml; ratio, at 1 pg/ml (m), and HIV protease inhibitor Ro 318959 at 0.05 ug/ml 30). Previous studies have indicated that for the TIBO (). derivative R 82913 this ratio was =5 when a 120- to 200-mg that the a-APA derivative R 87303 (5-methyl-substituted dose was given i.v. (3). derivative of R 87231, IC50 in MT-4 cells (IIIB, MTT method, 20 nM) lost its protective capacity during a time frame that DISCUSSION coincided with that of TIBO compound R 82913-i.e., 6-10 The a-APA derivatives represent another chemical family of hr p.i. This period fell shortly after that ofother RT inhibitors, anti-HIV-1 agents with selectivity indexes up to 4 to 5 orders such as AZT and ddl, and clearly differed from those periods of magnitude and with antiviral potencies, demonstrated in sensitive to the adsorption inhibitors dextran sulfate and different host cell types, in the nanomolar range. These aurintricarboxylic acid and the protease inhibitor Ro 318959. agents exhibit stereospecificity toward inhibition of HIV-1 Treatment of acutely infected MT cells resulted in a dose- replication and RT activity. That this class of drugs is dependent inhibition ofHIV-1 DNA formation as determined targeted at the RT was further shown through (i) identifica- by PCR analysis (data not shown). In addition R 89439 did not tion of the drug-sensitive phase of the replicative cycle, mask the CD4 receptor on MT-4 cells, it did not prevent which corresponded to that of other RT inhibitors; (ii) the binding of HIV-1 particles to the host cells, and it did not concentration-dependent inhibition ofHIV-1 DNA formation affect syncytium formation between uninfected MOLT-4 in acutely infected target cells; (iii) the correlation between cells and persistently HIV-1-infected HuT-78 cells. In the cellular IC50 values and IC50 values for RT inhibition; and (iv) latter, HIV-1 progeny formation, as measured by p24 pro- the HIV-1 specificity found both in cell culture and at the RT duction, was not affected by 4 days of incubation with 30 ,uM level. All tested HIV-1 strains, originating from different R 89439. geographical regions, proved sensitive to a-APA inhibition; RT Inhibition by a-APA Derivatives. R 89439 inhibited a all tested HIV-2 or SIV strains proved insensitive to a-APA poly(A)-(dT)12_18- and a poly(C) (dG)12.18-driven RT reaction inhibition. This specific spectrum is reminiscent of the spec- by 50% at 3.1 and 0.1 AM, respectively (Table 2). A trum of the nonnucleoside RT inhibitors with TIBO-like poly(dC) (dG)12_18-directed polymerase reaction was inhib- properties [TIBO, HEPT, dipyridodiazepinones, pyridino- ited by 50% at 1.3 ,uM. The RT inhibition by a-APA deriv- nes, and bis(heteroaryl)piperazines]. The most potent atives was stereospecific because the (+)-isomer of R 18893 a-APAs also act as noncompetitive inhibitors of HIV-1 RT, (R 87231) was >150-fold less active (IC50, 67 ,AM) than the exhibit a similar template preference for poly(C), and seem to (-)-isomer R 87232 (IC50, 0.4 ,uM) (Fig. 1). When a tritium preferentially inhibit RNA-directed DNA polymerase activ- label was introduced in the latter three a-APA derivatives ity. Photoaffinity labeling studies (22), characterization of and their RT-binding properties were investigated in equilib- drug-resistant HIV-1 isolates obtained in cell culture (23), Table 2. Inhibition of HIV RT IC50, AM Template/primer RT R 89439 R 82150 R 86183 ddGTP AZT-TP Endogenous (viral RNA)t HIV-1 virion 0.2 0.2 0.05 0.01 0.02 Poly(A)-(dT)12_18 HIV-1 virion 3.1 10 >5.0 0.05 Poly(C)-(dG)12_1s HIV-1 virion 0.1 0.2 0.06 0.02 Poly(C)-(dG)12-18 HIV-1 p66/p66 0.2 0.3 0.06 0.04 Poly(dC)-(dG)12-18 HIV-1 p66/p66 1.3 12 0.5 0.004 Poly(C)'(dG)1218 HIV-1 p66/pSi Y181 - Ct >50 45 0.01 Poly(C)-(dG)12.18 HIV-1 p66/pSi yl81 I§ >250 >250 0.03 Poly(C).(dG)12_18 HIV-1 p66/pSi yl88 - LI >250 >250 0.01 Poly(C)-(dG)12_18 HIV-2 p68/p68 >250 >250 >250 0.06 AZT-TP, AZT-5'-triphosphate. tTritium-labeled dGTP (2.5 ,M) was used as the substrate. The other dNTPs were used at saturating concentrations (100 ,uM). Recombinant mutagenized HIV-1 RT with a Tyr181 -) Cys (t) or -* Ile (§) mutation or a Tyr188 -. Leu mutation (¶). Medical Sciences: Pauwels et al. Proc. Natl. Acad. Sci. USA 90 (1993) 1715 construction of chimeric HIV-1/HIV-2 RTs (24), and site- 4. Baba, M., Tanaka, H., De Clercq, E., Pauwels, R., Balzarini, directed mutagenesis of RT (19, 24) all point to the important J., Schols, D., Nakashima, H., Perno, C.-F., Walker, R. T. & role of amino acids Tyr"8' and Tyr'88 for interaction with the Miyasaka, T. (1989) Biochem. Biophys. Res. Commun. 165, TIBO-like compounds. These residues flank the highly con- 1375-1381. served "polymerase signature sequence" (YXDD), ofwhich 5. Merluzzi, V. J., Hargrave, K. D., Labadia, M., Grozinger, K., the first aspartic residue may well participate in catalysis, as Skoog, M., Wu, J. C., Shih, C.-K., Eckner, K., Hattox, S., Adams, J., Rosenthal, A. S., Faanes, R., Eckner, R. J., Koup, shown for the analogous Asp882 in Escherichia coli Klenow R. A. & Sullivan, J. L. (1990) Science 250, 1411-1413. fragment (25). Here we demonstrate, at the RT level as well 6. Goldman, M. E., Nunberg, J. H., O'Bnren, J. A., Quintero, as with an HIV-1 strain containing the Tyr"8" -3 Cys muta- J. C., Schleif, W. A., Freund, K. F., Gaul, S. L., Saari, W. S., tion, that Tyr181 and Tyrt88 are also crucial for interaction Wai, J. S., Hoffman, J. M., Anderson, P. S., Hupe, D. J., with a-APA derivatives. All TIBO-like compounds, includ- Emini, E. A. & Stem, A. M. (1991) Proc. Natl. Acad. Sci. USA ing a-APA derivatives, seem to have a common target on RT. 88, 6863-6867. On the other hand, the a-APA derivative R 89439 proved 7. Romero, D. L., Busso, M., Tan, C.-K., Reusser, F., Palmer, inactive against agm3, the first SIV strain shown to be J. R., Pope, S. M., Aristoff, P. A., Downey, K. M., So, A. G., susceptible to TIBO and HEPT inhibition (21, 26). Further- Resnick, L. & Tarpley, W. G. (1991) Proc. Natl. Acad. Sci. more, an HIV-1 strain resistant to TIBO, R 82913, that USA 88, 8806-8810. contains a Lys00 -- Ile mutation was still sensitive to R 8. Debyser, Z., Pauwels, R., Andries, K., Desmyter, J., Kukla, 89439, indicating that resistance to a particular TIBO-like M. J., Janssen, P. A. J. & Clercq, E. (1991) Proc. Natl. Acad. Sci. USA 88, 1451-1455. compound does not automatically lead to cross-resistance to 9. Debyser, Z., Pauwels, R., Andries, K., Desmyter, J., other nonnucleoside RT inhibitors. These observations sug- Engelborghs, Y., Janssen, P. A. J. & De Clercq, E. (1992) Mol. gest that inside the binding site for the TIBO-like compounds Pharmacol. 41, 203-208. ("TIBO-pocket"), which we previously indicated to be func- 10. Dimroth, K. & Aurich, H. G. (1965) Chem. Ber. 98, 3902-3906. tionally and possibly spatially related to the substrate-binding 11. Janssen Pharmaceutica (1992) Eur. Patent W092/00952; (1992) site (9), the nonnucleoside RT inhibitors, depending on their Chem. Abstr. 116, 214161. chemical structure, display quantitative and/or qualitative 12. Spire, B., Sire, J., Zachar, V., Rey, F., Barre-Sinoussi, F., differences in their interaction with the amino acids that Galibert, F., Hampe, A. & Chermann, J.-C. (1989) Gene 81, constitute this putative pocket and/or may induce different 275-284. conformational changes. This hypothesis could be confirmed 13. Rey, M., Krust, B., Laurent, A. G., Guetard, D., Montagnier, when the high-resolution structure of RT becomes available, L. & Hovanessian, A. G. (1989) Virology 173, 258-267. and these findings can be placed in a three-dimensional 14. Franchini, G., Gurgo, C., Guo, H. G., Gallo, R. C., Collalti, context. this E., Fargnoli, K. A., Hall, L. F., Wong-Staal, F. & Reitz, M. J., During preparation of manuscript two groups Jr. (1987) Nature (London) 328, 539-543. reported significant advances in the elucidation of the struc- 15. Baier, M., Garber, C., Muller, C., Cichutek, K. & Kurth, R. ture of RT p66/pSi heterodimer (27, 28). One of these (1990) Virology 176, 216-221. crystallographic studies showed that the nonnucleoside RT 16. Tsujimoto, H., Cooper, R. W., Kodama, T., Fukasawa, W., inhibitor Nevirapine binds in a pocket of the p66 subunit on Miura, T., Ohta, Y., Ishikawa, K., Nakai, M., Frost, E., top of a f-hairpin motif containing Asp185 and Asp186, two Roelants, G. E., Roffi, J. & Hayami, N. (1988) J. Virol. 62, residues that are just underneath the active site for polymer- 4044-4050. ization (28). The side chains of Tyr181 and Tyr'" also con- 17. Pauwels, R., Balzarini, J., Baba, M., Snoeck, R., Schols, D., tacted this inhibitor (28), explaining why their mutation led to Herdewijn, P., Desmyter, J. & De Clercq, E. (1988) J. Virol. resistance. Other amino acids, such as Lys'03 and the above- Methods 20, 309-321. 18. Pauwels, R., De Clercq, E., Desmyter, J., Balzarini, J., mentioned Leu'00, identified by characterization of HIV Goubau, P., Herdewijn, P., Vanderhaeghe, H. & Vandeputte, strains resistant to TIBO-like compounds, all seem to be part M. (1987) J. Virol. Methods 16, 171-185. of this pocket on p66. Having a totally different chemical 19. De Vreese, K., Debyser, Z., Vandamme, A., Pauwels, R., structure, the a-APA derivatives may, therefore, contribute Desmyter, J., De Clercq, E. & Anne, J. (1992) Virology 188, to a better understanding of the mechanism of action of the 900-904. TIBO-like compounds, which seems to involve a highly 20. D'Aquila, R. T. & Summers, W. L. (1989) J. AcquiredImmune vulnerable site on the RT that is now inhibited by at least six Defic. Syndr. 2, 579-587. different chemical families. The a-APA derivatives rank 21. Pauwels, R., Andries, K., Debyser, Z., Kukla, M. J., Schols, among the most potent and selective agents ofthese different D., Breslin, H. J., Woestenborghs, R., Desmyter, J., Janssen, M. A. C., De Clercq, E. & Janssen, P. A. J. (1993)Antimicrob. families. The relatively easy chemical synthesis of R 89439, Agents Chemother., in press. its good oral availability, and its favorable pharmacokinetics 22. Cohen, K. A., Hopkins, J., Ingraham, R. H., Pargellis, C., Wu, make R 89439 a valid candidate for clinical studies. J. C., Palladino, D. E. H., Kinkade, P., Warren, T. C., We thank Hilde Azijn, Barbara Van Remoortel, and Sonia Van Rogers, S., Adams, J., Farina, P. R. & Grob, P. M. (1991) J. Dooren for excellent technical assistance. Work at the Rega Institute Biol. Chem. 266, 14670-14674. was funded by the Janssen Research Foundation and also supported 23. Nunberg, J. H., Schleif, W. A., Boots, E. J., O'Brien, J. A., by the Belgian Fonds voor Geneeskundig Wetenschappelijk Onder- Quintero, J. C., Hoffman, J. M., Emini, E. A. & Goldman, M. zoek, the Belgian Geconcerteerde Onderzoeksacties, and the AIDS (1991) J. Virol. 65, 4887-4892. Basic Research Programme of the European Community. 24. Shih, C.-K., Rose, J. M., Hansen, G. L., Wu, J. C., Bacolla, A. & Griffin, J. A. (1991) Proc. Natl. Acad. Sci. USA 88, 1. Pauwels, R., Andries, K., Desmyter, J., Schols, D., Kukla, 9878-9882. M. J., Breslin, H. J., Raeymaeckers, A., Van Gelder, J., 25. Polesky, A., Steitz, T. A., Grindley, N. D. F. & Joyce, C. M. Woestenborghs, R., Heykants, J., Schellekens, K., Janssen, (1990) J. Biol. Chem. 265, 14579-14591. M. A. C., De Clercq, E. & Janssen, P. A. J. (1990) Nature 26. Debyser, Z., Vreese, K., Pauwels, R., Yamamoto, N., Anne, (London) 343, 470-474. J., De Clercq, E. & Desmyter, J. (1992) J. Gen. Virol. 73, 2. Pauwels, R., Andries, K., Desmyter, J., Kukla, M. J., Hey- 1799-1804. kants, J., De Clercq, E. & Janssen, P. A. J. (1990) in Design of 27. Arnold, E., Jacobo-Molina, A., Nanni, R. G., Williams, R. L., Anti-AIDS Drugs, ed. De Clercq, E. (Elsevier, Amsterdam), Lu, X., Ding, J., Clark, A. D., Zhang, A., Ferris, A. L., Clark, pp. 103-122. P., Hizi, A. & Hughes, S. H. (1992) Nature (London) 357, 3. Pialoux, G., Youle, M., Dupont, B., Gazzard, B., Cauwen- 85-89. bergh, G. F. M. J., Stoffels, P. A. M., Davies, S., De Saint 28. Kohlstaedt, L. A., Wang, J., Friedman, J. M., Rice, P. A. & Martin, J. & Janssen, P. A. J. (1991) Lancet 338, 140-143. Steiz, T. A. (1992) Science 256, 1783-1790.