Design and Synthesis of a Covalently Linked HIV-1 Dimer Analog and Peptidomimetic Inhibitors

Yoshiaki Kiso

Department of Medicinal Chemistry, Kyoto Pharmaceutical University,

Abstract: The HIV-1 protease analogs were synthesized by the solid-phase method and the dimer analog covalently linked by a disulfide bridge was constructed using the thioether chemical ligation method. The HIV-1 protease analogs effectively cleaved the Tyr-Phe-type , but had weak affinity to the Tyr-Pro-type substrate. Consequently, the molecular recognition of the analogs differs from the wild-type . Based on the substrate transition-state mimetic concept, allophenylnorstatine-containing HIV protease inhibitors were designed and synthesized. Among them, a dipeptide-based HIV protease inhibitor, KNI-764, exhibited potent antiviral activities, low cytotoxicity and good pharmacokinetic properties. Also, KNI-764 showed strong inhibition against both wild-type HIV-1 protease and synthetic HIV protease analogs, whereas saquinavir, ritonavir, and indinavir weakly inhibited the synthetic protease analogs. This study suggests that the small-sized dipeptide HIV protease inhibitor, KNI-764, is a good candidate for anti-HIV drugs

1. Introduction

Protection/deprotection steps and amide-forming reactions are important in synthesis. We have developed new methods and reagents such as reductive acidolysis (ref. 1), 2-(benzotriazole-1-y1)- oxy-1,3-dimethylimidazolidinium hexafluorophosphate (BOI) (ref. 2), and 2-chloro-1,3-dimethylimi- dazolidinium hexafluorophosphate (CIP) (Fig.1) (ref. 3). These new methods are successfully applied to the syntheses of biologically important . The modern synthetic methods (ref. 4) of peptide chemistry have greatly contributed to the fields important for human being, such as AIDS research.

Fig. 1. Structure of BOI and CIP.

896 ( 32 ) J . Synth . Org . Chem . , Jpn . Fig. 2. The overlapping gag/pol reading frame of HIV-1, the polyprotein translation products Pr55gag and Pr160gag-pol, and the proteolytic cleavage sites. (gag, group- specific antigen; pol, polymerase; TF, transframe ; PR, protease; RT, reverse transcriptase; RN, ribonuclease H; IN, integrase.).

The human immunodeficiency virus type-1 (HIV-1) codes for a virus-specific responsible for processing the gag and gag-pol polyproteins and for the proliferation of the retrovirus (Fig. 2). The HIV-1 protease, a 99 amino residues peptide, functions in homodimeric form which is associated non-covalently. The HIV-1 protease is a potential target for the development of anti-AIDS agents (ref. 5). Mam- malian aspartic such as and pepsin have two characteristic Asp-Thr-Gly sequences at the active center of the and both side-chain carboxyl groups are important in the catalysis of the peptide . In contrast, the retroviral protease has only one Asp-Thr-Gly sequence and functions as a homodimer. The HIV-1 protease can recognize Phe-Pro and Tyr-Pro sequences as the retrovirus-specific cleavage site, but mammalian aspartic proteases do not have such specificity. These features provided a basis for the rational design of selective HIV protease-targeted drugs for the treat- ment of AIDS and related complex.

2. Synthesis of HIV-1 Protease Analogs

HIV-1 protease is formed from two identical 99 peptides (Table 1) by means of hydro- philic and hydrophobic interactions, not by a disulfide bond. Kent et al. (ref. 6) synthesized an HIV-1 protease analog, in which the two Cys residues were replaced by the isosteric L-a-amino-n-butyric acid, whereas we replaced the two Cys residues by a natural amino acid, L-alanine. The [Ala67'95]derivative (1) of HIV-1 protease (NY-5 isolate) was synthesized by the solid-phase method based on the Boc-Bzl strategy using PAM resin (ref. 7). Each Boc-amino acid derivative was incorporated by the efficient

Vol . 56, No .11 (November 1998) ( 33 ) 897 Table 1. Amino acid sequences of HIV proteases and analogs .

method of the solid-phase peptide synthesis (SPPS) (Fig. 3)(refs 2, 4) , which consists of Nƒ¿-selective deprotection by dilute methanesulfonic acid, in situ neutralization, and rapid coupling reaction using BOI reagent. The final deprotection was performed by the HF method, and the product was purified by the usual method (ref. 6). The chemically synthesized [Ala67'95]-HIV- 1 protease had sufficient cleaving effect on the synthetic nonapeptide substrate, Ac-Arg-Ala-Ser-Gln-Asn-Tyr-Pro-Val-Val-NH2 (pH 5 .75, 37•Ž)(ref. 8). We examined the protease inhibitory activities of the peptide compounds containing the hydroxymethylcarbonyl (HMC) structure by the assay system using these enzyme analogs and the syn- thetic nonapeptide substrate.

Fig. 3. Coupling pathway of efficient SPPS employing in situ neutralization . DIEA = diisopropylethylamine.

898 ( 34 ) J. Synth. Org. Chem., Jpn .

Fig. 4. Synthesis of N-terminal fragment of HIV-1protease analog . MSA = methanesulfonic acid, DCM = dichioromethane, DOX = dioxane, HBTU = 2- (1H-benzotriazole-1 -yI)-oxy-1 ,1 ,3,3-tetramethyluronium hexafluorophosphate, DMF = dimethylformamide, DIEA = diisopropyl ethylamine.

2.1. Thioester Linkage Chemical Ligation

Furthermore, we synthesized [Tyr6,42,Nle36,46, (NHCH2COSCH2C0)51-52, Ala67,95]HIV-1 protease (2) (NY-5 isolate) using the thioester linkage chemical ligation method (ref. 9). The N-terminal portion was synthesized as shown in Figure 4, whereas the C-terminal portion was synthesized as shown in Figure 5 by the efficient Boc strategy of SPPS stated above. The chemical ligation reaction of [Tyr6,42, Nle36,46]-HIV-1 protease (1-50)-Gly-SH and BrCH2C0-[Ala67,95]-HIV-1 protease (53-99) was carried out

Fig. 5. Synthesis of C-terminal fragment of HIV-1 protease analog. DIPCDI = diisopropylcarbodiimide, HOBt = N1-hydroxybenzotriazole.

Vol.56, No.11 (November 1998) ( 35 ) 899 in pH 4.0 medium. The MALDI-TOF/MS value (m/z 10694.9) of the final product purified by high- performance liquid chromatography (HPLC) agreed well with the theoretical value. After folding, the ligated compound was found to exhibit comparable enzymatic activity to the recombinant wild-type HIV-1 protease using the synthetic nonapeptide substrate.

2.2. Thioether Linkage Chemical Ligation

Englebretsen et al. (ref. 10) reported a synthesis of an HIV-1 protease analog employing thioether forming ligation which was a reaction of peptide-mercaptoethylamide with bromoacetyl-segment. In the synthesis, they prepared the mercaptoamide segment using a thiol-releasing resin by acid treatments. However, this strategy sometimes incurred trouble due to the released thiol group. We have investigated various linkages releasing the thiol function by acid treatment, which contained substituted benzyl thioethers. However, large peptide segments could not be obtained in good purity using these linkages. We suspected that the results were due to the thiol involving side reactions and attempted to protect the thiol group during the acid treatment. In order to obtain the thiol segment without difficulty, we employed a new strategy using a multi- detachable linker to avoid contact of thiol with strong (ref. 11). An aminoethyldithio-2-isobutyric acid (AEDI), which can connect the mercaptoamide peptide to solid support by a disulfide bond, is stable under the conditions during Fmoc-based SPPS and acid treatments. We introduced AEDI on an acid-labile resin and used it as a multi-detachable linker. In this strategy, the peptide containing a masked thiol moiety was cleaved from the resin by acid deprotection, and then the thiol was generated by reduc- tion. A mercaptoethylamide peptide segment, HIV-1 protease (1-50)-NHCH2CH2SH, was prepared start- ing from an Fmoc-AEDI-O-Clt-resin (Fig. 6). The SPPS was achieved by ABI 431A synthesizer with standard DCC-HOBt protocol, and then the peptide segment containing the linker moiety was cleaved by HF-dimethyl sulfide-m-cresol (3:6:1)(0•Ž, 1h). The treatment of the resultant with dithiothreitol

(DTT) in 6M guanidine•HC1 (pH8) gave the desired product in good purity. The purification by RP-

Fig. 6. Synthetic scheme for HIV-1protease analog (3) using thioether forming ligation. Clt: 2-chiorotrityl

900 ( 36 ) J . Synth . Org . Chem . , Jpn

HPLC gave a homogeneous product in 29.9% overall yield, which was characterized by amino acid analysis of the hydrolyzate and MALDI-TOF MS. Another segment, BrCH2C0-[Ala67'95]HIV-1 protease (52-99) was also prepared by the usual SPPS, deprotection by HF-m-cresol and purification on RP-HPLC (8.0%). The ligation of the above two segments was carried out in 6M guanidine•HBr, 200mM Tris buffer

(pH8.5) at r.t. for 5h with vigorous stirring. The product was purified on Hiload 16/60 superdex 75pg in 8M urea and then gel-filtered to remove urea. The desired product, [NHCH2CH2SCH2CO51-52, Ala67,95]- HIV-1 protease (3) was obtained in 50% yield and characterized by MALDI-TOF MS (m/z: 10759.2•}0.1%, calcd: 10758.6). After folding, the synthetic protease analog possessed enzymatic activity, though it showed a slight difference in the recognition of substrates. The synthetic protease had comparable enzymatic activity to the wild-type enzyme using a substrate, Lys-Ala-Arg-Val-Tyr*Phe(NO2)-Glu-Ala-Nle-NH2. The Km values of the synthetic and wild-type proteases were 22.1 and 18.3 .tM, respectively. On the other hand, the protease analog had larger Km (23.5 mM) than the wild-type protease (6.46 mM) to Ser-Gln- Asn-Tyr*Pro-Ile-Val substrate. The non-peptide region at the flap of the synthetic protease may affect the recognition of the substrate.

2.3. A Covalently-linked HIV-1 Protease Dimer Analog

Interestingly, single-chain tethered dimers of HIV-1 proteases, artificial peptides having the same topology as pepsin-like typical aspartic proteases, also show similar enzymatic activity and are more stable than the natural enzyme (ref. 12). We synthesized an HIV-1 protease dimer analog which consists of two identical peptide chains covalently linked by a disulfide bridge (Fig. 7). In the synthesis of a protease dimer analog covalently linked by a disulfide bridge, we adopted the chemical ligation strategy forming a thioether for the preparation of the monomer unit. In order to make

Fig. 7. A dimer analog of HIVprotease covalently linked by disulfide bridge (4).

V01.56, No. 11 (November 1998) ( 37 ) 901 Fig. 8. Enzyme reaction mechanism of HIV protease.

a disulfide bridge between two protease monomers, we introduced a Cys residue as a substitute for Asn98, because Asn98 and Asn198 are located very close in active conformation (distance between two Cp of Asn is 5.2A). Cys residues at positions 67 and 95 having a free sulfhydryl group were replaced by Ala. This substitution did not influence the enzymatic activity as described above. A monomer unit, [NHCH2CH2SCH2CO5152, A1a67,95, Cys(Acm)98HIV-1 protease was prepared by the same method mentioned above. The ligation reaction proceeded within 4h and the product was easily purified by gel-filtration (yield: 55%). Then the disulfide formation was achieved using iodine

(r.t., 1h). The homogeneous dimer analog (4) was obtained by the purification on ODS-column (yield: 11%, MALDI-TOF MS; m/z=21494.8 •} 0.1%, calcd for [M+H]=21494.2). After the folding procedure, the synthetic peptide possessed proteolytic activity comparable to that of the recombinant wild-type HIV-1 protease in an assay system using KARVY*Phe(NO2)-EA-Nle-NH2 as a substrate (Km=10.611M). Similarly to the case of the monomer analog, this dimer analog had weak affinity (Km = 26.0 mM) to the Tyr-Pro-type substrate, SQNY*PIV. Consequently, the molecular recog- nition of the analogs differs from that of the wild-type enzyme. These enzyme analogs should be useful tools for evaluation of protease inhibitors.

3. Substrate-Based Peptidomimetic HIV Protease Inhibitors

The transition state of amide hydrolysis by an aspartic protease is proposed as illustrated in Fig. 8. The hydrogen bonds between the aspartic acid carboxyl groups of protease and the hydroxyl groups of the substrate transition state are significant in the design of tight-binding inhibitors. Based on the tran- sition-state mimic concept, we designed and synthesized a novel class of substrate-based HIV protease inhibitors containing an unnatural amino acid, (2S,3S)-3-amino-2-hydroxy-4-phenylbutyric acid, named allophenylnorstatine (Apns), with an HMC isostere (Fig.9). The stereochemistry of the hydroxyl group was significant for the enzyme inhibition (ref.13) and the HMC group interacted excellently with the aspartic acid carboxyl groups of the HIV protease in a similar manner as the transition state(ref. 14).

3.1. A Potent Tripeptide Inhibitor, KNI-272

Inhibition of HIV protease has become a major target for AIDS chemotherapy. Among Apns-

902 ( 38) J. Synth . Org . Chem . , Jpn

containing HIV protease inhibitors, the tripeptide KNI-272 was a highly selective and superpotent HIV protease inhibitor (Ki = 5.5 pM). KNI-272 exhibited potent in vitro and in vivo antiviral activities with low cytotoxicity (ref. 15, 16). The NMR, X-ray crystallography and molecular modeling studies showed that the HMC group in KNI-272 interacted excellently with the aspartic acid carboxyl groups of the HIV protease active site(Fig. 9) (ref. 17). These results imply that the HMC isostere is an ideal transition- state mimetic.

3.2. A Resistance Surmountable Inhibitor, KNI-241

HIV-1 develops in vitro a high level of resistance to KNI-272 by acquiring mutations in the pro- tease-encoding gene, though it takes a relatively long time. Mutations conferring KNI-272-resistance were V32I, L33F, K45I, F53L, A71V and I84V. The EC50 values of KNI-272 against wild-type HIV-1 and KNI-272-resistant HIV-1 were 0.04 iM and 2 1.04, respectively. The sensitivity of KNI-272 re- duced 50-fold against resistant HIV-1. In order to overcome the resistance, the Apns-containing pep- tides were screened. Among them, KNI-241 (Fig. 10) was found to be active against both wild-type HIV-1 (EC50=0.04 ƒÊM) and KNI-272-resistant HIV-1 (EC50=0.04 ƒÊM) (ref. 18). KNI-241 showed strong inhibition against both wild-type HIV-1 protease and synthetic HIV-1 protease analogs (3 and 4) (Fig.11), whereas KNI-272, saquinavir, ritonavir and indinavir weakly inhibited the enzyme analogs (ref. 19).

Fig. 9. Design of HIV protease inhibitors.

Vol.56, No.11 (November 1998) ( 39 ) 903 3.3. Small-Sized Dipeptide Inhibitors

The bioavailability of KNI-241 was very low. The P3 moiety of KNI-241 interacts with both 53- Phe of wild-type protease and 53-Leu of the mutant protease . The solution, crystalline and complex structure of KNI-272 were similar except for the P3 moiety (Fig. 12) (ref. 20). Therefore, we considered that the P2-P2' moiety might be the core group for enzyme inhibition and studied small-sized dipeptide inhibitors as advantageous compounds. P2 was replaced by several alkyl and hydrophilic groups. Among them, KNI-549 containing dimethyl and carboxyl groups exhibited HIV protease inhibitory activities (HIV protease inhibition = 76.3% at 50 nM).

Fig. 10. Substrate-based peptidomimetic HIV protease inhibitors .

904 ( 40 ) J . Synth . Org . Chem . , Jpn Fig. 11. HIV protease activity in the presence of 50nM inhibitors.

In order to enhance the inhibitory activity, we cyclized the P2 moiety in KNI-549 to obtain the rigid form. The resulting KNI-577 was a potent HIV protease inhibitor (87.6% inhibition at 50 nM) and showed remarkable anti-HIV-1 (MB) activity (EC50=0.02 in CEM-SS cells). This antiviral activity of KNI-577 was higher than that of the tripeptide KNI-272. KNI-577 had very low cytotoxicity (CC50•„

200 ƒÊM) and the bioavailability after intraduodenal administration in rats was 48% (ref . 21). Furthermore, our design took into consideration the symmetry of the HIV protease dimer and introduced a 2-methylbenzylamine group in P2' . The resulting KNI-764 exhibited highly potent en-

Fig. 12. Superposition of KNI-272 structures in solution(white), in single crystal(dark gray) and in complex (light gray).

Vol.56, No.11 (November 1998) ( 41 ) 905

Fig.13. Synthetic scheme of KNI-764.

zyme inhibition (Ki = 0.035 nM) and surprisingly potent anti-HIV activity (EC50 = 0.007 ƒÊ,M) with low cytotoxicity (CC50•„ 20ƒÊM). KNI-764 had a favorable oral bioavailability (42 %) and a long plasma half life (94 min) in dogs (ref. 22). KNI-764 showed strong inhibition against both wild-type HIV-1 protease and synthetic HIV-1 protease analogs (3 and 4). The synthetic scheme of KNI-764 is simple and short as shown in Fig. 13. All couplings were performed by N,N'-dicyclohexyl carbodiimide (DCC) /HOBt and all deprotections by HC1.

4. Conclusion

The hydroxymethylcarbonyl isostere was an ideal transition-state mimetic in HIV protease inhibi- tor design. Small-sized dipeptides containing the hydroxymethylcarbonyl isostere exhibited potent HIV protease inhibitory activities. Among them, a dipeptide-based HIV protease inhibitor, KNI-764, exhib- ited potent antiviral activities, low cytotoxicity and good pharmacokinetic properties. Also, KNI-764 showed strong inhibition against both wild-type HIV-1 protease and synthetic HIV-1 protease analogs (3 and 4), whereas saquinavir, ritonavir, and indinavir weakly inhibited the synthetic protease analogs. Synthetic HIV protease analogs were useful tools for evaluation of HIV protease inhibitors. This study suggests that the small-sized dipeptide HIV protease inhibitor, KNI-764, is a good candidate for anti- HIV drugs.

Acknowledgment

The author wishes to thank his many colleagues for performing much of the work described in this paper. This research was supported in part by grants from the Ministry of Education, Science and Cul- ture of Japan and Japan Health Science Foundation.

References (1) (a) Kiso, K.; Kimura, T.; Yoshida, M.; Shimokura, M.; Akaji, K.; Mimoto, T. J. Chem. Soc., Chem. Commun. 1989, 1511. (b) Kimura, T.; Fukui, T.; Tanaka, S.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull.

906 ( 42 ) J . Synth . Org . Chem . , Jpn

1997, 45, 18. (2) Kiso, Y.; Fujiwara, Y.; Kimura, T.; Nishitani, A.; Akaji, K. Mt. J. Peptide & Protein Res. 1992, 40, 308. (3) (a) Akaji, K.; Kuriyama, N.; Kiso, Y. Tetrahedron Letters 1994, 35 , 3315. (b) Akaji, K.; Kuriyama, N.; Kiso, Y. J. Org. Chem. 1996, 61, 3350. (4) Kiso, Y.; Yajima, H. In Gutte, B. (Ed) Peptides: Synthesis, Structures, and Applications, Academic Press, San Diego, 1995, pp. 39-91. (5) DeClercq, E. J. Med. Chem. 1995, 38, 2491. (6) Wlodawer, A.; Miller, M.; Jaskolski, M.; Sathyanarayana, B. K.; Baldwin, E.; Weber, I. T.; Selk, L. M.; Clawson, L.; Schneider, J.; Kent, S. B. H. Science 1989, 245, 616. (7) Mimoto, T.; Imai, J.,;Tanaka, S.; Hattori, N.; Takahashi, O.; Kisanuki, S.; Nagano, Y.; Shintani, M.; Hayashi, H.; Sakikawa H.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1991, 39, 2465. (8) Moore, M. L.; Bryan, W. M.; Fakhoury, S A.; Magaard, V. W.; Huffman, W. F.; Dayton, B. D.; Meek, T. D.; Hyland, L.; Dreyer, G. B.; Metfacf, B. W.; Strickler, J. E.; Gomiak, J. G.; Debouck, C. Biochem. Biophys. Res. Commun. 1989,159, 420. (9) Schnoelzer, M.; Kent, S. B. H. Science 1992, 256, 221. (10) Englebretsen, D. R.; Garnham, B. G.; Bergan, D. A.; Alewood, P. F. Tetrahedron Lett. 1995, 36, 8871. (11) Nishizawa, N.; Kimura, T.; Matsumoto, H.; Fujiwara, T.; Takaku, H.; Akaji, K.; Kiso, Y. In Kitada, C. (Ed.) Peptide Chemistry 1996, Protein Research Foundation, Osaka, 1997, pp61-64. (12) (a) Dilanni, C. L.; Davis, L. J.; Holloway, M. K.; Darke, P. L.; Kohl, N. E.; Dixon R. A. F. J. Biol. Chem., 1990, 265, 17438. (b) Bhat, T. N.; Baldwin, E. T.; Liu, B.; Cheng, Y. S. E.; Erickson, J. W. Nature Str. Biol. 1994, 1, 552. (13) Mimoto, T.; Imai, J.; Tanaka, S.; Hattori, N.; Kisanuki, S.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1991, 39, 3088. (14) Kiso, Y. Biopolymers 1996, 40, 235. (15) Mimoto, T.; Imai, J.; Kisanuki, S.; Enomoto, H.; Hattori, N.; Akaji, K.; Kiso, Y. Chem. Pharm. Bull. 1992, 40, 2251. (16) Kageyama, S.; Mimoto, T.; Murakawa, Y.;Nomizu, M.; Ford, H. J.; Shirasaka, T.; Gulnik, S.; Erickson, J.; Takada, K.; Hayashi, H.; Broder, S.; Kiso, Y.; Mitsuya, H. Antimicrob. Agents Chemother 1993, 37, 810. (17) (a) Baldwin, E. T.; Bhat, T. N.; Gulnik, S.; Liu, B.; Topol, I. A.; Kiso, Y.; Mimoto, T.; Mitsuya, H.; Erickson, J. W. Structure 1995, 3, 581. (b) Wang, Y. X.; Freedberg, D. I.; Yamazaki, T.; Wingfield , P. T.; Stahl, S. J.; Kaufman, J. D.; Kiso, Y.; Torchia, D. A. Biochemistry 1996, 35, 9945. (c) Wang, Y. X.; Freedberg, D. I.; Wingfield, P. T.; Stahl, S. J.; Kaufman, J. D.; Kiso, Y.; Bhat, T. N.; Erickson, J. W.; Torchia, D. A. J. Am. Chem. Soc. 1996, 118, 12287. (18) Tanaka, M.; Mimoto, T.; Anderson, B.; Gulnik, S.; Bhat, T. N.; Yusa, K.; Hayashi, H.; Kiso, Y.; Erickson, J. W.; Mitsuya, H. 11th Int. Conf. AIDS 1996, Abst. Tu.A.260. (19) Kishida, M.; Matsumoto, H.; Nishizawa, N.; Kimura, T.; Fujiwara, Y.; Akaji, K.; Kiso, Y Ab- stracts, 118th Annu. Meeting of Pharm. Soc. Jpn. 1998, Abst. 31 [XG]13-3. (20) Ohno, Y.; Kiso, Y.; Kobayashi, Y. Bioorg. Med. Chem. 1996, 4, 1565. (21) Kiso, Y.; Yamaguchi, S.; Matsumoto, H.; Mimoto, T.; Kato, R.; Nojima, S.; Takaku, H.; Fukazawa, T.; Kimura, T.; Akaji, K. Arch. Pharm. Pharm. Med. Chem. 1998, 331, 87. (22) Mimoto, T.; Kato, R.; Takaku, H.; Misawa, S.; Fukazawa, T.; Nojima, S.; Terashima, K.; Satoh, H.; Shintani, M.; Kiso, Y.; Hayashi, H. Program & Abstracts, First Int. Peptide Symp. 1997, Abst. No. P-202. (Received June 29, 1998)

Vol.56, No .11 (November 1998) ( 43 ) 907