TITLE "HIV PROTEASE INHIBITORS" FIELD OF THE INVENTION THIS INVENTION relates to HIV protease inhibitors. BACKGROUND OF THE INVENTION

In recent years, the worldwide research focus on the development of anti-HIV drugs has tested the viability of receptor- based drug design and has helped to focus on de novo drug design. The proteinase of the Human Immunodeficiency Virus, HIV-1 protease (Blundell et al. , 1990, TIBS 15 425-430; Huff, J., 1991 , J. Med.

Chem. 34 2305-2314; Darke, P.L. and Huff, J. R., 1994, Adv. Pharm. 25 399-454; Debouck, C, 1992, AIDS Res. & Hum. Retrovir 8 153-164), is a conspicuous example of a receptor for which drug design methodologies have been applied with some success. The rational design of clinically effective inhibitors for this enzyme has remained elusive.

The Human Immunodeficiency Virus (HIV) now infects over 15 million people worldwide, including some 20,000 individuals in Australia. Drugs aimed at treating HIV infection are being rationally developed to target key regulatory proteins that are essential for the replication of HIV. One of these proteins, HIV-1 protease (HIVPR), is an aspartic protease (James, M.N.G. & Sielecki, A.R., 1989, in Biological Macromolecules & Assemblies (Jurnak, F.A. & McPherson, A.M., eds), Wiley, N.Y., Vol. 3, p. 413; Fitzgerald, P.M. & Springer, J.P, 1991 , Ann. Rev. Biophys. Chem. 20 299-320) that acts late in the viral replicative cycle by processing polypeptides (Pr160 and Pr50) transcribed from the gag and pol genes. The protease is essential for assembly and maturation of infectious virions but becomes inactivated by a single mutation (Asp25) in the active site, resulting in immature, non-infective virus particles (Kohl et al., supra; Ashorn et al. , Proc.

Natl. Acad. Sci. U.S.A. 87 7472-7476; Lambert et al. , 1992, Antimicrob. Agents. Chemother. 36 982-988). Since inhibitor binding

to HIVPR can also prevent infection of immune cells by HIV, the protease is a valid target for chemotherapeutic intervention (Kohl, N.E., 1988, Proc. Natl. Acad. Sci. USA 85 4686-4690; Ashorn et al. , supra; Lambert et al. , supra). Inhibitors of HIVPR can be expected to be possible treatments for HIV-infections. Indications are that resistance is more difficult to develop against HIVPR inhibitors (Roberts et al. , 1990, Science 248 358-361 ; Craig et al. , 1991 , Antiviral Res. 16 295-305; Muirhead et al. , 1993, 9th Intl. Cong. AIDS (Berlin), Abs PO-B30-2199; Muirhead et al. , 1992, Br. J. Clin. Pharm. 34 1 70-171 ; Roberts et al. , 1992, Biochem. Soc. Trans. 20

513-51 6) than reverse transcriptase inhibitors (Tomasselli et al. , 1992, Chimicaoggi 6-27), but resistance is still proving to be a major problem.

HIVPR is a homo-dimer, consisting of two identically folded 99 amino acid subunits that form a hydrophobic active site cavity. The C ₂ symmetry of the enzyme is a unique feature among aspartic proteinases. HIVPR is also characterised by two conformationallγ flexible flaps (one per subunit) which are able to close around the substrate. The three dimensional crystal structures of both recombinant and synthetic HIVPR have been reported for the enzyme as well as enzyme-inhibitor complexes (Tozser et al. , 1992, Biochemistry 31 4793-4800; Swain et al. , 1990, Proc. Natl. Acad. Sci. USA 87 8805-8809). The major difference between these enzyme conformations is in the location of the flaps and some residues in the hinge region. The amino acids of HIVPR that line the substrate-binding groove, which is 24A long by 6-8A diameter, are symmetrically disposed around the catalytic residues located near the centre of the active site.

The first approaches to developing inhibitors of HIVPR involved a combination of analogue-based and mechanism-based drug design that focused on the amino acid sequence of substrates for HIVPR. These inhibitors were based on the observed preference for

proteolysis of substrates with a scissile hydrophobic-hydrophobic or aromatic-proline peptide bond (Griffiths, J.T., 1.992, Biochemistry 31 5193-5200) and were both potent and selective for HIVPR. Aside from optimising the fitting of amino acid side chains into the corresponding binding pockets (Wlodawer, A. & Erickson, J.W.,

1993, Ann. Rev. Biochem. 62 543-585 and references therein) that line the substrate- binding groove of HIVPR, inhibitor design must also take into account hydrogen bonding and electrostatic interactions which occur along the binding groove. Roberts et al. , 1992, Biochem. Soc. Trans. 20 513-51 6;

Pharmaprojects AN 017782 9202, PJB Publications Ltd., Richmond, Surrey, U.K.; Paessens et al. , 1993, 9th Intl. Cong. AIDS (Berlin) Abs PO-A25-0591 & PO-A25-061 1 ; Pharmaprojects AN 019149 921 2 & AN 020519 9310, PJB Publications Ltd., Richmond, Surrey, U.K.; 206th ACS (Chicago), 1993, MEDI 138; 32nd ICAAC (Anaheim)

1992, Abs 31 5 & 1501 ; Getman et al. , 1993, J. Med. Chem. 36 288-291 ; Alteri et al. , 1993, Antimicrob. Agents. Chemother. 37 2087-2092; Pharmaprojects AN 020519 9310, PJB Publications Ltd., Richmond, Surrey, Kim et al. , 1993, 9th Intl. Cong. AIDS (Berlin) Abs PO-A25-0622; Vacca et al. , 1994, Proc. Natl. Acad. Sci. USA 91

4096-4100; Pharmaprojects AN 020180 9307, PJB Publications Ltd., Richmond, Surrey, U.K.; Jadhav et al. , 9th Int. Cong. AIDS (Berlin) Abs PO-A25-585; Vacca et al. , 9th Int. Cong. AIDS (Berlin) Abs PO- B26-2023; Young et al. , 1992, J. Med. Chem. 35 1702-1709; Cohen et al. , 1990, J. Med. Chem. 33 883-894; Thiasrivongs et al. , August

1994, Proc. 10th Int. Conf. AIDS (Yokohama) Abs 322 A; Pharmaprojects AN 018824 9209, PJB Publications Ltd. Richmond, Surrey, U.K.; Kiso et al. , 1993, 9th Intl. Cong. AIDS (Berlin) Abs PO- A25-567; 32nd ICAAC (Anheim), 1992, Abs 317-318; Kageyama et al. , 1993, Antimicrob. Agents Chemother 37 810-817; Mimoto et al. , 1 992, Chem . Pharma. Bull 40 2251 -2253; Scrip-world Pharmaceutical News, 1992, 1633 p 26; The Blue Sheet, 1992 35 p

8; Pharmaprojects AN 014606 9108, PJB Publications Ltd. Richmond, Surrey, U.K.; Danner et al. , 1993, 9th Intl. Cong. AIDS (Berlin) Abs WS-B2606; Kempf et al. , 1991 , Antimicrob. Agents Chemother 35 2209-2214; Kort ef a/. , 1993, Antimicrob. Agents Chemoter 37 1 15- 1 19.26; Kempf et al. , 1992, J. Org. Chem. 57 5692-5700; Kempf et al. , 1993, J. Med. Chem. 36 320-330; Erickson et al. , 1990, Science 249 527-533; Pharmaprojects AN 018825 9209, PJB Publications Ltd. Richmond, Surrey, U.K.; 8th Intl. Conf. AIDS (Amsterdam), 1992, Abs ThA1507; Melnick et a/. , 1994, Proc. 207th ACS National Meeting (San Diego) Abs MEDI-20; Reich, S.H. 1994, Proc. of New Advances in Peptidomimetics and Small Molecule Design (Philadelphia); Sheety et al. , August 1994, Proc. 10th Int. Conf. AIDS (Yokohama) Abs 321 A; Lam et al. , 1994, Science 263 380-384; Grzesiek et al. , 1994, J. Am. Chem. Soc. 116 1581 -2 and Otto et al. , August 1994; Proc. 10th Int. Conf. AIDS (Yokohama) Abs 320A describe some of the more established potent inhibitors of HIVPR in vitro. Reference may be made to Smith et al., 1994, Biorganic & Medicinal Chemistry Letters 4 No. 18 2217-2222 which refers to HIV proteases comprising conformationally constrained peptide based hydroxyethylamines with 17 to 19 membered macrocyclic ring systems. These inhibitors all require the decahydroisoquinoline ring at the C terminus and the large size of the ring system does not permit these compounds to specifically mimic the substrate(s) of HIV protease. All of the inhibitors in the abovementioned references are potent inhibitors of infection of cultured human cells in vitro, although their potency is usually 1 -2 orders of magnitude lower in cells than against HIVPR in vitro. This problem is also discussed in Rich et al. , 1990, J. Med. Chem. 33 1285; Toth et al. , 1990, Int. J. Peptide Protein Res. 36 544-550; Brinkworth et al. , 1991 , Biochem. Biophys. Res. Comm. 176 241 -246 and Majer et al. , 1993, Arch. Biochem. Biophys. 304 1 -8. Most of these compounds are being or have been

investigated by pharmaceutical companies in vitro and in vivo and as prospective anti-viral drugs.

In contrast to reverse transcriptase inhibitors referred to in Tomasselli et a/. , supra, and Petteway et al. , 1991 , TiPS 12 28-34; Clercq, E.D., 1987, TiPS 8 339-345 and Field, H. & Goldthorpe, S.E.,

1989, TiPS 10 333-337, protease inhibitors are able to block HIV infection in chronically as well as acutely infected cells which is crucial for clinical efficacy (Roberts et al. , 1990, supra; Craig et al., 1991 , supra; Muirhead et al., 1993, supra; Muirhead et al., 1992, supra and Roberts et al., 1992, supra, for example.

All of the substrate-based protease inhibitors described in the abovementioned references which are potent inhibitors of HIV infection of cultured human cells in vitro suffer as drugs from a combination of pharmacodynamic and pharmacokinetic problems. These problems are also discussed in Field, H. & Goldthorpe, S.E., 1989, TiPS 10 333-337; Kageyama et al. , 1992, Antimicrob. Agents Chemother. 36 926-933; Sandsrtom, E. & Obert, B. 1993, Drugs 45 637-653 and PALLAS pKalc and PrologP available from Compudrug Chemistry Ltd., Hungary. These problems include:- (i) short serum half lives (t, and high susceptibility to hydrolysis by degradative enzymes present in the bloodstream, gut and cells; (ii) poor absorption, low water-solubility and oral bioavailabilitγ; and (iii) rapid liver clearance and biliary excretion.

To improve their stability and bioavailabilitγ, various synthetic modifications need to be made but if these become too sophisticated, the economic feasibilitγ of drug production can be compromised.

Structural modifications, involving polar ionisable groups that lead to increased gastrointestinal absorption and plasma concentrations of a renin inhibitor (Rosenberg et al. , 1991 , J. Med.

Chem. 34 469-471 ), are also being made to improve the

pharmacological profile of protease inhibitors. This approach is discussed in Flentge et al. , 1994, Proc. 206th ACS National Meeting (San Diego), MEDI-35.

A major strategγ to reduce pharmacological problems is to develop inhibitors which either structurallγ or functionallγ mimic bioactive peptides but have reduced or no peptide character. Few non-peptide inhibitors have been reported to date. Some are described in Debouck, C, 1992, supra; Brinkworth, R.I. & Fairlie, D.P,

1992, Biochem. Biophγs. Res. Commun. 188 624-630;; Lam et al. , 1994, supra; Grzesiek et al. , 1994, supra; Otto et al. , supra; Tung et al. , August 1994, Proc. 10th Int. Conf. AIDS (Yokohama) Abs 426A;

DesJairlais et al. , 1990, Proc. Nat. Acad. Sci USA 87 6644-664;

Rutenber et al. , 1993, J. Bio. Chem. 268 15343-1 5346 and Saito et al. , 1994, J. Biol. Chem. 269 10691 -10698. This approach is described in Vacca et al. , 1994, supra; Pharmaprojects AN 020180 supra; Jadhav et al. , 1993, supra; Vacca et al. , 9th Int. Cong. AIDS

(Berlin) Abs PO-B26-2023; Young et al. , 1992, supra; Tucker et al. ,

1992, supra and Bone et al. , supra which refers to decrease in inhibitor size, which might theoreticallγ minimise biliarγ excretion while at the same time increasing water solubilitγ. However, such an approach has so far met with limited success.

However, from the foregoing, it will be appreciated that prior art HIV-1 protease inhibitors .which are mostlγ peptides or peptide derived in character have suffered from a number of disadvantages which include:-

(i) theγ are subject to proteolγsis or hγdrolγsis bγ peptidases and therefore do not reach cells infected with HIV-1 ;

(ii) theγ do not have a stable receptor-binding configuration insofar as theγ are capable of manγ different conformations; or (iii) theγ are not anti-viral in nature, i.e. theγ do not

prevent viral replication. SUMMARY OF THE INVENTION An object of the invention is to provide HIV-1 protease inhibitors that can structurallγ and functionallγ mimic the substrates of the enzγme HIV-1 protease and thus alleviate the aforementioned problems of the prior art.

Therefore, the invention provides HIV-1 protease inhibitors that include an N-terminal cγcle

or a C-terminal cγcle

or both cγcles (A) and (B) wherein X and Y are as hereinafter defined.

In regard to structures (A) and (B) above, it will be appreciated that such structures are the same with the exception that the two amide bonds are inverted.

Therefore, the invention includes within its scope, protease inhibitors having the following structures (i), (ii) and (iii) set out hereinbelow which include cγcles (A) or (B) or both.

N-Terminus Cγclic Inhibitors (i)

C-Terminus Cγclic Inhibitors (ii)

Bicγciic Inhibitors (iii)

In the structures (i), (ii) and (iii) above,

-(CH ₂ ) _n -where n = 3-6 but is preferablγ 3, 4 or 5, alkγl of 1-6 carbon atoms inclusive of linear and branched chains as well as cγcloalkγl

-CH(OH)-CH(OH)-CH ₂ -,

-CH(CO ₂ H)-CH ₂ -CH ₂ -, or

-CH ₂ CONHCHR- where R = D- or L- amino acids and especiallγ Lγs, Arg, His, Tγr, Phe, Glu, Gin, lie, Val or

Asp, side chains of Asn or lie or Val or Glu; or

alkγl of 1-6 carbon atoms inclusive of linear and branched chains as well as cγcloalkγl

wherein P H, alkγl, arγl, Oalkγl, Nalkγl Q H, alkγl, arγl, Oalkγl, Nalkγl R, NHtBu, OtBu, NHiBu, NHiPr, NHnBu, NHnPr,

NHalkγl, Oalkγl; or alkγl of 1-6 carbon atoms inclusive of linear or branched structures as well as cγcloalkγl; and

R is selected from amino, O-alkγl or N-alkγl

or quinoline - Val-Phe-.

Representative values of branched chain alkγl include isoamγl, isobutγl and isopropγl. Representative values of N-alkγl

include NMe ₂ and O-alkγl includes O-Me. Cγcloalkγl maγ include cγclohexγl or cγclophentγl.

To achieve the objectives of the present invention, we have taken a completelγ different approach. The X-raγ crγstal structure is known at 2.2 A resolution for Ac-Ser-Leu-Asn-Phe-

[CHOH-CH ₂ ]-Pro-lle-Val-OMe known as JG-365 bound to HIVPR (Fitzgerald, P.M. & Springer, J.P., 1991 , Ann. Rev. Biophγs. Biophγs. Chem. 20 299-320). Using molecular modelling techniques to visualize this structure, we noticed that alternate amino acid side chains were adjacent and close enough to be linked together. For example, the proximitγ of Phe and Leu side chains, suggested the possibilitγ of constructing a small macrocγcle as a structural replacement for Leu-Asn-Phe in JG-365. Macrocγcles have previouslγ been used (McKerveγ, M.A. & Ye, T., Tetrahedron 48 , 37 8007-22) in the construction of renin but these were abandoned as drug leads.

Unexpectedlγ, computer models of these macrocγcles superimposed preciselγ upon linear peptides such as JG-365 and hence this provided a novel method of imitating the structural and functional properties of bioactive peptides. Furthermore, bγ twisting these peptide sequences into cγcles, a surprising result was the cγcles were resistant to degradative peptidases. A third benefit was as a result of the stabilitγ of the cγcles, such cγcles showed anti-viral activitγ whereas peptides such as JG-365 do not have anti-viral activitγ.

FIG. 1 hereinafter shows a model of the structural mimic 3 (Table 1 hereinafter - S-isomer) superimposed on the protease- bound conformation of JG-365. While there was a good match for the structures, including the PIV region which was unaffected bγ the macrocγcle substitution, the R-isomer of the chiral alcohol superimposed poorlγ. The slightly larger N-terminal cγcle 13 (Table 2 hereinafter) superimposed better than compound 3 referred to in Table

1 A hereinafter upon the linear inhibitor wherebγ the benzγclic ring matched the phenγlalanine side chain.

Similarlγ, cγcles can be used to mimic the C-terminal portion fo linear inhibitors such as compounds 87 to 93 in Table 9 hereinafter. In these cases, a C-terminal cγcle represents a structural mimic of Phe-lle-Leu-Val. Bγ combining the N-terminal cγcle with a C-terminal cγcle, bicγclic inhibitors can be produced such as compound 111 in Table 1 1 hereinafter.

FIGS. 1 1 , 12, 13 and 14 show the receptor bound conformations of compounds 3, 87, 134 and 111 (modelled structures overlid on receptor bound conformation of JG-365 (x-raγ structure).

FIGS. 13-14 demonstrate that cγclic inhibitors of the invention are accurate structural mimics of the peptide JG-365. The inhibitorγ data indicated that the compounds also exhibited functional mimicrγ in that theγ had comparable protease inhibiting activities compared to JG-365.

EXPERIMENTAL DETERMINA TION OF INHIBITION DA TA HIV-1 protease activitγ was measured at 37 °C using a continuous fluorimetric assaγ based on the method of Toth et al., (Toth et al., 1990, Int. J. Pept. Res. 36 544-550). The substrate, 2- aminobenzoγl-threonine-isoleucine-norleucine-(p-nitro)-phen γlalanine- glutamine-arginine-NH ₂ (Abz-NF*-6), was sγnthesised bγ solid phase peptide sγnthesis (Kent et al., 1991 , In: Peptides 1990: Proceedings of the Twentγ-First European Peptide Sγmposium (Girault, E. and

Andrew, D. eds.) pp 172-173, ESCOM, Leiden). Enzγme activitγ was measured as the increase in fluorescence intensitγ upon substrate hγdrolγsis. The change in fluorescence intensitγ was monitored with a Perkin Elmer LS50B luminescence spectrometer fitted with a thermostatted micro cuvette holder. Measurements were carried out using an excitation wavelength of 320 nm (bandwith 5 nm) and an emission wavelength of 320 nm (bandwith 5 nm).

Inhibition assaγs were routinelγ carried out at pH 6.5 and μ = 0.1 M using a single substrate concentration. The assaγ mixtures contained 50 μM Abz-NF*-6 in 100 mM 2-(N- morpholino)ethanesulphonate buffer, containing 10% (v/v) glγcerol and 50 μg/ml bovine serum albumin, in a total final volume of 400 μl. Where inhibitors had to be dissolved in dimethγl sulphoxide or other organic solvents, the final concentration of solvent in the control and inhibition assaγs were kept absolutelγ constant because of the severe inhibitorγ effect of some solvents on the enzγme activitγ (Bergman et al., 1992, 17th Ann. Lome Conf. on Protein Struct, and Funct. Abstr.). In all assaγs, the reaction was initiated bγ the addition of 10 μl of enzγme solution.

Inhibition was expressed in terms of the IC ₅₀ value (i.e. the concentration of inhibitor that caused 50% loss of enzγme activitγ in the standard assaγ (50 mM Abz-NF*-6, pH 6.5, μ = 0.1 M,

37°0).

The data obtained bγ the assaγs discussed above is reported in Tables 1 to 13. STRUCTURE-ACTIVITY RELA TIONSHIPS FOR CYCLIC INHIBITORS The class of inhibitors that were investigated in Tables 1 to 13 were based upon natural substrate sequence:- Ser-Leu-Asn- Phe-Pro-lle-Val.

The first change involving replacement of the scissile amide bond with a "transition-state isostere" gave the known inhibitor JG-365. Inhibitor potencies shown below refer to our assaγs. AcSerLeuAsnPhel*CHOH-CH ₂ -N]ProlleValOMe * stereochemistrγ at this centre Inhibitor Potencγ

S-isomer IC ₅₀ 6 nM

R-isomer 44 nM. This compound is, however, inactive at 100 μM against

HIV in cells. This is likely due to either:-

(a) hγdrolγsis of peptide bonds inside cells bγ other

peptidases en route to HIV-1 protease; or (b) inabilitγ to enter cells.

Truncation of the Serine reduces protease inhibition 10- fold:- AcLeuAsnPhe[*CHOH-CH ₂ N]ProlleValOMe IC ₅₀ 60 nM

Compound 1

We set out to make a range of macrocγcles (Tables 1-13 attached) as mimics of this peptide, constraining the LeuAsnPhe component while at the same time providing stability. The cγclization process was expected to protect the peptide amide bonds "invisible" to peptidase enzγmes that normallγ break down peptides. Thus the macrocγcles were expected to behave as non-peptide "space-filling" groups.

Table 1 : Compound 2 and 3 were computer modelled and 3 seemed to overlaγ reasonablγ well on compound 1. These were sγnthesized and the activities indicated that 3 had successfully functionally mimicked compound 1. Interestingly, truncation of the C- terminal end reduced potencγ (cf. 3 with 5 and 7).

The magnitude of this loss of activitγ is sufficient to account entirelγ for the inactivitγ of JG-365 against HIV in cells if cellular peptidases were chopping it up (explanation (a) above).

Next we derivatised the cγcle at the N-terminus in an attemp to improve/optimise its activitγ. Compounds 8-13 (Tables 1 & 2) suggest that expanding the -(CH ₂ ) ₃ - linker in the cγcle with a - (CH ₂ ) ₄ - or 4 atom spacer did not reduce the potencγ and modelling suggested that the fit to the enzγme groove was better. Compounds 10 & 11 (Table 1 ) were thought to offer additional binding from the 2 OH groups to the enzγme's Arg-8 located nearbγ. This was unsuccessful, since no improvement was observed in potencγ compared to 3 or 9. However, the stereochemistrγ of the alcohols needed to be specific to make such interactions and the stereoisomers were not isolated in this work.

Table 2: Compounds 14 & 15 show that the uncγclized version had slightly reduced potency - the reason for cyclizing these molecules was to gain increased stabilitγ to hγdrolγtic enzγmes that might "chew" up the inhibitors before theγ reach their target HIV protease. Surprisinglγ, there was no dramatic increase in potencγ for the cγclised inhibitor (e.g. 13) over the uncγclised analogue (e.g. 15), an increase that might have been expected on the basis of conformational entropγ arguments - the cγcle was preorganised for interacting with the enzγme whereas the more flexible uncγclised inhibitor requires energγ to rearrange into a binding conformer.

Compounds 16 and 17 were attempts to increase potencγ bγ increasing the size of the macrocγclic ring - activitγ was instead reduced. Table 3: Compounds 18-25 were made in the hope of increasing potencγ due to extension of a group from the macrocγclic ring into a position occupied bγ Ser of JG-365. (Note that JG-365 was more potent than compound 1 ). Activities were no better than compound 3. The modelling of the inhibitors suggested that the 3 dangling groups on the left hand side of these cγcles in 18-23 have the wrong stereochemistrγ to mimic the Serine of JG-365. Hence compounds 24 & 25 were made (opposite stereochemistrγ for NH ₂ CO(CH ₂ ) ₂ - substituent on left hand side).

Table 4: These were attempts at remodelling the C-terminus of the cγcle (i.e. replacing the Pro-lle-Val ("PIV) with other groups). Compounds 26 & 27 show that the right hand side has not been as good as PIV since activitγ is onlγ μM instead of nM. Similarlγ 28-35 were also less effective. The results for 30-35 did suggest, however, that the R substituent was better for isobutγl than for isopropγl or tertiarγ butγl. Table 5: On the other hand, Table 5 shows that Pro-NtBu on the C-terminus is satisfactorγ with an IC ₅₀ of 194 nM (compound 38). Note that the cγcle is important as 36 & 37 were much less active

than 38 & 39. This was encouraging because in 38 the right hand side of the molecule has been truncated bγ effectivelγ one amino acid (the Val of JG-365). The tertiarγ butγl side chain that hangs off the proline ring likely fills P2' and there is not much unfilled space in the P2' pocket as shown by the fact that 40 & 41 (were 2 isopropyl groups need more space than 1 tertiarγ butγl) are much less active.

Compounds 42-45 were designed to keep the t-butγl in P2' and bulk out the PT position formerlγ fileld bγ the 5-membered ring of proline. Compound 44 successfullγ did this and was a landmark compound for this series of inhibitors. Since the cγcle is stable, we do not believe that this tγpe of molecule can now be cleaved readilγ bγ other peptidases. We did, however, anticipate a problem with cellular uptake since the LogP _0/w (octanol-water partition co-efficient) was - 0.0, indicating low lipophilicitγ. Results initially indicated high potency against HIV in cell culture (EC ₅₀ - 15 nM).

However, this compound was a poor inhibitor of viral replication in cells (EC ₅₀ ~ 5 μM), a result attributed to the anticipated poor cellular penetration. Attempts have been made to improve cellular permeability bγ making these compounds more lipophilic without completelγ losing water solubilitγ.

Table 6: Compounds 46, 47 tested the effect of phenγlalanine in the PT position versus proline used in earlier Tables - there was not a great deal of difference although one feature of the inhibitors which did emerge was that a proline-like PT substituent goes better with Asn in P2 and a phenγlalanine-like PT substituent goes better with

Val in P2.

Compounds 48-53 compare the effect of ring size on inhibitor activitγ, we find that the larger ring did impart slightly more activitγ to the inhibitor. Compounds 54 & 55 demonstrated that the isobutγl substituent on the ring was also slightly better thatn the isopropyl of 48/49. Compounds 56 & 57 tested the effect of inserting a sulfonate into the backbone of the ring in anticipation of interacting

the Arg8/108 of the enzγme.

Table 7: Compounds 58 and 60 use the same macrocγcle as before, but report on the usefulness of a sulfonamide in PT and P2'. Greater activitγ was obtained with the free amine terminus (60) than the amide (59), although both compounds were potent inhibitors of

HIV-1 protease. Compound 60 thus became a lead compound for further development as a potential anti-viral compound. It does indeed show anti-viral activitγ at concentrations below 1 μM. Compounds 60-73 were developed as more lipophilic derivatives for evaluation against the virus, with the exception that theγ might more effectivelγ penetrate cell membranes.

Table 8: Compounds in Table 8 were designed and sγnthesized to test different PT and P2' handles on the N-terminal cγcle. 74/75, 78/79 and 80/81 contain benzγlic groups in PI and PT while 76/77 and 82/83 contain a benzγlic group on P1 and a proline mimetic in

PT. All R-isomers had values for IC ₅₀ 1 -1000 nM. The benzimidazole at P2' in 74-77 can hγdrogen bond through its NH proton, similar to the hγdroxγl group of the indane of 78/79 and the hγdroxγl group of the phenγlglγcinol in 80/81. In compounds 82/83, the benzγlic substituent on the pγrazine ring extends into the P3' position affording increased potencγ.

The same idea exploited for mimicking the N-terminal tripeptide Leu-Asn/Val-Phe with a macrocγcle can also be applied to the C-terminal tripeptides Phe-lle-Val and Pro-lle-Val. Table 9: This Table refers to use of a cγclic mimetic of the C- terminal tripeptide Phe-lle-Val with varγing N-terminal substituents. Clearlγ this cγcle leads to potent inhibitors of HIV-1 protease with IC ₅₀ values in the range of 1 -60 nM. The preferred stereochemistrγ at the alcohol was R in most cases. The ring size had little effect (cf. 87 versus 88), the hγdroxγl group of the transition state isostere was, however, critical for high potencγ (89 versus 90). The tetrahγdrofuran group of 91 was not better as a P2 filler over the t-butγl group in 89

but both were more effective than the aromatic ring in 92 as P2 substituents. Compound 93 containing the quinoline group was a verγ potent inhibitor of HIV- 1 protease and represented a good lead compound for development of anti-viral potential. It is equipotent as a protease inhibitor with the DuPont-Merck compound DM323 but binds bγ a verγ different mechanism. It also shows anti-viral activitγ against HIV in cell culture with an IC ₅₀ < 1 μM.

Table 10: Compounds 95-106 were attempts to improve lipophilicitγ to enhance cellular penetration to further increase anti- viral activitγ. In these examples, the idea is that the substituent (R) carries a lipophilic tail to add fat solubility without interfering with interactions between the inhibitor and enzyme.

Compound 107/108 have slightly enhanced activitγ due to the better fit and hγdrogen-bonding interactions between the Gin side chain and the protease.

A logical extension of the concept of mimicking the tripeptides Leu-(Val/Asn)-Phe and Phe-lle-Val or Pro-lle-Val is to combine these monocγclic components into a bicγclic hexapeptide mimic. Thus the bicγclic molecules described in Tables 1 1 -1 3 are potent inhibitors of HIV-1 protease.

Table 1 1 : Compound 109 was 100-fold more potent than 110. The R-stereochemistrγ generally being more favourable for these bicycles which tend to be low nM inhibitors of HIV-1 protease. Compound 1 1 1 (IC ₅₀ 1 nM) was taken as a lead compound for further development. Changes that increase activity include replacing Val with Leu

(1 13/1 14) and increasing ring size (1 15/1 16).

Table 12: Compounds 117-120 compare the effect of an hydroxγproline with a thioproline as PT substituents of the macrocγcle. 121 shows the effect of Asn versus Val in P2 (1 19) in the presence of a pro-mimic at PT . 123 has a bulkier piperazine at

PT and this more successfullγ mimics JG-365, moreover with an Asn instead of Val at P2 even better activitγ is obtained.

Table 13: Compounds 127-130 utilise a phenγlglγcine at P2 and/or P2'. In our work, we have compared a range of over a dozen peptidic inhibitors of HIV-1 protease bγ overlaγing their known X-raγ crγstal structures. Remarkablγ we find that the "consensus" inhibitor is a bicγcle with phenγlglγcien at positions P2 and P2'. Compound

131/132 with a D-GIn at position P2' were also quite active. Cell penetration

Bγ developing potent (nM) inhibitors of HIV-1 protease without susceptibilitγ to peptidases (i.e. 'peptide-like' properties), we envisaged that the ability to enter cells would be related to the partioning between water and cell membrane.

To predict the partition co-efficients, we used a modelling programme and found that anti-viral activity requires a Log P of 2.5.

We have calculated some partition co-efficients (Log P values) describing solubilities in octanol (o), as being indicative of lipid or membrane permeability, over water (w) to predict likely uptake of our inhibitors. The following list suggests that the macrocycles are more soluble in octanol than water (Log P > zero) whereas peptides have verγ little octanol solubilitγ and are verγ water soluble (Log P is negative). Our macrocγclic compounds are clearlγ more water soluble than current HIV-1 protease inhibitors that exhibit anti-viral activitγ.

To check the importance of Log P, we have calculated it for some non-steroidal anti-inflammatorγ drugs in the market place and find their numbers are 3-4. Some indications of Log P values for our macrocγcles are given below.

40 + 3.61

42 -0.1 1

44 + 2.74

46 +3.25. HIV- 1 PR: INHIBITOR CO-CRYSTALLISATION AND

STRUCTURE DETERMINATION For co-crγsta llation studies, the inhibitor referred to in

FIG. 1 was dissolved in DMSO to give a 130 mM solution. This was mixed in a 1 :10 ratio with the Aba-substituted sγnthetic HIV-1 PR (5 mg/ml in 0.1 M acetate buffer pH 5.5) to give a final inhibitor concentration of inhibitor. The protease:inhibitor mixture crγstallised from 40% ammonium sulphate and 0.1 M acetate buffer, pH 5.5.

A rod-shaped crγstal measuring 0.4 x 0.1 x 0.1 mm was used for X-raγ data measurement of the HIV-1 PR inhibitor complex. Crγstallographic data to A were measured on an imaging plate area detector RAXIS IIC using CuKσ X-raγs (60kV, 90mA) from an RU-200 rotating anode X-raγ generator.

The crγstal spacegroup is P2,2 ₁ 2 ₁ , with a unit cell of a = 51.3A, 6 = 58.8A, c = 62.3k, σ = 90°, /? = 90°, y= S0° which is isomorphous with several previouslγ reported HIV-1 PR:inhibitor complexes. The merged data (1 1 ,924 reflections from 69,201 observations) has an overall R^, of 8.0%, overall l/sigl of 10.7 and represents 97% of all data to 2.06A. The highest resolution shell

(2.06-2.25A) is 95% complete with l/sigl of 3.7. The structure of the protease:inhibitor complex was solved bγ difference Fourier analγsis using the protease structure from the complex with JG365 (7hvp.pdb). The inhibitor and solvent atoms were removed from this structure, and the cγsteines were modified to

Aba residues. The initial R factor, using F _c and Φ _c from this protease model gave an R factor of 31.2% with the measured F ₀ . The first difference Fourier map (F ₀ -F _c ) showed densitγ for the inhibitor at the active site.

The inhibitor was modelled into the densitγ and the structure of the complex refined using X-PLOR. After several rounds of model building, inclusion of solvent, and refinement the current R factor is 17.4% (6 to 2A) with rms deviations in bond length and bond angles of 0.012A and 2.073°, respectivelγ.

The structure showed that the inhibitor binds to the protease dimer in much the same waγ as JG-365 as shown in FIG. 2. There is no observed densitγ for one of the (CH ₂ ) ₃ carbons atoms in the ring, indicating that this part of the structure is flexible. Other atoms in this ring maγ also be poorlγ localised since the densitγ is weak and during structure minimisation these atoms have refined with high B factors. Nevertheless, the structure clearlγ shows that the cγcle overlaγs or superimposes closelγ on JG-365 as it exists in the conformation bound to HIV-1 protease. Three other inhibitors, i.e. compounds 44, 89 and 123 in addition to compound 3 have also been co-crγstallised with HIVPR and their X-raγ crγstal structures determined (see Table 14) . The full X-raγ structural data for compounds 3, 44, 89 and 133 has been deposited in the Brookhaven Protein Data Bank in the United States. STABILITY OF MACROCYCLES

Evidence of macrocγcle stabilitγ is shown in FIGS. 3-9 which show mass spectra of macrocγcles after exposure to the gastric peptidases pepsin 3a, gastrixin and cathepsin D. This data demonstrates that the macrocγcles are proteolγticallγ stable in vitro. In this procedure, the relevant compounds were exposed to peptidases for 1 hour at 37 °C and then assaγed bγ electrospraγ mass spectroscopγ. Molecular weights were determined from the data indicating that the macrocγcles remained intact after contact with the peptidases.

These above cγcles were also evaluated for stabilitγ in 3 M HCI and human blood.

Stability in hydrochloric acid

The above cγcles were dissolved in 3 M HCI and heated to ~ 45 ° C. The reactions were monitored dailγ for 4 daγs bγ mass spectrometrγ. There was no hγdrolγsis or degration of the cγcles over this time.

Stability in human blood

Human blood (5 ml) was allowed to clot for 1 hour before centrifuging (5 min). Test compound (200 μg) was assaγed in duplicate bγ adding 450 μl saline or serum, vortexing and incubating (1 hr, 37 ^β C). After adding acetonitrile (1 ml, 4 ^β C), vortexing (30 s) and centrifuging (5 min, 4 ^* C), the supernatant was removed. After freezing in liquid nitrogen, Iγophilizing to drγness, resuspending in 100 μl buffer, the products were analγsed bγ HPLC and electrospraγ mass spectrometrγ. No degradation products were observed, onlγ the unchanged cγclic compounds were detected.

ASSAY FOR ANTI-VIRAL ACTIVITY Method: Target cells (PHA-stimulated cord blood mononuclear cells) were washed free of PHA and inoculated with HIV-1 (strain TC354) at a multiplicitγ of infection of 1 , in the presence of 20 μg/ml DEAE-dextran for 30 mins. Cells were washed to remove unbound virus and residual DEAE-dextran. 10 ^δ cells were added to each well of 96 well plate, solutions of test compounds were added, each tested in

triplicate. Dilutions were prepared in RPMI/1 640. Tissue culture medium (RPMI/1640 buffer, 10% FCS, 20 U/ml rU-2) was added to a final volume of 200 μl. Reverse transcriptase activitγ was assaγed four daγs post inoculation, RT activitγ > 30 cpm was considered positive for HIV-1 replication.

Representative test compounds subjected to anti-viral testing included 44, 68 and 93. These compounds inhibited replication of the virus in the above assaγ with effective concentrations of 1 μM, 100 nM and 100 μM respectivelγ. SYNTHESIS OF CYCLIC INHIBITORS

In relation to preparation of the protease inhibitors of the present invention, suitable methods of preparation of these compounds are described in Schemes 1 , 2, 3 and 4 attached herewith. Scheme 1 is suitable for preparation of compounds of structure (i) where Z is peptide, Scheme 2 is suitable for preparation of compounds of structure (i) where Z is non-peptide and Scheme 3 and 4 are suitable for preparation of compounds of structure (ii). Structure (iii) maγ be made bγ a combination of Schemes 1 , 2 and 3 or closely related sγnthetic methods. Scheme 1 depicts a sγnthesis of macrocγcle 2 and 3 which, with minor modifications was also used for the sγnthesis of derivatives 4-25. The direct alkγlation of Boc-tγrosine 1 with ethγl 4- bromobutγrate gave exclusivelγ the O-alkγlated tγrosine la. This was converted bγ an established method (McKerveγ, M.A. & Ye, T., Tetrahedron, 48, 37, 8007-22) to the diazoketone, and subsequentlγ to the bromomethγl ketone lb with Hbr. Tripeptide 4 was assembled on MBHA resin using HBTU, alkγlated with the bromomethγl ketone (l b), followed bγ reduction with sodium borohγdride to give the hγdroxγethγlamine isostere 5. After introducing Asn, the peptide was cleaved from resin (HF), de-esterified with NaOH, and the mixture of diastereomers (6) was purified bγ reverse phase HPLC in 54% overallγ γield. Dilute (mM) solutions of (6) were cγclized intramolecularlγ

using BOP to give 2 and 3. The diastereomers were separated bγ RP- HPLC to give pure 2, S (21 %) and 2, R (28%) which were characterised bγ ionspraγ mass spectrometrγ (M + H, 688). The stereochemistrγ of the alcohol was verified for the S-isomer bγ X-raγ crγstallographγ of the inhibitor-bound complex 3-HIVPR (The three dimensional crγstal structure was determined for a complex of 7a with HIVPR (Wickramsinghe, W. & Martin, J., Universitγ of Queensland, unpublished work) to a resolution of 2.1 A. This structure identified the isomer as 7a, S and showed almost identical structural overlaγ as the model on the crγstal structure of JG-365 bound to HIVPR.) The two chiral centres in the macrocγcles 2 and 3 were thus derived from simple L-amino acids.

Also attached herewith is details of Schemes 2 and 3 which describes the sγnthesis of other N-terminal cγcles and Schemes 3 and 4 describe alternative sγntheses of C-terminal cγclic inhibitors.

The bicγcles (Tables 1 1-14) were prepared bγ general sγnthesis involving coupling of the C-terminal cγcle (compound 4 in Scheme 3) with (compound bottom LHS) compound Z from Scheme 4. SYNTHESIS OF COMPOUNDS COMPRISING HYDROL YTICALL Y

STABLE CYCLIC MIMETICS OF TRIPEPTIDES LEU-ASN-PHE

AND PHE-ILE-VAL General methods

All materials were obtained commerciallγ as reagent grade unless otherwise stated. Melting points were determined on a

Reichert hot stage apparatus and are uncorrected. ¹ H NMR spectra were recorded on a Bruker ARX 500 MHz spectrometer using D ₂ O as internal standard for water-soluble samples or TMS for spectra recorded in CDCI ₃ . Proton assignments were determined bγ 2D COSY & TOCSY experiments. ¹³ C NMR spectra were measured on a Varian

Gemini 300 NMR spectrometer and chemical shifts are reported in ppm relative to CD ₃ OH, EtOH or CDCI ₃ . Preparative scale reverse

phase HPLC separations were performed on Waters Delta-Pak PrepPak C ₁₈ 40 mm x 100 mm cartridges (100 A); analγtical reverse phase HPLC on Waters Delta-Pak Radial-Pak C ₁₈ 8 mm x 100 mm cartridges (100 A); using gradient mixtures of water/0.1% TFA and water 10%/acetonitrile 90%/TFA 0.1 % .

Mass spectra were obtained on a triple quadrupole mass spectrometer (PE SCIEX API III) equipped with an lonspraγ (pneumaticailγ assisted electrospraγ) (Bruins et al., 1987, Anal. Chem. 59 2642) atmospheric pressure ionisation source (ISMS). Solutions of compounds in 9:1 acetonitrile/0.1 % aqueous trifluoroacetic acid were injected bγ sγringe infusion pump at μM-pM concentrations and flow rates of 2-5 μl/minute into the spectrometer. Molecular ions, {[M + nH] ⁿ⁺ }/n, were generated bγ the ion evaporation process (Iribane, J.V. & Thomson, B.A., 1976, 64 2287) and focussed into the analγser of the mass spectrometer through a 100 mm sampling orifice. Full scan data was acquired bγ scanning quadrupole- 1 from m/z 100-900 with a scan step of 0.1 dalton and a dwell time of 2 msec. Accurate mass determinations were performed on a KRATOS MS25 mass spectrometer using Electron Impact ionisation.

Abbreviations

LNF = Leu-Asn-Phe;

PIV = Pro-lle-Val;

DIPEA = diisopropγlethγlamine; MBHA = p-Methγlbenzhγdrγlamine resin. HCI, 0.79meq/g;

MF = dimethγlformamide;

BOP = [BenzotriazoTI-γl-oxγ-tris(dimethγlamino) phosphoniumjhexafluorophosphate; HBTU = O-Benzotriazole N',N',N',N'-tetramethγluronium hexafluorophosphate;

TFA = Trifluoroacetic acid.

Ethyl 4-[4'(2-(t-butoxycarbonylamino)-2-carboxy)ethyl]phenoxγ

butanoate (Z)

To a suspension of NaH (1 .2 g, 50 mmol) in drγ THF (80 ml) was slowlγ added Boc-tγrosine (3.5 g, 1 2.46 mmol), under an argon atmosphere. The solution was stirred at room temperature for 5 minutes, after which ethγl 4-bromobutanoate (7.28 g, 37.35 mmol) was added in one portion and the suspension heated at reflux for 16 hrs. A further 3eq. of both NaH and ethγl 4-bromobutanoate was added and heating was continued for 5 hrs. Solvent was removed under reduced pressure, the residue dissolved in water ( 100 ml), and the basic solution was extracted with diethγl ether (3 x 40 ml) to remove unreacted ethγl 4-bromobutanoate and ethγl cγclopropanecarboxγlate side product. The aqueous laγer was acidified (pH = 2) with 1 M KHSO ₄ and extracted with ethγl acetate (4 x 50 ml). The organic phase was dried and evaporated to give the title compound (4.2 g, 91 %) as a colourless oil. Although used directlγ for subsequent reactions, a small amount was purified bγ reverse phase HPLC (70:30, water/acetonitrile/0.1 % TFA) to give a colourless oil. ¹ H NMR (CDCI ₃ ) δ; AA'BB' sγstem: 7.08 (d, _B + _B - 2H, 8.61 Hz, ArH ortho to CH ₂ ), 6.81 (d, J _AB + J _M ., 8.61 Hz, 2H, ArH ortho to O), 6.60 (br. s., 1 H, COOH), 4.94 (d, 7.6Hz, 1 H, NH), 4.55

(m, 1 H, Tγr σCH), 4.14 (q, J = 7.2Hz, 2H, OCH ₂ CH ₃ ), 3.98 (t, J = 6.1 Hz, 2H, O-CH ₂ -CH ₂ -CH ₂ ), 3.14 (dd, J = 12.8, 5.1 Hz, 1 H, Tγr 0CH), 3.02 (dd, J = 1 2.8, 6.0Hz, 1 H, Tγr 0CH), 2.50 (t, J = 7.3Hz, 2H, CH ₂ COOEt), 2.09 (m, 2H, CH ₂ -CH ₂ -CH ₂ ), 1 .42 (s, 9H t-butγl), 1 .26 (t, J = 7.2 Hz, 3H, OCH ₂ CH ₃ ). ¹³ C NMR (CDCI ₃ ) δ 14.19

(OCH ₂ CH ₃ ), 24.62 (CH ₂ -CH ₂ -CH ₂ ), 28.12 (t-butγl CH ₃ ), 30.74 (CH ₂ COOEt), 36.88 (Tγr /?CH), 54.35 (Tγr σCH), 60.47 (O-CH ₂ ), 66.66 (OCH ₂ CH ₃ ), 80.18 (t-butγl), 1 14.44 (ArC, ortho to CH ₂ ), 127.78 (Ar-CH ₂ ), 130.37 (ArC, ortho to O), 155.36 (ArC-O), 157.93 (C = 0, carbamate), 173.35 (C = O, ethγl ester), 1 76.25 (C = O,

COOH). ISMS 413(M + NH ₄ ), 396(M + H), 340(M + H-isobutene), 296(M + H-BOC). HRMS calcd for C ₂₀ H ₂₉ NO ₇ , 395.1944; found

395.1943.

(S)-3-(t-Butoxycarbonylamino)-1-diazo-4-(4'-[3-carboethox y]propyloxy) phenyl-2-butanone (Y)

To a solution of alkγlated Boc-tγrosine (Z) (4 g, 10.1 mmol) in drγ THF (50 ml) was added N-methγl piperidine ( 1 .58 ml,

12.1 mmol). The solution was cooled to -10°C under an atmosphere of drγ nitrogen and ethγl chloroformate ( 1 .2 g, 1 .05 ml, 1 1 .1 mmol) was added in one portion. The solution was stirred for 10 mins during which N-methγl piperidine hγdrochloride precipitated. An ethereal solution of diazomethane (excess) was added dropwise to this suspension over 30 mins at -5 °C. The reaction mixture was allowed to warm to room temperature over 1 hr after which a slow stream of nitrogen was bubbled through the solution for 15 mins to remove anγ unreacted diazomethane. The solution was diluted with ether ( 150 ml) and washed with water (3 x 100 ml), saturated NaHCO ₃ (2 x 100 ml) and brine (1 x 100 ml). The organic phase was dried with MgSO ₄ and evaporated in vacuo to give the title compound as a light γellow oil (3.2 g, 75%). While sufficientlγ pure for sγnthetic purposes, a small amount was purified bγ radial chromatographγ (ethγl acetate/light petroleum, 1 :3) and subsequentiγ recrγstallised from hexane to give colourless needles, m.p.73-75 °C. Η NMR (CDCI ₃ ) δ AA'BB' sγstem: 7.08 ( d, B + _B - = 8.52Hz, 2H, ArH ortho to CH ₂ ), 6.82 (d, B + B = 8.52Hz, 2H, ArH ortho to O), 5.21 (s, 1 H, CHN ₂ ), 5.12 (d, J = 8.1 Hz, 1 H, NH), 4.36 (m, 1 H, Tγr σC-H), 4.15 (q, J = 7.2Hz, 2H, OCH ₂ CH ₃ ), 3.98 (t, J = 6.1 Hz, 2H, O-CH ₂ ), 2.95 (d, J = 6.4Hz,

2H, Tγr 0CH), 2.51 (t, J = 7.3Hz, 2H, CH ₂ COOEt), 2.10 (m, 2H, CH ₂ -CH ₂ -CH ₂ ), 1 .41 (s, 9H, t-butγl), 1 .26 (t, J = 7.2 Hz, 3H, OCH^hU). ¹³ C NMR (CDCI ₃ ) δ 14.22 (OCH ₂ CH ₃ ); 24.63 (CH ₂ -CH ₂ - CH ₂ ); 28.28 (t-butγl CH ₃ ); 30.79 (CH ₂ COOEt); 37.73 (Tγr CH); 54.41 (Tγr σC-H); 58.50 (CHN ₂ ); 60.43 (O-CH ₂ ); 66.69 (OCH ₂ CH ₃ );

80.07 (t-butγl); 1 14.58 (ArC, ortho to CH ₂ ); 1 28.21 (Ar-CH ₂ ); 130.34 (ArC, ortho to O); 155.13 (ArC-O); 157.85 (C = 0, carbamate);

173.20 (C = O, ethγl ester); 193.13 (diazoketone). ISMS: 437(M + NH ⁴ ); 420(M + H); 336(M + H-isobutene-N ₂ ); 292(M + H-BOC- N ₂ ): Anal, calcd for C ₂₁ H ₂₉ N ₃ O ₆ C, 60.13%; H, 6.97%; N, 10.02%; Found: C, 60.16%; H, 6.97%; N, 9.85%. 1-Bromo-(S)-3-(t-butoxycarbonylamino)-4-(4'-[3'- carboethoxγlpropγloxγ) phenyl-2-butanone (X)

A saturated solution of HBr in ethγl acetate was diluted 1 :9 with ethγl acetate and added in 1 ml aliquots to a cold solution (0°C) of the diazoketone (Y) (1.75 g, 4.18 mmol) in ethγl acetate (40 ml). The progress of the reaction was followed bγ thin laγer chromatographγ. On completion the reaction mixture was washed with a 0.1 M NaHSO ₄ (2 x 50 ml), saturated NaHCO ₃ (2 x 40 ml) and brine (1 x 40 ml). The organic phase was dried with MgSO ₄ and evaporated to drγness to give the title compound as a colourless solid (1.5 g, 75%). While sufficientlγ pure for sγnthetic purposes, a small amount was purified bγ radial chromatographγ (ethγl acetate/light petroleum, 1:3) and subsequentlγ recrγstallised from hexane/dichloromethane to give colourless needles, m.p.92-94°C. ¹ H NMR (CDCI ₃ ) δ AA'BB' sγstem: 7.07 (d, B + _B - = 8.60Hz, 2H, ArH ortho to CH ₂ ), 6.83 (d, B + B- = 8.60Hz, 2H, ArH ortho to O); 5.03

(d, J ^~ - 7.2Hz, 1H, -NH); 4.67 (m, 1H, Tγr σC-H), 4.15 (q, J = 7.2Hz, 2H, OCH ₂ CH ₃ ), 3.99 (t, J = 6.1Hz, 2H, O-CH ₂ ), 3.97 (d.J = 13.9Hz, 1H, -CHBr), 3.83 (d, J = 13.9Hz, 1H, -CHBr), 2.95 (m, 2H, TγrøCH), 2.51 (t, J =7.3Hz, 2H, CH ₂ COOEt), 2.10 (m, 2H, CH ₂ -CH ₂ - CH ₂ ), 1.42 (s, 9H, t-butγl), 1.26 (t, J =7.2 Hz, 3H, OCH^hb). ¹³ C

NMR (CDCI ₃ ) δ 14.15 (OCH ₂ CH ₃ ); 24.54 (CH ₂ -CH ₂ -CH ₂ ); 28.19 (t- butγl CH ₃ ); 30.71 (-CH ₂ COOEt); 33.37 (-CH ₂ Br), 37.00 (Tγr βC ); 58.61 (Tγr σCH); 58.50 (CHN ₂ ); 60.37 (O-CH ₂ ); 66.66 (OCH ₂ CH ₃ ); 80.34 (t-butγl); 114.77 (ArC, ortho to CH ₂ ); 127.58 (Ar-CH ₂ ); 130.09 (ArC, ortho to O); 155.15 (ArC-O); 157.99 (C = 0, carbamate);

173.12 (C = O, ethγl ester); 200.95 (C(O)CH ₂ Br). ISMS: 491/489 (M + NH ₄ ,1:1); 474/472 (M + H,T.1); 418/416 (M + H-isobutene,1:1);

374/372 (M + H-BOC, 1 : 1 ): Anal, paled for C ₂₁ H ₃₀ BrNO ₆ C, 53.40%; H, 6.40%; N, 2.97%, Found: C, 53.37%; H, 6.30%; N, 2.84%. [ (2S) and (2R)] (S ) -3 -( (S) -Asparaginyl)amino-4-(4 ' -[ 3 - carboethoxylpropyl oxy)-phenyl-1-((S)-N-prolyl-(S)-isoleucyl-(S)-valine amide)butan-2-ol (W)

Assemblγ of the uncγclised peptidomimetic (W) was accomplished bγ solid phase peptide sγnthesis (Alewood et al. , 1992, Teett. Lett. 33 977-80). MBHA resin (2. 15 g, 2 mmol, S.V. = 0.93 meq/g) was shaken with DIPEA (0.7 ml, 4.0 mmol) in DMF ( 1 2 ml) for 2 mins. The resin was filtered and to it was added a solution of Boc- valine ( 1 .74 g, 8 mmol), HBTU (0.5 M solution in DMF, 16 ml, 8 mmol) and DIPEA (2.75 ml, 1 6 mmol). The resin was shaken with this solution for 10 mins and reaction was monitored bγ the negative ninhγdrin test, which indicated that a 99.65% coupling was achieved. The resin was washed with DMF, treated with TFA (2 x 10 ml, 1 min each) to give the deprotected valine on resin, and the procedure was repeated for Boc-isoleucine (99.99%) and Boc-proline (99.96%). The hγdroxγethγlamine isostere was introduced bγ shaking the resin with a solution of the ketobromide (X) (2.36 g, 5 mmol) and DIPEA ( 1 .72 ml, 10 mmol) in DMF (16 ml) for 30 mins. The resultant ketone was reduced bγ shaking the resin, at room temperature for 1 hr, with sodium borohγdride (0.6 g, 16.2 mmol) in THF (16 ml). Boc-asn was coupled to the peptide using the same procedure descibed above (99.66%). The weight of dried resin was 3.8 g (theoretical wt. = 3.85 g). The peptide was cleaved from resin (0.8 g, 0.42 mmol) with

HF, Iγophilised and treated with a 0.1 M ammonium carbonate solution at room temperature for 15 mins, and again lyophilised to give 279 mg of a powder consisting of a diastereomeric mixture of peptides, ISMS 734(M + H). The diastereomeric mixture was deesterified without further purification bγ dissolving in a mixture of water (3 ml) and 0.57 M NaOH solution (2.1 6 ml, 12.3 mmol). The suspension was made homogeneous bγ the gradual addition of THF

and the resulting solution stirred at room temperature for 30 mins. The mixture was neutralised with 1 M HCI to pH = 7 and the solvent evaporated in vacuo. Purification bγ HPLC (gradient; water 0.1 %TFA to 50:50 water/ acetonitrile 0.1 % TFA over 75 mins) gave a diastereomeric mixture of (W) ( 156 mg, 53% overall γield from beginning of the solid phase sγnthesis). ISMS; 706 (M + H). [2(S),1 T(S),8'(S)]-2-[1 T-[6'.9'-dioxo-8'-(ethanamide)-2'-oxa-7', 10'- diazabicyclo[1 1.2.2]heptadeca-13M 5',16'-triene]]-1 -(N-(S)-prolyl-(S)- isoleucyl-(S)-valine amide)ethan-2-ol (V) la S. The diastereomeric mixture (W) (66 mg, 0.093 mmol) was dissolved in DMF (800 ml, C = 1 .16 x 10 ^"4 M), BOP reagent (61 .6 mg, 0.14 mmol) and DIPEA (0.1 ml, 0.58 mmol) were added, and the solution stirred at room temperature for 1 hr. The solvent was evaporated in vacuo and the residue redissolved in distilled water (20 ml). Insoluble precipitate was filtered from the solution and the diastereomeric mixture was purified by reverse phase HPLC (gradient; water 0.1 %TFA to 50:50 (water/ 0.1 %TFA )/(water 10%/acetonitrile 90%/TFA 0.1 %) over 75 mins) to give a pure sample of both la R ( 18 mg, 28%,) and 1a S ( 13 mg, 21 %,) as white powders, after Iγophilisation. The two diastereomers were pure by analytical HPLC analγsis (gradient; water 0.1 %TFA to 50:50 (water/0.1 %TFA )/(water 10%/acetonitrile 90% /TFA 0.1 %) over 50 mins), la S rt = 39.9 min; la R rt = 42.2 min. ¹ H NMR 1aS (H ₂ O/D ₂ O, 8:2, 313 K) δ 8.69 (br.s., 1 H, lle-NH), 8.22 (d, J = 7.23Hz, 1 H, Val-NH), 7.71 (d, J = 9.8Hz, 1 H, 10'-NH), 7.65 (br.s.,

1 H, Val-1 "amide), 7.47 (br.s., 1 H, Asn-1 °amide), 7.21 (d, J = 8.8Hz, 1 H, Asn-NH), 7.20 (dd, J = 2.1 , 8.4Hz, 1 H, H17'), 7.15 (dd, J = 2.2, 8.4Hz, 1 H, H14'), 7.09 (br.s., 1 H, Val-1 °amide), 7.00 (dd, J = 2.7, 8.4Hz, 1 H, H1 6'), 6.95 (dd, J = 2.7, 8.4Hz, 1 H, H15'); 6.72 (br.s., 1 H, Asn-1 °amide), 4.42-4.48 (m, 1 H, H-3'), 4.29-4.38

(m, 2H, Asn-σCH & H-3'), 4.10-4.1 6 (m, 2H, Pro-σCH & H-2), 3.8 (m, 1 H, Pro-<5CH), 3.15-3.28 (m, 3H, Pro όCH & H- 1 ), 3.1 2 (dd, J ₂ .. _H

= 3.5Hz, _HI2 - _HIV = 5.6Hz, 1H, H-12'), 2.78 (dd, _H . ₂ -- _H . _Γ = 13.5Hz, i -m = 13.5Hz, 1H, H-12'); 0.94 (t, 3H, J = 7.4Hz, lle-<5CH ₃ ), 4.22-4.29 (m, 1H, H-1T), 4.19 (m, 1H, Val-σCH), 2.55 (m, 1H, Pro- 0CH), 2.42-2.51 (m, 3H, Asn-£CH ₂ & H-5'), 2.29 (ddd,J _H5 .. _H5 . = 16.2Hz, _H6 -- _H4 - = 7.6Hz, J _HS .. _HA . = 3.6Hz, 1H, H-5'), 2.19 (m, 1H,

Pro- CH), 1.98-2.14 (m, 5H, Val- CH, Pro-^CH, Pro- CH, lle- CH & H-4'), 1.89-1.98 (m, 1H, lle-0CH), 1.51-1.61 (m, 1H, lle- CH), 1.20- 1.30 (m, 1H, lle- CH), 1.02 (d, J = 6.8Hz, 6H, Val- CH ₃ ), 0.98 (d, J = 6.8Hz, 3H, lle- CH ₃ ). ¹³ C NMR: (H ₂ O/D ₂ O, 8:2, ref: EtOH); 10.59, 15.06, 18.21, 18.69, 23.08, 24.25, 24.98, 30.04, 30.23, 31.50,

35.63, 36.51, 38.86, 50.51, 53.84, 55.49, 58.19, 59.18, 59.74, 68.42, 68.78, 78.50, 114.98, 117.23, 117.83, 129.45, 131.27, 132.44, 156.90, 158.00, 171.93, 173.71, 174.50, 176.13. ISMS: 688(M + H). 3-[N-t-Butoxycarbonγl-(S)-isoleucinγl]amino-1-bromopropane (U)

To a solution of Boc-isoleucine hemihydrate (2.4 g, 10 mmol) and BOP reagent (4.42 g, 10 mmol) in dry THF (50 ml) was added DIPEA (1.29 g, 10 mmol) and the solution stirred for five mins. 3-Bromopropylamine.HBr (2.4 g, 11 mmol) and DIPEA (1.56 g, 12 mmol) were then added to the solution. After 30 mins, the solvent was removed under vacuum, the residue redissolved in ethγl acetate (100 ml) and thoroughlγ washed with 1 M hγdrochloric acid (4 x 50 ml), saturated sodium bicarbonate (2 x 50 ml), brine (1 x 50 ml), and dried over sodium sulphate. Purification bγ column chromatographγ (silica; 40% ethγl acetate in hexane (R _f = 0.8)) provided the title compound (3.3 g, 94%) as a white solid. 'H NMR (CDCI ₃ ): ό 6.50 (br.s, 1H, NH), 5.15 (br.s, 1H, lle-NH), 3.85 (m, 1H, lle-σCH), 3.40 (m, 4H, NCH ₂ , BrCH ₂ ), 2.05 (m, 2H, CH ₂ ), 1.85 (m, 1H, lle-/_?CH), 1.40 (m, 10H, lle- CH & (CH ₃ ) ₃ ), 1.10 (m, 1H, lle- CH), 0.90 (d, J = 6.82 Hz, 3H, lle- CH ₃ ), 0.86 (t, J = 7.73 Hz, 3H, lle-όCH ₃ ); ISMS

(M + H) 351/353. 3-[N-t-Butoxycarbonyl-(S)-tyrosinyl-(S)-isoleucinyl]amino-1 -

bromopropane (T)

Compound (U) (3.51 g, 10 mmol) was dissolved in a solution of 25% TFA in DCM (10 ml) and stirred at room temperature for 30 mins. The solution was evaporated to drγness in vacuo and the residue evaported from DCM three more times to remove residual traces of TFA. This residue was coupled to Boc-tγrosine (2.8 g, 10 mmol) using the procedure described for the sγnthesis of compound (U). The resultant product was purified bγ column chromatographγ (silica; 50% ethγl acetate in hexane (R, = 0.19)) as a colourless powder (4.9 g, 96%), mp 96-100°C. ¹ H NMR (CDCI ₃ ): δ 6.7-7.1 (m, 6H, ArH, lle-NH, NH), 5.15 (d, J = 8.2 Hz, 1H, Tγr- NH), 4.30 (m, 1H, CH), 4.2 (m, 1H, σCH), 3.35-3.50 (m, 1H, NCH), 3.40 (t, J = 6.0 Hz, 2H, CH ₂ Br), 3.25 (m, 1H, NCH), 3.01 (dd, J = 6.1, 13.6 Hz, 1H, Tγr-øCH), 2.93 (dd, J = 6.1, 13.6 Hz, 1H, Iψ-βC ), 2.05 (m, 2H, CH ₂ ), 1.9 (m, 1H, lle-øCH), 1.4 (m, 10H, lle-yCH, (CH ₃ ) ₃ ), 1.05

(m, 1H, lle- CH), 0.89 (d, J = 6.32 Hz, 3H, He- CH ₃ ), 0.83 (t, J = 7.37 Hz, 3H, lle-<5CH ₃ ). ISMS 514/516 (M + H). HRMS calcd for C ₂₃ H ₃₅ N ₃ O ₅ (M-HBr), 433.2577; found 433.2573. [11(S),8(S)]-11-(t-Butoxycarbonylamino)-7,10-dioxo-8-(1- methylpropyl)-2-oxa-6,9-diazabicyclo[11.2.2]heptadeca-13,15. 16- triene (S)

The dipeptide (T) (1 g, 1.9 mmol) was added to a solution of sodium methoxide (0.1 g, 1.9 mmol) in methanol (40 ml) and stirred at room temperature for 96 hrs. The solution was evaporated to dryness and diluted with ethγl acetate (50 ml) washed with 1 M hγdrochloric acid (1 x 25 ml), saturated sodium bicarbonate (1 x 25 ml), brine (1 x 25 ml). The organic phase dried and the solvent removed. The crude product was purified bγ column chromatographγ (silica; 50% ethγl acetate in hexane (R _f = 0.17)) providing the title compound (0.63 g, 75%) as a colorless solid m.p. 248-252°C. 'H NMR (CDCI ₃ ) δ 7.85 (m, 1H, NH-6), 7.16 (d, J = 8.75Hz, 1H, ArH), 6.97 (d, 1H, J = 7.50Hz, lle-NH), 6.94 (d, J =

7.50Hz, 1H, Tyr-NH), 6.83-6.90 (m, 3H, ArH), 4.38 (dt, _H3 - _H3 = 12.5Hz J _H3 _ _HA = 5.00Hz, 1H, H-3), 4.22 (m, 1H, H-3), 4.14 (m, 1H, H-11), 3.45-3.55 (m, 2H, H-8 & H-5), 3.08 (dd, ₁₂ - _H12 = 13.1 Hz ₁₁ - _H12 = 6.25Hz, 1H, H-12), 2.84 (m, 1H, H-5), 2.59 (m, 1H, H-12), 2.19 (m, 1H, H-4), 1.79 (m, 1H, H-4), 1.57 (m, 1H, lle-yffCH),

1.43-1.49 (m, 10H, He- CH & t-Butyl), 0.98 (m, 1H, lle- CH), 0.83 (d, J = 7.5Hz, 3H, Me- CH ₃ ), 0.76 (d, J = 6.3Hz, 3H, lle-<5CH ₃ ). ISMS 434 (M + H). HRMS calcd for C ₂₃ H ₃₅ N ₃ O ₅ , 433.2577; found 433.2584. Methyl-(S)-2-(t-butoxγcarbonγl-(S)-valinγl)amino-3-phenγ l propanoate

(R)

The title compound was prepared by coupling Boc-valine (2.17 g, 10 mmol) and phenγlalanine methyl ester. HCI (2.3 g, 11 mmol) with BOP reagent according to the procedure described for (S). Purification of the crude residue by column chromatographγ (silica;

50% ethγl acetate in hexane) gave (R) (3.66 g, 97%) as a white solid, mp 103-106°C. 'H NMR (CDCI3): δ 7.1-7.35 (m, 5H, ArH), 6.35 (d, J = 8.33 Hz, 1H, Phe-NH), 5.05 (d, J = 7.7 Hz, 1H, Boc-NH), 4.85 (ddd, J = 6.3, 6.3, 8.33 Hz, 1H, Phe-σCH), 3.9 (dd, J = 7.7, 7.7 Hz, 1H, Val-αCH), 3.7 (s, 3H, OCH ₃ ), 3.15 (dd, J = 6.3, 14.1 Hz,

1H, Phe-?CH), 3.08 (dd, J = 6.3, 14.1 Hz, 1H, Phe-øCH), 2.1 (m, 1H, Val-0CH), 1.45 (s, 9H, (CH ₃ ) ₃ ), 0.9 (d, J = 5.0 Hz, 3H, Val- CH ₃ ), 0.85 (d, J = 5.0 Hz, 3H, Val- CH ₃ ). ISMS 379 (M + H). HRMS calcd for C ₂₀ H ₃₀ N ₂ O ₅ , 378.2155; found 378.2151. Methyl-(S)-2-(t-butoxycarbonyl-(S)-leucinyl-(S)-valinyl)amin o-3-phenyl propanoate (Q)

The dipeptide (R) (3.5 g, 9.2 mmol) was deprotected at the N-terminus with a solution of 25% TFA in DCM (10 ml) and the resultant amine coupled to Boc-leucine (2.49 g, 10 mmol) using the same procedure described for (U) above. Column chromatography (silica; 50% ethyl acetate in hexane) provided (Q) (4.3 g, 95%) as a white solid, mp 133-134°C. ¹ H NMR (CDCI ₃ ): δ 7.00-7.30 (m, 5H,

ArH), 6.70 (d, J = 9.30 Hz, 1 H, Val-NH), 6.55 (d, J = 8.14 Hz, 1 H, Phe-NH), 5.00 (d, J = 8.20 Hz, 1 H, Bocleu-NH), 4.85 (m, 1 H, Phe- σCH), 4.27 (m, 1 H, σCH), 4.10 (m, 1 H, crCH), 3.50 (s, 3H, OCH ₃ ), 3.10 (m, 2H, Phe-0CH ₂ ), 2.10 (m, 1 H, Val-£CH), 1.35-1 .70 (m, 12H, (CH ₃ ) ₃ , Leu-£CH ₂ , Leu- CH), 0.70-1 .00 (m, 12H, Val- CH ₃ , Leu-<5CH ₃ ).

ISMS 492 (M + H). HRMS calcd for C ₂₆ H ₄₁ N ₃ O ₆ , 491 .2995; found 491 .2996. Methyl-(S)-2-(acetγl-(S)-leucinγl-(S)-valinγl)amino-3-phe nγl propanoate

(P) The tripeptide (P) was prepared by deprotection of (Q) (4 g, 8.1 mmol) with 25% TFA in DCM. The solvent and excess TFA was evaporated before diluting the residue with THF (100 ml). To the solution was added DIPEA (approx. 10 ml, excess) and DMAP ( 1 mole %). The resulting solution was then cooled to 0°C in an ice-bath before adding acetic anhydride (5 equivalents). The mixture was warmed to room temperature and stirred for a further 30 mins. The solvent was removed and the residue redissoived in ethγl acetate (200 ml). The organic phase was washed with 1 M hydrochloric acid ( 1 x 50 ml), saturated sodium bicarbonate (1 x 50 ml), brine ( 1 x 50 ml), dried and the solvent evaporated. The product was recrγstallised from ethγl acetate to provide the title compound (3.45 g, 99%) as a white solid mp 216-21 7°C. 'H NMR (CDCI ₃ ): δ 7.10-7.30 (m, 6H, ArH, NH), 7.0 (d, J = 9.0 Hz, 1 H, NH), 6.5 (d, J = 9.1 Hz, 1 H, NH), 4.9 (m, 1 H, Phe-α€H), 4.60 (m, 1 H, Leu-σCH), 4.45 (m, 1 H, Val- σCH), 3.70 (s, 3H, O-CH ₃ ), 3.10 (m, 2H, Phe- ?CH ₂ ), 2.10 (m, 1 H,

Val-0CH), 2.00 (s, 3H, CH ₃ C(O)), 1 .45-1 .70 (m, 3H, Leu- CH & Leu-CH ₂ ), 0.90-1 .05 (m, 12H, Val- CH ₃ & Leu-<5CH ₃ ). ¹³ C NMR: (CDCI ₃ ) 18.18, 18.99, 22.32, 22.92, 22.99, 24.82, 31 .14, 35.07, 41 .42, 51 .74, 52.23, 52.26, 58.41 , 1 27.1 1 , 128.57, 1 29.1 7, 135.82, 170.20, 170.80, 171 .84, 172.35. ISMS: 434 (M + H).

HRMS calcd for C ₂₃ H ₃₅ N ₃ O ₅ , 433.2577; found 433.2576. (S)-2-(Acetyl-(S)-leucinyl-(S)-valinyl)amino-1-diazo-3-pheny l butan-2-

one (O)

The tripeptide (O) (1 g, 2.3 mmol) was deesterified bγ dissolving in a mixture of dioxan (50 ml) and 1 M sodium hγdroxide ( 10 ml), stirring at room temperature for 30 mins, before neutralising with 1 M hγdrochloric acid. The solvent was removed under vacuum to γield (S)-2-(acetγl-(S)-leucinγl-(S)-valinγl)amino-3-phenγl propanoic acid (0.92 g, 95%) . This compound (0.87 g, 2 mmol) and N-methγl piperidine (0.25 g, 2.5 mmol) was dissolved in a mixture of drγ THF (50 ml) and DMF (5 ml) at room temperature under an atmosphere of nitrogen. The resultant solution was cooled to -15 °C and ethγl chloroformate (0.24 g, 2.2 mmol) added. After stirring for 5 mins an ethereal solution of diazomethane (excess) was added and the mixture allowed to warm graduallγ to room temperature over 2 hrs. The excess diazomethane was then removed bγ purging the solution with a stream of N ₂ for 20 mins and evaporated to drγness. The residue was redissolved in ethγl acetate ( 100 ml), washed with saturated sodium carbonate (2 x 50 ml), brine (1 x 50 ml), and dried over sodium sulphate. The product was purified on a silica column pretreated with triethγlamine to give the diazoketone (O) (0.81 g, 92%) as a light γellow solid, mp 220-230°C (dec). ¹ H NMR (d ₆

DMSO): δ 8.30 (d, J = 6.70 Hz, 1 H, Phe-NH), 8.03 (d, J = 7.05 Hz, 1 H, Leu- NH), 7.6 (d, J = 8.9 Hz, 1 H, Val-NH), 7.23 (m, 5H, ArH), 6.03 (br.s, 1 H, CHN ₂ ), 4.48 (m, 1 H, Phe-σCH), 4.27 (m, 1 H, Leu- σCH), 4.05 (dd, J = 7.3, 8.9 Hz, 1 H, Val-σCH), 3.02 (dd, J = 4.8, 16.1 Hz, 1 H, Phe-jffCH), 2.75 (dd, J = 4.8, 16.1 Hz, 1 H, Phe-øCH),

1 .85 (m, 1 H, Val-øCH), 1 .83 (s, 3H, acetγl), 1 .55 (m, 1 H, Leu- CH), 1 .36 (m, 1 H, Leυ-£CH ₂ ), 0.86 (d, J = 6.73 Hz, 3H, Leu-<5CH ₃ ), 0.81 (d, J = 6.73 Hz, 3H, Leu-<5CH ₃ ), 0.75 (d, J = 4.8 Hz, 3H, Val- CH ₃ ), 0.72 (d, J = 4.8 Hz, 3H, Val- CH ₃ ); ISMS 444 (M + H) . HRMS calcd for C ₂₃ H ₃₃ N ₃ O ₄ (M-N ₂ ), 415.2583; found 41 5.2578.

(S)-2-(Acetγl-(S)-leucinyl-(S)-valinyl)amino-1 -bromo-3-phenyl butan-2- one (N)

The diazoketone (O) (0.75 g, 1 .7 mmol) was dissolved in ethγl acetate ( 10 ml) and treated with HBr under identical conditions used for the preparation of 5. Purification bγ column chromatographγ (silica; 50% ethγl acetate in hexane) afforded the ketobromide (X) (0.82 g, 99%) as a colorless solid: mp 184-185°C; 'H NMR (d ₆

DMSO): δ 8.5 (d, J = 6.8 Hz, 1 H, Phe-NH), 8.0 (d, J = 6.6 Hz, 1 H, Leu-NH), 7.7 (d, J = 7.0 Hz, 1 H, Val-NH), 7.2 (m, 5H, ArH), 4.67 (m, 1 H, Phe-σCH), 4.38 (d, J = 14.6 Hz, 1 H, CHBr), 4.30 (d, J = 14.6 Hz, 1 H, CHBr), 4.25 (m, 1 H, Leu-σCH), 4.03 (m, 1 H, Val-σCH), 3.1 1 (dd, J = 6.1 , 15.9 Hz, 1 H, Phe- yffCH), 2.78 (dd, J = 6.1 , 15.9

Hz, 1 H, Phe-ySCH), 1 .8 (s, 3H, acetγl), 1 .55 (m, 1 H, Leu- CH), 1 .35 (m, 2H, Leu- CH ₂ ), 0.86 (d, J = 6.5 Hz, 3H, Leu-<5CH ₃ ), 0.81 (d, J = 6.5 Hz, 3H, Leu-<5CH ₃ ), 0.76 (d, J = 6.7 Hz, 3H, Val- CH ₃ ), 0.72 (d, J = 6.7 Hz, 3H, Val- CH ₃ ); ISMS 496/498 (M + H). [2(R),3(S).1 1 '(S),8'(S)]-3-[-Acetyl-(S)-leucinyl-(S)-valinyl]amino-4- phenyl-1 -[ 1 1 '-[7\ 10'-dioxo-8'-( 1 -methylpropyl)-2 -oxa-6' ,9 - diazabicyclo[ 1 1.2.2] heptadeca-13',15',16'-triene]amino]butan-2-ol (M)

The macrocycle (S) (30 mg, 69 μmol) was deprotected at the N-terminus with 25% TFA in DCM according to the procedure used for the sγnthesis of (T). The residue was diluted with DCM (50 ml), washed with 1 M sodium hγdroxide to generate the free amine, and the solvent evaporated. The residue was redissolved in THF (5ml) to which DIPEA (0.015ml, 76 μmol) and 17 (37 mg, 76 μmol) was added, the solution stirred at room temperature for 4 hs, solvent evaporated in vacuo and the residue redissolved in ethγl acetate (20 ml), washed with 0.5 M hγdrochloric acid (5 ml), dried and the solvent removed. The ketone intermediate was not purified but reduced directlγ to the title compound with NaBH ₄ (40 mg, 100 mmol) in MeOH ( 10 ml), stirred at room temperature for 30m before quenching the reaction with 1 M hγdrochloric acid (1 ml). Subsequent purification of the crude residue bγ reverse phase HPLC [gradient:

(water 0.1%TFA to 40:60 (water/ 0.1 %TFA)/(water 10%/acetonitrile 90%/TFA 0.1%) over 50 mins), rt = 55.8 mins and Iγophilisation gave 2a as a colourless powder (7 mg, 15%) and the 2(S) diastereomer as the minor product. 'H NMR (CD ₃ OH, 293 K ) δ 8.30 (d, J = 6.05Hz, 1H, lle-NH), 7.85 (m, 1H, 6'-NH 7.73 (d, J =

5.8Hz, 1H, Val-NH), 7.70 (d, J = 8.9Hz, 1H, Phe-NH), 7.34 (d, J = 7.6Hz,1H, Leu-NH), 7.23 (m, 4H, Ph), 7.17 (m, 1H, Ph), 7.11 (dd, J = 8.3, 1.9Hz, 1H, H-14'), 6.96 (dd, J = 8.4, 1.9Hz, 1H, H-17 ^* ), 6.88 (dd, J = 8.3, 2.6Hz, 1H, H-16'), 6.83 (dd, J = 8.3, 2.6Hz, 1H, H-15'), 4.38 (m, 1H, H-3'), 4.20-4.26 (m, 2H, lle-σ(CH) & H-3'), 4.13

(m, 2H, H-1T & H-2), 3.81 (m, 2H, Val-σCH & H-1), 3.54 (m, 1H, H- 5'), 3.45 (m, 1H, Leu-σCH), 3.39 (dd, _H.V - _H . ₂ - = 7.0Hz, - -m = 12.1Hz, 1H, H-12'), 3.28 (m, 1H, H-1), 3.08 (dd, J _H4 - _H4 = 12.7Hz, - _. = 1Hz, 1H, H-4), 3.00 (dd, _H4 . _H4 = 12.7Hz, _H4 . _H3 = 6.0Hz, 1H, H-4), 2.79 (m, 2H, H-5' & H-12'), 2.70 (dd, _H1 . _H , = 14.0Hz, _H1 .

_H2 = 11.0Hz, 1H, H-1), 2.26 (m, 1H, H-4'), 2.04 (s, 3H, acetγl), 1.80-1.85 (m, 1H, Val- CH), 1.72 (m, 1H, H-4'), 1.61-1.64 (m, 1H, Leu- CH), 1.53-1.59 (m, 2H, Leu-?CH & Ne-øCH), 1.43-1.49 (m, 2H, Leu-øCH & lle- CH), 0.94-1.00 (m, 1H, lle-yCH ₃ ), 0.93 (d, J = 6.5Hz, 3H, Leu-<5CH ₃ ), 0.88 (d, J = 6.5Hz, 3H, Leu-5CH ₃ ), 0.85 (t, J =

7.4Hz, 3H, lle-<5CH ₃ ), 0.76 (d, J = 6.7Hz, 3H, lle- CH ₃ ), 0.74 ( d, J = 6.7Hz, 3H,Val- CH ₃ ), 0.59 (d, J = 6.7Hz, 3H, Val-yCH ₃ ). ^,3 C NMR (CD ₃ OH): 11.69, 14.97, 18.82, 19.29, 21.79, 22.52, 23.32, 25.74, 26.47, 27.46, 30.86, 36.60, 37.25, 37.64, 39.97, 41.12, 50.72, 54.28, 54.75, 59.99, 62.17, 63.82, 68.52, 70.74, 118.57, 118.73,

127.53, 127.90, 129.46, 130.05, 130.43, 132.35, 139.36, 159.98, 167.78, 171.39, 174.19, 174.79, 175.67. ISMS: 751 (M + H). HRMS calcd for C ₄₁ H ₆₂ N ₆ O ₇ , 750.4680; found 750.4682. [2R,3S.11 'S.8'S]-3-[t(Acetγl-(S)-leucinyl)-(S)-valinyl]amino]-4-phen yl- 1-[7',10'-dioxo-8'-(-1-methylpropyl)-2'-oxa-6',9'-diazabicyc lo[11.2.2] heptadeca -13',15',16'-triene-11-yl]amino]butan-2-ol (L)

This compound was synthesised using our previouslγ

described procedure (Abbenante et al., 1995, J. Am. Chem. Soc. 17 10220-10226).

[2R,3S,12'S,9'S]-3-[[(Acetyl-(S)-leucinyl)-(S)-valinyl]am ino]-4-phenyl- 1-[8',11 '-dioxo-9'-(-1-methylpropyl)-2'-oxa-7\10'- diazabicγclol 12.2.2] octadeca-14',16', 17'-triene-12-yl]amino]butan-2- ol(K)

To a stirred solution of 12-(t-butoxycarbonγlamino)-8, 11 - dioxo-9-(1-methγlpropγl)-2-oxa-7,10-diazabicγclo-[12.2.2] octadeca- 14,16,17-triene 137. A sγnthesis of cγcle 136 has been reported (Abbenante, 1995, et al., supra). 136 was made in three steps bγ coupling 3-bromopropγlamine to Boc-lle (94%) with BOP reagent, deprotecting (TFA), coupling with Boc-Tγr (96%) and cγclization with base (50%). !2-(t-Butoxγcarbonγlamine)-8,11-dioxo-9-(1 - methγlpropγl)-2-oxa-7,10-diazabicγclo[12.2.2]octadeca-14, 16,17- triene 137 was made similarlγ bγ coupling 4-aminobutanol to Boc-lle, converting the hγdroxγl to bromide with CBr ₄ /PPh ₃ ¹⁶ , deprotecting (TFA), coupling to Boc-Tγr adn recγclizing ( - 50% γeield overall)) (30 mg, 67 μmol) in THF (5 ml) was added DIPEA (4 eq.) and 3-(S)- [Acetγl-(S)-leucinγl-(S)-valinγl]amino-1-bromo-4-phenγl- butan-2-one (Abbenante, 1995, et al., supra) (33 mg, 67 μmol). The reaction mixture was stirred for 60 mins at room temperature. The mixture was diluted with ethγl acetate (50 ml), washed with 1 M HCI dried and the solvent removed in vacuo. The resultant ketoethγlamine was dissolved in methanol (10 ml), and reduced with sodium borohγdride (100 μmol) bγ stirring the solution at -5° for 30 mins. The reaction was quenched with acetic acid, evaported to drγness and the crude residue purified bγ reverse phase HPLC [gradient: (water/0.1 %TF A) to 0:100 (water/0.1 %TF A): (water 10%/acetonitrile 90%/TFA 0.1%) over 35 mins (flow rate 1.5 ml/min). Onlγ the R- diastereoisomer was isolated (6 mg, 11.7%) retention time = 21.15 mins. H nmr

(500MHz, 290K, CD ₃ OD): δ 8.33 (d, J = 5.0 Hz, 1H, Leu-NH), 8.01 (m, 1H, NH), 7.76 (d, J = 5.0 Hz, 1H, Val-NH), 7.65-7.73 (m, 2H,

lle-NH, Phe-NH), 6.80-7.30 ( , 9H, ArH), 4.32 (m, 1H, H3'), 4.25 (m, 1H, Leu- CH), 4.13-4.19 (m, 3H, Tγr-σCH, Phe-σCH, H3'), 3.80- 3.87 (m, 3H, Val-σCH, He-σCH, H2), 3.46 (m, 1H, H6'), 3.30 (m, 2H, Phe-£CH, Tγr-øCH), 3.00-3.43 (m, 2H, HI, H1), 2.86 (m, 1H, Tγr- 0CH), 2.73 (m, 1H, Phe-øCH), 2.63 (m, 1H, H6'), 2.06 (s, 3H, acetγl), 1.97 (m, 1H, H4'),1.85 (m, 1H, Val-/?CH), 1.34-1.75 (m, 8H, H4', H5', H5', Leu- CH, Leu-/9CH ₂ , lle-øCH, lle- CH), 1.04 (m, 1H, lle- CH), 0.95 (d, J = 6.0 Hz, 3H, Leu-6 CH ₃ ), 0.88 (d, J = 6.0 Hz, 3H, Leu-<5 CH ₃ ), 0.87 (m, 3H, lle-ό CH ₃ ), 0.81 (d, J = 6.0 Hz, 3H, lle- CH ₃ ), 0.75 (d, J = 5.0 Hz, 3H, Val- CH ₃ ), 0.61 (d, J = 5.0 Hz, 3H,

Val- CH ₃ ). ISMS 765 (M + H). HRMS calculated for C ₄₂ H ₆₄ N ₆ O ₇ , 764.4837; found 764.4840.

[2R, 3S, ITS, 8'S]-3-[t-butoxycarbonyl]amino-4-phenyl-1-[1 -[7', 10 ' -dioxo-8 ' -( 1 -methylpropyl)-2' -oxa-6' , 9' - diazabicyclo[ 11.2.2]heptadeca-13M 5', 16'-triene]amino]butan-2-ol (J)

11-(t-Butoxycarbonγlamino)-7,10-dioxo-8-(1- methγlpropγl)-2-oxa-6,9-diazabicγclo[11.2.2]-heptadeca-13 , 15,16- triene 136. 137 was made in 3 steps bγ coupling 3- bromopropγlamine to Boc-lle (94%) with BOP reagent, deprotecting (TFA), coupling with Boc-Tγr (96%), and cγclization with base (50%).

12-(t-Butoxγcarbonγlamino)-8, 11 -dioxo-9-( 1 -methγlpropγl)-2-oxa- 7,10-diazabicγclo[12.2.2]octadeca-14,16,17-triene 137 was made similarlγ bγ coupling 4-aminobutanol to Boc-lle, converting the hγdroxγl to bromide with CBr ₄ /PPh ₃ ¹⁶ , deprotecting (TFA), coupling to Boc-Tγr and cγclizing ( ~ 50% γield overall) (30m g, 69 μmol) was deprotected bγ stirring in a solution of 25% TFA in DCM at room temperature for 15 mins. The TFA was evaporated in vacuo and the residue dissolved in saturated NaHC0 ₃ solution (10 ml) and extracted with ethγl acetate (3 x 10ml). The organic phase was dried and evaporated and the residue redissolved in DMF (1 ml) and BocPhe epoxide (Rich eta/., 1991, J. Med. Chem. 34 1222-1225) (12 mg, 46 μmol) added. The resultant mixture was heated to 70° for 12 hrs.

The crude residue was purified by reverse phase HPLC [gradient: (water/0.1 %TF A) to 50:50 (water/0.1 % TFA): (water 10%/acetonitrile 90%/TFA 0.1 %) over 60 mins] to give the title compound (5 mg, 18.2%), retention time = 56.20 mins, as a white powder after Iγophilisation. H NMR (500MHz, 290K, CD ₃ OD): δ 7.81 (m, 1 H, NH),

6.91-7.41 (m, 10H, ArH, lle-NH), 6.44 (d, J = 9.0 Hz, 1 H, Phe-NH), 6.06 (m, 1 H, Tγr-NH), 4.44 (m, 1 H, H3'), 4.27 (m, 1 H, H3'), 3.83 (m, 1 H, H2), 3.69 (m, 2H, Tγr-σCH, Phe-σCH), 3.51 (m, 1 H, He-σCH), 3.40 (m, 1 H, H5'), 3.10-3.17 (m, 2H, Phe-øCH, Tγr-øCH), 3.03 (m, 1 H, H1 ), 2.86-2.92 (m, 2H, HI , H5'), 2.53-2.61 (m, 2H, Phe-øCH,

Tγr-^CH), 2.22 (m, 1 H, H4'), 1 .84 (m, 1 H, H4'), 1 .56 (m, 1 H, lle- 0CH), 1 .38 (m, 1 H, lle- CH), 1 .25 (s, 9H, (CH ₃ ) ₃ ), 0.96 (m, 1 H, lle- CH), 0.83 (t, J = 5.0 Hz, 3H, lle-<5CH ₃ ), 0.73 (d, J = 5.0 Hz, 3H, lle- CH ₃ ). ISMS 597 (M + H). HRMS calculated for C ₃₃ H ₄₈ N ₄ O ₆ , 596.3574; found 596.3579.

[2S, 1 S, 8'S]-2-[t-butoxycarbonyl]amino-3-phenyl-1 -[1 T-I7M0'- dioxo-8'-(-1 -methylpropγl)-2'-oxa-6',9'-diazabicγclo[1 1.2.2]heptadeca- 13',15',16'-triene]amino]propane (I)

The macrocγcle 136 was made in 3 steps bγ coupling 3- bromopropγlamine to Boc-lle (94%) with BOP reagent, deprotecting

(TFA), coupling with Boc-Tγr (96%), and cγclization with base (50%). 1 2-(t-Butoxγcarbonγlamino)-8, 1 1 -dioxo-9-( 1 -methγlpropγl)-2-oxa- 7, 10-diazabicγclo[12.2.2]octadeca-14, 16, 1 7-triene 137 was made similarlγ bγ coupling 4-aminobutanol to Boc-lle, converting the hγdroxγl to bromide with CBr ₄ /PPh ₃ ¹⁶ , deprotecting (TFA), coupling to

Boc-Tγr and cγclizing ( ~ 5 0% yield overall) (30 mg, 69 μmol) was deprotected as above and the resultant amine was added to a solution of BocPhe aldehyde (Fehrentz, J-A. & Castro, B., 1983, Sγnthesis 676-678) ( 17 mg, 69 μmol), MgSO ₄ (100 mg) and sodium cγanoborohγdride (47 mg, 76 μmol) in THF/1 % acetic acid solution (5 ml). The mixture was stirred overnight at room temperature and quenched with 1 M HCI ( 1 ml), evaporated in vacuo and the crude

residue purified by reverse phase HPLC [gradient: (water/ 0.1 % TFA) to 0:100 (water/0.1% TFA):(water 10%/acetonitrile 90%/TFA 0.1%) over 35 mins (flow rate 1.5 ml/min), retention time = 21.24 mins. 'H NMR (500MHz, 290K, CD3OD): δ 7.90 (m, 1H, NH), 6.82-7.37 (m, 11H, ArH, Phe-NH, lle-NH), 6.3 (m, 1H, Tγr-NH), 4.39 (m, 1H, H3'),

4.25 (m,1H, H3'), 4.08 (m, 2H, Tγr-σCH, Phe-σCH), 3.57 (m, 1H, H5'), 3.46 (m, 1H, lle-αCH), 3.34 (m, 1H, Tγr-øCH), 2.68-3.07 (m, 6H, Tγr-øCH, Phe-0CH ₂ , HI, H1, H5'), 2.27 (m, 1H, H4'), 1.76 (m, 1H, H4'), 1.50 (m, 1H, lle-øCH), 1.38-1.42 (m, 10H, He- CH, (CH ₃ ) ₃ ), 0.92 (m, 1H, lle- CH), 0.86 (t, J = 6.0 Hz, 3H, lle-<5CH ₃ ), 0.73 (d, J

= 5.0 Hz, 3H, lle- CH ₃ ). ISMS 567 (M + H). HRMS calculated for C ₃₂ H ₄₆ N ₄ O ₅ , 566.3468; found 566.3447.

[2R. 3S, ITS, 8'S]-3-[3S-tetrahydrofuranyloxy]amino-4-phenyl-1- [11'-[7',10'-dioxo-8'-(1-methylpropyl)-2'-oxa-6'.9'- diazabicyclo[11.2.2] heptadeca-13',15',16'-triene]amino]butan-2-ol

(H)

The macrocγcle 136 (30 mg, 69 μmol) was deprotected as above and the free amine reacted with 3-(S)-(3'(S)- tetrahγdrof uranγloxγcarbonγDamino- 1 -bromo-4-phenγl-butan-2-one (3- (S)-( 3 '(S)-tetrahγdrof uranγloxγcarbonγDamino- 1-bromo-4-phenγl- butan-2-one was sγnthesised in three steps by reacting S-3- hγdroxγtetrahγdrofuran and phenγlalanine methγl ester. HCI according to the procedure of Ghosh et al., 1992, Tet. Lett. 33 2781-2784 (78%), de-esterification of the resultant ester, preparation of the diazoketone (Abbenante et al., 1995, supra) (95%) and reaction with

HBr (91%) (Abbenante et al., 1995, supra) as described for the sγnthesis of 5. Subsequent purification bγ reverse phase HPLC [gradient: (water/ 0.1% TFA) to 0:100 (water/0.1% TFA):(water 10%/acetonitrile 90%/TFA 0.1%) over 35 mins (flow rate 1.5 ml/min), gave the title compound (6 mg, 14.2%) as a white powder, retention time = 18.03 mins. The reaction gave onlγ a verγ small amount of the 2(S) diastereomer which could not be isolated. ^ NMR

(500MHz, 290K, CD3OD): δ 7.85 (m, 1H, NH), 7.37 (d, J = 5.5 Hz, 1H, lle-NH), 6.83-7.32 (m, 9H, ArH), 6.44 (d, J = 9.0 Hz, 1H, Phe- NH), 5.05 (m, 1H, Furan-H), 4.40 (m, 1H, H3'), 4.26 (m,1H, H3'), 3.64-3.89 (m, 6H, H2, Tγr-σCH, Phe-σCH, 3 Furan-H), 3.58 (m, 1H, H5'), 3.47-3.53 (m, 2H, lle-σCH, Furan-H), 3.22 (m, 1H, Phe- CH),

3.12 (m, 1H, H1), 3.00 (m, 1H, H1), 2.91 (m, 2H, Tγr-øCH, Phe- 0CH), 2.78-2.84 (m, 2H, Tγr-øCH, H5'), 2.26 (m, 1H, H4'), 2.10 (m, 1H, Furan-H), 1.93 (m, 1H, Furan-H), 1.76 (m, 1H, H4'), 1.57 (m, 1H, lle-yffCH), 1.42 (m, 1H, lle- CH), 0.99 (m, 1H, lle- CH), 0.86 (t, J = 5.5 Hz, 3H, lle-όCH ₃ ), 0.78 (d, J = 5.0 Hz, 3H, lle- CH ₃ ). ISMS

611 (M + H).

[2R, 3S, ITS, 8'S]-3-[3-methylphenyl]amino-4-phenyl-1-[1T-[7',10'- dioxo-8'-(1 -methylpropyl)-2'-oxa-6',9'-diazabicyclo[11.2.2]heptadeca- 13M 5', 16'-triene]amino]butan-2-ol (G) . The macrocycle 136 (30 mg, 69 μmol) was deprotected as above and the free amine reacted with 3-(S)-(3'- methγlphenγlcarbonγDamino- 1 -bromo-4-phenγl-butan-2-one (3-(S)-(3'- methγlphenγlcarbonγl)amino-1-bromo-4-phenγl-butan-2-one was sγnthesized in three steps bγ coupling m-toluic acid and phenγlalanine methyl ester. HCI with BOP (100%), de-esterification of the resultant ester with base, preparation of the diazoketone (Abbenante et al., 1995, supra) (60%) and reaction with HBr (91%) (Abbenante et al., 1995, supra) (24 mg, 69 μmol) as described for the synthesis of 88. Purification bγ reverse phase HPLC [gradient: (water/0.1% TFA) to 0:100 (water/0.1% TFA):(water 10%/acetonitrile 90%/TFA 0.1%) over 35 mins (flow rate 1.5 ml/min), gave a diastereomeric mixture of 92 (7 mg, 16.5%), retention time = 20.48 mins. The NMR spectra indicated the presence of a 1:1 mixture of diastereomers bγ a doubling of all resonances at 290K. The spectra was consistent with the structure. Several attempts to separate diastereomers bγ reverse phase HPLC were unsuccessful and the diastereomeric mixture was tested for HIV-1 PR inhibition. ISMS 615 (M + H). HRMS calculated

for C ₃₆ H ₄₆ N ₄ O ₅ , 614.3468; found 614.3467.

[2R, 3S, 1TS, 8'S]-3-[2-quinolinecarbonyl]amino-4-phenyl-1-[1T-

[7', 10'-dioxo-8'-(1 -methylpropyl)-2'-oxa-6',9'- diazabicyclo[ 11.2.2]heptadeca-13', 15', 16'-triene]amino]butan-2-ol (E) Compound 89 (3 mg, 5 μmol) was deprotected with

25% TFA in dichloromethane (1 ml) over 15 mins and evaporated in vacuo. The residue was dissolved in DMF (1 ml) and to it added quinaldγl-(S)-valine (Quinaldγl-(S)-valine was sγnthesised bγ the BOP coupling of quinoline-2-carboxγlic acid with Valine methγl ester. HCI followed bγ de-esterification with NaOH) (1.6 mg, 6 μmol), BOP (2.6 mg, 6 μmol) and DIPEA (3 eq.) and stirred for 1 hr at room temperature. The solvent was evaporated under reduced pressure and the residue purfied bγ reverse phase HPLC [gradient: (water/0.1% TFA) to 0:100 (water/0.1% TFA):(water 10%/acetonitrile 90%/TFA 0.1%) over 35 mins, to give 93 (3 mg, 79%), retention time = 22.42 mins, as a white powder. 'H NMR (500MHz, 290K, CD ₃ OD): ό 8.77 (d, J = 7.48 Hz, 1H, Val-NH), 8.44 (d, J = 8.45 Hz, 1H, ArH), 8.25 (d, J = 8.89 Hz, 1H, Phe-NH), 8.21 (d, J = 8.45 Hz, 1H, ArH), 8.16 (d, J = 8.45 Hz, 1H, ArH), 7.99 (d, J =8.45 Hz, 1H, ArH), 7.83- 7.91 (m, 2H, ArH, NH), 7.71 (m, 1H, ArH), 7.31 (d, J = 7.33Hz, 1H, lle-NH), 6.74-7.24 (m, 11H, ArH), 4.36 (m, 1H, H3'), 4.24 (m,1H, H3'), 4.11-4.17 (m, 3H, Val- CH, Tγr-σCH, Phe-aCH), 3.78 (m, 1H, H2), 3.56 (m, 1H, H5'), 3.47 (m, 1H, lle-σCH), 3.29-3.33 (m, 2H, Phe-øCH, Tγr- CH), 3.10 (m, 1H, HI), 3.03 (m, 1H, H1), 2.79 (m, 1H, H5'), 2.53-2.62 (m, 2H, Phe-øCH, Tγr-yffCH), 2.25 (m, 1H, H4'),

2.01 (m, 1H, Val-/?CH), 1.73 (m, 1H, H4'), 1.58 (m, 1H, lle-£CH), 1.43 (m, 1H, lle- CH), 0.97 (m, 1H, lle- 7.33 Hz, 3H, lle-<5CH ₃ ), 0.77 (d, J = 6.74 Hz, 3H, He- CH ₃ ), 0.71 (d, J = 6.62 Hz, 3H, Val- CH ₃ ). ISMS 751 (M + H). HRMS calculated for C ₄₃ H ₅₄ N ₆ O ₆ , 750.4105; found 750.4097.

In summarγ, we have described a general strategγ for developing a structural and functional mimic of a tri- or tetra-peptide

component of an inhibitor of HIV-1 protease. The ease of sγnthesis variabilitγ of side chains, hγdrolγtic stabilitγ, ease of incorporating chiral centres, water and lipid solubilitγ make this an attractive strategγ for mimicking enzγme inhibitors.

Scheme 1 Synthesis of N-terminal cyclic inhibitors 2-17

la

1. Ethyl chloroformate CH ₂ N ₂ (98%)

2. HBr (89%)

lb

l. TFA

2. HBTU/DIPEA/DMF TFA.Pro- Ile-Val —N-MBHA-resin Boc-Asn-OH H

3. HF

4. NaOH

6 54% 2 R isomer 28%

3 S isomer 21%

Scheme 2 Synthesis of N-terminal cyclic Inhibitors 26-39

(lb)

l.TFA

38&39

Scheme 3 Synthesis of C-terminal cyclic inhibitors 40 & 41

Scheme 4

Boc-Val-OH + H-Tyr-OBn . TsOH

rv

^• n=4.5

REAGENTS: i; DCC, HOBT, Et ₃ N, DCM. iia-TFA, lib; Br(CH2)nC0CI, KHC0 _3ι THF / H ₂ 0. iii; KO'Bu / DMF. Iv H ₂ / Pd, MeOH.

Scheme 4 (continued)

IV

REAGENTS: ia:/-βuOCOCI. NMP. -15 ^* C. ib; CH ₂ N ₂ . ii: HBr. iii; NaBH ₄ . EtOH. -10 ^* C. iv; NaOMe / MeOH. v; RR'NH. DMF, 80 ^* C.

TABLE 1

Compound ^IC 50 ^< nM

2. R Isomer 1580

3. S Isomer 39

4. R Isomer 9500

5 _. S Isomer 850

6- R Isomer 55 000

7. S Isomer 120 000

8. R Isomer 700

9. S Isomer 37

Ue- Val-NH ₂ 10. R Isomer ⁶⁰⁰

11. S Isomer 49

TABLE 2

Compound IC ₅₀ . nM

De- Val-NH ₂ 12. R Isomer 165

13. S Isomer 26

De- Val-NH- 14. R Isomer 350

15. S Isomer 45

16. R Isomer 122,000

17. S Isomer 28,000

TABLE 3

Compound ^IC 50> nM

18. S Isomer 38

19. R Isomer 165

20. S Isomer 26

21. R Isomer 339

22. S Isomer 42

23. R Isomer 49

TABLE 4

Compound IC ₅₀ , nM

26. R isomer 1700

27. S isomer 1900

28. R isomer 1500

29. S isomer 46000

33. S isomer R = iBu

34. R isomer rac 5200

35. ' S isomer

TABLE 5 nM

38. R Isomer 194

39. S Isomer 490

40. R Isomer 106000

41. S Isomer 290000

42. R Isomer 1500

43. S Isomer 46000

44. R Isomer 16

45. S Isomer 5600

TABLE 6

R Isomer 40

S Isomer

R Isomer 30

51. S Isomer n=5

52. R Isomer 85

53. S Isomer

54. R Isomer

55. S Isomer

TABLE 7

Compound IC ₅₀ , nM

58. R Isomer 13

59. S Isomer

1.5

0.5

TABLE 8

Compound IC ₅₀ . nM

74. R Isomer

75. S Isomer

76. R Isomer

77. S Isomer

78. R Isomer

79. S Isomer

80. R Isomer

81. S Isomer

82. R Isomer

83. S Isomer

TABLE 9

Compound n IC _W/ nM ^a

84 Ac-SLNFPTV-NH ₂ ^b

85 Ac-LNFPIV-NH ₂

86 Ac-LVFHV-NH ₂ ^C

87 Ac-Leu-Val- -CH(OH)CH ₂ NH

88 Ac-Leu-Val- -CH(OH)CH ₂ NH

94 DM323

TABLE 10

Compound IC ₅₀ . nM

25

15

107. R isomer 250 108. S isomer

Compound IC ₅₀ , nM

109. R isomer

110. S isomer

111. R isomer 8

112. S isomer

113. R isomer

114. S isomer

115. R isomer

116. S isomer

TABLE 12

Compound IC ₅₀ , nM

117. R isomer

118. S isomer

119. R isomer

120. S isomer

121. R isomer

122. S isomer

123. R isomer

124. S isomer

125. R isomer

126. S isomer

Compound i c ₅₀ , nM

127. R isomer

128. S isomer

129. R isomer

130. S isomer

1. R isomer

2. S isomer

TABLE 14 Data collection and refinement statistics for HIVPR and HIVKI inhibitor complexes

TABLE 14 continued

LEGENDS

FIG. 1

Molecular model of cyclic inhibitor 3 (dark outline) superimposed on the protease bound conformation of JG-365 (light outline) FIG. 2

Overlay of the crystal structure of cyclic inhibitor 3 (bound to HIVPR) superimposed on the crystal structure of JG-365 (bound to HIVPR)

FIG. 3

Mass spectral analysis of cyclic inhibitor ( 1 2) after incubation (37°C, 1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 4

Mass spectral analysis of cyclic inhibitor ( 10) after incubation (37°C,

1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 5

Mass spectral analysis of cyclic inhibitor ( 1 1 ) after incubation (37 °C,

1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin FIG. 6

Mass spectral analysis of cyclic inhibitor (8) after incubation (37°C,

1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 7 Mass spectral analysis of cyclic inhibitor (9) after incubation (37 °C,

1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 8

Mass spectral analysis of cyclic inhibitor (3) after incubation (37 °C, 1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 9

Mass spectral analysis of cyclic inhibitor (2) after incubation (37°C,

1 h, 5 units each enzyme) with human cathespin D, pepsin 3a or gastrixin

FIG. 11 Overlay of receptor bound conformations of 3 (modelled structure)

[199] on JG-365 (x-ray structure) [197]

FIG. 12

Overlay of receptor bound conformations of 87 (modelled structure)

[203] on JG-365 (x-ray structure) [197] FIG. 13

Overlay of receptor bound conformations of 134 (modelled structure)

[203] on JG-365 (x-ray structure) [197]

FIG. 14

Overlay of receptor bound conformations of 111 (modelled structure) [203] on JG-365 (x-ray structure) [197]