Substrate-based cyclic peptidomimetics of Phe-Ile-Val that inhibit HIV-1 protease using a novel enzyme-binding mode

Article abstract of DOI:10.1021/ja953790z

Results are presented for inhibitors of HIV-1 protease that demonstrate a new strategy for developing peptidomimetics, involving the replacement of flexible segments of peptide substrates with conformationally constrained hydrolytically-stable macrocyclic

Full text of DOI:10.1021/ja953790z

J. Am. Chem. Soc. 1996, 118, 3375-3379
3375
Substrate-Based Cyclic Peptidomimetics of Phe-Ile-Val That
Inhibit HIV-1 Protease Using a Novel Enzyme-Binding Mode
Darren R. March, Giovanni Abbenante, Douglas A. Bergman, Ross I. Brinkworth,
Wasantha Wickramasinghe, Jake Begun, Jennifer L. Martin, and David P. Fairlie*
Contribution from the Centre for Drug Design and DeVelopment, The UniVersity of Queensland,
Brisbane Qld 4072, Australia
ReceiVed NoVember 13, 1995X
Abstract: Results are presented for inhibitors of HIV-1 protease that demonstrate a new strategy for developing
peptidomimetics, involving the replacement of flexible segments of peptide substrates with conformationally constrained
hydrolytically-stable macrocyclic structural mimics. A 15-membered macrocycle that imitates the tripeptide Phe-
Ile-Val was designed and incorporated into the C-terminus of Ac-Leu-Val-Phe-CHOHCH2-{Phe-Ile-Val}-NH2, an
inhibitor of HIV-1 protease derived from a substrate sequence. Advantages of the macrocycle over the acyclic
peptide include constraining its components into their bioactive conformation and protecting the amide bonds from
enzymatic degradation, the cycle being stable to acid, gastric proteases, and plasma. Molecular modeling and X-ray
structural studies reveal that the cyclic inhibitors have a unique enzyme-binding mode, the sterically unencumbered
hydroxyethylamine isostere binds via both its hydroxyl and protonated nitrogen to the anionic Asp25 catalytic residues.
The novel macrocycle superimposes well on the linear peptidic inhibitor for which it was designed as a structural
mimic. Structural mimicry led to functional mimicry as shown by comparable inhibition of the protease by cyclic
and acyclic molecules. Further modification of the acyclic N-terminus (Leu-Val-Phe) gave stable, water-soluble,
potent inhibitors of HIV-1 protease. This approach may have general application to the development of mimetics
of other bioactive peptides, including inhibitors of other enzymes.
Lytic replication of the Human Immunodeficiency Virus Type
1 (HIV-1) can be inhibited by interfering with viral enzymes
required for replication, such as HIV-1 protease (HIVPR),1
which processes polypeptides containing structural and func-
tional proteins that make up the virus. Either mutagenesis2a of
the catalytic Asp25 residue of this aspartic protease or binding
of inhibitors2b to the active site of the protease prevents assembly
of viral proteins and results in immature, noninfective virions.
Many peptidic and non-peptidic inhibitors of HIVPR are now
known,1 but most suffer as drug candidates from low bioavail-
ability which is sometimes due to hydrolytic degradation. Here
we describe a rapid approach to the development of hydrolyti-
cally-stable enzyme inhibitors, based upon designing novel
macrocycles that structurally mimic portions of the receptor-
bound conformation of a peptide substrate. A conformationally
constrained cycle that mimics the C-terminal tripeptide of a
substrate is used as a template here to produce a range of potent
inhibitors of HIVPR. These inhibitors interact with HIVPR
using a previously unreported binding mode, involving a
positively charged hydroxyethylamine transition state isostere
contacting negatively charged catalytic Asp residues of the
enzyme.
easily converted to potent and competitive inhibitors by replac-
ing the scissile amide bond (-CONH-) with non-cleavable
transition state isosteres (e.g. -CHOHCH2-N-). When the
sequence Ac-Ser-Leu-Asn-Phe-{CHOHCH2}-Pro-Ile-Val-NH2
(1, JG365), a potent heptapeptidic inhibitor of HIVPR with a
known binding mode,3 was truncated to the hexapeptidic
inhibitor Ac-Leu-Asn-Phe-{CHOHCH2}-Pro-Ile-Val-NH2 (2),
there was a loss in activity which is recovered in the modified
hexapeptide Ac-Leu-Val-Phe-{CHOHCH2}-Phe-Ile-Val-NH2 (3)
(Table 1). For synthetic expediency we chose to mimic the
C-terminus of the shorter inhibitor 3.
Figure 1a compares the modeled receptor-bound structures
of cyclic inhibitor 6 with the acyclic inhibitor 3, assuming that
the hydroxyethylamine nitrogen is not protonated (uncharged).
On the basis of the close proximity of the Phe and Val side
chains in the putative enzyme-bound conformation of 3 (Figure
1a, red), we decided to link them together in a 15-membered
macrocycle. The resulting cycle in 6 (Figure 1a, white) strongly
suggests good structural mimicry with Phe-Ile-Val. The ty-
rosine, isoleucine, and trimethylene units in 6 superimpose well
on the Phe-Ile-Val residues of 3, both side- and main-chain
atoms matching corresponding atoms of the protease-bound
acyclic peptide.
The peptides Ac-Ser-Leu-Asn-Phe-Pro-Ile-Val-NH2 and Ac-
Leu-Val-Phe-Phe-Ile-Val-NH2 are HIVPR substrates which are
However, at physiological pH, the hydroxyethylamine nitro-
gen is likely to be protonated (pKa ∼10 for a secondary amine).
When 6 is modeled as the protonated form (Figure 1b, green),
the enzyme-binding mode matches that observed in the X-ray
crystal structure of 6 bound to HIVPR (Figure 1b, blue, rmsd
) 0.4 Å) more closely than does the unprotonated form (white,
rmsd ) 2.8 Å). Further, when the crystal structure (blue) was
deliberately energy minimized with its amine nitrogen un-
charged, the modeled structure reverted back to the conformation
X Abstract published in AdVance ACS Abstracts, March 1, 1996.
(1) (a) Blundell, T. L.; Lapatto, R.; Wilderspin, A. F.; Hemmings, A.
M.; Hobart, P. M.; Danley, D. E.; Whittle, P. J. TIBS 1990, 15, 425-430.
(b) Tomasselli, A. G.; Howe, W. J.; Sawyer, T. K.; Wlodawer, A.;
Heinrikson, R. L. Chimicaoggi 1992, 6-27. (c) Wlodawer, A.; Erickson,
J. W. Ann. ReV. Biochem. 1993, 62, 543-585 and references therein. (d)
Darke, P. L.; Huff, J. R. AdV. Pharmacol. 1994, 25, 399-454. (e) West,
M. L.; Fairlie, D. P. Trends Pharmacol. Sci. 1995, 16, 67-75.
(2) (a) Kohl, N. E.; Emini, E. A.; Schleif, W. A.; Davis, L. J.; Heimbach,
J. C.; Dixon, R. A. F.; Scolnick, E. M.; Sigal, I. S. Proc. Natl. Acad. Sci.
U.S.A. 1988, 85, 4686-4690. (b) Lambert, D. M.; Petteway, S. R., Jr.;
McDanal, C. E.; Hart, T. K.; Leary, J. J.; Dreyer, G. B.; Meek, T. D.;
Bugelski, P. J.; Bolognesi, D. P.; Metcalf, B. W.; Matthews, T. J. Antimicrob.
Agents Chemother. 1992, 36, 982-988.
(3) Swain, A.; Miller, M. M.; Green, J.; Rich, D. H.; Schneider, J.; Kent,
S. B. H.; Wlodawer, A. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 8805-
8809.
0002-7863/96/1518-3375$12.00/0 © 1996 American Chemical Society

3376 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
March et al.
Table 1. Inhibition of HIV-1 Protease by Cyclic Peptidomimetics
a pH 6.5, I ) 0.1 M, 37 °C, 50 µM substrate [Abz-NF*-6];
continuous fluorometric assay7 using synthetic enzyme.6 Amino acid
content of inhibitors was quantified, after decomposition (6 N HCl, 24
h, 110 °C), by HPLC with Nle as the internal standard. Stereochemistry
refers to the chiral alcohol. b JG-365, acetyl-Ser-Leu-Asn-Phe-{CH-
OHCH2}-Pro-Ile-Val-NH2.3 c Acetyl-Leu-Asn-Phe-{CHOHCH2}-Pro-
Ile-Val-NH2. d Acetyl-Leu-Val-Phe-{CHOHCH2}-Phe-Ile-Val-NH2.
e See ref 8. f Predicted partition coefficients in octanol-water (ref 14).
Figure 1. Computer-modeled and X-ray structures for R-diasteromeric
inhibitors bound to HIV-1 protease: (a, top) inhibitors 3 (red) and 6
(white) modeled as unprotonated amines. (b, middle) modeled structures
for 6 with unprotonated amine (white) and protonated amine (green)
and compared with its X-ray structure (blue); (c, bottom) superposition
of X-ray crystal structures of 6 (blue) and JG-365 (red).3
shown in white (Figure 1). These results strongly suggest that
the nitrogen of the hydroxyethylamine isostere in inhibitor 6 is
protonated when bound to the enzyme and that this charged
nitrogen plays a key role in positioning the inhibitor in the active
site of HIVPR.
A novel consequence of creating the cycle one atom removed
from the hydroxyethylamine nitrogen, rather than incorporating
the nitrogen in the cycle, is that the nitrogen becomes sterically
unencumbered and can interact with the catalytic Asp residues
in the enzyme. The X-ray crystal structure of 6 and modeling
studies on 4-10 indicate that the nitrogen, protonated and
cationic at pH 5-7, is indeed within hydrogen (or ionic)-bonding
distance from both catalytic Asp carboxylates (Figure 2). At
the same time the hydroxyl substituent is still close enough to
form a hydrogen bond to one Asp. To our knowledge, this
binding mode is unique among inhibitors of HIVPR although
it has been proposed theoretically for an amino-diol isostere.4
The effect of this shift in the transition state isostere is that P1
and P1′ residues of 6 (Figure 1c, blue, X-ray structure) occupy
very different positions from those of Phe and Pro in JG365
(Figure 1c, red, X-ray structure3). However, side chain residues
more remote from the transition state isostere (P2, P2′, P3′)
occupy positions similar to those of JG365.
Figure 2. Hydrogen bonding distances (Å) between transition state
isostere of 6 and active site residues of HIVPR from the X-ray crystal
structure of the HIVPR-6 complex.
cycle 1215 and Boc-Phe-CHO.5a Inhibitors 6 and 10 were made
by solution phase coupling of the known epoxide from Boc-
Phe5b with the C-terminal cycle 12, followed by removal of the
Boc group and capping with one of the N-terminal groups
depicted in Table 1. The use of the (2S,3S)-Boc-Phe-epoxide
ensured a diastereoselective synthesis of the chiral alcohol.
Compounds 4, 5, 8, and 9 were made as diastereomeric mixtures
by coupling the cycle to the N-terminal bromomethyl ketone
Compounds 4-10 in Table 1 were synthetically accessible
by straightforward methods outlined in Scheme 1 for compounds
6 and 10 and Scheme 2 for compound 5. Compound 7 was
made by reductive amination of an equimolar mixture of the
(5) (a) Fehrentz, J.-A.; Castro, B. Synthesis 1983, 676-678. (b) Rich,
D. H.; Sun, C.-Q.; Prasad, J. V. N. V.; Pathiasseril, A.; Toth, M. V.;
Marshall, G. R.; Clare, M.; Mueller, R. A.; Houseman, K. J. Med. Chem.
1991, 34, 1222-1225.
(4) Bisacchi, G. S.; Ahmad, S.; Alam, M.; Ashfaq, A.; Barrish, J.; Cheng,
P. T. W.; Greytok, J.; Hermsmeier, M.; Lin, P.-F.; Merchant, Z.; Skoog,
M.; Spergel, S.; Zahler, R. Bioorg. Med. Chem. Lett. 1995, 5, 459-464.

Peptidomimetics That Inhibit HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3377
Scheme 1
presumably because the cycle in both cases shifts the furan
oxygen slightly out of hydrogen-bonding contact distance. On
the other hand the longer quinoline-Val-Phe of 10, which can
make more interactions with HIVPR, increased potency sig-
nificantly. Compound 9 has a benzamide substituent that
appears to be comparable to the other N-termini since even the
diastereomeric mixture had Ki ∼ 15 mM.
Macrocycles have previously been used11 as renin inhibitors,
and there are HIV-1 protease inhibitors containing cycles at the
N-terminus,12 but only one receptor-bound structure of these
molecules is known.12a In this work, the receptor-bound X-ray
crystal structure3 of an acyclic peptidic inhibitor (1) has been
used to design inhibitors containing a C-terminal macrocycle
as a structural mimic that locks the otherwise flexible peptide
into a protein-binding conformation. The predicted structural
mimicry has been confirmed for the cycle in 6 by X-ray
crystallography and can confidently be inferred for other
molecules here. The modeling studies indicate that, despite the
presence of the flexible trimethylene linker, the 15-membered
ring is highly constrained. This is supported by the NMR
inequivalence of the four aromatic protons in the cycle, implying
that the aromatic ring is not able to rotate freely due to steric
congestion.
Scheme 2
Like the cyclic renin inhibitors, we find that the macrocycle
in compounds 4-10 is quite stable to hydrolysis and proteolysis.
For example, the cycle survives intact in 3 M HCl solutions at
40 °C for at least 4 days. Similarly, when this 15-membered
macrocycle was incubated (37 °C, 1 h) with human cathepsin
D or pepsin A (mg/mL) and analyzed by rp-HPLC and
electrospray mass spectrometry, there was no evidence of
degradation. When the cycle in 4-10 was incubated with
human plasma (37 °C, 1 h), HPLC-MS analysis indicated no
significant cleavage of the macrocycle. The stability of the cycle
is also suggested by preliminary antiviral studies, where
exposure of 10 for 4 days to cultured cells infected with HIV-1
resulted in appreciable antiviral activity.13 In each of these
experiments, the acyclic peptidic inhibitors (1-3) of HIVPR are
completely degraded under the same conditions and show no
antiviral activity at 10 µM concentrations.
In summary, we have described a general strategy for
developing a proteolytically stable structural and functional
mimic of a tripeptide component of an enzyme inhibitor. The
ease of synthesis and control over chiral centers, ready variability
of side chains, hydrolytic and proteolytic stability, and the high
solubility (in water and octanol) make this a potentially useful
strategy for mimicking other bioactive peptides, including
peptidic substrates of other enzymes.
derivatives, followed by reduction with NaBH4. The R-
diastereomer was the major product for 4, 5, and 8 and was
easily purified by reverse phase HPLC. Compound 9 was
obtained as an equal mixture of diastereomeric alcohols that
did not separate well under a range of HPLC conditions.
Table 1 reports inhibitor potencies of some molecules
containing the macrocyclic mimetic of Phe-Ile-Val against
synthetic6 HIVPR in Vitro. Under the assay conditions,7
reference inhibitors JG3653 and DM3238 potently inhibit
cleavage of a synthetic fluorogenic substrate.7 The larger 16-
membered ring in 5 was no more effective than the 15-
membered ring of 4, and neither ring on its own was a significant
inhibitor (IC50 ∼ 1-10 µM). Compounds 6-10 combine
different N-terminal ends with the C-terminal cycle. Truncating
the N-terminal tripeptide LVF to Boc-Phe reduced the potency
by 1 order of magnitude (6 vs 4), and removing the important
hydroxyl substituent of the transition state isostere further
reduced potency by a second order of magnitude (7), despite
the presence of a likely protonated amine in the active site.
Compounds 8 and 10 couple the C-terminal cycle to N-terminal
groups of known inhibitors.9,10 The tetrahydrofuran terminus
in 8 did not improve potency over the Boc group in 6,
(9) Kim, E. E.; Baker, C. T.; Dwyer, M. D.; Murcko, M. A.; Rao, B.
G.; Tung, R. D.; Navia, M. A. J. Am. Chem. Soc. 1995, 117, 1181-1182.
(10) Roberts, N. A.; Martin, J. A.; Kinchington, D.; Broadhurst, A. V.;
Craig, J. C.; Duncan, I. B.; Galpin, S. A.; Handa, B. K.; Kay, J.; Krohn,
A.; Lambert, R. W.; Merret, J. H.; Mills, J. S.; Parkes, K. E. B.; Redshaw,
S.; Ritchie, A. J.; Taylor, D. L.; Thomas, G. J.; Machin, P. J. Science 1990,
248, 358-361.
(11) See partial review in: Fairlie, D. P.; Abbenante, G.; March, D. R.
Curr. Med. Chem. 1995, 2, 672-705.
(6) (a) Schneider, J.; Kent, S. B. H. Cell 1988, 54, 363-8. Except that
Aba replaces the Cys residues in the SF2 isolate as footnoted in: (b)
Wlodawer, A.; Miller, M.; Jasko´lski, M.; Sathyanarayana, B. K.; Baldwin,
E.; Weber, I. T.; Selk, L. M.; Clawson, L.; Schneider, J.; Kent, S. B. H.
Science 1989, 245, 616-621. To minimize autocatalytic degradation two
mutations (Q7K and L33I) were made for X-ray work.
(7) (a) Toth, M.; Marshall, G. R. Int. J. Pept. Protein Res. 1990, 36,
544-550. (b) Bergman, D. A.; Alewood, D.; Alewood, P. F.; Andrews, J.
L.; Brinkworth, R. I.; Engelbretsen, D. R.; Kent, S. B. H. Lett. Pept. Sci.
1995, 2, 99-107.
(8) Lam, P. Y. S.; Jadhav, P. K.; Eyermann, C. J.; Hodge, C. N.; Ru,
Y.; Bacheler, L. T.; Meek, J. L.; Otto, M. J.; Rayner, M. M.; Wong, Y. N.;
Chang, C.-H.; Weber, P. C.; Jackson, D. A.; Sharpe, T. R.; Erickson-
Viitanen, S. Science 1994, 263, 380-384.
(12) (a) Abbenante, G.; March, D. R.; Bergman, D. A.; Hunt, P. A.;
Garnham, B.; Dancer, R. J.; Martin, J. L.; Fairlie, D. P. J. Am. Chem. Soc.
1995, 117, 10220-10226. (b) Podlogar, B. L.; Farr, R. A.; Friedrich, D.;
Tarnus, C.; Huber, E. W.; Cregge, R. J.; Schirlin, D. J. Med. Chem. 1994,
37, 3684-3692. (c) Smith, R. A.; Coles, P. J.; Chen, J. J.; Robinson, V. J.;
Macdonald, I. D.; Carriere, J.; Krantz, A. Bioorg. Med. Chem. Lett. 1994,
4, 2217-2222.
(13) Compound 10 has similar antiviral activity (EC50 ∼100 nM; PHA-
stimulated cord blood mononuclear cells infected with HIV-1 (strain TC354);
RT activity assessed 4 days after inoculation) to DM3238 (EC50 110 nM)
under the same conditions.
(14) PALLAS pKalc and PrologP available from Compudrug Chemistry
Ltd, Hungary.

3378 J. Am. Chem. Soc., Vol. 118, No. 14, 1996
March et al.
(m, 1H, Val-âCH), 1.34-1.75 (m, 8H, H4′, H5′, H5′, Leu-γCH, Leu-
âCH2, Ile-âCH, Ile-γCH), 1.04 (m, 1H, Ile-γCH), 0.95 (d, J ) 6.0 Hz,
3H, Leu-δ CH3), 0.88 (d, J ) 6.0 Hz, 3H, Leu-δ CH3), 0.87 (m, 3H,
Ile-δ CH3), 0.81 (d, J ) 6.0 Hz, 3H, Ile-γCH3), 0.75 (d, J ) 5.0 Hz,
3H, Val-γCH3), 0.61 (d, J ) 5.0 Hz, 3H, Val-γCH3). ISMS: m/z 765
(M + H). HRMS: calcd for C42H64N6O7 764.4837, found 764.4840.
(2R,3S,11′S,8′S]-3-(tert-Butoxycarbonylamino)-4-phenyl-1-[[7′,-
10′-dioxo-8′-(1-methylpropyl)-2′-oxa-6′,9′-diazabicyclo[11.2.2]hep-
tadeca-13′,15′,16′-trien-11′-yl]amino]butan-2-ol (6). 11-Amino-7,10-
dioxo-8-(1-methylpropyl)-2-oxa-6,9-diazabicyclo[11.2.2]heptadeca-
13,15,16-triene (12)15 (23 mg, 69 µmol) was dissolved in DMF (1 mL),
and BocPhe epoxide5b (12 mg, 46 µmol) added. The resultant mixture
was heated to 70° for 12 h. The crude residue 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 60 min] to give the
title compound (5 mg, 18.2%), retention time ) 56.20 min, as a white
powder after lyophilization. 1H NMR (500 MHz, 290 K, CD3OD): δ
7.81 (m, 1H, NH), 6.91-7.41 (m, 10H, ArH, Ile-NH), 6.44 (d, J ) 9.0
Hz, 1H, Phe-NH), 6.06 (m, 1H, Tyr-NH), 4.44 (m, 1H, H3′), 4.27 (m,-
1H, H3′), 3.83 (m, 1H, H2), 3.69 (m, 2H, Tyr-RCH, Phe-RCH), 3.51
(m, 1H, Ile-RCH), 3.40 (m, 1H, H5′), 3.10-3.17 (m, 2H, Phe-âCH,
Tyr-âCH), 3.03 (m, 1H, H1), 2.86-2.92 (m, 2H, H1, H5′), 2.53-2.61
(m, 2H, Phe-âCH, Tyr-âCH), 2.22 (m, 1H, H4′), 1.84 (m, 1H, H4′),
1.56 (m, 1H, Ile-âCH), 1.38 (m, 1H, Ile-γCH), 1.25 (s, 9H, (CH3)3),
0.96 (m, 1H, Ile-γCH), 0.83 (t, J ) 5.0 Hz, 3H, Ile-δCH3), 0.73 (d, J
) 5.0 Hz, 3H, Ile-γCH3). ISMS: m/z 597 (M + H). HRMS: calcd
for C33H48N4O6 596.3574, found 596.3579.
(2S,11′S,8′S)-2-(tert-Butoxycarbonylamino)-3-phenyl-1-[[7′,10′-di-
oxo-8′-(1-methylpropyl)-2′-oxa-6′,9′-diazabicyclo[11.2.2]heptadeca-
13′,15′,16′-trien-11′-yl]amino]propane (7). The macrocycle 1215 (23
mg, 69 µmol) was added to a solution of BocPhe aldehyde5a (17 mg,
69 µmol), MgSO4 (100 mg), and sodium cyanoborohydride (47 mg,
76 µmol) in THF/1% acetic acid solution (5 mL). The mixture was
stirred overnight at room temperature quenched with 1 M HCl (1 mL),
and evaporated in vacuo, and the crude residue was 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 min (flow rate
1.5 mL/min), retention time ) 21.24 min. 1H NMR (500 MHz, 290
K, CD3OD): δ 7.90 (m, 1H, NH), 6.82-7.37 (m, 11H, ArH, Phe-NH,
Ile-NH), 6.3 (m, 1H, Tyr-NH), 4.39 (m, 1H, H3′), 4.25 (m,1H, H3′),
4.08 (m, 2H, Tyr-RCH, Phe-RCH), 3.57 (m, 1H, H5′), 3.46 (m, 1H,
Ile-RCH), 3.34 (m, 1H, Tyr-âCH), 2.68-3.07 (m, 6H, Tyr-âCH, Phe-
âCH2, H1, H1, H5′), 2.27 (m, 1H, H4′), 1.76 (m, 1H, H4′), 1.50 (m,
1H, Ile-âCH), 1.38-1.42 (m, 10H, Ile-γCH, (CH3)3), 0.92 (m, 1H, Ile-
γCH), 0.86 (t, J ) 6.0 Hz, 3H, Ile-δCH3), 0.73 (d, J ) 5.0 Hz, 3H,
Ile-γCH3). ISMS: m/z 567 (M + H). HRMS: calcd for C32H46N4O5
566.3468, found 566.3447.
(2R,3S,11′S,8′S)-3-[[[(3(S)-Tetrahydrofuranyl)oxy]carbonyl]amino]-
4-phenyl-1-[[7′,10′-dioxo-8′-(1-methylpropyl)-2′-oxa-6′,9′-diazabicyclo-
[11.2.2]heptadeca-13′,15′,16′-trien-11′-yl]amino]butan-2-ol (8). The
macrocycle 12 (23 mg, 69 µmol) was reacted with 3(S)-[[[(3′(S)-
tetrahydrofuranyl)oxy]carbonyl]amino]-1-bromo-4-phenylbutan-2-
one17 (24 mg, 69 µmol) as described for the synthesis of 5. Subsequent
purification 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 min (flow rate 1.5 mL/min) gave the title compound (6 mg,
14.2%) as a white powder, retention time ) 18.03 min. The reaction
gave only a very small amount of the 2(S) diastereomer which could
not be isolated. 1H NMR (500 MHz, 290 K, CD3OD): δ 7.85 (m,
1H, NH), 7.37 (d, J ) 5.5 Hz, 1H, Ile-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, Tyr-RCH, Phe-
RCH, 3 furan-H), 3.58 (m, 1H, H5′), 3.47-3.53 (m, 2H, Ile-RCH, furan-
H), 3.22 (m, 1H, Phe-âCH), 3.12 (m, 1H, H1), 3.00 (m, 1H, H1), 2.91
(m, 2H, Tyr-âCH, Phe-âCH), 2.78-2.84 (m, 2H, Tyr-âCH, 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, Ile-âCH), 1.42 (m, 1H, Ile-γCH), 0.99 (m, 1H,
Experimental Section
Abbreviations: PIV ) Pro-Ile-Val; DIPEA ) diisopropylethyl-
amine; DMF ) dimethylformamide; BOP ) (benzotriazol-1-yloxy)-
tris(dimethylamino)phosphonium hexafluorophosphate; HBTU ) [(ben-
zotriazolyl)oxy]-N′,N′,N′,N′-tetramethyluronium hexafluorophosphate;
TFA ) trifluoroacetic acid; PHA ) phytohaemagglutinin.
General Methods. 1H NMR spectra were recorded on a Bruker
ARX 500 MHz NMR spectrometer using CD3OH as an internal
standard. Proton assignments were made using 2D COSY and TOCSY
experimental data. Preparative scale HPLC separations were performed
on Waters Delta-Pak Prep-Pak C18 40 × 100 mm cartridges (100 Å);
analytical reverse phase HPLC was performed on Waters Delta-Pak
Radial-Pak C18 8 × 100 mm cartridges (100 Å); gradient mixtures of
water/0.1% TFA and water (10%)/acetonitrile (90%)/0.1%TFA were
used.
Mass spectra were obtained on a triple-quadrupole mass spectrometer
(PE SCIEX API III) equipped with an Ionspray (pneumatically assisted
electrospray) atmospheric pressure ionization source (ISMS). Solutions
of compounds in 9:1 acetonitrile/0.1% aqueous trifluoroacetic acid were
injected by syringe infusion pump at micromolar to picomolar
concentrations and flow rates of 2-5 mL/min into the spectrometer.
Molecular ions, {[M + nH]n+/n, were generated by the ion evaporation
process and focused into the analyzer of the mass spectrometer through
a 100 mm sampling orifice. Full scan data were acquired by scanning
quadrupole-1 from m/z 100-900 with a scan step of 0.1 Da and a dwell
time of 2 ms. Accurate mass determinations were performed on a
KRATOS MS25 mass spectrometer using electron impact ionization.
Amino acid analyses were performed by decomposition of inhibitors
with 6 M HCl (24 h, 110 °C) and quantified by rp-HPLC with Nle as
an internal standard.
Syntheses. (2R,3S,11′S,8′S)-3-[[(Acetyl-(S)-leucinyl)-(S)-valinyl]-
amino]-4-phenyl-1-[[7′,10′-dioxo-8′-(1-methylpropyl)-2′-oxa-6′,9′-
diazabicyclo[11.2.2]heptadeca-13′,15′,16′-trien-11-yl]amino]butan-
2-ol (4). This compound was synthesized using our previously
described procedure.12a
(2R,3S,12′S,9′S)-3-[[(Acetyl-(S)-leucinyl)-(S)-valinyl]amino]-4-
phenyl-1-[[8′,11′-dioxo-9′-(1-methylpropyl)-2′-oxa-7′,10′-diazabicyclo-
[12.2.2]octadeca-14′,16′,17′-trien-12-yl]amino]butan-2-ol (5). To a
stirred solution of 12-Amino-8,11-dioxo-9-(1-methylpropyl)-2-oxa-7,-
10-diazabicyclo[12.2.2]octadeca-14,16,17-triene (13)15 (23 mg, 67
µmol) in THF (5 mL) was added DIPEA (4 equiv) and 3(S)-[[(acetyl-
(S)-leucinyl)-(S)-valinyl]amino]-1-bromo-4-phenylbutan-2-one12a (33
mg, 67 µmol). The reaction mixture was stirred for 60 min at room
temperature. The mixture was diluted with ethyl acetate (50 mL),
washed with 1 M HCl, and dried and the solvent removed in vacuo.
The resultant ketoethylamine was dissolved in methanol (10 mL) and
reduced with sodium borohydride (100 µmol) by stirring the solution
at -5° for 30 min. The reaction was quenched with acetic acid and
evaporated to dryness, and the crude residue was 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 min (flow rate
1.5 mL/min). Only the R-diastereoisomer was isolated (6 mg, 11.7%);
retention time ) 21.15 min. 1H NMR (500 MHz, 290 K, CD3OD): 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, Ile-NH, Phe-NH), 6.80-7.30
(m, 9H, ArH), 4.32 (m, 1H, H3′), 4.25 (m, 1H, Leu-RCH), 4.13-4.19
(m, 3H, Tyr-RCH, Phe-RCH, H3′), 3.80-3.87 (m, 3H, Val-RCH, Ile-
RCH, H2), 3.46 (m, 1H, H6′), 3.30 (m, 2H, Phe-âCH, Tyr-âCH), 3.00-
3.43 (m, 2H, H1, H1), 2.86 (m, 1H, Tyr-âCH), 2.73 (m, 1H, Phe-
âCH), 2.63 (m, 1H, H6′), 2.06 (s, 3H, acetyl), 1.97 (m, 1H, H4′),1.85
(15) A synthesis of cycle 12 has been reported.12a In this work, 12 was
made in three steps by coupling 3-bromopropylamine to Boc-Ile (94%) with
BOP reagent, deprotecting (TFA), coupling with Boc-Tyr (96%), and
cyclization with base (50%). 12-Amino-8,11-dioxo-9-(1-methylpropyl)-2-
oxa-7,10-diazabicyclo[12.2.2]octadeca-14,16,17-triene (13) was made simi-
larly by coupling 4-aminobutanol to Boc-Ile, converting the hydroxyl to
bromide with CBr4/PPh3,16 deprotecting (TFA), coupling to Boc-Tyr, and
cyclizing (∼50% yield overall). Final deprotection of both cycles was
achieved by stirring in a solution of 25% TFA in DCM at room temperature
for 15 min. The TFA was evaporated in vacuo and the residue dissolved in
saturated NaHCO3 solution and extracted with ethyl acetate.
(17) 3(S)-[[[(3′(S)-Tetrahydrofuranyl)oxy]carbonyl]amino]-1-bromo-4-
phenylbutan-2-one was synthesized in three steps by reacting 3(S)-
hydroxytetrahydrofuran and phenylalanine methyl ester‚HCl according to
the procedure of Ghosh et al.19 (78%), de-esterification of the resultant ester,
preparation of the diazoketone12a (95%), and reaction with HBr (91%).12a
(16) Kocienski, P. J.; Cernigliaro, G.; Feldstein, G. J. Org. Chem. 1977,
42, 353-355.

Peptidomimetics That Inhibit HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 14, 1996 3379
Ile-γCH), 0.86 (t, J ) 5.5 Hz, 3H, Ile-δCH3), 0.78 (d, J ) 5.0 Hz, 3H,
Ile-γCH3). ISMS: m/z 611 (M + H).
Hz, 3H, Ile-γCH3), 0.71 (d, J ) 6.62 Hz, 3H, Val-γCH3). ISMS: m/z
751 (M + H). HRMS: calcd for C43H54N6O6 750.4105, found
750.4097.
(2R,3S,11′S,8′S)-3-[[(3-Methylphenyl)carbonyl]amino]-4-phenyl-
1-[[7′,10′-dioxo-8′-(1-methylpropyl)-2′-oxa-6′,9′-diazabicyclo[11.2.2]-
heptadeca-13′,15′,16′-trien-11′-yl]amino]butan-2-ol (9). The mac-
rocycle 12 (23 mg, 69 µmol) was reacted with 3(S)-[[(3′-methyl-
phenyl)carbonyl]amino]-1-bromo-4-phenylbutan-2-one18 (24 mg, 69
µmol) as described for the synthesis of 5. Purification 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 min (flow rate 1.5 mL/
min) gave a diastereomeric mixture of 9 (7 mg, 16.5%), retention time
) 20.48 min. The NMR spectra indicated the presence of a 1:1 mixture
of diastereomers by a doubling of all resonances at 290 K. The spectra
was consistent with the structure. Several attempts to separate
diastereomers by reverse phase HPLC were unsuccessful, and the
diastereomeric mixture was tested for HIV-1 PR inhibition. ISMS:
m/z 615 (M + H). HRMS: calcd for C36H46N4O5 614.3468, found
614.3467.
X-ray Crystallography. Synthetic6 HIVPR (mutant Q7K, L33I)
was mixed with a DMSO solution of 6 (ratio 1:10). Crystals of the
HIVPR-6 complex were grown at 20 °C by vapor diffusion from 45%
(NH4)2SO4 in 0.1 M acetate buffer (pH 5.4) and were isomorphous
with those of the HIVPR-JG365 complex.6 Crystallographic data were
measured and data refined as described12a with X-PLOR21 using an
inhibitor-HIVPR complex12a (pdb accession code 1CPI) as the starting
model. The program O22 was used for modeling and rebuilding. The
final refined structure, including 118 solvent molecules, has an R-factor
of 0.179 with rms deviations from ideality of 0.010 Å and 1.66°,
respectively, for bond distances and bond angles. This structure
identified the R-configured alcohol of 6 and showed the unusual binding
mode depicted in Figures 1 (blue) and 2. A detailed analysis of this
crystal structure will be published separately. Atomic coordinates for
the structure of the HIVPR-6 complex have been deposited with the
Brookhaven Protein Data Bank (accession code 1MTR).
(2R,3S,11′S,8′S)-3-[(2-Quinolinylcarbonyl)amino]2-propylmethyl]-
carbonyl]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 (10). Compound 6 (3 mg, 5 µmol) was deprotected with
25% TFA in dichloromethane (1 mL) over 15 min and evaporated in
vacuo. The residue was dissolved in DMF (1 mL), and to it was added
quinaldoyl-(S)-valine20 (1.6 mg, 6 µmol), BOP (2.6 mg, 6 µmol), and
DIPEA (3 equiv). Then the mixture was stirred for 1 h at room
temperature. The solvent was evaporated under reduced pressure and
the 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 min to give 10 (3 mg, 79%), retention time ) 22.42
min, as a white powder. 1H NMR (500 MHz, 290 K, CD3OD): δ
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,
Ile-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-RCH, Tyr-RCH, Phe-RCH), 3.78 (m, 1H,
H2), 3.56 (m, 1H, H5′), 3.47 (m, 1H, Ile-RCH), 3.29-3.33 (m, 2H,
Phe-âCH, Tyr-âCH), 3.10 (m, 1H, H1), 3.03 (m, 1H, H1), 2.79 (m,
1H, H5′), 2.53-2.62 (m, 2H, Phe-âCH, Tyr-âCH), 2.25 (m, 1H, H4′),
2.01 (m, 1H, Val-âCH), 1.73 (m, 1H, H4′), 1.58 (m, 1H, Ile-âCH),
1.43 (m, 1H, Ile-γCH), 0.97 (m, 1H, Ile-γCH), 0.85 (d, J ) 6.62 Hz,
3H, Val-γCH3), 0.82 (t, J ) 7.33 Hz, 3H, Ile-δCH3), 0.77 (d, J ) 6.74
Computer Modeling. Energy minimization was carried out using
Discover (version 2.9.6 23a); the CVFF forcefield23b parameters were
automatically assigned. Of the two catalytic aspartic acids (Asp25 and
Asp125),3 Asp25 was selectively ionized23c along with the hydroxy-
ethylamine NH of the inhibitor when the charged model was examined.
For the uncharged modeling all acids and bases were neutral. Protein
atoms were fixed according to the JG365-bound conformation3 of
HIVPR, a dielectric of 1.0 was employed, and the inhibitor atoms and
crystallographic waters were free to move. Calculations were performed
using algorithms steepest descents (down to a gradient of e10 kcal
mol-1 Å-1) followed by conjugate gradients (<1 kcal mol-1 Å-1) and
then Newton Raphson (<0.001 kcal mol-1 Å-1). Charged and
uncharged models were compared with the crystal structure using an
rms comparison over all heavy atoms using Insight II (version 2.3.523a).
Acknowledgment. Financial support was provided by the
Commonwealth AIDS Research Grant Scheme, National Health
and Medical Research Council, Australian Research Council,
and an Australian R&D Syndicate. Dianne Alewood (enzyme
synthesis), Trudy Bond (amino acid analyses), and Drs. Martin
Scanlon (NMR spectra), Michael Dooley and Mark Smythe
(modeling advice) are gratefully acknowledged for their assistance.
JA953790Z
(18) 3(S)-[[(3′-Methylphenyl)carbonyl]amino]-1-bromo-4-phenylbutan-
2-one was synthesized in three steps by coupling m-toluic acid and
phenylalanine methyl ester‚HCl with BOP (100%), de-esterification of the
resultant ester with base, preparation of the diazoketone12a (60%), and
reaction with HBr (91%).12a
(19) Ghosh, A. K.; Duong, T. T.; McKee, S. P.; Thompson, W. J.
Tetrahedron Lett. 1992, 33, 2781-2784.
(20) Quinaldyl-(S)-valine was synthesised by the BOP coupling of
quinoline-2-carboxylic acid with valine methyl ester‚HCl followed by de-
esterification with NaOH.
(21) Bru¨nger, A. T. X-PLOR (Version 3.1) Manual, The Howard Hughes
Medical Institute and Department of Molecular Biophysics and Biochem-
istry, Yale University, CT 06511, 1992.
(22) Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard, M. Acta
Crystallogr., Sect. A 1991, 47, 110-119.
(23) (a) Biosym Technologies Inc., San Diego. (b) Dauber-Osguthorpe,
P.; Roberts, V. A.; Osguthorpe, D. J.; Wolff, J.; Genest, M.; Hageler, A. T.
Proteins 1988, 4, 31-47. (c) Chen, X.; Tropsha, A. J. Med. Chem. 1995,
38, 42-48.