Irreversible inhibition of the HIV-1 protease: Targeting alkylating agents to the catalytic aspartate groups

Article abstract of DOI:10.1021/ja954069w

Irreversible inhibition of the HIV-1 protease by agents that specifically alkylate its catalytic aspartate residues is a potentially useful approach for circumventing the evolution of HIV strains that are resistant to protease inhibitors. Five haloperidol- and two FMOC-based epoxides of differing reactivities have been synthesized and tested as irreversible inhibitors of the HIV-1 protease (HIV-1 PR). Of these, two trisubstituted epoxides, a cis-1,2-disubstituted epoxide, a 1,1-disubstituted epoxide, and a monosubstituted epoxide function as irreversible inhibitors, but two trans-1,2-disubstituted epoxides do not. The most effective of the epoxides (6) inactivates HIV-1 PR with K(inact) = 65 μM and V(inact) = 0.009 min^-1. 1,2-Epoxy-3-(p-nitrophenoxy)propane (EPNP), a nonspecific inactivating agent for aspartyl proteases, has been used to validate a protocol for establishing the stoichiometry and site of protein alkylation. Mass spectrometric analysis of the inactivated enzyme shows that one molecule of either EPNP or the cyclic 1,2-disubstituted epoxide 6 is covalently bound per HIV-1 PR dimer. Mass spectrometric sequencing of labeled proteolytic peptides shows that both inhibitors are covalently bound to a catalytic aspartate residue. The covalent binding of three α,β-unsaturated ketone derivatives of haloperidol has been similarly examined. Analysis of HIV-1 PR inactivated by these agents establishes that they bind covalently to the two cysteines and the N-terminal amino group but not detectably to the catalytic aspartate residues. The results indicate that aspartate-targeted inactivation of HIV-1 PR depends on (a) matching the reactivity of the alkylating functionality to that of the aspartates, preferably by exploiting the two-aspartate catalytic motif of the protease to activate the alkylating agent, and (b) appropriate positioning of the alkylating functionality within the active site. These requirements are readily met by a monosubstituted, 1,1-disubstituted, or cyclic cis-1,2-disubstituted epoxide but not by trans-1,2-disubstituted epoxides or α,β-unsaturated ketones.

Full text of DOI:10.1021/ja954069w

5846
J. Am. Chem. Soc. 1996, 118, 5846-5856
Irreversible Inhibition of the HIV-1 Protease: Targeting
Alkylating Agents to the Catalytic Aspartate Groups
Zhonghua Yu,† Patricia Caldera,† Fiona McPhee,† James J. De Voss,†
Patrick R. Jones,‡ Alma L. Burlingame,† Irwin D. Kuntz,† Charles S. Craik,† and
Paul R. Ortiz de Montellano*,†
Contribution from the Department of Pharmaceutical Chemistry, UniVersity of California,
San Francisco, California 94143-0446, and Department of Chemistry, UniVersity of the Pacific,
Stockton, California 95211
ReceiVed December 4, 1995X
Abstract: Irreversible inhibition of the HIV-1 protease by agents that specifically alkylate its catalytic aspartate
residues is a potentially useful approach for circumventing the evolution of HIV strains that are resistant to protease
inhibitors. Five haloperidol- and two FMOC-based epoxides of differing reactivities have been synthesized and
tested as irreversible inhibitors of the HIV-1 protease (HIV-1 PR). Of these, two trisubstituted epoxides, a cis-1,2-
disubstituted epoxide, a 1,1-disubstituted epoxide, and a monosubstituted epoxide function as irreversible inhibitors,
but two trans-1,2-disubstituted epoxides do not. The most effective of the epoxides (6) inactivates HIV-1 PR with
Kinact ) 65 µM and Vinact ) 0.009 min-1. 1,2-Epoxy-3-(p-nitrophenoxy)propane (EPNP), a nonspecific inactivating
agent for aspartyl proteases, has been used to validate a protocol for establishing the stoichiometry and site of protein
alkylation. Mass spectrometric analysis of the inactivated enzyme shows that one molecule of either EPNP or the
cyclic 1,2-disubstituted epoxide 6 is covalently bound per HIV-1 PR dimer. Mass spectrometric sequencing of
labeled proteolytic peptides shows that both inhibitors are covalently bound to a catalytic aspartate residue. The
covalent binding of three R,â-unsaturated ketone derivatives of haloperidol has been similarly examined. Analysis
of HIV-1 PR inactivated by these agents establishes that they bind covalently to the two cysteines and the N-terminal
amino group but not detectably to the catalytic aspartate residues. The results indicate that aspartate-targeted
inactivation of HIV-1 PR depends on (a) matching the reactivity of the alkylating functionality to that of the aspartates,
preferably by exploiting the two-aspartate catalytic motif of the protease to activate the alkylating agent, and (b)
appropriate positioning of the alkylating functionality within the active site. These requirements are readily met by
a monosubstituted, 1,1-disubstituted, or cyclic cis-1,2-disubstituted epoxide but not by trans-1,2-disubstituted epoxides
or R,â-unsaturated ketones.
The AIDS epidemic has triggered a massive effort to develop
to the treatment of AIDS. This search is currently concentrated
on the development of inhibitors of the HIV protease,4,5
therapeutic strategies to halt the progress of the disease. The
focus of these efforts has been the human immunodeficiency
virus (HIV), the etiologic agent for the syndrome.1 As for all
retroviruses, the RNA of HIV is transcribed into DNA by a
virus-encoded reverse transcriptase and is incorporated into the
host genome by a viral integrase. Subsequent expression of
the virus by the host cell produces polyprotein precursors that
are processed by an HIV-encoded protease into functional capsid
proteins and enzymes.2 The maturation of the precursor proteins
occurs at the plasma membrane during particle release, ulti-
mately resulting in infectious viral particles.3
a
relatively small protein for which many crystal structures are
available.6 The HIV protease is a member of the aspartyl
protease family7 but differs from most members of the family
in that it exists as a dimer of two identical 99-amino acid chains
rather than as a monomeric polypeptide.8,9 Each monomer
provides one of the two active-site aspartates required for the
catalytic action of the dimer. Many competitive inhibitors of
this protease are now available,10-13 several of which are in
clinical trials, but there is growing evidence for the rapid rise
of strains that encode mutant proteases resistant to reversible
protease inhibitors.14,15
HIV infections are currently treated primarily with inhibitors
(e.g., AZT) of the viral reverse transcriptase. The toxicity of
these reverse transcriptase inhibitors and the rapid development
of resistant strains have led to a search for alternative approaches
(4) West, M. L.; Fairlie, D. P. Trends Pharmacol. Sci. 1995, 16, 67-
75.
(5) Tomasselli, A. G.; Howe, W. J.; Sawyer, T. K.; Wlodawer, A.;
Heinrikson, R. L. Chimicaoggi, 1991, 9, 6-27.
* To whom correspondence should be addressed. Phone: (415) 476-
2903. FAX: (415) 502-4728 or (415) 476-0688. E-mail: ortiz@cgl.ucsf.edu.
† University of California.
(6) Darke, P. L.; Huff, J. R. AdV. Pharmacol. 1994, 25, 399-454.
(7) Seelmeier, S.; Schmidt, H.; Turk, V.; von der Helm, K. Proc. Nat.
Acad. Sci. U.S.A. 1988, 85, 6612-6616.
‡ University of the Pacific.
(8) Navia, M. A.; Fitzgerald, P. M. D.; McKeever, B. M.; Leu, C. T.;
Heimbach, J. C.; Herbver, W. K.; Sigal, I. S.; Darke, P. L.; Springer, J. P.
Nature 1989, 337, 615-620.
X Abstract published in AdVance ACS Abstracts, June 1, 1996.
(1) (a) Barre-Sinoussi, F.; Chermann, J.-C.; Rey, F.; Nugeyre, M.;
Chamaret, S.; Gruest, J. G.; Dauguet, C.; Axler-Blin, C.; Brun-Vezinet, F.;
Rouzioux, C.; Rozenbaum, W.; Montagnier, L. Science 1983, 220, 868-
871. (b) Gallo, R.; Salahuddin, S.; Popovic, M.; Shearer, G.; Kaplan, M.;
Haynes, B.; Palker, T.; Redfield, R.; Oleske, J.; Safai, B.; White, G.; Foster,
P.; Markham, P. Science 1984, 224, 500-503.
(9) Wlodawer, A.; Miller, M.; Kaskolski, 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.
(10) Rose´, J. R.; Craik, C. S. Am. J. Respir. Crit. Care Med. 1994, 150,
S176-S182.
(2) Debouck, C.; Gorniak, J. G.; Strickler, J. E.; Meek, T. D.; Metcalf,
B. W.; Rosenberg, M. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 8903-8906.
(3) Stephens, E.; Compons, R. W. Annu. ReV. Microbiol. 1988, 42, 489-
516.
(11) Meek, T. Enzyme Inhib. 1992, 6, 65-98.
(12) Wlodawer, A.; Erickson, J. W. Annu. ReV. Biochem. 1993, 62, 543-
585.
(13) Clare, M. Perspect. Drug DiscoVery Des. 1993, 1, 49-68.
S0002-7863(95)04069-8 CCC: $12.00 © 1996 American Chemical Society

IrreVersible Inhibition of the HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5847
the key features required to target irreversible inhibitors to the
active site aspartyl residues of the HIV proteases, we have (a)
synthesized seven additional epoxides with a range of chemical
reactivities and evaluated their activities as irreversible inhibitors
of HIV-1 PR and (b) undertaken mass spectrometric analyses
of HIV-1 PR inactivated by five different agents to define the
stoichiometry of the inactivation process and identify the amino
acid(s) alkylated by each agent. The sensitivity and ability to
provide unambiguous answers make mass spectrometry the
method of choice for this purpose.24 The results provide
important information for the development of irreversible
inhibitors that specifically target the catalytic aspartates rather
than amino acids that can be mutated with the concomitant
development of viral strains resistant to the inhibitors.
Figure 1. Structures of the R,â-unsaturated ketone derivatives of
haloperidol and EPNP.
One strategy for minimizing the development of resistance
is to develop irreversible rather than reversible inhibitors of the
HIV protease. Irreversible inhibitors are less sensitive to
mutations that decrease reversible binding affinity because even
a low degree of active site occupancy can lead in time to
complete inactivation of the protein. The catalytic aspartate
residues are the ideal target for such irreversible inhibitors
because their mutation results in complete loss of catalytic
activity. This is clearly shown by the early demonstration that
a D25N mutation of HIV-1 PR (PR ) protease) results in
complete loss of protease activity and hence viral maturation.16
Irreversible inhibition of HIV-1 PR was first achieved with 1,2-
epoxy-3-(p-nitrophenoxy)propane (EPNP),17 a small epoxide
molecule that is a general irreversible inhibitor of aspartyl
proteases.18 EPNP (Figure 1) is a nonspecific and relatively
weak irreversible inhibitor of both HIV-1 PR (Kinact ) 9.9 mM,
Vinact ) 0.060 min-1) and the HIV-2 protease (HIV-2 PR; Kinact
) 6.7 mM, Vinact ) 0.048 min-1).19 An analysis of the binding
of EPNP to HIV-2 PR19 and a crystal structure of the adduct of
EPNP with the simian immunodeficiency virus (SIV) protease20
indicate that EPNP is bound to a catalytic aspartate residue in
both of these proteins. A tripeptide with a cis-1,2-disubstituted
epoxide functionality has also been reported to inactivate HIV-1
Results
Inhibitor Design and Synthesis. Seven new epoxide deriva-
tives have been synthesized. Four of the epoxides were
constructed using the haloperidol framework for delivery of the
reactive group to the HIV-1 PR active site. We have shown in
earlier work that a wide range of haloperidol derivatives inhibit
both HIV-1 PR and HIV-2 PR.22,25 Two crystal structures have
been reported which show that a single haloperidol derivative
can bind in two very different orientations within the active
site of the protein.26 The other three epoxides were incorporated
into frameworks based on a new class of FMOC derivatives
identified as reversible inhibitors of the HIV proteases (unpub-
lished results). A crystal structure of the lead compound for
this series demonstrates that the FMOC agents are also bound
within the active site of the protein (unpublished result). The
parent haloperidol and FMOC compounds are modest reversible
inhibitors of the proteases, with IC50 values in the high
nanomolar to low micromolar range. These frameworks were
chosen for the work described here in preference to frameworks
with tighter reversible binding affinities because they better
model the affinities that might be expected in inhibitor-resistant
variants of the HIV protease.
Of the four epoxides based on the haloperidol framework,
two (1, 2) are trans-1,2-disubstituted epoxides, one (3) is a
trisubstituted epoxide, and one (4) is a monosubstituted epoxide.
Of the three additional compounds based on the FMOC motif
one (5) is a trisubstituted epoxide, one (6) a cyclic cis-1,2-
disubstituted epoxide, and one (7) a 1,1-disubstituted epoxide.
In general, the chemical reactivities of epoxides toward SN2-
type reactions decrease in the order monosubstituted > 1,1-
disubstituted > 1,2-cis-disubstituted > 1,2-trans-disubstituted
due to steric factors. However, for steric reasons, 1,2-disub-
stituted epoxides in which the substituents are part of a six-
membered ring are less reactive than their acyclic 1,2-
disubstituted counterparts. Within each steric reactivity class,
the reactivities of the epoxides are further modulated by
electronic factors, epoxides with conjugating or electron-
releasing functionalities being more reactive than epoxides with
simple alkyl substituents.
PR with Kinact ) 20 µM and Vinact ) 0.31 min-1 21
.
Our search for irreversible inhibitors of the HIV proteases
has concentrated on derivatives of haloperidol, a lead structure
identified as a reversible inhibitor through computer-assisted
inhibitor design methods.22,23 We have shown in earlier work
that haloperidol analogues with a 1,1-disubstituted epoxide or
an R,â-unsaturated carbonyl group as the reactive center (Figure
1) irreversibly inhibit HIV-1 PR and HIV-2 PR.19,23 The
generally higher activity of these compounds as inactivators of
HIV-1 PR than HIV-2 PR suggests that they bind covalently to
several amino acids, but the actual nature of the irreversible
inhibition was not established. As part of an effort to identify
(14) Otto, M. J.; Garber, S.; Winslow, D. L.; Reid, C. D.; Aldrich, P.;
Jadhav, P. K.; Patterson, C. E.; Hodge, C. N.; Cheng, Y.-S. E. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 7543-7547.
(15) Condra, J. H.; Schleif, W. A.; Blahy, O. M.; Gabryelski, L. J.;
Grayam, D. J.; Quintero, J. C.; Rhodes, A.; Robbins, H. L.; Roth, E.;
Shivaprakash, M.; Titus, D.; Yang, T.; Teppler, H.; Squires, K. E.; Deutsch,
P. J.; Emini, E. A. Nature 1995, 374, 569-571.
(16) 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.
(17) Meek, T. D.; Dayton, B. D.; Metcalf, B. W.; Dreyer, G. B.; Strickler,
J. E.; Gorniak, J. G.; Rosenberg, M.; Moore, M. L.; Magaard, V. W.;
Debouck, C. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 1841-1845.
(18) Tang, J. J. Biol. Chem. 1971, 246, 4510-4517.
(19) Salto, R.; Babe´, L. M.; Li, J.; Rose, J. R.; Yu, Z.; Burlingame, A.;
De Voss, J. J.; Sui, Z.; Ortiz de Montellano, P.; Craik, C. S. J. Biol. Chem.
1994, 269, 10691-10698.
(20) Rose´, J. R.; Salto, R.; Craik, C. S. J. Biol. Chem. 1993, 268, 11939-
11945.
trans-Epoxides. Direct epoxidation of unsaturated haloperi-
dol analogues is difficult because the reaction also results in
oxidation of the piperidine ring nitrogen. The two trans-
epoxides were therefore synthesized by prefabricating a fragment
bearing the epoxide moiety and attaching it to 4-(p-chlorophen-
(21) Grant, S. K.; Moore, M. L.; Fakhoury, S. A.; Tomaszek, T. A.;
Meek, T. D. Bioorg. Med. Chem. Lett. 1992, 2, 1441-1445.
(22) DesJarlais, R.; Seibel, G.; Kuntz, I.; Furth, P.; Alvarez, J.; Ortiz de
Montellano, P.; DeCamp, D.; Babe´, L.; Craik, C. S. Proc. Natl. Acad. Sci.
U.S.A. 1990, 87, 6644-6648.
(23) De Voss, J. J.; Sui, Z.; DeCamp, D. L.; Salto, R.; Babe´, L. M.;
Craik, C. S.; Ortiz de Montellano, P. R. J. Med. Chem. 1994, 37, 665-
673.
(24) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Anal.Chem. 1994,
66, 634R.
(25) Lee, K. H.; McPhee, F.; Craik, C. S.; Ortiz de Montellano, P. R. J.
Org. Chem. 1994, 59, 6194-6199.
(26) Rutenber, E.; Fauman, E. B.; Keenan, R. J.; Fong, S.; Furth, P. S.;
Ortiz de Montellano, P. R.; Meng, E.; Kuntz, I. D.; DeCamp, D. L.; Salto,
R.; Rose´, J. R.; Craik, C. S.; Stroud, R. M. J. Biol. Chem. 1993, 268, 15343-
15346.

5848 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Scheme 1. Synthesis of Epoxides 1 and 2
Yu et al.
Scheme 3. Synthesis of Epoxides 4 and 7
Table 1. Inactivation of HIV-1 PR by Epoxides Synthesized in
This Study
inhibitor
[I] (µM)
t1/2 min
Kinact (µM)
Vinact (min-1
)
3
4
5
6
7
200
1000
200
100
120
700
400
600
110
800
NDa
1800
ND
65
ND
0.001
ND
0.009
0.0003
140
a ND ) not done.
Scheme 2. Synthesis of Epoxides 3, 5, and 6
80% yield by mCPBA epoxidation of the FMOC derivative of
1,2,5,6-tetrahydropyridine (Scheme 2).
Monosubstituted and 1,1-Disubstituted Epoxides. The
monosubstituted epoxide 4 was prepared by oxalyl chloride-
mediated condensation of 4-(p-chlorophenyl)-4-hydroxypiperi-
dine with 4-vinylbenzoic acid (Scheme 3). The vinyl group
was then oxidized to the epoxide with mCPBA (40% yield).
Commercially available FMOC-4,5-dehydro-Leu-OH was con-
verted to the corresponding methyl ester with diazomethane.
Epoxidation of the resulting ester with mCPBA provided 7 as
a mixture of two stereoisomeric products (83% yield).
Enzyme Inactivation. The trans-1,2-disubstituted epoxides
1 and 2 do not detectably inactivate HIV-1 PR, but surprisingly,
the trisubstituted epoxides 3 and 5 do slowly inactivate the
enzyme in a time-dependent fashion. The t1/2 values for
inactivation of HIV-1 PR by these two agents are 700 and 600
min, respectively (Table 1). These values are comparable to
those for the terminally unsubstituted epoxides 4 and 7. The
FMOC-derived cis-epoxide 6 inactivates the enzyme ap-
proximately 4-8 times more rapidly than the trisubstituted
epoxides (Table 1). Solubility limitations restrict the concentra-
tions of these agents that can be achieved, but approximate Kinact
and Vinact values have been obtained for compounds 4, 6, and 7
(Table 1). These values confirm that compound 6 is the most
effective inhibitor within this series of epoxides.
yl)-4-hydroxypiperidine (Scheme 1). Thus, oxidation of cinammyl
bromide provided the corresponding bromo epoxide (80% yield)
which condensed with the piperidine fragment under conditions
mild enough to prevent decomposition of the epoxide to give 1
in 16% isolated yield. In an analogous reaction sequence,
epoxidation of the double bond of 4-bromo-1-phenyl-1-butene
generated from R-cyclopropylbenzyl alcohol and trimethylsilyl
bromide (Scheme 2), followed by reaction with the piperidine
fragment, produced 2 in low yield. The major product (60%)
is formed by addition of the piperidine to the benzylic terminus
of the epoxide followed by base-catalyzed elimination of water
and HBr to give a diene product (Scheme 1).
cis-1,2-Disubstituted and Trisubstituted Epoxides. De-
hydration of 4-(p-chlorophenyl)-4-hydroxypiperidine yielded
4-(p-chlorophenyl)-1,2,5,6-tetrahydropyridine which was con-
verted to its FMOC derivative (Scheme 2). Epoxidation of the
dihydropyridine double bond with mCPBA produced epoxide
5 in 65% yield. Deprotection of 5 with piperidine in DMF
followed by K2CO3-catalyzed reaction of the deprotected
nitrogen with 4-iodo-4′-fluorobutyrophenone afforded a one-
pot synthesis of epoxide 3 (64% yield). 4-Iodo-4′-fluorobutyro-
phenone was employed because 4-chloro-4′-fluorobutyrophe-
none was not sufficiently reactive. The iodo derivative was
generated in situ by reaction of cyclopropyl 4-fluorophenyl
ketone with trimethylsilyl iodide. Epoxide 6 was obtained in
The kinetic values for the inactivation of HIV-1 PR by EPNP
and conjugated ketones 8-10 were determined in an earlier
study and are as follows: EPNP, Kinact ) 9900 µM, Vinact
0.06 min-1; 8, Kinact ) 62 µM, Vinact ) 0.380 min-1; 9, Kinact
10.7 µM, Vinact ) 0.057 min-1; 10, Kinact ) 57 µM, Vinact
)
)
)
0.232 min-1 19
.
These results show that the Kinact value for 6,
the best epoxide, is comparable to those for 8 and 10 but
somewhat worse than that for 9. On the other hand, the Vinact
value for 6 is smaller than those for all three unsaturated ketones.
Inactivation by EPNP. EPNP is a nonspecific irreversible
inhibitor of classic aspartyl proteases that binds to the active-
site aspartates.18 However, it is not known whether the
inactivation of HIV-1 PR, as found for the HIV-2 and SIV
proteases,19,20 involves covalent binding to the catalytic aspar-
tates. A detailed analysis of the HIV-1/EPNP adduct has
therefore been carried out both to resolve this ambiguity and to
validate with a well-understood agent the methodology for the
characterization of protease/inhibitor complexes.

IrreVersible Inhibition of the HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5849
Figure 3. CID spectrum of the EPNP-modified peptide isolated after
digestion of the modified protease. Peak assignments are based on the
nomenclature of Roepstorff42 and Biemann.43 The interpretation, shown
on the panel above the spectrum, identifies Asp-25 as the site of covalent
modification.
195 Da higher than that of the intact protease monomer, a
difference that corresponds exactly to the molecular weight of
EPNP. This product clearly corresponds to an adduct of one
molecule of EPNP with a subunit of HIV-1 PR. The presence
of the two peaks in approximately equal amounts suggests that
covalent attachment of EPNP to one subunit of the protease
homodimer inactivates the protein and prevents modification
of the second subunit. The stoichiometry of inactivation of
HIV-1 PR by EPNP is thus one inhibitor molecule per protease
dimer. In accord with this conclusion, HPLC analysis of
partially inactivated enzyme shows that both peaks are still
present but the peak that corresponds to the unmodified subunit
is larger than that for the modified subunit and the ratio of these
two peaks approaches a value of 1 as the enzyme approaches
complete inactivation.
In order to determine the site to which EPNP is covalently
attached, the inactivated protease was digested with trypsin or
pepsin. The resulting peptides were separated by HPLC, and
their molecular masses were determined by HPLC/ESMS.
Pepsin, which is not as specific as trypsin, generated peptides
of appropriate size for further analysis. Identification of the
modified peptides was achieved both by comparing the HPLC
profile obtained from the modified protease with that from the
unmodified enzyme and by using mass spectrometric peptide
mapping. Figure 3 shows the high-energy CID spectrum of an
EPNP-modified peptide generated by digestion of the modified
protease. This spectrum establishes the sequence of the peptide
and identifies the site of the modification. As shown in the
upper panel, the fragmentation pattern of this peptide unambigu-
ously establishes its sequence as Leu-Asp-Thr-Gly-Ala-Asp,
with EPNP bound to the second residue. The y series of ions
at 134, 205, 262, 363, and 673 and the b series of ions at 424,
525, 582, and 653 define the sequence and the site of
modification. In addition, the ion at m/z 283, labeled as J,
represents the immonium ion of the modified aspartic acid. As
this sequence encompasses amino acids 24-29 of HIV-1 PR,
the site of the modification is the catalytic residue Asp-25.
Figure 2. (a, top) HPLC profile of the HIV-1/EPNP complex. Two
major peaks are observed by HPLC UV detection. The masses of these
peaks have been determined by ESIMS, as shown in panels b and c.
(b, middle) ESIMS spectrum of the first HPLC peak shown in panel a.
A molecular mass of 10 789 Da is obtained. (c, bottom) ESIMS
spectrum of the second peak in panel a. A molecular mass of 10 984
Da is obtained.
HPLC analysis of HIV-1 PR after it is fully inactivated by
EPNP yields two well-resolved peaks (Figure 2a). The mass
spectrometrically determined molecular mass of the first HPLC
peak is 10 789 Da (Figure 2b), which corresponds to the mass
of the intact protease monomer (theoretical MW 10 790). The
molecular mass of the second peak, 10 984 Da (Figure 2c), is
Inactivation by Compound 6. HPLC of HIV-1 PR after
complete inactivation by compound 6 demonstrates the presence
of two peaks of equal size (Figure 4a). The first peak has the

5850 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Yu et al.
Figure 5. CID spectrum of a labeled peptide from the HIV-1 protease
inactivated by 6. The peptide sequence is Asp-Thr-Gly-Ala-Asp, with
the inhibitor attached to the first Asp (Asp-25).
established that the first peak has a molecular mass of 11 144
Da (Figure 6b), 354 Da higher than that of the intact protease
monomer. As this corresponds to the molecular weight (MW
353) of compound 8, it appears that one molecule of 8 is
attached to the protease in the most rapidly eluting adduct. The
second peak has a molecular mass of 11 498 Da (Figure 6c),
708 Da higher than that of the intact protease. This corresponds
to the covalent attachment of two molecules of 8 to the protease
monomer. The third peak exhibits a molecular mass of 11 854
Da (Figure 6d). The difference from the mass of the intact
protease monomer is 1064 Da, which corresponds to covalent
binding of three molecules of 8. In contrast to the reaction with
EPNP, no intact protease is detected.
Figure 4. (a, top) HPLC profile of the product of the reaction of HIV-1
with compound 6. Mass spectrometry identifies the first of the two
peaks as the intact protease monomer. (b, bottom) ESIMS spectrum of
the second peak in Figure 5a, which has a molecular mass of 11 111
Da.
The three HPLC fractions were digested by pepsin, and the
resulting peptides were separated by HPLC and analyzed by
HPLC/ESMS. One modified peptide of m/z 1349.6 was
obtained from the first fraction, and its sequence, as determined
by CID analysis (Figure 7a), is Thr-Gln-Ile-Gly-Cys-Thr-Leu-
Asn-Phe. A molecule of inhibitor 8 is bound to the cysteine
residue. This sequence, which corresponds to amino acids 91-
99 of the protease, identifies Cys-95 as the site of the
modification. Analysis of the second HPLC fraction provided,
in addition to the m/z 1349.6 peptide, a new modified peptide
of m/z 1252.6. The sequence of this second peptide was
established by CID analysis as Ile-Cys-His-Lys-Ala-Ile-Gly-
Thr, with the inhibitor bound to the cysteine (Figure 7b). As
this sequence matches that of residues 66-74 of HIV-1 PR,
the site of the alkylation is Cys-67. Analysis of the third HPLC
fraction provided, in addition to the two modified peptides
already discussed, a modified peptide of m/z 923.5. CID
analysis identifies this peptide as a Pro-Gln-Ile-Thr-Leu with a
molecule of the inhibitor bound to the proline. This sequence
is that of the N-terminus of the protease, and the site of
alkylation is the N-terminal proline amino function.
same retention as the intact protease monomer, and mass
spectrometric analysis showed that it is indeed the intact
protease. The second peak shows a molecular mass of 11 111
Da (Figure 4b), which is 321 Da higher than that of the intact
protease monomer. The difference in mass corresponds exactly
to the molecular weight of 6. Covalent binding of one inhibitor
molecule to only one of the two subunits of the fully inactivated
protease dimer again establishes a stoichiometry of one inhibitor
per protease dimer. Cleavage of the intact and modified protease
with pepsin and separation of the resulting peptides by HPLC
reveal differences in the two HPLC profiles. The peaks unique
to the modified protease were analyzed by MS and sequenced
by tandem MS (Figure 5). As shown by the CID spectrum of
the modified peptide, its sequence is Asp-Thr-Gly-Ala-Asp, with
6 attached to the first aspartic acid residue. This sequence,
which corresponds to residues 25-29 of HIV-1 PR, establishes
that the inhibitor is attached to Asp-25. As with EPNP, analysis
of the partially inactivated protease shows that the ratio of the
modified and unmodified subunits is lower than 1 but ap-
proaches that value as the enzyme becomes fully inactivated.
This confirms that modification of Asp-25 is responsible for
the inactivation.
The extent of modification could be deduced from the size
of the HPLC peaks. Cys-95 was modified completely since it
was present in all three peaks. Cys-67 was present in both the
second and third peaks, which represent about 70% of the total
material recovered. Therefore, roughly 2/3 of Cys-67 was
modified. The third HPLC peak represented less than 10% of
the total area, so the N-terminus of the protease was only
modified by 8 to a limited extent. No modification of Asp-25
was detected.
Inactivation by Compound 8. HPLC analysis of the fully
inactivated protease after reaction with 8 reveals the presence
of three peaks of different sizes (Figure 6a). Mass spectrometry

IrreVersible Inhibition of the HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5851
Figure 6. (a, top left) HPLC profile of the product of reaction of HIV-1 with compound 8. The molecular masses of the three major peaks,
determined by ESIMS, are shown in panels b-d. (b, top right) ESIMS spectrum of the first peak in panel a, which has a molecular mass of 11 144
Da. (c, bottom left) ESIMS spectrum of the second peak in panel a, which has a molecular mass of 11 498 Da. (d, bottom right) ESIMS spectrum
of the third peak in panel a, which has a molecular mass of 11 854 Da.
Inactivation by Compounds 9 and 10. HPLC purification
of the complex after inactivation by compound 9 showed one
major HPLC peak with an ESIMS molecular mass of 11 651
Da. The difference of 861 Da between this mass and that of
the protease monomer corresponds to the mass of two molecules
of the inhibitor. Trypsin digestion followed by HPLC provided
two modified peptides, one with a mass spectrometric MH+ of
m/z 1958.9 and the other with a mass spectrometric MH+ of
m/z 1765.8. Mass mapping results indicate that these two
masses correspond to the peptides Gln-58 to Lys-70 and Asn-
88 to Phe-99, each with one covalently bound inhibitor moiety.
MS/MS experiments were carried out to identify the modifica-
tion sites. The CID spectrum of the precursor ion at m/z 1958.9
confirmed the sequence Gln-Tyr-Asp-Gln-Ile-Pro-Val-Glu-Ile-
Cys-Gly-His-Lys, with the inhibitor attached to Cys-67. The
b-series ions revealed the sequence, whereas several fragment
peaks defined the modification site. Similarly, the CID spectrum
of the precursor ion at m/z 1765.8 showed its sequence to be
Asn-Leu-Leu-Thr-Gln-Ile-Gly-Cys-Thr-Leu-Asn-Phe, with com-
pound 8 bound to Cys-95.
HPLC analysis of the protease after inactivation by compound
10 provided one major product peak whose molecular mass
(11 540 Da) was 750 Da higher than that of the intact monomer,
a difference that corresponds to the mass of two inhibitor
molecules. Tryptic digestion and MS analysis yielded two
modified peptides with the sequences Gln-58 to Lys-70 and Asn-
88 to Phe-99, each covalently attached to one inhibitor molecule.
CID analysis of these peptides identified the sites of modification
of these peptides as Cys-67 and Cys-95, respectively.
Discussion
A shadow has been cast on the promise of the potent
inhibitors of the HIV proteases now available by the finding
that HIV becomes resistant to the inhibitors due to the evolution
of strains encoding protease variants that are resistant to the
inhibitors.14,15 In light of this development, we are investigating
approaches to irreversible inhibition of the enzyme because
irreversible inhibition is less susceptible to the development of
resistance. Mutation-dependent decreases in the affinity of the
enzyme for inhibitors are less critical for irreversible than
reversible agents because the degree of reversible active site

5852 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Yu et al.
Figure 7. (a, top left) CID spectrum of a modified peptide isolated from the peptic digest of the first peak in Figure 6a. This peptide covers amino
acid residues 91-99 of HIV-1 PR and includes a modified Cys-95. (b, top right) CID spectrum of a modified peptide isolated from the peptic digest
of the second peak in Figure 6a. The peptide corresponds to amino acids 66-74 of the protease with covalent modification of Cys-67. (c, bottom)
CID spectrum of a modified peptide isolated from the third peak in Figure 6a. This peptide corresponds to the N-terminus of the protease with the
inhibitor bound to the N-terminus.
occupancy need not be high and need not be continuous to
achieve full inactivation of the enzyme. This requires, of course,
that the protein residues targeted by irreversible inhibitors not
be replaced through mutations by unreactive residues. For this
reason, the most desirable amino acids for reaction with
irreversible inhibitors are the catalytic aspartate groups of HIV-1
PR because mutations of these residues suppress catalytic
activity. It is therefore important to define the parameters that
determine whether an irreversible inhibitor reacts with the
catalytic aspartates as opposed to other residues of the HIV
proteases.
Irreversible inhibitors are constructed by incorporating a
reactive functionality, or a latent version of such a functionality,
into the framework of a ligand for the enzyme. The bioavail-
ability problems associated with peptide substrates have led us
to concentrate on non-peptide inhibitor frameworks. In this
study we have synthesized seven new potentially irreversible
inhibitors, four based on a haloperidol framework (1-4) and
three FMOC derivatives (5-7). An epoxide group has been
incorporated into each of the seven structures as the reactive
locus for potential covalent attachment to HIV-1 PR. However,
the intrinsic reactivity of the epoxide group toward SN2-type
addition reactions differs in the seven agents. The least reactive
are the trisubstituted epoxide moieties of 3 and 5, followed by
the trans-1,2-disubstituted epoxide moieties of 1 and 2. The
most reactive are the 1,1-disubstituted epoxide 7 and, particu-
larly, the monosubstituted epoxide 4. The cyclic cis-1,2-
disubstituted epoxide group of 6 is expected to be of interme-
diate intrinsic reactivity. Studies of the activities of the seven
structures show that 1 and 2 have no detectable activity as
irreversible inhibitors of HIV-1 PR, either because the epoxide
is too unreactive or because it is incorrectly positioned when
the compound is bound in the active site. The other five
compounds are slow but irreversible HIV-1 PR inhibitors (Table
1). Of the five active compounds, the cyclic cis-1,2-disubsti-
tuted epoxide 6 is the most effective. The other four com-
pounds, including the trisubstituted epoxides 3 and 5 and the
terminally unsubstituted epoxides 4 and 7, exhibit weaker
activities as irreversible inhibitors.
In order to characterize the inactivation process, we have used

IrreVersible Inhibition of the HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5853
mass spectrometric approaches to define the stoichiometry of
the reaction and to identify the amino acids that are modified
in the process. The major advantages of mass spectrometry
for this purpose are its high efficiency, sensitivity, and ability
to directly characterize protein modifications. The protocol used
in these studies was validated by studying the inactivation of
HIV-1 PR by EPNP, a nonselective irreversible inhibitor of
aspartyl proteases.18,20,27 The same protocol was then used to
characterize the inactivation of HIV-1 PR by 6, the most
effective inhibitor of the series prepared in this study. For both
EPNP and 6, the data clearly establish that irreversible inactiva-
tion involves modification of only one of the two subunits in
the homodimeric enzyme by covalent attachment of one
molecule of the inhibitor. Mass spectrometric sequencing
establishes, furthermore, that both EPNP and 6 bind exclusively
to Asp-25, the catalytic aspartate group.
Figure 8. Proposed mechanism for hydrogen-bonding-assisted alky-
lation of the catalytic aspartate group by 1,2-disubstituted epoxides.
The steric basis for the higher reactivity of cis- relative to trans-
disubstituted epoxides is the steric effect shown schematically in the
figure, in which the alkyl substituents are represented by spheres.
We reported previously that haloperidol derivatives 8-10
with an R,â-unsaturated ketone in the flexible side chain are
irreversible inhibitors of HIV-1 PR.19,23 It was not possible at
that time, however, to define the nature of the inactivation
reaction. We have therefore subjected the protease inactivated
by these agents to the same mass spectrometric analyses used
to investigate inactivation of the enzyme by the epoxides. The
surprising finding from these studies is that inactivation of HIV-
PR by the R,â-unsaturated ketones 8-10 involves (a) modifica-
tion of both subunits of the protease dimer rather than only one
subunit, (b) binding of more than one molecule of the inhibitor
to the protease monomer, and (c) modification of the cysteine
residues (Cys-67 and Cys-95) and to a small extent the
N-terminal amino group rather than the catalytic aspartate group.
As shown by analysis of the enzyme inactivated by 8, Cys-95
is completely alkylated and Cys-67 is alkylated in most of the
subunits of the inactive enzyme. Alkylation of the N-terminal
proline amino group occurs in only a minor fraction of the
protein. Cys-95 is within the substrate binding channel of the
protease, and modification of this residue would be expected
to interfere with substrate binding. It is possible that alkylation
of this residue alone accounts for the loss of catalytic activity.
Cys-67 and the terminal amino group are surface residues, and
their alkylation could interfere with dimerization to give the
catalytically active homodimer. It is possible, however, that
alkylation of Cys-67 does not contribute significantly to the
observed loss of catalytic activity because only the alkylation
of Cys-95 correlates with complete loss of catalytic activity.
No evidence was found for covalent modification of the catalytic
aspartate group by compounds 8-10. However, the earlier work
with these compounds established that they inactivate HIV-2
PR as well as HIV-1 PR, although the inactivation of HIV-2
PR is much less effective.19,23 Thus, for the acetylenic ketone
8 Vinact ) 380 × 10-3 min-1 for HIV-1 PR and 7.6 × 10-3
min-1 for HIV-2 PR. Inactivation of HIV-2 PR, which has no
cysteines, must involve alkylation of other residues. It is not
known whether the low rate of inactivation of HIV-2 PR by
8-10 results from reaction with the catalytic aspartate, but if it
does the differential rates of inactivation of the HIV-1 and HIV-2
enzymes suggest, as found here, that very little if any aspartate
alkylation is likely to occur with HIV-1 PR.
indication that irreversible inhibitors that target the cysteines
of HIV-1 PR may be subject to the development of resistance
due to mutation of the cysteines to unreactive amino acids.19
The trans-1,2-disubstituted epoxide functions in 1 and 2 appear
to be of too low a reactivity to alkylate either the aspartates or
the cysteines. EPNP and the cyclic cis-1,2-disubstituted epoxide
6, on the other hand, appear to be insufficiently reactive to
alkylate the cysteines but readily alkylate one, and only one,
active site aspartate residue. These results support a mechanism
in which the epoxide reactivity is enhanced by hydrogen bonding
to the protonated aspartate in the active site with concomitant
addition of the second aspartate, which is ionized, to the epoxide
ring (Figure 8). The difference in the ability of cis- and trans-
1,2-disubstituted epoxides to inactivate the enzyme presumably
is due to the higher steric hindrance to backside addition of the
carboxyl group to the trans-disubstituted epoxide. This steric
factor, however, only accounts for part of the reactivity
difference because two trisubstituted epoxides also slowly
inactivate the enzyme. The unsaturated ketones 8-10, in
contrast, are too reactive toward sulfhydryl groups and primarily
inactivate HIV-1 PR by reacting with the cysteine residues. The
orientation of the inhibitor when reversibly bound in the active
site is likely to contribute to the specificity of the inactivation
reaction, although substrate orientation is unlikely to be of much
importance in the reactions of a compound as small as EPNP.
Furthermore, crystallographic evidence that a haloperidol de-
rivative is readily bound in at least two almost orthogonal
orientations in the active site of HIV-1 PR suggests that there
is sufficient freedom of movement for these compounds within
the active site to allow the catalytic aspartate residues to interact
with most regions of the inhibitors.26
A model for the alkylation specificity found here for HIV-1
PR is provided by the recent finding that the hydrolysis of
epoxides by epoxide hydrolase is a two-step process, the first
step of which is addition of an aspartate to the epoxide to give
an ester adduct.28-31 In the second step, the ester is hydrolyzed
to the diol metabolite with concomitant regeneration of the
enzyme catalyst. Earlier data on the hydrolysis of epoxides
One key factor established by the present studies as an
important determinant of the mechanism of inactivation of
HIV-1 PR by agents that alkylate protein residues is the intrinsic
reactivity of the alkylating functionality. This reactivity must
not be too high, as that favors reaction with the cysteine residues,
or too low, as that prevents reaction with the protein. The
catalytic activity of HIV-2 PR in which both cysteines have
been replaced by unreactive amino acids provides a clear
(28) Lacourciere, G. M.; Armstrong, R. N. J. Am. Chem. Soc. 1993, 115,
10466-10467.
(29) Lacourciere, G. M.; Armstrong, R. N. Chem. Res. Toxicol. 1994,
7, 121-124.
(30) Pinot, F.; Grant, D. F.; Beetham, J. K.; Parker, A. G.; Borhan, B.;
Landt, S.; Jones, A. D.; Hammock, B. D. J. Biol. Chem. 1995, 270, 7968-
7974.
(27) Rose, R. B.; Rose´, J. R.; Salto, R.; Craik, C. S.; Stroud, R. M.
Biochemistry 1993, 32, 12498-12507.
(31) Borhan, B.; Jones, A. D.; Pinot, F.; Grant, D. F.; Kurth, M. J.;
Hammock, B. D. J. Biol. Chem. 1995, 270, 26923-26930.

5854 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Yu et al.
100), 224 (20), 212 (25); HRMS for C20H22ClNO2, calcd 343.1339,
found 343.1337.
showed that cis-disubstituted epoxides are hydrolyzed more
readily than trans-disubstituted epoxides by the membrane-
bound enzyme. For example, the rates for hydrolysis of the
epoxides from cis- and trans-stilbene by the microsomal epoxide
hydrolase differ by a factor of 785.32 In view of the recently
elucidated mechanism of epoxide hydrolysis, the differences in
rate for the processing of cis- and trans-disubstituted epoxides
provide a model for the selectivity observed in this study of
HIV-1 PR.
The present results indicate that efforts to construct clinically
useful inactivating agents for HIV-1 PR should incorporate a
functionality with the reactivity of a cyclic cis-1,2-disubstituted
epoxide, preferably one which, like the epoxide, is activated
toward nucleophilic addition by specific hydrogen bonding with
one of the catalytic aspartate groups. Moieties with the
reactivity of a conjugated ketone are intrinsically too reactive
to target the aspartate groups and are unsuitable, and those with
the reactivity of a trans-1,2-disubstituted epoxide are too
unreactive and are likely to be ineffective.
trans-4-Phenyl-1-bromo-3-butene. Chlorotrimethylsilane (1.24 g,
11.5 mmol) was slowly added at 0 °C to a suspension of lithium
bromide (1.1 g, 12.6 mmol) in a solution of R-cyclopropylbenzyl alcohol
(1.56 g, 10.5 mmol) in CH2Cl2.34 The mixture was then stirred at room
temperature for 3 h before it was poured into water. The organic phase
was separated, washed with water and brine, and dried over anhydrous
Na2SO4. Solvent evaporation yielded a crude product that appears as
a single spot on TLC. The crude product was filtered through silica
gel to remove polar impurities, yielding 1.41 g (63%) or the desired
product: 1H NMR (CDCl3) 7.20-7.4 (m, 5H, aromatic), 6.4-6.5 (d,
J ) 15 Hz, 1H, ArCH), 6.13-6.23 (dt, J ) 15 Hz, J ) 7 Hz, 1H,
ArCHCH), 3.46 (t, J ) 7 Hz, 2H, CH2Br), and 2.77 (AB q, J ) 7 Hz,
CH2CH2Br) ppm; MS (LREI) m/z (%) 212, 210 (M+, 28, 30), 131 (M+
- Br, 100), 117 (95), 91 (48). The analytical data are consistent with
literature data.35
trans-4-Phenyl-3,4-epoxy-1-bromobutane. A solution of trans-4-
phenyl-1-bromo-3-butene (811 mg, 3.8 mmol) and 800 mg (3.8 mmol)
of mCPBA in a biphasic mixture of CH2Cl2 and 0.5 N NaHCO3 was
stirred for 18 h at ∼25 °C. After workup as described for 1-phenyl-
1,2-epoxy-3-bromopropane, 770 mg of product was isolated (90%
yield): IR (neat) 3037, 2987, 2358, 1771, 1728, 1505, 1468, 1437,
Experimental Section
1
1277 cm-1; H NMR (CDCl3) 7.2-7.4 (m, 5 H, arom H), 3.72 (d, 1
General Procedures. 1,2-Epoxy-3-(p-nitrophenoxy)propane (EPNP)
was supplied by Sigma (St. Louis, MO). The syntheses of the
haloperidol derivatives 8-10 were reported elsewhere.23 Tetrahydro-
furan and ether were dried over sodium/benzophenone and distilled
under argon immediately prior to use. HIV-1 protease was expressed
and purified as described previously.19 The Q7K HIV-1 mutant was
used instead of the wild type enzyme since it is more resistant to
autoproteolysis.27
Melting points were determined with a Thomas capillary melting
point apparatus and are uncorrected. 1H and 13C NMR spectra were
obtained at 300 MHz on a GE QE-300 instrument. Infrared spectra
were recorded on a Nicolet 5DX FT-IR instrument. Mass spectra were
obtained at the University of California, San Francisco, Mass Spec-
trometry Facility. Elemental analyses were performed by the Mi-
croanalytical Laboratory, University of California, Berkeley.
1-Phenyl-1,2-epoxy-3-bromopropane. A solution of 80% m-
chloroperbenzoic acid (mCPBA) (8.5 mmol) in 10 mL of CH2Cl2 was
added to a solution of cinnamyl bromide (571 mg, 2.9 mmol) in 30
mL of CH2Cl2. To the resulting solution was added 20 mL of 0.5 N
NaHCO3, and the heterogeneous mixture was stirred at ∼25 °C for 18
h. A few drops of methyl ethyl sulfide were then added to destroy the
excess peroxide. The resulting mixture was washed with sodium
bicarbonate several times. Evaporation of the solvent provides a crude
residue (787 mg): 1H NMR (CDCl3) 3.31 (m, 1H, CHCH2Br), 3.50-
3.52 (dd, 2H, CH2Br), 3.82 (d, 1H, J ) 2 Hz, CHOCHCH2Br), and
7.3-7.5 (m, 5H, aromatic) ppm; MS m/z (%) 215 (MH+, 6), 213, 171
(18). The analytical data are consistent with literature data.33
4-Hydroxy-4-(p-chlorophenyl)-N-(2,3-epoxy-3-phenylpropyl)pi-
peridine (1). K2CO3 (386 mg, 2.7 mmol) and 4-(p-chlorophenyl)-4-
hydroxypiperidine (485 mg, 2.3 mmol) were added to a solution of
the crude 1-phenyl-1,2-epoxy-3-bromopropane (700 mg, 12.3 mmol)
in DMF, and the mixture was refluxed for 1.5 h. The suspension was
then filtered, and the precipitate was washed with ethyl acetate.
Evaporation of the solvent from the filtrate followed by flash chroma-
tography on silica gel with ethyl acetate as the eluant provided 128
mg of purified material (16% yield): IR (neat) 3383 (br), 2950, 2821,
1678, 1635, 1604, 1493, 1462, 1370, 1098, 1048, 1011, 906, and 820
cm -1; 1H NMR (CDCl3) 7.3-7.5 (m, 9H, aromatic), 3.66 (dd, 1 H, J
) 2 Hz, CHOCHPh), 3.21 (m, 1 H, CHOCHPh), 3.01 (br d, 1 H, J )
11.25 Hz, CHeq HN), 2.87 (dd, 2H, J ) 3.8 Hz, J ) 13 Hz, CHeq-
NCHH), 2.62 [m, 3H, (CHax)2NCHH], 2.15 [m, 2H, (CHax)2COH],
and 1.77 [m, 2H, (CHeq)2COH] ppm; 13C NMR (CDCl3) 146.82, 137.01,
132.81, 128.50, 128.42, 126.08, 125.65 (aromatic), 70.75 (COH), 60.7,
60.5 (CHOCH), 56.9 (NCH2), 49.7, 49.9 [piperidine (CH2)2N], and
38.3 [piperidine (CH2CH2)2N] ppm; MS (LRCI) m/z (%) 344 (MH+,
H, J ) 1.8 Hz, ArCHO), 3.54 (m, 2 H, CH2 Br), 3.10 (td, J ) 1.8 Hz,
J ) 5.3 Hz, 1 H, ArCHOCH), and 2.24 (m, 2 H, CH2CH2Br) ppm;
MS (HREI) m/z (%) 228 (M+, 25), 226, 147 (M+ - Br, 12), 133 (28),
119 (30), 105 (26), 90 (100), 77(30); HRMS for C10H11BrO, calcd
227.9975, found 227.9973.
4-(p-Chlorophenyl)-4-hydroxy-N-(3,4-epoxy-4-phenylbutyl)pi-
peridine (2). A mixture of 702 mg (3.1 mmol) of 4-(p-chlorophenyl)-
4-hydroxypiperidine, 675 mg (3.2 mmol) of trans-4-phenyl-3,4-epoxy-
1-bromobutane, and 441 mg (3.2 mmol) of K2CO3 in DMF was refluxed
for 2.5 h. The suspension was then filtered and the solvent removed
from the filtrate. The residue was redissolved in ether, and the solution
was washed with water and brine and then dried over anhydrous
Na2SO4. Evaporation of solvent provided a residue that was subjected
to flash chromatography on silica gel with ethyl acetate as the solvent.
The desired product was isolated in low yield (62.7 mg, 5%): 1H NMR
(CDCl3) 7.28-7.5 (m, 9 H, aromatic), 4.27 (t, 2 H, J ) 5.9 Hz, NCH2-
CH2), 4.03 [br, 2 H, (CHeq)2N], 3.64 (d, 1 H, J ) 2 Hz, OCHPh), 3.26
(br, 2 H, (CHax)2], 3.05 (br t, 1 H, J ) 5.7 Hz, J ) 2 Hz, CHOCHPh),
2.02 (m, 2 H, NCH2CH2), 1.9 [td, 2 H, J ) 13.4 Hz, J ) 4.55 Hz,
(CHax)2CH2N], and 1.6-1.7 [d, 2 H, J ) 13.4 Hz, (CHeq)2CH2N] ppm;
13C NMR (CDCl3) 155.1, 146.44, 137.17, 132.97, 128.16, 128.43,
125.93, 125.51, 71.03, 62.15, 60.20, 58.33, 39.91, 37.85, and 32.12
ppm; MS (LSIMS) m/z (%) 358 (MH+, 100); (LREI) 357 (M+, 5),
339 (M+ - 18, 10), 238 (M+ - 119, 20), 176 (100), 146 (65). HRMS
for C21H24ClNO2, calcd 357.1496, found 357.1507.
4-(p-Chlorophenyl)-1,2,5,6-tetrahydropyridine. A mixture of 4-(p-
chlorophenyl)-4-hydroxypiperidine (1 g, 4.7 mmol), acetic acid (12 mL),
and hydrochloric acid (6 mL) was refluxed for 2 h. After pouring the
reaction mixture into an ice-water mixture, solid NaOH was added to
make the solution basic. The white precipitate that formed was
collected and washed extensively with water. After drying, 794 mg
1
of white flakes was isolated (85% yield): mp 124-126 °C; H NMR
(CDCl3) 7.29 (m, 4H, aromatic), 6.12 (br s, 1 H, CHCH2NH), 3.28 (d,
2 H, J ) 2.5 Hz, CHCH2NH), 3.10 (dd, 2 H, J ) 5.6 Hz, CH2CH2-
NH), and 2.42 (br s, 2 H, CH2CH2NH) ppm. NMR data were identical
to those of the free base of a commercial sample.
N-FMOC-4-(p-Chlorophenyl)-1,2,5,6-tetrahydropyridine. A so-
lution of 4-(p-chlorophenyl)-1,2,5,6-tetrahydropyridine (456 mg, 2.4
mmol) and 4-(dimethylamino)pyridine (2.4 mmol) in CHCl3 was cooled
to 0 °C. To this solution was slowly added a CHCl3 solution of
9-fluorenylmethyl chloroformate (619 mg, 2.4 mmol). The reaction
mixture was stirred at 0 °C for 30 min and then at ∼25 °C for 2 h
before it was sequentially washed with 1 N HCl, water, and brine.
Evaporation of the solvent after drying over sodium sulfate gave 650
mg of a light yellow solid (65% yield) that was recrystallized from
CHCl3/MeOH: mp 118-120 °C: IR (CHCl3) 3024, 1703, 1493, 1456,
(32) Watabe, T.; Akamatsu, K. Biochim. Biophys. Acta 1972, 279, 297-
305.
(33) Dickinson, J. M.; Murphy, J. A.; Patterson, C. W.; Wooster, N. J.
J. Chem. Soc., Perkin Trans. 1 1990, 1179-1184.
(34) Balme, G.; Fournet, G.; Gore, J. Tetrahedron Lett. 1986, 27, 1907-
1908.
(35) Friedrich, E. C. J. Org. Chem. 1969, 34, 528-534.

IrreVersible Inhibition of the HIV-1 Protease
J. Am. Chem. Soc., Vol. 118, No. 25, 1996 5855
1
1431, 1295, 1215, 1116 cm-1; H NMR (CDCl3) 7.25-7.42 (m, 8 H,
DMF to initiate the reaction. After 30 min a solution of 4-(p-
chlorophenyl)-4-hydroxypiperidine (338 mg, 1.6 mmol) in CH2Cl2 was
added, and the reaction mixture was stirred at ∼25 °C for 1.5 h. The
solution was sequentially washed with 5% NaHCO3, water, and brine
before it was dried and the solvent removed. Purification of the crude
residue by flash chromatography on silica gel (hexane/ethyl acetate,
aromatic), 7.6 (d, 2 H, J ) 7.3 Hz, aromatic), 7.77 (d, 2 H, J ) 7.3
Hz, aromatic), 6.03 (br s, 1 H, NCH2CH), 4.47 (d, 2 H, J ) 6.78 Hz,
FMOC-CH2), 4.27 (t, 1 H, J ) 6.5 Hz, FMOC-CH), 4.13 (br s, 2 H,
NCH2CH), 3.68 (br s, 2 H, NCH2CH2), and 2.49 (br s, 2 H, NCH2CH2)
ppm; 13C NMR (CDCl3) 174.9, 144.4, 141.3, 138.9, 133.09, 128.54,
127.66, 127.01, 126.18, 124.94, 119.96, 67.3, 47.36, 43.72, 42.67, 40.47,
and 27.13 ppm; MS (LSIMS) m/z (%) 416 (MH+, 100), 238 (78);
HRMS for C26H22ClNO2, calcd 415.1339, found 415.1329.
1
6:4) provided 172 mg of product (30% yield): mp 138-140 °C; H
NMR (CDCl3) 7.2-7.5 (m, 8 H, aromatic), 6.6-6.7 (dd, 1 H, J )
10.9 Hz, J ) 17.6 Hz, vinyl H), 5.79 (d, 1 H, J ) 17.6 Hz, vinyl
CHCHcisCHtrans), 5.32 (d, 1 H, J ) 10.9 Hz, vinyl CHCHcisCHtrans),
4.6 (br, 1 H, OH), 3.2-3.7 [br m, 4 H, (CH2)2N)], and 1.8-2.8 [br m,
4 H, (CH2CH2)2N] ppm; 13C NMR (CDCl3) 38.17, 38.95, 43.80, 71.39,
115.30, 125.96, 126.23, 127.23, 128.63, 135.10, 136.04, 138.98, 146.08,
and 170.19 ppm; MS (HREI) m/z (%) 341 (15), 323 (15), 131 (100);
HRMS for C20H20ClNO2, calcd 341.1183, found 341.1179.
N-FMOC-4-(p-chlorophenyl)-3,4-epoxypiperidine (5). To a solu-
tion of N-FMOC-4-(p-chlorophenyl)-1,2,5,6-tetrahydropyridine (200
mg, 0.48 mmol) in CH2Cl2 was added a solution of 80% mCPBA (126
mg, 0.7 mmol) in CH2Cl2, and the reaction mixture was stirred for 18
h. A few drops of methyl ethyl sulfide were added to destroy excess
peroxide. The solution was then washed sequentially with 5% NaHCO3,
water, and brine. Drying over Na2SO4 and evaporation of the solvent
provided a crude solid that was purified by flash chromatography on
silica gel (hexane/ethylacetate, 6:4). The desired product (163.7 mg,
79% yield) was thus obtained: IR (CHCl3) 3024, 1697, 1456, 1431,
4-(p-Chlorophenyl)-4-hydroxy-N-[4-(1,2-epoxyethyl)benzoyl]pi-
peridine (4). 4-(p-Chlorophenyl)-4-hydroxy-N-(4-ethenylbenzoyl)-
piperidine (102 mg, 0.29 mmol) was epoxidized as described above.
Flash chromatography provided 41.5 mg (40%) of the product: mp 90
1339, 1246, 1221, 1123, 1098, 1018 cm-1; H NMR (CDCl3) 7.77-
1
1
°C; H NMR (CDCl3) 7.2-7.45 (m, 8 H, aromatic), 4.64 (br s, 1 H,
7.30 (m, 12 H, aromatic), 4.47 (d, 2 H, J ) 6.5 Hz, OCH2CH), 4.26 (t,
1 H, J ) 6.4 Hz, OCH2CH), 3.8-4.2 (br m, 2 H, CH2NCH2), 3.6-3.8
(br d, 1 H, NCH2CHOC), 3.1-3.3 (br m, 2 H, CH2NCH2), and 2.0-
2.5 (br m, 2 H, NCH2CH2) ppm; 13C NMR (CDCl3) 143.94, 141.33,
133.84, 128.64, 127.67, 127.03, 126.66, 124.83, 119.94, 67.36, 58.52,
and 47.53 ppm; MS (LSIMS) m/z (%) 432 (MH+, 90), 391 (54); HRMS
for C12H11NClO3, calcd 252.0427, found 252.0418. Anal. Calcd for
C26H22ClNO3: C, 73.3; H, 5.13; N, 3.24. Found: C, 73.78; H, 5.37;
N, 3.24.
1-(4′-Fluorophenyl)-4-iodobutane. To a solution of 850 mg (5.18
mmol) of p-fluorophenyl cyclopropyl ketone in CHCl3 were added 4.6
g (31 mmol) of NaI and a catalytic amount of tetrabutylammonium
iodide followed by 1.6 g (15.5 mmol) of (CH3)3SiCl.36 The solution
was stirred for 18 h and was then poured into a solution of NaHCO3.
The excess iodide was removed by the addition of a saturated solution
of sodium thiosulfate. The organic layer was separated and washed
with water and brine. Evaporation of the solvent after drying over
anhydrous Na2SO4 yielded a crude product that was dissolved in hexane/
ethyl acetate (9:1) and filtered through a short column of silica gel.
Evaporation of the solvent yields a dark brown liquid, 1.41 g (94%
yield), that is best stored at 0 °C because it decomposes at room
temperature: 1H NMR (CDCl3) 8.00 (m, 2 H, aromatic), 7.14 (m, 2 H,
aromatic), 3.32 (t, 2 H, J ) 6.7 Hz, CH2I), 3.11 (m, 2 H, CH2CO), and
2.25 (m, 2 H, CH2CH2I) ppm; 13C NMR (CDCl3) 196.56, 130.41,
130.29, 115.56, 115.27, 38.57, 27.25 and 6.58 ppm.
CHHN), 3.88 (br t, 1 H, PhCHOCHcisHtrans), 3.2-3.7 (br m, 3 H,
piperidine H), 3.17 (t, 1 H, J ) 4.6 Hz, PhCHOCHcisHtrans), 2.7 (dd, 1
H, J ) 2.3 Hz, J ) 5.3 Hz, PhCHOHcisHtrans), and 1.6-2.2 (br m, 4
H, piperidine H) ppm; 13C NMR (CDCl3) 169.97, 146.07, 139.36,
135.77, 133.21, 128.58, 127.16, 125.94, 125.6, 71.27, 51.95, 51.26,
43.81, 38.87, 38.29, and 37.72 ppm.
N-FMOC-1,2,5,6-tetrahydropyridine. To a solution of 455 mg
(5.4 mmol) of 1,2,5,6-tetrahydropyridine in CH2Cl2 was added 668 mg
(5.4 mmol) of 4-(dimethylamino)pyridine, and the resulting solution
was cooled to 0 °C. A solution of 1.41 g (5.4 mmol) of 9-fluorenyl-
methyl chloroformate in CH2Cl2 was slowly added, after which the
cooling bath was removed and the mixture was stirred for 2 h. The
solution was washed with 1 N hydrochloric acid, water, and brine.
Evaporation of solvent after drying over anhydrous Na2SO4 leaves a
residue that upon recrystallization from CHCl3/CH3OH gave 1.3 g (77%
yield) of fine yellow needles: mp 94 °C; IR (CHCl3) 3024, 1696, 1455,
1431, 1283, 1239, 1215, 1116, and 1029 cm-1; 1H NMR (CDCl3) 7.31-
7.75 (m, 8 H, aromatic), 5.6-5.8 (br d, 2 H, HCCH), 4.41 (m, 2 H,
OCH2), 4.24 (m, 1 H, OCH2CH), 3.96 (br s, 2 H, NCH2CH), 3.56 (br
s, 2 H, NCH2CH2), and 2.15 (br s, 2 H, NCH2CH2) ppm; 13C NMR
(CDCl3) 174.69, 144.10, 141.31, 127.61, 126.99, 124.98, 119.93, 67.29,
47.37, 43.44, 40.57, 33.68, and 24.94 ppm; MS (LSIMS) m/z 306 (MH+,
100). Anal. Calcd for C20H19NO2: C, 78.66; H, 6.27; N, 4.59.
Found: C, 78.48; H, 6.35; N, 4.51.
N-FMOC-3,4-epoxypiperidine (6). To a solution of 300 mg (0.98
mmol) of N-FMOC-1,2,5,6-tetrahydropyridine in CH2Cl2 was added a
solution of 700 mg (3.2 mmol) of mCPBA in CH2Cl2 and 10 mL of
0.5 M NaHCO3. The reaction mixture was stirred for 18 h at ∼25 °C.
Excess peroxide was destroyed by the addition of a few drops of methyl
ethyl sulfide (starch-iodide paper test). The organic phase was washed
sequentially with aqueous NaHCO3, water, and brine and was then dried
over anhydrous Na2SO4. Solvent removal and purification by flash
chromatography on silica gel with hexane/ethyl acetate (6:4) gave 280
mg (89% yield) of crystalline material: mp 110 °C; IR (CHCl3) 3018,
1703 1481, 1450, 1431, 1339, 1289, 1252, 1215, 1110, 1055, 962, and
N-(4-p-Fluorophenyl-4-oxobutyl)-4-(p-chlorophenyl)-3,4-epoxypi-
peridine (3). To a solution of 138 mg (0.32 mmol) of N-FMOC-4-
(p-chlorophenyl)-3,4-epoxypiperidine in 1.5 mL of DMF was added
0.1 mL of piperidine. The solution was stirred for 0.5 h at ∼25 °C
before the solvent was removed on a rotary evaporator. The solid
residue was redissolved in 10 mL of DMF before 44 mg (0.32 mmol)
of K2CO3 and 93.4 mg (0.32 mmol) of 1-(4′-fluorophenyl)-4-iodobutane
were sequentially added. The reaction mixture was gently refluxed
for 0.5 h. The solvent was then evaporated, and the residue was
redissolved in CHCl3 and washed with 5% NaHCO3, water, and brine.
The solution was dried over anhydrous Na2SO4. Filtration and
evaporation of solvent gave a dark brown residue that, after flash
chromatography on silica gel with ethyl acetate, provided 76.8 mg (64%
yield) of the desired product: IR (neat) 3018, 2950, 1684, 1604, 1221,
1
752 cm-1; H NMR (CDCl3) 7.2-7.8 (m, 8 H, aromatic), 4.41 (m, 2
H, OCH2), 4.24 (m, 1 H, OCH2CH), 3.70-4.00 (m, 2 H, NCHHCHO),
3.18-33.6 (m, 4 H, CH2NCHHCHOCH), and 1.8-2.2 (m, NCH2CH2)
ppm; 13C NMR (CDCl3) 155.28, 143.92, 141.27, 127.61, 126.98,
124.87, 119.90, 67.31, 50.44, 47.28, 42.39, 37.30, and 24.10 ppm; MS
(LSIMS) m/z (%) 322 (MH+, 35), 179 (99), 178 (100). Anal. Calcd
for C20H19NO3: C, 74.10; H, 5.96; N, 4.29. Found: C, 74.10; H, 5.98;
N, 4.29.
1
1160, and 1098 cm-1; H NMR (CDCl3) 7.0-8.1 (m, 8 H, aromatic),
3.18 (m, 1 H, epoxide), 3.12 (m, 1 H, CHCHaxHeqN), 2.98 (t, 2 H,
NCH2CH2CH2CO), 2.55 (d, 1 H, J ) 11.2 Hz, CHCHaxHeqN), 2.3-
2.55 (m, 5 H, CHHCHHNCH2CH2CH2CO), 2.08 (m, 1 H, CHHCH2N),
and 1.95 (m, 2 H, NCH2CH2CH2CO) ppm; 13C NMR (CDCl3) 133.42,
130.57, 130.68, 128.46, 126.85, 115.76, 115.46, 60.54, 58.56, 56.98,
52.15, 46.52, 36.06, 29.07, and 21.56 ppm; MS (HREI) m/z (%) 373
(5), 356 (12), 235 (100), and 222 (52); HRMS for C21H21ClFNO2, calcd
373.1235, found 373.1245.
FMOC-4,5-epoxy-Leu Methyl Ester (7). A methanol solution of
FMOC-4,5-dehydro-Leu-OH (100 mg, 0.28 mmol) was treated with
an excess of diazomethane in ether. Evaporation of the solvent gave
a quantitative yield of the ester: 1H NMR (CDCl3) 7.25-7.7 (m, 8 H,
aromatic), 5.27 (d, 1 H, J ) 7.8 Hz, NH), 4.7-4.8 (d, 2 H, J ) 30 Hz,
vinyl H), 4.52 (m, 1 H, NHCHCOOH), 4.38 (d, 2 H, J ) 7.2 Hz,
CHCH2O), 4.22 (m, 1 H, J ) 7.2 Hz, CHCH2O), 3.74 (S, 3 H, OCH3),
2.55 (dd, 1 H, J ) 5.2 Hz, J ) 13.2 Hz, CHCHHC), 2.40 (dd, 1 H, J
) 8.3 Hz, J ) 13.2 Hz CHCHHC), and 1.75 (s, 3 H, -CH3) ppm; MS
(HREI) m/z (%) 365.2 (M+, 12), 178.1 (100). The crude FMOC-4,5-
4-(p-Chlorophenyl)-4-hydroxy-N-(4-ethenylbenzoyl)piperidine. To
a CH2Cl2 solution of 4-ethenylbenzoic acid (340 mg, 2.3 mmol) was
added 2.3 mmol of oxalyl chloride at 0 °C followed by a few drops of
(36) Olah, G. A.; Narang, S. C.; Field, L. D., Salem, G. F. J. Org. Chem.
1980, 45, 4792-4793.

5856 J. Am. Chem. Soc., Vol. 118, No. 25, 1996
Yu et al.
dehydro-Leu methyl ester (100 mg, 0.27 mmol) was subjected to
epoxidation as described above. Purification by flash chromatography
yielded 85.4 mg (83%) of the epoxide: 1H NMR (CDCl3) 7.32-7.8
(m, 8H, aromatic), 5.5 (dd, 1 H, NH), 4.42 (m, 1 H, NHCHCOOCH3),
4.3 (m, 2 H, CHCH2O), 4.15 (m, 1 H, CHCH2O), 3.68 (d, 3 H, OCH3),
22.52 (S, 2 H, epoxide HH), 2.05-2.2 (2 dd, 1 H, J ) 5 Hz, J ) 14
Hz, CHCHHC), 1.72-1.85 (m, 1 H, CHCHHC), and 1.3 (s, 3 H,
-CH3) ppm; MS (HREI) m/z (%) 381 (M+, 8), 178 (100); HRMS for
C22H23NO5, calcd 381.1576, found 381.1568.
HIV Protease Expression and Purification. Recombinant HIV-1
PR was expressed in Escherichia coli strain X90 using the pT2Cb3
PR vector.10 The enzyme was purified to homogeneity as previously
described.20 Concentrations of active HIV-1 PR were determined by
active site titration using the peptidomimetic inhibitor U-85548 (a gift
of Dr. A. Tomasselli, Upjohn), Val-Ser-Gln-Asn-Leu-ψ-[CH(OH)CH2]-
Val-Ile-Val.5
a sweep rate of 7 s/scan. Approximately 10 scans were accumulated
into one spectrum for a better signal to noise ratio. The average
molecular weight39 of each species was calculated on the basis of the
multiple charge nature of electrospray ionization.
Enzymatic Digestion. The HPLC fractions containing the modified
protease were subjected to enzymatic digestion to determine the site
of modification of the inhibitor. For tryptic digestion, the samples were
dissolved in 20 µL of 8 M urea and then diluted by adding 80 µL of
100 mM ammonium bicarbonate, pH 8.3, to a final concentration of
80 mM NH4HCO3 and 1.6 M urea. Enzymes were then added to a
concentration of ∼4% of the sample. The digestion was carried out at
37 °C for 4 h but in some cases was continued overnight. For digestion
by pepsin, the samples were first dissolved in 70% formic acid, and
then 20 volumes of 1 mM HCl containing pepsin was added to the
sample. The ratio of protein to enzyme was about 50 to 1. The
digestion was carried out at room temperature for 0.5-2 h and stopped
by lyophilization or by injection onto a C-18 reversed phase HPLC
column.
Peptide Molecular Weight Measurement. Molecular weights of
the enzymatic peptides were determined by LC/ESIMS. These
experiments were performed on the VG Platform mass spectrometer
equipped with an ABI Model 140B HPLC system, using a microbore
C-18 reversed phase HPLC column (1 × 100 mm, ABI). A 50-min
linear gradient from 2% to 52% solvent B was run at a flow rate of 50
µL/min. Solvent A was 0.1% formic acid in H2O, and solvent B was
0.05% formic acid in 10% 2-propanol/90% ethanol. The mass
spectrometer was set to a m/z range of 350-2100 with a sweep speed
of 5 s/scan. Subsequent isolation of the modified peptides was carried
out using an analytical C-18 HPLC column (Vydac 4.6 × 250 mm)
employing conditions as described above.
Inactivation of HIV Proteases. HIV-1 PR (concentration ∼50 µg/
mL) was inhibited by EPNP by preincubation at ∼25 °C in a buffer
containing 250 mM sodium acetate buffer (pH 5.5), 0.5 M NaCl, 1
mM EDTA, 0.5 mM DTT, and 10% DMSO in the presence of 5 mM
EPNP. For quantitation of the irreversible inactivation of HIV proteases
by haloperidol and FMOC derivatives, HIV-1 Q7K PR (15 µg/mL final
concentration) was preincubated at 25 °C in 50 mM HEPES, pH 8.0,
containing 1 M NaCl, 1 mM EDTA, 0.5 mM DTT, and 5% DMSO in
the presence of 10 µM to 1 mM of each inhibitor. Baseline
measurements were carried out using 5% DMSO in the absence of
inhibitors. At various times, aliquots were removed and assayed for
activity. The HIV protease was assayed against the fluorescent substrate
ABZ-TInLF(NO2)QR-NH2. The enzyme was assayed at pH 5.5 in 50
mM acetate buffer containing 1 M NaCl, 1 mM EDTA, and 1 mM
DTT as previously described.37 For calculation of the rates of
inactivation, kinetic data were fit to a pseudo-first-order equation
according to the formula ln(Vi/Vo) ) -kobst. From a plot of inactivation
rates (kobs) Versus the square of the inhibitor concentrations, kinact, the
Tandem Mass Spectrometry. Selected peptides were subjected
to high-energy CID analysis on a Kratos Concept IIHH four-sector
tandem mass spectrometer. Samples were introduced into the instru-
ment through a flow probe with a flow rate of 3 µL/min. The delivery
solvent contained 5% acetonitrile, 2% thioglycerol, and 0.1% trifluo-
roacetic acid in H2O. The precursor ion was subjected to CID process
in the collision cell at a helium gas pressure which resulted in an
attenuation of its abundance to 30% of the original value. The resulting
CID spectra were recorded on a rapid scanning multichannel array
detector.40,41
inhibitor concentration resulting in half-maximal inactivation, and Vinact
,
the maximum inactivation rate, were calculated. In preparing the
proteins for digestion and mass spectrometric work, the loss of activity
was monitored by a UV assay using the peptide Arg-Val-nLeu-Phe-
(p-NO2)-Glu-Ala-nLeu-Ser-NH2 (Bioserv, San Diego, CA). Aliquots
of the protease/inhibitor complex were removed, and the protease
activity was assessed by measuring the change in UV absorbance at
300 nm, the characteristic absorbance maximum of the substrate.
HPLC Separation of the Inhibited Proteases. The inhibition
products were isolated by reversed phase HPLC on a C-4 column (4.6
× 250 mm, Vydac) using a 75-min gradient from 30% to 55% solvent
B (acetonitrile containing 0.08% trifluoroacetic acid) against solvent
A (water containing 0.1% trifluoroacetic acid). The flow rate was 1
mL/min, and the products were monitored by a UV/vis detector.
Fractions were manually collected and lyophilized for further charac-
terization.
Acknowledgment. Financial support was provided by
National Institutes of Health Grant GM39522, NIH NCRR
Biomedical Research Technology Program Grant RR01614,
National Institute of Environmental Health Sciences Grant
ES04705, and National Science Foundation Grant DIR 8700766.
F.M. was a recipient of an AmFAR Fellowship, award number
70361-14-RF. We thank Fisons/VG for the loan of the platform
electrospray mass spectrometer.
Protein Molecular Weight Determination. The HPLC-purified
fractions were analyzed by electrospray ionization mass spectrometry38
on a VG Platform single quadrupole mass spectrometer equipped with
an electrospray ion source. The samples were dissolved in an aqueous
solvent (40% acetonitrile and 1% acetic acid), and 5 µL aliquots
(approximately 10 pmol/µL) were injected into the mass spectrometer.
The solvent system for batch mode electrospray contained 10%
2-propanol/40% acetonitrile in H2O and 50 pmol/µL gramicidin. The
flow rate was 5 µL/min, and the mass range was set at 400-2000 with
JA954069W
(39) Carr, S. A.; Burlingame, A. L. In Mass Spectrometry in the
Biological Sciences; Burlingame, A. L., Carr, S. A., Eds.; Humana Press:
Totowa, NJ, 1996; pp 546-553.
(40) Walls, F. C.; Hall, S. C.; Medzihradsky, K. F.; Yu, Z.; Burlingame,
A. L.; Evans, S.; Hoffman, A. D.; Buchanan, R.; Glover, S. Proceedings
of the 41st ASMS Conference on Mass Spectrometry and Allied Topics,
San Francisco, 1994; p 937.
(41) Burlingame, A. L. In Biological Mass Spectrometry: Present and
Future; Matsuo, T., Caprioli, R. M., Gross, M. L., Seyama, Y., Eds.; John
Wiley & Sons: New York, 1994; pp 147-164.
(42) Roestorff, P.; Fohlman, J. J. Biomed. Mass Spectrom. 1984, 11,
601.
(37) Toth, M. V.; Marshall, G. R. Int. J. Peptide Protein Res. 1990, 36,
544-550.
(38) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C.
M. Science 1989, 246, 64-71.
(43) Biemann, K. Methods Enzymol. 1990, 193, 886-887.