Variations of the P2 group in HIV-1 protease inhibitors containing a tertiary alcohol in the transition-state mimicking scaffold

Article abstract of DOI:10.1039/b606859f

The development of synthetic protocol leading to HIV-1 protease inhibitors with a tertiary alcohol based transition-state mimicking unit and different P2 side chains was investigated. (2S)-2-benztloxirane-2-carboxylic acid ((S)-5) was used as a key interm

Full text of DOI:10.1039/b606859f

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Variations of the P2 group in HIV-1 protease inhibitors containing a tertiary
alcohol in the transition-state mimicking scaffold†
Jenny K. Ekegren,a Johan Gising,a Hans Wallberg,b Mats Larhed,a Bertil Samuelssonb and Anders Hallberg*a
Received 15th May 2006, Accepted 23rd June 2006
First published as an Advance Article on the web 6th July 2006
DOI: 10.1039/b606859f
A short synthetic protocol leading to HIV-1 protease in-
hibitors with a tertiary alcohol based transition-state mim-
icking unit and different P2 side chains has been developed.
HIV-1 protease inhibitors (PIs) are important in the most
frequently used regimen for the treatment of HIV/AIDS, the
highly active antiretroviral therapy (HAART), and have improved
the quality of life and prolonged the lifetimes of infected
patients.1 Seven PIs have reached the market so far,2,3 and in
addition, tipranavir has been granted accelerated approval by
the FDA.4 All PIs currently available for therapy are peptide-like
and suffer from shortcomings related to their pharmacokinetic
properties. Low membrane permeability, high protein binding and
rapid metabolism have been reported for almost all approved
inhibitors.3,5 In addition, considering the increasing numbers of
HIV-1 strains resistant to the present PI generation,6,7 the need
for new structural themes and unique chemical entities to combat
HIV/AIDS is obvious.8
Fig. 1 Structures of HIV-1 protease inhibitor 1, containing a tertiary
We recently described a new class of HIV-1 PIs containing
a tertiary alcohol, a component which has rarely been used
in transition-state analogs.9,10 These inhibitors, exemplified by
1 in Fig. 1, exhibited low Ki values and, notably, excellent
membrane permeation properties in a Caco-2 assay. The latter
observation was accounted for by anticipating that participation
in intramolecular hydrogen bonding could mask the tertiary
hydroxyl group. The stereochemistry of the tertiary alcohol was
critical for inhibition, with the (S)-isomer being considerably
more potent than the corresponding (R)-isomer. The transition-
state mimicking unit in these inhibitors was synthesized by ring
opening of a 2,2-disubstituted epoxide comprising an amido-
indanol moiety, present in the approved PI indinavir as shown
in Fig. 1.11,12 Unfortunately, results from the first two series
of inhibitors indicated high intrinsic clearance (Clint) in liver
microsomes. This could be attributed to degradation of the
P2 indanol group, known to be metabolized by CYP enzymes
via benzylic oxidation in related systems.13,14 We were therefore
encouraged to develop an alternative synthetic route that would
allow the incorporation of other P2 groups in a straightforward
manner. Herein we present a new protocol and enzyme inhibition
data from a series of P2-varied compounds.
alcohol-derived transition-state mimic, and indinavir.
We were interested in using (2S)-2-benzyloxirane-2-carboxylic
acid ((S)-5) as a key intermediate in the synthesis of the new
HIV-1 protease inhibitors (Scheme 1). Amide coupling of (S)-5 to
different amines followed by epoxide ring opening would afford
inhibitors encompassing various P2 groups. Thus, a synthetic
protocol starting from 2-benzylacrylic acid (2), preparedaccording
to a literature procedure,15 was developed (Scheme 1). Compound
2 was activated by SOCl2 and then coupled with ethyl-(S)-
lactate producing ester 3 in 75% isolated yield. Epoxidation by
3-chloroperoxybenzoic acid (mCPBA) followed by flash chro-
matography gave the pure diastereomers (S)-4 and (R)-4, in
aDepartment of Medicinal Chemistry, Organic Pharmaceutical Chemistry,
BMC, Uppsala University, Box 574, SE-751 23, Uppsala, Sweden. E-mail:
Anders.Hallberg@orgfarm.uu.se; Fax: +46-18-4714474; Tel: +46-18-
4714284
bMedivir AB, Lunastigen 7, SE-141 44, Huddinge, Sweden
† Electronic supplementary information (ESI) available: Experimental
details and spectroscopic data for all new compounds, elemental analysis
data, procedures for enzyme assays and liver microsome stability evalua-
tion. See DOI: 10.1039/b606859f
Scheme 1 Reagents and conditions: (i) SOCl2, room temp and then
ethyl-(S)-lactate, DMAP, CH2Cl2, room temp, 75%; (ii) mCPBA, CH2Cl2,
reflux, 77%; (iii) NaOH, THF, room temp, quantitative yield.
3040 | Org. Biomol. Chem., 2006, 4, 3040–3043
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
addition to a mixed fraction (in a 4 : 5 : 1 ratio) in a total yield
of 77%. The lactate unit was cleaved from the epoxide part using
NaOH in water–THF, providing the free acid in quantitative yield.
The absolute configuration of (S)-5 was confirmed by coupling
with (1S,2R)-2-aminoindanol followed by comparison of NMR
data and optical rotation of the isolated compound with the
reported values.9
Initially, (S)-5 was coupled with different amines using
EDC (1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydro-
chloride), NMM (N-methylmorpholine) and HOBT (1-hydroxy-
benzotriazole), resulting in the corresponding amides at low to
moderate yields (6a, b, f–h, Scheme 2). For the more sluggish
aromatic amines this system did not promote any coupling and in-
stead PyBOP ((benzotriazol-1-yloxy)tripyrrolidinophosphonium
hexafluorophosphate) and (i-Pr)2NH were used, which resulted
in compounds 6c and 6d in 46% and 39% isolated yields, re-
spectively (Scheme 2). With L-valine-methylamide, HATU (O-(7-
azabenzotriazole-1-yl)-N,N,Nꢀ,Nꢀ-tetramethyluronium hexafluo-
rophosphate) together with (i-Pr)2EtN provided the most efficient
coupling reaction in this series and amide 6e was generated in
60% yield (Scheme 2). The amino acid a-carbon partly racemized
during the amide couplings and mixtures of amides 6e–h with
minor amounts of the corresponding (R)-isomers (as determined
by NMR) were isolated. These impurities were removed by reverse-
phase LC-MS of the final products. Coupling between 2 and
L-tert-leucine-methylamide, followed by epoxidation yielded a 50 :
50 mixture of (S)-8 and (R)-8, which was separated by flash
chromatography (Scheme 2). The absolute configuration at the
quaternary carbon atom in (S)- and (R)-8 was determined using
X-ray crystallographic data from the corresponding (S)-inhibitor
18, obtained by ring opening of (S)-8 (Fig. 2, Table 1).16
Interestingly, in this X-ray structure, a hydrogen bond was revealed
between the tertiary alcohol and the prime side hydrazide carbonyl
Fig. 2 X-Ray crystal structure of compound 18.
alcohol in the transition-state mimic is present in these molecules,
which, in turn, could contribute to efficient membrane permeation.
Hydrazide 9, prepared as previously reported9 was used to open
disubstituted epoxides 6a–h and 8 (Table 1). Surprisingly, attempts
to use Ti(Oi-Pr)4 as a catalyst, successfully applied in the earlier
studies using indanol-amide epoxides,9,10 rendered the extensive
formation of by-products. An uncatalyzed protocol provided clean
product patterns, but required longer reaction times (4–8 days).
Purification by reverse-phase LC-MS resulted in moderate to
good yields of most target compounds. However, the products
derived from epoxides 6c and d comprising aromatic amides were
recovered in very poor yields (Table 1).
Enzyme inhibition data and anti-HIV activity in cell culture
for structures 10–19 are summarized as Ki and EC50 values
in Table 1. The amido-indanol derivative, 1, is included as a
reference compound. Evaluation of inhibitors 10–13, with cyclic
P2 groups, resulted in poor to moderate inhibitory potencies,
˚
group (distance 1.8 A). This observation further supports our
hypothesis that intramolecular hydrogen bonding to the tertiary
Scheme 2 Reagents and conditions: (i) 6a, b, f–h: P2–NH2, EDC, HOBT, NMM, EtOAc, room temp, 17–45%; 6c, d: P2–NH2, PyBOP, (i-Pr)2NH,
CH2Cl2, room temp, 46%, 39%; 6e: P2–NH2, HATU, (i-Pr)2EtN, CH2Cl2, room temp, 60%; (ii) L-tert-leucine-methylamide, EDC, HOBT, NMM, EtOAc,
room temp, 54%; (iii) mCPBA, AIBN (2,2’-azobis(2-methylpropionitrile), ClCH2CH2Cl, reflux, 76%.
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3040–3043 | 3041
©

View Article Online
Table 1 Synthesis and enzyme inhibition data for compounds 10–19a
Compound
P2 group
From epoxide
—
Yield (%)
—
Ki (nM)
EC50 (lM)
1
2.4
1.1
10
11
6a
6b
54
52
88
>10
>10
110
12
6c
7
480
>10
13
6d
6e
11
46
1140
7.4
>10
7.3
14b
15c
16
6f
73d
35
5.1
81
3.1
6g
>10
17
6h
39
860
>10
18
19
(S)-8
(R)-8
49
69
180
>10
>10
>5000
a Conditions: i-PrOH, 80 ◦C, 4–8 days. b Clint = 160 lL min−1 mg−1
3042 | Org. Biomol. Chem., 2006, 4, 3040–3043
.
c Clint = 230 lL min−1 mg−1
.
d Based on the amount of converted epoxide.
This journal is
The Royal Society of Chemistry 2006
©

View Article Online
Ki = 88–1140 nM (Table 1). Compounds 14–18, comprising amino
acid-derived P2 units, displayed large variability of the Ki values
depending on the size of the amino acid side chain (Table 1).
The iso-propyl group in compounds 14 and 15 seemed to be well
tolerated in the enzyme S2 pocket affording highly potent HIV-1
protease inhibitors with Ki = 7.4 and 5.1 nM, respectively. On
the other hand, the iso-butyl, tert-butyl and benzyl side chains
afforded several times less potent inhibitors as deduced from the
corresponding Ki values (16–18, Table 1). As expected, the (R)-
analogue 19 was not active in the enzyme assay at concentrations
below 5000 nM. Surprisingly, most of the inhibitors did not exhibit
any cellular antiviral activity (Table 1). Only the best enzyme
inhibitors, 14 and 15, were active in this assay with EC50 = 7.3 and
3.1 lM, respectively. Compound 14 and 15 were further evaluated
for stability in the presence of liver microsomes (Table 1). Slightly
lower intrinsic clearance was observed for these two compounds
(Clint = 160 and 230 lL min−1 mg−1, respectively), lacking the
metabolically unstable indanol-amide P2 group compared to
inhibitor 1 (Clint = 266 lL min−1 mg−1).
alcohol in the transition-state mimicking scaffold and comprising
various P2 groups. The inhibitors were prepared applying four
or five synthetic steps and no protecting groups were required.
Compound 15 exhibited the lowest Ki value (5.1 nM) in the series
and also demonstrated the highest activity in cell culture (EC50
3.1 lM).
=
We thank the Swedish Research Council (VR), the Swedish
Foundation for Strategic Research (SSF), Dr Gunnar Lindeberg
and Mr Christian Sko¨ld.
Notes and references
1 R. J. Pomerantz and D. L. Horn, Nat. Med., 2003, 9, 867–873.
2 A. Brik and C.-H Wong, Org. Biomol. Chem., 2003, 1, 5–14.
3 J. T. Randolph and D. A. DeGoey, Curr. Top. Med. Chem., 2004, 4,
1079–1095.
4 S. R. Turner, J. W. Strohbach, R. A. Tommasi, P. A. Aristoff, P. D.
Johnson, H. I. Skulnick, L. A. Dolak, E. P. Seest, P. K. Tomich, M. J.
Bohanon, M.-M. Horng, J. C. Lynn, K. Chong, R. R. Hinshaw, K. D.
Watenpaugh, M. N. Janakiraman and S. Thaisrivongs, J. Med. Chem.,
1998, 41, 3467–3476.
Several of the approved HIV-1 protease inhibitors comprise
relatively small, cyclic P2 structural elements. This is the case
for example with amprenavir, containing tetrahydrofuran as the
P2 group and for nelfinavir with a phenol-related structure in
this position.3 On the other hand, the most recently launched
inhibitor atazanavir, carries N-derivatized amino acid residues in
the P2/P3 and P2ꢀ/P3ꢀ positions.17 We were encouraged to evaluate
structural units representing both these types of substituents as
potential P2 groups in our new inhibitors. The HIV-1 protease
inhibition data summarized in Table 1 suggest that the size and
polarity of the P2 substituent are crucial to allow proper accom-
modation in the S2 sub-site. Small P2 groups, unable to reach the
enzyme S3 pocket, furnish poor to moderate inhibitory potencies
(10–13). Furthermore, the distance between the transition-state
mimicking tertiary hydroxyl group and P2 aromatic ring structures
in 10, 12 and 13 proved to be of importance. A methylene spacer
between the amide bond and the P2 aryl group as in compound
10 afforded a 5 to 13 times more potent inhibitor than 12 and 13
(Table 1). The amino acid-derived P2 substituents in compounds
14–18 have the potential of reaching both the enzyme S2 and S3
pockets, which could be beneficial for efficient binding. However,
the bulkiness of the P2 side chain strongly affected inhibition
and only the iso-propyl group present in compounds 14 and
15 provided highly potent inhibitors (Table 1). The fact that
compound 18 was devoid of activity in the cellular assay was
somewhat surprising and is difficult to rationalize since a similar
tert-leucine-derived P2/P3 group present in the approved inhibitor
atazanavir has been reported to afford both excellent potency in
cell culture and high oral bioavailability.17
5 F. Rodr´ıguez-Barrios and F. Gago, Curr. Top. Med. Chem., 2004, 4,
991–1007.
6 F. Clavel and A. J. Hance, New Engl. J. Med., 2004, 350, 1023–1035.
7 C. de Mendoza and V. Soriano, Curr. Drug Metab., 2004, 5, 321–
328.
8 E. De Clercq, J. Med. Chem., 2005, 48, 1297–1313.
9 J. K. Ekegren, T. Unge, M. Z. Safa, H. Wallberg, B. Samuelsson and
A. Hallberg, J. Med. Chem., 2005, 48, 8098–8102.
˚
10 J. K. Ekegren, N. Ginman, A. Johansson, H. Wallberg, M. Larhed,
B. Samuelsson, T. Unge and A. Hallberg, J. Med. Chem., 2006, 49,
1828–1832.
11 J. P. Vacca, B. D. Dorsey, W. A. Schleif, R. B. Levin, S. L. McDaniel,
P. L. Darke, J. Zugay, J. C. Quintero, O. M. Blahy, E. Roth, V. V.
Sardana, A. J. Schlabach, P. I. Graham, J. H. Condra, L. Gotlib, M. K.
Holloway, J. Lin, I.-W. Chen, K. Vastag, D. Ostovic, P. S. Anderson,
E. A. Emini and J. R. Huff, Proc. Natl. Acad. Sci. U. S. A., 1994, 91,
4096–4100.
12 B. D. Dorsey, R. B. Levin, S. L. McDaniel, J. P. Vacca, J. P. Guare, P. L.
Darke, J. A. Zugay, E. A. Emini, W. A. Schleif, J. C. Quintero, J. H.
Lin, I.-W. Chen, M. K. Holloway, P. M. D. Fitzgerald, M. G. Axel,
D. Ostovic, P. S. Anderson and J. R. Huff, J. Med. Chem., 1994, 37,
3443–3451.
13 S. K. Balani, B. H. Arison, L. Mathai, L. R. Kauffman, R. R. Miller,
R. A. Stearns, I. W. Chen and J. H. Lin, Drug Metab. Dispos., 1995, 23,
266–270.
14 J. H. Lin, Adv. Drug Delivery Rev., 1999, 39, 33–49.
15 X. Liu, E. Hu, X. Tian, A. Mazur and F. H. Ebetino, J. Organomet.
Chem., 2002, 646, 212–222.
16 Crystal data. C32H46BrN5O6xC2H5OH, M = 722.72, monoclini◦c, a =
˚
˚
˚
10.6340(3) A, b = 12.1440(4) A, c = 15.9560(8) A, b = 108.102(1) , U =
3
−1
˚
1958.6(1) A , T = 293(2) K, space group P21, Z = 2, l = 1.098 mm
,
8473 reflections measured, 8463 unique (Rint = 0.0252). Final R1
=
0.0830 (for 4316 reflections with I > r(I)), wR2 (on F∧2) = 0.2307 (for
all data). CCDC reference number 607406. For crystallographic data
in CIF format see DOI: 10.1039/b606859f.
17 G. Bold, A. Fa¨ssler, H.-G. Capraro, R. Cozens, T. Klimkait, J. Lazdins,
J. Mestan, B. Poncioni, J. Ro¨sel, D. Stover, M. Tintelnot-Blomley, F.
Acemoglu, W. Beck, E. Boss, M. Eschbach, T. Hurlimann, E. Masso,
S. Roussel, K. Ucci-Stoll, D. Wyss and M. Lang, J. Med. Chem., 1998,
41, 3387–3401.
In summary, an enantiomerically pure epoxy carboxylic acid
was identified as the key building block in a novel synthetic
strategy delivering HIV-1 protease inhibitors with a tertiary
This journal is
The Royal Society of Chemistry 2006
Org. Biomol. Chem., 2006, 4, 3040–3043 | 3043
©