Design, synthesis, evaluation, and crystallographic-based structural studies of HIV-1 protease inhibitors with reduced response to the V82A mutation

Article abstract of DOI:10.1021/jm701170f

In our quest for HIV-1 protease inhibitors that are not affected by the V82A resistance mutation, we have synthesized and tested a second generation set of C₂-symmetric HIV-1 protease inhibitors that contain a cyclohexane group at P1 and/or P1′

Full text of DOI:10.1021/jm701170f

852
J. Med. Chem. 2008, 51, 852–860
Design, Synthesis, Evaluation, and Crystallographic-Based Structural Studies of HIV-1 Protease
Inhibitors with Reduced Response to the V82A Mutation§
José C. Clemente,† Arthur Robbins,† Paula Graña,‡ M. Rita Paleo,‡ Juan F. Correa,‡ M. Carmen Villaverde,‡ F. Javier Sardina,‡
Lakshmanan Govindasamy,† Mavis Agbandje-McKenna,† Robert McKenna,† Ben M. Dunn,† and Fredy Sussman*,‡
Departamento de Química Orgánica, Facultad de Química, UniVersidad de Santiago de Compostela, 15782-Santiago de Compostela, Spain,
and Department of Biochemistry and Molecular Biology, UniVersity of Florida College of Medicine, GainesVille, Florida 32610
ReceiVed September 18, 2007
In our quest for HIV-1 protease inhibitors that are not affected by the V82A resistance mutation, we have
synthesized and tested a second generation set of C2-symmetric HIV-1 protease inhibitors that contain a
cyclohexane group at P1 and/or P1′. The binding affinity results indicate that these compounds have an
improved response to the appearance of the V82A mutation than the parent compound. The X-ray structure
of one of these compounds with the V82A HIV-1 PR variant provides the structural rationale for the better
resistance profile of these compounds. Moreover, scrutiny of the X-ray structure suggests that the ring of
the Cha side chain might be in a boat rather than in the chair conformation, a result supported by molecular
dynamics simulations.
Table 1. Inhibitors Studied in this Work
Introduction
Many research laboratories have developed HIV-1 protease
(HIV-1 PR) inhibitors that have become lead compounds for
antiviral drugs, and some of these compounds have been
approved for therapeutical use.1–4 It has been found that the
most effective drug treatment, referred to as highly active
antiretroviral therapy or HAARTa, is based on drug combina-
tions, such as HIV-1 PR inhibitors and HIV-1 reverse tran-
scriptase inhibitors.5 Still, one of the pervading problems found
in these therapies is the development of resistance mutations in
the protease. During therapy with antiretroviral drugs, the
pretherapy forms of the virus are inhibited from replication,
while variant forms, created by the activity of the error-prone
reverse transcriptase, are allowed to propagate. The primary
sequence modifications that are encoded in this way are called
“resistance” mutations and have been observed both in vitro
and in vivo.3–6 These sequence changes occur in a variety of
residues located in the binding site as well as on the surface of
the enzyme. The resistance mutations keep or enhance the
catalytic activity of the enzyme, allowing the virus to replicate,
but they lower the affinity of the HIV-1 PR for the antiviral
drug, thus reducing its potential therapeutic power. One of the
residues that is consistently mutated as a consequence of the
HAART treatment, especially with Indinavir and Ritonavir,7 is
V82, which can be replaced by a variety of amino acids ranging
from Ala to Phe. The V82A variant has been identified as a
resistant mutant in vitro5 as well as in clinical isolates.4,6 X-ray
crystallographic studies of a C2 symmetric inhibitor bound to
the V82A HIV-1 PR variant indicate that the V82A mutant
breaks down the almost symmetric binding modes to the S1
and S1′ sites by reducing the interaction of the enzyme only
with the P1 fragment.8 This can be considered as a “hydrophobic
a The names PP, PC, and CC came from the P1/P1′ fragments, where
P stands for Phe, the phenylalanine side chain, and C for Cha, the
cyclohexylalanine side chain. b See ref 25.
defect” in the packing of the inhibitor-enzyme complex, similar
to that produced by mutations in the core of the protein.9 The
design of a second generation C2 pseudosymmetric inhibitors
that are less affected by the V82A mutation could include more
bulky groups in P1 or P1′ that could “repair” the above-
mentioned “hydrophobic hole”. To probe this hypothesis, we
have synthesized and tested the known inhibitor PP (A-76928),
a C2-pseudosymmetric inhibitor with a diol motif based isos-
tere,10 as well as the related inhibitors PC and CC that replace
one or two phenyl groups, respectively, by the larger cyclo-
hexane fragment in this inhibitor (see Table 1). Our second
generation inhibitors have a better resistance profile for the
V82A mutations, demonstrating the validity of our design
hypothesis.
§
Coordinates for HIV-1 PRV82A complexed with inhibitor PC (9) have
been deposited in the Protein Data Bank, with access code 2FLE.
* To whom correspondence should be addressed. Phone: 0034981563100,
ext 14402. Fax: 0034981591014. E-mail: fsussman@usc.es.
†
University of Florida.
Universidad de Santiago de Compostela.
X-ray-based structural studies (combined with a variety of
other techniques) have afforded many insights into the mech-
anisms of resistance mutations, which have been used to produce
and test second generation inhibitors or to design totally new
inhibitors with improved resistance profile against a variety of
‡
a Abbreviations: HAART, highly active antiretroviral therapy; Cha,
cyclohexylalanine side chain; Cy, cyclohexyl; Cby, 2,2,4,4-tetramethyl-
1,3-oxazolidine-3-carbonyl; IPTG, isopropylthio-D-galactopyranoside; dTT,
dithiothreitol.
10.1021/jm701170f CCC: $40.75
2008 American Chemical Society
Published on Web 01/24/2008

Studies of HIV-1 Protease Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4 853
Scheme 1. Total synthesis of HIV-1 protease inhibitors CC (8), PC (9), and PP (10)a
a Reagents and conditions: (a) 1. s-BuLi/TMEDA/Et2O, -78 °C; 2. 2/Et2O; (b) 1. MeSO3H, MeOH, reflux; 2. LiOH, dioxane/H2O, reflux; (c) H2, Pd(OH)2/
C, (Boc)2O; (d) TFA; (e) K2CO3, THF, rt, 20 h.
Table 2. Binding Constants for the Inhibitors Bound to the Native and
HIV-1 PR mutations.11–14 To get some insight into the binding
V82A Mutant HIV-1 PR Strains
mode of our new inhibitors to the resistant mutant V82A HIV-1
inhibitor
Ki (nM) native
Ki (nM) mutant
Ki mutant/Ki native
PR, we obtained the structure of the inhibitor PC bound to this
variant by X-ray crystallography. The structural results provide
a rationale for the better resistance profile of inhibitors PC and
CC. Moreover, our crystallographic model suggests the exist-
ence of a hitherto unseen induced fit phenomenon.
PP
PC
CC
1.3 ( 0.2
3.2 ( 0.3
6 ( 1
9 ( 1
14 ( 3
12 ( 1
7
4
2
mutant HIV-1 PR strains have reduced binding affinity for the
Cha-containing compounds. Nevertheless, the sensitivity to the
mutation, as given by the ratio of the Ki values (mutant/native)
for the studied inhibitors, exhibits a pattern in which PP > PC
> CC, indicating that the new inhibitors are less sensitive to
the V82A strain. The reduced ratio in affinity between the native
and mutant enzymes for the compounds containing Cha
demonstrates that the Cha containing compounds are less
sensitive to the V82A mutation, suggesting that this group is
better able to fill the larger S1/S1′ pocket left by the V82A
mutation.
To gain an understanding about the structural modifications
that make the Cha containing inhibitors less sensitive to the
V82A mutation, the X-ray structure of the PC-HIV-1 PRV82A
complex was modeled and refined to 1.9 Å resolution in space
group P61 (PDB entry 2FLE). When CNS-SA omit maps were
used (see Experimental Section), it was not possible to assign
the Phe or Cha side chain to only the S1 or S1′ pocket. This is
indicative of a 2-fold (180°) bimodal binding mode that has
been previously observed for other inhibitors bound to the HIV-1
PR.1,3 To visualize the structural modifications introduced by
the V82A mutation and by the replacement of a Phe side chain
in P1 or P1′ by a Cha group, we have compared the structures
of the PC-HIV-1 PRV82A complex (PDB entry 2FLE) with
the PP-HIV-1 PRwt complex (PDB entry 1HVK).10 The V82A
mutation and the single side chain change in the inhibitor
produced significant structural changes in the protein and the
ligand, when compared to the 1HVK structure (see Figure 1).
The most significant changes are located in the 80s loop of the
Chemistry
L-Cyclohexylalanine and L-phenylalanine have been used as
starting materials for the preparation of the central core of the
HIV-1 protease inhibitors CC, PC, and PP. The synthetic
approach, depicted in Scheme 1, is based on the methodology
developed by Hoppe for the selective deprotonation of carbam-
ates, followed by addition to the appropriate electrophile.15 Thus,
substrate-directed deprotonation of carbamate 1 mediated by
s-BuLi/TMEDA,16 followed by reaction with amino aldehyde
2,17 allowed the preparation of 3a and 3b18 in 69% yield.
Hydrolysis of the carbamate moiety to 2,5-dibenzyldiamino-
3,4-diols 4 was performed by acid-catalyzed cleavage of
the oxazolidine ring followed by base-catalyzed cleavage of the
N-hydroxyalkylcarbamate intermediate (82–97% yield).19 The
N-benzyl protecting groups in diamines 4 were successfully
removed by a two-step sequence involving Pd(OH)2/C catalyzed
hydrogenolysis in the presence of (Boc)2O followed by TFA
treatment of the resulting dicarbamate, this sequence of reactions
afforded diamino-diols 6 in 78–55% overall yield.20 Finally,
acylation of 6a, 6b, or 6c with p-nitrophenylester 7,21 in the
presence of K2CO3, gave 8, 9, or 10 in 36, 52, and 81%
unoptimized yields, respectively.
Results and Discussion
Structural Factors behind Improved Resistance Profile.
The binding constants of inhibitors PP (10), PC (9), and CC
(8) with the native and V82A mutant strains of the HIV-1 PR
are listed in Table 2. The results indicate that both native and

854 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4
Clemente et al.
Other protein features that differ between these two crystal-
lographic structures are the loops involved in the flaps. As seen
from the lower panel of Figure 1, the conformation of residues
46–50 change from one complex to the other, both in the main
as well as in the side chains. The number of van der Waals
contacts also increase in the PC-HIV-1 PRV82A complex when
compared to the PP-HIV-1 PRwt complex (see Table 3).
Conformation of the Inhibitor Cha Group. Refinement of
the model against the density data suggested that the Cha group
is in a boat conformation. Restricting the Cha group to the chair
conformation and refining the model against the density results
in the C25 sp3 atom adopting a highly strained angle of 121.1°
compared to 113.9° for the boat conformation (Figure 2). This
supports the notion that the cyclohexane exists in a boat
conformation when the PC ligand is bound to the enzyme.
The likelihood that either the boat or the chair conformation
would occur when the ligand is bound to the enzyme depends
on the difference between the energy minima of these two
conformations when bound to the protein in reference to the
unbound conformations. In solution, the difference favors the
chair over the boat conformation.23 Binding of the inhibitor to
the enzyme could shift the balance between these two confor-
mations by stabilizing energetically either of them. For instance,
a higher enzyme affinity for the boat conformation could help
to compensate for the strain energy involved in docking this
conformer. Because the difference in affinity for the enzyme of
both the chair and the boat conformations cannot be obtained
experimentally, we have estimated the values through molecular
dynamics (MD)-based simulations, starting from the crystal-
lographic structure of the inhibitor PC bound to the V82A
HIV-1 PR variant (see Experimental Section). The calculated
energy histogram suggests that the interaction energies favor
the boat over the chair conformation (Figure 3). The highest
ranked inhibitor’s boat conformation outpaces the chair con-
formation by up to 10 kcal/mol. This energy difference may be
sufficient to offset the internal strain energy necessary to convert
the chair to the boat conformation. This ranking is borne out as
well by an energy minimization protocol directly applied to the
crystallographic refined structures, which indicates that the boat
containing inhibitor is favored over the one with a chair by 7.0
kcal/mol (see Experimental Section).
Figure 1. Some noteworthy features resulting from the superimposition
of the PP-HIV-1 PRwt (light blue) and PC-HIV-1 PRV82A (red)
complexes. The upper panel displays a close-up of the superimposed
80s loop, while the lower panel displays the superimposed flaps. The
inhibitors are shown as bold sticks.
protein, a feature that was observed in other structures of
inhibitors bound to the V82A HIV-1 PR variant.22 The van der
Waals contacts, below 4 Å, between the inhibitors and the active
site residues for both structures are reported in Table 3. As seen
from this table, some of the inhibitor-protein van der Waals
contacts are lost due to V82A mutation but are compensated
by the formation of a number of new ones in the PC bound
structure, keeping the binding free energy of both complexes
close. For instance, the loss of the van der Waals contacts
with the Val 82 CG1 atom at residues A82 and B82 is
compensated by new contacts made by the Cha fragment,
like the ones with the Ala 82 CB, Pro 81 CD, Ile A50 CA,
and CD1 atoms (see Table 3).
The central premise for the design of these inhibitors was that
the replacement of a Phe by a bulkier Cha would help to fill the
hydrophobic void left by the V82A mutation. Our results indicate
that the hypothesis used for inhibitor design is satisfied experi-
mentally. This is born out by the comparison of the shortest
distances between the CΒ atom of Val82 to the Phe side chain
and the CΒ atom of Ala82 to the Cha side chain. The closest atom
to the CB atom (in Val82) is C44 in the Phe group at P1, located
at 4.62 Å from the CB of Val B82. When the Phe group is replaced
by a Cha, the closest atom to the CB (in Ala 82) is C28, from the
Cha fragment at P1′, located at 3.48 Å from the CB of Ala A82,
demonstrating that the Cha side chain at P1′ fills some of the empty
space left by the V82A mutation.
The Cha boat conformation in the enzyme could be brought
about by trying to avoid steric clashes with the pyridine fragment
at P3. It has been shown previously that an increase in the size
of the P3 and P3′ fragments in C2-symmetric inhibitors with a
Phe group in P1 and P1′ led to a reduction in the binding to the
native HIV-1 PR, because P1 and P3 (as well as P1′ and P3′)
are positioned at close quarters into pockets with limited
capacity.24 Our binding data also showed that replacement of a
single Phe group at P1 or P1′ by the bulkier Cha fragment leads
to a lower affinity for HIV-1 PRV82A (see Table 2). We
hypothesize that the reduction in affinity may be due in part to
the penalty paid for the chair to boat transition in the complex
affected by steric interactions with the P3 fragments. Hence,
inhibitors that lack the P3 and P3′ fragments could, in principle,
accommodate the Cha fragment in its chair conformation, the
one preferred in solution. This could be the case of CGP 53820,
a pseudosymmetric HIV-1 PR inhibitor with a Cha fragment in
P1′, and lacking the P3′ fragment (see Table 1).25 We have
carried out MD-based calculations for this inhibitor bound to
the HIV-1 PR, similar to those performed on the PC-bound
complex described above. The starting point was the crystal-
lographic structure of the complex (PDB entry 1HII). The results
for the ligand-HIV-1 PR interaction energy distribution for the

Studies of HIV-1 Protease Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4 855
Table 3. Van der Waals Contacts with Distances below 4 Å between the Inhibitor’s Side Chain Fragments at P1/P1′ and the Protein Atoms in the
Complexes PP-HIV-1 PRwt (PDB entry 1HVK) and PC-HIV-1 PRV82A (PDB entry 2FLE)a
protease atom
PP (Phe) atom
distance (Å)
PCb (Cha) atom
distance (Å)
Ile A84/B84 CD1
Val A82/B82 CG1
Val A82/B82 CG1
Val A82/B82 CG1
Ala A82/B82 CB
Pro A81/B81 CB
Pro A81/B81 CG
Pro A81/B81 CD
Ile A50/B50 N
Ile A50 CA
C53/C41
C55, C56/C43
C57/C44
C58/C45
3.97/3.83
3.95, 3.69/3.56
3.62/3.49
3.83/3.93
C31/C31
-
-
3.68/3.60
-
-
-
-
-
-
C28/C28
C29/C29
-
C30/C30
C30/C31
C30
3.48/3.61
3.83/3.82
-
3.82/3.85
3.46/3.90
3.84
C56/-
C56/C43
-
C41, C42/C55
-
3.85/-
3.89/3.78
-
3.78, 3.41/3.65
-
Ile A50 CD1
Ile A50/B50 CG1
Gly B49 N
-
-
C30, C31
-/C31
C30
3.92, 3.90
-/3.83
3.51
C41, C42/-
-
3.78, 3.81/-
-
Gly A49/B49 CA
Gly A49/B49 C
Gly A49 O
C42/C55
C42/C55
-
-
3.91/3.93
3.71/3.71
-
-
C30/C30
C30/C30
C30
3.94/3.63
3.50/3.58
3.87
Leu A23/B23 CD2
C27/C27
3.82/3.93
a The slash (/) is used to differentiate the contacts between the chains A and B of the protein. In those cases where, in the same row, there are more than
one atom contacts for the inhibitor, we use commas (,) for the separation. b In this case, the van der Waals distances are for the same Cha fragment in the
two orientations of the inhibitor inside the protein.
Figure 4. Ligand-HIV-1 protease interaction energy histogram
obtained from the production stage of MD simulations of the complex
between the native HIV-1 PR and the CGP 53820 inhibitor with the
Figure 2. Overlap of the Cha side chain boat (in red) and chair (in
Cha fragment in its boat conformation (black) and chair conformation
green) models resulting from the refinement protocol of the crystal
(indigo).
structure of the PC ligand bound to the V82A mutant of HIV-1 PR.
2Fo - Fc density at 1.5 σ is drawn as a blue mesh. Valence angles for
In general, ligands rarely bind in their lowest energy
C25 in both conformations are shown in black.
conformation, and in many cases the bound conformation of
the ligand is at least 5 kcal/mol above its absolute minimum in
the unbound conformation in solution.26 Nevertheless, to the
best of our knowledge, this is the first report of a possible
enzyme “trapping” a cyclohexane in a boat conformation
through an induced-fit process.27,28
Conclusions
We have designed, synthesized, and tested a set of almost
C2-symmetric HIV-1 PR inhibitors, with a Cha group in P1,
P1′, or both. The design objective was to compensate for the
hydrophobic cavity created in the S1/S1′ pocket by the V82A
Figure 3. Ligand-HIV-1 protease interaction energy histogram
mutation. The binding data indicates that the second generation
inhibitors are less sensitive to the V82A mutation, in agreement
with our working hypothesis. To understand the structural
features underlying the response of these inhibitors for the V82A
mutation, we have solved the structure of the PC inhibitor bound
to the V82A HIV-1 PR by X-ray crystallography. Comparison
of this structure with that of the PP inhibitor bound to the native
HIV-1 PR indicates that the Cha group fills at least in part the
void left by the V82A mutation. An analysis of the alternative
ring conformations in the Cha group indicates the possibility
that this group may adopt the boat rather than the chair
conformation, which is prevalent in solution. MD simulations
support the conformation assignment and provide a rationale
for this outcome, indicating that the boat conformation can result
obtained from the production stage of the MD simulations of the
complex between the V82A protease variant and the PC inhibitor with
the Cha fragment in its boat conformation (black) and the chair
conformation (indigo).
Cha in its boat and chair conformation are shown in Figure 4.
These results indicate that the interaction energy favors the chair
over the boat conformation. Comparison of the ligand-protein
interaction energy histograms shown in Figures 3 and 4 suggests
that the presence of P3/P3′ fragments may be essential for
selecting the boat conformation for the P1/P1′ fragment of PC
when bound to the HIV-1 PR. In principle, the larger pocket
created by the V82A mutation could influence the chair/boat
equilibrium by accommodating better the chair conformation,
but our results indicate that this is not the case.

856 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4
Clemente et al.
from avoidance of steric clashes of the Cha group with the
pyridine fragment at P3/P3′. The existence of an induced-fit
phenomenon of this kind would have to be validated through
high resolution X-ray structures or neutron diffraction studies.
55.9, 55.6, 55.5, 54.4, 54.2, 54.1, 53.6, 52.9, 35.1, 35.0, 34.7,
34.5, 34.4, 34.2, 33.4, 33.2, 32.8, 32.4, 31.5, 31.2, 27.2, 26.8,
26.7, 26.4, 26.3, 26.0, 25.9, 25.4, 24.8, 24.3, 24.1. Anal.
(C54H73N3O4) C, H, N.
(2S,3S,4S,5S)-6-Cyclohexyl-2,5-bis(dibenzylamino)-4-hy-
droxy-1-phenylhexan-3-yl 2,2,4,4-tetramethyl-1,3-oxazolidine-
3-carboxylate (3b) and (2S,3S,4S,5S)-1-Cyclohexyl-2,5-
bis(dibenzylamino)-4-hydroxy-6-phenylhexan-3-yl 2,2,4,4-
tetramethyl-1,3-oxazolidine-3-carboxylate (3b′). A solution
of 1b (739 mg, 1.52 mmol) and TMEDA (342 µL, 2.28 mmol)
in Et2O (11 mL) at -78 °C was treated with s-BuLi (1.90 mL,
2.28 mmol, 1.2 M in cyclohexane). The resulting solution was
stirred at -78 °C for 5 h, then treated with a solution of
aminoaldehyde 2a (764 mg, 2.28 mmol) in Et2O (11 mL) and
stirred for 1.5 h. The cooling bath was removed and stirred for
30 min, then partitioned between EtOAc (20 mL) and H2O (30
mL). The organic phase was washed with brine (3 × 20 mL),
dried, and evaporated to give a mixture of two isomers, arising
from a partial migration of the Cby group to the newly
introduced hydroxyl group in the basic reaction medium, in a
60/40 ratio. Column chromatography (SiO2, EtOAc/hexane 1:8)
gave 3 in a combined yield of 69% (861 mg). Main isomer:
Experimental Section
Chemistry. General. All reactions were performed under an
inert atmosphere of dry argon in glassware that had been oven-
dried or flame-dried. Reaction temperatures are reported as the
temperature of the bath surrounding the vessel. Tetrahydrofuran
(THF) and diethyl ether (Et2O) were distilled from sodium/
benzophenone immediately before use, methylene chloride
(CH2Cl2), EtOAc, triethylamine, tetramethylethylenediamine
(TMEDA), and DMSO were distilled from CaH2, and methanol
was distilled from Mg(OMe)2. NMR spectra were recorded in
CDCl3 unless otherwise noted. Chemical shifts are reported in
ppm (δ) relative to the internal TMS (tetramethylsilane)
reference signal. J values are reported in hertz (Hz). Final
solutions were dried over anhydrous Na2SO4 before evaporation.
2,2,4,4-Tetramethyl-1,3-oxazolidine,29 2,2,4,4-tetramethyl-1,3-
oxazolidine-3-carbonyl chloride29 (CbyCl), 1b,30 2a,17b 2b,17a
3c,18 4c,18 and 721 were prepared according to the literature.
(2S)-3-Cyclohexyl-2-(dibenzylamino)propyl 2,2,4,4-tetram-
ethyl-1,3-oxazolidine-3-carboxylate (1a). To a suspension of
NaH (265 mg, 6.65 mmol) in THF (5 mL), a solution of (2S)-
3-cyclohexyl-2-(dibenzylamino)propan-1-ol17b (1.50 g, 4.44
mmol) in THF (2.5 mL) was added, and the resulting mixture
was stirred at rt for 30 min. A solution of CbyCl (1.02 g, 5.33
mmol) in THF (2 mL) was added and then stirred at rt for 5
days. The reaction was quenched by addition of 2 M HCl to
reach pH 7 and then extracted with EtOAc (3 × 15 mL). The
combined organic phase was washed with satd aq NaHCO3 (30
mL), H2O (30 mL), and brine (30 mL), dried, and evaporated
to give 1a, after column chromatography (SiO2, EtOAc/
hexane 1:15), as a colorless oil (1.73 g, 79% yield, 97% based
499 mg of 3b as a white foam (40%): [R]23 -27.9° (c 1.1,
D
1
CH2Cl2); IR (NaCl) 1732 cm-1; H NMR (C6D6, rotamers) δ
7.38 (d, J ) 7.0, 4H), 7.23–6.98 (m, 21H), 5.70 (bs, 1H), 4.45
(bs, 1H), 4.01 (d, J ) 14.0, 2H), 3.81 (d, J ) 7.0, 4H), 3.66 (d,
J ) 14.0, 2H), 3.44 (m, 3H), 3.15 (m, 1H), 3.00 (m, 2H),
1.80–0.81 (m, 25H), 0.28 (m, 1H); 13C NMR (rotamers) δ 152.9,
152.2, 140.4, 139.7, 139.6, 129.8, 128.7, 128.3, 128.2, 128.1,
128.0, 126.8, 126.7, 126.2, 96.3, 94.9, 76.3, 76.0, 75.5, 75.4,
71.0, 70.9, 61.2, 60.0, 59.0, 58.9, 55.9, 55.8, 54.6, 53.7, 34.4,
33.7, 33.4, 33.3, 31.2, 27.3, 27.0, 26.9, 26.3, 26.1, 25.8, 25.7,
25.1, 24.4, 24.2. Anal. (C54H67N3O4) C, H, N. Minor isomer:
362 mg of 3b′ as a white foam (29%): [R]20 -17.3° (c 1.5,
D
1
CH2Cl2); IR (NaCl) 1689 cm-1; H NMR (C6D6, rotamers) δ
7.39 (m, 5H), 7.20–6.98 (m, 20H), 5.75 (bs, 1H), 4.37 (bs, 1H),
3.91 (m, 4H), 3.77 (m, 4H), 3.37 (m, 3H), 3.20 (m, 2H), 3.00
(t, J ) 12.2, 1H), 2.80 (dd, J ) 3.1 and 13.9, 1H), 1.89–0.81
(m, 24 H), 0.27 (m, 1H); 13C NMR (rotamers) δ 153.1, 152.4,
140.03, 140.01, 129.9, 128.7, 128.3, 128.0, 127.9, 126.9, 126.5,
125.8, 96.3, 94.9, 76.3, 76.0, 72.3, 72.2, 61.18, 61.15, 61.0, 59.9,
54.5, 54.3, 54.2, 53.8, 35.4, 35.3, 34.6, 33.59, 33.56, 32.3, 32.2,
31.6, 31.5, 27.2, 26.9, 26.8, 26.34, 26.32, 26.1, 26.0, 25.4, 24.9,
24.4, 24.3. Anal. (C54H67N3O4) C, H, N.
on recovered starting material): [R]22 -22.6° (c 1.25,
D
CH2Cl2); IR (NaCl) 1695 cm-1; H NMR (rotamers) δ 7.32
1
(m, 10H), 4.36 (m, 1H), 4.13 (m, 1H), 3.79 (m, 4H), 3.60
(d, J ) 13.9, 2H), 2.99 (m, 1H), 1.75–1.05 (m, 23H), 0.92
(m, 1H), 0.62 (m, 1H); 13C NMR (rotamers) δ 152.7 and
152.0, 140.0, 128.8, 127.9, 126.7, 95.8 and 94.5, 76.2 and
75.8, 60.6 and 59.4, 53.6 and 53.0, 36.84 and 36.78, 34.0,
33.8, 32.6, 26.44, 26.40, 26.0, 25.34, 25.28, 25.12, 25.07,
24.0. Anal. (C31H44N2O3) C, H, N.
(2S,3S,4S,5S)-1,6-Dicyclohexyl-2,5-bis(dibenzylamino)hex-
ane-3,4-diol (4a). A solution of 3a (485 mg, 0.59 mmol) and
MeSO3H (305 µL, 2.93 mmol) in MeOH (7 mL) was refluxed
for 16 h in a sealed tube. After cooling to rt, it was treated with
a mixture of LiOH ·H2O (197 mg, 4.68 mmol), dioxane (7 mL),
and H2O (7 mL). The resulting mixture was refluxed for 2 h,
cooled to rt, diluted with CH2Cl2 (15 mL), and neutralized with
10% H3PO4. The organic layer was washed with H2O (15 mL)
and brine (15 mL), dried, filtered, and evaporated to give a
residue which was chromatographed (SiO2, EtOAc/hexane 1:6)
(2S,3S,4S,5S)-1,6-Dicyclohexyl-2,5-bis(dibenzylamino)-4-
hydroxyhexan-3-yl 2,2,4,4-tetramethyl-1,3-oxazolidine-3-car-
boxylate (3a). To a solution of 1a (435 mg, 0.88 mmol) and
TMEDA (200 µL, 1.32 mmol) in Et2O (7 mL) at -78 °C was
added s-BuLi (1.0 mL, 1.32 mmol, 1.3 M in cyclohexane). The
resulting solution was stirred at -78 °C for 5 h, and then treated
with a solution of aldehyde 2a (444 mg, 1.32 mmol) in Et2O (6
mL). The resulting mixture was stirred for 1.5 h at -78 °C,
then the cooling bath was removed and stirring was continued
for an additional 30 min. The mixture was partitioned between
H2O (20 mL) and EtOAc (30 mL). The organic layer was
washed with brine, dried, filtered, and evaporated to give a
residue that was purified by column chromatography (SiO2,
to yield 4a as a white foam (323 mg, 82%): [R]20 -18.4° (c
D
1.0, CH2Cl2); 1H NMR δ 7.27 (m, 20H), 3.90 (d, J ) 4.3, 2H),
3.71 (d, J ) 14.3, 4H), 3.64 (d, J ) 14.3, 4H), 2.78 (m, 2H),
2.27 (bs, 2H), 1.75–1.12 (m, 22H), 0.93 (m, 2H), 0.75 (m, 2H);
13C NMR δ 140.1, 128.8, 128.2, 126.9, 71.4, 56.7, 54.5, 34.6,
34.4, 33.9, 33.5, 26.6, 26.5, 26.2. Anal. (C46H60N2O2) C, H, N.
EtOAc/hexane 1:8) to yield 3a as a white foam (505 mg, 69%):
1
[R]20 -42.4° (c 1.0, CH2Cl2); IR (CsI) 1692 cm-1; H NMR
D
(rotamers) δ 7.21 (m, 20H), 5.37 (bs, 1H), 4.16 (bs, 1H),
3.92–3.20 (m, 11 H), 2.85 (m, 1H), 2.66 (bd, J ) 9.7, 1H),
1.77–0.76 (m, 36H), 0.04 (m, 2H); 13C NMR (rotamers) δ 152.9,
152.2, 140.6, 140.2, 139.3, 128.9, 128.4, 128.3, 128.2, 128.0,
127.0, 126.8, 126.6, 96.1, 94.7, 76.2, 75.8, 75.4, 71.6, 61.0, 59.7,
(2S,3S,4S,5S)-6-Cyclohexyl-2,5-bis(dibenzylamino)-1-phe-
nylhexane-3,4-diol (4b). Same procedure as above applied to
both isomers 3b and 3b′ yielded the same product 4b in 97%
yield: [R]22D -6.8° (c 1.2, CH2Cl2); 1H NMR δ 7.23–7.07 (m,

Studies of HIV-1 Protease Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4 857
25H), 3.90 (d, J ) 3.8, 1H), 3.86 (d, J ) 6.3, 1H), 3.65 (d, J
) 13.3, 2H), 3.54 (t, J ) 14.3, 4H), 3.49 (d, J ) 13.3, 2H),
2.98 (m, 3H), 2.69 (q, J ) 6.4, 1H), 1.97 (bs, 2H), 1.64–1.39
(m, 7H), 1.27 (m, 1H), 1.19–1.04 (m, 3H), 0.81 (m, 1H), 0.66
(m, 1H); 13C NMR δ 141.6, 140.0, 139.6, 129.7, 128.9, 128.7,
128.3, 128.2, 127.03, 127.01, 125.8, 72.2, 71.9, 61.9, 56.5, 54.7,
54.6, 34.9, 34.0, 33.6, 32.8, 26.7, 26.5, 26.3. Anal. (C46H54N2O2)
C, H, N.
(2S,3S,4S,5S)-2,5-Diamino-1,6-diphenylhexane-3,4-diol Bis-
trifluoracetate (6c). Following the same procedure as above,
6c was prepared from 286 mg of 5c (0.57 mmol) in quantitative
1
yield (302 mg): H NMR (CD3OD) δ 7.29 (m, 6H), 7.17 (m,
4H), 4.09 (d, J ) 3.1, 2H), 3.70 (m, 2H), 3.02 (dd, J ) 7.1,
14.3, 2H), 2.89 (dd, J ) 8.4, 14.3, 2H); 13C NMR (CD3OD) δ
162.7 (q, JC-F ) 35.5), 136.5, 130.2, 129.7, 128.5, 118.0 (q,
JC-F ) 290.7), 68.4, 58.4, 35.8.
(2S,3S,4S,5S)-2,5-Bis[(tert-butoxycarbonyl)amino]-1,6-di-
cyclohexylhexane-3,4-diol (5a). To a deoxygenated solution
of 4a (292 mg, 0.43 mmol) and (Boc)2O (260 mg, 1.19 mmol)
in EtOAc (10 mL) was added 20% Pd(OH)2/C (117 mg), and
the mixture was stirred under hydrogen for 4 days. The catalyst
was removed by filtration through Celite and washed thoroughly
with EtOAc and warm CH2Cl2. The filtrate and washings were
evaporated to give a white solid that was recrystallized from
CH2Cl2/EtOH (137 mg). The mother liquors were evaporated
and the residue was chromatographed (SiO2, EtOAc/hexane 1:6
to 1:3) to yield 36 mg of 5a (173 mg, 78% combined yield):
mp 239–241 °C (CH2Cl2/EtOH); [R]20D -50.8° (c 0.5, CH2Cl2);
(2S,3S,4S,5S)-2,5-Bis[N-[N-[[N-methyl-N-(2-pyridinylme-
thyl)amino]carbonyl]valinyl] amino]-1,6-dicyclohexylhexane-
3,4-diol (8). A solution of 721 (132 mg, 0.34 mmol) in THF
(1.6 mL) was added to a mixture of 6a (74 mg, 0.14 mmol)
and K2CO3 (38 mg, 0.27 mmol) in THF (1.6 mL). The resulting
mixture was stirred at room temperature for 20 h, diluted with
H2O (1 mL), basified to pH 12 with 1 M NaOH, and stirred for
an additional hour. The mixture was partitioned between H2O
(5 mL) and EtOAc (2 × 10 mL). The combined organic phase
was washed with satd aq K2CO3 (5 × 5 mL), dried, evaporated,
and purified by column chromatography (SiO2, CH2Cl2/Et2O
90:10 to CHCl3/MeOH 95:5) to yield 8 as a white solid, which
was recrystallized from CH2Cl2/Et2O (40 mg, 36%): mp
136–138 °C (CH2Cl2/Et2O); [R]25D -11.2° (c 1.0, CH2Cl2); IR
1
IR (CsI) 1662 cm-1; H NMR δ 4.47 (d, J ) 8.8, 2H), 3.74
(m, 4H), 3.27 (m, 2H), 1.88–1.54 (m, 10H), 1.44 (s, 18H),
1.50–0.75 (m, 16H); 13C NMR δ 157.2, 80.0, 73.0, 50.0, 39.5,
34.3, 32.2, 28.3, 26.5, 26.4, 26.1. Anal. (C28H52N2O6) C, H, N.
1
(CsI) 1655, 1633, 1522 cm-1; H NMR δ 8.55 (d, J ) 4.7,
2H), 7.74 (bt, J ) 7.1, 2H), 7.31 (bd, J ) 7.9, 2H), 7.27 (m,
2H), 6.73 (bs, 2H), 6.39 (bs, 2H), 4.66 (bd, J ) 12.4, 2H), 4.43
(d, J ) 15.9, 2H), 4.08 (m, 2H), 4.04 (bt, J ) 6.4, 2H), 3.42 (d,
J ) 6.0, 2H), 3.01 (s, 6H), 2.32 (oct, J ) 6.7, 2H), 1.76 (bd, J
) 12.7, 2H), 1.63 (m, 10H), 1.29 (m, 4H), 1.21–0.81 (m, 10H),
0.99 (d, J ) 7.0, 6H), 0.97 (d, J ) 7.0, 6H); 13C NMR δ 173.8,
159.2, 157.2, 149.3, 137.4, 122.9, 122.5, 72.0, 60.3, 54.9 (CH2),
49.1, 39.3 (CH2), 35.2, 34.3, 34.0 (CH2), 32.0 (CH2), 29.4, 26.42
(CH2), 26.39 (CH2), 26.1 (CH2), 19.7, 17.4. Anal.
(C44H70N8O6 ·H2O) C, H, N.
(2S,3S,4S,5S)-2,5-Bis[(tert-butoxycarbonyl)amino]-6-cyclo-
hexyl-1-phenylhexane-3,4-diol (5b). Following the above
procedure, 209 mg of 5b (55% yield) was prepared from 4b
(500 mg, 0.75 mmol) as a white solid after recrystallization from
CH2Cl2/EtOH: mp 210–212 °C (CH2Cl2/EtOH); [R]25D -26.4°
(c 1.0, CH2Cl2); IR (CsI) 1664 cm-1; 1H NMR δ 7.25 (m, 5H),
4.53 (d, J ) 8.8, 1H), 4.30 (d, J ) 6.1, 2H), 4.01 (m, 1H), 3.85
(d, J ) 4.8, 1H), 3.71 (m, 1H), 3.32 (m, 1H), 3.13 (m, 2H),
2.98 (dd, J ) 3.6 and 13.4, 1H), 1.99–0.76 (m, 13H), 1.45 (s,
9H), 1.38 (s, 9H); 13C NMR δ 157.3, 156.8, 137.4, 129.8, 128.5,
126.4, 80.2, 80.1, 73.1, 70.3, 51.4, 49.6, 39.7, 36.4, 34.3, 32.1,
28.34, 28.27, 26.5, 26.4, 26.1. Anal. (C28H46N2O6) C, H, N.
(2S,3S,4S,5S)-2,5-Bis[N-[N-[[N-methyl-N-(2-pyridinylme-
thyl)amino]carbonyl]valinyl] amino]-6-cyclohexyl-1-phenyl-
hexane-3,4-diol (9). A total of 88 mg of 9 (52% yield) was
obtained as a white solid from 721 (191 mg, 0.49 mmol) and
6b (105 mg, 0.20 mmol) by following the same procedure as
(2S,3S,4S,5S)-2,5-Bis[(tert-butoxycarbonyl)amino]-1,6-diphe-
nylhexane-3,4-diol (5c).20 Following the same procedure as for
5a, 5c was prepared from 4c (1.37 g, 2.07 mmol) in 74% yield
(767 mg) as a white solid after recrystallization: mp 213–215
for 8: mp 106–108 °C (CH2Cl2/Et2O); [R]25 +10.6° (c 1.0,
D
CH2Cl2); IR (CsI) 1647, 1572 cm-1; 1H NMR δ 8.58 (m, 2H),
7.85 (m, 2H), 7.38 (m, 4H), 7.21 (m, 4H), 7.12 (m, 1H), 6.36
(bs, 1H), 6.18 (bs, 1H), 4.88 (bs, 2H), 4.36 (m, 3H), 4.14 (m,
1H), 3.96 (bs, 2H), 3.56 (bs, 2H), 3.02 (s, 3H), 3.00 (s, 3H),
2.98 (m, 1H), 2.90 (dd, J ) 8.4 and 14.1, 1H), 2.30 (oct, J )
6.7, 1H), 2.22 (oct, J ) 6.7, 1H), 1.77 (d, J ) 13.0, 2H),
1.68–1.42 (m, 6H), 1.25 (m, 2H), 1.12 (m, 3H), 0.98 (m, 6H),
0.85 (d, J ) 6.8, 3H), 0.79 (d, J ) 6.8, 3H); 13C NMR δ 174.0,
173.9, 159.2, 159.0, 157.2, 157.0, 149.3, 149.2, 138.3, 137.4,
137.3, 129.1, 128.4, 126.2, 123.0, 122.9, 122.7, 122.4, 72.1,
70.7, 60.3, 59.8, 54.9 (CH2), 54.8 (CH2), 52.7, 49.0, 39.4 (CH2),
37.6 (CH2), 35.2, 35.1, 34.3, 34.1 (CH2), 31.8 (CH2), 29.5, 29.3,
26.4 (CH2), 26.3 (CH2), 25.9 (CH2), 19.8, 19.4, 17.4, 16.5. Anal.
(C44H64N8O6 ·1.5H2O) C, H, N.
°C (CH2Cl2/EtOH); [R]25 -13.4° (c 1.0, CH2Cl2); IR (CsI)
D
1662 cm-1; 1H NMR δ 7.22 (m, 10H), 4.42 (m, 4H), 4.02 (m,
2H), 3.20 (m, 4H), 2.97 (dd, J ) 4.3 and 13.7, 2H), 1.41 (s,
18H); 13C NMR δ 156.9, 137.0, 129.8, 128.6, 126.5, 80.2, 70.1,
50.8, 36.3, 28.3. Anal. (C28H40N2O6) C, H, N.
(2S,3S,4S,5S)-2,5-Diamino-1,6-dicyclohexylhexane-3,4-di-
ol Bistrifluoracetate (6a). A solution of 5a (150 mg, 0.29
mmol) in TFA-CH2Cl2 (1:2, 6 mL) was stirred for 15 min at
rt and then evaporated to give a pale yellow foam (158 mg,
1
100%) that was used without further purification: H NMR
(CD3OD) δ 4.02 (d, J ) 3.9, 2H), 3.44 (m, 2H), 1.86–0.69 (m,
22H), 0.97 (m, 4H); 13C NMR (CD3OD) δ 161.5 (q, JC-F
37.3), 117.3 (q, JC-F ) 289.4), 69.0, 54.5, 37.6, 35.0, 34.2, 33.9,
27.4, 27.2, 27.1.
)
(2S,3S,4S,5S)-2,5-Bis[N-[N-[[N-methyl-N-(2-pyridinylme-
thyl)amino]carbonyl]valinyl] amino]-1,6-diphenylhexane-3,4-
diol (10). A mixture of 6c (200 mg, 0.40 mmol) and K2CO3
(110 mg, 0.80 mmol) in THF (1.5 mL) at 0 °C was treated
with a solution of 721 (217 mg, 0.66 mmol) in THF (2 mL).
The resulting mixture was stirred at 0 °C for 10 h, then warmed
to rt, and stirred for an additional 10 h. A second addition of 7
(200 mg, 0.61 mmol) in THF (2 mL) was carried out, and the
resulting mixture was stirred for 20 h at room temperature. H2O
(2 mL) was added, followed by 1 M NaOH until pH 12 was
reached, and then stirred for an additional hour. The reaction
(2S,3S,4S,5S)-2,5-Diamino-6-cyclohexyl-1-phenylhexane-
3,4-diol Bistrifluoroacetate (6b). Following the same procedure
as above, 6b was prepared from 5b (100 mg, 0.20 mmol) as a
pale yellow foam in quantitative yield (105 mg): 1H NMR
(CD3OD) δ 7.35 (m, 5H), 4.15 (m, 1H), 3.87 (m, 1H), 3.78 (m,
1H), 3.42 (m, 1H), 3.21 (dd, J ) 6.6, 14.2, 1H), 2.99 (dd, J )
9.6, 14.2, 1H), 1.81–1.06 (m, 10H), 0.88 (m, 3H); 13C NMR
(CD3OD) δ 158.8 (q, JC-F ) 41.7), 136.4, 130.1, 129.7, 128.3,
115.8 (q, JC-F ) 283.9), 68.2, 67.8, 57.8, 54.5, 37.6, 35.9, 34.32,
34.27, 33.6, 27.3, 26.8, 26.7.

858 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4
Clemente et al.
Table 4. Crystallographic Statistics
Data Collection Statistics
mixture was extracted with EtOAc (2 × 15 mL), the combined
organic phase was washed with satd aq K2CO3 (5 × 10 mL),
dried, evaporated, and chromatographed (SiO2, CH2Cl2/Et2O 90:
10 to CHCl3/MeOH 95:5) to give 10 as a white solid (256 mg,
81%), which was recrystallized from CH2Cl2/Et2O: mp 110 °C
(softens) (CH2Cl2/Et2O); [R]25D +28.0° (c 1.0, CH2Cl2); IR (CsI)
wavelength (Å)
resolution range (Å)
space group
1.1
20–1.9
P61
62.2, 82.8
76487
14390
100 (100)
13.3
7.1 (48.6)
unit-cell parameters a, c (Å)
number of reflections
number of unique reflections
overall completeness (%)
average I/s
1
1646, 1571 cm-1; H NMR δ 8.53 (d, J ) 4.6, 2H), 7.74 (t, J
) 7.1, 2H), 7.27 (m, 4H), 7.16 (m, 8H), 7.09 (m, 2H), 6.73
(bs, 2H), 6.14 (bs, 2H), 4.59 (bd, J ) 15.9, 2H), 4.35 (d, J )
15.9, 2H), 4.31 (m, 2H), 4.08 (dd, J ) 5.4 and 7.1, 2H), 3.46
(d, J ) 8.0, 2H), 3.20 (dd, J ) 4.2 and 14.2, 2H), 2.96 (s, 6H),
2.77 (dd, J ) 9.0 and 14.3, 2H), 2.23 (oct, J ) 6.8, 2H), 0.85
(d, J ) 6.7, 6H), 0.72 (d, J ) 6.7, 6H); 13C NMR δ 174.0,
159.1, 157.0, 149.3, 138.7, 137.4, 129.2, 128.3, 126.1, 123.0,
122.4, 71.4, 59.9, 54.7 (CH2), 52.1, 37.5 (CH2), 35.1, 29.3, 19.4,
16.3. Anal. (C44H58N8O6 ·H2O) C, H, N.
Rsym (%)a
Refinement Statistics
Rwork (%)b
19.9 (22.5)
23.5 (26.8)
0.009
Rfree (%)c
rmsd bond length (Å)
rmsd bond angle (deg)
1.5
Average B Factor (Å2)
Mutagenesis and Expression of the Protease. A complete
description of the cloning, expression, and purification proce-
dures can be found in Ido et al.31 and Goodenow et al.32 In
brief, the HIV-1 protease DNA was subcloned into the pET23a
expression vector (Novagen).33 The construct was transformed
into the E. coli strain BL21 Star DE3 PlysS from Invitrogen.
The V82A mutation was introduced into the HIV-PR back-
ground using the QuikChange Mutagenesis Kit from Stratagene.
Protease expression in bacteria was initiated when the OD600
reached 0.6 by addition of 1 mM IPTG to a culture grown at
37 °C in M9 Media (6.8 g Na2HPO4, 3 g KH2PO4, 0.5 g NaCl,
1 g NH4SO4, and 5 g casamino acids were autoclaved together
in 987 mL of H2O, and then 1 mL of 0.1 M CaCl2, 2 mL of 1.0
M MgSO4, 10 mL of 20% glucose, and 50 µg/L of ampicillin
were added). After three hours, cells were harvested by
centrifugation at 16000 g for five minutes and resuspended in
TN buffer (0.05 M Tris, 0.15 M NaCl, 0.001 M MgCl2, pH
7.4). Inclusion bodies containing the protease were isolated by
centrifugation through a 27% sucrose cushion. The inclusion
bodies were solubilized in 8 M urea, and the protease was
refolded by dialysis against 0.05 M sodium phosphate buffer
(0.05 M Na2HPO4, 0.005 M EDTA, 0.3 M NaCl, and 0.001 M
dTT, pH 7.3). The protease was purified through ammonium
sulfate precipitation and gel filtration chromatography using a
Superdex 75 16/60 column from Amersham Pharmacia attached
to an FPLC LCC 500 Plus, also from Pharmacia. The protease
was eluted using potassium phosphate buffer (50 mM K2HPO4,
2 mM EDTA, 150 mM NaCl, 2 mM dTT, 5% glycerol, and
5% isopropanol, pH 7.3).
protein main/side chains
inhibitor atoms
water molecules
30.7/35.7
32.1
39.5
Ramachandran Plot Quality
most favored (%)
94.9
5.1
0
additionally allowed (%)
generally allowed (%)
disallowed (%)
0
a Rsym
)
hkl (| I(hkl) - Ii(hkl)|⁄ Ii(hkl) )×100, where Ii(hkl) is the ith
i
∑
observation of the intensity of a reflection with indices h, k, and l and I
(hkl) is the average intensity of all symmetry equivalent measurements of
b
∑
those reflections. Rwork
)
hkl (Fobs(hkl) - Fcalc(hkl))⁄Fobs(hkl)×100 ,
where Fobs(hkl) and Fcalc(hkl) are the observed and calculated structure factor
amplitudes respectively. Rfree was calculated using 5% of data excluded
c
during the refinement process. In parenthesis is the highest resolution shell.
by centrifugation at 10000 g at 4 °C. The enzyme–inhibitor
complex was mixed with reservoir solution (Hampton Research
Cryo Crystallization Kit #18) in a 1:1 (v/v) ratio to set up 4 µL
hanging drops at 25 °C. Rod-shaped crystals grew in two days.
X-ray Data Collection, Structure Determination, and
Refinement. X-ray diffraction images were collected at 100 K
at the Brookhaven National Laboratory beamline X29 on an
ADSC quantum 315 CCD detector. The crystal was dipped in
cryoprotectant solution (30% glycerol in reservoir solution) prior
to data collection. A complete data set was collected from a
single crystal. The data were indexed, scaled, and reduced using
DENZO and SCALEPACK.37 The complex crystallized in the
P61 space group.
Molecular replacement was used to calculate initial phases
employing the coordinates of PDB entry 1SGU as a search
model after removing inhibitor and waters. Standard methods
of structure refinement were then employed using programs in
the CNS Suite.38 Electron density maps with coefficients 2Fo-Fc
and Fo-Fc were used to guide manual fitting of the protease
and bound inhibitor followed by real space refinement using
the molecular graphics program WinCoot.39 The inhibitor and
water molecules were initially modeled into Fo-Fc density at
2σ. Detailed refinements of the cyclohexyl conformations were
done utilizing CNS SA-omit maps holding the protease residue
coordinates fixed. Eight possible conformations were sampled,
four chair and four boat conformations, including 2-fold
disordered inhibitor conformations. A 2-fold boat conformation
was found to provide the best fit to the density. During model
building and refinement, 5% of the data was reserved for cross
validation of the refinement progress. The atomic coordinates
have been deposited in the Protein Data Bank (PDB entry
2FLE). The X-ray data collection and refinement statistics are
shown in Table 4.
Determination of Inhibition Constants. Ki values were
determined for each inhibitor with the HIV-PR strains as
previously described.34 Cleavage of the chromogenic substrate
(K-A-R-V-L*Nph-E-A-nL-G, Nph ) p-nitrophenylalanine, nL
) norleucine) was monitored using a Hewlett-Packard 8452A
spectrophotometer equipped with a seven-cell sample handling
system at 37 °C in sodium acetate buffer (0.05 M NaOAc, 0.15
M NaCl, 0.002 M EDTA, 0.001 M dTT, pH 4.7), as described
by Dunn et al.35 The inhibition constants Ki were determined
by monitoring the inhibition of hydrolysis of the chromogenic
substrate, as described by Bhatt et al.36
Crystallization of Protein-Inhibitor Complex. After the
HIV-PRV82A was purified, the enzyme purification buffer was
exchanged for 50 mM sodium acetate, 1 mM EDTA, and 1
mM dTT at pH 4.7 during concentration to 2 mg/mL. The
inhibitor PC was dissolved at a concentration of 20 mM in 100%
DMSO and mixed with the HIV- PRV82A in a molar ratio of
3:1. The inhibitor and enzyme were allowed to equilibrate for
one hour at 4 °C, after which precipitated material was removed

Studies of HIV-1 Protease Inhibitors
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4 859
(3) Wlodawer, A.; Vondrasek, J. Inhibitors of HIV-1 protease: A major
success of structure-assisted drug design. Annu. ReV. Biophys. Biomol.
Struct. 1998, 27, 249–284.
Molecular Modeling. (a). Energy Calculations from the
Modeled Complexes with Cha in the Boat and Chair
Conformations. The starting point for our calculations was the
crystal structures of the HIV-1 PR bound to two ligands: the
inhibitor PC studied in the present work and the inhibitor CGP
53820, from PDB entry 1HII25 (see Table 1 for chemical
structures). In every case, hydrogen atoms were added using
the Biopolymer module in InsightII.40 We chose to have the
active site aspartic acid dyad monoprotonated by neutralizing
the residue A25. The position of the protons was optimized using
an energy minimization protocol with a gradient tolerance of
0.01 kcal/Å2 in the CHARMM suite of programs.41 The solvent
was modeled by surrounding these complexes by 10 Å layer
with about 2000 water molecules by using the module Soak in
InsightII.40 The water molecules were equilibrated by a 0.5 ns
molecular dynamics run at 300 K and a NVT ensemble using
the CHARMM module with CHARMM potentials available in
InsightII. To predict the relative stability of either chair or boat
conformations in both systems, we built copies of these
complexes with their alternative Cha conformations. In this way,
we generated the HIV-1 PRV82A-PC complex with the Cha in
a chair conformation and a HIV-1 PR-CGP 58320 complex
with the inhibitor Cha side chain in a boat conformation. The
Cha fragments were built using their original ꢀ1 and ꢀ2 side chain
torsion angles found in the crystal structures. All systems were
then subject to a two-step molecular mechanics protocol. The
first step contained an energy minimization with an ABNR
minimizer and a gradient tolerance of 0.01. In the second step,
more than 350 ps of dynamics were performed on a subsystem
of the complex containing all atoms at a 10 Å radius from the
ligand, at 300 K in a NVT ensemble. The affinity of the ligands
for the enzyme was gauged using their molecular mechanics-
based interaction energy, evaluated in the last 200 ps of the
MD production stage.
(4) Condra, J. H.; Schleif, W. A.; Blahy, O. M.; Gabryelski, L. J.; Graham,
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. In vivo emergence of HIV-1 variants
resistant to multiple protease inhibitors. Nature 1998, 374, 569–571.
(5) Ho, D. D.; Toyoshima, T.; Mo, H.; Kempf, D. J.; Norbeck, D.; Chen,
C.-M.; Wideburg, N. E.; Burt, S. K.; Erickson, J. W.; Singh, M. K.
Characterization of human immunodeficiency virus type 1 variants
with increased resistance to a C2-symmetric protease inhibitor. J. Virol.
1994, 68, 2016–2020.
(6) Ohtaka, H.; Muzammil, S.; Schön, A.; Velazquez-Campoy, A.; Vega,
S.; Freire, E. Thermodynamic rules for the design of high affinity
HIV-1 protease inhibitors with adaptability to mutations and high
selectivity towards unwanted targets. Int. J. Biochem. Cell Biol. 2004,
36, 1787–1799.
(7) Clemente, J. C.; Moose, R. E.; Hemrajani, R.; Whitford, L. R. S.;
Govindasamy, L.; Reutzel, R.; McKenna, R.; Agbandje-McKenna, M.;
Goodenow, M. M.; Dunn, B. M. Comparing the accumulation of
active- and nonactive-site mutations in the HIV-1 protease. Biochem-
istry 2004, 43, 12141–12151.
(8) Baldwin, E. T.; Bhat, T. N.; Liu, B.; Pattabiraman, N.; Erickson, J. W.
Structural basis of drug resistance for the V82A mutant of HIV-1
proteinase. Nat. Struct. Biol. 1995, 2, 244–249.
(9) Matthews, B. W. Studies on protein stability with T4 lysozyme. AdV.
Protein Chem. 1995, 46, 249–278.
(10) Hosur, M. V.; Bhat, T. N.; Kempf, D. J.; Baldwin, E. T.; Liu, B.;
Gulnik, S.; Wideburg, N. E.; Norbeck, D. W.; Appelt, K.; Erickson,
J. W. Influence of stereochemistry on activity and binding modes for
C2 symmetry-based diol inhibitors of HIV-1 protease. J. Am. Chem.
Soc. 1994, 116, 847–855.
(11) Ohtaka, H.; Freire, E. Adaptive inhibitors of the HIV-1 protease. Prog.
Biophys. Mol. Biol. 2005, 88, 193–208.
(12) Yin, P. D.; Das, D.; Mitsuya, H. Overcoming HIV drug resistance
through rational drug design based on molecular, biochemical, and
structural profiles of HIV resistance. Cell. Mol. Life Sci. 2006, 63,
1706–1724.
(13) King, N. M.; Prabu-Jeyabalan, M.; Nalivaika, E. A.; Schiffer, C. A.
Combating susceptibility to drug resistance: lessons from HIV-1
protease. Chem. Biol. 2004, 11, 1333–1338.
(14) Vega, S.; Kang, L.-W.; Velazquez-Campoy, A.; Kiso, Y.; Amzel,
L. M.; Freire, E. A structural and thermodynamic escape mechanism
from a drug resistant mutation of the HIV-1 protease. Proteins 2004,
55, 594–602.
(15) Hoppe, D.; Marr, F.; Brüggemann, M. Enantioselective synthesis by
lithiation adjacent to oxygen and electrophile incorporation. In Topics
in Organometallic Chemistry; Hodgson, D. M., Ed.; Springer: Berlín,
2003; Vol. 5, pp. 61–138.
(16) Schwerdtfeger, J.; Hoppe, D. Stereoselective generation of 1-acyloxy-
2-amino carbanions by deprotonation; synthesis of enantiomerically
and diastereomerically pure ꢁ-amino alcohols. Angew. Chem., Int. Ed.
1992, 31, 1505–1507.
(17) (a) Reetz, M. T.; Drewes, M. W.; Schwickardi, R. Preparation of
enantiomerically pure R-N,N-dibenzylamino aldehydes: S-2-(N,N-
dibenzylamino)-3-phenylpropanal. Org. Synth. 1999, 76, 110–122. (b)
Alexander, C. W.; Liotta, D. C. A diastereoselective synthesis of
(2S,3R,4S)-2-amino-1-cyclohexyl-6-methylheptane-3,4-diol, the Abbott
aminodiol. Tetrahedron Lett. 1996, 37, 1961–1964.
(b). Energy Evaluation of the Structures Resulting
from the X-ray Refinement. The X-ray refinement allows the
fitting of the Cha fragment in both boat and retracted chair
conformation (see Figure 2). We have used the coordinates of
these two structures as the starting point for their energy
optimization. The protons were added and optimized as above.
Solvation was taken into account by the equilibrated 10 Å layer
with about 2000 water molecules described above. At this point,
the whole system was energy-minimized to alleviate the clashes
resulting from X-ray refinement with a protocol starting with
200 steps of SD minimization followed by an ABNR minimiza-
tion with a gradient tolerance of 0.05 kcal/Å2. The affinity of
the ligands was evaluated as above.
(18) For the preparation of 3c and 4c see: Weber, B.; Kolczewski, S.;
Fröhlich, R.; Hoppe, D. Facile and highly stereoselective synthesis of
the C2-symmetrical diamino diol core-unit of HIV-1 protease inhibitors
and of their symmetrical and unsymmetrical analogs from lithiated
2-(dibenzylamino)alkyl carbamates: oxidative dimerization. Synthesis
1999, 1593–1606.
(19) van Bebber, J.; Ahrens, H.; Fröhlich, R.; Hoppe, D. Efficient
desymmetrization of meso-cis-1,2-cyclohexanedimethanol with dif-
ferentiation between diastereotopic and enantiotopic C-H bonds by
(-)-sparteine-mediated deprotonation. Chem.sEur. J. 1999, 5, 1905–
1916.
Acknowledgment. This work was supported by grants from
the NIH (NIH Grant AI28571 to B.M.D. and Post-Doctoral
Fellowship from Training Grant T32-CA09126 to J.C.C.), the
Spanish Ministry of Education and Science (Grant CTQ2006-
07854/BQU to F.J.S. and a fellowship to P.G.), and Xunta de
Galicia (to F.S. and Grant PGIDT06PXIB209109PR to F.J.S.).
We thank the CESGA for computer time.
(20) For a different preparation of 5c see: Kempf, D. J.; Sowin, T. J.;
Doherty, E. M.; Hannick, S. M.; Codavoci, L.; Henry, R. F.; Green,
B. E.; Spanton, S. G.; Norbeck, D. W. Stereocontrolled synthesis of
C2-symmetric and pseudo-C2-symmetric diamino alcohols and diols
for use in HIV protease inhibitors. J. Org. Chem. 1992, 57, 5692–
5700.
(21) The side chain was prepared following the literature procedure: Kempf,
D. J.; Codacovi, L.; Wang, X. C.; Kohlbrenner, W. E.; Wideburg,
N. E.; Saldivar, A.; Vasavanonda, S.; Marsh, K. C.; Bryant, P.; Sham,
H. L.; Green, B. E.; Betebenner, D. A.; Erickson, J.; Norbeck, D. W.
Symmetry-based inhibitors of HIV protease. Structure–activity studies
of acylated 2,4-diamino-1,5-diphenyl-3-hydroxypentane and 2,5-
Supporting Information Available: Results from elemental
analysis. This material is available free of charge via the Internet
at http://pubs.acs.org.
References
(1) Wlodawer, A.; Erickson, J. W. Structure-based inhibitors of HIV-1
protease. Annu. ReV. Biochem. 1993, 62, 543–585.
(2) Erickson, J. W.; Burt, S. K. Structural mechanisms of HIV drug
resistance. Annu. ReV. Pharmacol. Toxicol. 1996, 36, 545–571.

860 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 4
Clemente et al.
diamino-1,6-diphenylhexane-3,4-diol. J. Med. Chem. 1993, 36, 320–
330.
(22) Tie, Y.; Boross, P. I.; Wang, Y.-F.; Gaddis, L.; Hussain, A. K.;
Leshchenko, S.; Ghosh, A. K.; Louis, J. M.; Harrison, R. W.; Weber,
I. T. High resolution crystal structures of HIV-1 protease with a potent
non-peptide inhibitor (UIC-94017) active against multi-drug-resistant
clinical strains. J. Mol. Biol. 2004, 338, 341–352.
(23) Clark, T.; McKervey, M. A. Saturated hydrocarbons. In ComprehensiVe
Organic Chemistry 1st ed.; Stoddart, J. F. Ed.; Pergamon Press Ltd.:
Oxford, 1979; Vol. 1, pp 59–60.
(24) Lee, T.; Le, V.-D.; Lim, D.; Lin, Y.-C.; Morris, G. M.; Wong, A. L.;
Olson, A. J.; Elder, J. H.; Wong, C.-H. Development of a new type
of protease inhibitors, efficacious against FIV and HIV variants. J. Am.
Chem. Soc. 1999, 121, 1145–1155.
(25) Priestle, J. P.; Fässler, A.; Rösel, J.; Tintelnot-Blomley, M.; Strop,
P.; Grütter, M. G. Comparative analysis of the X-ray structures of
HIV-1 and HIV-2 proteases in complex with CGP 53820, a novel
pseudosymmetric inhibitor. Structure 1995, 3, 381–389.
(26) Perola, E.; Charifson, P. S. Conformational analysis of drug like
molecules bound to proteins: an extensive study of ligand reorganiza-
tion upon binding. J. Med. Chem. 2004, 47, 2499–2510.
(27) Koshland, D. E. Jr. Application of a theory of enzyme specificity to
protein synthesis. Proc. Natl. Acad. Sci. U.S.A. 1958, 44, 98–104.
(28) Bosshard, H. R. Molecular recognition by induced fit: How fit is the
concept. News. Physiol. Sci. 2001, 16, 171–173.
(29) Hintze, F.; Hoppe, D. Enantioselective synthesis of (S)-1-methyldode-
cyl acetate, a pheromone of Drosophila mulleri, via (-)-sparteine-
assisted deprotonation of 1-dodecanol. Synthesis 1992, 1216–1218.
(30) Schwerdtfeger, J.; Kolczewski, S.; Weber, B.; Fröhlich, R.; Hoppe,
D. Stereoselective deprotonation of chiral and achiral 2-aminoalkyl
carbamates: Synthesis of optically active ꢁ-amino alcohols via 1-oxy-
substituted alkyllithium intermediates. Synthesis 1999, 1573–1592.
(31) Ido, E.; Han, H.; Kezdy, F. J.; Tang, J. Kinetic studies of human
immunodeficiency virus type 1 protease and its active-site hydrogen
bond mutant A28S. J. Biol. Chem. 1991, 266, 24359–24366.
(32) Goodenow, M. M.; Bloom, G.; Rose, S. L.; Pomeroy, S. M.; O’Brien,
P. O.; Perez, E. E.; Sleasman, J. W.; Dunn, B. M. Naturally ocurring
amino acid polymorphisms in human immunodeficiency virus type 1
(HIV-1) Gag p7 (NC) and the C-cleavage site impact Gag-Pol
processing by HIV-1 protease. Virology 2002, 292, 137–149.
(33) Wain-Hobson, S.; Sonigo, P.; Danos, O.; Cole, S.; Alizon, M.
Nucleotide sequence for the AIDS virus, LAV. Cell 1985, 40, 9–17.
(34) Clemente, J. C.; Hemrajani, R.; Blum, L. E.; Goodenow, M. M.; Dunn,
B. M. Secondary mutations M36I and A71V in the human immuno-
deficiency virus type 1 protease can provide an advantage for the
emergence of the primary mutation D30N. Biochemistry 2003, 42,
15029–15035.
(35) Dunn, B. M.; Gustchina, A.; Wlodawer, A.; Kay, J. Subsite preferences
of retroviral proteinases. Methods Enzymol. 1994, 241, 254–278.
(36) Bhatt, D.; Dunn, B. M. Chimeric aspartic proteinases and active site
binding. Bioorg. Chem. 2000, 28, 374–393.
(37) Otwinowski, Z.; Minor, W. Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol. 1997, 276, 307–326.
(38) Brünger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros,
P.; Grosse-Kunstleve, R. W.; Jiang, J.-S.; Kuszewski, J.; Nilges, M.;
Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L.
Crystallography and NMR system: A new software suite for macro-
molecular structure determination. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 1998, 54, 905–921.
(39) Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaard, M. Improved
methods for building protein models in electron density maps and the
location of errors in these models. Acta Crystallogr., Sect. A: Found.
Crystallogr. 1991, 47, 110–119.
(40) InsightII, Biopolymer and Soak are trademarks of Accelrys Inc., San
Diego, CA.
(41) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. CHARMM: A program for macromo-
lecular energy, minimization, and dynamics calculations. J. Comput.
Chem. 1983, 4, 187–217.
JM701170F