Investigation of the noncovalent binding mode of covalent proteasome inhibitors around the transition state by combined use of cyclopropylic strain-based conformational restriction and computational modeling

Article abstract of DOI:10.1021/jm400542h

To develop potent covalent inhibitors, the noncovalent interactions around the transition state to form covalent bonding should be optimized because the potency of the inhibitor can be depending on the energy of the transition state. Here, we report an efficient analysis of the noncovalent binding mode of a potent covalent proteasome inhibitor 3a around the transition state by a combined use of the chemical approach, i.e., the cyclopropylic strain-based conformational restriction, and the computational docking approach. Furthermore, we calculated the binding energy of a series of salinosporamide derivatives in the predicted noncovalent complex around the transition state with the simulation model of proteasome constructed in this study, which was well correlated to their pIC₅₀. Thus, the proposed docking methods to predict the noncovalent binding mode around the transition state of covalent inhibitors will be helpful toward the development of covalent inhibitors.

Full text of DOI:10.1021/jm400542h

Article
pubs.acs.org/jmc
Investigation of the Noncovalent Binding Mode of Covalent
Proteasome Inhibitors around the Transition State by Combined Use
of Cyclopropylic Strain-Based Conformational Restriction and
Computational Modeling
Shuhei Kawamura,† Yuka Unno,§ Motohiro Tanaka,∥ Takuma Sasaki,∥ Akihito Yamano,⊥
Takatsugu Hirokawa,# Tomoshi Kameda,# Akira Asai,§ Mitsuhiro Arisawa,† and Satoshi Shuto*,†,‡
‡
†Faculty of Pharmaceutical Sciences and Center for Research and Education on Drug Discovery, Hokkaido University, Kita-12,
Nishi-6, Kita-ku, Sapporo 060-0812, Japan
§Graduate School of Pharmaceutical Sciences, University of Shizuoka, Yada, Shizuoka 422-8526, Japan
∥School of Pharmacy, Aichi Gakuin University, 1-100 Kusumoto-cho, Chikusa-ku, Nagoya 464-8650, Japan
⊥Rigaku Corporation, X-ray Institute, 3-9-12 Matsubara-cho, Akishima, Tokyo 196-8666, Japan
#Computational Biology Research Center (CBRC), National Institute of Advanced Industrial Science and Technology (AIST),
2-4-7 Aomi, Koutou-ku, Tokyo 135-0064, Japan
S
* Supporting Information
ABSTRACT: To develop potent covalent inhibitors, the non-
covalent interactions around the transition state to form
covalent bonding should be optimized because the potency of
the inhibitor can be depending on the energy of the transition
state. Here, we report an efficient analysis of the noncovalent
binding mode of a potent covalent proteasome inhibitor 3a
around the transition state by a combined use of the chemical
approach, i.e., the cyclopropylic strain-based conformational
restriction, and the computational docking approach. Fur-
thermore, we calculated the binding energy of a series of salinosporamide derivatives in the predicted noncovalent complex
around the transition state with the simulation model of proteasome constructed in this study, which was well correlated to their
pIC50. Thus, the proposed docking methods to predict the noncovalent binding mode around the transition state of covalent
inhibitors will be helpful toward the development of covalent inhibitors.
X-ray structure of the complex might not be so helpful compared
with the design of noncovalent inhibitors because conformation
of the covalent inhibitors in the binding site can be significantly
changed along with the covalent bond formation. In such cases,
the noncovalent binding mode around the transition state cannot
be effectively predicted by the X-ray crystallographic analysis.
Thus, it is difficult to analyze the noncovalent binding mode
of covalent inhibitors around the transition state for the
optimization process of covalent inhibitors.
In recent years, proteasome inhibitors have been extensively
studied from the viewpoint of antitumor drug discovery.4
Because the systematic degradation of intracellular proteins by
proteasome is essential for cellular functions such as cell cycle
progression,5 signal transduction,6 and endoplasmic reticulum-
associated protein degradation (ERAD),7 proteasome inhibition
causes cell cycle arrest and induces apoptosis.4a,b In fact, the
proteasome inhibitors bortezomib and carfilzomib were
approved by FDA for the treatment of multiple myeloma,8 and
INTRODUCTION
■
Much attention has been focused on covalent inhibitors of
proteins due to their strong and prolonged inhibitory effects
based on the stable covalent bonding.1 Many covalent inhibitors
are useful as clinical drugs1 and also as tools for investigating
biological pathways.2 Even if an inhibitor binds covalently to
its target protein, it should first be recognized by its target via
noncovalent interactions to form a reversible noncovalent
complex (P·I). After that, it reacts with the reacting group of
the target protein via transition state (P···I) to form the covalent
complex (P−I)1a as shown in Figure 1.
Accordingly, the potency of covalent inhibitors can be signif-
icantly affected by their binding affinity for the target protein in
the noncovalent binding mode, especially around its transition
state to form covalent bonding. Therefore, to design optimized
covalent inhibitors, it is desirable to know the noncovalent
binding mode of the lead inhibitor around the transition state,
which can be the “bioactive conformation” of covalent inhibitors.
X-ray crystallographic structures of inhibitors in complex with
their targets are often effectively used for designing further active
inhibitors.3 For the design of covalent inhibitors, however, the
Received: April 15, 2013
Published: June 25, 2013
© 2013 American Chemical Society
5829
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 1. Inhibitory mechanism of covalent inhibitors.
Figure 2. Development of potent proteasome inhibitors using belactosin A as a lead by the stereochemical diversity-oriented strategy.
Figure 4. The structure and proteasome inhibitory activity of β-lactam
congener 3a′.
A11 and subsequent optimization studies (Figure 2).12 These
studies revealed that the two hydrophobic moieties (yellow in
Figure 2) on the left side of the molecules are important for the
high potency of 3a−5a, compared with the lead 2.
Figure 3. X-ray crystallographic analysis of 3a in complex with
proteasome. The structure of 3a was colored according to its B-factor
value, in which the areas with low B-factors are colored blue and the
areas with high B-factors are colored red.
We also analyzed the X-ray crystal structure of 3a in complex
with proteasome,12b but, the B-factors, which represent smearing
of atomic electron densities around their equilibrium positions
due to thermal motion and positional disorder,13 of 3a were
unexpectedly high; the average ligand B-factor was greater than
80 Å2, especially in the region containing the two hydrophobic
moieties (Figure 3). This suggests that, in the covalent complex,
the region containing the two hydrophobic moieties of 3a is
highly flexible and would not interact effectively with
proteasome, while the two moieties are clearly necessary for
high inhibitory activity. This contradictory result indicates that
the region containing two hydrophobic moieties effectively
interacts with proteasome to facilitate noncovalent complex
formation and/or covalent bond formation. Once the covalent
bond is formed, the conformation of 3a would be significantly
several proteasome inhibitors are currently in clinical trials.4b
Notably, all of these inhibitors bind covalently to proteasome,
showing the effectiveness of covalent inhibitors as anticancer
agents.
Belactosin A (1) is a naturally occurring tripeptide identified
as a proteasome inhibitor by Asai and co-workers.9 It inhibits
proteasome covalently by acylating the active site Thr residue via
ring-opening of its β-lactone moiety.10 We previously developed
highly potent proteasome inhibitors 3a−5a by the three-
dimensional structure−activity relationship (SAR) of belactosin
5830
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
should effectively interact with proteasome in its transition state
to facilitate covalent bond formation.
Therefore, to effectively optimize the noncovalent interactions
of 3a with proteasome by further structural modifications, the
analyses of the noncovalent binding mode of 3a with proteasome
around the transition state are necessary (Figure 5).
The organic chemistry-based approach by synthesizing con-
formationally restricted analogues of a lead is often effective for
investigating the bioactive conformation of the lead.14 On the
other hand, a computational approach by docking simulations is
also useful for investigating the noncovalent binding mode of
compounds.15 We thought that effective investigation of the
noncovalent binding mode of covalent inhibitors around the
transition state would be possible by combined use of the organic
chemistry-based conformational restriction approach and the
computational modeling.
Thus, in this report, we describe the design, synthesis, confor-
mational analysis, and pharmacological effect of cyclopropylic
strain-based conformationally restricted analogues 3b−5b and
3c−5c (Figure 6) and docking simulations of them and their
parent compounds 3a−5a to identify the noncovalent binding
mode of these covalent proteasome inhibitors around the transi-
tion state. We present a useful docking concept for simulation of
the noncovalent binding mode of covalent inhibitors around the
transition state. On the basis of this concept, we investigated
the correlation between the calculated binding energies of the
simulated noncovalent complexes and the actual inhibitory
effects of a series of covalent proteasome inhibitors to show that
the obtained noncovalent binding mode of covalent inhibitors
around the transition state can be well related to their inhibitory
potency.
Figure 5. Plausible binding mode change of 3a with proteasome before
and after covalent bond formation.
altered due to its strained β-lactone-opening, which would
weaken the interaction of the two hydrophobic moieties with
proteasome.
RESULTS AND DISCUSSION
■
Cyclopropylic Strain-Based Design of the Conforma-
tionally Restricted Analogues. Because of its small and rigid
ring structure, cyclopropane is effective for restricting the
conformation of a molecule without changing the chemical and
To further analyze the contribution of noncovalent interac-
tions in these inhibitors, the β-lactam congener 3a′ was synthe-
sized and evaluated its proteasome inhibitory potency (Figure 4).
Because the reactivity of β-lactam is much lower than the
corresponding β-lactone, covalent bond formation should be
significantly retarded in 3a′. In stark contrast to 3a, the β-lactam
congener 3a′ showed no proteasome inhibitory activity (IC50
>
10000 nM). This clearly suggests that, in these compounds, the
noncovalent complex prior to covalent bond formation is not
stable enough to exhibit proteasome inhibitory activity by itself
and proteasome inhibitory activity of belactosin derivatives is
predominantly owing to their covalent bond formation ability.
Thus, the region containing the two hydrophobic moieties of 3a
Figure 7. The cyclopropylic strain-based conformational restriction.
Figure 6. Previously reported proteasome inhibitors 3a−5a and their conformationally restricted analogues 3b−5b and 3c−5c.
5831
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 8. Plausible stable conformations of 3a−5a (a), 3b−5b (b), and 3c−5c (c).
Figure 10. NOE of conformationally restricted analogues 3b and 3c and
their parent compound 3a in CHCl3.
(Figure 8a). Importantly, in the preferable conformers A (anti)
and B (syn), the positioning of the side chain containing two
hydrophobic moieties (X) that are essential for the strong
binding to proteasome is significantly different each other, so that
one of the two conformers is thought to be the bioactive form in
the noncovalent complex around the transition state.
To analyze the bioactive conformation of 3a−5a in the non-
covalent complex around the transition state, we designed the
C1′-methyl-substituted derivatives 3b−5b and 3c−5c as con-
formationally restricted analogues of 3a−5a (Figure 6).
Depending on the configuration at the C1′ position,
conformation of the compounds can be restricted due to the
cyclopropylic strain between the introduced methyl group and
the cis-oriented amide moiety; the syn-conformer would be stable
in 3b−5b (1′R, Figure 8b); conversely, the anti-conformer would
be stable in 3c−5c (1′S, Figure 8c). Pharmacological evaluations
of these conformationally restricted analogues would allow us
to clarify the bioactive conformation because the analogues
restricted in the bioactive conformation would be active as their
Figure 9. Stable conformations of 3a and its conformationally restricted
analogues 3b and 3c obtained by calculations: orange, the region con-
taining two hydrophobic moieties; green, the β-lactone moiety.
physical properties of the lead compound.16 A characteristic
structural feature of cyclopropane is that cis-oriented adjacent
substituents on the ring exert significant mutual steric repulsion
because they are fixed in the eclipsed orientation, which we
previously termed “cyclopropylic strain”.17 Consequently,
conformation of the substituents on a cyclopropane can be
restricted so that the steric repulsion due to the strain is minimal,
as indicated in Figure 7.
In compounds 3a−5a, bond rotation between the cyclo-
propane (C1) and its adjacent carbon (C1′) would be restricted
by the cyclopropylic strain. Thus, the two conformers A (anti, the
cyclopropane ring “down”/the side chain (X) “up”) and B (syn,
the cyclopropane ring “down”/the side chain (X) “down”) would
be preferable to conformer C due to the significant steric repul-
sion with the adjacent cis-oriented amide moiety in conformer C
5832
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 11. X-ray crystal structure of 3c.
Figure 12. The concept of the docking simulation of covalent inhibitors
performed by us.
parent compounds, while the analogues restricted in the different
conformation would be not.
The conformations of 3a−c were analyzed by molecular
mechanics calculations with MacroModel 9.9 (Schrodinger,
the conformations predicted by molecular mechanics calcu-
lations and that the cyclopropylic strain-based conformational
restriction strategy seems to work effectively in these com-
pounds, as we expected. Therefore, pharmacological evaluations
of these compounds would allow us to identify the bioactive
conformation of 3a−5a in the noncovalent complex around the
transition state.
Pharmacological Effects of the Conformationally
Restricted Analogues. The inhibitory effects of synthesized
conformationally restricted analogues 3b−5b and 3c−5c on the
chymotrypsin-like (ChT-L) activity of purified human 20S
proteasome were investigated using succinyl-Leu-Leu-Val-Tyr-4-
methylcoumaryl-7-amide as a substrate (Table 1). These con-
formationally restricted analogues showed proteasome inhib-
itory activity weaker than that of their parent compounds without
a methyl group at C1′. In all the C1′-stereoisomeric pairs (3b/3c,
4b/4c, and 5b/5c), however, the 1′R-isomer is significantly more
potent than the corresponding 1′S-isomer. Therefore, the bio-
active conformation of 3a−5a in the noncovalent complex
around the transition state should be the syn-form. The decreased
activity of the 1′R-isomers compared with their parent
compounds might be explained by steric repulsion between the
introduced methyl group and proteasome.
The cell growth inhibitory effects of 3a and its conformation-
ally restricted analogues 3b and 3c against several tumor cells
were also investigated. Cell growth inhibitory activity of 3c,
whose conformation is restricted in the anti-form, was lower
than that of the parent compound 3a, consistent with its un-
ambiguously lower proteasome inhibitory effect than 3a. How-
ever, the cell growth inhibitory activity of 3b, whose confor-
mation is restricted in the syn-form, is similar or even higher than
that of its parent compound 3a in all cell lines examined despite
its lower proteasome inhibitory activity compared with 3a.
To investigate these contradictory results, we focused on the
stability of 3a−3c in aqueous medium. Because these com-
pounds were almost insoluble in water, we removed their Cbz
group to obtain the water-soluble analogues 6a−6c, respectively,
for the stability evaluations. The compounds 6a−6c were in-
cubated in 0.1 M TEAA buffer (pH 7.4) or human AB serum at
̈
LLC). In the conformational analysis of 3a, two types of
particularly stable conformers were obtained, which correspond
to the anti- and the syn-conformers in Figure 8a, respectively
(Figure 9a), while the anti-conformer is slightly more stable than
the syn-conformer (ΔE = 1.2 kcal/mol). In these two calculated
conformers, positioning of the side chain containing two
hydrophobic moieties (orange) relative to its β-lactone moiety
(green) differs significantly. On the other hand, in the confor-
mational analysis of 3b and 3c, the most stable structure was
calculated as a syn-conformer for 3b and an anti-conformer for
3c, as expected (Figure 9b). The energy difference between the
two conformers was rather large, which were 2.4 kcal/mol for 3b
and 7.1 kcal/mol for 3c. The calculations of 4a−c and 5a−c were
also carried out, and the results were similar to those for 3a−c
(see Supporting Information).
Thus, conformational analysis by molecular mechanics
calculations supported our molecular design of the conforma-
tionally restricted analogues based on the cyclopropylic strain.
We therefore synthesized these conformationally restricted anal-
ogues to analyze the bioactive conformation of their parent com-
pounds in the noncovalent complex around the transition state.
Conformational Analysis of the Conformationally
Restricted Analogues. We investigated the stable conforma-
tions of the synthesized conformationally restricted 1′R and 1′S-
methyl derivatives 3b and 3c and their parent compound 3a in
CDCl3 by NOE experiments (Figure 10). Irradiation of Hd of the
1′R isomer 3b gave NOE at Ha, suggesting that it is stable in its
syn-conformation. On the other hand, irradiation of Hb and Hc of
the 1′S isomer 3c gave NOEs at Ha, suggesting that it is stable in
its anti-conformation. Furthermore, irradiation of Ha of the
parent compound 3a, without methyl group, gave NOEs at both
Hb and Hc, suggesting that it might be rather stable in its anti-
conformation.
We successfully analyzed the X-ray crystal structure of 3c,
which unambiguously showed its anti-conformation in the solid
state, as expected (Figure 11).
These experimental results suggested that the conformation-
ally restricted analogues 3b−5b and 3c−5c are actually stable in
5833
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Table 1. Proteasome and Cell Growth Inhibitory Effects of Conformationally Restricted Analogues 3b−5b and 3c−5c and their
parent compounds 3a−5a
a
Based on three experiments.
37 °C, and the time courses were analyzed by HPLC to obtain
the half-life (t1/2) as summarized in Table 2. Interestingly, both
chemical and biological stabilities of the conformationally restricted
analogues 6b and 6c are superior to those of their parent com-
pound 6a. Thus, although the proteasome inhibitory activity of
conformationally restricted analogue 3b was lower than its parent
compound 3a, the increased stability of 3b might make it a more
effective cell growth inhibitor than 3a. The decreased structural
flexibility of the compounds by the conformational restriction
might be related to the increased stability.
effectively toward designing compounds for further optimization.
In the case of belactosin derivatives, the strained β-lactone
moiety reacts with the hydroxyl group of the N-terminal Thr
of the proteasome, which would lead to a significant conforma-
tional change of the compound via its β-lactone ring-opening.10
Consequently, although the X-ray crystal structure of 3a
covalently complexed with proteasome is available, its non-
covalent binding mode around the transition state could not be
effectively predicted from the X-ray analysis. To investigate the
noncovalent binding mode of 3a around the transition state, we
planned to perform docking simulations of 3a and its con-
formationally restricted analogues 3b and 3c, according to the
scheme shown in Figure 12.
Docking Study of 3a−3c. As described above, the
noncovalent binding mode of the covalent inhibitors, especially
around the transition state to form a covalent bond, can be used
5834
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
structure of the covalent complex (A in Figure 12), it should only
produce noncovalent binding mode far from its transition state
(C in Figure 12). Thus, to predict the noncovalent binding mode
of 3a around the transition state, we devised a docking method
through steps B−F, depicted in Figure 12. First, the reacting
group of the protein is removed to avoid steric repulsion to
simulate the noncovalent complex around the transition state
(B−D). Then, the plausible position of the ligand’s reacting
group (orange closed circle in Figure 12) in the noncovalent
complex around the transition state is hypothesized based on the
X-ray crystal structure of the covalent complex (A), which can be
effectively used to restrict the positioning of ligand reacting
group in docking simulations (D−E). The constructed model
(E) consists of the trimmed protein with constraint on the ligand
reacting group position would be useful to simulate the non-
covalent binding mode around the transition state (E−F).
To perform the docking simulations of belactosin derivatives
according to the scheme, reasonable positioning of the react-
ing group β-lactone is needed. However, as mentioned above,
belactosin derivatives react with the proteasome Thr residue via
its strained β-lactone ring-opening, which would induce
significant conformational change of the belactosin derivatives10
(Figure 13a). Accordingly, positioning of the β-lactone moiety in
the noncovalent complex around the transition state is difficult to
predict exactly based on the covalent complex structure. Thus, in
the docking simulation of belactosin derivatives, construction of
the model structure used for docking simulation (Figure 12E)
based on the X-ray crystal structure (Figure 12A) is problematic.
To resolve the problem, we focused on the X-ray crystal
structure of fluorosalinosporamide A covalently complexed with
the proteasome.18 Salinosporamide A and its derivatives are
Table 2. Half-life (t1/2) of 6a−6c in 0.1 M TEAA Buffer
(pH 7.4) or Human AB Serum at 37 °C Investigated by HPLC
Analysis
t1/2
compd
no.
0.1 M TEAA buffer human AB serum
R2
conformation
(h)
(min)
6a
6b
6c
H
syn/anti
syn
10
14
21
2.3
4.2
4.7
R-CH3
S-CH3
anti
In the noncovalently interacting state of a covalent inhibitor
and its target protein around the transition state to form a
covalent bond, the two atoms to react each other in the inhibitor
and the protein should locate closer than the sum of their van der
Waals radii. Because docking programs simulate the “non-
covalent interaction” (P·I in Figure 1) between the ligand and
protein, they cannot simulate the binding mode including such
quite close interactions around the reaction transition state (P···I
in Figure 1). Therefore, if we perform docking simulations of
covalent inhibitors as in the case of noncovalent inhibitors using
the protein structure (B in Figure 12), in which the covalently
bound inhibitor is simply removed from the X-ray crystal
Figure 13. Inhibitory mechanism of 3a (a), salinosporamide A (b), and fluorosalinosporamide A (c).
5835
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
covalent inhibitors of proteasome that have a β-lactone ring as a
reacting group like belactosin derivatives,19 so that the β-lactone
ring of salinosporamide seems to occupy a space similar to that
occupied by belactosin derivatives in their noncovalent binding
state around the transition state. Importantly, in comparison
with belactosin derivatives, the conformational change of the
salinosporamide derivatives induced by the β-lactone ring-
opening should not be so significant because the β-lactone of
salinosporamide A derivatives is fused with a γ-lactam to restrict
the conformational change of the β-lactone region by the ring-
opening19c (Figure 13b,c). In the case of salinosporamide A, after
the β-lactone ring-opening reaction with the proteasome Thr
residue, the resulting hydroxyl group substitutes the chloride on
its chloroethyl side chain to form the five-membered ether
ring,19c which would further differentiate its structure from that
in the transition state (Figure 13b). Fortunately, in the case of
fluorosalinosporamide A, having a less reactive fluoroethyl group
instead of a chloroethyl group, its X-ray crystal structure in
complex with the proteasome without formation of the ether ring
was analyzed.18 In the structure, conformations of both the ligand
and the proteasome would be rather analogous to those in the
noncovalent binding state (Figure 13c*). Therefore, we thought
that the structure of fluorosalinosporamide A in complex with
proteasome (Figure 13c*) would be effectively used to construct
the model structure (corresponding to Figure 12E) for the
docking simulations to investigate the noncovalent complex
around the transition state.
Thus, as shown in Figure 14, we modified the X-ray crystallo-
graphic structure of fluorosalinosporamide A in complex with
proteasome (Figure 14a)18 to construct the model structure used
for docking simulation of 3a−3c. First, the ester bond between
the fluorosalinosporamide A and the Thr was cleaved, and then,
the side-chain of Thr was trimmed to remove the steric repul-
sion between the Thr side-chain and the β-lactone of 3a−3c
(Figure 14b).20 Next, we reconstructed the β-lactone structure
by ligating the hydroxyl oxygen and the carbonyl carbon, which
was minimized by molecular mechanics calculation to give the
putative noncovalent binding mode of the fluorosalinosporamide
A around the transition state (Figure 14c). From the recon-
structed structure of fluorosalinosporamide A, moieties other
than the β-lactone ring were removed to obtain the model
structure used for docking simulation, which consists of the
trimmed proteasome structure and the β-lactone ring used as a
core structure to restrict the β-lactone positioning of the ligand in
docking simulations (Figure 14d). Thus, in this model, it is
hypothesized that the reacting β-lactone is located at the position
near to that in the transition state.
Using the model, we carried out the docking simulation of
3a−3c to predict their noncovalent binding state around their
transition states, and the results are shown in Figure 15. The
predicted noncovalent binding mode of 3a was the syn-form
(Figure 15a) in accord with the experimental results of 3a−3c
described above, suggesting that the docking simulation is
reliable. As expected, the predicted noncovalent binding mode of
3a around the transition state was significantly different from the
X-ray analyzed covalent binding mode (Figure 3), especially in
the region containing the two hydrophobic moieties essential
for its strong proteasome inhibition. Furthermore, the two
aromatic groups and proteasome surface were very close to be
accommodated precisely in the binding site, compared with
those in the X-ray analyzed covalent complex, which further
supports our hypothesis that the two hydrophobic moieties
effectively interact with proteasome in the noncovalent complex
Figure 14. Modification of the X-ray crystallographic structure of the
fluorosalinosporamide A in complex with proteasome.
around the transition state rather than in the covalent complex, as
shown in Figure 5.
The predicted noncovalent binding mode of 3b (Figure 15b)
around the transition state was also the syn-form, and the region
containing the two hydrophobic moieties interacts with
proteasome almost the same as in 3a (Figure 15d). On the
other hand, in the docking simulation of 3c (Figure 15c), none of
plausible binding modes similar to that of 3a was obtained. These
docking results of 3b and 3c are consistent with their proteasome
inhibitory activity, which again indicates that the docking simula-
tion was performed properly.
Correlation between Calculated Energy of the Non-
covalent Complex and Inhibitory Activity. As mentioned
above, it is important to know the noncovalent binding mode
around the transition state for the effective optimization of covalent
inhibitors, because the inhibitory activity of covalent inhibitors can
be related to the binding energy of the noncovalent complex around
the transition state. On the basis of this hypothesis, we performed
docking simulations of a series of salinosporamide A derivatives
5836
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 15. (a−c) The noncovalent binding mode of 3a−3c around the transition state predicted by docking simulations: gray tube, predicted binding
mode of each compound; green wire, structure of 3a in covalent complex analyzed by X-ray crystallography. (d) Superimposed predicted structures of
3a (orange tube) and 3b (green tube).
using the proteasome model constructed in this study and calculated
the binding energy of the predicted noncovalent complexes around
the transition state by Prime MM-GBSA.21 In this trial, 11
salinosporamide A derivatives, for which the IC50 values for yeast
20S proteasome were reported previously,22 were used, and the
correlation between their pIC50 and the calculated Prime MM-
GBSA ΔG bind was investigated.
As shown in Figure 16a, calculated binding energies of the
predicted noncovalent complexes around the transition state are
well correlated to their pIC50 values (r = 0.82). Because each of
these derivatives has different hydrophobic residues (see
Supporting Information Figure S6), we further investigated the
correlation between clogP and pIC50 of them (Figure 16b). How-
ever, no correlation was observed between them (r = 0.096),
5837
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 16. Calculated prime MM-GBSA ΔG bind vs pIC50 (a) and clogP vs pIC50 (b) of salinosporamide A derivatives.
Scheme 1. Synthetic Plan of 3b−5b
which clearly suggests that the calculated binding energies in the
predicted noncovalent complexes around the transition state
does not reflect the hydrophobicity of the inhibitors but effec-
tively reflects specific hydrophobic interactions between the
protein and inhibitors around the transition state. To our
knowledge, this is the first demonstration of the good correlation
between the inhibitory activities of covalent inhibitors and the
calculated binding energies of predicted noncovalent complexes
around the transition state by docking simulations.
Chemistry. The synthetic plan for 3b−5b is summarized in
Scheme 1. In the synthesis, construction of the sequential tertiary
chiral carbons (C1′ and C2′) is a key. The C2′ chiral center is
thought to be constructed by Grignard reaction with the
sulfinylimine 10b, as we reported previously for the synthesis of
3a−5a.12b The sulfinylimine 10b is prepared from the conforma-
tionally restricted chiral cyclopropane unit 9b. The C1′ chiral
center of 9b seemed to be constructed by the stereoselective
methylation by chiral oxazolidinone method23 using substrate
8b. Starting from chiral cyclopropane unit 7, developed by us as a
chiral cyclopropane unit for the synthesis of stereochemically
diverse cyclopropane compounds,24 substrate 8b is prepared.
Similarly, their diastereomers 3c−5c are prepared (see Supporting
Information about the synthesis of 3c−5c).
Pinnick oxidation afforded the corresponding carboxylic acid 13.
Condensation of 13 and (4R)-4-benzyl-2-oxazolidinone23 with
the mixed acid anhydride method was carried out to form 8b to
regulate the stereochemistry of the following methylation of the
α-position (C1′) to the carbonyl group. Thus, 8b was treated
with CH3I/NaHMDS in THF at −78 °C to stereoselectively
afford the desired methylated product 9b (dr 97:3). The
undesired diastereomer was removed at the purification step of
15b to afford it as a single diastereomer. The chiral auxiliary was
removed reductively with DIBAL to yield the corresponding
25
aldehyde 14b, condensation of which with (S)-t-BuSONH2
gave the conformationally restricted chiral cyclopropane unit
10b. The absolute configuration of the introduced methyl
group was determined by the PGME method26 (see Supporting
Information).
Grignard reaction of 10b with phenethylmagnesium chloride
afforded the alkylated product diastereoselectively.25 The
absolute configuration of the introduced phenethyl group was
determined by the modified Mosher’s method27 (see Supporting
Information). The t-butylsulfinyl group and TBDPS group were
then removed under acidic conditions in methanol, and the
resulting amino group was protected with an Fmoc group to
afford 15b. Successive oxidations of 15b under Dess−Martin and
Pinnick oxidation conditions gave the corresponding carboxylic
acid, and subsequent treatment of it with DPPA gave the cor-
responding acyl azide, which was heated in t-BuOH to form the
Curtius rearrangement product 12b as a key intermediate.
Synthesis of the unit 10b and the key intermediate 12b are
shown in Scheme 2. Cyclopropane unit 7 was subjected to a
Wittig reaction with MeOCH2PPh3Cl, and subsequent hydrol-
ysis of the product gave its homologous aldehyde, of which
5838
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of the Conformationally Restricted Chiral Cyclopropane Unit 10b and the Key Intermediate 12b
a
Reagents and conditions: (a) MeOCH2PPh3Cl, NaHMDS, THF, 0 °C; (b) HCl, THF/H2O; (c) Pinnick ox, 90% (3 steps from 7); (d) PivCl,
Et3N, CH2Cl2, 0 °C; (e) (4R)-4-benzyl-2-oxazolidinone, n-BuLi, THF, −78 °C, 83% (2 steps from 13); (f) CH3I, NaHMDS, THF, −78 °C, 83% (dr
97:3); (g) DIBAL, THF, −78 °C, 69%; (h) (S)-t-BuSONH2, CuSO4, CH2Cl2, 96%; (i) PhCH2CH2MgCl, CH2Cl2; (j) HCl, AcOEt/MeOH; (k)
FmocOSu, Na2CO3, THF/H2O, 66% (3 steps from 10b); (l) Dess−Martin ox; (m) Pinnick ox; (n) DPPA, Et3N, CH2Cl2, 0 °C to rt; (o) t-BuOH,
reflux, 72% (4 steps from 15b).
a
Scheme 3. Synthesis of 3b−5b
a
Reagents and conditions: (a) K2CO3, MeOH; (b) Cbz-Ala-OPiv, Et3N, CH2Cl2, 0 °C to rt, 100% (2 steps from 12b); (c) TFA/CH2Cl2; (d) PivCl,
Et3N, CH2Cl2, 0 °C to rt, (3b, two steps 82% from 16b; 4b, two steps 91% from 18b); (e) Pd/C, H2, TFA/THF, 0 °C; (f) 2-naphthoyl chloride,
Et3N, CH2Cl2, 0 °C, 70% (2 steps from 3b); (g) 2-naphthoyl chloride, Et3N, CH2Cl2, 0 °C, 100% (2 steps from 12b).
The synthesis of 3b−5b is shown in Scheme 3. After removal
of the Fmoc group of 12b with K2CO3 in methanol, Cbz-Ala-OH
was condensed with the mixed anhydride method to yield 16b.
After removal of the Boc group of 16b with TFA in DCM, the
β-lactone unit 1728 was condensed by the mixed anhydride
method to yield 3b. The Cbz group of 3b was removed by
hydrogenolysis, and the product was treated with 2-naphthoyl
chloride to afford 5b. Compound 4b was prepared from 12b as
the preparation of 3b.
group adjacent to the imino moiety. In these reactions, the 1′R
substrate 10b gave 11b as a single isomer, while the 1′S substrate
10c gave 11c as a diastereomeric mixture in a ratio of 5:1.
The different stereochemical outcome between these substrates
might be explained by the reaction mechanism shown in
Figure 17. The Grignard reaction of the t-butylsulfinylimines
is thought to proceed through a six-membered transition
state due to coordination of the sulfinyl oxygen to the Mg2+,
in which the bulky t-butyl group is in the equatorial posi-
tion to determine the stereochemistry of the reaction (Ellman
model).25 When the substrate is the 1′S isomer 10c, the
methyl group adjacent to the imino moiety should be
During these synthetic studies, we observed that the
diastereoselectivity of the Grignard reactions of 10b and 10c
was significantly affected by the configuration of the C1′ methyl
5839
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
Figure 17. Plausible reaction mechanisms of diastereoselective Grignard reactions.
positioned on the reagent-accessible side of the imine plane in
the Ellman model, and therefore, its steric effect might lower
the stereoselectivity.
ACKNOWLEDGMENTS
■
This investigation was supported by Grant-in-Aids for Scientific
Research (21390028) from the Japan Society for the Promotion
of Science.
CONCLUSIONS
■
ABBREVIATIONS USED
■
We successfully analyzed the noncovalent binding mode of 3a
with proteasome around the transition state using both a
conformational restriction approach based on the cyclopropylic
strain and docking simulations. Although precise prediction of
the noncovalent binding mode of covalent inhibitors around the
transition state based on X-ray crystallographic analysis is often
unsuccessful, our findings suggest that the combined use of a
conformational restriction approach and docking simulations can
be effective to investigate the noncovalent binding mode of
covalent inhibitors around the transition state in various cases.
Furthermore, we calculated the binding energy of a series of
salinosporamide derivatives in the predicted noncovalent
complex around the transition state with the simulation model
of proteasome constructed in this study, and the results are well
correlated to the pIC50. These findings indicate that the
noncovalent binding mode around the transition state can be a
key in the binding between covalent inhibitors and their target
molecules. Thus, reliable docking methods to predict the
noncovalent binding mode of covalent inhibitors around the
transition state can be helpful toward the development of potent
covalent inhibitors, and an organic chemistry approach by the
conformational restriction is very effective to verify the
simulation results.
Boc, tert-butoxycarbonyl; Cbz, benzyloxycarbonyl; ChT-L,
chymotrypsin-like; DIBAL, diisobutylaluminium hydride;
DCM, dichloromethane; DPPA, diphenylphosphoryl azide;
ERAD, endoplasmic reticulum-associated protein degradation;
Fmoc, 9-fluorenylmethyloxycarbonyl; MM-GBSA, molecular
mechanics generalized Born surface area; PGME, phenylglycine
methyl ester; Piv, pivaloyl; TBDPS, tert-butyldiphenylsilyl;
TEAA, triethylammonium acetate
REFERENCES
■
(1) (a) Singh, J.; Petter, R. C.; Baillie, T. A.; Whitty, A. The resurgence
of covalent drugs. Nature Rev. Drug Discovery 2011, 10, 307−317.
(b) Smith, A. J.; Zhang, X.; Leach, A. G.; Houk, K. N. Beyond picomolar
affinities: quantitative aspects of noncovalent and covalent binding of
drugs to proteins. J. Med. Chem. 2009, 52, 225−233. (c) Potashman, M.
H.; Duggan, M. E. Covalent Modifiers: An Orthogonal Approach to
Drug Design. J. Med. Chem. 2009, 52, 1231−1246.
(2) Evans, M. J.; Cravatt, B. F. Mechanism-Based Profiling of Enzyme
Families. Chem. Rev. 2006, 106, 3279−3301.
(3) (a) Davis, A. M.; Teague, S. J.; Kleywegt, G. J. Application and
limitations of X-ray crystallographic data in structure-based ligand and
drug design. Angew. Chem., Int. Ed. Engl. 2003, 42, 2718−2736.
(b) Hajduk, P. J.; Greer, J. A decade of fragment-based drug design:
strategic advances and lessons learned. Nature Rev. Drug Discovery 2007,
6, 211−219. (c) Congreve, M.; Langmead, C. J.; Mason, J. S.; Marshall,
F. H. Progress in structure based drug design for G protein-coupled
receptors. J. Med. Chem. 2011, 54, 4283−4311.
ASSOCIATED CONTENT
■
S
* Supporting Information
(4) (a) Adams, J. The proteasome: a suitable antineoplastic target.
Nature Rev. Cancer 2004, 4, 349−360. (b) Kisselev, A. F.; van der
Linden, W. A.; Overkleeft, H. S. Proteasome inhibitors: an expanding
army attacking a unique target. Chem. Biol. 2012, 19, 99−115. (c) Dick,
L. R.; Fleming, P. E. Building on bortezomib: second-generation
proteasome inhibitors as anti-cancer therapy. Drug Discovery Today
2010, 15, 243−249. (d) Orlowski, R. Z.; Kuhn, D. J. Proteasome
inhibitors in cancer therapy: lessons from the first decade. Clin. Cancer
Res. 2008, 14, 1649−1657. (e) Schneekloth, J. S.; Crews, C. M. Natural
Product Inhibitors of the Ubiquitin−Proteasome Pathway. Curr. Drug
Targets 2011, 12, 1581−1594. (f) Myung, J.; Kim, K. B.; Crews, C. M.
The ubiquitin−proteasome pathway and proteasome inhibitors. Med.
Res. Rev. 2001, 21, 245−273.
Experimental details of synthesis, stability testing, biological
evaluations, computational simulations, and a table listing
combustion analysis data for target compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*Phone/Fax: +81-11-706-3769. E-mail: shu@pharm.hokudai.ac.
jp. Address: Faculty of Pharmaceutical Sciences, Hokkaido
University, Kita-12, Nishi-6, Kita-ku, Sapporo 060-0812, Japan.
Notes
(5) King, R. W.; Deshaies, R. J.; Peters, J. M.; Kirschner, M. W. How
proteolysis drives the cell cycle. Science 1996, 274, 1652−1659.
The authors declare no competing financial interest.
5840
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
(6) (a) Fuchs, S. Y. The role of ubiquitin−proteasome pathway in
oncogenic signaling. Cancer Biol. Ther. 2002, 1, 337−341. (b) Chen, J. J.;
Lin, F.; Qin, Z. H. The roles of the proteasome pathway in signal
transduction and neurodegenerative diseases. Neurosci. Bull. 2008, 24,
183−194.
(7) (a) Vembar, S. S.; Brodsky, J. L. One step at a time: endoplasmic
reticulum-associated degradation. Nature Rev. Mol. Cell Biol. 2008, 9,
944−957. (b) Ron, D.; Walter, P. Signal integration in the endoplasmic
reticulum unfolded protein response. Nature Rev. Mol. Cell Biol. 2007, 8,
519−529. (c) Hiller, M. M.; Finger, A.; Schweiger, M.; Wolf, D. H. ER
degradation of a misfolded luminal protein by the cytosolic ubiquitin−
proteasome pathway. Science 1996, 273, 1725−1728.
(8) (a) Paramore, A.; Frantz, S. Bortezomib. Nature Rev. Drug Discovery
2003, 2, 611−612. (b) Bross, P. F.; Kane, R.; Farrell, A. T.; Abraham, S.;
Benson, K.; Brower, M. E.; Bradley, S.; Gobburu, J. V.; Goheer, A.; Lee,
S. L.; Leighton, J.; Liang, C. Y.; Lostritto, R. T.; McGuinn, W. D.; Morse,
D. E.; Rahman, A.; Rosario, L. A.; Verbois, S. L.; Williams, G.; Wang, Y.
C.; Pazdur, R. Approval summary for bortezomib for injection in the
treatment of multiple myeloma. Clin. Cancer Res. 2004, 10, 3954−3964.
(c) Kane, R. C.; Farrell, A. T.; Sridhara, R.; Pazdur, R. United States
Food and Drug Administration approval summary: bortezomib for the
treatment of progressive multiple myeloma after one prior therapy. Clin.
Cancer Res. 2006, 12, 2955−2960. (d) Abraham, J. Carfilzomib and
bortezomib therapy in patients with multiple myeloma. Community
Oncol. 2012, 9, 278−282.
(9) (a) Asai, A.; Hasegawa, A.; Ochiai, K.; Yamashita, Y.; Mizukami, T.
Belactosin A, a novel antitumor antibiotic acting on cyclin/CDK
mediated cell cycle regulation, produced by Streptomyces sp. J. Antibiot.
2000, 53, 81−83. (b) Asai, A.; Tsujita, T.; Sharma, S. V.; Yamashita, Y.;
Akinaga, S.; Funakoshi, M.; Kobayashi, H.; Mizukami, T. A new
structural class of proteasome inhibitors identified by microbial
screening using yeast-based assay. Biochem. Pharmacol. 2004, 67,
227−234.
(10) Groll, M.; Larionov, O. V.; Huber, R.; De Meijere, A. Inhibitor-
binding mode of homobelactosin C to proteasomes: new insights into
class I MHC ligand generation. Proc. Natl. Acad. Sci. U. S. A. 2006, 103,
4576−4579.
(11) Yoshida, K.; Yamaguchi, K.; Sone, T.; Unno, Y.; Asai, A.;
Yokosawa, H.; Matsuda, A.; Arisawa, M.; Shuto, S. Synthesis of 2,3- and
3,4-methanoamino acid equivalents with stereochemical diversity and
their conversion into the tripeptide proteasome inhibitor belactosin a
and its highly potent cis-cyclopropane stereoisomer. Org. Lett. 2008, 10,
3571−3574.
(12) (a) Yoshida, K.; Yamaguchi, K.; Mizuno, A.; Unno, Y.; Asai, A.;
Sone, T.; Yokosawa, H.; Matsuda, A.; Arisawa, M.; Shuto, S. Three-
dimensional structure−activity relationship study of belactosin A and its
stereo- and regioisomers: development of potent proteasome inhibitors
by a stereochemical diversity-oriented strategy. Org. Biomol. Chem.
2009, 7, 1868−1877. (b) Kawamura, S.; Unno, Y.; List, A.; Mizuno, A.;
Tanaka, M.; Sasaki, T.; Arisawa, M.; Asai, A.; Groll, M.; Shuto, S. Potent
Proteasome Inhibitors Derived from the Unnatural cis-Cyclopropane
Isomer of Belactosin A: Synthesis, Biological Activity, and Mode of
Action. J. Med. Chem. 2013, 56, 3689−3700.
(13) (a) Parthasarathy, S.; Murthy, M. Protein thermal stability:
insights from atomic displacement parameters (B values). Protein Eng.
2000, 13, 9−13. (b) Yuan, Z.; Zhao, J.; Wang, Z.-X. Flexibility analysis of
enzyme active sites by crystallographic temperature factors. Protein Eng.
2003, 16, 109−114. (c) Radivojac, P.; Obradovic, Z.; Smith, D. K.; Zhu,
G.; Vucetic, S.; Brown, C. J.; Lawson, J. D.; Dunker, A. K. Protein
flexibility and intrinsic disorder. Protein Sci. 2009, 13, 71−80.
(14) (a) Kozikowski, A. P. Drug Design for Neuroscience; Raven Press:
New York, 1993; (b) Silverman, R. B. The Organic Chemistry of Drug
Design and Drug Action; Academic Press: New York, 2004. (c) Wermuth,
C. G. The Practice of Medicinal Chemistry; Academic Press: New York,
2008.
R. D.; Jewsbury, P. J.; Essex, J. W. A review of protein−small molecule
docking methods. J. Comput.-Aided Mol. Des. 2002, 16, 151−166.
(16) (a) Armstrong, P. D.; Cannon, J. G.; Long, J. P. Conformationally
Rigid Analogues of Acetylcholine. Nature 1968, 220, 65−66. (b) Kazuta,
Y.; Hirano, K.; Natsume, K.; Yamada, S.; Kimura, R.; Matsumoto, S.;
Furuichi, K.; Matsuda, A.; Shuto, S. Cyclopropane-based conforma-
tional restriction of histamine. (1S,2S)-2-(2-aminoethyl)-1-(1H-imida-
zol-4-yl)cyclopropane, a highly selective agonist for the histamine H3
receptor, having a cis-cyclopropane structure. J. Med. Chem. 2003, 46,
1980−1988. (c) Watanabe, M.; Kazuta, Y.; Hayashi, H.; Yamada, S.;
Matsuda, A.; Shuto, S. Stereochemical diversity-oriented conformational
restriction strategy. Development of potent histamine H3 and/or H4
receptor antagonists with an imidazolylcyclopropane structure. J. Med.
Chem. 2006, 49, 5587−5596. (d) Watanabe, M.; Kobayashi, T.;
Hirokawa, T.; Yoshida, A.; Ito, Y.; Yamada, S.; Orimoto, N.; Yamasaki,
Y.; Arisawa, M.; Shuto, S. Cyclopropane-based stereochemical diversity-
oriented conformational restriction strategy: histamine H3 and/or H4
receptor ligands with the 2,3-methanobutane backbone. Org. Biomol.
Chem. 2012, 10, 736−745.
(17) (a) Shuto, S.; Ono, S.; Hase, Y.; Kamiyama, N.; Takada, H.;
Yamasihita, K.; Matsuda, A. Conformational restriction by repulsion
between adjacent substituents on a cyclopropane ring: Design and
enantioselective synthesis of 1-phenyl-2-(1-aminoalkyl)-N,N-diethylcy-
clopropanecarboxamide as potent NMDA receptor antagonists. J. Org.
Chem. 1996, 61, 915−923. (b) Shuto, S.; Ono, S.; Hase, Y.; Ueno, Y.;
Noguchi, T.; Yoshii, K.; Matsuda, A. Synthesis and biological activity of
conformationally restricted analogs of milnacipran: (1S,2R)-1-phenyl-2-
[(S)-1-aminopropyl]-N,N-diethylcyclopropanecarboxamide, an effi-
cient noncompetitive N-methyl-D-aspartic acid receptor antagonist. J.
Med. Chem. 1996, 39, 4844−4852. (c) Shuto, S.; Ono, S.; Imoto, H.;
Yoshii, K.; Matsuda, A. Synthesis and biological activity of conforma-
tionally restricted analogues of milnacipran: (1S,2R)-1-phenyl-2-[(R)-1-
amino-2-propynyl]-N,N-diethylcyclopropanecarboxamide is a novel
class of NMDA receptor channel blocker. J. Med. Chem. 1998, 41,
3507−3514. (d) Ono, S.; Ogawa, K.; Yamashita, K.; Yamamoto, T.;
Kazuta, Y.; Matsuda, A.; Shuto, S. Conformational analysis of the
NMDA receptor antagonist (1S,2R)-1-phenyl-2-[(S)-1-aminopropyl]-
N,N-diethylcyclopropanecarboxamide (PPDC) designed by a novel
conformational restriction method based on the structural feature of
cyclopropane ring. Chem. Pharm. Bull. 2002, 50, 966−968. (e) Yama-
guchi, K.; Kazuta, Y.; Hirano, K.; Yamada, S.; Matsuda, A.; Shuto, S.
Synthesis of 1-arylpiperazyl-2-phenylcyclopropanes designed as anti-
dopaminergic agents: cyclopropane-based conformationally restricted
analogs of haloperidol. Bioorg. Med. Chem. 2008, 16, 8875−8881.
(f) Watanabe, M.; Hirokawa, T.; Kobayashi, T.; Yoshida, A.; Ito, Y.;
Yamada, S.; Orimoto, N.; Yamasaki, Y.; Arisawa, M.; Shuto, S.
Investigation of the bioactive conformation of histamine H3 receptor
antagonists by the cyclopropylic strain-based conformational restriction
strategy. J. Med. Chem. 2010, 53, 3585−3593.
(18) Groll, M.; McArthur, K. A.; Macherla, V. R.; Manam, R. R.; Potts,
B. C. Snapshots of the fluorosalinosporamide/20S complex offer
mechanistic insights for fine tuning proteasome inhibition. J. Med. Chem.
2009, 52, 5420−5428.
(19) (a) Feling, R. H.; Buchanan, G. O.; Mincer, T. J.; Kauffman, C. A.;
Jensen, P. R.; Fenical, W. Salinosporamide A: a highly cytotoxic
proteasome inhibitor from a novel microbial source, a marine bacterium
of the new genus salinospora. Angew. Chem., Int. Ed. Engl. 2003, 42,
355−357. (b) Macherla, V. R.; Mitchell, S. S.; Manam, R. R.; Reed, K. A.;
Chao, T. H.; Nicholson, B.; Deyanat-Yazdi, G.; Mai, B.; Jensen, P. R.;
Fenical, W. F.; Neuteboom, S. T.; Lam, K. S.; Palladino, M. A.; Potts, B.
C. Structure−activity relationship studies of salinosporamide A (NPI-
0052), a novel marine derived proteasome inhibitor. J. Med. Chem. 2005,
48, 3684−3687. (c) Groll, M.; Huber, R.; Potts, B. C. M. Crystal
structures of salinosporamide A (NPI-0052) and B (NPI-0047) in
complex with the 20S proteasome reveal important consequences of
beta-lactone ring opening and a mechanism for irreversible binding. J.
Am. Chem. Soc. 2006, 128, 5136−5141.
(15) (a) Kitchen, D. B.; Decornez, H.; Furr, J. R.; Bajorath, J. Docking
and scoring in virtual screening for drug discovery: methods and
applications. Nature Rev. Drug Discovery 2004, 3, 935−949. (b) Taylor,
(20) In the docking simulation using the native proteasome structure
with positional constraint of β-lactone moiety, no pose was obtained,
5841
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842

Journal of Medicinal Chemistry
Article
probably due to significant steric repulsion between β-lactone and Thr
side-chain.
(21) (a) Guimaraes, C. MM-GB/SA Rescoring of Docking Poses.
̃
Methods Mol. Biol. 2012, 819, 255. (b) Guimaraes, C. R. W.; Cardozo, M.
̃
MM-GB/SA Rescoring of Docking Poses in Structure-Based Lead
Optimization. J. Chem. Inf. Model. 2008, 48, 958−970. (c) Lyne, P. D.;
Lamb, M. L.; Saeh, J. C. Accurate Prediction of the Relative Potencies of
Members of a Series of Kinase Inhibitors Using Molecular Docking and
MM-GBSA Scoring. J. Med. Chem. 2006, 49, 4805−4808.
(22) (a) Nett, M.; Gulder, T. A. M.; Kale, A. J.; Hughes, C. C.; Moore,
B. S. Function-oriented biosynthesis of β-lactone proteasome inhibitors
in Salinispora tropica. J. Med. Chem. 2009, 52, 6163−6167. (b) Manam,
R. R.; McArthur, K. A.; Chao, T. H.; Weiss, J.; Ali, J. A.; Palombella, V. J.;
Groll, M.; Lloyd, G. K.; Palladino, M. A.; Neuteboom, S. T. C. Leaving
groups prolong the duration of 20S proteasome inhibition and enhance
the potency of salinosporamides. J. Med. Chem. 2008, 51, 6711−6724.
(23) Evans, D. A.; Wu, L. D.; Wiener, J. J. M.; Johnson, J. S.; Ripin, D.
H. B.; Tedrow, J. S. A General Method for the Synthesis of
Enantiomerically Pure β-Substituted, β-Amino Acids through α-
Substituted Succinic Acid Derivatives. J. Org. Chem. 1999, 64, 6411−
6417.
(24) Kazuta, Y.; Matsuda, A.; Shuto, S. Development of versatile cis-
and trans-dicarbon-substituted chiral cyclopropane units: synthesis of
(1S,2R)- and (1R,2R)-2-aminomethyl-1-(1H-imidazol-4-yl)-
cyclopropanes and their enantiomers as conformationally restricted
analogues of histamine. J. Org. Chem. 2002, 67, 1669−1677.
(25) Cogan, D. A.; Liu, G.; Ellman, J. Asymmetric synthesis of chiral
amines by highly diastereoselective 1,2-additions of organometallic
reagents to N-tert-butanesulfinyl imines. Tetrahedron 1999, 55, 8883−
8904.
(26) Yabuuchi, T.; Kusumi, T. Phenylglycine methyl ester, a useful tool
for absolute configuration determination of various chiral carboxylic
acids. J. Org. Chem. 2000, 65, 397−404.
(27) (a) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. High-field
FT NMR application of Mosher’s method. The absolute configurations
of marine terpenoids. J. Am. Chem. Soc. 1991, 113, 4092−4096.
(b) Kusumi, T.; Fukushima, T.; Ohtani, I.; Kakisawa, H. Elucidation of
the absolute configurations of amino acids and amines by the modified
mosher’s method. Tetrahedron Lett. 1991, 32, 2939−2942.
(28) Armstrong, A.; Scutt, J. N. Total synthesis of (+)-belactosin A.
Chem. Commun. 2004, 510−511.
5842
dx.doi.org/10.1021/jm400542h | J. Med. Chem. 2013, 56, 5829−5842