Phosphoramidate derivatives of hydroxysteroids as inhibitors of prostate-specific membrane antigen

Article abstract of DOI:10.1016/j.bmcl.2007.10.096

Prostate-specific membrane antigen (PSMA) is a membrane-bound cell surface peptidase which is over-expressed in prostate cancer cells. The enzymatic activities of PSMA are understood but the role of the enzyme in prostate cancer remains conjectural. We previously confirmed the existence of a hydrophobic binding site remote from the enzyme's catalytic center. To explore the specificity and accommodation of this binding site, we prepared a series of six glutamate-containing phosphoramidate derivatives of various hydroxysteroids (1a-1f). The inhibitory potencies of the individual compounds of the series were comparable to a simple phenylalkyl analog (8), and in all cases IC₅₀ values were sub-micromolar. Molecular docking was used to develop a binding model for these inhibitors and to understand their relative inhibitory potencies against PSMA.

Full text of DOI:10.1016/j.bmcl.2007.10.096

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 281–284
Phosphoramidate derivatives of hydroxysteroids as inhibitors
of prostate-specific membrane antigen
Lisa Y. Wu,b Jacinda C. Do,a Marat Kazak,a Helen Page,a Yoko Toriyabe,a
Marc O. Andersona and Clifford E. Berkmanb,*
aDepartment of Chemistry & Biochemistry, San Francisco State University, San Francisco, CA 94132-4163, USA
bDepartment of Chemistry, Washington State University, Pullman, WA 99164-4630, USA
Received 21 July 2007; revised 24 October 2007; accepted 25 October 2007
Available online 30 October 2007
Abstract—Prostate-specific membrane antigen (PSMA) is a membrane-bound cell surface peptidase which is over-expressed in pros-
tate cancer cells. The enzymatic activities of PSMA are understood but the role of the enzyme in prostate cancer remains conjectural.
We previously confirmed the existence of a hydrophobic binding site remote from the enzyme’s catalytic center. To explore the spec-
ificity and accommodation of this binding site, we prepared a series of six glutamate-containing phosphoramidate derivatives of var-
ious hydroxysteroids (1a–1f). The inhibitory potencies of the individual compounds of the series were comparable to a simple
phenylalkyl analog (8), and in all cases IC50 values were sub-micromolar. Molecular docking was used to develop a binding model
for these inhibitors and to understand their relative inhibitory potencies against PSMA.
Ó 2007 Elsevier Ltd. All rights reserved.
A notable discovery in prostate cancer research has been
the identification of an over-expressed membrane-bound
cell surface protein on prostate cancer cells, namely,
prostate-specific membrane antigen (PSMA). PSMA,
also known as folate hydrolase I (FOLH1) and gluta-
mate carboxypeptidase II (GCPII),1,2 is a 750-amino
acid type II membrane glycoprotein3 and was discovered
during the development of the LNCaP cell line; one
which retains most of the known features of prostate
cancer.4
tance, PSMA is also reported to possess two
predominant, yet poorly understood enzymatic activi-
ties: the hydrolytic cleavage and liberation of glutamate
from c-glutamyl derivatives of folic acid12,13 and the
proteolysis of the neuropeptide N-acetylaspartylgluta-
mate (NAAG).1 With respect to its function, recent
studies suggest that PSMA plays a regulatory role in
angiogenesis.14
Recently, we identified the presence of a hydrophobic
binding site remote from the catalytic center of PSMA
using a series of phenylalkylphosphonamidate deriva-
tives of glutamic acid.15 Based upon ongoing substrate
and inhibitor-based studies in our laboratory, we have
determined that PSMA can accommodate a variety of
structural motifs remote from the catalytic center. The
focus of this work was to determine if PSMA could spe-
cifically accommodate steroidal motifs in auxiliary bind-
ing sites remote from the catalytic center. To this end,
we prepared a series of putative hydroxysteroid-contain-
ing phosphoramidate inhibitors of PSMA (Fig. 1) and
determined their inhibitory potency against purified
PSMA.
Although PSMA is primarily expressed in normal hu-
man prostate epithelium, the importance of this enzyme
lies with the fact that it is upregulated and strongly ex-
pressed in prostate cancer cells, including those of the
metastatic disease state.5 It has also been demonstrated
that PSMA expression is present in the endothelium of
tumor-associated neovasculature of multiple non-pros-
tatic solid malignancies,6 including metastatic renal car-
cinoma.7 As such, it is not surprising that PSMA has
attracted a great deal of attention as a target for immu-
notherapy.8–11 In addition to its immunological impor-
A general method (Scheme 1) was employed for the
preparation of hydroxysteroid phosphoramidate inhibi-
tors (Fig. 1). Briefly, bis(diisopropylamino)-chlorophos-
phine (2) was conjugated with benzyl alcohol to generate
Keywords: PSMA; Prostate-specific membrane antigen; Glutamate
carboxypeptidase II; Steroid; Molecular docking; Phosphoramidate.
*
Corresponding author. Tel.: +1 415 338 6495; fax: +1 415 405
0377; e-mail: cberkman@sfsu.edu
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.096

282
L. Y. Wu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 281–284
Figure 1. Structures and inhibitory potencies of steroid-derived phosphoramidate inhibitors (1a–1f) of PSMA and non-steroidal reference compound
(8). Standard deviation for IC50 values in parentheses.
Scheme 1. Synthesis of hydroxysteroid-derived phosphoramidate inhibitors of PSMA. Reagents and conditions: (a) benzyl alcohol (3.0 equiv), Et3N
(3.0 equiv), CH2Cl2, 0 °C, 1 h; (b) diisopropylammonium tetrazolide (1.05 equiv), steroid-derived alcohols 4a–4f, CH2Cl2; (c) 5-(ethylthio)-1H-
tetrazole (1.1 equiv), CH3CN, H2O; (d) CH3CN, Et3N (2.0 equiv), p-TsOHÆH-Glu(OBn)-OBn (1.3 equiv); then CCl4 (10 equiv); (e) H2, cat. Pd (10%
on C), K2CO3 (1.5 equiv), THF–H2O, 3 h, rt.
benzyloxyphosphine 3, and then coupled with steroid-
derived alcohols 4a–4f to generate phosphoramidites
5a–5f (see Supporting Information for the structures,
synthesis, and characterization of 4a–4f), using conven-
tional methodology.16,17 These molecules were hydro-
lyzed to the corresponding steroid-conjugated
phosphites 6a–6f. Oxidative coupling with benzyl pro-
tected glutamate was accomplished with CH3CN–CCl4
to generate 7a–7f. Lastly, removal of benzyl protecting
groups was performed under hydrogenolysis conditions
to yield the phosphoramidate inhibitors 1a–1f as the tri-
potassium salts.
the enzyme can accommodate organic structures of con-
siderable size and lipophilicity in proximity to the cata-
lytic center. In at least one case, the propyl ether linker
had an interesting effect on the potency of the inhibitors
toward PSMA. For example, propyl ether linked inhib-
itor 1a was significantly more potent than the unlinked
analog of the same steroid 1e. The effect was much less
pronounced in linked inhibitor 1b versus its unlinked
counterpart 1f. Unfortunately, the unlinked form of
the relatively potent inhibitor 1c could not be readily
prepared with our synthetic scheme.
Overall, the inhibitors appear to be broadly differenti-
ated into two classifications based on their inhibitory
potency against PSMA. Inhibitors 1a and 1c exhibit
greater potency than 1b, 1d, 1e, and 1f, and within
these groups, the compounds attain fairly narrow
ranges of IC50 values. The activity of the more active
compounds (1a and 1c) is comparable to that of the
simple phenylalkyl phosphoramidate inhibitor 8. That
all compounds achieved sub-micromolar potency,
including the less active inhibitors in the series, sug-
Once prepared, the steroid-containing inhibitors 1a–1f
were assayed for inhibitory potency against purified
PSMA
using
methods
described
previously
(Fig. 1).15,18,19 For comparison, the IC50 value for a sim-
ple phenylalkyl phosphoramidate 8 identified in a previ-
ous study is also presented.20 All of the inhibitors
maintained sub-micromolar inhibitory potency against
PSMA, while two inhibitors (1a and 1c) achieved a
slight improvement over 8. These results suggest that

L. Y. Wu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 281–284
283
gests that the general glutamyl phosphoramidate scaf-
fold dominates the interactions with PSMA for this
class of compounds. The significance of these results
is that this scaffold appears to be capable of
accommodating a wide range of linked hydrophobic
molecular fragments while still retaining notable
potency.
Upon docking this series of inhibitors, the most nota-
ble observation was the effect of the propyl ether lin-
ker on the steroid conformation in compounds 1a (vs
1e) and 1b (vs 1f). Interestingly, the presence of the
linker caused the steroid nucleus to flip its docked ori-
entation relative to the inhibitor without a linker. A
result of this phenomenon is that it dictated the posi-
tions of the steroidal angular methyl groups, which
have a propensity to sterically clash with residues in
the PSMA active site. Thus, the docked conformations
of 1a (linked) and 1e (unlinked) show opposite orien-
tations of the steroid systems, and in the case of 1e,
the A-B ring angular methyl is oriented in a manner
which potentially clashes with the Tyr700 aromatic
ring. We speculate that such a steric effect could ex-
plain the significant potency difference between 1a
and 1e. The propyl linker also extends 1a through
the PSMA binding cavity in a manner that could al-
low the steroid alkyl chain to be involved in favorable
hydrophobic interactions with distant residues Pro502
and Tyr460, further explaining the improved potency
of this compound. A similar conformational flip was
observed with 1b (linked) and 1f (unlinked), but in
this case 1b exhibited potential steric clashing with
Tyr700, while 1f was observed in a relatively unhin-
dered conformation. This model could explain the rel-
atively poor potency of 1b, but the lack of inhibitory
To establish a tentative mode of binding and to rational-
ize the observed differentiation of the inhibitors by their
potencies against PSMA, phosphoramidates 1a–1f were
computationally docked into the active site of PSMA
(Fig. 2 and Supporting Information). The results were
obtained employing a recently determined high-resolu-
tion X-ray crystal structure21 in which the enzyme was
co-crystallized
with
a
phosphonate
inhibitor
(PDB = 2C6C). Docking of each inhibitor was per-
formed with FRED2 (OpenEyes) employing a library
of ligand conformations generated by OMEGA (Open-
Eyes). To filter docking poses non-productive for enzy-
matic inhibition,
constraint was utilized requiring the phosphoramidate
a
SMARTS pharmacophore
˚
oxygen atoms to be within 3 A of the two catalytic zinc
atoms of PSMA. Subsequently, the top consensus scor-
ing pose for each inhibitor was then minimized, without
constraint, in the MMFF94 force field as implemented
in SZYBKI (OpenEyes).
Figure 2. Results from docking inhibitors 1a and 1e into the active site of PSMA (PDB = 2C6C). Dark gray spheres represent the zinc atoms in the
PSMA active site. Docking results of other inhibitors are in Supporting Information. Structures were visualized with PYMOL.

284
L. Y. Wu et al. / Bioorg. Med. Chem. Lett. 18 (2008) 281–284
potency in 1f is not readily explained with this model.
However, the docked pose of the relatively potent
inhibitor 1c (linked) displayed an orientation of the
steroid framework in which little steric hindrance
was observed overall. Lastly, compound 1d (linked)
is unique relative to the other inhibitors due to the
linker attachment at the steroid D ring. The docked
conformation of this inhibitor presented the angular
methyl in a manner which could potentially clash with
Tyr700, which may explain its relative low potency.
References and notes
1. Carter, R. E.; Feldman, A. R.; Coyle, J. T. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 749.
2. Tiffany, C. W.; Lapidus, R. G.; Merion, A.; Calvin, D. C.;
Slusher, B. S. Prostate 1999, 39, 28.
3. Holmes, E. H.; Greene, T. G.; Tino, W. T.; Boynton, A.
L.; Aldape, H. C.; Misrock, S. L.; Murphy, G. P. Prostate
Suppl. 1996, 7, 25.
4. Horoszewicz, J. S.; Leong, S. S.; Kawinski, E.; Karr, J. P.;
Rosenthal, H.; Chu, T. M.; Mirand, E. A.; Murphy, G. P.
Cancer Res. 1983, 43, 1809.
In summary, a small library of hydroxysteroid-contain-
ing phosphoramidate PSMA inhibitors was prepared
and evaluated for inhibitory potency against PSMA.
The relative potencies were explained by a model gener-
ated through computational docking studies. Namely,
the least potent compounds in the series docked with
their steroidal angular methyl groups projected in an
opposite orientation from Tyr700 avoiding a potential
steric clash. In contrast, the remaining compounds pre-
senting docked poses in which the angular methyls pro-
jected away from Tyr700 all exhibited improved potency
(with the exception of 1f). The activity of the most po-
tent inhibitor (1a) may be explained by additional
hydrophobic contacts with Pro502 and Tyr460. Notably,
the entire series of inhibitors exhibited sub-micromolar
potency against PSMA suggesting that the binding of
these analogs may be driven by the N-phosphogluta-
mate core. Results from the designs of subsequent gen-
erations of PSMA inhibitors will be reported in due
course.
5. Bacich, D. J.; Pinto, J. T.; Tong, W. P.; Heston, W. D.
Mamm. Genome 2001, 122, 117.
6. Chang, S. S.; O’Keefe, D. S.; Bacich, D. J.; Reuter, V. E.;
Heston, W. D.; Gaudin, P. B. Clin. Cancer Res. 1999, 5,
2674.
7. Chang, S. S.; Reuter, V. E.; Heston, W. D.; Gaudin, P. B.
Urology 2001, 57, 801.
8. Tasch, J.; Gong, M.; Sadelain, M.; Heston, W. D. Crit.
Rev. Immunol. 2001, 21, 249.
9. Salit, R. B.; Kast, W. M.; Velders, M. P. Front. Biosci. J.
Virtual Library 2002, 7, e204.
10. Lu, J.; Celis, E. Cancer Res. 2002, 6220, 5807.
11. Fracasso, G.; Bellisola, G.; Cingarlini, S.; Castelletti, D.;
Prayer-Galetti, T.; Pagano, F.; Tridente, G.; Colombatti,
M. Prostate 2002, 531, 9.
12. Heston, W. D. Urology 1997, 49, 104.
13. Pinto, J. T.; Suffoletto, B. P.; Berzin, T. M.; Qiao, C. H.;
Lin, S.; Tong, W. P.; May, F.; Mukherjee, B.; Heston, W.
D. Clin. Cancer Res. 1996, 2, 1445.
14. Conway, R. E.; Petrovic, N.; Li, Z.; Heston, W.; Wu, D.;
Shapiro, L. H. Mol. Cell Biol. 2006, 26, 5310.
15. Maung, J.; Mallari, J. P.; Girtsman, T. A.; Wu, L. Y.;
Rowley, J. A.; Santiago, N. M.; Brunelle, A. N.; Berkman,
C. E. Bioorg. Med. Chem. 2004, 12, 4969.
16. Whalen, L. J.; McEvoy, K. A.; Halcomb, R. L. Bioorg.
Med. Chem. Lett. 2003, 13, 301.
17. Cohen, S. B.; Halcomb, R. L. J. Org. Chem. 2000, 65,
6145.
This work was supported in part by grants from the Na-
tional Institutes of Health, MBRS SCORE Program-
NIGMS (Grant No. S06-GM052588) and the Depart-
ment of Defense (Grant No. PC051060). The authors
would also like to extend their gratitude to W. Tam
and the NMR facility at SFSU for their expert
assistance.
18. Wone, D. W.; Rowley, J. A.; Garofalo, A. W.; Berkman,
C. E. Bioorg. Med. Chem. 2006, 14, 67.
19. Liu, T.; Toriyabe, Y.; Berkman, C. E. Protein Expr. Purif.
2006, 49, 251.
20. Wu, L. Y.; Anderson, M. O.; Toriyabe, Y.; Maung, J.;
Campbell, T. Y.; Tajon, C.; Kazak, M.; Moser, J.;
Berkman, C. E. Bioorg. Med. Chem. 2007, 15, 7434.
21. Mesters, J. R.; Barinka, C.; Li, W.; Tsukamoto, T.; Majer,
P.; Slusher, B. S.; Konvalinka, J.; Hilgenfeld, R. EMBO J.
2006, 25, 1375.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.10.096.