Improving the Specificity of the Prostate-Specific Antigen Substrate Glutaryl-Hyp-Ala-Ser-Chg-Gln as a Promoiety

Article abstract of DOI:10.1111/cbdd.12559

To develop PSA peptide substrates with improved specificity and plasma stability from the known substrate sequence glutaryl-Hyp-Ala-Ser-Chg-Gln, systematic replacements of the N-terminal segment with D-retro-inverso-peptides were performed with the incorporation of 7-amino-4-methylcoumarin (7-AMC) after Gln for convenient fluorometric determination and ranking of the PSA substrate activity. The D-retro-inverso-peptide conjugates with P2-P5 D-amino acid substitutions were moderate but poorer PSA substrates as compared to the original peptide, suggesting that inversion of the amide bonds and/or incorporation of the additional atom as in the urea linker adversely affected PSA binding. However, P5 substitution of Hyp with Ser showed significant improvements in PSA cleavage rate; the resulting AMC conjugate, glutaryl-Ser-Ala-Ser-Chg-Gln-AMC (11), exhibited the fastest PSA cleavage rate of 351 pmol/min/100 nmol PSA. In addition, GABA←mGly-Ala-Ser-Chg-Gln-AMC (conjugate 6) was the second best PSA substrate and released 7-AMC at a rate of 225 pmol/min/100 nmol PSA as compared to 171 pmol/min/100 nmol PSA for the control conjugate glutaryl-Hyp-Ala-Ser-Chg-Gln-AMC. Incubations of selected AMC conjugates with mouse and human plasma revealed that GABA←D-Ser-ψ[NH-CO-NH]-Ala-Ser-Chg-Gln-AMC (5) and GABA←mGly-Ala-Ser-Chg-Gln-AMC (6) were most stable to non-PSA-mediated proteolysis. Our results suggest that the PSA specificity of glutaryl-Hyp-Ala-Ser-Chg-Gln is improved with Ser and mGly substitutions of Hyp at the P5.

Full text of DOI:10.1111/cbdd.12559

Chem Biol Drug Des 2015; 86: 837–848
Research Article
Improving the Specificity of the Prostate-Specific
Antigen Substrate Glutaryl-Hyp-Ala-Ser-Chg-Gln as a
Promoiety
1
1,2,3,
Herve Aloysius and Longqin Hu
*
c-aminobutyric acid; HPLC, high-performance liquid chroma-
tography; IPCF, isopropyl chloroformate; LC-MS, liquid chro-
matography–mass spectrometry; mGly, malonyl analog of
glycine; NaSeH, sodium hydrogen selenide; NMP, N-methyl-
2-pyrrolidone; NMR, nuclear magnetic resonance; PSA, pros-
tate-specific antigen; TFA, trifluoroacetic acid; TLC, thin-layer
chromatography.
1
Department of Medicinal Chemistry, Ernest Mario School
of Pharmacy, Rutgers, The State University of New Jersey,
Piscataway, NJ 08854, USA
School of Pharmaceutical Sciences, Shanxi Medical
University, Taiyuan 030001, China
The Cancer Institute of New Jersey, New Brunswick, NJ
2
3
0
8901, USA
Received 15 January 2015 and accepted for publication 5
February 2015
*
Corresponding author: Longqin Hu, longHu@rutgers.edu
To develop PSA peptide substrates with improved
specificity and plasma stability from the known sub-
strate sequence glutaryl-Hyp-Ala-Ser-Chg-Gln, system-
atic replacements of the N-terminal segment with
D-retro-inverso-peptides were performed with the
incorporation of 7-amino-4-methylcoumarin (7-AMC)
after Gln for convenient fluorometric determination and
ranking of the PSA substrate activity. The D-retro-inv-
erso-peptide conjugates with P2-P5 D-amino acid sub-
stitutions were moderate but poorer PSA substrates as
compared to the original peptide, suggesting that
inversion of the amide bonds and/or incorporation of
the additional atom as in the urea linker adversely
affected PSA binding. However, P5 substitution of Hyp
with Ser showed significant improvements in PSA
cleavage rate; the resulting AMC conjugate, glutaryl-
Ser-Ala-Ser-Chg-Gln-AMC (11), exhibited the fastest
PSA cleavage rate of 351 pmol/min/100 nmol PSA. In
addition, GABA mGly-Ala-Ser-Chg-Gln-AMC (conju-
gate 6) was the second best PSA substrate and
released 7-AMC at a rate of 225 pmol/min/100 nmol
PSA as compared to 171 pmol/min/100 nmol PSA for
the control conjugate glutaryl-Hyp-Ala-Ser-Chg-Gln-
AMC. Incubations of selected AMC conjugates with
mouse and human plasma revealed that GABA
D-Ser-w[NH-CO-NH]-Ala-Ser-Chg-Gln-AMC (5) and
GABA mGly-Ala-Ser-Chg-Gln-AMC (6) were most
stable to non-PSA-mediated proteolysis. Our results
suggest that the PSA specificity of glutaryl-Hyp-Ala-
Ser-Chg-Gln is improved with Ser and mGly substitu-
tions of Hyp at the P5.
Prostate cancer is the most common type of cancer
reported and the second leading cause of morbidity and
death among American males next to lung cancer.
According to the American Cancer Society, there will be
around 233 000 newly diagnosed cases of prostate can-
cer and about 29 480 estimated deaths from prostate
cancer in 2014 (1,2). Although hormonal treatment through
androgen ablation may initially reduce tumor growth and
metastasis rates, it is rarely curative because tumor cells
become refractory to hormonal therapy within 1–2 years
(3). Furthermore, the propensity of prostate cancer to pro-
gress through osteoblastic metastasis remains a clinical
challenge. While alternative treatment modalities involving
platinum compounds, radiopharmaceuticals and monoclo-
nal antibodies have been explored recently (4–8), the clini-
cal use of chemotherapeutics such as docetaxel,
mitoxantrone, and doxorubicin is limited by systemic toxic-
ity stemming from non-discriminatory drug exposure to
normal tissues (9–11). To mitigate this issue, one of the
approaches developed to improve the selectivity of che-
motherapy toward tumor cells involves the targeted pro-
drug strategy (12–15). Secretion of active prostate-specific
antigen (PSA) in the tumor microenvironment can be
exploited for targeted activation of peptide prodrugs in
cancer cells (15). PSA is a serine protease with chymo-
trypsin-like activity and is predominantly expressed in the
prostate although present in other normal and tumor tis-
sues at low concentrations (16–18). It exhibits high sub-
strate specificity being the only endopeptidase known to
cleave peptide substrates at the amide bond after Gln.
Moreover, its enzymatic activity is restricted to the tumor
microenvironment due to inactivation by a1-antichymotryp-
sin and a2-macroglobulin in the systemic circulation
(19,20), which makes PSA an attractive target for tumor-
selective prodrug activation. Several prodrugs consisting of
Key words: 7-amino-4-methylcoumarin, D-retro-inverso-pep-
tide, prostate-specific antigen
Abbreviations: 7-AMC, 7-amino-4-methylcoumarin; 7-AZD,
-azido-4-methylcoumarin; CPM, counts per min; DCM, di-
chloromethane; FCC, flash column chromatography; GABA,
7
ª 2015 John Wiley & Sons A/S. doi: 10.1111/cbdd.12559
837

Aloysius and Hu
PSA-specific peptide substrates coupled to cytotoxic
drugs have been reported (21–23).
reagent or HPLC grade and used without further purifica-
tion. Reactions were monitored by TLC and/or on an LC-
MS system consisting of PE 200 Series autosampler and
pumps (Perkin Elmer, Waltham, MA, USA) coupled to an
LCQ ion trap mass spectrometer (Thermo Scientific, Wal-
tham, MA, USA). Accurate mass values of final peptide
conjugates were determined by direct inlet infusion of
10 lg/mL solutions using a LTQ Orbitrap (Thermo Scien-
tific). Flash column chromatography (FCC) was performed
on a Teledyne ISCO CombiFlash Companion Automated
Flash Chromatographic System (Teledyne Technologies,
Thousand Oaks, CA, USA) with prepacked silica gel col-
Glutaryl-Hyp-Ala-Ser-Chg-Gln-Ser-Leu-Dox
(L-377202,
prodrug I) is one example of a PSA-activated prodrug for
prostate cancer that was advanced to phase I clinical
studies (24). Based on the cleavage sequences of PSA’s
natural substrates, semenogelins I and II, the highly spe-
cific PSA substrate, glutaryl-Hyp-Ala-Ser-Chg-Gln was
developed and coupled to the aminoglycoside of doxorubi-
cin via a Ser-Leu linker (21,25). Following intravenous
administration of prodrug I to mice, rats, dogs, and mon-
keys in preclinical studies, roughly 30–40% of the parent
compound was metabolized to doxorubicin by non-PSA-
mediated mechanisms (25). In tumor-bearing mice, the
dose fraction converted to doxorubicin increased to 87%
versus 28% in control mice (25). Following intravenous
administration of prodrug I to 19 patients with advanced
1
13
umns. All H and C NMR spectra were recorded on a
400 MHz Bruker spectrometer at ambient temperature
and calibrated using residual undeuterated solvents as the
internal reference.
hormone-refractory prostate cancer at escalating doses
Synthesis of 7-Azido-4-methylcoumarin (7-AZD,
12)
2
(up to 315 mg/m ), circulating levels of the active metabo-
lites Leu-doxorubicin (AUC = 4.3 lM h, Cmax = 5 lM) and
doxorubicin (AUC = 1.2 lM h, max = 0.12 lM) were
7-azido-4-methylcoumarin was prepared as a yellow solid
in 79% yield (234 mg) starting from 7-amino-4-methyl cou-
marin (257 mg) as previously described (28). H NMR
C
2
1
detected at the efficacious dose of 225 mg/m (24,26,27).
Dose-limiting neutropenia was also observed in 1 of 6
patients at this dose (24). The protease responsible for the
non-specific hydrolysis of prodrug I was not identified.
Although the observed toxicity was not correlated with
systemic levels of active metabolites, the clinical develop-
ment of prodrug I was not pursued beyond phase I stud-
ies most likely due to toxicity concerns following
intravenous administration at the efficacious dose. Thus,
we decided to further improve the PSA specificity of gluta-
ryl-Hyp-Ala-Ser-Chg-Gln by introducing D-retro-inverso-
peptide fragments in its sequence using urea and malo-
nate (mGly) linkers. Although previous optimization studies
suggest that amino acid substitution in P1-P4 for the
sequence of prodrug I may compromise substrate speci-
ficity for PSA (25), we anticipated that D-retro-inverso-pep-
tide fragment insertion will mitigate non-PSA-mediated
hydrolysis while maintaining or improving PSA cleavage
rate. This is based on the assumption that amino acid
side-chain orientation is most significant to optimal PSA
substrate binding. Sequence optimization was streamlined
with the choice of a fluorescent tag that was readily cou-
pled to a variety of peptides. The optimized sequences
can then be coupled to cytotoxic drugs via appropriate
linkers and evaluated. In this study, peptide conjugates of
the fluorescent dye 7-amino-4-methylcoumarin (7-AMC)
were synthesized as convenient surrogates for developing
peptide substrates with improved PSA specificity because
enzyme-mediated cleavage of AMC conjugates at the Gln-
AMC bond results in quantifiable fluorescence increase.
(DMSO-d
6
, 400 MHz): d 7.83 (d, 1H, J = 8.5 Hz), 7.2 (m,
1H), 7.18 (d, 1H, J = 2.1 Hz), 6.39 (s, 1H), 2.47 (s, 3H);
13
6
C NMR (DMSO-d , 100 MHz): d 159.5, 154.0, 152.9,
143.3, 126.9, 116.7, 115.5, 113.2, 106.8, 18.0, MS
(ESI+): m/z (intensity) 201.9 ([M + H] , 100%).
+
a
Synthesis of 7-(N -t-butyloxycarbonyl-L-glutaminyl)
amino-4-methylcoumarin (Boc-Gln-AMC, 13)
Boc-Gln-AMC was synthesized using our previously
reported amidation procedure between Boc-Gln seleno-
carboxylate and 7-AZD (28). Briefly, a fresh solution of
4
NaHSe was obtained by reacting NaBH and an isopro-
panolic suspension of Se as previously described (29). In
parallel, the mixed anhydride of Boc-Gln-OH was prepared
by adding a 1.0 M solution of IPCF in toluene (2.0 mL,
2.0 mmol) to a solution of Boc-Gln-OH (2.0 mmol) and N-
methylpiperidine (244 lL, 2.0 mmol) in 20 mL of anhy-
drous THF at ꢀ15 °C under nitrogen and allowing the mix-
ture to stir for 20 min at ꢀ15 °C. The selenocarboxylate
was generated in situ by adding the freshly prepared
NaHSe reagent above to the mixed anhydride via cannula
over a period of 5 min and stirring the reaction mixture for
an additional 30 min below ꢀ10 °C under nitrogen atmo-
sphere. A solution of 7-AZD (1.62 mmol) in 2 mL of anhy-
drous THF was cannulated into the selenocarboxylate
solution, and the reaction carried out at room temperature
under nitrogen for 2 h. Organic solvents were removed;
the residue was suspended in 15 mL of a saturated
sodium bicarbonate solution and extracted with 45 mL
(39) of ethyl acetate. The organic extracts were combined,
Methods and Materials
washed sequentially with water and brine, and dried over
anhydrous sodium sulfate. After filtering out sodium sul-
fate, the crude mixture was dry-loaded onto silica gel and
purified by FCC. The product (142 mg) was obtained as a
Fmoc-protected amino acids were purchased from Chem-
Impex (Wood Dale, IL, USA). Solvents were either ACS
8
38
Chem Biol Drug Des 2015; 86: 837–848

Improving Specificity of PSA Substrate
1
light-yellow solid in 84% yield.
00 MHz): d 7.70 (d, 1H, J = 2.0 Hz), 7.60 (d, 1H,
J = 8.7 Hz), 7.39 (d, 1H, J = 8.2 Hz), 6.13 (s, 1H), 4.12
H
NMR (CD
3
OD,
100 MHz): d 142.6, 139.4, 137.4, 128.9, 127.2, 124.2,
121.3, 120.0, 109.6, 77.5, 61.8, 51.1, 31.6; MS (ESI+): m/
z (intensity) 325.8 ([M+H]+, 100%).
4
(
s, 1H), 2.35 (s, 3H), 2.28 (t, 2H, J = 7.5 Hz), 2.02 (m,
13
1
1
1
5
4
H), 1.87 (m, 1H), 1.36 (s, 9H);
3
C NMR (CDCl ,
00 MHz): d 177.8, 173.4, 163.2, 157.9, 155.4, 155.2,
49.9, 143.4, 126.7, 117.3, 117.2, 113.7, 108.1, 81.0,
6.5, 32.5, 29.1, 28.8, 18.5; MS (ESI+): m/z (intensity)
Loading of Fmoc-GABA on 2-chloro-trityl chloride
resin
Attempts to load Fmoc-GABA on Wang resin failed using
DIC, IPCF, or PyBOP in the presence of catalytic DMAP
+
03.7 ([M + H] , 12%), 706.7 ([2M + H-Boc]+, 100%).
(0.1 eq.). Alternatively, Fmoc-GABA (1.2 g, 3.6 mmol) was
effectively coupled to 2-chloro-trityl chloride resin (1 g,
1.7 mmol) in 10 mL of anhydrous DCM/NMP (9:1) in the
presence of 4 eq. DIEA at room temperature for 3 h. The
loading levels were determined to be 0.48 mmol/g resin
by Fmoc-deprotection with 20% piperidine and UV analy-
sis (A290). The unreacted 2-chloro-trityl chloride on the
resin was end-capped with 2 mL of methanol/DIEA (9:1) in
10 mL of anhydrous DCM under nitrogen for 60 min at
room temperature, and the resin was ready for further
coupling after 20% piperidine deprotection.
Synthesis of glutaryl-Hyp-Ala-Ser-Chg-Gln-AMC (1)
The solid-phase synthesis of Fm-glutaryl-Hyp-Ala-Ser-
Chg-OH was carried out using standard solid-phase pep-
tide synthesis procedures. Loading of Fmoc-Chg-OH on
1
g of Wang resin (0.35 mmol/g resin) yielded 148 mg of
peptide which was used without further purification. In
order to couple Gln-AMC to the peptide, Boc-Gln-AMC
(60 mg, 0.15 mmol) was deprotected with 1 mL of TFA/
DCM/H O (50:45:5) in 30 min at room temperature. Sol-
2
vents were removed under a gentle stream of nitrogen;
the residue was reconstituted in 400 lL of methanol and
treated with Amberlyst A-26 (0.5 g, 0.4 mmol) to remove
residual TFA. After filtering out Amberlyst A-26, methanol
was removed and the light-yellow residue resuspended in
Synthesis of GABA D-Ser D-Ala D-Ser D-Chg-
w[NH-CO-NH]-Gln-AMC (2)
GABA D-Ser(tBu) D-Ala D-Ser(tBu) D-Chg-H
was
5
0
00 lL of NMP ready for the next step. Gln-AMC (17 mg,
.042 mmol) was coupled to Fm-glutaryl-Hyp-Ala-Ser-
generated on 2-chloro-trityl chloride resin (1.8 g,
0.48 mmol/g resin) starting with the GABA resin prepared
above and following standard automated peptide synthesis
Chg-OH (9.6 mg, 0.013 mmol) in 600 lL of NMP using
HBTU (10 mg, 0.027 mmol) preactivation of the peptide
for 30 min in the presence of 1 eq. diethylisopropylamine
procedures.
Resin-GABA D-Ser(tBu) D-Ala D-Ser
(tBu) D-Chg-H (53 mg, 0.026 mmol) was activated with
CDI (22 mg, 0.14 mmol) in 200 lL of anhydrous DCM for
3 h at room temperature to generate the carbonyl-imidaz-
ole intermediate (Scheme 1). After washing the resin with
three 1-mL portions of anhydrous DCM, coupling to H-
Gln-AMC (43 mg, 0.14 mmol) was carried out in 600 lL
of DCM/NMP (60:40) for 48 h at room temperature. The
protected peptide was cleaved off the resin with 5% TFA/
DCM. After HPLC purification, treatment with 50% TFA/
DCM at room temperature for 2 h afforded the desired
(DIEA) at room temperature, followed by the addition of
Gln-AMC. The reaction was allowed to proceed for 16 h
at room temperature, after which 10% piperidine depro-
tection was carried out followed by HPLC purification. The
desired peptide–AMC conjugate, Glutaryl-Hyp-Ala-Ser-
Chg-Gln-AMC (1), was obtained in 52% yield (5.9 mg) and
+
was > 98% pure by LC-UV analysis. HRMS (ESI ) m/z
+
calc’d for
28.3806.
C
39
H
54
N
7
O
13 [M + H] : 828.3780, found:
8
peptide–AMC
conjugate,
GABA D-Ser D-Ala D-
Ser D-Chg-[NH-CO-NH]-Gln-AMC (2), in 6% yield
c
+
Synthesis of N -(9-Fluorenyloxycarbonyl)-c-
aminobutyric Acid (Fmoc-GABA, 14)
(1.5 mg, 95% pure by LC-UV analysis). HRMS (ESI ) m/z
+
calc’d for
817.3753.
C
37
H
53
N
8
O
13 [M + H] : 817.3732, found:
To a solution of GABA (2.00 g, 19.4 mmol, in 14 mL 10%
3
NaHCO ), Fmoc-OSu (4 g, 11.7 mmol, in 40 mL acetoni-
trile) was added drop-wise over a period of 2 h at room
temperature. The mixture was allowed to stir at room tem-
perature for an additional hour. Acetonitrile was removed
under reduced pressure and the aqueous layer acidified to
pH 1 with 10% HCl. The precipitate was washed with two
Synthesis of H-Chg-Gln-AMC (16)
Fmoc-Chg-OH (104 mg, 0.27 mmol) was preactivated
with HBTU (107 mg, 0.28 mmol) in the presence of 5 eq.
DIEA in 200 lL of NMP for 1 h and then coupled to H-
Gln-AMC (72 mg, 0.24 mmol) at room temperature for 2 h
under nitrogen atmosphere. The crude mixture was sus-
20-mL portions of water, 20 mL of ethyl acetate, and dried
under reduced pressure. Fmoc-GABA was obtained as a
1
white solid in 73% yield (2.8 g). H NMR (DMSO-d6,
pended in 10 mL of saturated NaHCO and extracted with
3
4
00 MHz): d 7.89 (d, 2H, J = 7.4 Hz), 7.44 (d, 2H,
J = 7.2 Hz), 7.42 (t, 2H, J = 7.5 Hz), 7.35 (s, 1H), 7.33 (t,
H, J = 7.0 Hz), 4.30 (d, 2H, J = 7 Hz), 4.21 (t, 1H,
J = 6.7 Hz), 3.01 (q, 2H, J = 5.6 Hz), 2.20 (t, 2H,
30 mL of DCM three times. The organic layer was dry-
loaded onto silica gel and purified by FCC. The protected
peptide intermediate (134 mg) was deprotected with
0.5 M TBAF and purified by HPLC to generate the desired
peptide (m/z = 443.1) in 57% yield (60 mg).
2
13
J = 7.3 Hz), 1.63 (q, 2H, J = 7.1 Hz); C NMR (DMSO-d6,
Chem Biol Drug Des 2015; 86: 837–848
839

Aloysius and Hu
Scheme 1: Synthesis of peptide–AMC conjugate 2.
Synthesis of GABA D-Ser D-Ala D-Ser-w[NH-
CO-NH]-Chg-Gln-AMC (3)
The amino end of resin-GABA D-Ser(tBu) D-Ala D-Ser
600 lL of DCM/NMP (60:40) for 24 h at room tempera-
ture. The protected peptide was cleaved off the resin with
5% TFA/DCM. After HPLC purification, treatment with
50% TFA/DCM at room temperature for 2 h afforded the
desired peptide–AMC conjugate, GABA D-Ser D-Ala-
[NH-CO-NH]-Ser-Chg-Gln-AMC (4), in 7.8% yield (1.6 mg,
(
(
tBu)-H (69 mg, 0.034 mmol) was preactivated with CDI
27 mg, 0.17 mmol) in 200 lL of anhydrous DCM at room
temperature for 3 h and, after three 1-mL DCM washes,
coupled to H-Chg-Gln-AMC in 600 lL of DCM/NMP
+
> 99% pure by LC-UV analysis). HRMS (ESI ) m/z calc’d
+
(60:40) for 48 h at room temperature. The protected pep-
for C H N O [M + H] : 817.3732, found: 817.3754.
3
7 53 8 13
tide was cleaved off the resin with 5% TFA/DCM. After
HPLC purification, treatment with 50% TFA/DCM at room
temperature for 2 h afforded the desired peptide–AMC
Me,Me
Synthesis of H-Ala-Ser(Ψ
pro)-Chg-Gln-AMC
conjugate,
GABA D-Ser D-Ala D-Ser-[NH-CO-NH]-
(18)
Fmoc-Ala-Ser(Ψ
Me,Me
Chg-Gln-AMC (3), in 10% yield (2.1 mg, > 99% pure by
pro)-OH (136 mg, 0.31) was preacti-
+
LC-UV analysis). HRMS (ESI ) m/z calc’d for C37
H
53
N
8
O
13
vated with HBTU (105 mg, 0.27 mmol) in the presence of
5 eq. DIEA in 400 lL of NMP for 30 min and then coupled
to H-Chg-Gln-AMC (97 mg, 0.22 mmol) at room tempera-
ture for 3 h under nitrogen atmosphere. One-pot Fmoc-
deprotection was carried out with 10% DEA and NMP
was removed with repeated hexane washes (5-mL por-
tions, at least 10 times). The crude mixture was dissolved
in 1 mL of 50% acetonitrile/water and purified by FCC.
The peptide 18 (m/z = 641.1) was obtained in 66% yield
(93 mg) and > 98% pure by LC-UV analysis.
+
[
M + H] : 817.3732, found: 817.3766.
Synthesis of H-Ser(tBu)-Chg-Gln-AMC (17)
Fmoc-Ser(tBu)-OH (116 mg, 0.30 mmol) was preactivated
with HBTU (120 mg, 0.32 mmol) in the presence of 5 eq.
DIEA in 400 lL of NMP for 30 min and then coupled to
H-Chg-Gln-AMC (119 mg, 0.27 mmol) at room tempera-
ture for 10 h under nitrogen atmosphere. The crude mix-
ture was suspended in 10 mL of saturated NaHCO3 and
extracted with 30 mL of DCM three times. The organic
layer was dry-loaded onto silica gel and purified by FCC.
The protected peptide intermediate (53 mg) was depro-
tected with 0.5 M TBAF and purified by HPLC to generate
the desired peptide (m/z = 586.0 in 28% yield (21 mg).
Synthesis of GABA D-Ser-w[NH-CO-NH]-Ala-Ser-
Chg-Gln-AMC (5)
Conjugate 5 was prepared by activating the amino end of
resin-GABA D-Ser(tBu)-H (54 mg, 0.025 mmol) with CDI
(20 mg, 0.12 mmol) in 200 lL of anhydrous DCM at room
temperature for 3 h and, after three 1 mL DCM washes,
Me,Me
Synthesis of GABA D-Ser D-Ala-w[NH-CO-NH]-
Ser-Chg-Gln-AMC (4)
coupled to H-Ala-Ser(Ψ
pro)-Chg-Gln-AMC (50 mg,
0.078 mmol) in 600 lL of DCM/NMP (60:40) for 24 h at
room temperature. The protected peptide was cleaved off
the resin with 5% TFA/DCM. After HPLC purification, treat-
ment with 50% TFA/DCM at room temperature for 2 h
afforded the desired peptide–AMC conjugate, GABA D-
Ser-[NH-CO-NH]-Ala-Ser-Chg-Gln-AMC (5), in 7.8% yield
Conjugate 4 was prepared by activating the amino end of
resin-GABA D-Ser(tBu) D-Ala-H (50 mg, 0.025 mmol)
with CDI (23 mg, 0.14 mmol) in 200 lL of anhydrous
DCM at room temperature for 3 h and, after three 1-mL
DCM washes, coupled to H-Ser(tBu)-Chg-Gln-AMC in
8
40
Chem Biol Drug Des 2015; 86: 837–848

Improving Specificity of PSA Substrate
+
(
2.6 mg, > 98% pure by LC-UV analysis). HRMS (ESI ) m/
Synthesis of GABA D-Ser mGly-Ser-Chg-Gln-
AMC (7)
Fluorenylmethyl malonate (8.2 mg, 0.029 mmol), H-Ser-
Chg-Gln-AMC (14 mg, 0.024 mmol), and EDC (12 mg,
+
z calc’d for C37
17.3754.
H
53
N
8
O
13 [M + H] : 817.3732, found:
8
0
.063 mmol) were dissolved in 400 lL of NMP in the
Synthesis of t-butyl fluorenylmethyl malonate (15)
Malonic acid (283 mg, 2.7 mmol) and fluorenyl methanol
presence of DIEA (5 eq.) at room temperature under nitro-
gen atmosphere. DMAP (7.6 mg, 0.063 mmol) was
added, and the reaction carried out at room temperature
for 16 h under nitrogen atmosphere. The reaction mixture
was washed with 2 mL of hexane (five times), reconsti-
tuted in 400 lL of NMP, and used for the second coupling
step without further purification. Additional portions of EDC
(25 mg, 0.13 mol) and DMAP (12 mg, 0.098 mmol) were
added to couple the peptide to resin-GABA D-Ser(tBu)-H
(20 mg, 0.01 mmol) at room temperature for 16 h in the
presence of 5 eq. DIEA. The protected peptide was
cleaved off the resin with 5% TFA/DCM. After HPLC purifi-
cation, treatment with 50% TFA/DCM at room temperature
for 2 h afforded the desired peptide–AMC conjugate,
GABA D-Ser mGly-Ser-Chg-Gln-AMC (7), in 32% yield
(
213 mg, 1.1 mmol) were dissolved in 2 mL of acetonitrile
at room temperature under nitrogen atmosphere. EDC
531 mg, 2.8 mmol) was added, and the reaction was car-
(
ried out at room temperature for 30 min under nitrogen
atmosphere. After removing acetonitrile under reduced
pressure, the residue was redissolved in 50 mL of DCM
and extracted with two 20-mL portions of saturated NaH-
CO
3
. The aqueous layers were combined, acidified to pH
1, and extracted with 35 mL of DCM three times. The
organic layers were combined, washed with water, brine
and dried over sodium sulfate. DCM was removed under
reduced pressure, and the product recovered as an off-
white solid in 57% yield (176 mg, > 99% pure by LC-UV
1
+
analysis). H NMR (CDCl
3
, 400 MHz): d 7.80 (d, 2H,
(2.3 mg, > 98% pure by LC-UV analysis). HRMS (ESI ) m/
+
J = 7.4 Hz), 7.61 (d, 2H, J = 6.8 Hz), 7.44 (t, 2H,
J = 8.0 Hz), 7.35 (t, 2H, J = 8.3 Hz), 4.52 (d, 2H,
z calc’d for C36
H
50
N
7
O
13 [M + H] : 788.3467, found:
788.3454.
13
J = 7.2 Hz), 4.27 (t, 1H, J = 7.2 Hz), 3.55 (s, 2H);
C
NMR (CDCl , 100 MHz): d 167.9, 166.8, 143.5, 140.7,
3
127.7, 127.1, 125.2, 120.1, 66.4, 46.3, 41.6.
Synthesis of GABA D-Lys mGly-Ser-Chg-Gln-
AMC (8)
Fluorenylmethyl malonate (20 mg, 0.071 mmol), H-Ser-
Chg-Gln-AMC (27 mg, 0.046 mmol), and EDC (22 mg,
0.12 mmol) were dissolved in 400 lL of NMP at room
temperature in the presence of 5 eq. DIEA under nitrogen
atmosphere. DMAP (14 mg, 0.12 mmol) was added, and
the reaction was carried out at room temperature for 16 h
under nitrogen atmosphere. The reaction mixture was
washed with 2 mL of hexane (five times), reconstituted in
400 lL of NMP, and used for the second coupling step
without further purification. Additional portions of EDC
(44 mg, 0.23 mol) and DMAP (28 mg, 0.13 mmol) were
added to couple the peptide to resin-GABA D-Lys(Boc)-
H (20 mg, 0.01 mmol) at room temperature for 24 h in the
presence of 5 eq. DIEA. The protected peptide conjugate
was cleaved off the resin with 5% TFA/DCM. After HPLC
purification, treatment with 50% TFA/DCM at room tem-
perature for 2 h afforded the desired peptide–AMC conju-
gate, GABA D-Lys mGly-Ser-Chg-Gln-AMC (8), in 11%
yield (1.1 mg, > 98% pure by LC-UV analysis). HRMS
Me,Me
Synthesis of HO-mGly-Ala-Ser(Ψ
Gln-AMC (16)
pro)-Chg-
Fluorenylmethyl malonate (14 mg, 0.050 mmol), H-Ala-Ser
Me,Me
(
Ψ
pro)-Chg-Gln-AMC (20 mg, 0.031 mmol), and
EDC (13 mg, 0.068 mmol) were dissolved in 400 lL of
NMP in the presence of 5 eq. DIEA at room temperature
under nitrogen atmosphere. DMAP (9 mg, 0.07 mmol)
was added, and the reaction was carried out for 16 h at
room temperature under nitrogen atmosphere. The reac-
tion mixture was washed with 2 mL of hexane (five times)
to remove NMP, reconstituted in 1 mL of 50% acetonitrile/
water and purified by HPLC. The peptide 16 (m/z = 641.1)
was obtained in 62% yield (14 mg) and was > 98% pure
by LC-UV analysis.
Synthesis of GABA mGly-Ala-Ser-Chg-Gln-AMC
(
6)
HO-mGly-Ala-Ser(Ψ
.017 mmol) was activated with EDC (9 mg, 0.05 mmol)/
Me,Me
+
+
pro)-Chg-Gln-AMC
(12 mg,
(ESI ) m/z calc’d for C39
H
57
N
8
O
12 [M + H] : 829.4096,
0
found: 829.4092.
DMAP (6 mg, 0.05 mmol) in the presence of DIEA (5 eq.)
in 400 lL of NMP and coupled to GABA resin (20 mg,
0
.0094 mmol) for 48 h at room temperature. The pro-
Synthesis of GABA D-Ser D-Ala mGly-Chg-Gln-
AMC (9)
tected peptide was cleaved off the resin with 5% TFA/
DCM. After HPLC purification, treatment with 50% TFA/
DCM at room temperature for 2 h afforded the desired
peptide–AMC conjugate, GABA mGly-Ala-Ser-Chg-Gln-
AMC (6), in 32% yield (2.5 mg, > 98% pure by LC-UV
analysis). HRMS (ESI ) m/z calc’d for
Fluorenylmethyl malonate (16 mg, 0.057 mmol), H-Chg-
Gln-AMC (21 mg, 0.047 mmol), and EDC (23 mg,
0.12 mmol) were dissolved in 400 lL of NMP at room
temperature in the presence of DIEA (5 eq.) under nitrogen
atmosphere. DMAP (13 mg, 0.11 mmol) was added, and
the reaction was carried out at room temperature for 16 h
+
36 50 7 12
C H N O
+
[M + H] : 772.3517, found: 772.3520.
Chem Biol Drug Des 2015; 86: 837–848
841

Aloysius and Hu
under nitrogen atmosphere. The reaction mixture was
washed with 2 mL of hexane (five times), reconstituted in
sure and the protected peptide conjugate purified by
HPLC. Treatment with 50% TFA/DCM at room temperature
for 2 h afforded the desired peptide–AMC, Glutaryl-Ser-
Ala-Ser-Chg-Gln-AMC (11), in 37% yield (93.9 mg, > 98%
4
00 lL of NMP, and used for the second coupling step
without further purification. Additional portions of EDC
45 mg, 0.24 mol) and DMAP (31 mg, 0.25 mmol) were
+
(
pure by LC-UV analysis). HRMS (ESI ) m/z calc’d for
+
added to couple the peptide to resin-GABA D-Ser(tBu)
C H N O [M + H] : 802.3623, found: 802.3656.
3
7 52 7 13
D-Ala-H (26 mg, 0.013 mmol) at room temperature for
2
4 h in the presence of 5 eq. DIEA. The protected peptide
conjugate was cleaved off the resin with 5% TFA/DCM.
After HPLC purification, treatment with 50% TFA/DCM at
room temperature for 2 h afforded the desired peptide–
AMC conjugate, GABA D-Ser D-Ala mGly-Chg-Gln-
PSA and blood stability assays
Measurement of peptide concentration by UV and
the PSA enzyme assay
AMC (9), in 20% yield (2.0 mg, > 95% pure by LC-UV
Peptide concentrations were determined by measuring the
UV absorbance at 324 nm of 20 lM solutions prepared
based on weight in methanol from 10 mM DMSO stock
solutions. Using the predetermined extinction coefficient of
Boc-Gln-AMC at 324 nm (12 800/M/cm), peptide concen-
trations were calculated and adjusted accordingly to pre-
pare 10 lM solutions in 50 mM Tris buffer, pH 8 containing
2 mM CaCl2 and 0.1% Tween-20. Each peptide solution
(4 lL) was added to 36 lL of Tris buffer (final concentra-
tion 1 lM) in triplicate wells of a white Costar 384-well
plate (Corning, Inc., Corning, NY) and centrifuged at
1600 9 g for 1 min. Reactions were initiated with the
addition of 4 lL aliquots of a 1.1 lM solution of human
PSA (final concentration 100 nM), and time-points recorded
over a period of 3 h. The stability of peptide conjugates in
50 mM Tris buffer was evaluated by measuring the fluores-
cence of 1 lM solutions throughout the course of the
assay. Buffer fluorescence was measured by adding
40 lL aliquots of 50 mM Tris buffer to separate wells in
triplicate. All wells were read using a 335/460 nm (ex/em)
filters on a Victor 3V 1420 Multilabel Counter (PerkinElmer
Life and Analytical Sciences, Shelton, CT, USA). PSA
cleavage rates were calculated by taking the slopes of the
linear portions of fluorescence counts–time curves and
+
analysis). HRMS (ESI ) m/z calc’d for
36 50 7 12
C H N O
+
[M + H] : 772.3517, found: 772.3536.
Synthesis of GABA D-Ser D-Ala D-Ser mGly-
Gln-AMC (10)
Fluorenylmethyl malonate (25 mg, 0.089 mmol), H-Gln-
AMC (21 mg, 0.047 mmol), and EDC (20 mg, 0.10 mmol)
were dissolved at room temperature in 400 lL of NMP in
the presence of 5 eq. DIEA under nitrogen atmosphere.
DMAP (11 mg, 0.90 mmol) was added, and the reaction
was carried out at room temperature for 16 h under nitro-
gen atmosphere. The reaction mixture was washed with
2
mL of hexane (five times), reconstituted in 400 lL of
NMP, and used for the second coupling step without fur-
ther purification. Additional portions of EDC (67 mg,
0
.35 mol) and DMAP (8 mg, 0.07 mmol) were added to
couple the peptide to resin-GABA D-Ser(tBu) D-
Ala D-Ser(tBu)-H (26 mg, 0.013 mmol) at room tempera-
ture for 48 h in the presence of 5 eq. DIEA. The protected
peptide was cleaved off the resin with 5% TFA/DCM. After
HPLC purification, treatment with 50% TFA/DCM at room
temperature for 2 h afforded the desired peptide–AMC
conjugate, GABA D-Ser D-Ala D-Ser mGly-Gln-AMC
using
a fluorescence standard curve generated with
(
10), in 21% yield (1.5 mg, > 99% pure by LC-UV analy-
known amounts of 7-AMC.
+
+
sis). HRMS (ESI ) m/z calc’d for C H N O [M + H] :
31 42 7 13
7
20.2841, found: 720.2869.
Measurement of non-PSA-mediated proteolysis in
plasma
Synthesis of glutaryl-Ser-Ala-Ser-Chg-Gln-AMC
11)
Fmoc-Ser(tBu)-OH (8.8 mg, 0.023 mmol) was preactivated
with HBTU (7.7 mg, 0.020 mmol) in the presence of 5 eq.
DIEA in 200 lL of NMP for 30 min at room temperature
Non-PSA-mediated hydrolysis of peptide conjugates was
determined following a procedure similar to the PSA
enzyme assay described above. Briefly, each peptide solu-
tion (4 lL) was added to 36 lL of phosphate buffer,
human, or mouse plasma (final peptide concentration
1 lM) in a white Costar 384-well plate (Corning, Inc., Corn-
ing, NY, USA). The stability of peptide conjugates in
100 mM phosphate buffer, pH 7.4 was evaluated by mea-
suring the fluorescence of 1 lM solutions throughout the
course of the assay. Sample evaporation rate was also
measured by adding mixtures of 36 lL of plasma and
4 lL of 100 mM phosphate to separate wells in triplicate
and recording fluorescence increase over time. All wells
were read using a 335/460 nm (ex/em) filters on a Victor
3V 1420 Multilabel Counter (PerkinElmer Life and Analyti-
cal Sciences) over a period of 24 h.
(
under nitrogen atmosphere and then coupled to H-Ala-Ser
Me,Me
(
Ψ
pro)-Chg-Gln-AMC (9.6 mg, 0.15 mmol) for 2 h.
Fmoc-deprotection was carried out without purification by
adding 20 lL of DEA to the reaction mixture and allowing
the reaction to proceed for 90 min at room temperature.
The reaction mixture was washed with 2 mL of hexane (five
times), dried under reduced pressure (to remove residual
DEA), and reconstituted in 200 lL of NMP. The reconsti-
tuted mixture was reacted with glutaric anhydride (5.1 mg,
0
.045 mmol) for 2 h at room temperature under nitrogen
atmosphere. Solvents were removed under reduced pres-
8
42
Chem Biol Drug Des 2015; 86: 837–848

Improving Specificity of PSA Substrate
Results and Discussion
Prodrug I featuring the peptide sequence glutaryl-Hyp-Ala-
Ser-Chg-Gln, a highly specific PSA substrate, coupled to
doxorubicin via a Ser-Leu linker, was evaluated in clinical
studies as a PSA-activated prodrug for the treatment of
advanced prostate cancer (24). However, its development
was discontinued probably due to toxicity concerns stem-
ming from significant systemic levels of the active metabo-
lite Leu-doxorubicin as well as free doxorubicin following
prodrug I administration (25). Subsequent metabolism
studies of prodrug I suggested that in vivo non-PSA-medi-
ated hydrolysis may be responsible for the extra-prostatic
release of doxorubicin (24,25,30). In the present study, we
further optimized the aforementioned peptide sequence to
improve PSA specificity. Sequence optimization was car-
ried out using the convenient release of fluorescent 7-
AMC from peptide conjugates following PSA cleavage of
the Gln-AMC bond. Thus, we synthesized a series of par-
tial D-retro-inverso-peptide conjugates which were evalu-
ated as PSA substrates.
Our general strategy for optimizing the sequence glutaryl-
Hyp-Ala-Ser-Chg-Gln involved systematic D-amino acid
substitution in P2-P5 using either urea or malonate (mGly)
linkage to generate various 7-AMC peptide conjugates
which were evaluated as PSA substrates (Figure 1). The D-
amino acid C-terminal ends of the peptide conjugates 2–10
were capped with c-aminobutyric acid (GABA) in lieu of glut-
aryl, and Hyp was replaced by Ser. For consistency and the
convenience of capping P5 D-amino acids through amide
bond formation when a linkage is inserted between P3 and
P5, glutaryl was replaced by GABA. In addition, Ser was
inserted in P5 to remove the conformational restriction intro-
duced by Hyp. Where urea linkage was used, peptides were
synthesized by a combination of solid-phase and solution-
phase chemistry using an Fmoc strategy. Urea linkages
(AMC conjugates 2–5) were inserted into the peptide
sequence of Ser-Ala-Ser-Chg-Gln by first activating the
amino end of the appropriate resin-bound peptide interme-
diate with carbonyldiimidazole (CDI) at room temperature,
followed by coupling of the corresponding AMC peptide
conjugate. The synthesis of conjugate 2 is shown in
Scheme 1 as an example. Due to challenges encountered
with the HPLC purification of final peptide conjugates, the t-
Figure 1: Structures of peptide–AMC conjugates designed and
synthesized. Glutaryl-Hyp-Ala-Ser-Chg-Gln-AMC (1), GABA D-
Ser D-Ala D-Ser D-Chg-w[NH-CO-NH]-Gln-AMC
(2),
GABA D-Ser D-Ala D-Ser-w[NH-CO-NH]-Chg-Gln-AMC (3),
GABA D-Ser D-Ala-w[NH-CO-NH]-Ser-Chg-Gln-AMC
GABA D-Ser-w[NH-CO-NH]-Ala-Ser-Chg-Gln-AMC
(4),
(5),
GABA mGly-Ala-Ser-Chg-Gln-AMC (6), GABA D-Ser mGly-
Ser-Chg-Gln-AMC (7), GABA D-Lys mGly-Ser-Chg-Gln-AMC
(
8), GABA D-Ser D-Ala mGly-Chg-Gln-AMC (9), GABA D-
Ser D-Ala D-Ser mGly-Gln-AMC (10), Glutaryl-Ser-Ala-Ser-
Chg-Gln-AMC (11).
Chem Biol Drug Des 2015; 86: 837–848
843

Aloysius and Hu
Scheme 2: Synthesis of peptide–AMC conjugate 7.
butyl-protected conjugates were purified to homogeneity
prior to TFA deprotection. To introduce the malonate linkage
featured in the sequences of conjugates 6–10, several con-
ditions were explored. For example, attempts to couple the
peptide, Ser(tBu)-Chg-Gln-AMC to resin-GABA D-Ser
selenocarboxylate of Boc-Gln with 7-azido-4-methylcouma-
rin (AZD), followed by TFA-detprotection of the Boc-group
at room temperature. As TFA can be readily activated by
HBTU in subsequent steps and couple to Gln-AMC, residual
TFA was removed using Amberlyst A-26.
(
tBu)-H to generate conjugate 7 were unsuccessful when
malonic acid was first activated with HBTU (or EDC) and
reacted with resin-GABA D-Ser(tBu)-H. The lack of reac-
tivity of activated malonic acid toward the amino group of
resin-bound peptides was surprising because malonic acid
could be effectively activated and trapped with benzylamine
in solution. The acylation of sterically hindered alcohols
through the ketene intermediate of malonic acid in the pres-
ence of dehydrating agents was previously reported (31). In
our case, the OBt-activated ester, acyl urea, or ketene inter-
mediate generated from malonic acid was anticipated to
react more readily with amino groups as compared to steri-
cally hindered alcohols. However, as the ketene intermedi-
ate is reactive and short-lived, its coupling with the amino
group on solid phase may not be the major pathway. For
example, the ketene could be trapped by EDC through a
To evaluate their PSA cleavage rates, all AMC conjugates
were incubated with human PSA and fluorescence mea-
sured over a period of 3 h (Table 1 and Figure 2). Using
the standard curve generated from known amounts of free
7-AMC, the rate of 7-AMC liberated upon PSA-mediated
cleavage of the Gln-AMC was determined from the slope
of the linear portions of the fluorescence counts–time
curve for each conjugate; fluorescence increase was insig-
nificant for Boc-Gln-AMC up to a concentration of 1 lM
(data not shown). From the urea-containing AMC conju-
gates synthesized, conjugate 5 exhibited modest substrate
specificity for PSA with a 7-AMC formation rate of 92
pmol/min/100 nmol PSA. As shown in Table 1, AMC con-
jugates 2–4 and 8–10 were poor substrates for PSA and
did not generate any significant amount of fluorescence.
[4+2] cycloaddition mechanism similar to DCC (31). Opti-
mized conditions involved activation of the fluorenylmethyl
malonate mono-ester with EDC in the presence of 2.5 eq.
DMAP to afford the mono-amide which was spontaneously
deprotected upon prolonged exposure to DMAP
The results in Table 1 and Figure 2 indicate that, although
amino acid side-chain orientation was maintained with D-
amino acid substitution (as compared to the original pep-
tide sequence of prodrug I), introducing urea in place of
any amide bond between P1 and P5 may have disrupted
critical peptide substrate binding interactions with PSA
such as backbone H-bonding and side-chain interactions.
Our findings support previous optimization studies which
suggested that further amino acid substitution in P1-P4 for
the optimized sequence of prodrug I may not be well toler-
ated and thereby compromise substrate specificity for PSA
(25). Our observations are consistent with M eꢀ nez and col-
(Scheme 2). Conjugates 6, 8–10 were synthesized following
this approach. Peptide segments were directly coupled to
Gln-AMC to synthesize the final AMC conjugates. The Gln-
AMC intermediate was obtained using our previously
reported methodology for the convenient preparation of
aminoacyl-AMC peptide conjugates (28) because standard
coupling methods using HBTU, IPCF, DCC, or PyBOP affor-
ded low product yields (often < 10%). Boc-Gln-AMC was
generated in excellent yield (84%) by the amidation of the
leagues analysis of key PSA substrate binding interactions
8
44
Chem Biol Drug Des 2015; 86: 837–848

Improving Specificity of PSA Substrate
Table 1: PSA-mediated hydrolysis of AMC peptide conjugates
7-AMC Fluorescence
PSA cleavage rate
Entry
Peptide–AMC conjugate
(counts/min)
(pmol/min/100 nmol PSA)
1
2
Glutaryl-Hyp-Ala-Ser-Chg-Gln-AMC (1)
GABA D-Ser D-Ala D-Ser D-Chg-w[NH-CO-NH]-
Gln-AMC (2)
156
<1
171
<1
3
4
5
GABA D-Ser D-Ala D-Ser-w[NH-CO-NH]-
Chg-Gln-AMC (3)
GABA D-Ser D-Ala-w[NH-CO-NH]-Ser-Chg-
Gln-AMC (4)
GABA D-Ser-w[NH-CO-NH]-Ala-Ser-Chg-
Gln-AMC (5)
<1
<1
84
<1
<1
92
6
7
8
9
1
1
GABA mGly-Ala-Ser-Chg-Gln-AMC (6)
GABA D-Ser mGly-Ser-Chg-Gln-AMC (7)
GABA D-Lys mGly-Ser-Chg-Gln-AMC (8)
GABA D-Ser D-Ala mGly-Chg-Gln-AMC (9)
GABA D-Ser D-Ala D-Ser mGly-Gln-AMC (10)
Glutaryl-Ser-Ala-Ser-Chg-Gln-AMC (11)
205
37
<1
<1
<1
225
41
<1
<1
<1
0
1
320
351
Conditions: Each peptide conjugate (1 lM) was subject to human PSA (100 nM) in 50 mM Tris buffer, pH 8 containing 2 mM CaCl
2
and
0
.1% Tween-20 in triplicate. Fluorescence was measured using a 335/460 (ex/em) filter over a period of 3 h. PSA cleavage rates were
computed by taking the slopes of the linear portions of fluorescence counts–time curves (see Figure 2) and using a standard curve gener-
ated with known amounts of 7-AMC.
Figure 2: PSA-mediated hydrolysis of AMC
peptide conjugates (PSA substrates).
using the first reported crystal structure of PSA in complex
with the monoclonal antibody 8G8F5 and the fluorogenic
semenogelin I-derived substrate Mu-Lys-Gly-Ile-Ser-Ser-
Gln-Tyr-AFC. Their findings suggest tight binding interac-
tions in the P1-P4 of the substrate within the active site of
PSA (32); thus it is not surprising that modifications in this
region of the substrate sequence translate to significant
changes in binding affinity. We also explored the strategic
insertion of malonate in place of one amino acid between
D- and L-amino acids in the peptide sequence of prodrug
I. Although P2-P5 substitution decreased PSA cleavage
rate, it was gratifying to find that peptide conjugate 6
showed a significant improvement over conjugate 1 (repre-
sentative sequence of prodrug I). Furthermore, AMC con-
jugate 6 exhibited the second fastest cleavage rate of
225 pmol/min/100 nmol PSA (Table 1, Figure 2). Introduc-
tion of urea linkages in the peptide sequence of AMC con-
jugate 1 increased the peptide chain length by one carbon
atom, whereas malonate preserves the original sequence
length but deleted the side chain of amino acid it replaces.
Moreover, the malonate linkage imparts less rotational
restriction to the peptide chain overall compared to urea.
Chem Biol Drug Des 2015; 86: 837–848
845

Aloysius and Hu
The lower rotational conformation restriction is demon-
strated with AMC conjugate 6 as the malonate linkage is
moved farther away from P1; only modest effects on PSA
cleavage rate were observed with the corresponding urea-
containing AMC conjugate 5 (Figure 2). The fastest cleav-
age rate of 351 pmol/min/100 nmol PSA was observed
with AMC conjugate 11. It is possible that the bend and
rigidity introduced by Hyp at P5 in conjugate 1 compro-
mised optimal substrate binding. Moreover, for conjugate
gates with PSA cleavage rates > 1 pmol/min/100 nmol
PSA were incubated with mouse or human plasma for
24 h and 7-AMC formation monitored by fluorescence.
Although 7-AMC formation was insignificant during the first
3 h of incubation, Gln-AMC hydrolysis increased to detect-
able levels (up to 10 nM) at 24 h, suggesting that non-
PSA-mediated hydrolysis did occur. The apparent delay in
7-AMC formation observed in Figure 4 is due to the fact
that non-specific endopeptidase cleavage must first occur
at scissile bonds other than the Gln-AMC before hydrolysis
of the Gln-AMC bond via subsequent aminopeptidase
action. Furthermore, all peptide conjugates exhibited
greater non-PSA-mediated hydrolysis in mouse plasma
than in human plasma (Figures 3 and 4) suggesting the
involvement of multiple proteases or isoforms of the same
protease in conjugate instability.
1
1, the hydroxyl of Ser may have contributed to additional
stabilizing H-bonding interactions in the active site of PSA.
Overall, we confirmed that PSA specificity was lost with
modifications in P2-P5 as suggested by previous substrate
optimization studies. Additionally, we showed that PSA
cleavage rate was maintained or improved by up to 2-fold
with the P5 substitutions.
To further assess the PSA specificity of the peptide
sequences generated, we measured conjugate stability in
mouse and human plasma (Table 2, Figure 3). The
selected peptide conjugates were examined in mouse
plasma because prodrugs generated from the peptide
substrate with the highest PSA specificity could further be
evaluated in nu/nu mouse PK/PD studies. Peptide conju-
Conjugates 5 and 6 were most stable in mouse plasma
indicating that urea and malonate substitutions in
P4-P5 effectively mitigated hydrolysis mediated by a yet
unidentified protease at scissile bond between P4 and
P5. P5 substitution appears to block cleavage by this
protease, the major contributor to non-PSA-mediated
hydrolysis for conjugates 1, 7, and 11 in mouse plasma.
Table 2: Non-PSA-mediated hydrolysis of AMC peptide conjugates in mouse and human plasma at 24 h
-AMC generated in
7
7-AMC generated in
human plasma (nM)
Entry
Peptide–AMC conjugate
mouse plasma (nM)
1
2
Glutaryl-Hyp-Ala-Ser-Chg-Gln-AMC (1)
GABA D-Ser-w[NH-CO-NH]-Ala-Ser-
Chg-Gln-AMC (5)
9.7
0
1.1
1.1
3
4
5
GABA mGly-Ala-Ser-Chg-Gln-AMC (6)
GABA D-Ser mGly-Ser-Chg-Gln-AMC (7)
Glutaryl-Ser-Ala-Ser-Chg-Gln-AMC (11)
4.7
8.1
2.4
1.4
0.98
0.82
Conditions: Each peptide conjugate (1 lM) was incubated in mouse or human plasma. Fluorescence was measured using a 335/460 (ex/
em) filter over a period of 24 h (see Figure 3).
Figure 3: Stability of AMC peptide
conjugates in mouse plasma.
8
46
Chem Biol Drug Des 2015; 86: 837–848

Improving Specificity of PSA Substrate
Figure 4: Stability of AMC peptide
conjugates in human plasma.
Although conjugate 5 was completely stable in mouse
plasma over 24 h, it did exhibit minor instability in human
plasma. Moreover, non-PSA-mediated hydrolysis appears
to be comparable for all conjugates in human plasma
for their human PSA cleavage rate, PSA-mediated cytotox-
icity and resistance to hydrolysis in biological matrices lack-
ing active PSA.
(Figure 4). Nevertheless, differences in hydrolysis rate for
conjugates 5 and 6 in mouse plasma suggest that at
least two additional proteases with different cleavage
maps may be responsible for the residual hydrolysis
observed for conjugate 5 in human plasma and conju-
gate 6 in both matrices. The protease(s) responsible for
the non-PSA-mediated hydrolysis of the sequences
examined herein and their PSA specificity are being fur-
ther investigated in additional studies by coupling the
optimized sequence GABA mGly-Ala-Ser-Chg-Gln to
select cytotoxic agents. It should be noted that the
extent of hydrolysis of conjugate 6 in mouse plasma
Acknowledgments
We gratefully acknowledge the financial support of a
research grant from the Elsa U. Pardee Foundation, a pilot
grant from the Gallo Prostate Cancer Center of the Cancer
Institute of New Jersey, and grant RSG-03-004-01-CDD
from the American Cancer Society.
Conflict of Interest
(
2.42 nM eq. 7-AMC over 24 h) was much lower as com-
The authors declare that they have no conflict of interest
that would influence their objectivity.
pared to conjugate 1 (9.74 nM eq. AMC over 24 h), fur-
ther suggesting that modifications at P5 also improved
peptide stability.
References
In summary, we carried out D-amino acid substitutions
between P1 and P4 and incorporated either urea or malonic
acid linkage between the D- and L-amino acids to improve
the specificity of the PSA substrate, glutaryl-Hyp-Ala-Ser-
Chg-Gln. Our results suggest that modifications between
P1 and P4 would adversely affect PSA specificity while sub-
stitutions at P5 improved PSA cleavage rate and resistance
to non-PSA-mediated hydrolysis. To assess the applicability
of the optimized sequences to peptide prodrug design, the
PSA specificity of the optimized sequences (conjugates 6
and 11) is being further investigated in additional studies. To
further evaluate the utility of our newly designed sequences,
glutaryl-Ser-Ala-Ser-Chg-Gln and GABA mGly-Ala-Ser-
Chg-Gln, as promoieties with improved PSA cleavage rate
and stability to non-PSA-mediated metabolism, they will be
coupled to doxorubicin or phosphoramide mustard using
appropriate linkers. The resulting prodrugs will be assessed
1. Brawley O.W. (2012) Prostate cancer epidemiology in
the United States. World J Urol;30:195–200.
2. Siegel R., Ma J., Zou Z., Jemal A. (2014) Cancer statis-
tics, 2014. CA-Cancer J Clin;64:9–29.
3. Robson M., Dawson N. (1996) How is androgen-
dependent metastatic prostate cancer best treated?
Hematol Oncol Clin North Am;10:727–747.
4. Kelland L.R. (2000) An update on satraplatin: the first
orally available platinum anticancer drug. Expert Opin
Investig Drugs;9:1373–1382.
5. Coleman R.E. (2008) Risks and benefits of bisphosph-
onates. Br J Cancer;98:1736–1740.
6. De Dosso S., Berthold D.R. (2008) Docetaxel in the
management of prostate cancer: current standard of
care and future directions. Expert Opin Pharmacoth-
er;11:1969–1979.
Chem Biol Drug Des 2015; 86: 837–848
847

Aloysius and Hu
7
. Jakobovits A. (2008) Monoclonal antibody therapy
for prostate cancer. Handb Exp Pharmacol;181:
22. Denmeade S.R., Nagy A., Gao J., Lilja H., Schally
A.V., Isaacs J.T. (1998) Enzymatic activation of a
doxorubicin-peptide prodrug by prostate-specific anti-
gen. Cancer Res;58:2537–2540.
23. Jiang Y., Hu L. (2007) Phenylalanyl-aminocyclophos-
phamides as model prodrugs for proteolytic activation:
synthesis, stability, and stereochemical requirements for
enzymatic cleavage. Bioorg Med Chem Lett;17:517–
521.
237–256.
8
. Tu S.M., Lin S.H. (2008) Current trials using bone-tar-
geting agents in prostate cancer. Cancer J;4:35–39.
. Amato R.J., Sarao H. (2006) A phase I study of paclit-
axel/doxorubicin/thalidomide in patients with androgen-
independent prostate cancer. Clin Genitourin Can-
cer;4:281–286.
9
1
1
0. Petrylak D.P. (2006) The treatment of hormone-refrac-
tory prostate cancer: docetaxel and beyond. Rev
Urol;8:S48–S55.
24. DiPaola R.S., Rinehart J., Nemunaitis J., Ebbinghaus
S., Rubin E., Capanna T., Ciardella M. et al. (2002)
Characterization of a novel prostate-specific antigen-
activated peptide-doxorubicin conjugate in patients
with prostate cancer. J Clin Oncol;20:1874–1879.
25. Wong B.K., DeFeo-Jones D., Jones R.E., Garsky
V.M., Feng D.M., Oliff A., Chiba M., Ellis J.D., Lin J.H.
(2001) PSA-specific and non-PSA-specific conversion
of a PSA-targeted peptide conjugate of doxorubicin to
its active metabolites. Drug Metab Dispos;29:313–318.
26. Desai R.B., Schwartz M.S., Matuszewski B.K. (2004)
The identification of three human metabolites of a pep-
tide-doxorubicin conjugate using HPLC-MS-MS in
positive and negative ionization modes. J Chromatogr
Sci;42:317–322.
27. Schwartz M.S., Matuszewski B.K. (2002) Determina-
tion of a peptide-doxorubicin, prostate-specific antigen
activated prodrug, and its active metabolites in human
plasma using high-performance liquid chromatography
with fluorescence detection. Stabilization of the pep-
tide prodrug with EDTA. J Chromatogr B;780:171–
182.
28. Wu X., Chen Y., Aloysius H., Hu L. (2011) A novel high-
yield synthesis of aminoacyl p-nitroanilines and amino-
acyl 7-amino-4-methylcoumarins: important synthons
for the synthesis of chromogenic/fluorogenic protease
substrates. Beilstein J Org Chem;7:1030–1035.
29. Klayman D.L., Griffin T.S. (1973) Reaction of selenium
with sodium borohydride in protic solvents. A facile
method for the introduction of selenium into organic
molecules. J Am Chem Soc;95:197–199.
1. Tannock I.F., Osoba D., Stockler M.R., Ernst D.S.,
Neville A.J., Moore M.J., Armitage G.R., Wilson J.J.,
Venner P.M., Coppin C.M., Murphy K.C. (1996) Che-
motherapy with mitoxantrone plus prednisone or pred-
nisone alone for symptomatic hormone-resistant
prostate cancer: a Canadian randomized trial with palli-
ative end points. J Clin Oncol;14:1756–1764.
2. Atkinson J.M., Siller C.S., Gill G.H. (2008) Tumour en-
doproteases: the cutting edge of cancer drug deliv-
ery?. Br J Pharmacol;153:1344–1352.
1
1
1
1
1
1
1
3. Choi K.Y., Swierczewska M., Lee S., Chen X. (2012)
Protease-activated drug development. Theranos-
tics;2:156–178.
4. Yang Y., Aloysius H., Inoyama D., Chen Y., Hu L.
(2011) Enzyme-mediated hydrolytic activation of pro-
drugs. Acta Pharm Sin B;1:143–159.
5. Aloysius H., Hu L. (2015) Targeted prodrug
approaches for hormone refractory prostate cancer.
Med Res Rev;35:554–585.
6. Diamandis E. (1995) New diagnostic applications and
physiological functions of prostate specific antigen.
Scand J Clin Lab Invest;55:105–112.
7. Levesque M., Yu H., D’Costa M., Diamandis E. (1995)
Prostate specific antigen expression by various tumors.
J Clin Lab Anal;9:123–128.
8. Yu H., Diamandis E., Sutherland D.J.A. (1994) Immu-
noreactive prostate specific antigen levels in female
and male breast tumors and its association with ste-
roid hormone receptors and patient age. Clin Bio-
chem;27:75–79.
30. Garsky V.M., Lumma P.K., Feng D.M., Wai J., Ramjit
H.G., Sardana M.K., Oliff A., Jones R.E., DeFeo-Jones
D., Freidinger R.M. (2001) The synthesis of a prodrug
of doxorubicin designed to provide reduced systemic
1
9. Otto A., Bar J., Birkenmeier G. (1998) Prostate specific
antigen forms complexes with human a2-macroglobu-
lin and binds to the a2-macroglobulin receptor/ldl
receptor-related protein. J Urol;159:297–303.
toxicity and greater target efficacy.
J
Med
Chem;44:4216–4224.
2
0. Leinonen J., Zhang W., Stenman U. (1996) Complex
formation between PSA isoenzymes and protease
inhibitors. J Urol;155:1099–1103.
31. Shelkov R., Nahmany M., Melman A. (2002) Acylation
through ketene intermediates. J Org Chem;67:8975–
8982.
2
1. DeFeo-Jones D., Garsky V.M., Wong B.K., Feng D.M.,
Bolyar T., Haskell K., Kiefer D.M. et al. (2000) A peptide-
doxorubicin prodrug activated by prostate-specific anti-
gen selectively kills prostate tumor cells positive for
prostate-specific antigen in vivo. Nat Med;6:1248–1252.
32. M eꢀ nez R., Michel S., Muller B.H., Bossus M., Ducan-
cel F., Jolivet-Reynaud C., Stura E.A. (2008) Crystal
structure of a ternary complex between human pros-
tate-specific antigen, its substrate acyl intermediate
and an activating antibody. J Mol Biol;376:1021–1033.
8
48
Chem Biol Drug Des 2015; 86: 837–848