Synthesis and enzymatic activation of N-[Nα-(4-amino-4-deoxypteroyl)-Nδ- hemiphthaloyl-L-ornithiny]-L-phenylalanine, a candidate for Antibody-Directed Enzyme Prodrug Therapy (ADEPT)

Article abstract of DOI:10.1016/S0968-0896(01)00298-X

N-[N^α-(4-Amino-4-deoxypteroyl)-N^δ- hemiphthaloyl-L-ornithinyl]-L-phenylalanine (1), a carboxypeptidase A (CPA) cleavable prodrug was synthesized for use in an antibody directed strategy to improve the therapeutic selectivity of N

Full text of DOI:10.1016/S0968-0896(01)00298-X

Bioorganic & Medicinal Chemistry 10 (2002) 493–500
Synthesis and Enzymatic Activation of N-[Nꢀ-(4-Amino-4-
deoxypteroyl)-Nꢁ-hemiphthaloyl-L-ornithiny]-L-phenylalanine, a
Candidate for Antibody-Directed Enzyme Prodrug Therapy
(ADEPT)
Joel E. Wright* and Andre Rosowsky
Dana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School,
Boston, MA, USA
Received 25 June 2001; accepted 7August 2001
Abstract—N-[Na-(4-Amino-4-deoxypteroyl)-Nd-hemiphthaloyl-l-ornithinyl]-l-phenylalanine (1), a carboxypeptidase A (CPA)
cleavable prodrug was synthesized for use in an antibody directed strategy to improve the therapeutic selectivity of Na-(4-amino-4-
deoxypteroyl)-Nd-hemiphthaloyl-l-ornithine (2), an extremely potent nonpoly-glutamatable DHFR inhibitor which is also highly
cytotoxic. Compound 1 was shown by HPLC analysis to give a >99% yield of 2 upon incubation with bovine CPA (bCPA) for 20
min at 25 ꢀC. In a spectrophotometric kinetic assay with 50 mM dihydrofolate as the competing substrate in the presence of 65 mM
NADPH, 1+bCPA stoichiometrically inhibited recombinant human DHFR (rhDHFR) with a Ki of 0.35 pM. In contrast, 1
without bCPA was a poor inhibitor of rhDHFR (Ki>10 mM). In a 72 h growth inhibition assay against cultured CCRF-CEM
human leukemic lymphoblasts, the growth inhibitory activities of 1+bCPA, 2+bCPA, and 2 alone were the same (IC50 1.3–1.4
nM), whereas 1 in the absence of bCPA was >100-fold less potent (IC50 155 nM). # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
Chemotherapeutic specificity is achieved by conjugating
the enzyme or abzyme to a tumor selective antibody or
antibody fragment via a linkage that preserves the
activities of both the targeting moiety (the antibody)
and the activating moiety (the enzyme). Activating and
targeting components may be linked by a variety of
synthetic chemical methods8 or assembled by recombi-
nant techniques as bifunctional fusion proteins.9 The lat-
ter approach can provide a reproducibly homogeneous
product in high yield.10
Antibody-directed enzyme prodrug therapy (ADEPT)1,2
is a strategy for improving the specificity of cancer
chemotherapy by surrounding tumor cells with a con-
centrated pool of a diffusible cytotoxic drug.3 The
ADEPT approach utilizes a weakly- or non-immuno-
genic exogenous macromolecule, typically an enzyme, to
catalyze the activation of a prodrug. Optimally, the
prodrug should not be a good substrate for human
enzymes.4
Some years ago, Huennekens and co-workers11 synthe-
sized a-l-phenylalanyl MTX (MTX-l-Phe), the first
example of an ADEPT based antifolate prodrug. MTX-
l-Phe was rapidly cleaved by a CPA-antibody con-
jugate, giving a high yield of MTX. The cytotoxicity of
MTX-l-Phe against UCLA-P3 human lung adeno-
carcinoma cells improved 350-fold in the presence of
conjugated bovine CPA (bCPA).11 However, the IC50 of
MTX against these cells, 28 nM, is higher than that of a
number of other antifolates that could have been used
for this purpose, and also is large in comparison
with other drugs that have been used to make ADEPT
prodrugs.12
A prokaryotic, genetically altered or chemically mod-
ified enzyme may serve as the activating catalyst.5 In a
novel variant of this approach, recent studies have
focused on abzymes, which are catalytic antibodies
capable of promoting reactions not ordinarily catalyzed
by natural enzymes.6,7
*Corresponding author at: 44 Binney Street, Boston, MA 02115, USA.
Tel.: +1-617-632-3123; fax: +1-617-632-3124; e-mail: joel_wright@
dfci.harvard.edu
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00298-X

494
J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
In this paper we describe the synthesis of a derivative of
Na-(4-amino-4-deoxypteroyl)-Na-hemiphthaloyl-l-orn-
ithine (2, PT523) and its testing as a possible ADEPT
prodrug. Compound 2 is a non-polyglutamatable ana-
logue of aminopterin that was discovered in our
laboratory and is currently in advanced preclinical test-
ing in preparation for a phase I trial. It is one of the
most potent known antifolates, with IC50 values in the
low to subnanomolar range against a wide variety of
tumor cells. In competition with MTX, the affinity of 2
for the reduced folate carrier is 10-fold greater. In the
presence of NADPH with dihydrofolate as the compet-
ing substrate, its affinity for human DHFR exceeds that
of MTX by 15-fold. The inability of 2 to form poly-
glutamates is due to the nearly absolute specificity of
folylpolyglutamate synthetase for substrates with a glu-
tamyl side chain. The 72-h cytotoxicity of 2 toward cul-
tured CCRF-CEM human lymphoblastic leukemia cells
surpasses that of MTX by 10-fold.13
group by hydrogenolysis over 10% Pd–C in glacial
AcOH for 24 h under 60 psi of H2 pressure unexpect-
edly led to reduction of the aromatic ring in the Phe
moiety, affording Nd-Boc-l-Orn-l-(b-cyclohexyl)alanine
tert-butyl ester (7) in 85% yield. However, over-reduc-
tion could be avoided by diluting the AcOH with two
volumes of MeOH, shortening the reaction time to 3 h
and decreasing the H2 pressure to 30 psi. These less
vigorous conditions led to Nd-Boc-l-Orn-l-Phe tert-
butyl ester (8) in 86% yield (Scheme 1). Compounds 7
and 8 were clearly distinguishable on the basis of their
1
C,H,N analyses and H NMR spectra.
Compound 1 was obtained from 4-amino-4-deoxy-
pteroic acid (9) as shown in Scheme 2. Two strategies
were investigated. The first, consisting of five steps, was
predicated on the expectation that regiospecific phtha-
loylation would require N10 protection. Thus, 9 was
treated with 98% HCO2H in the presence of a catalytic
amount of 4-N,N-dimethylaminopyridine (DMAP) to
obtain 4-amino-4-deoxy-10-formylpteroic acid (10), as a
tetrahydrate, in 99% yield. This compound was pre-
viously obtained in somewhat lower yield by reaction of
9 with a mixture of 98% HCO2H and Ac2O.15 Coupling
Of the many possible derivatives of 2 that would be
suitable as candidate ADEPT prodrugs, we chose to
synthesize and study N-[Na-(4-amino-4-deoxypteroyl)-
Nd-hemiphthaloyl-l-ornithiny]-l-phenylalanine (1). Our
aim was to obtain a prodrug with CPA substrate activ-
ity similar to that of MTX-l-Phe. The structures of 1
and 2 are shown in Figure 1.
Results
Chemistry
Compound 8, the dipeptide precursor of 1 was synthe-
sized according to Scheme 1. The protected amino acids
Na-Cbz-Nd-Boc-l-Orn (3), Na-Cbz-Nd-Boc-l-Orn penta-
fluorophenyl ester (4), and l-Phe tert-butyl ester (5)
were prepared in ꢁ85% yield by standard peptide
synthesis methods.14 Condensation of 3 and 4 in the
presence of N-methylmorpholine (NMM) afforded the
blocked dipeptide 6 in 93% yield. Removal of the Cbz
Figure 1. Structures of Na-(4-amino-4-deoxypteroyl)-N?-hemi-
phthaloyl-l-ornithinyl-l-phenylalanine (1) and Na-(4-amino-4-deox-
ypteroyl)-N@-hemiphthaloyl-l-ornithine (PT523, 2).
Scheme 1. Synthesis of 8, the dipeptide precursor of Na-(4-amino-4-deoxypteroyl)-N@-hemiphthaloyl-l-ornithinyl-l-phenylalanine (1).

J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
495
of 8 and 10 in the presence of diethyl phosphor-
i
described in the literature.11 The reactions were quen-
ched after 20 min by heating for 30 min at 95 ꢀC to
denature the enzyme. The extent of cleavage was mon-
itored by reversed phase HPLC. Elution with a 20 min
gradient of 5–25% v/v CH3CN buffered with 0.1 M
NH4OAc at pH 7.5 gave baseline separation of 1 and 2.
As expected from its less hydrophilic character, prodrug
1 had the greater retention volume, 18.8 mL, whereas 2
had a retention volume of 13.6 mL. The co-product, l-
phenylalanine was not observed at 370 nm, the detector
ocyanidate (DEPC) and Pr2EtN16 led to an 87% yield
of 11, which on cleavage of the Boc group with 1:2
TFA–CH2Cl2 was converted to 12 as a TFA salt in 98%
yield. Acylation of 12 with phthalic anhydride in DMF
and iPr2EtN afforded the hemiphthaloyl derivative 13 in
50% yield, but removal of the 10-formyl group by
treatment with a stoichiometric amount of NaOH at
room temperature for 30 min gave a poor yield of 1,
<7%. Thus the overall yield for the five-step sequence
was only 3%. Removal of the 10-formyl group with
10% N2H2 in DMF was also tried without success.14
The second, and more concise route to 1 was made
possible by the subsequent discovery that acylation of
unprotected N10 did not occur when 1.1 molar equiva-
lents of phthalic anhydride were used. This eliminated
the problem of N10-formyl deprotection in the presence
of an N?-hemiphthaloyl group. Thus, as indicated in
Scheme 2, compound 12 was converted to the fully
deprotected dipeptide 14 in 61% yield by a one-pot
method in which the formyl group was cleaved with
ethanolic HCl and the Boc and tert-butyl ester groups
were cleaved with TFA. Acylation of 14 with phthalic
anhydride occurred in 24% yield. Thus the overall yield
of 1 from APA by this improved four-step route was
10%.
Elemental analyses (C, H, N) for 1 and its intermediates
1
were within ꢂ0.4% of calculated values. H NMR and
FABMS data were consistent with the desired struc-
tures. The target compound was >99% pure by HPLC.
Enzyme studies
Figure 2. Plot of rhDHFR activity versus concentration of 1 after 20
min incubation in the presence (triangles) or absence (circles) of bovine
carboxypeptidase A. DHFR substrates: 50 mM dihydrofolate and 65
mM NADPH. Data obtained by spectrophotometric titration at 340 nm.
Cleavage of 1 by bCPA was performed at 25 ꢀC in pH
7.4 Tris–HCl buffer in the presence of 10 mM ZnCl2 as
Scheme 2. Synthesis of 1 from 4-amino-4-deoxypteroic acid (APA, 9).

496
J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
wavelength chosen for its sensitivity to the pteridine
nucleus. After 20 min of exposure to bCPA the amount
of unchanged 1 was estimated to be <1%. The reten-
tion volume of the enzymatic cleavage product was the
same as that of compound 2.
CPA of this type has been reported by Smith and cow-
orkers.17 They conjugated the mutant enzyme with
ING-1, an antibody against Ep-CAM1, an epithelial
cell adhesion protein that is overexpressed in colorectal
and other tumor cells. The conjugate, ING1-hCPA1-
T268G selectively targeted HT-29 human colon adeno-
carcinoma cells and cleaved terminal Phe and Tyr from
MTX18 and other antifolates.19 Synthetic adducts with
bulky substituents on the aromatic ring of Phe or Tyr
were also good substrates.
Typical titration curves showing inhibition of rhDHFR
activity by 1 in the presence and absence of bCPA are
given in Figure 2. The Ki of 1+bCPA for competitive
inhibition of DHFR activity was not significantly dif-
ferent from that of 2, 0.35ꢂ0.15 pM. After incubation
with 1 in the absence of bCPA, rhDHFR retained
>98% of its native activity.
On an optimal therapeutic schedule, binding of the
conjugate to the tumor and clearance of nonbound
conjugate should precede administration of the
prodrug.20 Then wherever the circulating prodrug
encounters the antibody-enzyme conjugate, a local con-
centration gradient of cytotoxic metabolite should
form3 and enter the targeted cells. If the activated drug
molecule is small and diffusible, it should also enter
neighboring cells and produce a ‘bystander effect’.21
Cell growth assay
The ability of 1 to inhibit the proliferation of cultured
CCRF-CEM human leukemic lymphoblasts was deter-
mined in the presence and absence of bCPA. Median
effect curves for cell population growth in the presence
of 1 alone, 2 alone, 1+bCPA, and 2+bCPA are shown
in Figure 3. The IC50 values derived from the median
effect curves were in the 1.3–1.4 nM range for 2,
2+bCPA, and 1+bCPA. In contrast, the IC50 of 1 in
the absence of bCPA was 155 nM. These results indi-
cated that the enzyme was able to activate the prodrug
quantitatively and did not influence the activity of the
released drug as an inhibitor of cell proliferation.
This phenomenon has been demonstrated in vivo by 19
F
NMR imaging with a glucuronic acid acetal derivative
of 5-fluorouracil.22
Heterogeneous tumors commonly contain some cells
that only express targeted antigens sparingly. Surface
inaccessibility of targeted epitopes may also occur.
Bulky immunotoxins have sometimes failed to achieve
multilog cell reduction because of their inability to
achieve adequate delivery to the cell surface. This pro-
blem may be overcome by the ability of small diffusible
drug molecules to elicit a robust bystander effect.23
Discussion
Ideally, the use of 1 for repeated cycles of ADEPT
chemotherapy in human patients would require that a
‘humanized’ carboxypeptidase be used in order to pre-
vent immune reactions. A genetically engineered mutant
Taking these considerations into account, candidate
prodrugs for ADEPT need to exhibit the following
characteristics:12
1. The cytotoxicity of the prodrug itself must be less
than that of the parent drug. As reported here, the
cytotoxicity ratio of 1 to 2 (prodrug/activated drug) is
similar to those of several other recently reported
ADEPT prodrugs,4,12 though not as high as the 800-
fold ratio found with sterically hindered conjugates such
as 3-cyclopentyltyrosyl- and 3-cyclobutylphenylalanyl-
MTX.17ꢃ19
2. There should be adequate substrate activity at physi-
ological pH in the presence of the activating enzyme. As
reported in this paper, 1 is quantitatively converted to 2,
its active product, by bCPA, and in all likelihood would
be similarly activated by other enzymes that cleave
C-terminal Phe residues.
3. The prodrug should have long-term chemical stability
at pH 7.4. We have found that the decomposition of a 1
mM stock solution of 1 after 72 h at 37 ꢀC in pH 7 .4
buffer is <1% (data not shown).
4. The prodrug should have the ability to transit the
tumor vasculature and diffuse throughout the tumor.
Unlike large macromolecules that have been used in the
construction of immunoconjugates for targeted chemo-
therapy, 1 is a small diffusible molecule.
Figure 3. CEM cell growth inhibition curves for 2 incubated in the
presence (squares) or absence (diamonds) of carboxypeptidase A and 1
incubated in the presence (triangles) or absence (circles) of carboxy-
peptidase A.

J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
497
5. The prodrug should have suitable pharmacological
and pharmacokinetic properties. Although the pharma-
cokinetics and in vivo pharmacology of 1 itself have not
been investigated, the pharmacokinetics and tissue dis-
tribution characteristics of 2 have been studied in mice
and rats,24,25 and are not likely to differ greatly between
1 and 2.
pentafluorophenol and N,N0-dicyclohexylcarbodiimide
and the tert-butyl ester group (OtBu) of 5 with iso-
butylene and dry HCl(g).
Nꢀ-Cbz-Nꢁ-Boc-L-ornithinyl-L-phenylalanine tert-butyl
ester (6). A solution of of NMM (0.225 mL) in DMF
(15 mL) was added to l-phenylalanine tert-butyl ester
hydrochloride (0.516 g, 2 mmol). When the solid dis-
solved, 4 (1.07g, 2 mmol) was added. It likewise dis-
solved rapidly, but
a precipitate formed shortly
Summary
thereafter. After 15 min at 23 ꢀC the mixture was eva-
porated in vacuo at 27 ꢀC. The residue was triturated for
1h with 0.8 M NaHCO3 (15 mL). The precipitate was
filtered and washed with 20 mL each of 0.2 M KHSO4
and H2O. The filter cake was dried in vacuo at 23 ꢀC to
obtain 6 as a white solid (1.06 g, 93%): mp 121–122 ꢀC;
Compound 1 is chemically stable at physiologic pH and
temperature, is quantitatively cleaved by CPA at pH
7.4, and is at least 100-fold less cytotoxic than its acti-
vated product, 2. While these are promising preliminary
results, and the efficient cellular transport, tight DHFR
binding, and potent cytotoxicity of 2 combine to make 1
a promising candidate for ADEPT chemotherapy,
additional studies would be needed in order to determine
whether 1 would be useful as an prodrug in vivo. These
would include an assessment of its pharmacokinetics and
toxicology in a suitable animal model.
1
TLC (CHCl3/MeOH/AcOH, 90:9:1) Rf 0.70; H NMR
(CDCl3, 500 MHz) d 1.14–1.54 (m, 4H, b- and g-CH2),
1.31 (s, 9H, OtBu), 1.35 (s, 9H, Boc), 2.88 (m, 2H, d-
CH2), 2.93 (m, 2H, PhCH2CH), 3.93 (m, 1H, a-
CHCO2), 4.31 (m, 1H, a-CHCO), 4.99 (s, 2H,
PhCH2O), 7.17–7.33 (m, 10H, 2C6H5). Anal. calcd for
C31H43N3O7: C, 65.36; H, 7.61; N, 7.38. Found: C,
65.54, H, 7.67, N, 7.44.
It may also be noted that, as part of our continuing
studies on second-generation analogues of 2, several
compounds modified at positions 5, 8, and 10 have been
discovered that are superior to 2 in terms of chemical
stability, cellular uptake kinetics, DHFR inhibition, and
cytotoxicity.26,27 Because of the high potency of these
analogues, it would seem reasonable to view them as
alternatives to 2 in the context of ADEPT chemotherapy.
Nꢁ-Boc-L-ornithinyl-L-(ꢂ-cyclohexyl)alanine tert-butyl
ester (7). To a solution of 6 (0.75 g, 1.27 mmol) in gla-
cial AcOH (40 mL) were added 0.3 g of 10% Pd/C. The
mixture was shaken for 24 h under 60 psi of H2 pressure
in a Parr apparatus. The mixture was filtered through
Celite and the filtrate evaporated under reduced pres-
sure to a volume of 2 mL. It was stirred with Et2O (10
mL) and chilled overnight at 4 ꢀC. The precipitated
solid, consisting of the acetate salt of 7, was filtered,
washed with Et2O (5 mL) and hexanes (5 mL), and
dried in vacuo at 21 ꢀC. The acetate salt was dissolved in
CH2Cl2 (10 mL) the solution washed with 4 M K2CO3
(15 mL) and the aqueous layer back-extracted with
CH2Cl2 (2Â10 mL). The CH2Cl2 extracts were pooled
and dried over anhydrous Na2SO4. Evaporation under
reduced pressure gave an oil, which was recrystallized
from Et2O–hexanes to obtain the free base 7 as soft
white crystals (0.50 g, 85%): mp (dec) 74–75 ꢀC; TLC
(CHCl3/MeOH/AcOH, 90:9:1) Rf 0.52; 1H NMR
(DMSO-d6, 500 MHz) d 0.91–2.52 (m, 17H, CH2C6H11,
b- and g-CH2), 1.31 (s, 9H, OtBu), 1.35 (s, 9H, Boc)
2.90 (m, 2H, d-CH2), 4.39 (m, 1H, a-CHCO), 3.97(m,
Experimental
Compounds 1 and 6–14 were characterized by elemental
analysis (QTI Laboratories, Whitehouse, NJ, USA) and
500 MHz H NMR on a Varian model ML500 instru-
1
ment; 1 by HRFABMS on a Vacuum Generators Ltd
instrument, Manchester, UK (resolutionꢂ2.6 ppm) and
12–14 by FABMS on a Perkin-Elmer Voyager system
4036 (resolutionꢂ8.0 ppm). The purity of compounds 1
and 12–14 was confirmed by HPLC with a Waters
(Milford, MA, USA) automated gradient controller and
two model 510 pumps giving a combined flow rate of
1.0 mL/min. The mobile phase was 1 M NH4OAc pH
7.5 with a linear concentration gradient of 0–20% v/v
CH3CN for 20 min followed by isocratic 20% CH3CN
for 10 min. The column was a Waters Novapak 8NV4m
C-18 radial compression cartridge. A Waters model 484
tunable absorbance detector was operated at 370 nm.
TLC of compounds 6–8 and 11–13 was performed with
MK6F glass plates (Whatman Ltd., Maidenhead, UK).
Eluent compositions for TLC and silica gel column
chromatography are given below as volume ratios.
.
1H, a-CHCO2). Anal. calcd for C23H43N3O5 H2O: C,
60.10; H, 9.87; N, 9.14. Found: C, 60.13; H, 10.12, N,
9.18.
Nꢁ-Boc-L-ornithinyl-L-phenylalanine tert-butyl ester (8).
To a solution of 6 (1.0 g, 1.75 mmol) in glacial AcOH
(40 mL) and MeOH (80 mL) were added 0.3 g of 10%
Pd/C. The mixture was shaken for 3 h under 30 psi of
H2 pressure in a Parr apparatus, and was then filtered
and evaporated under reduced pressure to a volume 2
mL. After addition of Et2O (15 mL), the mixture was
chilled overnight at 4 ꢀC. The solid was filtered, washed
with EtOAc (5 mL), and dried in vacuo at 21 ꢀC for 18 h
to obtain the acetate salt of 8 as a white solid (1.74 g,
99%): mp 104–106 ꢀC. The acetate salt was dissolved in
CH2Cl2 (15 mL), the solution washed with 4 M K2CO3
Amino acid protection and activation
Standard peptide synthesis methods were used:14 start-
ing with Nd-Boc-l-ornithine (Bachem Bioscience, Inc.,
King of Prussia, PA, USA), the benzyloxycarbonyl
(Cbz) group of 3 was introduced with N-benzyloxy-
succinimide, the pentafluorophenyl ester group of 4 with

498
J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
(25 mL), and the aqueous layer back-extracted twice
with CH2Cl2 (2Â15 mL). The CH2Cl2 extracts were
pooled and dried over anhydrous Na2SO4. Evaporation
under reduced pressure gave an oil, which was dissolved
in dry Et2O (10 mL). The solution was chilled on an ice
bath and 6.6 M anhydrous ethereal HCl (0.3 mL) was
added. The precipitated solid was filtered, washed
immediately with dry Et2O (5 mL) and hexanes (15 mL)
at 25 ꢀC, and dried in vacuo overnight to obtain the HCl
salt of 8 as a white crystalline solid (740 mg, 86%): mp
(dec) 140–150 ꢀC; TLC (CHCl3/MeOH/AcOH, 90:9:1)
Rf 0.15; 1H NMR (DMSO-d6, 500 MHz) d 1.22–1.43 (m,
4H, b- and g-CH2), 1.31 (s, 9H, OtBu), 1.35 (s, 9H,
Boc), 2.93 (m, 2H, PhCH2), 3.04 (m, 2H, d-CH2), 3.11
(m, 1H, a-CHCO), 4.38 (m, 1H, a-CHCO2), 7.17–7.28
(m, 2H, PhCH2), 4.31 (m, 1H, a-CHCO2), 4.41 (m, 1H,
a-CHCO), 5.18 (s, 2H, 9-CH2), 7.12–7.86 (m, 9H, C6H5,
C6H4), 8.63 (s, 1H, C7-H), 8.78 (s, 1H, CH¼O). Anal.
.
calcd for C38H48N10O7 3/4H2O: C, 59.25; H, 6.48; N,
18.18. Found: C, 59.24, H, 6.50, N, 17.89.
Nꢀ-(4-Amino-4-deoxy-10-formylpteroyl)-L-ornithinyl-L-
phenylalanine (12). Compound 11 (120 mg, 0.15 mmol)
was dissolved in 4 mL of ice-cold CF3CO2H/CH2Cl2,
1:2 v/v and allowed to warm to 21 ꢀC. After 6 h the
solution was evaporated under reduced pressure. TLC
(silica gel, n-BuOH/AcOH/H2O, 3:1:1) showed the
absence of 11 (Rf 0.66) and formation of 12 (Rf 0.21) as
the only product. The residue was dissolved in CH2Cl2
(5 mL) and the solution added dropwise to rapidly stir-
red Et2O (100 mL). The precipitated solid was filtered
and washed with Et2O (10 mL) and hexanes (10 mL).
When exposed to room air the tan powder darkened
slightly and appeared to absorb atmospheric moisture.
Drying in vacuo at 50 ꢀC afforded 12 as a beige solid
(108 mg, 98%): mp (dec) >240 ꢀC; HPLC retention
volume: 26.5 mL; HRFABMS m/z [M+H]+ 601.2616
(C29H33N10O5 requires 601.2635). 1H NMR (DMSO-d6,
500 MHz) d 1.46–1.76 (m, 4H, b- and g-CH2), 2.89 (m,
2H, d-CH2), 2.92 (m, 2H, PhCH2), 4.39 (m, 1H, a-
CHCO), 4.48 (m, 1H, a-CHCO2), 5.32 (s, 2H, 9-CH2),
7.15–7.70 (m, 9H, C6H5, C6H4), 8.68 (s, 1H, C7–H), 8.70
.
.
(m, 5H, C6H5). Anal. calcd for C23H37N3O5 HCl H2O:
C, 56.37; H, 8.23; N, 8.57. Found: C, 56.13, H, 8.12, N,
8.48.
4-Amino-4-deoxy-10-formylpteroic acid (10). A suspen-
sion of 1.5 g, 4.0 mmol of 4-amino-4-deoxypteroic acid
3.5H2O (9)28 and DMAP (25 mg) in 98% HCO2H (100
mL) was stirred at 46 ꢀC for 14 h. The mixture was
cooled to 18 ꢀC and poured into rapidly stirred anhy-
drous Et2O (900 mL). The white precipitate was col-
lected on a fritted glass funnel. The filter cake was
washed with dry Et2O (4Â100 mL) and then dissolved
in 0.05 M NH4OH (760 mL). The pale amber-colored
solution was freeze-dried under a 7 mbar vacuum for 120
h to obtain 10 as a yellow solid (1.33 g, 99%): mp (dec)
>240 ꢀC; HPLC retention volumes: 10, 15.3; 9, 16.4
.
(s, 1H, CH¼O). Anal. calcd for C29H32N10O5 CF3-
.
CO2H H2O: C, 50.82; H, 4.81; N, 19.12. Found: C,
50.72, H, 5.03, N, 18.98.
1
mL; H NMR (DMSO-d6, 500 MHz) d 5.23 (s, 2H, 9-
Nꢀ-(4-Amino-4-deoxy-10-formylpteroyl)-Nꢁ-hemiphtha-
loyl-L-ornithinyl-L-phenylalanine (13). Compound 12 (74
mg, 0.10 mmol) was stirred in DMF (1.5 mL) contain-
CH2), 7.58–7.92 (m, 4H, C6H4), 8.72 (s, 1H, C7-H), 8.83
.
(s, 1H, CH¼O). Anal. calcd for C15H13N7O3 4H2O: C,
i
43.80; H, 5.15; N, 23.83. Found: C, 43.77, H, 5.32, N,
23.60.
ing Pr2NEt (0.18 mL, 1.0 mmol) until all the solid dis-
solved (about 45 min). Phthalic anhydride (18 mg, 0.12
mmol) was then added and stirring was continued for
1.5 h. The solution was evaporated under reduced pres-
sure, and the residue was subjected to flash chromato-
graphy on a 24 mm IDÂ145 mm long column of Baker
C18 silica gel. The column was eluted with a 0.5 L gra-
dient of 4–20% EtOH in 0.1M NH4OAc buffer, pH 8.6,
followed by 1 L of 20% EtOH in the same buffer.
Fractions (25 mL) were monitored by HPLC with gra-
dient elution as described above. Fractions containing
pure product were pooled, and evaporated. The residue
was taken up in H2O (20 mL) and the solution was
freeze-dried to obtain the trifluoroacetate salt of 13 as a
yellow powder (39 mg, 50%): mp (dec) >240 ꢀC; HPLC
retention volume: 24 mL; HRFABMS m/z [M+H]+
749.2787 (C37H37N10O8 requires 749.2796; 1H NMR
(DMSO-d6, 500 MHz) d 1.46–1.76 (m, 4H, b- and g-
CH2), 2.89 (m, 2H, d-CH2), 3.01 (m, 2H, PhCH2), 4.43
(m, 1H, a-CHCO), 4.53 (m, 1H, a-CHCO2), 5.19 (s, 2H,
9-CH2), 7.10–8.38 (m, 13H, 2C6H4, C6H5), 8.64 (s, 1H,
C7-H), 8.77 (s, 1H, CH¼O). Anal. calcd for
Nꢀ-(4-Amino-4-deoxy-10-formylpteroyl)-Nꢁ-Boc-L-orni-
thinyl-L-phenylalanine tert-butyl ester (11). To a suspen-
sion of 10 (410 mg, 1.0 mmol) in dry DMF (250 mL)
i
were added Pr2NEt (0.45 mL, 2.5 mmol) and DEPC
(0.3 mL, 1.7mmol). The mixture was stirred at 21 ꢀC
until the solids slowly dissolved. After 4 h, compound 8
(520 mg, 1.1 mmol) was added and stirring was con-
tinued for 3 days. A 400 mg portion of NaHCO3 was
then added and the mixture was evaporated under
reduced pressure. The residue was stirred for 45 min
with CH2Cl2/EtOH (50 mL, 9:1, v/v), the mixture was
filtered, and the solid on the funnel was washed with the
same solvent mixture (2Â25 mL). The filtrate and com-
bined washings were evaporated. The residue was pur-
ified by flash chromatography on a 38 mm IDÂ160 mm
silica gel column with CH2Cl2/MeOH (9:1) as the elu-
ent. Homogeneous product fractions were identified by
TLC, pooled and evaporated. The residue was dissolved
in acetone (5 mL) and the product was precipitated by
stepwise addition of Et2O (20 mL) and hexanes (20 mL).
The precipitate was filtered and dried in vacuo to obtain
11 as a pale-yellow solid (675 mg, 87%): mp (dec) 149–
150 ꢀC; TLC (CHCl3/MeOH, 9:1) Rf 0.27; (n-BuOH/
AcOH/H2O, 3:1:1), Rf 0.66; 1H NMR (DMSO-d6,
500 MHz) d 1.24–1.67(m, 4H, b- and g-CH2), 1.25 (s,
9H, OtBu), 1.33 (s, 9H, Boc), 2.90 (m, 2H, d-CH2), 2.92
.
.
C37H36N10O8 CF3CO2H 3H2O: C, 51.09; H, 4.73; N,
15.28. Found: C, 50.83, H, 5.00, N, 15.10.
Nꢀ-(4-Amino-4-deoxypteroyl)-L-ornithinyl-L-phenylala-
nine (14). A solution of 11 (120 mg, 0.15 mmol) in
ethanolic 1 M HCl (10 mL) was stirred at 21 ꢀC for 16 h.
The mixture was evaporated and the residue dissolved

J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
499
in 25% v/v CF3CO2H in CH2Cl2 (4 mL). After another
16 h the solution was evaporated under reduced pres-
sure. TLC (silica gel, n-BuOH/AcOH/H2O, 3:1:1)
showed some unchanged 11 (Rf 0.66) along with the
desired product 14 (Rf 0.17). The mixture was chroma-
tographed on an 18 mm IDÂ220 mm long column of
cellulose DE-52 (Whatman Ltd., Maidenhead, UK)
with a 0.01–0.08 M gradient of CF3CO2NH4 buffer, pH
6.5. TLC homogeneous product fractions were pooled
and evaporated. The deliquescent residue was dissolved
in H2O (350 mL) and the solution was freeze-dried to
obtain 14 as a yellow powder (58 mg, 61%): mp (dec)
>165 ꢀC; HPLC retention volume: 27.5 mL;
HRFABMS m/z [M+H]+ 573.2640 (C28H33N10O4
and 80 mL of 0.05M TRIS–HCl pH 7.4+10 mM ZnCl2.
A fresh 1.3 mU/mL bCPA stock solution was prepared
in the same buffer and reactions were initiated by addi-
tion of a 10 mL aliquot. For negative controls, 10 mL of
buffer were added. After 20 min at 25 ꢀC, the test sample
and control tubes were tightly sealed and heated for 30
min at 95 ꢀC to denature the enzyme and stop the reac-
tion. The solutions were analyzed by HPLC as descri-
bed above. HPLC retention volumes were: 1, 25; 2, 21
mL.
DHFR inhibition was analyzed as described pre-
viously29 in a quartz cuvette in which 5.5–6.0 mU of
rhDHFR-NADPH were preincubated with 0–45 nM
1ꢂbCPA. The reaction was initiated by addition of
dihydrofolate (FAH2) and the rate of decrease of the
340 nm absorption of 50 mM FAH2 and 65 mM
NADPH monitored at 21 ꢀC. The Ki of 2 was calculated
by the method of Strauss and Goldstein.30
1
requires 573.2686). H NMR (DMSO-d6, 500 MHz) d
1.54–1.76 (m, 4H, b- and g-CH2), 2.89 (m, 2H, d-CH2),
2.92 (m, 2H, PhCH2), 4.49 (m, 1H, a-CHCO), 4.50 (m,
1H, a-CHCO2), 5.32 (s, 2H, 9-CH2), 7.11–8.43 (m, 9H,
C6H5, C6H4), 8.68 (s, 1H, C7–H), Anal. calcd for
.
.
C28H32N10O4 1/4CF3CO2H 2H2O: C, 53.72; H, 5.73; N,
21.98. Found: C, 53.63, H, 5.81, N, 21.78.
Cell growth inhibition assays were performed to deter-
mine the ability of 1 or 2 alone, 1+bCPA, and
2+bCPA to inhibit the population growth of CCRF-
CEM human leukemic lymphoblasts in a 72 h assay.26
IC50 values were calculated by median effect analysis.31
Nꢀ-(4-amino-4-deoxypteroyl)-Nꢁ-hemiphthaloyl-L-orni-
thinyl-L-phenylalanine (1). Method A. Compound 13 (90
mg, 0.11 mmol) was stirred with 0.4 mL of 1 N NaOH
(0.4 mL) until it dissolved, H2O (2.1 mL) was added,
and stirring was continued for 30 min at 21 ꢀC. The
mixture was chilled to 4 ꢀC and 1 N AcOH (0.5 mL) was
added. The precipitated solid was rapidly filtered,
washed with ice-cold H2O (2 mL), and redissolved in
0.05 N NH4OH (25 mL). The amber-colored solution
was freeze-dried for 96 h to obtain 1 as a yellow powder
(7.5 mg, 6.8%). The HPLC retention volume of this
sample was 25 mL, identical to the sample prepared by
Method B.
Acknowledgements
This work was supported in part by grant RO1-
CA25394 from the National Cancer institute, NIH,
DHHS. The technical assistance of Ying-Nan Chen in
performing cytotoxicity assays against CCRF-CEM
cells in culture is also acknowledged.
Method B. Compound 14 (30 mg, 0.047mmol) was
suspended in DMF (0.5 mL) and the mixture was stir-
red for 45 min. The solid did not fully dissolve, but
when phthalic anhydride (7.7 mg, 0.052 mmol) was
added a clear solution was obtained. After 1.5 h the
DMF was evaporated under reduced pressure, and the
residue was chromatographed by the gradient method
described for compound 13. Fractions were monitored
by HPLC, and fractions containing pure product were
pooled and evaporated. Volatile buffer salts were eva-
porated by entrainment with H2O (5Â200 mL) at 30 ꢀC.
The residue was dissolved in H2O (20 mL) and the
solution freeze-dried to obtain 1 as a yellow powder (8.8
mg, 24%); HRFABMS m/z [M+H]+ 721.2827
(C36H37N10O7 requires 721.2846). 1H NMR (DMSO-d6,
500 MHz) d 1.46–1.76 (m, 4H, b- and g-CH2), 2.89 (m,
2H, d-CH2), 3.00 (m, 2H, PhCH2), 4.53 (m, 1H, a-
CHCO2), 4.48 (m, 1H, a-CHCO), 5.19 (s, 2H, 9-CH2),
7.10–8.38 (m, 13H, 2C6H4, C6H5), 8.68 (s, 1H, C7–H).
References and Notes
1. Bagshaw, K. D. Advan. Pharmacol. 1993, 24, 99.
2. Melton, R. G.; Knox, R. J. Enzyme-prodrug strategies for
cancer therapy; Kluwer Academic/Plenum: New York, 1999.
3. Senter, P. D.; Wallace, P. M.; Svensson, H. P.; Vrudhula,
v. M.; Kerr, D. E.; Hellstrom, I.; Hellstrom, K. E. Bioconj.
Chem. 1993, 4, 3.
4. Jungheim, L. N.; Shepherd, T. A. Chem. Rev. 1994, 94,
1553.
5. Weiner, L. M.; Adams, G. P. Oncogene 2000, 19, 6144.
6. Shabat, D.; Rader, C.; List, B.; Lerner, R. A.; Barbas, C. F.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 6925.
7. Wentworth, P.; Datta, A.; Blakey, D.; Boyle, T.; Partridge,
L. J. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 799.
8. King, D. J.; Turner, A.; Farnsworth, A. P. H.; Adair, J. R.;
Owens, R. J.; Pedley, R. B.; Baldock, D.; Proudfoot, K. A.;
Lawson, A. D. G.; Beeley, N. R. A.; Millar, K.; Millican,
T. A.; Boyce, B. A.; Pari, A.; Mountain, A.; Begent, R. H. J.;
Shochat, D.; Yarranton, G. T. Cancer Res. 1994, 54, 6167.
9. Cooke, S. P.; Pedley, R. B.; Boden, R.; Hollinger, P.; Win-
ter, G.; Begent, R. H. J.; Chester, K. A. Brit. J. Cancer 1999,
80, 228.
.
Anal. calcd for C36H36N10O7 2H2O: C, 57.14; H, 5.33;
N, 18.51. Found: C, 56.97, H, 5.17, N, 18.54.
10. Bhatia, J.; Sharma, S. K.; Chester, K. A.; Pedley, R. B.;
Boden, R. W.; Read, D. A.; Boxer, G. M.; Michael, N. P.;
Begent, R. H. J. Int. J. Cancer 2000, 85, 571.
Enzymatic cleavage of 1
Reactions were carried out in 1 mL conical freezing
tubes. Solutions contained 10 mL of 1.0–1.2 mM 1 or 2
11. Vitols, K. S.; Haag-Zieno, B.; Baer, T.; Montejano, Y. D.;
Huennekens, F. M. Cancer Res. 1995, 55, 478.

500
J. E. Wright, A. Rosowsky / Bioorg. Med. Chem. 10 (2002) 493–500
12. Niculescu-Duvaz, I.; Springer, C. J. Current Med. Chem.
1995, 2, 687.
13. Rosowsky, A.; Wright, J. W.; Vaidya, C.; Forsch, R. A.
Pharmacol. Therap. 2000, 85, 191.
14. Bodanszky, M.; Bodanszky, A. The Practice of Peptide
Synthesis; Springer: Berlin, 1984.
15. Rosowsky, A.; Freisheim, J. H.; Moran, R. G.; Solan,
V. C.; Bader, H.; Wright, J. E.; Radike-Smith, M. J. J. Med.
Chem. 1986, 29, 655.
22. Guerquin-Kern, J.-L.; Volk, A.; Chenu, E.; Lougerstay-
Madec, R.; Monneret, C.; Florent, J.-C.; Carrez, D.; Croisy,
A. NMR in Biomed. 2000, 13, 306.
´
23. Hamstra, D. A.; Page, M.; Maybaum, J.; Rehemtulla, A.
Cancer Res. 2000, 60, 657.
24. Wright, J. E.; Pardo, M.; Tretyakov, O.; Alperin, W. L.;
Trites, D.; Rosowsky, A. Cancer Chemother. Pharmacol. 1998,
42, 300. For erratum, see: Cancer Chemother. Pharmacol.
1999, 43, 522.
16. Rosowsky, A.; Wright, J. E.; Ginty, C.; Uren, J. J. Med.
Chem. 1982, 25, 960.
25. Lin, T.; Noker, P. E.; Johnson, J. D.; Rhie, J. K. Proc.
Am. Assoc. Cancer Res. 2000, 41, 490.
17. Wolfe, L. A.; Mullin, R. J.; Laethem, R.; Blumenkopf,
T. A.; Cory, M.; Miller, J. F.; Keith, B. R.; Humphries, J.;
Smith, G. K. Proc. Am. Assoc. Cancer Res. 1996, 37, 467.
18. Smith, G. K.; Banks, S.; Blumenkopf, T. A.; Cory, M.;
Humphreys, J.; Laethem, R. M.; Miller, J.; Moxham, C. P.;
Mullin, R.; Ray, P. H.; Walton, L. M.; Wolfe, L. A., III. J.
Biol. Chem. 1997, 272, 15804.
19. Wolfe, L. A.; Mullin, R. J.; Laethem, R.; Blumenkopf,
T. A.; Cory, M.; Miller, J. F.; Keith, B. R.; Humphries, J.;
Smith, G. K. Bioconj. Chem. 1999, 10, 38.
26. Rosowsky, A.; Wright, J. E.; Vaidya, C.; Forsch, R. A.;
Bader, H. J. Med. Chem. 2000, 43, 1620.
27. Rosowsky, A.; Wright, J. E.; Vaidya, C. M.; Bader, H.;
Forsch, R. A.; Mota, C. E.; Pardo, J.; Chen, C. S.; Chen, Y.-
N. J. Med. Chem. 1998, 41, 5310.
28. Rosowsky, A.; Freisheim, J. H.; Bader, H.; Forsch, R. A.;
Susten, S. A.; Cucchi, C. A.; Frei, E., III J. Med. Chem. 1985,
28, 660.
29. Johnson, J. M.; Meiering, E. M.; Wright, J. E.; Pardo, J.;
Rosowsky, A.; Wagner, G. Biochemistry 1997, 36, 4399.
30. Strauss, O. H.; Goldstein, A. J. Gen. Physiol. 1943, 26,
559.
20. Stribbling, S. M.; Martin, J.; Pedley, R. B.; Boden, J. A.;
Sharma, S. K.; Springer, C. J. Cancer Chemother. Pharmacol.
1997, 40, 277.
31. Chou, T. C.; Talalay, P. Adv. Enzyme Regul. 1984, 22,
27.
21. Jain, R. K. Microcirculation 1997, 4, 1.