Synthesis and evaluation of pteroic acid-conjugated nitroheterocyclic phosphoramidates as folate receptor-targeted alkylating agents

Article abstract of DOI:10.1021/jm000306g

A novel nitroheterocyclic bis(haloethyl)phosphoramidate prodrug linked through lysine to a pteroic acid has been prepared and evaluated as a potential alkylating agent to target tumor cells that overexpress the folate receptor. The prodrug exhibited IC

Full text of DOI:10.1021/jm000306g

J . Med. Chem. 2001, 44, 69-73
69
Syn th esis a n d Eva lu a tion of P ter oic Acid -Con ju ga ted Nitr oh eter ocyclic
P h osp h or a m id a tes a s F ola te Recep tor -Ta r geted Alk yla tin g Agen ts
Gali Steinberg and Richard F. Borch*
Department of Medicinal Chemistry and Molecular Pharmacology and the Cancer Center, Purdue University,
West Lafayette, Indiana 47907
Received J uly 21, 2000
A novel nitroheterocyclic bis(haloethyl)phosphoramidate prodrug linked through lysine to a
pteroic acid has been prepared and evaluated as a potential alkylating agent to target tumor
cells that overexpress the folate receptor. The prodrug exhibited IC50 values in the micromolar
range and was 10-400-fold less cytotoxic in vitro than the phosphoramidate that lacks the
lysine-pteroyl moiety. The data does not support a contribution of the folate receptor to
cytotoxicity. In an attempt to determine the basis for the decreased cytotoxicity in the pteroyl-
lysyl analogue, compounds were prepared in which the lysine-pteroyl moiety was replaced with
lysine alone or with an n-propyl group. The n-propyl and the lysyl analogues were on average
3.8- and 21-fold less potent than the unsubstituted bis(haloethyl)phosphoramidate, respectively.
Chemical reduction of the prodrugs followed by 31P NMR kinetics demonstrated that all of the
phosphoramidate anions cyclized to the aziridinium ion at similar rates and gave comparable
product distributions, suggesting that changes in chemical activation did not account for the
differences in cytotoxicity. It is likely that folate receptor-mediated transport is not sufficient
to deliver adequate intracellular concentrations of the cytotoxic phosphoramide mustard.
In tr od u ction
the alkylating moiety. The pteroic acid ligand for the
folate receptor is introduced to achieve folate receptor
selectivity. It has been established that folate receptors
have a structural requirement of a pteroyl moiety
appended to the R-amino group of an amino acid
containing a free R-carboxyl group that is essential for
ligand binding.7 With respect to the amino acid sub-
stituent, the receptor will tolerate extensive substitu-
tion. Lysine has been selected to connect the phospho-
ramidate to the pteroyl moiety. Lysine fulfills the
R-amino acid requirement of the receptor when it is
coupled to the pteroyl moiety, and at the same time the
ꢀ-amino group can be exploited to attach the phospho-
ramidate. Here we report the synthesis, evaluation in
vitro, and activation mechanisms of substituted nitro-
furyl phosphorodiamidates.
A recent approach to cancer treatment involves
enhancement of the differential specificity of anticancer
agents by selective targeting mechanisms. The vitamin
folic acid has attracted considerable attention as a
potential means of delivering covalently bound drug
conjugates.1 Folate uptake is mediated by a family of
membrane receptors termed the folate-binding proteins
(FBP). FBPs have KD values for folate that vary from
10-9 to 10-11 M and exhibit a strong preference for folic
acid over its reduced counterparts. Many human cancer
cell lines have highly overexpressed levels of the protein
that binds folic acid,2 a finding which is being exploited
with the preparation of folate-drug conjugates.3
Bioreductive activation is a well-established concept
for achieving selective toxicity of antitumor agents to
hypoxic cells.4 Recently we reported that nitrohetero-
cyclic phosphoramidates can be activated under both
aerobic and hypoxic conditions.5 These prodrugs were
highly potent but only moderately hypoxia-selective.
The activation of these prodrugs involves bioreduction
of the nitro group and subsequent release of the cyto-
toxic phosphoramide mustard (PDA).5,6 We have at-
tempted to exploit the selective targeting afforded by
folate-drug conjugates by combining a pteroic acid
moiety with the high potency of the nitroheterocyclic
phosphoramidate prodrugs to design a folate receptor-
selective cytotoxic agent.
Resu lts a n d Discu ssion
Ch em istr y. The mechanistic hypothesis underlying
the proposed activation of these compounds is shown
in Scheme 1. Enzymatic reduction is expected to proceed
to the corresponding hydroxylamine derivative. Lone
pair-assisted expulsion of the phosphorodiamidate anion
from this intermediate will activate the alkylating
moiety.
The nitrofuryl phosphorodiamidates 1a and 1b were
prepared as outlined in Scheme 2. Lysine methyl ester
(7) was reacted with the phosphoramidic chlorides 2a
and 2b in the presence of diisopropylethylamine to give
the lysine phosphoramidates 3a and 3b. The methyl
esters were then hydrolyzed using pig liver esterase in
acetone/phosphate buffer to give the free acids 4a and
4b. To confirm that phosphorylation took place on the
ꢀ-amino group of the lysine, an NMR decoupling experi-
ment was carried out on 3b. If phosphorylation occurred
on the ꢀ-amino group, irradiation of the protons on the
The prodrug includes several components: a nitro-
heterocyclic delivery group, an alkylating group, and a
ligand for the folate receptor. When the nitroheterocyclic
delivery group is bioreduced a phosphoramide mustard
is released intracellularly; this process serves to activate
* To whom correspondence should be addressed. Tel: (765) 494-
1403. Fax: (765) 494-1414. E-mail: rickb@pharmacy.purdue.edu.
10.1021/jm000306g CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/05/2000

70 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 1
Steinberg and Borch
Sch em e 1
9b were cleaved in a palladium-catalyzed reaction in
the presence of sodium toluenesulfinate to give 4a and
4b, which were identical to the products obtained from
the enzymatic cleavage of the methyl esters 3a and 3b.
Coupling of the nitrofuryl lysylphosphoramidates 4a
and 4b to 10-trifluoroacetylpteroic acid 5 was carried
out using isobutyl chloroformate and triethylamine to
give the pteroyllysyl intermediates 6a and 6b. Finally,
the trifluoroacetyl group was removed using 3 equiv of
potassium carbonate to give the prodrugs 1a and 1b.
Compounds 10a and 10b were synthesized as previously
described.5 They are known to be highly cytotoxic and
were used as positive controls in the cytotoxicity studies.
Biologica l Activity. The cytotoxicities of 1a and 10a
were assessed against cell lines with overexpressed
folate receptors (IGROV human ovarian carcinoma8 and
MDA231 human breast adenocarcinoma9) and against
cell lines with low levels of folate receptors (A549 human
lung carcinoma cells3c and A498 human kidney carci-
noma cells9). To assess the contribution of folate trans-
port to drug activity, cytotoxicity was compared using
overexpressed receptor cell lines in folate-containing and
folate-deficient media. Folate competes for the receptor
and therefore inhibits receptor-mediated drug transport,
so differences between folate-deficient and folate-
containing media reflect the contribution of the folate
transport system to drug activity. Cell viability was
measured after 72-h drug incubation using the MTT
assay;10 the results are presented in Table 1. The
cytotoxicity of 1a is modest, with IC50 values in the low-
micromolar range, and is 1-2 orders of magnitude less
toxic than the unsubstituted prodrug 10a . Although 1a
is somewhat more toxic to IGROV and MDA231 cells
in folate-deficient compared to folate-containing media,
the differences are much smaller than one would expect
δ-carbon should give a doublet for the ꢀ-protons arising
from three-bond proton-phosphorus coupling. If phos-
phorylation occurred on the R-amino group, however,
the ꢀ-protons should appear as a singlet following
irradiation. In fact, irradiation of the δ-methylene at 1.6
ppm caused collapse of the ꢀ-proton peak at 2.9 ppm to
a doublet with J ) 9.7 Hz, consistent with phosphory-
lation of the ꢀ-amino group. Unfortunately, the yields
for the enzymatic hydrolysis were not reproducible and
ranged between 5% and 65%. An alternative route was
then adopted (Scheme 2). The dicyclohexylammonium
salt of di-t-Boc-lysine was converted to the allyl ester
with allyl bromide in the presence of diisopropylethy-
lamine, and the amine groups were deprotected with
trifluoroacetic acid. The resulting lysine allyl ester 8 was
then phosphorylated under the same conditions as
described for the methyl ester. The allyl esters 9a and
Sch em e 2a
a
Reagents: (a) allyl bromide, i-Pr2NEt, 70 °C; (b) TFA-CH2Cl2; (c) lysine methyl ester 7 or allyl ester 8, i-Pr2NEt; (d) pig liver esterase,
0.1 M KH2PO4, pH 7, 10% acetone; (e) Pd(PPh3)4, ArSO2Na, THF/H2O; (f) n-PrNH2, i-Pr2NEt; (g) i-BuOCOCl, Et3N, then 4a and 4b; (h)
K2CO3, MeOH/THF.

Pteroic Acid Nitroheterocyclic Phosphoramidates
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 1 71
Ta ble 1. Growth Inhibition of Human Tumor Cell Lines in
ppm); the phosphoramidate anion subsequently disap-
peared with a half-life of 10.9 min. Similarly, reduction
of 11 with sodium dithionite led to rapid disappearance
of the resonance corresponding to 11 (-5.5 ppm) and
the appearance of the analogous phosphoramidate anion
(-11.9 ppm). This intermediate subsequently disap-
peared with a half-life of 13.0 min. Finally, dithionite
reduction of 10b (-5.5 ppm) led to formation of phos-
phoramide mustard (-12.3 ppm) which disappeared
with a half-life of 18 min. These modest differences in
rate of phosphoramidate activation are unlikely to
account for the differences in cytotoxicity observed
among the prodrugs. Furthermore, in each case the bis-
(chloroethyl)phosphoramidate reacts to give a mixture
of products arising from initial formation of the aziri-
dinium ion and subsequent reactions of water and other
nucleophiles at both carbon and phosphorus.12,13 Al-
though there are minor changes in the product distribu-
tion of the phosphoramidate anions derived from 4b,
10b, and 11, none of these changes would account for
the observed differences in cytotoxicity. Therefore it is
unlikely that differences in activation chemistry account
for the variation in cytotoxicity among these prodrugs.
Vitroa
IC50 (µM)
cell line
1a
10a
4b
10b
11
IGROVb,c
IGROVfdb,c
MDA231c
MDA231fdb,c
A549d
0.83
0.57
8.0
3.6
11
23
11
19
5.2
0.30
11
16
0.041
0.049
0.032
0.022
0.27
16
0.52
0.33
1.1
1.9
41
26
1.9
A549fdb,d
A498d
0.26
0.045
0.029
0.013
0.060
0.25
1.4
A498fdb,d
UMUC3
HT29
PC3
PACA2
0.62
20
21
0.055
0.38
1.0
0.72
17
2.3
0.45
0.16
2.4
0.13
a
See Experimental Section for details of the in vitro assays.
All cell lines designated “fd” were grown in folate-deficient
b
medium. c IGROV and MDA231 cell lines have overexpressed
d
folate receptors. A549 and A498 cell lines have low levels of folate
receptors.
if selective transport were essential for activity. We do
not understand why folate receptor overexpression does
not enhance the toxicity of these compounds. A diverse
array of small and large molecules can be linked to folic
acid to give conjugates with excellent receptor binding
affinity, but the γ-carboxyl group is the only linkage site
that provides high-affinity conjugates. Folate conjugates
penetrate cells exclusively by receptor-mediated endocy-
tosis;1 although sequestration in vesicular structures is
known to be a problem with folate-protein conjugates,
free folic acid and low-molecular-weight conjugates are
not sequestered.3 Folate-mediated tumor targeting has
been exploited for other low-molecular-weight antitumor
agents, but these compounds must be linked via the
γ-carboxyl group of pteroylglutamic acid.11 Conceivably
the γ-carboxyl group contributes to receptor binding,
and the lysine conjugates are not transported because
they lack the ester carbonyl group. The compounds were
then examined in seven cell lines in folate-containing
medium where folate-specific transport should not occur
(see Table 1). The pteroate conjugate 1a is significantly
less cytotoxic than the unsubstituted nitrofuryl phos-
phoramidate 10a in all of the cell lines tested. To
determine the possible effects of phosphoramidate sub-
stitution on cytotoxicity, the unsubstituted, the lysyl,
and the n-propyl bis(chloroethyl)phosphoramidates 10b,
4b, and 11, respectively, were prepared and evaluated
in vitro. Introduction of the n-propyl, the lysine, and
the pteroyllysine groups increases the relative IC50
values by approximately 4-, 20-, and 64-fold, respectively
(Table 1). The cytotoxicity of the nitrofuryl phosphora-
midate prodrugs requires that the compounds undergo
intracellular conversion to the phosphoramidate anion
followed by cyclization to the reactive aziridinium ion.5
To assess whether substitution influences the activation
of these compounds, the chemical reduction and subse-
quent phosphoramidate anion chemistry were assessed
using 31P NMR.
Con clu sion s
The pteroyllysine-conjugated nitroheterocyclic phos-
phoramidate 1a showed modest cytotoxicity that was
significantly diminished compared to that of the unsub-
stituted phosphoramidate 10a ; the basis for this dimin-
ished toxicity is not understood. Differences in cytotox-
icity under folate-deficient conditions to cells that are
high and low in folate receptor levels were modest, and
addition of folate to the medium gave only a modest
increase in the IC50 for the folate receptor-overexpressed
cell lines. These results suggest that folate receptors are
not contributing significantly to the intracellular deliv-
ery of the cytotoxin. The rates of phosphoramidate anion
cyclization and product distribution were not signifi-
cantly altered by the nature of the phosphoramidate
substitution, indicating that the decreased cytotoxicity
is not related to changes in prodrug activation. Finally,
it is likely that successful development of a pteroate-
conjugated drug that can exploit folate receptor over-
expression will require delivery of a cytotoxin that is
much more potent than a phosphoramide mustard.
Exp er im en ta l Section
1
All H NMR spectra were measured on a 250-MHz Bruker
NMR system equipped with a Tecmag data acquisition system
and 5-mm multinuclear probe. The NMR program MacNMR
1
was used for acquisition and processing of data. H chemical
shifts are reported in parts per million from tetramethylsilane.
All 31P NMR spectra were obtained on the same instrument
using broadband gated decoupling. Chemical shifts are re-
ported in parts per million from a coaxial insert containing
5% phosphoric acid in H2O (or in D2O), except for 31P kinetics
experiments where a coaxial insert containing 1% triph-
enylphosphine oxide in benzene-d6 was used. The chemical
shift difference between phosphoric acid and triphenylphos-
phine oxide inserts is 25.2 ppm. 31P NMR kinetics were carried
out as described previously.12
Chromatographic separations were done on silica gel grade
60 or on 40-µm octadecyl (C18) supports. High-performance
liquid chromatography (HPLC) analyses were performed using
a Beckman System Gold with a 126 solvent module, a 168
detector set to 313 nm, and an Econosphere C18 5-µm column
(250 mm; Alltech Associates). Mass spectral data were ob-
31P NMR Kin etics. Compound 4b was reduced with
sodium dithionite (cacodylate buffer, pH 7.4, 37 °C), and
the reaction was monitored by 31P NMR. Addition of
sodium dithionite to the solution resulted in rapid
disappearance of the resonance corresponding to 4b
(-5.7 ppm) and the appearance of the resonance for the
substituted bis(chloroethyl)phosphoramidate anion (-11.5

72 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 1
Steinberg and Borch
tained from the Purdue University Mass Spectrometry Service,
West Lafayette, IN, using electrospray (ESI) or fast atom
bombardment (FAB) ionization. All mass spectral data for
halogenated compounds show the appropriate isotope ratios;
however only the lowest-molecular-weight peak is reported.
All anhydrous reactions were carried out in oven-dried flasks
under argon. Reagents and solvents were introduced via
syringe where appropriate. All solvents were distilled prior to
use over calcium hydride or sodium.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-(O-Meth yllysyl)-N,N-bis(2-
br om oeth yl)p h osp h or od ia m id a te (3a ). Lysine methyl ester
dihydrochloride was dried over P2O5 at 45 °C for 48 h and then
dissolved (244 mg, 1.05 mmol) in dry acetonitrile (15 mL) and
cooled to -40 °C under argon. Diisopropylethylamine (0.60 mL,
3.45 mmol) was added, followed by the addition of a small
amount of 18-crown-6 to improve solubility. Phosphoramidic
chloride 2a 5 (476 mg, 1.05 mmol) was dissolved in dry
acetonitrile (3 mL) and added to the reaction mixture. The
mixture was warmed to -10 °C and allowed to stir for 3 h.
The reaction was quenched with water at -10 °C and extracted
with CH2Cl2 (3×). The combined organic layers were dried over
MgSO4, filtered and concentrated under reduced pressure.
Reverse-phase chromatography using a gradient of acetoni-
trile/water:acetonitrile 20% to 80% afforded 3a as a mixture
of diastereomers (440 mg, 72%): 1H NMR (CDCl3) δ 7.31 (d,
1Η), 6.75 (d, 1H), 5.08 (dd, 2H), 3.25-3.71 (bm, 12H), 2.95 (bs,
2H), 1.34-1.94 (bm, 6H); 31P NMR (CDCl3) δ 16.82; MS (ESI)
m/z 577 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-(O-m eth yllysyl)-N,N-bis(2-
ch lor oeth yl)p h osp h or od ia m id a te (3b) was prepared as
described for 3a (1.7 mmol scale) using bis(chloroethyl)-
phosphoramidic chloride 2b.5 The crude product was purified
by reverse-phase chromatography using a gradient of aceto-
nitrile/water:acetonitrile 20% to 80% to give 3b (582 mg, 70%)
as a mixture of diastereomers: 1H NMR (CDCl3) δ δ 7.30 (d,
1Η), 6.75 (d, 1H), 5.07 (dd, 2H), 3.26-3.70 (bm, 12H), 2.93 (bs,
2H), 1.34-1.93 (bp, 6H); 31P NMR (CDCl3) δ 17.26;. MS (ESI)
m/z 489 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-Lysyl-N,N-bis(2-br om oeth -
yl)p h osp h or od ia m id a te (4a ). Methyl ester 3a (180 mg, 0.31
mmol) was dissolved in acetone (0.6 mL) and KH2PO4 (5.4 mL,
0.1 M, pH 7) was added with stirring. Pig liver esterase (800
units) was added to the reaction mixture, and it was stirred
at room temperature for 36 h. During that time the pH was
maintained by dropwise addition of 0.5 M NaOH (to a total of
1 equiv). The reaction progress was monitored using HPLC.
When the reaction was complete, the solvent was removed
under reduced pressure and the residue redissolved in meth-
ylene chloride. The suspension was filtered, evaporated and
the residue was redissolved in acetonitrile and purified by
reverse-phase chromatography using a gradient of acetonitrile/
water:acetonitrile 20% to 80% to give 4a as a mixture of
diastereomers (116 mg, 66%): 1H NMR (CDCl3) δ 7.28 (d, 1Η),
6.82 (d, 1H), 5.02 (dd, 2H), 3.18-3.79 (bm, 9H), 2.91 (bs, 2H),
1.30-1.90 (bm, 6H); 31P NMR (CDCl3) δ 17.07; MS (ESI) m/z
563 (M + H)+.
trifluoroacetic acid-methylene chloride (60 mL, 1:1 v/v), and
the solution was stirred at room temperature for 50 min. The
solution was concentrated under reduced pressure and the
residue triturated repeatedly with diethyl ether. The product
was isolated as a yellow oil (1.64 g, 86% overall yield): Rf )
1
0.28 (CHCl3-MeOH-HOAc, 75:23:2); H NMR (DMSO-d6) δ
5.94 (m, 1H), 5.35 (dd, 2H), 4.70 (d, 2H), 4.06 (t, 1H), 2.75 (m,
2H), 1.17-1.92 (m, 6H); MS (ESI) m/z 187 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-(O-Allyllysyl)-N,N-bis(2-br o-
m oeth yl)p h osp h or od ia m id a te (9a ). Lysine ester 8 (445 mg,
1.08 mmol) was dissolved in acetonitrile (5 mL) and cooled to
-30 °C, and diisopropylethylamine (0.62 mL, 3.55 mmol) was
added to the solution. Phosphoramidic chloride 2a (489 mg,
1.08 mmol) was dissolved in dry acetonitrile (3 mL) and added
to the reaction mixture, which was then warmed to -10 °C
and stirred for 2.5 h. The reaction was quenched with water
at -10 °C and extracted with CH2Cl2 (3×). The combined
organic layers were dried over MgSO4 and concentrated under
reduced pressure to give 9a as an oil (565 mg, 87%): 1H NMR
(CDCl3) δ 7.30 (d, 1Η), 6.74 (m, 1H), 5.90 (m, 1H), 5.35 (dd,
2H), 5.02 (m, 2H), 4.63 (d, 2H); 4.10 (t, 1H), 3.24-3.65 (bm,
8H), 2.92 (m, 2H), 1.28-1.99 (m, 6H); 31P NMR (CDCl3) δ 17.06
(diastereomers, broad peak); MS (ESI) m/z 603 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-(O-allyllysyl)-N,N-bis(2-ch lo-
r oeth yl)p h osp h or od ia m id a te (9b) was prepared as de-
scribed for 9a (2.86 mmol scale) using phosphoramidic chloride
2b. The crude product was used without further purification
(1.33 mg, 90%): 1H NMR (CDCl3) δ 7.30 (d, 1Η), 6.73 (m, 1H),
5.91 (m, 1H), 5.34 (dd, 2H), 5.02 (m, 2H), 4.61 (d, 2H), 3.89 (t,
1H), 3.24-3.75 (bm, 8H), 2.92 (m, 2H), 1.29-2.01 (m, 6H); 31
P
NMR (CDCl3) δ 17.11 (diastereomers, broad peak); MS (ESI)
m/z 515 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-Lysyl-N,N-bis(2-br om oeth -
yl)p h osp h or od ia m id a te (4a ). Pd(PPh3)4 (47 mg, 0.041 mmol)
was added to a solution of phosphoramidate 9a (490 mg, 0.81
mmol) in THF (10 mL). p-Toluenesulfinic acid (173 mg, 0.97
mmol) was dissolved in water (0.5 mL) and added to the THF
solution. The reaction mixture was stirred for 30 min at room
temperature, diluted with water, washed with diethyl ether
(5×), and the water layer was then concentrated under reduced
pressure. Reverse-phase chromatography using a gradient of
acetonitrile/water:acetonitrile 20% to 80% afforded 4a as a
yellow oil (244 mg, 53%). This product was identical in all
respects to that prepared by enzymatic hydrolysis of the
methyl ester.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-lysyl-N,N-bis(2-ch lor oeth -
yl)p h osp h or od ia m id a te (4b) was prepared as described for
4a (2.16 mmol scale) using phosphorodiamidate 9b. The crude
residue was purified by reverse-phase chromatography using
a gradient of acetonitrile/water:acetonitrile 20% to 80%) to give
4b as a yellow oil (0.564 g, 55%). This product was identical
in all respects to that prepared by enzymatic hydrolysis of the
methyl ester.
(5-Nitr o-2-fu r yl)m eth yl NE-(N10-(Tr iflu or oa cetyl)p ter -
oyllysyl)-N,N-bis(2-br om oeth yl)ph osph or odiam idate (6a).
N10-(Trifluoroacetyl)pteroic acid 5 was dried over P2O5 for 48
h at 45 °C. A solution of 5 (10 mg, 0.03 mmol) in anhydrous
DMF (0.5 mL) under argon was cooled to 0 °C, and triethy-
lamine (0.007 mL, 0.05 mmol) was added, followed by isobutyl
chloroformate (0.005 mL, 0.04 mmol). The reaction mixture
was protected from light and stirred at 0 °C for 1 h. Phospho-
rodiamidate 4a (21 mg, 0.04 mmol) was dissolved in anhydrous
DMF (0.5 mL), added to the reaction mixture, and stirred for
2 h at 0 °C. Water (3 mL) was added to the reaction mixture
and the precipitated solid was collected by centrifugation to
give 6a as a yellow powder (18 mg, 76%): 1H NMR (DMSO-
d6) δ 8.60 (s, 1Η), 7.91 (d, 2H), 7.49-7.76 (bm, 3H), 6.90 (d,
1H), 5.10 (s, 2H), 4.92 (bd, 2H); 4.25 (bs, 1H), 2.88 (bd, 2H),
1.24-1.90 (m, 6H); 31P NMR (DMSO-d6) δ 17.67; MS (ESI) m/z
953 (M + H)+.
(5-Nitr o-2-fu r yl)m eth yl Nꢀ-lysyl-N,N-bis(2-ch lor oeth -
yl)p h osp h or od ia m id a te (4b) was prepared as described for
4a (0.37 mmol scale). The crude residue was purified by
reverse-phase chromatography using a gradient of acetonitrile/
water:acetonitrile 20% to 80% to give 4b as a mixture of
diastereomers (112 mg, 64%): 1H NMR (CDCl3) δ 7.25 (d, 1Η),
6.69 (d, 1H), 4.96 (dd, 2H), 3.09-3.82 (bm, 9H), 2.93 (bs, 2H),
1.29-1.91 (bm, 6H); 31P NMR (CDCl3) δ 17.45; MS (ESI)
(C15H25N4O7PCl2) calcd 475.0916, obsd 475.0903.
Allyl Lysin a te Bistr iflu or oa ceta te (8). NR,Nꢀ-Di-t-Boc-
lysine dicyclohexylammonium salt (2 g, 3.8 mmol) was dis-
solved in allyl bromide (10 mL), and diisopropylethylamine
(1.32 mL, 7.58 mmol) was added to the solution. The mixture
was heated for 2 h at 70-75 °C. It was diluted with ethyl
acetate (200 mL) and washed with 0.1 N HCl (3 × 100 mL),
saturated NaHCO3 (3 × 100 mL), and saturated NaCl (3 ×
100 mL), dried over MgSO4, and concentrated under reduced
pressure. The resultant yellow oil (1.8 g) was dissolved in
(5-Nitro-2-furyl)methylNꢀ-(N10-(trifluoroacetyl)pteroylly-
syl)-N,N-bis(2-ch lor oeth yl)p h osp h or od ia m id a te (6b) was
prepared as described for 6a (0.025 mmol scale) using phos-
phorodiamidate 4b: 1H NMR (DMSO-d6) δ 8.50 (s, 1Η), 7.77

Pteroic Acid Nitroheterocyclic Phosphoramidates
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 1 73
(d, 2H), 7.30-7.64 (bm, 3H), 6.78 (d, 1H), 5.00 (s, 2H), 4.83
(bd, 2H), 4.17 (bs, 1H), 2.74 (bd, 2H), 1.38-1.82 (m, 6H); 31P
NMR (DMSO-d6) δ 17.87; MS (ESI) m/z 865 (M + H)+.
(5-Nit r o-2-fu r yl)m et h yl Nꢀ-(P t er oyllysyl)-N,N-b is(2-
b r om oet h yl)p h osp h or od ia m id a t e (1a ). (5-Nitro-2-furyl)-
methyl Nꢀ-(N10-(trifluoroacetyl)pteroyllysyl)-N,N-bis(2-bromo-
ethyl)phosphorodiamidate 6a (13 mg, 0.014 mmol) was
suspended in THF/MeOH (1.6 mL, 1:1), and aqueous K2CO3
(5.8 mg, 0.04 mmol; 0.02 mL) was added. The resulting
homogeneous solution was stirred for 6 days at room temper-
ature. Aqueous trifluoroacetic acid (3 mL, 0.1%) was added to
the reaction mixture, and the precipitated solid was collected
by centrifugation to give 1a as a yellow powder (10.4 mg,
89%): 1H NMR (DMSO-d6) δ 11.40 (s, 1H), 8.63 (s, 1H), 7.65
(bm, 3H), 6.88 (bm, 2H), 6.61 (d, 2H), 4.93 (bd, 2H), 4.46 (s,
2H), 4.26 (bs, 1H), 2.72 (bd, 2H), 1.38-1.82 (m, 6H); 31P NMR
(DMSO-d6) δ 17.82; MS (FAB) (C29H35N10O9PBr2) calcd
857.0771, obsd 857.0789.
(5-Nit r o-2-fu r yl)m et h yl Nꢀ-(p t er oyllysyl)-N,N-b is(2-
ch lor oeth yl)p h osp h or od ia m id a te (1b) was prepared as
described for 1a (0.012 mmol scale) using phosphorodiamidate
6b: 1H NMR (DMSO-d6) δ 11.40 (s, 1Η), 8.64 (s, 1Η), 7.66 (bm,
3H), 6.89 (bm, 2H), 6.24 (d, 2H), 4.93 (bd, 2H), 4.47 (s, 2H),
4.26 (bs, 1H), 2.73 (bd, 2H), 1.38-1.79 (m, 6H); 31P NMR
(DMSO-d6) δ 17.98; MS (negative ion FAB) (C29H35N10O9PCl2)
calcd 767.1625, obsd 767.1621.
(5-Nitr o-2-fu r yl)m eth yl N,N-Bis(2-ch lor oeth yl)-N-p r o-
p ylp h osp h or od ia m id a te (11). Propylamine (0.07 mL, 0.85
mmol) was dissolved in methylene chloride (10 mL) and cooled
to -20 °C under argon. Diisopropylethylamine (0.18 mL, 1.02
mmol) was added, followed by addition of phosphoramidic
chloride 2b (310 mg, 0.85 mmol) in methylene chloride (1 mL).
The mixture was warmed to -10 °C and allowed to stir for 3
h. The reaction was quenched by addition of water at -10 °C
and extracted with CH2Cl2 (3×). The combined organic layers
were dried over MgSO4 and concentrated under reduced
pressure. The product was isolated as a yellow oil after
purification by flash chromatography (123 mg, 77%): 1H NMR
(CDCl3) δ 7.30 (d, 1Η), 6.67 (d, 1H), 5.06 (dd, 2H), 3.29-3.70
(bm, 8H), 2.88 (m, 2H), 1.54 (m, 2H), 0.93 (t, 3H); 31P NMR
(CDCl3) δ 16.40; MS (FAB) (C12H20N3O5PCl2) calcd 388.0596,
obsd 388.0595.
31P NMR Stu d ies. Compounds 4b, 10b and 11 were
dissolved in 100 µL of CH3CN, and cacodylate buffer (400 µL,
0.4M, pH7.4) was added. The solution was placed in a 5-mm
NMR tube, and sodium dithionite (3 equiv) in 100 µL of buffer
was added to the NMR tube. The acquisition was started and
the start time recorded. Spectra were taken every 2.5 min for
0.5 h, then every 5 min for 0.5 h, then every 10 min for 1 h,
and time points for each spectrum were assigned from the
beginning of the data acquisition. The temperature of the probe
was maintained at 37 °C, using the Bruker variable temper-
ature unit. Chemical shifts are reported in parts per million
from a coaxial insert containing 1% triphenylphosphine oxide
in benzene-d6. The relative concentrations of the intermediates
were determined from integration of the peak areas.
Croy in the Drug Development Shared Resource, Pur-
due University Cancer Center. Financial support from
Grant CA34619, provided by the National Cancer
Institute, DHHS, is gratefully acknowledged.
Refer en ces
(1) (a) Brigle, K. E.; Spinella, M. J .; Westin, E. H.; Goldman, I. D.
Increased expression and characterization of two distinct folate
binding proteins in murine erythroleukemia cells. Biochem.
Pharmacol. 1994, 47, 337-345. (b) Shen, F.; Ross, J . F.; Wang,
X.; Ratnam, M. Identification of a novel folate receptor, a
truncated receptor, and receptor type beta in hematopoietic
cells: cDNA cloning, expression, immunoreactivity, and tissue
specificity. Biochemistry 1994, 33, 1209-1215. (c) Leamon, C.
P.; Low, P. S. Membrane folate-binding proteins are responsible
for folate-protein conjugate endocytosis into cultured cells.
Biochem. J . 1993, 291, 855-860. (d) Leamon, C. P.; Low, P. S.
Delivery of macromolecules into living cells: A method that
exploits folate receptor endocytosis. Proc. Natl. Acad. Sci. U.S.A.
1991, 88, 5572-5576.
(2) (a) Mantovani, L. T.; Miotti, S.; Menard, S.; Canevari, S.;
Raspagliesi, F.; Bottini, C.; Bottero, F.; Colnaghi, M. I. Folate
binding protein distribution in normal tissues and biological
fluids from ovarian carcinoma patients as detected by the
monoclonal antibodies MOv18 and MOv19. Eur. J . Cancer 1994,
30A, 363-369. (b) Ross, J . F.; Chaudhuri, P. K.; Ratnam, M.
Differential regulation of folate receptor isoforms in normal and
malignant tissues in vivo and in established cell lines. Physi-
ologic and clinical implications. Cancer 1994, 73, 2432-2443.
(3) (a) Leamon, C. P.; Low P. S. Cytotoxicity of momordin-folate
conjugates in cultured human cells. J . Biol. Chem. 1992, 267,
24966-24971. (b) Leamon, C. P.; Low P. S. Selective targeting
of malignant cells with cytotoxin-folate conjugates. J . Drug
Targeting 1994, 2, 101-112. (c) Leamon, C. P.; Pastan, I.; Low
P. S. Cytotoxicity of folate-Pseudomonas exotoxin conjugates
toward tumor cells. Contribution of translocation domain. J . Biol.
Chem. 1993, 268, 24847-24854.
(4) (a) Sartorelli, A. Therapeutic attack of hypoxic cells of solid
tumors: Presidential address. Cancer Res. 1988, 48, 775-778.
(b) Denny, W. A.; Atwell, G. J .; Anderson, R. F.; Wilson, W. R.
Hypoxia-selective antitumor agents. Relationships between
structure, physicochemical properties, and hypoxia-selective
cytotoxicity for nitracrine analogues with varying side chains:
The “iminoacridan hypothesis”. J . Med. Chem. 1990, 33, 1288-
1295.
(5) Borch, R. F.; Liu, J .; Schmidt, J . P.; Marakovits, J .; J oswig, C.;
Gipp, J . J .; Mulcahy, T. R. Synthesis and evaluation of nitro-
heterocyclic phosphoramidates as hypoxia-selective alkylating
agents. J . Med. Chem. 2000, 43, 2258-2265.
(6) (a) Firestone, A.; Mulcahy, R. T.; Borch, R. F. Nitroheterocycle
reduction as a paradigm for intramolecular catalysis of drug
delivery to hypoxic cells. J . Med. Chem. 1991, 34, 2933-2935.
(b) Mulcahy, R. T.; Gipp, J . J .; Schmidt, J . P.; J oswig, C.; Borch,
R. F. Nitrobenzyl phosphorodiamidates as potential hypoxia-
selective alkylating agents. J . Med. Chem. 1994, 37, 1610-1615.
(7) Wang, S.; Low, P. S. Folate-mediated targeting of antineoplastic
drugs, imaging agents, and nucleic acids to cancer cells. J .
Controlled Release 1998, 53, 39-48.
(8) Leamon, C. P.; Weigl, D.; Hendren, R. W. Folate copolymer-
mediated transfection of cultured cells. Bioconjugate Chem.
1999, 10, 947-957.
Gr ow th In h ibition Stu d ies in Vitr o. Stock solutions of
drugs were prepared in DMSO and added to cell suspensions
so there was <1% DMSO in the final medium. The cells were
inoculated into each well of a 96-well plate (0.2 mL/well) and
cultured at 37 °C for 24 h in McCoys 5A media supplemented
with 10% heat-inactivated fetal bovine serum, penicillin (100
U/mL), and streptomycin (100 µg/mL). Supplemented McCoys
5A media (0.1 mL) containing the drug in DMSO was added,
and culture of the cells was continued at 37 °C for 72 h. Five
serial 10-fold dilutions were tested for each drug sample using
a maximum final drug concentration of 100 µg/mL. Adriamycin
and compound 10 were used as positive controls. The number
of viable cells was estimated by the MTT method10 to deter-
mine the IC50 for each drug.
(9) Turk, M. J .; Low, P. S. Unpublished results.
(10) Gorelik, E.; Ovejera, A.; Shoemaker, R.; J arvis, A.; Alley, M.;
Duff, R.; Mayo, J .; Herberman, R.; Boyd, M. Microencapsulated
tumor assay: new short-term assay for in vivo evaluation of the
effects of anticancer drugs on human tumor cell lines. Cancer
Res. 1987, 47, 5739-5747.
(11) Reddy, J . A.; Low, P. S. Folate-mediated targeting of therapeutic
and imaging agents to cancers [Review]. Crit. Rev. Ther. Drug
Carrier Sys. 1998, 15, 587-627.
(12) Shulman-Roskes, E. M.; Noe D. A.; Gamcsik, M. P.; Marlow, A.
L.; Hilton, J .; Hausheer, F. H.; Colvin, M. O.; Ludeman, S. M.
The partitioning of phosphoramide mustard and its aziridinium
ions among alkylation and P-N bond hydrolysis reactions. J .
Med. Chem. 1998, 41, 515-529.
(13) Flader, C.; Liu, J .; Borch, R. F. Development of novel quinone
phosphorodiamidate prodrugs targeted to DT-diaphorase. J .
Med. Chem. 2000, 43, 3157-3167.
Ack n ow led gm en t. Mass spectral data were pro-
vided by the Purdue University Mass Spectrometry
Facility. In vitro drug studies were carried out by Vicki
J M000306G