Synthesis and in vitro antitumor activity of thiophene analogues of 5-chloro-5,8-dideazafolic acid and 2-methyl-2-desamino-5-chloro-5,8-dideazafolic acid

Article abstract of DOI:10.1016/S0968-0896(02)00018-4

N-[5-[N-(2-Amino-5-chloro-3,4-dihydro-4-oxoquinazolin-6-yl)methylamino]-2- thenoyl]-L-glutamic acid (6) and N-[5-[N-(5-chloro-3,4-dihydro-2-methyl-4-oxoquinazolin-6-yl)methylamino]-2- thenoyl]-L-glutamic acid (7), the first reported thiophene analogues of 5-chloro-5,8-dideazafolic acid, were synthesized and tested as inhibitors of tumor cell growth in culture. 4-Chloro-5-methylisatin (10) was converted stepwise to methyl 2-amino-5-methyl-6-chlorobenzoate (22) and 2-amino-5-chloro-3,4-dihydro-6-methyl-4-oxoquinazoline (19). Pivaloylation of the 2-amino group, followed by NBS bromination, condensation with di-tert-butyl N-(5-amino-2-thenoyl)-L-glutamate (28), and stepwise cleavage of the protecting groups with ammonia and TFA yielded 6. Treatment of 9 with acetic anhydride afforded 2,6-dimethyl-5-chlorobenz[1,3-d]oxazin-4-one (31), which on reaction with ammonia, NaOH was converted to 2,6-dimethyl-5-chloro-3,4-dihydroquinazolin-4-one (33). Bromination of 33, followed by condensation with 28 and ester cleavage with TFA, yielded 7. The IC₅₀ of 6 and 7 against CCRF-CEM human leukemic lymphoblasts was 1.8±0.1 and 2.1±0.8 μM, respectively.

Full text of DOI:10.1016/S0968-0896(02)00018-4

Bioorganic & Medicinal Chemistry 10 (2002) 2067–2076
Synthesis and In Vitro Antitumor Activity of Thiophene
Analogues of 5-Chloro-5,8-dideazafolic Acid and
2-Methyl-2-desamino-5-chloro-5,8-dideazafolic Acid
Ronald A. Forsch, Joel E. Wright and Andre Rosowsky*
Dana-Farber Cancer Institute and the Department of Biological Chemistry and Molecular Pharmacology,
Harvard Medical School, Boston, MA02115, USA
Received 21 September 2001; accepted 8 November 2001
Abstract—N-[5-[N-(2-Amino-5-chloro-3,4-dihydro-4-oxoquinazolin-6-yl)methylamino]-2-thenoyl]-l-glutamic acid (6) and N-[5-[N-
(5-chloro-3,4-dihydro-2-methyl-4-oxoquinazolin-6-yl)methylamino]-2-thenoyl]-l-glutamic acid (7), the first reported thiophene
analogues of 5-chloro-5,8-dideazafolic acid, were synthesized and tested as inhibitors of tumor cell growth in culture. 4-Chloro-5-
methylisatin (10) was converted stepwise to methyl 2-amino-5-methyl-6-chlorobenzoate (22) and 2-amino-5-chloro-3,4-dihydro-6-
methyl-4-oxoquinazoline (19). Pivaloylation of the 2-amino group, followed by NBS bromination, condensation with di-tert-butyl
N-(5-amino-2-thenoyl)-l-glutamate (28), and stepwise cleavage of the protecting groups with ammonia and TFA yielded 6. Treat-
ment of 9 with acetic anhydride afforded 2,6-dimethyl-5-chlorobenz[1,3-d]oxazin-4-one (31), which on reaction with ammonia,
NaOH was converted to 2,6-dimethyl-5-chloro-3,4-dihydroquinazolin-4-one (33). Bromination of 33, followed by condensation
with 28 and ester cleavage with TFA, yielded 7. The IC50 of 6 and 7 against CCRF-CEM human leukemic lymphoblasts was
1.8ꢀ0.1 and 2.1ꢀ0.8 mM, respectively. # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
ation.5aꢁe Interestingly, with the exception of a series of
papers by Hynes and co-workers6aꢁh on 5,8-dideazaiso-
folic acid derivatives (N9–C10 bridge), among which
was 5-chloro-5,8-dideazaisofolic acid, the effect of
5-chloro substitution on the in vitro antitumor activity
of 2-amino-3,4-dihydro-4-oxoquinazoline antifolates
with a para-aminobenzoyl-l-glutamate side chain has
remained unexplored. Moreover, 5-chloro-3,4-dihydro-
4-oxoquinazoline antifolates with a conventional C9–
N10 bridge and a heterocyclic ring in place of the
phenyl ring have not been described. Thus it was of
interest to synthesize compounds 6 and 7, the first
reported examples of 5-chloro-5,8-dideazafolic acid
analogues in which the phenyl ring is replaced by thio-
phene. The structures of 1–7 are shown in Figure 1.
5-Chloro-5,8-dideazafolic acid (1) was first synthesized
by Davoll and Johnson1 and later reported by Jones
and co-workers2a,b to bind tightly to dihydrofolate
reductase (DHFR) and thymidylate synthase (TS), two
key enzymes used by cells to synthesize the pyrimidine
and purine nucleotide precursors of DNA that are
essential for normal growth. The potency of 1 as an
inhibitor of the growth of leukemia cells in culture was
essentially the same as that of 5,8-dideazafolic acid (2).
A major synthetic effort was subsequently undertaken
to optimize the activity of 5-unsusbtituted-5,8-dideaza-
folates, led to the development of N-[4-[N-(2-amino-3,4-
dihydro-4-oxoquinazolin-6-yl)methyl]-N-propargylami-
no]benzoyl-l-glutamic acid (3)3aꢁc and its more soluble
2-desamino-2-methyl analogue (4, ICI 198583).4a,b An
even better third-generation analogue, N-[5-[N-(3,4-di-
hydro-2-methyl-4-oxoquinazolin-6-yl)methyl]-N-methyl-
amino]-2-thenoyl-l-glutamic acid (5) was then
developed and subjected to intensive clinical evalu-
Chemistry
The general plan we adopted for the synthesis of 6 and 7
was similar to one that had been used earlier to prepare
the 5,8-dideazafolic acid analogues 11,2a and 85a (Fig. 1),
the sole difference being that the starting material we
chose to use was the commercially unavailable
compound 2-amino-5-methyl-6-chlorobenzoic acid (9)
*Corresponding author at: 44 Binney Street, Boston, MA 02115,
USA. Tel.: +1-617-632-3117; fax: +1-617-632-2410; e-mail:
andre_rosowsky@dfci.harvard.edu
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00018-4

2068
R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
Scheme 1. Synthesis of compound 6. Reagents: (a) CCI3CH(OH)2, H2NOH, Na2SO4; (b) (i) H2S04; (ii) separate isomers; (c) H2O2, NaOH; (d)
ꢂ
.
(CF3CO)2O; (e) (i) NaOEt, (ii) aq HCI; (iii) (t-BuO)2O, NaOH, dioxane; (g) Mel, Cs2CO3, DMF; (h) TFA; (i) C1C(¼NH)NH2 HC1, 200 C.
(Scheme 1). We envisaged that it would be possible to
obtain 9 from the isatin 10, which Baker and co-work-
ers7 described many years ago as one of two isomers
produced when the a-isonitrosoanilide 11 was heated at
65 ꢂC under strongly acidic conditions. The two isomers
were said to precipitate at different rates upon gradual
acidification of an alkaline solution of the mixture with
HCl, and were distinguishable by their color, which was
red in the case of 10 and orange in the case of the more
soluble isomer 12. In addition, recrystallized samples of
the two compounds had different melting points, which
were 242–244 ꢂC in the case of 10 and 256–258 ꢂC in the
case of 12. The structures of 10 and 12 were assigned by
oxidizing them to the corresponding anthranilic acids 9
and 13 with alkaline H2O2, followed by reductive dea-
mination with NaNO2 and H3PO2, which yielded the
previously known compound 3-methyl-4-chlorobenzoic
acid from 12 and the positional isomer 2-chloro-3-
methylbenzoic acid from 10. Although a sample of the
latter acid was not synthesized by an alternate route for
comparison, the structure of the isatin oxidation pro-
duct was eventually confirmed unequivocally by 1H
NMR,8 proving that the original formulation of 10 had
been correct. This was obviously of critical importance,
since it meant that the ultimate product of our synthesis
had to be a 5-chloro and not a 7-chloro derivative, the
latter of which would be expected to have lower anti-
folate activity.
When the original synthesis of 10 and 12 was repeated,
separation of the isomers was unexpectedly found to be
less straightforward than reported.7 In our hands, gra-
dual addition of HCl to an alkaline solution of the
crude product from the cyclization of 11 never yielded
any crops consisting only of the desired isomer 10.
Instead, mixtures of 10 and 12 invariably co-pre-
cipitated during gradual acidification and cooling, and
it was only when the crude mixture was recrystallized
directly from glacial AcOH or acetone that crops of
Figure 1. Structures of compounds 1–7.

R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
2069
pure individual isomers could be isolated. Moreover the
same isomer did not always crystallize first, and the
order of crystallization seemed to be independent of the
relative amount of each isomer in the mixture. Thus it
was necessary to carefully analyze each crop from every
crystallization by TLC in order to make sure that it
consisted of the desired isomer. Confirmation that the
ing material was recovered from the aqueous layer and
the only neutral product in the organic layer was
obviously not the desired methyl ester, since its 1H
NMR spectrum lacked an OMe signal and the two well-
defined doublets in the aromatic region were replaced
by a complex multiplet. We therefore concluded that
acid-catalyzed decarboxylation had probably occurred
even at low temperature. Although Webber and co-
workers9 had made the same observation when they
tried to esterify 2-amino-5-methyl-6-bromobenzoic acid
by this method, it was striking to see that such a facile
loss of CO2 occurs even with a smaller Cl substituent at
C6.
1
slower-moving isomer was indeed 10 came from the H
NMR spectrum which featured a pair of doublets at d
6.77 (C7-H) and d 7.53 (C6-H). By contrast, the 1H
NMR spectrum of the faster-moving isomer showed
only singlets at d 6.95 (C7-H) and d 7.55 (C4-H), prov-
ing that it was 12. The larger chemical shift of the C6
proton relative to the C7 proton in 10 and of the C4
proton relative to the C7 proton in 12 can be explained
on the basis of an electron-withdrawing effect by the 3-
keto group.
In another approach, we prepared the oxazinone 14
from 9 and trifluoroacetic anhydride in pyridine in the
hope that the oxazinone could be cleaved to 15 with
NaOEt in dry EtOH. In the event, the only product
isolated after these reactions was not 15, but rather the
N-trifluoroacetylated acid 16, probably arising via
attack at the imine carbon of 14, followed by hydrolysis
of the putative imino ether intermediate 17 upon neu-
tralization with aqueous acid. The structure of 16 was
In the course of trying to optimize the yield from the
ring closure reaction, we made the interesting observa-
tion, not noted earlier,7 that the temperature at which
ring closure was carried out could markedly influence
regioselectivity. Thus, when the reaction was performed
at 60 ꢂC the ratio of 10 to 12 was ca. 2:1, whereas at
70 ꢂC the two isatins formed in nearly equal amount.
The fact that the sterically more hindered isomer 10 was
favored at the lower temperature suggested that the
formation of 10 may be kinetically controlled whereas
the less hindered product 12 is thermodynamically
preferred.
1
evident from its H NMR spectrum, in which the aro-
matic doublets appeared at d 7.73 and d 7.93, whereas
the corresponding doublets in the anthranilic acid 9
were much further upfield at d 6.63 and d 7.08.
Although this unfavorable outcome was disappointing,
it could be explained by increased chemical reactivity of
the imine carbon next to the electron-withdrawing CF3
group.
Oxidation of 10 with alkaline H2O2 afforded anthranilic
acid 9 as reported,7,8 but there was again a small dis-
crepancy between our results and those of the earlier
investigators. Whereas our observed melting point for
this compound after recrystallization was 167–168 ꢂC,
the value given in the literature is approximately eleven
degrees lower,7,8 suggesting that a completely pure
sample may not have been used in the earlier papers. In
order to ensure that we had the right compound, isatin
12 was oxidized to 13, and the melting point of the latter
was found to be more or less in agreement with the
published value of 210–211 ꢂC, although our sample
actually had a double melting point, first at 210–211 ꢂC
and then, after resolidifying briefly, at 213–214 ꢂC. The
yield of both 9 and 13 was 80%. The most interesting
difference in the behavior of the two anthranilic acids,
not mentioned by the earlier workers, was that 9
underwent vigorous gas evolution at its melting tem-
perature, whereas 13 did not. The facile decarboxylation
of 9 was consistent with the crowded 1,2,3,4-tetra-
substituted structure of this molecule. The effect of
steric hindrance on the behavior of the carboxyl group
in 9 also became evident from some unexpected chem-
istry which is discussed below.
4-Trifluoromethylisatin has been converted to 5-tri-
fluoromethylisatoic anhydride upon oxidation with
monoperphthalic acid, and the anhydride has been
converted to 2-amino-5-trifluoromethyl-3,4-dihydro-4-
oxoquinazoline by further reaction with guanidine car-
bonate.10 We reasoned that if we could isolate 5-chloro-
6-methylisatoic anhydride (18), it might be possible to
convert it directly to 2-amino-5-chloro-3,4-dihydro-6-
methyl-4-oxoquinazoline (19) by reaction with guani-
dine. As expected, oxidation of 10 with mono-
perphthalic acid or m-chloroperbenzoic acid was rapid
as judged by the disappearance of the characteristic red
color of the isatin. However subsequent heating with
guanidine did not yield 19, but instead a product that
was highly water-soluble at pH 8 and whose identity
was not determined.
The nettlesome problems discussed above were finally
resolved when 9 was found to be readily converted to
the Boc derivative 20 with di-tert-butyl dicarbonate and
then to the N-protected ester 21 with Cs2CO3 and MeI
in DMF with a combined two-step yield of 70%
(Scheme 1). Deprotection of 21 with TFA was quanti-
tative, affording the amino ester 22. Although our
expectations that heating 22 with guanidine or guani-
dine carbonate would yield quinazoline 19 failed to
materialize, when 22 was heated with chlor-
oformamidine hydrochloride in diglyme at 200 ꢂC,9 the
desired product was obtained in 79% yield. Acylation of
the amino group with pivaloyl chloride afforded the
more soluble and easily recrystallized derivative 23
Because we had first intended to cyclize 9 to a 2-amino-
quinazolin-4(3H)-one by reaction of its methyl ester
with guanidine, we attempted to esterify the acid via a
standard reaction with MeOH in the presence of SOCl2
at 10 ꢂC. To our surprise, when the reaction product
was partitioned between aqueous NaHCO3 and an
organic solvent, approximately two-thirds of the start-

2070
R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
Scheme 2. Synthesis of compound 7. Reagents: (a) PivCl, Et3N, THF; (b) NBS, Bz2O2, CHCI3 (c) KMnO4, NaH2PO4, aq Me2CO; (d) (i) SOCI2 (ii)
di-tert-butyl l-glutamate, Et3N, CH2Cl2 (e) Fe powder, FeSO4, aq MeOH; (f) 24, NaHCO3, DMF; (g) 1:3 TFA–CH2Cl2, 5 ꢂC.
(86% yield), which was then brominated with NBS/
Bz2O2 in CHCl3 to obtain bromide 24 (Scheme 2).
Although it was possible to obtain an analytical sample
of 24 by recrystallization from MeOH, there was sub-
stantial loss of material, and we therefore took this
compound to the next step without purification. That
bromination of 23 had occurred on the 6-Me group was
evident from the 1H NMR spectrum of 24, in which the
Me singlet at d 2.47 was replaced by a singlet at d 4.72.
Moreover, the doublet we had observed at d 7.57 for the
C7 aromatic proton in the spectrum of 23 was now at d
7.72. Interestingly, the other aromatic proton in com-
pounds 23 and 24 had almost the same chemical shift
(ca. d 7.3).
a gum that could not be induced to crystallize even
though its purity was of microanalytical quality and its
identity was fully supported by its H NMR spectrum,
1
in which the protons of the thiophene ring were readily
discerned as a pair of doublets at d 7.50 and d 7.88,
respectively. Reduction of the nitro group was per-
formed with a mixture of activated Fe powder and
FeSO4 in refluxing MeOH to obtain the amino ester 28
(60% crude yield). Complete reduction of the NO2
1
group was verified by the H NMR spectrum, in which
the doublets for protons on the thiophene ring were
shifted upfield to d 6.08 and d 7.22. The doublet with the
high chemical shift (d 7.88) in the spectrum 27 was
assigned to C3-H because this proton is next to the
electron-withdrawing NO2 group. Using analogous rea-
soning, the doublet with the low chemical shift (d 6.08)
in the spectrum of 28 was assigned to C3-H on the basis
that this proton would now be shielded rather than
deshielded. By a process of exclusion the doublets at d
7.50 in 27 and d 7.22 in 28 were therefore assigned to
C4-H.
With the synthesis of 24 completed, the next task was to
synthesize the right-hand fragment of the target com-
pound 6. Although dimethyl or diethyl esters had gen-
erally been used by other workers in the past to protect
the glutamate moiety during the synthesis of quinazo-
line antifolates, pilot experiments showed that, perhaps
because of the absence alkyl substitution on nitrogen,
the aminothiophene ring would be extensively damaged
in aqueous base. Moreover, while the thiophene moiety
appeared to be stable in methanolic ammonia at room
temperature, we remained concerned that even these
mild conditions might convert methyl esters to amides,
and thus felt that protection of the glutamate side chain
would be more appropriate with tert-butyl esters.
Accordingly, we synthesized the previously unknown di-
tert-butyl ester 28 (Scheme 2). Oxidation of commer-
cially available 5-nitro-2-thiophenecarboxaldehyde (25)
with KMnO4 in the presence of NaH2PO4 afforded acid
26, which on reaction with SOCl2 followed by addition
of di-tert-butyl l-glutamate was converted to the nitro
ester 27 (93% crude yield). The successful synthesis of
26 from 25 with KMnO4 was in surprising contrast to
an earlier report11 that this oxidation does not work,
but instead requires alkaline AgNO3.11,12 Diester 27 was
Coupling of the pivaloylated bromide 24 with amine 28
to form 29 was performed by stirring equimolar
amounts of the reactants with two mols of NaHCO3 in
DMF at room temperature for 2 days. The yield of 29
after chromatography was 69%. Attempted acidolysis
of the pivaloyl and ester groups in a single reaction
using TFA were unsuccessful, as only the latter were
cleaved. Thus, 29 was deprotected by treatment with
anhydrous methanolic ammonia at room temperature
for 20 h to remove the pivaloyl group, followed by TFA
in CH2Cl2 at 5 ꢂC for 20 h to cleave the esters. The
resulting diacid (5) was purified by a two-stage process
involving preparative HPLC on a C18 column, followed
by ion-exchange chromatography on a DEAE-cellulose
(HCO3- form). Because a number of side products
formed during the two-stage deprotection scheme, the
final yield of highly pure material suitable for biological

R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
2071
Scheme 3. Reagents: (a) Ac2O, reflux ; (b) NH3, ꢁ33 ꢂC; (c) 1 N NaOH, reflux; (d) NBS/Bz2O2, CHCl3 (e) 28, NaHCO3, DMF; (1) 1:3 TFA–
CH2Cl2, 5 ꢂC.
testing was still only 15%. Although protection of the
side chain with acid-cleavable tert-butyl groups appar-
ently did not provide the advantage we had hoped for, it
should be noted that in the only example published thus
far of the use of tert-butyl groups during the synthesis
of a 5,8-dideazafolate containing a 4-aminothiophene
ring,5a the bridge nitrogen was substituted with a cya-
nomethyl group which may have had a protective effect.
All the other thiophene analogues in the series, includ-
ing one in which the nitrogen on the ring was unsub-
stituted, were made from diethyl l-glutamate.
the expectation that any non-brominated starting mate-
rial would be easy to remove by chromatography at the
next stage. The spectrum of 30 showed a singlet at d
2.33, attributable to the 2-Me group, and a two-proton
singlet at d 4.87 which we assigned to the CH2Br group
at C6 because it was approximately in the same region
as that of the CH2Br group in 24 (d 4.72). That bromi-
nation had occurred at the 6-position rather than the 2-
position was also consistent with an earlier study which
unequivocally established that bromination of 3,4-dihy-
dro-2,6-dimethyl-4-oxoquinazoline yields the 6-bromo-
methyl derivative.13 Moreover, as in the bromination of
23, the doublet for the C7 aromatic proton was dis-
placed downfield from d 7.73 to d 7.97, and the C8
proton likewise showed a downfield shift, albeit a smal-
ler one, from d 7.42 to d 7.52. The magnitude of this
shift was greater than would have been expected if the
2-Me group had undergone bromination. In agreement
with earlier data for 2-amino-3,4-dihydro-4-oxoquin-
azolines14a,b versus 3,4-dihydro-2-methyl-4-oxoquina-
zolines4a lacking a C5 substituent, when the chemical
shifts of the C7 and C8 protons in 33 were compared
with those of the C7 and C8 protons in 19, the latter
both had higher d values consistent with the ability of
the 2-amino group to increase electron density at C8 via
a resonance effect. Moreover, the absence in either 19 or
33 of a doublet below d 8.0 confirmed that the values
assigned originally to the C5 proton in 5-unsubstituted
6-bromomethyl-3,4-dihydro-4-oxoquinazolines14a,14b,15
was correct.
An alternative approach to the synthesis of 5 in which
closure of the quinazolinone ring would be left to the
end was also briefly explored. Thus, bromination of 21
with NBS/Bz2O2 and condensation of the resulting
bromide with aminothiophene 28 afforded a carbamate
triester which we proposed to subject to acidolysis of the
Boc and tert-butyl ester groups, followed by condensa-
tion with guanidine. In the event, even though pilot
experiments showed that the synthesis of the carbamate
triester quite feasible, this approach was abandoned
after a model ring closure reaction with guanidine and
the non-brominated amino ester 22 was found to be
completely unsuccessful (see above).
We next turned our attention to the synthesis of the 2-
methyl-2-desamino analogue
6 via the previously
unknown bromide 30, which was prepared according to
Scheme 3. Treatment of 9 with Ac2O under reflux
afforded the benz[1,3-d]oxazine 31 in 87% yield, and
further reaction with liquid ammonia presumably yiel-
ded the anthranilamide 32, which was directly cyclized
to 33 with hot aqueous NaOH. The structure of 31 was
The synthesis of 7 was completed by condensing bro-
mide 30 with aminothiophene 28 in DMF in the pre-
sence of NaHCO3 to obtain diester 34 (56%) and
deprotecting the latter with TFA in CH2Cl2 at 5 ꢂC.
Whereas the yield of 6 from 29 had been only 15%, that
of 7 from 34 after purification by HPLC and ion-
exchange chromatography was more than 2-fold higher
(39%). In agreement with the method of isolation of
these compounds by freeze-drying of NH4HCO3 eluates
from ion-exchange columns, microchemical analysis
indicated that 6 and 7 were both solvated with 3.5 mol
of H2O, and also contained 0.5 and 1.0 mol of ammo-
nia, respectively.
1
easily identified by a pair of H NMR singlets at d 2.43
and d 2.50 for the Me groups. Two singlets were like-
wise observed in the spectrum of 33, this time at d 2.30
and d 2.40, and it was assumed, as has been done with
the corresponding 5-unsubstituted compounds,13 that
the upfield signal in each case corresponded to the Me
group at C6. Bromination of 33 with NBS/Bz2O2 in
CHCl3 under reflux afforded bromide 30, which pre-
cipitated directly from the reaction mixture (crude yield
77%) and was used directly for the next reaction with

2072
R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
Biological Activity
Synthesis and separation of 4-chloro-5-methylisatin (10)
and 6-chloro-5-methylisatin (12). Step 1. Solid chloral
hydrate (45 g, 0.27 mol) was added to a suspension of
Na2SO4 (286 g, 2.01 mol) in H2O (965 mL). In another
flask, a solution of 3-chloro-4-methylaniline hydrochlo-
ride was prepared by adding the free amine (35.4 g, 0.25
mol) to 12 N HCl (21.5 mL) and H2O (150 mL), and
warming the mixture gently until all the solid dissolved.
This warm solution was then added, all in one portion,
to the solution of chloral while stirring manually to
Compounds 6 and 7 were tested as inhibitors of the
growth of CCRF-CEM human leukemic lymphoblasts
as described by McGuire and co-workers.16 The con-
centrations at which 6 and 7 inhibited cell growth by
50% (IC50) after 120 h of drug exposure in medium
containing 10% horse serum were 1.8ꢀ0.1 and 2.1ꢀ0.8
mM, respectively. The IC50 values reported for 5 against
human WI-L2 lymphoblastic leukemia cells, which are
likewise T-cells, was 0.0035 mM.16 In assays against
L1210 murine leukemic cells the N10-methyl group in 5
is known to increase growth-inhibitory potency versus
the corresponding N10-unsubstituted compound at
least 100-fold.5d,e While it may be speculated that simi-
lar modification of 6 and 7 would similarly have a
favorable effect on biological activity, we did not
address this question in the present work.
.
obtain a fine dispersion. A solution of H2NOH HCl (55
g, 0.79 mol) in H2O (250 mL) was then added in one
portion, and manual stirring continued until all the
Na2SO4 dissolved and the organic solids were finely
dispersed. The mixture was heated and stirred for 40
min until it came to a vigorous boil, kept at this tem-
perature for 2 min, cooled by immersion in cold water,
and left to stand overnight to facilitate filtration. The
solid was collected, washed with H2O, and dried in a
lyophilization apparatus to obtain the isonitrosoanilide
11 as a beige solid (median crude yield from several
Experimental
runs: 50.3 g, 95%), mp 168–169 ꢂC. H NMR (DMSO-
1
IR spectra were obtained on a Perkin Elmer model 281
spectrophotometer and UV spectra on a Varian model
210 instrument. For the sake of brevity, only IR peaks
with wave numbers greater than 1200 cmꢁ1 are repor-
ted, and very weak peaks and shoulders are omitted. 1H
NMR spectra were recorded at 60 MHz with a Varian
model EM360L instrument using Me4Si as the refer-
ence, or in the case of 6 and 7 at 200 MHz with a Varian
model Mercury 200 instrument. The very broad amide
NH signal in several compounds was barely discernible
at 60 MHz is therefore recorded only for 6 and 7.
Analytical TLC was performed on fluorescent What-
man MK6F silica gel-coated glass slides, with spots
viewed under 254 nm illumination. Column chromato-
graphy was on Baker silica gel (regular grade, 60–200
mesh; flash grade 40-mm particle size) or on Whatman
DE-52 preswollen DEAE-cellulose (HCO3-form).
HPLC separations were on C18 silica gel radial
compression cartridges (Millipore, Milford, MA, USA;
analytical, 5-mm particle size, 5 ꢃ 100 mm; preparative,
15-mm particle size, 24 ꢃ 100 mm). In those instances
where preparative HPLC was followed by ion-
exchange chromatography, the latter step was in
order to ensure that the sample was not con-
taminated with traces of silica gel from the C18 col-
umn. MOISTURE-sensitive reactions were carried out
in solvents that were of Sure-Seal grade (Aldrich, Mil-
waukee, WI, USA) or had been stored over Linde 4A
molecular sieves. Solids were generally dried over P2O5
at 50–80 ꢂC in a vacuum oven or Abderhalden app-
aratus. Melting points (not corrected) were obtained on
a Fisher–Johns hot-stage microscope or in open Pyrex
capillary tubes in a Mel-Temp apparatus (Cam-
bridge Laboratory Devices, Cambridge, MA, USA).
Starting materials and other reagents and chemicals
were purchased from Aldrich (Milwaukee, WI, USA),
Lancaster (Windham, NH, USA), or Fisher (Boston,
MA, USA). Microanalyses were performed by Quanti-
tative Technologies, Whitehouse, NJ, USA and were
within ꢀ0.4% of calculated values unless otherwise
noted.
d6) d 2.28 (s, 3H, Me), 7.33 (m, 2H, C5-H and C6-H),
7.65 (s, 1H, CH¼N), 7.90 (d, J=2 Hz, 1H, C2-H), 10.28
(br s, 1H, OH, exchangeable with D2O). Recrystalliza-
tion of a small sample from aqueous MeOH raised the
melting point to 170–172 ꢂC (lit.8 177–179 ꢂC), but
because there was considerable loss of material during
purification the freeze-dried crude product was used in
the next reaction directly.
Step 2. Crude 11 (50.3 g, 0.237 mol) was added in
small portions to concentrated H2SO4 (223 mL) at 60–
65 ꢂC at such rate that the internal temperature did not
exceed 65 ꢂC. When addition was complete, the solution
was warmed to 80 ꢂC and kept at this temperature for
10 min, then cooled to room temperature by immersion
in cold water, and poured into ten volumes of ice. After
being kept at room temperature overnight to facilitate
filtration, the solid was collected, washed with H2O to
remove the acid, and taken up in 0.4 N NaOH (1 L).
The insoluble portion was separated by filtration, and
the filtrate was acidified, left to stand overnight, and
filtered to obtain an orange-red solid (41.1 g, 89% crude
yield) consisting of a mixture of 10 and 12. In some
experiments in which the acidified mixture was filtered
after only 1 h, the filter cake was washed with H2O, and
the filtrate was left to stand overnight, a small amount
of solid precipitated which turned out to be pure 12 (2%
yield).
A 10 g portion of the foregoing mixture of 10 and 12
was dissolved in warm acetone (250 mL) and the solu-
tion was left to cool in a beaker covered with aluminum
foil. The first crop was a red powder consisting of pure
10 (3.73 g) mp 245–247 ꢂC (lit.7 242–244 ꢂC); TLC: Rf
0.4 (silica gel, 1:1 isooctane–Me2CO). 1H NMR
(DMSO-d6) d 2.23 (s, 3H, Me), 6.77 (d, J=8 Hz, 1H,
C7-H), 7.53 (d, J=8 Hz, 1H, C6-H). Reduction of the
volume to 175 mL yielded another 0.66 g of red powder
consisting mostly of 10 but also containing some 12.
TLC of the mother liquor showed that it now contained

R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
2073
mostly 12. Although it can be estimated that the
amount of pure 10 theoretically recoverable from the
original 41.1 g mixture if the entire sample were recrys-
tallized in one batch would be 15.5 g (33%), actual
recoveries of pure 10 from different recrystallizations
were not always the same.
J=8 Hz, 1H, C3- or C4-H), 8.02 (br s, 1H, NH), 8.68
(br s, 1H, OH). Anal. calcd for C13H16ClNO4: C, 54.65;
H, 5.64; N, 4.90; Cl. 12.41. Found: C, 54.71; H, 5.52; N,
4.94; C1, 12.71.
Methyl 2-(N-tert-butyloxycarbonyl)amino-5-methyl-6-
chlorobenzoate (21). A stirred suspension of Cs2CO3
(8.06 g, 0.0247 mol) in dry DMF (100 mL) was treated
first with 20 (14.1 g, 0.0495 mol) and then, after 15 min,
with MeI (3.10 mL, 7.05 g, 0.0493 mol). Stirring was
continued for 24 h, during which all the solids dissolved.
The DMF was evaporated and the residue partitioned
between EtOAc and brine. Rotary evaporation of the
EtOAc layer afforded a beige solid which was recrys-
tallized from MeOH to obtain 21 as an off-white pow-
der (13.1 g, 88%): mp 135 ꢂC sharp. IR (KBr) nmax
3210, 3170, 3110, 2980, 2930, 1735, 1705, 1600, 1570,
1495, 1460, 1390, 1370, 1360, 1300, 1280, 1265, 1235
A 4.7 g portion of the isomer mixture was recrystallized
from glacial AcOH (275 mL). The first crop (2.1 g)
consisted of pure 12 as an orange powder: mp 255–
256 ꢂC (lit.7 mp 256–258 ꢂC). TLC Rf 0.5 (silica gel, 1:1
isooctane–Me2CO). 1H NMR (DMSO-d6) d 2.26 (s, 3H,
Me), 6.95 (s, 1H, C7-H), 7.55 (s, 1H, C4-H). The second
crop (1.8 g) was less pure by TLC, and the supernatant
contained mostly 10. The estimated amount of recover-
able pure 12 from the first crop of the 41.1 g mixture if
the entire sample had been recrystallized as a single
batch was 18.4 g (45%).
1
cmꢁ1. H NMR (CDCl3) d 1.52 (s, 9H, t-Bu), 2.35 (s,
2-(N-tert-Butyloxycarbonyl)amino-5-methyl-6-chloroben-
zoic acid (20). Step 1. A stirred solution of 10 (3.91 g,
0.02 mol) in 5% NaOH (100 mL) was treated dropwise
over 10 min with 30% H2O2 (5.7 mL, calculated to
contain 1.71 g, 0.05 mol). After another 20 min of being
stirred, during which it became warm and effervesced,
the solution was cooled in an ice-bath and acidified to
pH 4 with 3 N HCl. The precipitate was collected and
dried in a lyophilizer to obtain 2-amino-5-methyl-6-
chlorobenzoic acid (9) as a beige powder (2.96 g, 80%):
3H, aromatic Me), 3.98 (s, 3H, OMe), 7.28 (d, J=9 Hz,
1H, C3-H), 7.78 (d, J=9 Hz, 1H, C4-H). Anal. calcd for
C14H18ClNO4: C, 56.10; H, 6.05; N, 4.67; Cl, 11.83.
Found: C, 55.90; H, 5.92; N, 4.61; Cl, 11.94.
Methyl 2-amino-5-methyl-6-chlorobenzoate (22). Com-
pound 21 (1.5 g, 0.005 mol) was dissolved in TFA (7
mL), the solution was evaporated to dryness under
reduced pressure, and the residue was partitioned
between EtOAc and 5% NaHCO3. Evaporation of the
EtOAc layer yielded an oil (1.1 g), a small portion of
which was purified by flash chromatography on silica
gel (5 g, 1.5 ꢃ 9 cm) using 2:1 isooctane–EtOAc as the
eluent. The resulting material was still an oil, but was
homogeneous by TLC (Rf 0.2, silica gel, 2:1 isooctane–
EtOAc). IR (thin film) nmax 3470, 3380 2950, 2920,
2850, 1710, 1620, 1565, 1485, 1460, 1445, 1400, 1295,
mp 163–164 ꢂC dec, gas evolution. H NMR (DMSO-
1
d6) d 2.13 (s, 3H, Me), 6.63 (d, J=9 Hz, 1H, C3-H), 7.08
(d, J=9 Hz, 1H, C4-H), 7.3–8.2 (m, 2H, NH2). Recrys-
tallization of part of this solid from EtOAc-isooctane
afforded glistening beige flakes: mp 167–168 ꢂC, vigor-
ous gas evolution (lit.8 mp 156–157 ꢂC; lit.7 mp 156–
158 ꢂC). Most of the product was used without recrys-
tallization.
1
1275, 1225 cmꢁ1. H NMR (CDCl3) d 2.23 (s, 3H, aro-
matic Me), 3.92 (s, 3H, OMe), 6.52 (d, J=9 Hz, 1H, C3-
H), 7.05 (d, J=9 Hz, 1H, C4-H). MS: m/e 199 (M,
35Cl), 200 (M+1, 35Cl), 201 (M, 37Cl), 202 (M+1, 37Cl).
Anal. calcd for C9H10ClNO2: C, 54.15; H, 5.05; N, 7.02;
Cl, 17.76. Found: C, 53.70; H, 4.89; N, 6.88; Cl, 17.49.
The same procedure with isatin 12 (3.91 g, 0.02 mol)
afforded 2-amino-5-methyl-4-chlorobenzoic acid (13) as
a beige solid (2.95 g, 80%) after recrystallization from
toluene: mp 210–211 ꢂC (lit.7 mp 210–211 ꢂC), with
resolidification and a second mp at 213–214 ꢂC. In con-
trast to 9, there was very little gas evolution at the mp of
13.
2-Amino-5-chloro-3,4-dihydro-4-oxo-6-methylquinazoline
(19). Chloroformamidine hydrochloride was freshly
prepared by bubbling HCl gas through a solution of
cyanamide (0.5 g, 0.012 mol) in Et2O (100 mL), collect-
ing the precipitate, and drying it under vacuum. This
solid was then added to a solution of the entire product
from another TFA hydrolysis of 21 (1.5 g, 0.005 mmol)
in diglyme (8 mL), and the resulting slurry was plunged
into an oil bath preheated to 200 ꢂC. The mixture
became homogenous after a few min, and then solidi-
fied. Heating was continued for another 10 min, and
after being allowed to cool to 80 ꢂC the solid was tritu-
rated with EtOH (15 mL), collected, and washed with
Et2O. The solid was then dissolved in a mixture of 1 N
HCl (17 mL), DMF (20 mL), and H2O (20 mL) at
80 ꢂC. The solution was left to cool to room tempera-
ture and basified to pH 8 with concentrated NH4OH (5
mL). The resulting precipitate was gelatinous but
became granular and easier to filter when the mixture
was reheated to 60 ꢂC and left to cool. The solid was
Step 2. A solution of 9 (11.7 g, 0.0633 mol) in 2 N
NaOH (32 mL) and dioxane (65 mL) was treated with
di-tert-butyl dicarbonate (15.2 g, 0.0696 mol) and stir-
red at room temperature for 2 days. The dioxane was
evaporated and the residue partitioned between EtOAc
and H2O. The aqueous layer was washed with EtOAc to
remove neutral impurities, then acidified to pH 2 with
3 N HCl and re-extracted with EtOAc. Evaporation to
dryness yielded a foam which solidified on prolonged
standing. Recrystallization from a mixture of benzene
and isooctane yielded a greenish-gray powder (14.2 g,
79%): mp 130–131 ꢂC, vigorous gas evolution. IR (KBr)
nmax 3200, 2980, 2920, 2620, 2550, 1710, 1665, 1600,
1575, 1490, 1455, 1415, 1390, 1370, 1290, 1255, 1225
1
cmꢁ1. H NMR (CDCl3) d 1.50 (s, 9H, t-Bu), 2.33 (s,
3H, Me), 7.32 (d, J=8 Hz, 1H, C3- or C4-H), 7.78 (d,

2074
R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
filtered, washed with H2O (2 ꢃ 15 mL) and EtOH (2
mL), and dried over P2O5 in vacuo at 75 ꢂC to obtain 19
as a white solid (0.825 g, 79%). When the reaction was
scaled up four-fold the yield was 68%: mp >300 ꢂC;
TLC: Rf 0.5 (silica gel, 30:4:1 CHCl3–MeOH–AcOH).
IR (KBr) nmax 3450, 3320, 3120, 1665, 1600, 1555, 1515,
moderate exotherm. After several min, the solvent was
evaporated under reduced pressure and the residue par-
titioned between EtOAc and H2O. The organic layer
was washed with 0.5 N HCl, H2O, 5% NaHCO3, and
brine, then evaporated to dryness to obtain diester 27
(11.4 g, 93%) as a brown gum suitable for use in the
next reaction. The analytical sample was obtained by
flash chromatography (30:1 w/w silica gel, 2:1 iso-
octane–EtOAc) and drying in vacuo at 65 ꢂC over P2O5;
TLC: Rf 0.3 (silica gel, 2:1 isooctane–EtOAc). IR (thin
film) nmax 3100, 2970, 2930, 2820, 1735, 1660, 1545,
1505, 1475, 1450, 1430, 1390, 1365, 1335, 1285, 1250
1
1470, 1370, 1325, 1300, 1230 cmꢁ1. H NMR (DMSO-
d6) d 2.33 (s, 3H, Me), 6.53 (br m, 2H, NH2), 7.12 (d,
J=8 Hz, 1H, C7-H), 7.55 (d, J=8 Hz, 1H, C8-H). MS:
m/e 210 (M+1, 35Cl), 37Cl peak obscured by the matrix.
Anal. calcd for C9H8ClN3O.0.25H2O: C. 50.48; H, 4.00;
N, 19.62; Cl, 16.56. Found: C, 50.21; H, 3.98; N, 19.31;
Cl, 16.91.
1
cmꢁ1. H NMR (CDCl3) d 1.45 (s, 9H, t-Bu), 1.50 (s,
9H, t-Bu), 1.90–2.67 (m, 4H, b- and g-CH2), 4.63 (m,
1H, a-CH), 7.50 (d, J=4 Hz, 1H, thiophene C4-H), 7.88
(d, J=4H, 1H, thiophene C3-H). Anal. calcd for
C18H26N2O7S: C, 52.16; H, 6.32; N, 6.76; S, 5.74.
Found: C, 52.00; H, 5.98; N, 7.17; S, 5.45.
5-Chloro-3,4-dihydro-6-methyl-4-oxo-2-(N-pivaloylamino)-
quinazoline (23). A suspension of 19 (2.78 g, 0.0133 mol)
in dry THF (200 mL) was treated with Et3N (2.78 mL,
2.02 g, 0.02 mol) and pivaloyl chloride (2.46 mL, 2.41 g,
0.02 mol), and the mixture was refluxed for 20 h, then
cooled and filtered. The filtrate was evaporated under
reduced pressure, and the residue crystallized from
MeOH to obtain 23 as white needles (2.55 g, 65%). The
mother liquor was evaporated to dryness, the residue
was extracted with hot EtOAc, and the solid which was
insoluble in EtOAc was recrystallized from MeOH
afford an additional amount of 23 (0.56 g (14%). The
filter cake from the original reaction mixture was dried
and the pivaloylation reaction repeated, yielding an addi-
tional 0.26 g (8%) of 23; total yield 3.37 g (86%): mp
232–233 ꢂC. IR (KBr) nmax 3240, 3170, 2960, 1655,
1640, 1595, 1550, 1515, 1490, 1455, 1400, 1380, 1370,
Di-tert-butyl N-(5-amino-2-thenoyl)-L-glutamate (28).
Iron powder (14 g, 0.25 mol)5a was kept under 2 N HCl
for 10 min with occasional swirling, then filtered,
washed with H2O, rinsed with acetone, and dried in
vacuo. The activated iron and a catalytic amount of
.
FeSO4 7H2O (4.63 g, 0.016 mol) were added to a solu-
tion of the nitro diester 27 (5.9 g, 0.014 mol) in a mix-
ture of MeOH (75 mL) and H2O (25 mL). After being
refluxed for 24 h, the mixture was filtered through Celite
on a bed of glass fiber paper, and the filtrate was con-
centrated to a small volume to remove the MeOH. The
product was then partitioned between EtOAc and H2O,
and the organic layer was evaporated to a gum (5.27 g),
which was purified by flash chromatography (silica gel,
60 g, 5 ꢃ 8 cm, 1:1 isooctane–EtOAc) to obtain 27 as a
foam (3.28 g, 60%). The analytical sample was obtained
by passing a 0.2 g portion of this material through a
second column of flash-grade silica gel (10 g, 20:1 w/w),
and drying the resulting gum in vacuo over P2O5 at
70 ꢂC until it formed a glass that could be ground to a
powder: mp 116–117 ꢂC; TLC: Rf 0.3, turning orange on
standing overnight (silica gel, 1:1 isooctane–EtOAc). IR
(KBr) nmax 3440, 3380, 3310, 3200, 2970, 2930, 1715,
1315, 1290, 1245 cmꢁ1 1H NMR (CDCl3+2 drops
.
DMSO-d6) d 1.35 (s, 9H, t-Bu), 2.47 (s, 3H, Me), 7.28
(d, J=8 Hz, 1H, C7-H), 7.57 (d, J=8 Hz, C8-H). Anal.
calcd for C14H16ClN3O2: C, 57.24; H, 5.49; N, 14.30; Cl,
12.07. Found: C, 57.32; H, 5.54; N, 14.11; Cl, 11.86.
Di-tert-butyl N-(5-nitro-2-thenoyl)-L-glutamate (27).
Step 1. A stirred suspension of KMnO4 (15.8 g, 0.10
mol) in 5% aq NaH2PO4 (150 mL) was added in a sin-
gle portion to a stirred solution of 25 (15.7 g, 0.10 mol)
in acetone (250 mL). The reaction was mildly exother-
mic. After the mixture was stirred at ambient tempera-
ture for 45 min, the brown MnO2 precipitate was
removed by filtration and the filter cake was washed
with H2O and acetone. Acetone was removed from the
combined filtrates, the mixture was acidified with 2 N
HCl, and the precipitated solid was filtered, dried in
vacuo, and used directly in the next step. After being
recrystallized from H2O, the product melted at 153–
155 ꢂC (lit.12a 157–158 ꢂC, lit.12b 157 ꢂC, lit.12b 155–
157 ꢂC).
1625, 1545, 1515, 1470, 1365, 1295, 1235 cmꢁ1 1H
.
NMR (CDCl3) d 1.43 (s, 9H, t-Bu), 1.48 (s, 9H, t-Bu),
2.23 (m, 4H, b- and g-CH2), 4.22 (m, 1H, a-CH), 6.08
(d, J=4 Hz, 1H, thiophene C3-H), 7.22 (d, J=4 Hz,
1H, thiophene C4-H). Anal. calcd for C18H28N2O5S: C,
56.23; H, 7.34; N, 7.29; S, 7.74. Found: C, 56.32; H,
7.16; N, 7.17; S, 7.64.
N-[5-[N-(2-Amino-5-chloro-3,4-dihydro-4-oxoquinazolin-
6-yl)methylamino]-2-thenoyl]-L-glutamic acid (6).
Step 1. A suspension of 23 (669 mg, 2.23 mmol) in
CHCl3 (10 mL, previously passed through silica gel to
remove EtOH) was treated with NBS (460 mg, 2.58
mmol) and 20 mg of Bz2O2, and the mixture was
refluxed for 1 h. A second 20 mg portion of Bz2O2 was
added, reflux was resumed for 20 h, and the solvent
was evaporated to dryness under reduced pressure. The
residue was partitioned between EtOAc and H2O, a
small amount of MeOH was added to dissolve any
remaining solid, and the organic layer was separated,
washed with brine, and evaporated to dryness to
Step 2. Compound 26 (6.92 g, 0.04 mol) was added
slowly (foaming!) to SOCl2 (30 mL), and the reaction
mixture refluxed for 30 min and evaporated under
reduced pressure to obtain the acid chloride as a gum.
The crude acid chloride was taken up directly in CH2Cl2
(50 mL) and added to a solution of di-tert-butyl l-glu-
tamate hydrochloride (8.88 g, 0.03 mol) in CH2Cl2 (100
mL). The resulting solution was treated immediately
with Et3N (6.9 mL, 5.06 g, 0.05 mol), producing a

R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
2075
obtain 24 as a white powder (837 mg, 99%), which was
used directly in the next step; mp 179–180 ꢂC (MeOH).
IR (KBr) nmax 3350, 3410, 3120, 2950, 2850, 1665, 1630,
1585, 1535, 1475, 1440, 1385, 1375, 1355, 1300, 1255,
5-Chloro-2,6-dimethylbenzo[d][1,3]oxaxine-4-one (31). A
mixture of the anthranilic acid 9 (5.57 g, 0.03 mol) and
Ac2O (50 mL) was refluxed for 1 h and evaporated to
dryness under reduced pressure. Recrystallization from
isooctane and EtOAc yielded straw-colored crystals
(5.13 g, plus a second crop of 0.29 g; total 87%); mp
180–181 ꢂC. IR (KBr) nmax 1925, 1750, 1660, 1590,
1555, 1465, 1335, 1395, 1370, 1295, 1275, 1255, 1215
cmꢁ1. 1H NMR (CDCl3) d 2.43 (s, 3H, 2- or 6-Me), 2.50
(s, 3H, 2- or 6-Me), 7.38 (d, J=8 Hz, C8-H), 7.68 (d,
J=8 Hz, C7-H). Anal. calcd for C10H8ClNO2; C, 57.30;
H, 3.85; N, 6.68; Cl, 16.91. Found: C, 57.41; H, 4.01; N,
6.51; Cl, 17.04.
1
1230 cmꢁ1. H NMR (CDCl3+2 drops DMSO-d6) d
1.35 (s, 9H, t-Bu), 4.72 (s, 2H, CH2Br), 7.30 (d, J=8
Hz, 1H, C7-H, partly obscured by CHCl3), 7.72 (d, J=8
Hz, 1H, C8-H).
Step 2. A solution of bromide 24 (372 mg, 1.0 mmol)
in DMF (10 mL) was treated with amino diester 28 (386
mg, 1.0 mmol) and solid NaHCO3 (168 mg, 2.0 mmol),
and the mixture was stirred at room temperature for 2
days. The solvent was evaporated, and the residue par-
titioned between EtOAc and H2O. The organic layer
was evaporated under reduced pressure, and the residue
purified by flash chromatography on silica gel (25 g, 2 ꢃ
20 cm, 3:2 isooctane–acetone) to obtain 29 as a glass
(479 mg, 69%), which was used directly in the next step;
TLC: Rf 0.3 (3:2 isooctane–acetone).
5-Chloro-2,6-dimethyl-3,4-dihydro-4-oxoquinazoline (33).
A suspension of the oxazinone 31 (2.10 g, 0.01 mol) in
anhydrous NH3 (35 mL) in a dry ice-acetone bath was
stirred for 1.5 h and allowed to warm to room tem-
perature overnight. The residue was then refluxed with
1 N NaOH (25 mL) for 1 h, and the solution cooled in
an ice-bath and acidified with glacial AcOH. The pre-
cipitate was collected, washed thoroughly with H2O,
and freeze-dried. Recrystallization from MeOH (250
mL) gave 33 as white needles (1.18 g). Concentration of
the mother liquor afforded a second crop (0.87 g); total
yield 2.05 g (99%): mp 296–299 ꢂC. IR (KBr) nmax 3170,
3060, 3030, 2920, 2890, 2675, 1635, 1600, 1545, 1500,
Step 3. A solution of diester 29 (439 mg, 0.916 mmol)
in MeOH (20 mL) was cooled in an ice-bath and satu-
rated with gaseous NH3. The cooling bath was removed
and the solution was kept at room temperature for 20 h
in a stoppered flask. The solvent was evaporated, and
the residue dissolved in CH2Cl2 (30 mL). The solution
was cooled in an ice-bath and stirred while TFA (10
mL) was added dropwise over 5 min. The solution was
kept at 5ꢂ for 20 h, then concentrated to dryness by
rotary evaporation. The residue was taken up into a
small volume water to which enough 1 N NaOH and
10% AcOH were added to bring the solution to pH 8.
After removing a small amount of insoluble material,
the filtrate was passed through a preparative HPLC
column (7% MeCN in 0.1 M NH4OAc, pH 6.9; flow
rate 10 mL/min; 335 nm). The largest peak, whose elu-
tion time on an analytical column (flow rate 1.0 mL/min)
was 21 min, was collected and freeze-dried. The result-
ing product was taken up dilute NH4OH and passed
through a DEAE-cellulose column (1.5 ꢃ 17 cm,
HCOꢁ3 1 form), which was eluted first with H2O, then
with 0.2 M NH4HCO3, and finally with 0.4 M
NH4HCO3 adjusted to pH >10 with ammonia. Pooled
fractions containing the desired product were con-
centrated by rotary evaporation and lyophilization to
obtain 6 as a white solid (76 mg, 15%): mp >300 ꢂC. IR
(KBr) nmax 3340, 2970, 1705, 1655, 1600, 1555, 1515,
1470, 1400, 1340, 1290, 1260 cmꢁ1. 1H NMR (DMSO-d6,
200 MHz) d 1.77 (m, 2H, b-CH2), 2.16 (m, 2H, g-CH2),
4.12 (m, 1H, a-CH), 4.25 (m, bridge CH2), 5.72 (d, J=4
Hz, thiophene C4-H), 6.46 (br s, 2H, NH2), 7.04 (d, J=8
Hz, 1H, C8-H), 7.29 (d, J=4 Hz, 1H, thiophene C3-H),
7.33 (m, 1H, CONH), 7.44 (d, J=8 Hz, 1H, C7-H), 7.64
(m, 1H, CONH). UV (pH 7.4) lmax 233 nm (e 36,900),
273sh (8890), 335 (19,000); (0.1 N NaOH) 234 (37,200),
279 (11,100), 337 (20,100); (0.1N HCl) 228 (32,200), 235
(30,300), 335 (15,300); MS: m/e 478 (Mꢁ1, 35Cl), 479 (M,
35Cl), 480 (Mꢁ1, 37Cl), 481 (M, 37Cl). Anal. calcd for
1465, 1445, 1375, 1305, 1270, 1230 cmꢁ1 1H NMR
.
(DMSO-d6) d 2.30 (s, 3H, 6-Me), 2.40 (s, 3H, 2-Me), 7.42
(d, J=8 Hz, C8-H), 7.73 (d, J=8 Hz, 1H, C7-H). Anal.
calcd for C10H9ClN2O: C, 57.57; H, 4.35; N, 13.43; Cl,
16.99. Found: C, 57.58; H, 4.32; N, 13.41; Cl, 16.77.
N-[5-[N-(2-Methyl-5-chloro-3,4-dihydro–2-methyl-4-oxo-
quinazolin-6-yl)methylamino]-2-thenoyl]-L-glutamic acid
(7).
Step 1. A mixture of 33 (209 mg, 1.0 mmol), NBS (178
mg, 1.0 mmol), and 20 mg of Bz2O2 (20 mg) in CHCl3
(20 mL) that had been passed through 7 g of silica gel to
remove EtOH was refluxed for 1 h, then treated with
another 10 mg of Bz2O2, and refluxed again for 8 h. The
heavy precipitate was collected and washed with CHCl3
to obtain bromide 30 (221 mg, 77%) as a white solid
that was used in the next step without purification: mp
>300 ꢂC. IR (KBr) nmax 3170, 3070, 3020, 2970, 2950,
2920, 2880, 1675, 1630, 1595, 1545, 1495, 1435, 1425,
1
1375, 1300, 1270, 1225, 1210 cmꢁ1. H NMR (DMSO-
d6) d 2.33 (s, 3H, 2-Me), 4.87 (s, 2H, CH2Br), 7.52 (d,
J=8 Hz, 1H, C8-H), 7.97 (d, J=8H, C7-H).
Step 2. A mixture of the impure bromide 30 (144 mg,
0.5 mmol), amine 28 (192 mg, 0.5 mmol), and NaHCO3
(84 mg, 1.0 mmol) in DMF (5 mL) was stirred at room
temperature for 4 days, during which a clear solution
was formed. After removal of the solvent under reduced
pressure, the residue was partitioned between EtOAc
and H2O. TLC (silica gel, 2:3 isooctane–Me2CO)
revealed a dark spot with Rf 0.25 (turning yellow on
standing overnight), corresponding to the desired die-
ster 34, and additional spots with Rf 0.6 (turning orange
on standing overnight), 0.4, and 0.05. The EtOAc layer
was evaporated and the residue chromatographed (flash
.
C19H18ClN5O6S.0.5NH3 3.5H2O: C, 41.38; H, 3.84; N,
13.97; S, 5.81. Found: C, 41.51; H, 4.41; N, 14.23; S, 5.72.

2076
R. A. Forsch et al. / Bioorg. Med. Chem. 10 (2002) 2067–2076
grade silica gel, 11 g, 1.5 ꢃ 16 cm, 2:3 isooctane–
Me2CO) to obtain an off-white solid (165 mg, 56%). IR
(KBr) nmax 3300, 3070, 2970, 2930, 1730, 1675, 1630,
1595, 1550, 1510, 1460, 1395, 1365, 1335, 1295 1255
Harland, S. J.; Robinson, B. A.; Jackman, A. L.; Jones, T. R.;
Newell, D. R.; Siddik, Z. H.; Wilshaw, E.; McElwain, T. J.;
Smith, I. E.; Harrap, K. R. J. Clin. Oncol. 1986, 4, 1245.
4. (a) Hughes, L. R.; Jackman, A. L.; Oldfield, J.; Smith,
R. C.; Burrows, K. D.; Marsham, P. R.; Bishop, J. A. M.;
Jones, T. R.; O’Connor, B. M.; Calvert, A. H. J. Med. Chem.
1990, 33, 3060. (b) Jackman, A. L.; Newell, D. R.; Gibson,
W.; Jodrell, D. I.; Taylor, G. A.; Bishop, J. A.; Hughes, L. R.;
Calvert, A. H. Biochem. Pharmacol. 1991, 42, 1885.
5. (a) Marsham, P. R.; Jackman, A. L.; Hayter, A. J.; Daw,
M. R.; Snowden, J. L.; O’Connor, B. M.; Bishop, J. A. M.;
Calvert, A. H.; Hughes, L. R. J. Med. Chem. 1991, 34, 1594.
(b) Jackman, A. L.; Taylor, G. A.; Gibson, W.; Kimbell, R.;
Brown, M.; Calvert, A. H.; Judson, I. E.; Hughes, L. R. Can-
cer Res. 1991, 51, 5579. (c) Jackman, A. L.; Marsham, P. R.;
Moran, R. G.; Kimbell, R.; O’Connor, B. M.; Hughes, L. R.;
Calvert, A. H. Adv. Enz. Regul. 1991, 31, 13. (d) Hughes, L.
R., Stephens, T. C., Boyle, F. T., Jackman, A. L. In Antifolate
Drugs in Cancer Therapy; Jackman, A. L., Ed., Humana:
Totowa, 1999; p 147. (e) Beale, P., Clarke, S. In Antifolate
Drugs in Cancer Therapy; Jackman, A. L., Ed., Humana:
Totowa, 1999; p 167.
6. (a) Hynes, J. B.; Garrett, C. M. J. Med. Chem. 1975, 18,
632. (b) Hynes, J. B.; Eason, D. E.; Garrett, C. M.; Colvin,
P. L., Jr.; Shores, K. E.; Freisheim, J. H. J. Med. Chem. 1977,
20, 588. (c) Scanlon, K. J.; Moroson, B. A.; Bertino, J. R.;
Hynes, J. B. Mol. Pharmacol. 1979, 16, 261. (d) Hynes, J. B.;
Yang, Y. C. S.; McGill, J. E.; Harmon, S. J.; Washtien, W. L.
J. Med. Chem. 1984, 27, 232. (e) Hynes, J. B.; Kumar, A.;
Tomazic, A.; Washtien, W. L. J. Med. Chem. 1987, 30, 1515.
(f) McGuire, J. J.; Sobrero, A. F.; Hynes, J. B.; Bertino, J. R.
Cancer Res. 1987, 47, 5975. (g) Hynes, J. B.; Kumar, A.;
Tomazic, A.; Washtien, W. L. J. Med. Chem. 1987, 30, 1515.
(h) Patil, S. A.; Shane, B.; Freisheim, J. H.; Singh, S. K.;
Hynes, J. B. J. Med. Chem. 1989, 32, 1559.
7. Baker, B. R.; Joseph, J. P.; Schaub, R. E.; McEvoy, F. J.;
Williams, J. H. J. Org. Chem. 1952, 17, 157.
8. Ghisalberti, E. L.; Jefferies, P. R.; Mori, T. A.; Skelton,
B. W.; White, A. H.; Whiteside, N. J. J. Chem. Res. (Micro-
fiche Ed.) 1993, 3155.
9. Webber, S. E.; Bleckman, T. M.; Attard, J.; Deal, J. G.;
Kathardekar, V.; Welsh, K. M.; Webber, S.; Janson, C. A.;
Matthews, D. A.; Smith, W. W.; Freer, S. T.; Jordan, S. R.;
Bacquet, R. J.; Howland, E. F.; Booth, C. L.; Ward, R. W.;
Hermann, S. M.; White, J.; Morse, C. A.; Hilliard, J. A.;
Bartlett, C. A. J. Med. Chem. 1993, 36, 733.
10. Singh, S. K.; Govingan, M.; Hynes, J. B. J. Heterocycl.
Chem. 1990, 27, 2101.
11. Threadgill, M. D.; Webb, P.; O’Neill, P.; Naylor, M. A.;
Stephens, M. A.; Stratford, I. J.; Cole, S.; Adams, G. E.;
Fielden, E. M. J. Med. Chem. 1991, 34, 2112.
12. (a) Several alternative approaches to the synthesis of 28
have also been described; see, for example: Rinkes, I. J.
Researches on thiophene derivatives. Red. Trav. Chim. Pay-
Bas 1932, 51, 1134. (b) Fournari, P.; Chane, J. P. Bull. Soc.
Chim. Fr. 1963, 479. (c) Hodson, S. J.; Bigham, E. C.; Duch,
D. S.; Smith, G. K.; Ferone, R. J. Med. Chem. 1994, 37, 2112.
13. Patil, S. D.; Jones, C.; Nair, M. G.; Galivan, J.; Maley, F.;
Kisliuk, R. L.; Gaumont, Y.; Duch, D.; Ferone, R. J. Med.
Chem. 1989, 32, 1284.
1
cmꢁ1. H NMR (CDCl3) d 1.40 (s, 9H, t-Bu), 1.47 (s,
9H, t-Bu), 2.10–2.40 (m, 4H, b- and g-CH2), 2.52 (s, 3H,
2-Me), 4.57 (m, 3H, bridge CH2 and a-CH), 5.88 (d,
J=4 Hz, 1H, thiophene C3-H), 6.75 (d, J=8 Hz, 1H,
C8-H), 7.28 (d, J=4 Hz, 1H, thiophene C4-H), 7.60 (d,
1
J=8 Hz, 1H, C7-H). A high-field H NMR signal at d
0.8 showed the sample to contain occluded isooctane.
This material was used directly in the next step without
elemental analysis.
Step 3. Diester 34 (311 mg, 0.526 mmol) was dissolved
in CH2Cl2 (15 mL), cooled in an ice-bath, and treated
dropwise with TFA (5 mL) over 5 min. After being kept
at 5 ꢂC for 20 h, the solution was evaporated to dryness
under reduced pressure and the residue was taken up in
dilute NaOH. The solution adjusted to pH <9 with
10% AcOH and the product isolated by preparative
HPLC (C18 silica gel, 7% MeCN in 0.1 M NH4OAc,
pH 6.9, 10 mL/min, 335 nm). The largest peak, whose
elution time on an analytical column (flow rate 1.0 mL/
min) was 33 min, was collected and concentrated to
dryness by rotary evaporation followed by freeze-dry-
ing. Further purification by ion-exchange chromato-
graphy, as described above for compound 6, afforded 7
(116 mg, 39%) as a pale-yellow solid; mp >300 ꢂC. IR
(KBr) nmax 3200, 3070, 2930, 1675, 1625, 1595, 1550,
1510, 1460, 1395, 1335, 1205 cmꢁ1. 1H NMR (200 MHz,
DMSO-d6) d 1.77 (m, 2H, b-CH2), 2.17 (m, 2H, g-CH2),
2.24 (s, 3H, 2-Me), 4.16 (m, 1H, a-CH), 4.36 (m, 2H, bridge
CH2), 5.75 (d, J=4 Hz, 1H, thiophene C4-H), 7.31 (d, J=4
Hz, 1H, thiophene C3-H), 7.43 (d, J=8 Hz, 1H, C8-H),
7.50 (m, 1H, CONH), 7.66 (d, J=8 Hz, 1H, C7-H). UV:
lmax (pH 7.4) 232 nm (e 32,800), 238sh (28,800), 270 (9200),
277 (9100), 331 (22,100); (0.1 N NaOH) 232 (28,200), 283sh
(10,600), 291 (12,100), 334 (23,000); (0.1 N HCl) 240
(22,200), 280 (6200), 336 (15,700); MS: m/e 477 (Mꢁ1,
35Cl), 478 (M, 35Cl), 479 (Mꢁ1, 37Cl), 480 (M, 37Cl). Anal.
.
.
calcd for C20H19ClN4O6S NH3 3.5H2O: C, 42.97; H, 5.23;
N, 12.53; S, 5.74. Found: C, 42.87; H, 4.74; N, 12.43; S, 5.45.
Acknowledgements
This work was supported in part by Grants RO1-
CA63064 and RO1-CA70349 from the National Cancer
Institute (NIH, DHHS). Technical assistance in cell
culture was provided by Dr. Ying-Nan Chen.
References and Notes
14. (a) Jones, T. R.; Smithers, M. J.; Taylor, M. A.; Jackman,
A. L.; Calvert, A. H.; Harland, S. J.; Harrap, K. R. J. Med.
Chem. 1986, 29, 468. (b) Jones, T. R.; Smithers, M. J.; Better-
idge, R. F.; Taylor, M. A.; Jackman, A. L.; Calvert, A. H.;
Davies, L. C.; Harrap, K. R. J. Med. Chem. 1986, 29, 1114.
15. Coll, R. J.; Cesar, D.; Hynes, J. B.; Shane, B. Biochem.
Pharmacol. 1991, 42, 833.
1. Davoll, J.; Johnson, A. M. J. Chem. Soc. (C) 1970, 997.
2. (a) Jones, T. R. Int. J. Cancer 1980, 16, 707. (b) Calvert,
J. H.; Jones, T. R.; Dady, P. J.; Grzelakowska-Sztabert, B.;
Paine, R. M.; Taylor, G. A.; Harrap, K. R. Int. J. Cancer
1980, 16, 713.
3. (a) Jones, T. R.; Calvert, A. H.; Jackman, A. L.; Brown,
S. J.; Jones, M.; Harrap, K. R. Int J. Cancer 1981, 17, 11. (b)
Jackson, R. C.; Jackman, A. L.; Calvert, A. H. Biochem.
Pharmacol. 1983, 32, 3783. (c) Calvert, A. H.; Alison, D. L.;
16. McGuire, J. J.; Heitzman, K. J.; Haile, W. H.; Russell,
C. A.; McCloskey, D. E.; Piper, J. R. Leukemia 1993, 7, 1996.