Synthesis of new 2,4-Diaminopyrido[2,3-d]pyrimidine and 2,4-Diaminopyrrolo[2,3-d]pyrimidine inhibitors of Pneumocystis carinii, Toxoplasma gondii, and Mycobacterium avium dihydrofolate reductase

Article abstract of DOI:10.1016/S0968-0896(02)00325-5

A concise new route allowing easy access to five previously unreported 2,4-diamino-6-(substituted benzyl)pyrido[2,3-d]pyrimidines (2a-e) was developed, involving condensation of 2,4-dipivaloylamino-5-bromopyrido[2,3-d]pyrimidine (6) with an organozinc halide in the presence of a catalytic amount of [1,1′-bis(diphenylphosphino)ferrocene]dichloropalladium(II)·CH ₂Cl₂, followed by removal of the pivaloyl groups with base. Also prepared via a scheme based on the Taylor ring expansion/ring annulation synthesis were three heretofore undescribed 2,4-diamino-5-(substituted benzyl)-7H-pyrrolo[2,3-d]pyrimidines (3b-c). Standard spectrophotometric assays were used to compare the ability of 2a-e and 3b-c to inhibit dihydrofolate reductase (DHFR) from Pneumocystis carinii, Toxoplasma gondii, and Mycobacterium avium, three examples of opportunistic pathogens to which AIDS patients are highly vulnerable because of their immunocompromised state. For comparison, 13 previously untested 2,4-diamino-6-(substituted benzyl)quinazolines (17a-m) were also evaluated as inhibitors of these enzymes, as well as the enzyme from rat liver. None of the quinazolines or pyridopyrimidines tested was more potent against the P. carinii enzyme than the structurally related reference compound piritrexim (1), and none showed selectivity for the P. carinii enzyme over the rat enzyme. One of the pyridopyrimidines (2c) showed 10-fold selectivity for T. gondii versus rat DHFR, and two of them (2b, 2c) showed selectivity for the M. avium enzyme. However, this gain in species selectivity was achieved at the cost of decreased in potency, as has been noted with many other lipophilic DHFR inhibitors.

Full text of DOI:10.1016/S0968-0896(02)00325-5

Bioorganic & Medicinal Chemistry 11 (2003) 59–67
Synthesis of New 2,4-Diaminopyrido[2,3-d]pyrimidine and
2,4-Diaminopyrrolo[2,3-d]pyrimidine Inhibitors of
Pneumocystis carinii, Toxoplasma gondii, and
Mycobacterium avium Dihydrofolate Reductase
Andre Rosowsky,a,* Han Chen,a Hongning Fua and Sherry F. Queenerb
aDana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology, Harvard Medical School,
Boston, MA02115, USA
bDepartment of Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, IN 46202, USA
Received 18 April 2002; accepted 16 July 2002
Abstract—A concise new route allowing easy access to five previously unreported 2,4-diamino-6-(substituted benzyl)pyrido[2,3-
d]pyrimidines (2a–e) was developed, involving condensation of 2,4-dipivaloylamino-5-bromopyrido[2,3-d]pyrimidine (6) with an
0
.
organozinc halide in the presence of a catalytic amount of [1,1 -bis(diphenylphosphino)ferrocene]dichloropalladium(II) CH2Cl2,
followed by removal of the pivaloyl groups with base. Also prepared via a scheme based on the Taylor ring expansion/ring annu-
lation synthesis were three heretofore undescribed 2,4-diamino-5-(substituted benzyl)-7H-pyrrolo[2,3-d]pyrimidines (3b–c). Stan-
dard spectrophotometric assays were used to compare the ability of 2a–e and 3b–c to inhibit dihydrofolate reductase (DHFR) from
Pneumocystis carinii, Toxoplasma gondii, and Mycobacterium avium, three examples of opportunistic pathogens to which AIDS
patients are highly vulnerable because of their immunocompromised state. For comparison, 13 previously untested 2,4-diamino-6-
(substituted benzyl)quinazolines (17a–m) were also evaluated as inhibitors of these enzymes, as well as the enzyme from rat liver.
None of the quinazolines or pyridopyrimidines tested was more potent against the P. carinii enzyme than the structurally related
reference compound piritrexim (1), and none showed selectivity for the P. carinii enzyme over the rat enzyme. One of the pyr-
idopyrimidines (2c) showed 10-fold selectivity for T. gondii versus rat DHFR, and two of them (2b, 2c) showed selectivity for the M.
avium enzyme. However, this gain in species selectivity was achieved at the cost of decreased in potency, as has been noted with
many other lipophilic DHFR inhibitors.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
rapidly through the immune system. The only 2,4-dia-
minopyrido[2,3-d]pyrimidine tested clinically against P.
carinii in AIDS patients to date has been piritrexim (1).2
Because this is an extremely potent inhibitor of dihy-
drofolate reductase (DHFR), and its affinity is higher
for human dihydrofolate reductase (DHFR) than it is
for the P. carinii or T. gondii enzyme, 1 is highly toxic to
the hematopoietic system and has to be used in combi-
nation with leucovorin (LV) to selectively protect the
bone marrow and other sensitive host tissues. LV is
readily taken up by mammalian cells, whereupon it
reconstitutes the tetrahydrofolate pool and thus acts as
an antidote to the DHFR inhibitor. The selectivity of
LV in this context is considered to reflect the fact that
its uptake is mediated by a reduced folate carrier (RFC)
protein that is present in the plasma membrane of
mammalian cells but not parasitic eukaroytic species
such as P. carinii.3 Lipophilic antifolates like 1 do not
2,4-Diaminopyrido[2,3-d]pyrimidines containing a sub-
stituted phenyl or other aromatic ring attached to C6
via by a short bridge comprise a family of lipophilic
small-molecule antifolates with potential applications in
the treatment and prophylaxis of AIDS-associated
opportunistic microbial infections. Among the often
fatal opportunistic pathogens frequently identified in
immunocompromised AIDS patients are Pneumocystis
carinii, Toxoplasma gondii, and the Mycobacterium
avium complex (MAC).1 Indeed, identification of one or
more of these organisms is often the earliest sign that
human immunodeficiency virus (HIV) is spreading
*Corresponding author. Tel.: +1-617-632-3117; fax: +1-617-632-
2410; e-mail: andre_rosowsky@dfci.harvard.edu
0968-0896/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00325-5

60
A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
depend on the RFC for uptake, and thus would be
selectively toxic to P. carinii in the absence of LV only if
their affinity for the DHFR of P. carinii were higher
than their affinity for human enzyme. This would of
course be true, as well, for other microbial organisms
that lack the RFC system of mammalian cells.
substituted phenyl group at C5 were subsequently
reported by our own group.11 In the present paper, in
addition to 2a–e, we report the synthesis and antifolate
activity of the previously unknown 5-(substituted ben-
zyl) derivatives 3a–c. The structures of 1 (piritrexim),
the pyrido[2,3-d]pyrimidines 2a–e, and the pyrrolo[2,3-
d]pyrimidines 3a–c are shown in Figure 1.
A large variety 2,4-diaminopyrido[2,3-d]pyrimidines
with a substituted aromatic ring joined to C6 by a short
bridge have been studied in vitro as part of a broad
search for compounds with a preferential binding affi-
nity for P. carinii and T. gondii DHFR. These have
included variously substituted 2,4-diamino-6-(N-aryl-
amino)methyl,4,5 2,4-diamino-6-(2-arylethyl),5,6 2,4-di-
amino-6-(N-aralkyl)amino,7 and 2,4-diamino-6-arylthio7
derivatives. The 6-(N-aralkyl)-amino derivatives have
also been tested against DHFR from M. avium.8 How-
ever, to our knowledge, inhibition of P. carinii, T. gon-
dii, and M. avium DHFR has not been studied with any
2,4-diamino-6-(substituted benzyl)pyrido[2,3-d]pyrimi-
dines other than 1 itself, in which a 5-Me group is also
present. In the present paper, we describe a synthesis of
the previously unknown 2,4-diamino-6-(substituted
benzyl)pyrido[2,3-d]pyrimidines 2a–e via a convenient
Pd(0) mediated carbon–carbon bond-forming reaction
developed recently in our laboratory as a route to
2,4-diamino-6-(substituted benzyl)quinazolines from
2,4-diamino-6-iodoquinazoline.9
Chemistry
As shown in Scheme 1, condensation of 2,4,6-triamino-
pyrimidine (4) with freshly prepared bromomalonalde-
hyde according to Gangjee and coworkers4c afforded
2,4-diamino-6-bromopyrido[2,3-d]pyrimidine (5). Upon
further reaction with pivalic anhydride in refluxing pyr-
1
idine, 5 was converted to a product whose H NMR
spectrum showed it to consist mostly of the desired
N2,N4-dipivaloyl derivative 6, along with a smaller
amount of material that appeared to be more exten-
sively pivaloylated. Silica gel chromatography followed
by recrystallization of pooled TLC-homogeneous
fractions from acetone afforded pure 6 in 29% overall
yield (two steps). In a typical coupling reaction, a
catalytic amount (2.5 mol.%) of 1,10-bis(diphenylphos-
phino)ferrocene]dichloropalladium(II).CH2Cl2
was
added to a commercially supplied solution of the sub-
stituted benzylzinc chloride in dry THF under an
atmosphere of dry N2. The solution was stirred at room
temperature for 5 min to allow the black palladium–zinc
complex to form, and a solution of 6 in dry THF was
added via a cannula while maintaining the system con-
tinuously under N2. When addition was complete, the
mixture was refluxed for ca. 3 h, during which the pro-
gress of the reaction was monitored by TLC. After
routine workup as described in Experimental, the piva-
loylated coupling products (7a–e) were passed through a
silica gel column and converted directly to 2a–e by
removal of the pivaloyl groups with NaOH in aqueous
MeOH, followed by recrystallization from mixtures of
aqueous DMF, mixtures of MeOH, EtOH, and H2O, or
mixtures of DMF, EtOH, and H2O. Although the non-
optimized combined yields for the coupling and depro-
tection reactions were <40%, the obvious merits of the
synthesis are that it requires only four steps from 4, and
that it has the potential to easily generate a library of 2,4-
More distantly related to 1, but nonetheless of potential
interest in the context of the search for newer lipophilic
antifolates for use in the treatment of AIDS-related
opportunistic infections, are 2,4-diamino-5-(anilino-
methyl) and 5-(arylthiomethyl-7H-pyrrolo[2,3-d]pyr-
imidines, of which several examples have been tested.10
Apart from the presence of an electron-donating instead
of electron-withdrawing nitrogen next to the pyrimidine
ring, a notable feature of these five-membered ring ana-
logues was that the side chain had a different spatial
orientation with respect to the heterocyclic moiety than
the side chain of pyrido[2,3-d]pyrimidine derivatives.
Several analogous pyrrolo[2,3-d]-pyrimidines with a
Scheme 1. Reagents: (a) bromomalonaldehyde/HCl/EtOH; (b) Piv2O/
.
C5H5N; (c) arymethylzinc chloride/(DPPF)2PdCl2 CH2Cl2/THF; (d)
NaOH/MeOH.
Figure 1. Structures of piritrexim (1), 2,4-diaminopyrido[2,3-d]pyr-
imidines 2a–e, and 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines 3a–c.

A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
61
diamino - 6 - (arylmethyl)pyrido[2,3 - d]pyrimidines from
commercially available arylmethylzinc halide reagents
by parallel synthesis.
pheny)propanal (14) with CrO3/pyridine, and the Tay-
lor approach was followed to convert 14 to 3c via
the a-hydroxyketone 15 and the furan aminonitrile 16.
As in the synthesis of 3a and 3c, the intermediates 13–16
were all oils or semi-solids whose identity and purity
after chromatography on silica gel were established
For the synthesis of 3a and 3b, the method we pre-
viously used for the synthesis of 2,4-diamino-5-benzyl-
7H-pyrrolo[2,3-d]pyrimidine itself11 was followed with
minor changes. As shown in Scheme 2, the 2-amino-4-
(substituted aralkyl)furan-3-carbonitriles 8a–c were
synthesized from 3,4,5-trimethoxybenzaldehyde, 3,4-
dichlorobenzaldehyde, and 3,4,5-trimethoxyphenyl-
acetaldehyde, respectively, via the ring transformation/
ring annulation method of Taylor and coworkers.12
Although the oily or semi-solid intermediates 9a,b (enol
ethers), 10a,b (phenylacetaldehdyes), and 11a,b (a-
hydroketones) were not analyzed, they were carefully
purified by column chromatography, and 1H NMR
spectra were obtained on the chromatographed pro-
ducts in order to confirm both their structure and pur-
ity. The enol ethers 9a,b were each used as mixtures of E
and Z isomers whose vinylic doublets could be readily
assigned from their coupling constants of 7.0 and 13.2
Hz, respectively, in the case of 9a and 7.2and 12.8 Hz,
respectively, in the case of 9b. The benzylic and alde-
hyde protons in 10b produced singlets at d 3.87 and d
9.66, whereas the corresponding singlets in 10a were
shifted downfield, as expected, to d 5.05 and d 9.97. In
contrast, upfield rather than downfield shifts of the
benzylic protons were observed in 11b (d 3.31) and 8b (d
3.55) relative to 11a (d 3.93) and 8a (d 3.65). As antici-
pated, the 2-furanyl proton in 8a and 8b (both at ca. d
6.8) was unaffected by the electron-donating or elec-
tron-withdrawing character of the phenyl substituents.
Critically important confirmation that the ring transfor-
mation/ring annulation reaction had proceeded in the
desired manner to form the pyrrolopyrimidines 3a and
3b, and not the corresponding furopyrimidines, was
1
from their H NMR spectra, which were essentially the
same as those of the corresponding 3,4,5-trimethoxy-
benzyl intermediates 8b–11b with the exception of an
extra peak in the CH2 region. Furthermore, as in the
1
case of 3b, the H NMR spectrum of 3c showed the
peaks at d 6.41 (pyrrole ¼CH) and d 10.35 (pyrrole NH)
expected from a pyrrolo[2,3-d]pyrimidine structure.
Bioassay
The ability of compounds 2a–e and 3a–c to inhibit
reduction of dihydrofolate by P. carinii, T. gondii, M.
avium, and rat DHFR in the presence of NADPH was
assayed spectrophotometrically at 340 nm as previously
described.13 Also tested for the purpose of comparison
were the 2,4-diamino-6-(substituted benzyl)quinazolines
(17a–m, Table 1) whose chemistry we published recently
without DHFR binding data.9 The results for both
groups of inhibitors are presented in Table 1. The IC50
values of 2a–e against P. carinii DHFR were only in the
0.2–0.5 mM range, and thus were approximately one
order of magnitude higher than that of 1 (IC50 0.03
mM). Compound 2a, with the identical 2,5-dimethoxy
substitution as 1, was a 10-fold poorer inhibitor, point-
ing to an important role for the 5-methyl group in
binding to the enzyme. The IC50 values of 17a–m were
in the 0.05–0.3 mM range, and in four of the five exam-
ples for which a side-by-side comparison could be made
the quinazolines tended to be slightly more potent than
the pyrido[2,3-d]pyrimidines. With regard to selectivity,
it can be seen that both types of diaminoheterocycles
were more potent against the rat enzyme than against
the P. carinii enzyme, resulting in SI values of <1.
1
provided by the H NMR spectrum, which revealed the
presence of the C2and N7 protons as singlets at d 6.40
and d 10.42in the case of 3a and d 6.4 and d10.54 in the
case of 3b.
The IC50 values of 2a–e against T. gondii DHFR were in
the 0.01–0.07 mM range, and thus were consistently
lower than their IC50 values against the P. carinii
enzyme. However, the potency of these compounds
against T. gondii DHFR was lower than that of 1. On
the other hand, the 2-chlorobenzyl analogue 2d dis-
For the synthesis of the analogue 3c with an extra CH2
group in the bridge (Scheme 3), the commercially
available starting material 3,4,5-trimethoxycinne-
maldehyde (12) was reduced with LiAlH4, the resulting
alcohol (13) was re-oxidized to 3-(3,4,5-trimethoxy-
Scheme 2. Reagents: (a) MeOCH¼PPh3; (b) aq HClO4/Et2O; (c) (CH2O)n/N-benzothiazolium bromide/Et3N; (d) CH2(CN)2/Et3N; (e)
H2NC(=NH)NH2.

62
A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
played a modest SI of 4.3 in comparison with the highly
unfavorable value of 0.077 observed with 1. An even
better SI of 10.3 was found with the 3,5-dimeth-
oxybenzyl analogue 2c, although this was unfortu-
nately accompanied by a small decrease in potency. As
in the case of the P. carinii enzyme, the quinazoline
inhibitors tended to be somewhat more potent than the
corresponding pyrido[2,3-d]pyrimidines. However the
increased potency of the quinazolines was, once
again, accompanied by decrease in selectivity; cf. for
example the SI values of 17f versus 2c and 17h versus
2d. Interestingly, the compound with the best combi-
nation of potency and selectivity against T. gondii,
relative to 1, was the 3,5-dimethoxy analogue 2c rather
than the 2,5-dimethoxy analogue 2a. This difference
relative to 1 may be due to the absence of a methyl
group at C5.
In general, the IC50 values of 2a–e and 17a–m against
M. avium DHFR were similar to those against T. gondii
Scheme 3. Reagents: (a) LiAlH4/THF; (b) CrO3/pyridine; (c) (CH2O)n/N-ethylbenzothiazolium bromide/Et3N; (d) CH2(CN)2/Et3N; (e)
H2NC(¼NH)NH2.
Table 1. Inhibition of P. carinii, T. gondii, M. avium, and rat liver dihydrofolate reductase by 2,4-diamino-6-(substituted benzyl)-pyrido[2,3-
d]pyrimidines, 2,4-diamino-5-(substituted benzyl)-7H-pyrido[2,3-d]pyrimidines, and 2,4-diamino-6-(substituted benzyl)quinazolines
Compd
R
IC50 (mM)a
Selectivity index (SI)b
P. carinii
T. gondii
M. avium
Rat liver
P. carinii
T. gondii
M. avium
1c
2a
2b
2c
2d
2e
3a
3b
0.013
0.28
0.24
0.21
0.51
0.52
29
7214
0.77
0.31
0.20
0.23
0.11
0.092
0.11
0.15
0.11
0.17
0.15
0.050
0.67
0.087
0.0043
0.034
0.064
0.036
0.017
0.040
3.3
0.0006
0.059
0.028
0.041
0.039
0.057
14
0.00033
0.054
0.21
0.026
0.19
0.88
1.8
0.14
0.25
0.30
0.077
1.6
3.3
10.3
4.3
0.55
0.92
7.5
9.0
1.9
2,5-(OMe)2
3,4-(OMe)2
3,5-(OMe)2
2-Cl
0.37
0.073
0.13
9.6
4-F
3.3
2.9
2.3
0.69
3,4-Cl2, n=1
3,4,5-(OMe)3, n=1
3,4,5-(OMe)3, n=2
2-OMe
67
52 0.723.7
0.78
3c
0.037
0.011
0.010
0.018
0.010
0.013
0.0099
0.017
0.0064
0.016
0.022
0.056
0.026
0.022
0.077
0.016
0.021
0.013
0.011
0.009
0.017
0.007
0.009
0.021
0.025
0.021
0.030
0.020
0.20
0.26
0.17
0.19
0.43
0.085
0.30
0.27
0.28
0.10
0.24
0.27
0.56
0.10
0.17
5.4
4.9
3.8
5.6
0.94
2.2
3.0
2.5
1.2
2.5
1.8
0.50
2.6
0.68
2.6
3.4
1.8
7.7
0.85
3.1
1.8
6.0
0.52
1.9
1.6
1.3
2.3
0.75
17a
17b
17c
17d
17e
17f
17g
17h
17i
17j
17k
17l
17m
0.054
0.038
0.10
3-OMe
4-OMe
2,5-(OMe)2
3,4-(OMe)2
3,5-(OMe)2
3,4,5-(OMe)3
2-Cl
0.0094
0.028
0.030
0.042
0.011
0.040
0.040
0.028
0.068
0.015
3-Cl
4-Cl
3,4-Cl2
4-F
d
aAssays of DHFR inhibition were performed and described in refs 13a and b (P. carinii, T. gondii, rat liver) and 13c (M. avium). The standard
deviation of these values, based on recent data obtained with 1 against rat liver DHFR is estimated to be ꢀ10%.
bSI=IC50 (rat liver)/IC50 (P. avium, T. gondii, or M. avium).
cIC50 values of 1 against P. carinii, T. gondii, and rat liver DHFR differ somewhat from those reported recently in ref 15, which had been obtained
with an older reference sample acquired from the NCI drug repository through the courtesy of Dr. Mohamed Nasr. The data reported here were
obtained on a different sample kindly provided by Dr. Gary Smith, Glaxo-Wellcome, Research Triangle Park, NC, USA.
d2-Naphthylmethyl.

A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
63
DHFR, but lower than those against P. carinii DHFR.
Moreover the quinazolines tended once again to be
more potent than the corresponding pyrido[2,3-d]pyr-
imidines; cf. for example, 17d–f versus 2a–c. The most
potent member of the series in the case of the M. avium
enzyme was 17g, with an IC50 of 0.007 mM. Unlike 1,
17g displayed a modest level of selectivity (SI=6.0).
Although there have been almost no systematic studies
focusing on the ability of compounds such as those
described here to inhibit the growth of human tumor
cells in culture, the results in Table 2generally met our
expectation that these compounds would probably have
IC50 values in the 0.1–1 mM range. As a point of refer-
ence, the IC50 of 1 against CCRF-CEM cells after 48 h
of drug exposure can be estimated from published
growth curves to be roughly in the 0.01–0.1 mM range.14
Thus, our somewhat higher IC50 for 2a as compared to
1 was consistent with a view of the 5-Me substituent as
being a positive contributor to the biological activity of
1 by virtue of its ability to either promote cellular
uptake or interact with nonpolar residues in the active
site of DHFR. Conversely, the finding that 2a and sev-
eral of its congeners were selective for the T. gondii and
M. avium DHFR, whereas 1 was not, lends support to
the view that the smaller active site of the DHFR from
these species may not be able to accommodate the 5-Me
group as readily as the more ‘open’ mammalian enzyme.
As a corollary, it is conceivable that 5-unsusbtituted
analogues would be less likely than 1 to require co-
administration of leucovorin when given to AIDS
patients with P. carinii pneumonia or other opportunis-
tic infections.
The pyrrolopyrimidine analogues 3a and 3b displayed a
dramatic decrease in potency relative to the corre-
sponding quinazolines (3a vs 17k, 3b vs 17g), indicating
that a bulky benzyl group at the 5-position is very
poorly tolerated when the six-membered aromatic B-
ring is replaced by a pyrrole. However, potency was
substantially restored when an extra CH2 group was
added to the bridge (3c vs 3b). This suggested that a
benzyl group at the 5-position, as in 3a or 3b, can
interfere with the proper positioning of the 4-amino
group in the active site, and that this steric constraint is
relieved in 3c because the phenethyl group is able to
rotate out of the way.
Although the selectivity of 2a–e and 17a–m against the
three species of microbial DHFR was less than had been
hoped, we felt that, in light of the well-known ability of
lipophilic DHFR inhibitors to overcome transport-
based resistance to classical antifolates like methotrex-
ate, it might be of interest to test these compounds as
inhibitors of the growth of tumor cells in culture.
Accordingly, a standard 72-h growth inhibition assay
was performed with these compounds against CCRF-
CEM human leukemic lymphoblasts. As shown in
Table 2, the IC50 values of pyridopyrimidines 2a–e ran-
ged from 0.20 mM (2b) to 1.9 mM (2c), and those of
quinazolines 17a–m ranged from 0.089 mM (17h) to 0.38
mM (17g). Where pairs of structures with identical aryl
substituents could be compared, the quinazolines were
more potent than the pyridopyrimidines in nearly every
case (e.g., 17d>2a, 17h>2d, and 17l>2e), in agreement
with their greater ability to bind to DHFR. However, in
one instance (17e vs 2b) no difference in activity between
the two heterocyclic systems was observed.
Experimental
IR spectra were obtained on a Perkin–Elmer model 781
double-beam spectrophotometer. Only peaks with wave-
1
numbers greater than 1400 cmꢁ1 are reported. H NMR
spectra were recorded at 200 MHz on a Varian VX200 or,
where specified, on a Varian EM360 instrument. Mass
spectra in the fast atom bombardment (FAB) mode
were obtained by staff of the Dana-Farber Cancer
Institute Molecular Biology Core Facility. TLC ana-
lyses were performed on Whatman MK6F silica gel
plates with UV illumination at 254 nm. Column
chromatography was performed on Baker 7025 flash
silica gel (40 mm particle size). Melting points were
measured in Pyrex capillary tubes in a Mel-Temp
‘Electrothermal’ apparatus (Fisher Scientific), and are
not corrected. [1,10-Bis(diphenylphosphino)ferrocene]
dichloropalladium(III) CH2Cl2 [(DPPF)2PdCl,2CH2Cl2]
.
was purchased from Aldrich, Milwaukee, WI, USA.
Other chemicals, including the Rieke organozinc
reagents as 0.5 M solutions in THF in ‘Sure-Seal’ bot-
tles, were also from Aldrich. The organozinc chloride
solutions were transferred to the reaction flask with the
aid of a cannula and dry N2, and glassware was thor-
oughly dried and purged with dry N2 before use. The
THF in the Pd(0) catalyzed cross coupling reactions was
freshly distilled from Na benzophenone ketyl under a
dry N2, atmosphere. Elemental analyses were performed
by Robertson Laboratories, Madison, NJ, USA, and
were within ꢀ0.4% of theoretical values unless other-
wise specified. The 2,4-diamino-6-benzylquinazolines
17a–m were prepared as previously reported.9 2,4-Di-
amino-6-bromopyrido[2,3-d]pyrimidine (5) and 2,4-di-
Table 2. Growth inhibition of CCRF-CEM human leukemic cells by
2,4-diamino-6-(substituted benzyl)pyrido[2,3-d][pyrimidines and 2,4-
diamino-6-(substituted benzyl)quinazolines
Compd IC50 (mM) Compd IC50 (mM) Compd IC50 (mM)
2a
2b
2c
2d
2e
0.31
0.20
1.89
0.45
1.4
17a
17b
17c
17d
17e
17g
0.21
0.30
0.33
0.13
0.20
0.38
17h
17i
17j
17k
17l
17m
0.089
0.32
0.28
0.32
0.21
0.27
Cells were grown in 96-well plates for 72h under a 5% CO humidi-
2
fied atmosphere in RPMI-1640 medium containing 10% fetal bovine
serum, 2mM l-glutamine, and antibiotics. IC50 values for each com-
pound were determined by generating a composite growth curve using
combined numbers from either two or three separate experiments on
different days. IC50 values determined from individual experiments
with a given compound did not differ from the composite value for
that compound by more than ꢀ20%.
pivaloyl-6-bromopyrido[2,3-d]pyrimidine
(6)
were
synthesized by a slight modification of the method of
Gangjee and coworkers4c as described below.

64
A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
2,4-Diamino-6-(20,50-dimethoxybenzyl)pyrido[2,3-d]pyri-
.
midine (2a). Step 1. (DPPF)2PdCl2 CH2Cl2 (42mg, 0.05
2840, 1650, 1620, 1580, 1555, 1500, 1460 cmꢁ1. Anal.
calcd for C16H17N5O2 0.3H2O 0.15 MeOH: C, 60.33; H,
5.70; N, 21.78. Found: C, 60.56; H, 5.56; N, 21.43.
.
.
mmol) was added to a solution of 2,5-dimethoxy-
benzylzinc chloride in THF (0.5 M, 4 mL, calculated to
contain 2mmol), and the mixture stirred at room tem-
perature for 5 min under dry N2. A solution containing
6 (203 mg, 0.5 mmol) in dry THF (2 mL) was added via
a cannula. The reaction mixture was stirred under reflux
and monitored periodically by TLC. After 3 h, the
reaction was quenched with saturated aqueous NH4Cl
(10 mL) followed by saturated aqueous Na2EDTA (10
mL). After being stirred at room temperature for 30
min, the brown mixture was extracted with CH2Cl2, and
the extracts were washed with brine, dried (Na2SO4)
and concentrated under reduced pressure. Chromato-
graphy of the residue on silica gel (3% MeOH in
CH2Cl2) yielded 7a as a pale-yellow solid that was used
directly in the next step (100 mg, 48%); 1H NMR
(DMSO-d6) d 1.21 (s, 9H, Me3C), 1.23 (s, 9H, Me3C),
3.66 (s, 3H, 20- or 50-OMe). 3.73 (s, 3H, 20- or 50-OMe),
4.00 (s, 2H, CH2), 6.73–6.90 (m, 3H, 30-, 40-, and 60-H),
8.51 (d, J=2.4 Hz, 1H, H-5), 8.82 (d, J=2.4 Hz, 1H, H-
7), 11.31 (br, s, CONH).
2,4-Diamino-6-(30,50 -dimethoxybenzyl)pyrido[2,3-d]pyri-
midine (2c). Step 1. Addition of 6 (203 mg, 0.5 mmol)
in dry THF (2mL) to 3,5-dimethoxybenzylzinc chloride
in THF (0.5 M, 4 mL, calculated to contain 2mmol),
followed by a workup similar to the one for 7a, yielded
7c as a pale yellow solid (117 mg, 56%); 1H NMR
(DMSO-d6) d 1.23 (s, 9H, Me3C), 1.25 (s, 9H, Me3C),
3.70 (s, 6H, 30- and 50-OMe), 4.04 (s, 2H, CH2), 6.35 (t,
J=2.4 Hz, 1H, 40-H), 6.48 (d, J=2.4 Hz, 2H, 20-H and
60-H), 8.54 (d, J=2.4 Hz, 1H, H-5), 8.85 (d, J=2.4
Hz,1H, H-7), 11.31 (br, s, CONH).
Step 2. Starting from 7c (103 mg, 0.245 mmol), 2c was
prepared by the same method as 2a except that it was
recrystallized from MeOH/H2O: white solid (29 mg,
38%); mp >300 ꢂC; MS m/z found 312, calcd 312
(M+1)+; H NMR (DMSO-d6) d 3.69 (s, 6H, 30- and
1
50-OMe), 3.85 (s, 2H, CH2), 6.30 (br s, 2H, NH2), 6.32
(t, J=2.2 Hz, 1H, H-60), 6.40 (d, J=2.2 Hz, 2H, 20-H
and 60-H), 7.50 (br s, 2H, NH2), 8.25 (d, J=2.4 Hz,
1H, H-5), 8.55 (d, J=2.4 Hz, 1H, H-7), IR (KBr) n
3410, 3340, 3130, 2960, 2840, 1610, 1585, 1555, 1500,
Step 2. A solution of 7a (90 mg, 0.214 mmol) in
MeOH (10 mL) was treated with 1 N aqueous NaOH
(10 mL) and stirred at room temperature under N2 for 5
days. The solid precipitate was filtered, washed to neu-
trality with distilled H2O, and recrystallized from DMF/
H2O to obtain 2a as a white solid (21 mg, 31%); mp
1460 cmꢁ1. Anal. calcd for C16H17N5O2 1.1H2O: C,
.
58.03; H, 5.84; N, 21.15. Found: C, 58.00; H, 5.81; N,
20.87.
>300 ꢂC; MS m/z: found 312, calcd 312 (M+1)+; H
2,4-Diamino-6-(20 -chlorobenzyl)pyrido[2,3-d]pyrimidine
(2d). Step 1. Addition of 6 (203 mg, 0.5 mmol) in dry
THF (2mL) to -2chlorobenzylzinc chloride in THF
(0.5 M, 4 mL, calculated to contain 2mmol), followed
by a workup similar to the one for 7a, yielded 7d as a
1
NMR (DMSO-d6) d 3.67 (s, 3H, 20- or 50-OMe), 3.73 (s,
3H, 20- or 50-OMe), 3.85 (s, 2H, CH2), 6.21 (br s, 2H,
NH2), 6.73 (d, J=2.8 Hz, 1H, 60-H), 6.75 (dd, J=9.6
Hz, 2.8 Hz, 1H, 40-H), 6.88 (d, J=9.6 Hz, 1H, 30-H),
7.41 (br s, 2H, NH2), 8.20 (d, J=2.4 Hz, 1H, H-5), 8.49
(d, J=2.4 Hz, 1H, H-7). IR (KBr) n 3410, 3340, 3180,
2940, 2840, 1640, 1620, 1580, 1555, 1500 cmꢁ1. Anal.
1
pale-yellow solid (107 mg, 54%); H NMR (DMSO-d6)
d 1.20 (s, 9H, Me3C), 1.24 (s, 9H, Me3C), 4.24 (s, 2H,
CH2), 7.28–7.47 (m, 4H, 30-, 40-, 50-, and 60-H), 8.50
(s,1H, H-5), 8.87 (s, 1H, H-7), 11.31 (br s, CONH).
.
calcd for C16H17N5O2 0.35H2O: C, 60.50; H, 5.62; N,
22.05. Found: C, 60.56; H, 5.46; N, 21.74.
Step 2. Starting from 7d (95 mg, 0.24 mmol), 2d was
prepared by the same method as 2a except that recrys-
tallization was from DMF/EtOH/H2O, white solid (24
mg, 34%); mp >300 ꢂC; MS m/z found 286, calcd 286
2,4-Diamino-6-(30,40 -dimethoxybenzyl)pyrido[2,3-d]pyri-
midine (2b). Step 1. Addition of 6 (203 mg, 0.5 mmol)
in dry THF (2mL) to 3,4-dimethoxybenzylzinc chloride
in dry THF (0.5 M, 4 mL, calculated to contain 2.0
mmol), followed by a workup similar to the one for 7a,
yielded 7b as a pale-yellow solid (121 mg, 58%); 1H
NMR (DMSO-d6) d 1.21 (s, 9H, CMe3), 1.25 (s, 9H,
CMe3), 3.64 (s, 3H, 30- or 40-OMe), 3.68 (s, 3H, 30- or 40-
OMe), 4.02(s, 2H, CH 2), 6.78 (dd, J=8.6 Hz, 3.0 Hz,
1H, 60-H), 6.91 (d, J=3.0 Hz, 1H, 20-H), 6.92(d, J=8.6
Hz, 50-H), 8.51 (d, J=2.4 Hz, 1H, H-5), 8.88 (d, J=2.4
Hz, 1H, H-7), 11.31 (br, s, CONH).
1
(M+1)+; H NMR (DMSO-d6) d 4.05 (s, 2H, CH2),
6.26 (br s, 2H, NH2), 7.25–7.42 (m, 3H, 40-, 50-, and 60-
H), 7.43 (d, J=8.8 Hz, 1H, 30-H), 7.47 (br s, 2H, NH2),
8.22 (d, J=2.4 Hz, 1H, H-5), 8.51 (d, J=2.4 Hz, 1H, H-
7); IR (KBr) n 3410, 3340, 3140, 1650, 1620, 1580, 1560,
1460, 1375 cmꢁ1. Anal. calcd for C14H12N5Cl 0.2H2O:
.
C, 58.12; H, 4.32; N, 24.20; Cl, 12.25. Found: C, 57.89;
H, 4.06; N, 23.80; Cl, 12.73.
2,4-Diamino-6-(40 -fluorobenzyl)pyrido[2,3-d]pyrimidine
(2e). Step 1. Addition of 6 (203 mg, 0.5 mmol) in dry
THF (2mL) to 4-fluorobenzylzinc chloride in THF
(0.5 M, 4 mL, calculated to contain 2mmol), followed
by a workup similar to the one for 7a, afforded 7e as a
Step 2. Starting from 7b (84 mg, 0.20 mmol), 2b was
obtained by the same method as 2a except that recrys-
tallization was from EtOH/MeOH/H2O; white solid (20
mg, 31%); mp >300 ꢂC; MS m/z found 312, found 312
(M+1)+; H NMR (DMSO-d6) d 3.68 (s, 3H, 3 - or 4 -
OMe), 3.71 (s, 3H, 30- or 40-OMe), 3.86 (s, 2H, CH2),
6.26 (br s, 2H, NH2), 6.71 (d, J=8.8 Hz, 1H, 60-H), 6.85
(m, 2H, 20 and 50-H), 7.46 (br s, 2H, NH2), 8.23 (s, 1H,
H-5), 8.53 (s, 1H, H-7); IR (KBr) n 3405, 3340, 3130,
0
0
1
1
pale-yellow solid (144 mg, 76%); H NMR (DMSO-d6)
d 1.20 (s, 9H, Me3C), 1.23 (s, 9H, Me3C), 4.12(s, 2H,
CH2), 7.09–7.18 (m, 2H, 20- and 60-H), 7.30–7.37 (m,
2H, 30- and 50-H), 8.46 (s, 1H, H-5), 8.84 (s, 1H, H-7),
11.32(br s, CONH).

A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
65
Step 2. Starting from 7e (140 mg, 0.37 mmol), 2e was
prepared by the same method as 2a except that recrys-
tallization was from DMF/EtOH/H2O; white solid (40
mg, 40%); mp >300 ꢂC; 1H NMR (DMSO-d6) d 3.91 (s,
2H, CH2), 6.20 (br s, 2H, NH2), 7.05–7.08 (m, 2H, 20-
and 60-H), 7.25–7.29 (m, 2H, 30- and 50-H), 7.40 (br s,
2H, NH2), 8.22 (d, J=2.4 Hz, 1H, H-5), 8.51 (d, J=2.4
Hz, 1H, H-7); IR (KBr) n 3410, 3340, 3140, 1650, 1620,
d6) d 5.05 (s, 1H, CH2), 7.27 (d, J=8.4 Hz, 1H, 6-H),
7.45 (d, J=8.4 Hz, 1H, 5-H), 9.97 (s, 1H, CHO).
Step 3. Paraformaldehyde (1.0 g, 34 mmol), N-ethyl-
benzothiazolium bromide (1.5 g, 6.2mmol), and Et N
3
(0.63 g, 6.2mmol) were added to a solution of 10a (6.5
g, 34 mmol) in absolute EtOH (30 mL), and the reaction
mixture was stirred under reflux for 24 h. After removal
of the solvent under reduced pressure, the residue was
shaken with EtOAc (50 mL) until a white solid formed,
which was collected, washed with EtOAc (3ꢃ20 mL),
and discarded. The pooled filtrates were evaporated and
the residue was chromatographed (silica gel, 7:3 hex-
ane–EtOAc) to obtain a-hydroxyketone 11a as a brown
semi-solid (4.4 g, 57%) that was used directly in the next
1580, 1560, 1460, 1375 cmꢁ1
. Anal. calcd for
C14H12N5F 0.2H2O), C, 61.62; H, 4.58; N, 25.66; F,
6.95. Found: C, 61.82; H, 4.47; N, 25.96; F, 6.84.
.
2,4-Diamino - 5 - (30,40 - dichlorobenzyl) - 7H - pyrrolo[2,3-
d]pyrimidine (3a). Step 1. Solid (methoxymethyl)-
triphenylphosphonium chloride (17.7 g, 50 mmol) was
added to a heat-dried 1000-mL three-necked flask con-
taining dry THF (300 mL) which had been freshly dis-
tilled from Na benzophenone ketyl under an argon
atmosphere. The reaction mixture was then stirred and
cooled in a dry ice/acetone bath while n-butyllithium
(2.5 M in hexane, 20 mL, 3.2 g, 55 mmol) was added
dropwise over 30 min. To the flask was then added
dropwise with vigorous stirring over 45 min a solution
of 3,4-dichlorobenzaldehyde (9.2g, 50 mmol) in dry
THF (100 mL). The solution changed slowly in color
from red to yellow and became turbid. After being left
to stir at room temperature overnight, the reaction
mixture was filtered to remove the triphenylphosphine
oxide. The solvent was evaporated under reduced pres-
sure, and the resulting yellow oil was stirred with EtOAc
(50 mL) until a second crop of triphenylphosphine oxide
crystallized out. The latter was filtered off, the filtrate
was evaporated, and the residue was subjected to flash
chromatography (silica gel, 95:5 EtOAc–hexane). A
nonpolar component identified as 1-(3,4-dichloro-
phenyl)butene was discarded, and the slower-moving
band, consisting of a mixture of the Z and E isomer of
9a (total yield 7.7 g, 76%), was used directly in the next
1
step; H NMR (DMSO-d6) d 3.93 (s, 2H, CH2), 4.04 (s,
2H, CH2OH), 4.12(t, 1H, CH OH), 7.22 (s, J=8.5 Hz,
2
1H, 60-H), 7.38 (s, 1H, 20-H), 7.41 (d, J=8.5 Hz, 1H, 50-H).
Step 4. A solution of malononitrile (0.6 g, 10 mmol)
and Et3N (0.38 g, 3.8 mmol) in anhydrous MeOH (5
mL) was added dropwise to a solution of 11a (2.2 g, 10
mmol) in anhydrous MeOH (30 mL) while keeping the
reaction mixture under N2. The solution was stirred at
room temperature overnight, and the solvent was eva-
porated under reduced pressure. The residue was chro-
matographed (silica gel, 7:3 EtOAc–hexane) and
appropriately pooled column fractions were recrystallized
from CH2Cl2 to obtain furan amino nitrile 8a as a brown
solid (1.2g, 45%) that was used directly in the next step;
mp 187–188 ꢂC; H NMR (DMSO-d6) d 3.65 (s, 2H,
1
CH2), 6.80 (s, 1H, furan ¼CH). 7.20 (dd, Jortho=8.4 Hz,
Jmeta=2.2 Hz, 1H, 60-H), 7.29 (broad peak, NH2), 7.46
(d, J=2.2 Hz, 1H, 20-H), 7.55 (d, J=8.4 Hz, 1H, 50-H).
Step 5. Aminonitrile 8a (400 mg, 1.5 mmol) was
added to a solution prepared from guanidine hydro-
chloride (215 mg, 2.3 mmol) and NaOMe (124 mg, 2.3
mmol) in anhydrous MeOH (25 mL). The reaction was
stirred under reflux for 24 h, the solvent was evapo-
rated, and the residue was chromatographed on silica
gel (4:1 CHCl3–MeOH). Evaporation of appropriately
pooled fractions and recrystallization from MeOH
afforded 3a as a brown solid (350 mg, 50%); mp 235–
236 ꢂC; MS m/e found 308, calcd 308 (M+1)+; IR
(KBr) n 3485, 3442, 3205, 1625, 1595, 1570, 1490, 1445
1
step; for the Z isomer: H NMR (DMSO-d6) d 3.63 (s,
3H, OMe), 5.23 (d, J=7.0 Hz, 1H, CH¼CHOMe), 6.40
(d, J=7.0 Hz, 1H, CH¼CHOMe), 7.22 (d, J=8.8 Hz,
1H, 6-H), 7.27 (d, J=8.8 Hz, 1H, 5-H), 7.34 (s, 1H, 2-
1
H); for the E isomer: H NMR (DMSO-d6) d 3.78 (s,
3H, OMe), 5.80 (d, J=13.2Hz, 1H, CH ¼CHOMe),
7.37 (d, J=13.2Hz, 1H, C H¼CHOMe), 7.40 (d, J=8.8
Hz, 1H, 6-H), 7.44 (s, 1H, 5-H), 7.53 (d, J=8.8 Hz, 1H,
2-H).
1
cmꢁ1; H NMR (DMSO-d6) d 4.03 (s, 2H, CH2), 5.44
(br s, 2H, NH2), 5.89 (br s, 2H, NH2), 6.44 (s, 1H, pyr-
role =CH), 7.19 (dd, Jortho=8.2Hz, Jmeta=2.0 Hz, 1H,
Step 2. A solution of 9a (10 g, 49 mmol) in Et2O (2 0
mL) was added dropwise to a stirred pre-cooled (ice
bath) mixture of 70% HClO4 (8.7 mL, 98 mmol), H2O
(8.7 mL), and Et2O (20 mL). After the addition was
complete, the yellow reaction mixture was stirred vigor-
ously at room temperature for 1 h. Additional H2O (30
mL) and Et2O (30 mL) were added, the layers were
separated, and the aqueous layer was extracted with
Et2O (2ꢃ20 mL). The combined organic layers were
washed sequentially with H2O (2ꢃ20 mL), 5%
NaHCO3 (2ꢃ20 mL), and brine (2ꢃ20 mL), and finally
were dried (MgSO4) and evaporated. Column chroma-
tography (silica gel, 7:3 hexane–EtOAc) afforded 10a as
an amber-colored oil (6.5 g, 70%) that was pure enough
0
60-H), 7.46 (d, 1H, 20-H), 7.51 (d, J=8.2Hz, 1H, 5 -H),
10.54 (br s, 1H, pyrrole NH). Anal. calcd for
C13H10N5Cl2: C, 50.83; H, 3.28; N, 22.80; Cl, 23.08.
Found, C, 50.49; H, 3.44; N, 22.53; Cl, 23.06.
2,4-Diamino-5-(30,40,50-trimethoxybenzyl)-7H-pyrrolo[2,3-
d]pyrimidine (3b). Step 1. Starting from (methox-
ymethyl)triphenylphosphonium chloride (17.7 g, 50
mmol), n-butyllithium (2.5 M in hexane, 20 mL, calcu-
lated to contain 3.2g, 55 mmol), and 3,4,5-trimethoxy-
benzaldehyde (10 g, 50 mmol), the same method as was
used to prepare 9a afforded enol ether 9b (estimated to
1
contain 65% of the E and 35% of the Z isomer by H
NMR) as a yellow oil (9.0 g, 80%, used directly in the
1
to be used directly in the next step; H NMR (DMSO-

66
A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
1
next step). For the E isomer: H NMR (DMSO-d6) d
3.61 (s, 3H, ¼CHOMe), 3.73 (s, 3H, 4-OMe), 3.75 (s,
6H, 3- and 5-OMe), 5.77 (d, J=7.2Hz, 1H,
CH¼CHOMe), 6,58 (s, 2H, aromatic), 7.25 (d, J=7.2
Hz, 1H, CH¼CHOMe); for the Z isomer: 1H NMR
(DMSO-d6) d 3.63 (s, 3H, ¼CHOMe), 3.74 (s, 3H, 4-
OMe), 3.76 (s, 6H, 3- and 5-OMe), 5.14 (d, J=12.8 Hz,
1H, CH¼CHOMe), 6,22 (d, J=12.8 Hz, 1H,
CH=CHOMe), 6.84 (s, 2H, 2- and 6-H).
g, 93%, used directly in the next step); 1H NMR (DMSO-
d6) d 1.67 (m, 2H, CH2O), 2.48 (t, 2H, benzylic CH2), 3.36
(m, 2H, CH2), 3.5 (s, 2H, 4-OMe), 3.69 (s, 6H, 3- and
5-OMe), 4.41 (t, 1H, OH), 6.43 (s, 2H, 2- and 6-H).
Step 2. CrO3 (1.2g, 12.2mmol) was added in small
portions to a stirred mixture of pyridine (2mL) and
anhydrous CH2Cl2 (15 mL) at room temperature. After
15 min, the brown suspension was treated with a solu-
tion of 13 (0.4 g, 2.0 mmol) in CH2Cl2 (2mL), and
stirring was continued for 1 h. The liquid was decanted
and the residue triturated with Et2O (3ꢃ50 mL). The
pooled Et2O extracts were washed successively with
H2O (4ꢃ20 mL), 5% NaOH (2ꢃ10 mL), 5% HCl (2ꢃ20
mL), and finally 50% saturated NaHCO3 (2ꢃ20 mL)
The organic layer was dried (MgSO4) and evaporated
under reduced pressure, and the residue was chromato-
graphed on silica gel (9:1 CHCl3–MeOH) to obtain
aldehyde 14 as a yellow oil (244 mg, 53%, used directly
Step 2. Starting from 9b (9 g, 40 mmol) and 70%
HClO4 (7 mL, 80 mmol), aldehyde 10b was obtained by
the same method as 10a, except that the column eluent
was 4:1 hexane–EtOAc; yellow oil (6.0 g, 71%, used
directly in the next step); H NMR (DMSO-d6) d 3.75
(s, 2H, 3- and 5-OMe), 3.77 (s, 3H, 4-OMe), 3.87 (s, 2H,
1
CH2), 7.25 (s, 2H, 2- and 6-H), 9.66 (s, 1H, CH¼O).
Step 3. Starting from paraformaldehyde (0.84 g, 29
mmol), N-ethylbenzothiazolium bromide (1.3 g, 5.2
mmol), Et3N (0.53 g, 5.2mmol), and 10b (6.0 g, 29 mmol),
a-hydroxyketone 11b was obtained by the same method as
11a; yellow oil (3.1 g, 45%, used directly in the next step);
1H NMR (DMSO-d6) d 3.31 (s, 2H, CH2), 3.70 (s, 3H,
4-OMe), 3.88 (s, 6H, 3- and 5-OMe), 3.92(s, 2H, C H2OH),
3.98 (br s, 1H, CH2OH), 7.21 (s, 2H, 2- and 6-H).
1
in the next step); H NMR (DMSO-d6) d 2.77 (m, 4H,
CH2CH2), 3.60 (s, 3H, 4-OMe), 3.74 (s, 6H, 3- and 5-
OMe), 6.52 (s, 2H, 2-H and 6-H).
Step 3. Starting from 14 (7 g, 31 mmol), paraf-
ormaldehyde (0.96 g, 32mmol), N-ethylbenzothiazo-
lium bromide (1.4 g, 5.6 mmol), and Et3N (0.58 g, 5.6
mmol), a-hydroxyketone 15 was obtained by the same
method as 11a; brown oil (5.0 g, 63%, used directly in
the next step); 1H NMR (CDCl3) d 2.72 (m, 4H,
CH2CH2), 3.60 (s, 3H, 4-OMe), 3.73 (s, 6H, 3- and
5-OMe), 4.80 (d, 2H, CH2OH), 5.10 (s, 2H, 2- and 6-H).
Step 4. Starting from malononitrile (0.25 g, 3.8
mmol), Et3N (0.38 g, 3.8 mmol), and 11b (0.90 g, 3.8
mmol), furan amino nitrile 8b was obtained by the same
method as 8a except that the column eluent was CHCl3;
brown solid (0.66 g, 60%, used directly in the next step);
mp 127–128 ꢂC; H NMR (DMSO-d6) d 3.55 (s, 2H,
Step 4. Starting from 15 (5 g, 20 mmol), malononitrile
(1.4 g, 22 mmol) and Et3N (2.2 g, 22 mmol), furan
amino nitrile 16 was obtained by the same method as 8a
except that the column eluent was 95:5 CHCl3–MeOH;
yellow solid (5 g, 84%, used directly in the next step);
1
CH2), 3.62(s, 3H, 4 -OMe), 3.73 (s, 6H, 30- and 50-
0
OMe), 6.55 (20- and 60-H), 6.77 (s, 1H, furan ¼CH).
Step 5. Starting from 8b (200 mg, 0.7 mmol), guani-
dine hydrochloride (100 mg, 1.05 mmol) and NaOMe
(57 mg, 1.05 mmol), 3b was obtained by the same
method as 3a except that the column eluent was 85:15
CHCl3–MeOH; yellowish-white solid (150 mg, 65%);
mp 222–223 ꢂC; MS m/e found 330, calcd 330 (M+1)+;
IR (KBr) n 3440, 3365, 3210, 1610, 1555, 1510, 1465,
mp 151–152 ꢂC; H NMR (DMSO-d6) d 2.56 (m, 2H,
1
CH2), 2.74 (m, 2H, benzylic CH2), 3.61 (s, 3H, 4-OMe),
3.74 (s, 6H, 3- and 5-OMe), 6.50 (s, 2H, 2- and 6-H),
6.74 (s, 2H, furan ¼CH), 7.22 (br s, 2H, NH2).
Step 5. Starting from 15 (200 mg, 0.66 mmol), guani-
dine hydrochloride (95 mg, 1.0 mmol) and NaOMe (54
mg, 1.0 mmol), 3c was obtained by the same method as
3a; brown powder (100 mg, 44%); mp 210–211 ꢂC; MS
m/z found 344, calcd 344 (M+1)+; IR (KBr) n 3440,
3260, 3000, 1620, 1575, 1530, 1480, 1445 cmꢁ1; 1H NMR
(DMSO-d6) d 2.77 (m, 2H, CH2), 2.93 (m, 2H, benzylic
CH2), 3.60 (s, 3H, 40-OMe), 3.73 (s, 6H, 30- and 50-OMe),
5.33 (br s, 2H, NH2), 5.93 (br s, 2H, NH2), 6.41 (s, 1H,
pyrrole ¼CH), 6.54 (s, 2H, 20- and 60-H), 10.35 (br s, 1H,
pyrrole NH). Anal. calcd for C17H21N5O3: C, 59.46; H,
6.16; N, 20.40. Found: C, 59.22; H, 6.22; N, 20.35.
1430 cmꢁ1 1H NMR (DMSO-d6) d 3.60 (s, 3H, 40-
;
OMe), 3.68 (s, 6H, 30- and 50-OMe), 3.92(s, 2H, CH 2),
5.35 (br s, 2H, NH2), 5.56 (br s, 2H, NH2), 6.40 (s, 1H,
pyrrole ¼CH), 6.56 (s, 2H, 20- and 60-H), 10.42(br s,
1H, pyrrole NH). Anal. calcd for C16H19N5O3: C, 58.35;
H, 5.81; N, 21.26; Found, C, 57.97; H, 5.88; N, 20.97.
2,4-Diamino-5-[2-(30,40,50 -trimethoxyphenyl)ethyl]-7H-
pyrrolo[2,3-d]pyrimidine (3c). Step 1. To a stirred sus-
pension of LiAlH4 (1.28 g, 32 mmol) in dry THF (50 mL,
freshly distilled from CaH2) at 0 ꢂC under argon was
added dropwise a solution of 12 (5 g, 20 mmol) in dry
THF (50 mL). The mixture was stirred at 0 ꢂC for 2 h and
then at room temperature overnight. After being cooled
again to 0 ꢂC, the excess hydride was quenched by drop-
wise addition of 20% H2O in THF (50 mL). The color of
the suspension changed from gray to yellow. The mixture
was filtered through Celite, and the filter cake was washed
with Et2O until the washings were colorless. Evaporation
of the pooled filtrates afforded 13 as a pale-yellow oil (4.2
Acknowledgements
This work was supported by Grant RO1-AI29904 from
the National Institute of Allergy and Infectious Dis-
eases, NIH, DHHS. We are grateful to Dr. Ying-Nan
Chen for carrying out the growth inhibition assays with
CCRF-CEM cells.

A. Rosowsky et al. / Bioorg. Med. Chem. 11 (2003) 59–67
67
References and Notes
8. Suling, W. J.; Seitz, L. E.; Pathak, V.; Westbrook, L.; Bar-
row, E. W.; Zywno-van Ginkel, S.; Reynolds, R. C.; Piper,
J. R.; Barrow, W. W. Antimicrob. Agents Chemother. 2000, 44,
2784.
9. Rosowsky, A.; Chen, H. J. Org. Chem. 2002, 66, 7522.
10. (a) Gangjee, A.; Mavandadi, F.; Queener, S. F.; McGuire,
J. J. J. Med. Chem. 1995, 38, 2158. (b) Gangjee, A.; Mavan-
dadi, F.; Queener, S. F. J. Med. Chem. 1997, 40, 1173.
11. Rosowsky, A.; Fu, H.; Queener, S. F. J. Heterocycl.
Chem. 2001, 38, 1197.
1. Klepser, M. E.; Klepser, T. B. Drugs 1997, 53, 40.
2. Falloon, J.; Allegra, C. J.; Kovacs, J.; O’Neill, D.; Ogata-
Arakaki, D.; Feuerstein, I.; Polis, M.; Davey, R.; Lane, H. C.;
LaFon, S.; Rogers, M.; Zunich, K.; Turlo, J.; Tuazon, C.;
Parenti, D.; Simon, G.; Masur, H. Clin. Res. 1990, 38, 361.
3. Kovacs, J. A.; Allegra, C. J.; Beaver, J.; Boarman, D.;
Lewis, M.; Parrillo, J. E.; Chabner, B.; Masur, H. J. Infect.
Dis. 1989, 160, 312.
4. (a) Gangjee, A.; Shi, J.; Queener, S. F.; Barrows, L. R.;
Kisliuk, R. L. J. Med. Chem. 1993, 36, 3437. (b) Gangjee, A.;
Vasudevan, A.; Queener, S. F.; Kisliuk, R. L. J. Med. Chem.
1995, 38, 1778. (c) Gangjee, A.; Devraj, R.; Queener, S. F. J.
Med. Chem. 1997, 40, 470.
5. Piper, J. R.; Johnson, C. A.; Krauth, C. A.; Carter, R. L.;
Hosmer, C. A.; Queener, S. F.; Borotz, S. E.; Pfefferkorn,
E. R. J. Med. Chem. 1996, 39, 1271.
12. (a) Taylor, E. C.; Patel, H.; Jun, J.-G. J. Org. Chem. 1995,
60, 6684. (b) Taylor, E. C.; Mao, Z. J. Org. Chem. 1995, 61,
7973.
13. (a) Broughton, M. C.; Queener, S. F. Antimicrob. Agents
Chemother. 1991, 35, 1348. (b) Chio, L.-C.; Queener, S. F.
Antimicrob. Agents Chemother. 1993, 37, 1914. (c) Rosowsky,
A.; Cody, V.; Galitsky, N.; Fu, H.; Papoulis, A. T.; Queener,
S. F. J. Med. Chem. 1999, 42, 4853.
6. Gangjee, A.; Vasudevan, A.; Queener, S. F.; Kisliuk, R. L.
J. Med. Chem. 1996, 39, 1438.
14. Taylor, I. W.; Slowiaczek, P.; Friedlander, M. L.; Tatter-
sall, M. H. N. Cancer Res. 1985, 45, 978.
7. Gangjee, A.; Adair, O.; Queener, S. F. Bioorg. Med. Chem.
2002, 9, 2929.
15. Rosowsky, A.; Forsch, R. F.; Queener, S. F. J. Med.
Chem. 2002, 45, 233.