Synthesis of new 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines via the Taylor ring transformation/ring annulation strategy

Article abstract of DOI:10.1002/jhet.5570380526

Selected examples 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines with a phenyl or benzyl group at the 5-position were synthesized as inhibitors of dihydrofolate reductase (DHFR) from Pneumocystis carinii and Toxoplasma gondii, two potentially life-threatening opportunistic pathogens associated with AIDS and other disorders of the immune system. Aldol condensation of paraformaldehyde with substituted benzaldehydes or with phenylacetaldehyde afforded α-hydroxyketones with a phenyl or benzyl group at the 4-position. Further reaction of the hydroxyketones with malononitrile afforded 2-aminofuran-3-carbonitriles, which upon heating with guanidine underwent ring transformation/ring annulation to produce 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines rather than 2,4-diaminofuro[2,3-d]pyrimidines. One of the target compounds obtained in this manner, 2,4-diamino-5-(3,4,5-trimethoxyphenyl)-7H-pyrrolo [2,3 -d]pyrimidine (1d), may be viewed as a conformationally restricted analogue of trimethoprim, an antimicrobial agent widely used in combination with a sulfa drug to treat P. carinii and T. gondii opportunistic infections in patients with AIDS. Compound 1d inhibited P. carinii and T. gondii DHFR with IC₅₀ values of 8.3 and 14 μM, respectively. This potency was somewhat greater than that of trimethoprim. However, because this compound was also more potent than trimethoprim against mammalian (rat liver) DHFR rat liver it lacked species selectivity. The other 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines synthesized were neither potent nor selective.

Full text of DOI:10.1002/jhet.5570380526

Sep-Oct 2001
Synthesis of New 2,4-Diamino-7H-pyrrolo[2,3-d]pyrimidines
via the Taylor Ring Transformation/Ring Annulation Strategy
Andre Rosowsky* [a], Hongning Fu [a] and Sherry F. Queener [b]
1197
[a] Dana-Farber Cancer Institute and the Department of Biological Chemistry and Molecular Pharmacology,
Harvard Medical School, Boston, MA 02115
[b] Department of Pharmacology and Toxicology, Indiana University School of Medicine, Indianapolis, IN 46202
Received May 11, 2001
Selected examples 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines with a phenyl or benzyl group at the
5-position were synthesized as inhibitors of dihydrofolate reductase (DHFR) from Pneumocystis carinii
and Toxoplasma gondii, two potentially life-threatening opportunistic pathogens associated with AIDS
and other disorders of the immune system. Aldol condensation of paraformaldehyde with substituted
benzaldehydes or with phenylacetaldehyde afforded α-hydroxyketones with a phenyl or benzyl group at
the 4-position. Further reaction of the hydroxyketones with malononitrile afforded 2-aminofuran-3-
carbonitriles, which upon heating with guanidine underwent ring transformation/ring annulation to
produce 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines rather than 2,4-diaminofuro[2,3-d]pyrimidines. One
of the target compounds obtained in this manner, 2,4-diamino-5-(3,4,5-trimethoxyphenyl)-7H-
pyrrolo[2,3-d]pyrimidine (1d), may be viewed as a conformationally restricted analogue of trimethoprim,
an antimicrobial agent widely used in combination with a sulfa drug to treat P. carinii and T. gondii
opportunistic infections in patients with AIDS. Compound 1d inhibited P. carinii and T. gondii DHFR
with IC values of 8.3 and 14 µM, respectively. This potency was somewhat greater than that of trimetho-
50
prim. However, because this compound was also more potent than trimethoprim against mammalian (rat
liver) DHFR rat liver it lacked species selectivity. The other 2,4-diamino-7H-pyrrolo[2,3-d]pyrimidines
synthesized were neither potent nor selective.
J. Heterocyclic Chem., 38, 1197 (2001).
As part of in an ongoing program of design, synthesis,
of tetrahydrofolate, namely leucovorin (5-formyl-
tetrahydrofolate). Experimental regimens using highly
toxic DHFR inhibitors to treat severe P. carinii and
T. gondii infections refractory to other classes of drugs are
based on this rationale. An alternative approach that would
obviate the need to co-administer leucovorin would be to
use a single agent that binds preferentially to the DHFR
from P. carinii or T. gondii, thereby decreasing the risk to
the patient. A number of newer DHFR inhibitors that bind
more tightly to P. carinii DHFR, and especially to T. gondii
DHFR, have been discovered in the last few years, but thus
far have not been tested clinically. A most vexing problem
in designing such drugs has been that, in most of the
examples studied thus far, selectivity of binding appears to
be inversely correlated with potency. Thus, a major goal of
our research and that of investigators working in this area
has been to find a DHFR inhibitor that is both potent and
selective.
and biochemical and biological evaluation of various types
of lipophilic condensed 2,4-diaminopyrimidine ring
systems as inhibitors of dihydrofolate reductase (DHFR)
and potential drugs against life-threatening opportunistic
infections in AIDS patients with a dysfunctional immune
system [1,2], we wish to report the synthesis of several
previously unknown 2,4-diaminopyrrolo[2,3-d]-
pyrimidines with a substituted phenyl and, in one example,
an unsubstituted benzyl substituent at the 5-position.
2,4-Diaminopyrrolo[2,3-d]pyrimidines are but one of a
number of condensed diaminopyrimidine ring systems
known to bind tightly to dihydrofolate reductase (DHFR),
the enzyme responsible for maintaining adequate levels of
essential tetrahydrofolate cofactors in dividing cells.
Mammalian cells take up reduced folates efficiently from
their exogenous environment by making use of membrane-
associated transporter proteins, and thus do not have to
synthesize these folates de novo. However these
transporter proteins are not present in Pneumocystis carinii
and Toxoplasma gondii, two common opportunistic
pathogens that often provide the earliest clue that someone
has AIDS, and can themselves prove fatal if left untreated,
especially if both organisms are present concurrently. The
rationale for using lipophilic DHFR inhibitors as drugs
against P. carinii, T. gondii and other opportunistic
parasites lacking folate transporter proteins on their
surface is that sensitive host tissues such as the bone mar-
row can be preferentially spared from the toxic effects of
the antifolate by co-administration of an exogenous source
Although a number of 2,4-diaminopyrrolo[2,3-d]-
pyrimidines with lipophilic aromatic substituents at the
5-position have been described [3,4], they all contain a
CH NH, CH N(Me), or CH S bridge between the pyrrole
2
2
2
and aryl moiety. In the present work we were interested in
preparing compounds in which the two halves of the
molecule are joined via a shorter bridge in order to assess
the effect of this particular structural modification on
potency and selectivity. To this end, we took advantage of
an elegant strategy first described by Taylor and coworkers
[5,6], called "one-step ring-transformation/ring
annulation". At the heart of this strategy is the intriguing

1198
A. Rosowsky, H. Fu and S. F. Queener
Vol. 38
proclivity of 2-aminofuran-3-carbonitriles with a
substituent at the 4-position to form 2,4-diamino-
pyrrolo[2,3-d]pyrimidines rather than 2,4-diamino-
furo[2,3-d]pyrimidines upon being heated in the presence
guanidine. Although two plausible pathways were
proposed for this novel transformation, a choice between
these pathways could not be made.
Among the simple model compounds reported by Taylor
[5] was 2,4-diamino-5-phenylpyrrolo[2,3-d]pyrimidine
(1a), which formed in 42% yield when 2-amino-4-phenyl-
furan-3-carbonitrile (1b) was heated with guanidine in
refluxing methanol for 24 hours (Scheme 1). Several other
2,4-diaminopyrrolo[2,3-d]pyrimidines with methyl, ethyl,
phenyl groups at the 5- and/or 6-position were obtained in
similar fashion, as well as analogs in which the
5-substituent was a 2-phenylethyl or 3-phenylpropyl group
As shown in Scheme 1, aldol condensation of
appropriately substituted benzaldehydes with para-
formaldehyde in the presence of N-ethylbenzothiazolium
bromide and triethylamine [5-7] afforded the desired
α-hydroxyketones 2a-d. Further reaction of 2a-d with
malononitrile and triethylamine in refluxing methanol
yielded the furans 3a-d, which upon heating with
1.5 molar equivalents of guanidine in refluxing ethanol,
typically for 30 hours, were converted to the pyrrolo-
pyrimidines 1a-d. In the case of 1e, phenylacetaldehyde
was used in place of benzaldehyde, giving rise to the
α-hydroxyketone 2e and the furan 3e.
with a CO Me group at the para position. The latter esters
2
were used to prepare classical DHFR inhibitors with a
glutamic acid side chain.
Because of the common presence of aromatic chloro
and methoxy groups in therapeutically useful lipophilic
DHFR inhibitors, a selected number of analogs of 1a
containing this pharmacophoric motif were prepared, the
form of compounds 1b-e (Scheme 1) In addition, because
Taylor and coworkers had not used the ring transforma-
tion/ring annulation method to make 2,4-diamino-5-
benzylpyrrolo[2,3-d]pyrimidine (1e, Scheme 1) their
method was used to also prepare this compound as the
simplest example of a 5-benzyl derivative. The 4-chloro-
and 3,4-dichlorophenyl derivatives 1b and 1c may be
viewed as analogs of pyrimethamine and metoprine, and
the 3,4,5-trimethoxy derivative 1d as an analog of
trimethoprim. The structures of trimethoprim,
pyrimethamine, metoprine, piritrexim, and trimetrexate
are given in Figure 1.
Compounds 1a-e were all obtained in analytically pure
form by column chromatography on silica gel, and were
characterized spectroscopically and by microchemical
analysis. A characteristic feature of each product was the
1
presence in the H-nmr spectrum of a singlet at ca. δ 6.8,
Figure 1. Structures of clinically used lipophilic DHFR inhibitors.
which could be assigned to the C6 proton of a

Sep-Oct 2001
Synthesis of New 2,4-Diamino-7H-pyrrolo[2,3-d]pyrimidines
1199
pyrrolo[2,3-d]pyrimidine and was in agreement with the
interest is that 2,4-diamino-(3,4,5-trimethoxy-
phenyl)furo[2,3-d]pyrimidine (4), the oxygen analog of
reported value for 1a [5]. The benzylic CH group in the
2
1e, is reported to have an IC >300 µM against both
furan 3e gave a singlet at δ 3.61, and thus was less
deshielded than the same proton in the pyrrole 1e (δ 4.00).
Interestingly, the vinylic C6 proton in 3e (δ 6.76) was less
deshielded than the same proton in 3a (δ 6.99) and the C6
proton in 1e (δ 6.32) was less deshielded than the same
proton in 1a (δ 6.78). Wahid and coworkers [8] recently
reported that the vinylic C6 proton in 2,4-diamino-5-
(3,4,5-trimethoxyphenyl)furo[2,3-d]pyrimidine (4) gives
rise to a singlet at δ 7.53, a chemical shift considerably
higher than the one we observed for 1d. Final proof that
1a-e were pyrrolo[2,3-d]pyrimidines rather than furo-
[2,3-d]pyrimidines, as would have been the case if the
guanidine reaction had occurred without rearrangement,
came from the observation that all these compounds
showed an exchangeable singlet at ca. δ 11.0 corres-
ponding to the pyrrole NH.
50
P. carinii and human DHFR [8]. From our results it would
appear that the acidic proton at N7 must play a significant
role in the binding of 1e to DHFR.
An interesting analog of trimethoprim and 1d,
synthesized and studied in detail in 1996 by Kuyper and
coworkers [9], is 2-amino-4-methyl-5-(3,4,5-trimetho-
xyphenyl)-7H-pyrrolo[2,3-d]pyrimidine (6), in which the
normal role of the 4-amino group in DHFR binding is now
played by the pyrrole nitrogen. The IC values of 6 against
50
rat DHFR was found to be 29 µM as compared with
350 µM for trimethoprim, a difference of 12-fold. The
corresponding values against Escherichia coli DHFR,
whose active site is more compact than that of mammalian
enzymes, were 0.44 and 0.007 µM respectively, a
difference of 63-fold. Crystallographic analysis of the
binary complex of 6 with Escherichia coli DHFR
suggested that the 2-amino and 3,4,5-trimethoxyphenyl
groups occupy approximately the same locations in the
enzyme active site as the 2-amino and 3,4,5-trimethoxy-
benzyl groups of trimethoprim, with a slight displacement
to accommodate the sterically hindered biphenyl-like
structure of 6, whose two halves are forced to lie in a nearly
perpendicular orientation to each another. Although
structural analysis of the ternary enzyme complex
containing NADPH as well as the ligand in the active was
not carried out, the authors speculated that the stronger
binding of 6 to both the bacterial and mammalian enzyme
Spectrophotometric assays of the ability of 1a-e to inhibit
partly purified DHFR from P. carinii, T. gondii, and rat
liver were carried out according to a standardized protocol
one of us (R.F.Q.) developed and extensively used in the
past to test compounds synthesized in this program [2].
With the exception of the 3,4,5-trimethoxy derivative 1d,
whose IC values were 8.3 µM against the P. carinii
50
enzyme, 14 µM against the T. gondii enzyme, and 27 µM
against the rat liver enzyme, all the compounds were weak
inhibitors, and appeared to lack species selectivity.
Gangjee and coworkers [3] have presented DHFR
binding data for 2,4-diamino-5-(3,4,5-trimethoxy-
anilino)methyl-7H-pyrrolo[2,3-d]pyrimidine (5), an ana-
log of 1d in which the two halves of the molecule are
joined via a CH NH bridge. The IC values of 5 were
might be due to its higher pK , which results in 6%
a
protonation at neutral pH as compared with 56% in the case
of trimethoprim. Because 1d is a 2,4-diamino derivative its
pK is likely to be closer to trimethoprim than to 6. Thus
2
50
a
>23 µM against P. carinii DHFR, 8.1 µM against T. gondii
DHFR, and 56 µM against rat liver DHFR. Thus our data
our finding that 1d and 6 have almost the same IC against
50
rat DHFR suggests that other factors than just basicity may
have to be considered in order to explain the improved
binding of these conformationally restricted analogs
relative to trimethoprim.
suggest that removal of the CH NH bridge from 5
2
increased potency against the P. carinii and rat enzyme
while decreasing it against the T. gondii enzyme. Also of

1200
A. Rosowsky, H. Fu and S. F. Queener
Vol. 38
Table 1
Dihydrofolate Reductase Inhibition by Pyrrolo[2,3-d]pyrimidines 1a-e
IC (µM) [a]
Selectivity Index [b]
50
Compound
P. carinii
T. gondii
Rat Liver
P. carinii
T. gondii
1a
1b
1c
1d
30% @ 186 µM
27% @ 161 µM
12
113
16
14
9.1
62
23
27
270
130
[c]
[c]
0.68
3.2
[c]
0.76
0.55
1.5
1.9
[c]
33
8.3
30% @ 189 µM
12
1e
44% @ 130
2.7
Trimethoprim [d]
11
48
[a] Enzyme activity was determined according to a standardized method used this program reliably for a number of years [12,13]. Each drug
concentration was tested in triplicate. As an example of the reproducibility of the assay, the IC values (mean ± standard error) in Dr. Queener's
50
laboratory over a five-year period using pyrimethamine and partially purified dihydrofolate reductase from P. carinii, T. gondii, and rat liver have been
2.39 ± 0.42, 0.50 ± 0.23, and 1.52 ± 0.32 µM, respectively; [b] Selectivity Index = IC (rat liver)/IC (P. carinii or T. gondii); [c] Not determined
50
50
because the solubility of the drug was exceeded in one or both assays; [d] Data cited in reference 2.
EXPERIMENTAL
The finding that 1d is more potent than trimethoprim
against P. carinii and rat liver DHFR while being less
potent against the T. gondii enzyme is also of interest.
While the distance between the phenyl ring and the
diaminopyrimidine moiety is similar in 1d and trimetho-
prim, the conformational flexibility of the 3,4,5-
trimethoxybenzyl side chain of trimethoprim is clearly
absent in 1d. Molecular models indicate that the only way
unfavorable steric interaction can be avoided between the
4-amino group and the ortho hydrogen of the phenyl ring is
for the latter to adopt a severely twisted conformation, as
has been found in the case of 6 [9] A consequence of this
nearly orthogonal configuration is that the methoxy groups
of 1d must lie in a different spatial location in the DHFR
active site than those of trimethoprim. Moreover, a
consequence of the sterically constrained biphenyl-like
structure of 2,4-diamino-5-aryl-7H-pyrrolo[2,3-d]pyrimi-
dines is that these compounds can in principle exist as pairs
of non-superimposable rotamers with different energies of
binding to dihydrofolate reductase. Our results suggest that
the orthogonal orientation of the rings in 1d may be toler-
ated better in the active site of the P. carinii and rat enzyme
than in the active site of the T. gondii enzyme.
Given the dissimilarity in their three-dimensional
structures, the finding that 1d and trimethoprim so closely
resembles each other in their ability to inhibit DHFR is a
novel and possibly exploitable lead. Substitution of a
pyrrole ring for one of the methoxy groups in trimethoprim,
as well replacement of one or more of the methoxy groups
by other alkoxy groups, has been reported to selectively
enhance binding to P. carinii versus mammalian DHFR
[10], and was the basis for the recent development of
2,4-diamino-5-(3,5-diethoxy-4-pyrrolobenzyl)pyrimidine
(7, epiroprim) as a promising new drug against various
resistant microbial organisms including P. carinii [10].
Inasmuch as 1d appears to represent a sparsely studied
class of DHFR inhibitors, this compound may be viewed as
a potential candidate for lead optimization studies aimed at
improving potency and selectivity.
Infrared spectra were obtained on a Perkin-Elmer Model 781
double-beam recording spectrophotometer. H-nmr spectra were
1
recorded at 60 MHz (Dana-Farber Cancer Institute) or at 200 or
500 MHz (Harvard Medical School). Low-resolution mass
spectra were obtained by the Molecular Biology Core Facility,
Dana-Farber Cancer Institute, using either the electron impact
(EI) or fast atom bombardment (FAB) mode. Analytical
thin-layer chromatography (tlc) was on Whatman MK6F silica
gel slides (60 µm layer, 1.5 x 4.5 cm, with fluorescent indicator),
using 254-nm illumination to visualize the spots. Column
chromatography was on Baker 7024 flash silica gel (40 µm
particle size). Melting points were determined in Pyrex capillary
tubes using a Mel-Temp apparatus (Laboratory Devices, Inc.,
Cambridge, MA) and are not corrected. Solvents for moisture-
sensitive reactions were of 'Sure-Seal' grade (Aldrich,
Milwaukee, WI). Other chemicals obtained from commercial
sources were of the best available purity. Elemental analyses
were performed by QTI Laboratories, Madison, NJ, and were
within ±0.4% of theoretical values unless otherwise indicated.
2,4-Diamino-5-phenyl-7H-pyrrolo[2,3-d]pyrimidine (1a).
Step 1 (Method A): Paraformaldehyde (0.63 g, 20 mmole),
(N-ethylbenzothiazolium bromide (0.90 g, 3.6 mmole), and
triethylamine (0.37 g, 3.6 mmole) were added with magnetic
stirring to a solution of benzaldehyde (2.14 g, 20 mmole). The
mixture was stirred under reflux overnight, the solvent was
evaporated under reduced pressure, and the residue was taken up
in ethyl acetate (30 mL). The white solid that formed gradually
on standing was collected and washed with ethyl acetate (10 mL).
The filtrate was evaporated to form a brown oil which slowly
solidified in the refrigerator. Recrystallization from a 7:3
hexanes-ethyl acetate mixture afforded 2a as colorless crystals
1
(1.2 g, 44%, used directly in the next step), mp 63-64°; H-nmr
(60 MHz, deuteriochloroform) δ 3.51 (s, 1H, CH OH),
2
4.85 (s, 2H, CH OH), 7.41-7.98 (m, 5H, aromatic protons).
2
Step 2 (Method B) A solution of malononitrile (0.60 g,
9.1 mmole) and triethylamine (0.92 g, 9.1 mmole) in anhydrous
methanol (10 m) was added dropwise under a nitrogen
atmosphere to a solution of 2a (1.1 g, 8.3 mmole) in anhydrous
methanol (30 mL). The mixture was stirred magnetically at room

Sep-Oct 2001
1201
Synthesis of New 2,4-Diamino-7H-pyrrolo[2,3-d]pyrimidines
temperature for 15 hours. After removal of the solvent by rotary
evaporation, the residue was chromatographed on silica gel with
chloroform as the eluent to obtain 3a as a yellow oil, which
1
92-93°) was used directly in the next step; H-nmr (200 MHz,
d -dimethylsulfoxide): δ 4.78 (s, 2H, NH ), 7.17 (s, 1H, C=CH),
6
2
7.50-7.85 (m, 3H, aromatic protons). Condensation of 3c with
guanidine was carried out by Method C except that the product
(1c) was chromatographed with 4:1 chloroform-methanol;
off-white powder (49%), mp 264° dec; ms: theory for
became partially solid on standing (0.58 g, 39%, used directly in
1
the next step); H-nmr (60 MHz, d -dimethylsulfoxide) δ 5.45
6
(br s, 2H, NH ), 6.99 (s, 1H, C=CH), 7.21-7.85 (m, 5H, aromatic
2
protons).
C
H N Cl m/z 294, 295, 296; found 294, 295, 296; ir (KBr) ν
12
9 5 2
Step 3 (Method C). Amino nitrile 3a (184 mg, 1 mmole) was
added to a solution prepared from guanidine hydrochloride
(140 mg, 1.5 mmole) and sodium methoxide (81 mg, 1.5 mmole)
in anhydrous methanol (25 mL). The reaction mixture was stirred
under reflux for 48 hours, the solvent was removed by rotary
evaporation, and the residue was chromatographed on silica gel
with 95:5 chloroform-methanol as the eluent. Fractions
containing the desired product were combined and evaporated
under reduced pressure to obtain 1a as a slightly purplish white
3470, 3370, 3330, 3140, 1605, 1570, 1530, 1465, 1130, 1025,
-1
1
980, 790 cm ; H-nmr (200 MHz, d -dimethylsulfoxide) δ 5.59
6
(br s, 2H, 2-NH ), 5.71 (br s, 2H, 2H, 4-NH ), 6.95 (s, 1H,
2
2
C=CH), 7.38 (dd, 1H, J = 8 Hz, J = 1.8 Hz, 5'-H), 7.61 (d, 1H,
ab
ac
J
= 1.8 Hz, 2'-H), 7.63 (d, 1H, J = 8.0 Hz, 4'-H), 11.08 (s, 1H,
ac
ab
pyrrole NH).
Anal. Calcd. for C H N Cl •0.55CH OH: C, 48.35; H, 3.62;
12
9
5
2
3
N, 22.46; Cl, 22.74. Found: C, 48.48; H, 3.22, N, 22.14; Cl,
22.60.
solid (128 mg, 57%), mp 215-216°; ms: theory for C H N m/z
12 11
5
2,4-Diamino-5-(3,4,5-trimethoxyphenyl)-7H-pyrrolo[2,3-d]-
pyrimidine (1d).
225; found 225; ir (potassium bromide) ν 3472, 3310, 3161,
-1
1
1600, 1565, 1540, 1474, 1418, 1304, 1070, 960 cm ; H-nmr
(500 MHz, d -dimethylsulfoxide): δ 5.40-5.70 (br s, 2- and
6
3,4,5-Trimethoxy-α-hydroxyacetophenone (2d) was obtained
from 3,4,5-trimethoxybenzaldehyde by Method A, and was used
4-NH ), 6.78 (d, 1H, C=CH), 7.30-7.43 (m, 5H, aromatic pro-
2
tons), 10.95 (broad s, 1H, pyrrole NH).
directly in the next step; colorless crystals (31%), mp 73-74°
Anal. Calcd. For C
H N •0.1H O: C, 63.48; H, 4.97; N,
1
12 11 5 2
(90:10 hexanes-ethyl acetate); H-nmr (200 MHz, d -dimethyl-
6
30.84. Found: 63.26; H, 4.79; N, 30.64.
sulfoxide): δ 3.72 (s, 3H, 4-OMe), 3.83 (s, 6H, 3- and 5-OMe),
4.79 (d, 2H, J = 5.8 Hz, CH OH), 5.02 (t, 1H, J = 5.8 Hz,
2,4-Diamino-5-(4-chlorophenyl)-7H-pyrrolo[2,3-d]pyrimidine
(1b).
2
CH OH), 7.21 (s, 2H, aromatic protons). 2-Amino-4-(3,4,5-
2
trimethoxyphenyl)furan-3-carbonitrile (3d) was obtained from
2d by Method B except that the product was chromatographed
4-Chloro-α-hydroxyacetophenone (2b) was obtained from
4-chlorobenzaldehyde according to Method A, and was used
directly for the next step; yellow crystals (48%), mp 111-112°
with 3:2 hexane-ethyl acetate; dark-brown powder (30%), mp
1
117-118°; H-nmr (200 MHz, d -dimethylsulfoxide): δ 3.66
6
1
(from 7:3 hexanes-ethyl acetate); H-nmr (60 MHz, deuterio-
(3H, 4-OMe), 3.79 (s, 6H, 3- and 5-OMe), 4.82, br s, 2H, NH )
2
chloroform): δ 3.85 (s, 1H, CH OH), 4.88 (s, 2H, CH OH), 7.45
2
2
6.86 (s, 2H, 2'- and 6'-H), 7.41 (s, 1H, C=CH). Condensation of
3d with guanidine was carried out by Method C except that the
product (1d) was chromatographed with 4:1 chloroform-
methanol; tan powder (55%), mp 265-266°; ir (potassium
bromide) ν 3480, 3380, 3170, 1605, 1565, 1410, 1360, 1120, 995
(d, 2H, aromatic protons), 7.88 (d, 2H, aromatic protons).
2-Amino-4-(4-chlorophenyl)furan-3-carbonitrile (3b) was
obtained from 2b and malononitrile by Method B except that the
product was chromatographed with ethyl acetate, then
recrystallized from dichloromethane. The greenish-white product
-1
1
cm ; ms: theory for C
H N O m/z 316, found 316; H-nmr
1
15 17 5 3
(53%, mp 155-156°) was used directly in the next step; H-nmr
(200 MHz, d -dimethylsulfoxide): δ 3.67 (s, 3H, 4-OMe), 3,80
6
(60 MHz, d -dimethylsulfoxide) δ 4.94 (s, 2H, NH ), 6.97 (s,1H,
6
2
(s, 6H, 3- and 5-OMe), 5.52 (br s, 2H, 2-NH ), 5.66 (br s, 2H,
2
C=CH), 7.25-7.71 (m, 4H, aromatic protons). Condensation of 2c
with guanidine was carried out by Method C to obtain 3c as an
off-white powder (51%), mp 226° dec; ms: theory for
4-NH ), 6.68 (s, 2H, 2- and 6-H), 6.81 (s, 1H, C=CH), 10.90
2
(s, 1H, pyrrole NH).
Anal. Calcd for C
H N O •0.1H O: C, 56.81; H, 5.47; N,
15 17 5 3 2
C
H N Cl m/z 260, 261; found 260, 261; ir (KBr) ν 3470,
12 10 5
22.08. Found: C, 56.79; H, 5.28; N, 22.02.
3390, 3320, 3185, 1610, 1590, 1570, 1540, 1420, 1090, 960
-1 1
cm ; H-nmr (500 MHz, d -dimethylsulfoxide): δ 5.56 (br s, 2H,
6
2,4-Diamino-5-benzyl-7H-pyrrolo[2,3-d]pyrimidine (1e).
2-NH ), 5.60 (br s, 2H, 4-NH ), 6.83 (s, 1H, C=CH), 7.33 (d, 4H,
2
2
1-Hydroxy-4-phenylpropan-2-one (2e) was obtained by
Method A except that the product was purified by silica gel chro-
matography using 7:3 hexanes-ethyl acetate as the eluent; yellow
aromatic protons), 11.00 (br s, 1H pyrrole NH).
Anal. Calcd. for C N Cl: C, 55.50; H, 3.88; N, 26.97; Cl,
H
12 10
5
13.66. Found: C, 55.53; H, 4.09; N, 26.68; Cl, 13.74.
1
oil (73%); H-nmr (60 MHz, deuteriochloroform): δ 3.65 (s, 2H,
2,4-Diamino-5-(3,4-chlorophenyl)-7H-pyrrolo[2,3-d]pyrimidine
(1c).
benzylic CH ), 4.80 (br m, 1H, CH OH), 4.22 (s, 2H, CH OH),
2
2
2
7.11-7.54 (m, 5H, aromatic protons). The ketone was condensed
directly with malononitrile by Method B to obtain 3e as a brown
3,4-Dichloro-α-hydroxyacetophenone (2c) was obtained from
3,4-dichlorobenzaldehyde by Method A, and was used directly in
the next step; tan crystals (34%), mp 115-116° (from 9:1
1
solid (43%), mp 105-106°; H-nmr (200 MHz, d -dimethylsul-
6
foxide) δ 3.31 (s, 2H, 4-NH ), 3.61 (s, 2H, benzylic CH ), 6.76
2
2
1
(s, 1H, C=CH), 7.21-7.32 (m, 5H, aromatic protons).
Condensation of 3e with guanidine was carried out by Method C;
colorless crystals (37%) from methanol, mp 266-267°; ms: the-
hexanes-ethyl acetate); H-nmr (200 MHz, deuteriochloroform):
δ 4.77 (s, 2H, CH OH), 5.22 (broad s, 1H, CH OH), 7.87 (s, 1H,
2
2
aromatic proton), 8.12 (m, 3H, aromatic protons). 2-Amino-4-
(3,4-dichlorophenyl)furan-3-carbonitrile (3c) was obtained from
2c by Method B except that the product was chromatographed
with ethyl acetate. The resultant yellow powder (41%, mp
+
+
ory for C
H N m/z 239 (M ), 240 (MH ), found 239, 240; ir
13 13 5
(potassium bromide) ν 3490, 3450, 3390, 3305, 3130, 1610,
-1
1
1575, 1545, 1480, 1420, 1400, 1320, 1100, 700 cm ; H-nmr

1202
A. Rosowsky, H. Fu and S. F. Queener
Vol. 38
[5] E. C. Taylor, H. H. Patel, and J.-G. Jun, J. Org. Chem., 60,
6684 (1995).
(200 MHz, d -dimethylsulfoxide): δ 4.00 (s, 2H, benzylic CH ),
6
2
5.34 (br s, 2H, 2-NH ), 5.70 (br s, 2H, 4-NH ), 6.32 (s, 1H,
2
2
[6] E. C. Taylor and Z. Mao, J. Org. Chem., 61, 7973 (1996).
[7] T. Matsumoto, M. Ohishi, and S. Inoue, J. Org. Chem., 50,
C=CH), 7.12-7.31 (m, 5H, aromatic protons), 10.43 (s, 1H,
pyrrole NH).
603 (1985).
[8] F. Wahid, C. Monneret, and D. Dauzonne, Chem. Pharm.
Bull., 47, 156 (1999).
Anal. Calcd. for C
Found: C, 65.00; H, 5.49; N, 29.49.
H N : C, 65.26; H, 5.48; N, 29.27.
13 13 5
[9] L. F. Kuyper, J. M. Garvey, D. P. Baccanari, J. N. Champness,
D. K. Stammers, and C. R. Beddell, Bioorgan. Med. Chem., 4, 593
(1996).
[10] R. L. Then, P. G. Hartman, I. Kompis, and D. Santi, Adv.
Exptl. Biol. Med., 338, 533 (1993).
[11] R. L. Then, P. G. Hartman, I. Kompis, M. Stephan-Güldner,
and K. Stöckel, Drugs Future, 19, 446 (1994).
[12] M. C. Broughton and S. F. Queener, Antimicrob. Agents
Chemother., 35, 1348 (1991).
REFERENCES AND NOTES
[1] For a review of this major clinical problem in AIDS
management, see: M. E. Klepser and T. B. Klepser, Drugs, 53, 40 (1997).
[2] For a recent listing of previous papers from our group and
others, see: A. Rosowsky, H. Fu, and S. F. Queener, J. Heterocyclic
Chem. 37, 921 (2000).
[3] A. Gangjee, F. Mavandadi, S. F. Queener, and J. J. McGuire,
J. Med. Chem., 38, 1422 (1995).
[13] L-C. Chio and S. F. Queener, Antimicrob. Agents Chemother.,
37, 1914 (1993).
[4] A. Gangjee, F. Mavandadi, and S. F. Queener, J. Med. Chem.,
40, 1173 (1997).