Synthesis of the lipophilic antifolate piritrexim via a palladium(0)-catalyzed cross-coupling reaction

Article abstract of DOI:10.1021/jo040268z

(Chemical Equation Presented) A regiospecific and convergent route the lipophilic antifolate piritrexim (PTX) is described in which a key step is a Pd(0)-catalyzed cross-coupling reaction between 2-amino-3-cyano-4-methyl-5- bromopyridine and 2,5-dimethoxy

Full text of DOI:10.1021/jo040268z

Synthesis of the Lipophilic Antifolate Piritrexim via a
Palladium(0)-Catalyzed Cross-Coupling Reaction
David C. M. Chan and Andre Rosowsky*
Dana-Farber Cancer Institute and Department of Biological Chemistry and Molecular Pharmacology,
Harvard Medical School, Boston, Massachusetts 02115
andre_rosowsky@dfci.harvard.edu
Received September 21, 2004
A regiospecific and convergent route the lipophilic antifolate piritrexim (PTX) is described in which
a key step is a Pd(0)-catalyzed cross-coupling reaction between 2-amino-3-cyano-4-methyl-5-
bromopyridine and 2,5-dimethoxybenzylzinc chloride to form 2-amino-4-methyl-5-(2,5-dimethoxy-
benzyl)nicotinonitrile. To complete the synthesis, the amino group is replaced by a more reactive
bromine atom via nonaqueous diazotization with tert-butyl nitrite, and the resultant bromo nitrile
is cyclized with guanidine.
Piritrexim (1, PTX) is a nonclassical antifolate which
was first synthesized by Grivsky and co-workers1a,b as a
lipophilic analogue of the anticancer drug methotrexate
(MTX) (Figure 1). Extensive in vitro and in vivo preclini-
cal studies were subsequently carried out by a number
of investigators with a view to exploring the mechanism
and scope of action of 1 relative to MTX.2a-g Phase I3a-c
as well as single- and multi-agent Phase II clinical trials
have been performed with 1 in patients with several
types of cancer including soft tissue sarcoma,4 malignant
melanoma,5a,b and carcinomas of the head and neck,6a-c
lung,7 breast,8 bladder,9a-c and brain.10 Other diseases
against which 1 has been used clinically include
FIGURE 1. Structures of piritrexim (PTX, 1) and metho-
trexate (MTX).
* To whom correspondence should be addressed. Phone: 617-632-
3117. Fax: 617-632-2410.
psoriasis11a,b and AIDS-related opportunistic parasitic
infections such as Pneumocystis carinii pneumonia.12,13
For this reason, new and/or improved synthetic routes
to 1 and its congeners continue to be of interest.
(1) (a) Grivsky E. M.; Lee S.; Sigel C. W.; Duch D. S.; Nichol, C. A.
J. Med. Chem. 1980, 23, 327. (b) Grivsky E. M. US Patent 4,959,474,
September 25, 1990.
(2) (a) Duch, D. S.; Edelstein, M. P.; Bowers, S. W.; Nichol, C. A.
Cancer Res. 1982, 42, 3987. (b) Sedwick, W. D.; Hamrell, M.; Brown,
O. E.; Laszlo, J. Mol. Pharmacol. 1982, 22, 766. (c) Taylor, J. W.;
Slowiaczek, P.; Friedlander, M. L.; Tattersall, M. H. N. Cancer Res.
1985, 45, 978. (d) Richards, R. G.; Brown, O. E.; Gillison, M. L.;
Sedwick, W. D. Mol. Pharmacol. 1986, 30, 651. (e) Sigel, C. W.;
Macklin, A. W.; Woolley, J. L., Jr.; Johnson, N. W.; Collier, M. A.; Blum,
M. R.; Clendennin, N. J.; Everitt, B. J. M.; Grebe, G.; Mackars, A.;
Foss, R. G.; Duch, D. S.; Bowers, S. W.; Nichol, V. A. NCI Monogr.
1987, 5, 111. (f) Sharma, R. C.; Assaraf, Y. G.; Schimke, R. T. Cancer
Res. 1991, 51, 2949. (g) Assaraf, Y. G.; Slotky, J. I. J. Biol. Chem. 1993,
268, 4556.
(3) (a) Weiss, G. R.; Sarosy, G. A.; Shenkenberg, T. D.; Williams,
T.; Clendeninn, N. J.; Von Hoff, D. D.; Woolley, J. L.; Liao, S. H.; Blum,
M. R. Eur. J. Cancer 1989, 25, 1867. (b) Feun L. G.; Savaraj, N.;
Benedetto, P.; Hanlon, J.; Sridhar, K. S.; Collier, M.; Richman, S.; Liao,
S. H.; Clendeninn, N. J. J. Natl. Cancer Inst. 1991, 83, 51. (c) Adamson,
P. C.; Balis, F. M.; Miser, J.; Arndt, C.; Wells, R. J.; Gillespie, A.;
Aronson, L, Penta, J. S.; Clendeninn, N. J.; Poplack, D. G. Cancer Res.
1992, 52, 521.
10.1021/jo040268z CCC: $30.25 © 2005 American Chemical Society
Published on Web 01/21/2005
1364
J. Org. Chem. 2005, 70, 1364-1368

Synthesis of the Lipophilic Antifolate Piritrexim
Three synthetic routes to 1 have been reported in the
literature,1a,14,15 all of which suffer from various limita-
tions. In the Grivsky synthesis,1a 2,4,6-triaminopyrimi-
dine is condensed with ethyl 2-(2,5-dimethoxybenzyl)-3-
oxobutanoate to afford the dicyclic lactam 2, whereupon
chlorination (SOCl2/DMF) followed by catalytic dechlo-
rination (H2/Pd-C/KOH) yields 3 and 1, respectively.
This route suffers from the capricious nature of the
chlorination step and from fact that the diaminohetero-
cyclic intermediates 2 and 3 are difficult to purify on a
large scale because of their poor solubility in organic
solvents other than DMF or DMSO.
and 9 (HBr/AcOH). After removal of the unwanted isomer
9, bromo nitrile 8 was cyclized with guanidine. The lack
of regiospecificity in the DEMA reaction necessitated
careful optimization of this step by HPLC and ultimately
separation of 8 and 9 by flash chromatography and
crystallization.
In the Hill synthesis,14 which was designed to avoid
some of the foregoing problems and also provide access
to[14C]PTX,Knoevenagelcondensationof4-(2,5-dimethoxy-
phenyl)-2-butanone (4) with malononitrile, followed by
reaction of the resulting ylidenemalononitrile 5 with
diethoxymethyl acetate (DEMA), yielded the isomeric
acetals 6 and 7. Without being separated, the latter were
then converted directly to the 2-bromonicotinonitriles 8
In the Troschu¨tz synthesis,15 4 was subjected to a
Vilsmeier reaction (POCl3/DMF) to obtain an approxi-
mately 3:1 mixture of 10a and 10b, the crude mixture
was condensed directly with cyanoacetamide to obtain
4-methylpyridone-3-carbonitrile (11), and after extensive
purification via a combination of chromatography and
recrystallization, 11 was subjected to chlorination (SOCl2/
DMF) followed by treatment with guanidine to form 12
and 1, respectively. However, the reaction of 10a/10b
with cyanoacetamide was said to afford more than one
product,15 one of which we surmised might be 13. Thus
a possible drawback of this route is that unless 11 and
12 are entirely free of their respective regioisomers 13
and 14 the final product could be a difficulty separable
mixture of PTX and the corresponding 7-methyl isomer,
i.e., “isoPTX”.16
(4) Schiesel, J. D.; Carabasi, M.; Magill, G.; Casper, E.; Cheng, E.;
Marks. L.; Feyzi, J.; Clendeninn, N. J.; Smalley, R. V. Invest. New
Drugs 1992, 10, 97.
(5) (a) Feun, L. G.; Gonzalez, R.; Savaraj, N.; Hanlon, J.; Collier,
M.; Robinson, W. A.; Clendeninn, N. J. Clin. Oncol. 1991, 9, 464. (b)
Feun, L. G.; Robinson, W. A.; Savaraj, N.; Gonzalez, R.; Liebmann,
A.; Offenhauser, K.; Clendeninn, N. J. Am. J. Clin. Oncol. 1995, 18,
488.
(6) (a) Vokes, E. E.; Dimery, I. W.; Jacobs, C. D.; Karp, D.; Molina,
A.; Collier, M. A.; Eble, M. L.; Clendeninn, N. J. Cancer 1991, 67, 2253.
(b) Vokes, E. E.; Haraf, D. J.; McEvilly, J. M.; Mick, R.; Kozloff, M. F.;
Goldman, M. D.; Moran, W. J.; Clendeninn, J. N.; Collier, M. A.;
Weichselbaum, R. R.; Panje, W. R. Ann. Oncol. 1992, 3, 79. (c) Uen
W.-C.; Huang, A. T.; Mennel, R.; Jones, S. E.; Spaulding, M. B.; Killion,
K.; Havlin, K.; Keegan, P.; Clendeninn, N. J. Cancer 1992, 69, 1008.
(d) Degardin, M.; Domenge, C.; Cappelaere, P.; Luboinski, B.; Navar-
rete, M. S.; Scott, J. E.; Lefebvre, J. L Eur. J. Cancer 1994, 30, 1044.
(7) Kris, M. G.; Gralla, R. J.; Burke, M. T.; Berkowitz, L. D.; Marks,
L. D.; Kelsen, D. P.; Heelan, R. T. Cancer Treat. Rep. 1987, 71, 763.
(8) de Vries, E. G.; Gietema, J. A.; Workman, P.; Scott, J. E.;
Crawshaw, A.; Dobbs, H. J.; Dennis, I.; Mulder, N. H.; Sleijfer, D. T.;
Willemse, P. H. Br. J. Cancer 1993, 68, 641.
(9) (a) De Wit, R.; Kaye, S. B.; Roberts, J. T.; Stoter, G.; Scott, J.;
Verweij, J. Br. J. Cancer 1993, 67, 388. (b) Khorsand, M.; Lange, J.;
Feun, L.; Clendeninn, N. J.; Collier, M.; Wilding, G. Invest. New Drugs
1997, 15, 157. (c) Roth, B. J.; Manola, J.; Dreicer, R.; Graham, D.;
Wilding, G. Invest. New Drugs 2002, 20, 425.
(10) Bleehen, N. M.; Newman, H. V.; Rampling, R. P.; Ramsay, J.
R.; Roberts, J. T.; Bedford, P.; Nethersell, A. B. Br. J. Cancer 1995,
72, 766.
(11) (a) Guzzo, C.; Benik, K.; Lazarus, G.; Johnson, J.; Weinstein,
G. Arch. Dermatol. 1991, 127, 511. (b) Perkins, W.; Williams, R. E.;
Vestey, J. P.; Tidman, M. J.; Layton, A. M.; Cunliffe, W. J.; Saihan, E.
M.; Klaber, M. R.; Manna, V. K.; Baker, H.; Mackie, R. M. Br. J.
Dermatol. 1993, 129, 584.
(12) 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, 361A.
(13) For a recent compilation of papers on the use of lipophilic
antifolates, including analogues structurally related to PTX, in the
prophylaxis and treatment of these opportunistic illnesses in HIV-
infected patients, see: Rosowsky, A.; Fu, H.; Chan, D. C. M.; Queener,
S. F. J. Med. Chem. 2004, 47, 2475.
(14) Hill, J. A.; Wisowaty, J. C.; Darnofall, M. E. J. Labelled Compds.
Radiopharm. 1993, 33, 1119.
(15) Troschu¨tz, R.; Zink, M.; Gnibl, R. J. Heterocycl. Chem. 1999,
36, 703.
An inefficient feature of all the previously described
routes1,14,15 to 1 is that they are nonconvergent. That is,
for the synthesis of every 2,4-diamino-5-methyl-6-(sub-
stituted benzyl)pyrido[2,3-d]pyrimidines, a different ethyl
2-(substituted benzyl)-3-oxobutanoate, 4-(substituted phen-
yl)-3-butanone, or 3-(substituted phenyl)-1-propanone
(16) Troschu¨tz, R.; Zink, M.; Dennstedt, T. Arch Pharm. 1995, 328,
535.
J. Org. Chem, Vol. 70, No. 4, 2005 1365

Chan and Rosowsky
SCHEME 1a
a Reagents: (a) CH2(CN)2/piperidine/AcOH; (b) NH3/MeOH; (c) NIS/DMF; (d) NBS/DMF; (e) 2,5-(MeO)2C6H3CH2ZnCl/PdCl2(dppf)‚CH2Cl/
THF; (f) SbBr3/t-BuONO/CH2Br2; (g) H2NC(dNH)NH2/C5H5N.
must be used. It therefore seemed of interest to devise a
strategy for the synthesis of 1, and potentially its
congeners,17 that would be both regiospecific and conver-
gent. To this end we took advantage of a Pd(0)-catalyzed
organozinc coupling reaction18 we had used earlier to
prepare 2,4-diamino-6-(substituted benzyl)quinazolines19
and 2,4-diamino-6-(substituted benzyl)pyrido[2,3-d]py-
rimidines.20 In the latter instance, the poor solubility of
unprotected 2,4-diaminopyrido[2,3-d]pyrimidines neces-
sitated the use of 2,4-bis(pivaloylamino) derivatives in
the coupling reaction. Accordingly, to avoid the extra
protection and deprotection steps that would be needed
with a preformed 2,4-diamino-6-halo-5-methylpyrido[2,3-
d]pyrimidine we decided to perform the Rieke reaction
at an early stage of the sequence where solubility would
be less of a problem.
As shown in Scheme 1, commercially available 4,4-
dimethoxy-2-butanone was condensed with malononitrile
to form the ylidenemalononitrile 1521 (cf. also refs 22-
24), and the latter was cyclized to 2-amino-4-methyl-3-
carbonitrile (16) with ammonia in methanol. Halogena-
tion of 16 with N-iodosuccinimide or N-bromosuccinimide
in anhydrous DMF23 afforded the 5-iodo and 5-bromo
derivatives 17 and 18, respectively. As expected, the pair
of doublets for C5 (δ 6.56) and C6 (δ 8.07) protons in the
1H NMR spectrum of 16 was replaced by a singlet at δ
8.42 in the case of 17 and δ 8.24 in the case of 18. The
larger downfield shift of the C6 proton in the iodide 17
(∆ ) -0.35 ppm) than in the bromide 18 (∆ ) -0.17 ppm)
relative to 16 was qualitatively consistent with the
chemical shifts of the corresponding singlet in 3-iodo-4-
methylpyridine (δ 8.9)25 and 3-bromo-4-methylpyridine
(δ 8.5),26 respectively.
Reaction of 17 in dry THF with 2,5-dimethoxybenzyl-
zinc chloride and 1,1′-bis[(diphenylphosphino)ferrocene]-
dichloropalladium(II)‚CH2Cl2 [PdCl2(dppf)‚CH2Cl2] under
an inert atmosphere afforded the desired 5-(2,5-dimethox-
ybenzyl) derivative 19. Interestingly, the reaction of 18
appeared to be much more efficient than that of 17. Thus,
when the reaction of the bromide was worked up after
just 4 h in refluxing THF the yield of 19 was 59%,
whereas in the case of the iodide the yield was only 36%
even after 24 h. Moreover, the 1H NMR spectrum of crude
19 obtained via 17 indicated the likely presence of some
16, presumably formed by reductive deiodination. It thus
appears that despite the better yields sometimes ob-
served with aryl iodides in Pd(0)-catalyzed coupling
reactions, the bromide was clearly superior to the iodide
in this case.
While we had hoped that treatment of 19 with guani-
dine would yield 1, we were disappointed to find that the
desired ring closure did not occur despite a number of
attempts to vary the temperature and/or duration of the
reaction. We suspect that this was probably due to a
combination of two factors. The first is that the nucleo-
philic reactivity of the NH2 group is diminished because
it is simultaneously flanked by a ring nitrogen and
CtN group. The other is that the electrophilic reactivity
of the nitrile may be compromised by steric hindrance
by the methyl group at C4 and the additional buttressing
effect of the arylmethyl group at C5. Faced with this
hurdle, we decided to replace the NH2 group by Br via a
Sandmeyer-type reaction in the hope that the bromo
nitrile would be more reactive. Accordingly, nonaqueous
diazotization of 19 with t-BuONO and SbBr3 in CH2Br2
via a slight modification of the excellent method of Robins
and Uznanski27 led to the required bromo nitrile inter-
mediate 20 (43%), which on further treatment with
guanidine was successfully cyclized to 1 (38%).
(17) See the following paper for a large group of piritrexim analogues
synthesized via the Grivsky route: Zhang, Y.; Zhou, J.; Li, R. Chin. J.
Med. Chem. 1993, 3, 85.
(18) Zhu, L.; Wehmeyer, R. M.; Rieke, R. D. J. Org. Chem. 1991,
56, 1445.
(19) Rosowsky, A.; Chen, H. J. Org. Chem. 2001, 66, 7522.
(20) Rosowsky, A.; Chen, H.; Fu H.; Queener, S. F. Bioorg. Med.
Chem. 2003, 11, 59.
(21) This reaction is reported in a patent22 to yield a mixture of 15
and a second compound identified as 4-methoxy-3-methyl-1,3-butadi-
enylidenemalonitrile, presumably formed by loss of a molecule of
MeOH during distillation. The products were used without separation,
but the reactions disclosed in the patent did not include cyclization
with ammonia. Interestingly, the 1H NMR spectrum of our double-
distilled product showed no evidence of an enol ether byproduct. The
diethyl acetal analogue of 15 is known23 and can also be used to make
16. However, unlike 15, the diethyl acetal is not commercially available
and its synthesis from 1,1-dicyano-2-methylpropene (acetonyliden-
emalonitrile), triethyl orthoformate, acetic anhydride, and ZnCl2
catalyst laboratory adds an extra step. Compound 16 has also been
synthesized previously by a multistep route from 2-amino-4-methylpy-
ridine,24 but the method described here is superior.
In summary, while we did not optimize the yields of
the various steps in Scheme 1, we believe that our route
to 1 is a novel alternative to other approaches used in
the past because it involves fewer overall steps, avoids
capricious Vilsmeier-type reactions, and is by its very
nature both regiospecific and convergent. Thus, an inter-
(22) Gupton, B. F. PCT International Publication WO 0230901, April
18, 2002.
(23) Henrie, R. N.; Peake, C. J.; Cullen, T. G.; Chaguturu, M. L.;
Ray, P. S.; Bennett, B. D. U.S. Patent 5,547,954, August 20, 1996.
(24) Dunn, A. D.; Norrie, R. J. Prakt. Chem. 1996, 338, 663.
(25) Wasicak, J. T.; Garvey, D. S.; Holladay, M. W.; Lin, N.-H.;
Ryther, K. B. U.S. Patent 5,733,912, March 31, 1998.
(26) Mallet, M.; Que´guiner, G. Tetrahedron 1982, 38, 3035.
(27) Robins, M. J.; Uznanski, B. Can. J. Chem. 1981, 59, 2608.
1366 J. Org. Chem., Vol. 70, No. 4, 2005

Synthesis of the Lipophilic Antifolate Piritrexim
259.05, found 259.80 (M + 1); IR (KBr) ν 3400, 3320, 3320,
2200 (CtN), 1640, 1570, 1470, 1450, 1360, 1330, 1260, 830,
770 cm-1; 1H NMR (CDCl3) δ 2.55 (s, 3H, CH3), 5.20 (br s, 2H,
NH2), 8.42 (s, 1H, H6). Anal. Calcd for C7H6IN2: C, 32.46; H,
2.33 N, 16.22. Found: C, 32.34; 2.43; N, 15.99.
esting feature of this strategy is that is has the potential
to be used in the construction of libraries of piritrexim
analogues by parallel synthesis.
Experimental Section
2-Amino-3-cyano-4-methyl-5-bromopyridine (18). A so-
lution of NBS (1.8 g, 10 mmol) in dry DMF (10 mL) was added
dropwise to a solution of 16 (1.3 g, 9.8 mmol) in dry DMF (10
mL) at 0 °C. The reaction mixture was warmed to room
temperature, stirred overnight, and poured slowly into 3 M
NaOH (150 mL). After dilution with H2O (150 mL), the solid
was collected and dried to obtain a product (1.9 g, 92%) that
was pure enough to be used directly in the next step. A small
sample recrystallized from cyclohexane gave fine beige
needles: mp 198-200 °C; MS calcd m/z 212.05 (based on
natural distribution of Br isotopes), found 211.85, 213.85 (M
+ 1); IR (KBr) ν 3520, 3360, 3200, 2220 (CtN), 1640, 1570,
Infrared (IR) spectra using KBr disks or NaCl plates were
obtained on a double-beam recording spectrophotometer. Mass
spectra (MS) were provided by the Molecular Biology Core
Facility of the Dana-Farber Cancer Institute. Proton magnetic
resonance (1H NMR) spectra were recorded at 200 MHz. Each
peak is denoted as a singlet (s), broad singlet (br s), doublet
(dd), or triplet (t). The 13C NMR spectrum of 1 was recorded
at 125 MHz field strength, with peak assignments made by
means of the attached proton test.28 TLC analyses were on
glass plates coated with silica gel containing a fluorescent dye,
and spots were visualized by illumination at 254 nm. Column
chromatography was on flash-grade silica gel (40 µm particle
size). Chemicals were of the best grade available from com-
mercial suppliers, and were used without additional purifica-
tion. Elemental analyses were performed by a commercial
laboratory.
1
1480, 1390 1250, 840, 770 cm-1; H NMR (DMSO-d6) δ 2.36
(s, 3H, CH3), 6.99 (br s, 2H, NH2, exchangeable with D2O),
1
8.20 (s, 1H, H6); H NMR (CDCl3) δ 2.52, (s, 3H, CH3), 5.19
(br s, 2H, NH2), 8.24 (s, 1H, H6). Anal. Calcd for C7H6BrN2:
C, 39.65; H, 2.85; N, 19.82; Br, 37.68. Found: C, 39.56; H, 2.65;
N, 19.58; Br, 37.57.
4,4-Dicyano-3-methyl-3-butenal Dimethyl Acetal (15).
Malononitrile (25 g, 0.38 mmol) was added in portions over
20 min to a stirred solution of 4,4-dimethoxy-2-butanone (50
g, 0.38 mol) in toluene (150 mL) containing acetic acid (2.2
mL, 0.038 mmol) and piperidine (3.8 mL, 0.038 mol). Stirring
was continued at room temperature overnight, and the result-
ing dark red solution was washed with H2O (50 mL). The
organic layer was evaporated and the residue double-distilled
to obtain a colorless oil (52 g, 76%).29 A center-cut (77-79 °C/
0.03 mm) was set aside for spectroscopic and microchemical
analysis, and the remainder was used directly in the next
step: IR (NaCl) ν 2930, 1840, 2240 (CtN), 1680, 1600, 1440,
2-Amino-3-cyano-4-methyl-5-(2,5-dimethoxybenzyl)py-
ridine (19). Method A. A mixture of the bromide 18 (1.6 g,
7.5 mmol) and PdCl2(dppf)‚CH2Cl2 (308 mg, 0.75 mmol) in dry
THF (10 mL) was stirred at room temperature for 5 min in a
thoroughlydriedthree-neckedflaskunderargon.2,5-Dimethoxy-
benzylzinc chloride (30 mL of 0.5 M solution in THF, 2 molar
equiv) was then added via a transfer needle, the mixture
refluxed for 4 h, and the solvent evaporated under reduced
pressure. The black solid was applied to the top of a flash silica
gel column, which was eluted with CH2Cl2 and then 8:1 CH2-
Cl2-EtOAc to obtain a reddish solid (1.3 g, 59%). Recrystal-
lization from i-PrOH gave colorless needles: mp 173-174 °C;
MS calcd m/z 283.33, found 283.96 (M + 1); IR (KBr) ν 3480,
3310, 3160, 2980, 2220 (CtN), 1640, 1480, 1250, 1220, 1210,
1
1360, 1340, 1120, 1070 cm-1; H NMR (CDCl3) δ 2.35 (s, 3H,
CH3), 2.87 (d, 1H, J ) 5.2 Hz, CH2), 3.39 (s, 6H, OCH3), 4.57
(t, 1H, J ) 5.2 Hz, CH). Anal. Calcd for C9H12N2O2: C, 59.99;
H, 6.71; N, 15.55. Found: C, 59.69; H, 6.72; N, 15.82.
1
1050, 1050, 1020, 800, 710 cm-1; H NMR (DMSO-d6) δ 2.21
(s, 3H, CH3), 3.60 (s, 3H, OCH3), 3.69 (s, 5H, OCH3 and CH2),
6.44 (d, 1H, J ) 2.8 Hz, H6′), 6.55 (br s, 2H, NH2), 6.69 (dd,
1H, J ) 2.8 Hz, J ) 8.8 Hz, H4′), 6.85 (d, 1H, J ) 9.0 Hz, H3′),
7.85 (s, 1H, pyridine H6); 13C NMR (DMSO-d6) δ 17.3 (CH3),
29.2 (CH2), 55.2 (CH3), 55.7 (CH3), 90.4 (C), 111.2 (CH), 111.5
(CH), 115.8 (CH), 116.5 (C), 122.3 (C), 28.7 (C), 150.6 (C), 151.0
(C), 152.9 (C), 153.0 (C), 159.3 (CH). Anal. Calcd for C16H17N3O2‚
0.55H2O: C, 65.54; H, 6.22; N, 14.33. Found: C, 65.74; H, 6.22;
N, 14.04.
Method B. The same procedure using iodide 17 (1.3 g, 5
mmol), PdCl2(dppf)‚CH2Cl2 (410 mg, 0.5 mmol) in dry THF (20
mL), and 2,5-dimethoxybenzylzinc chloride (50 mL of 0.05 M
solution in THF, 5 molar equiv) afforded 21 (510 mg, 36%).
Microanalytical and spectroscopic data indicated that this
product and the one obtained via method A were the same.
2-Amino-3-cyano-4-methylpyridine (16). Ammonia gas
was bubbled without external cooling through a solution of
15 (8.0 g, 44 mmol) in MeOH (300 mL), and the deep-red
solution was stirred at room temperature overnight. The
solvent was evaporated, and the residue partitioned between
1 N HCl (300 mL) and EtOAc (300 mL). The aqueous layer
was added carefully (caution: foaming) to ice-cold concd
NaHCO3 (600 mL), and the precipitate was filtered to obtain
a beige solid (3.0 g, 33%). Recrystallization of a portion of the
solid from EtOAc-cyclohexane gave yellow needles: mp 154-
155 °C (lit.24 mp 150-152.5 °C); MS calcd m/z 134.06, found
133.94 (M + 1); IR (KBr) ν 3400, 3340, 3160, 2220 (CtC), 1650,
1
1560, 1480, 1370, 1330, 1260, 1250, 800, 770 cm-1; H NMR
(CDCl3) δ 2.44 (s, 3H, CH3), 5.18 (bs s, 2H, NH2), 6.56 (d, 1H,
J ) 5.2 Hz, H5), 8.07 (d, 1H, J ) 5.2 Hz, H6). Anal. Calcd for
C7H7N3: C, 63.14 H, 5.30; N, 31.56. Found: C, 63.13; H, 5.32;
N, 31.56.
2-Amino-3-cyano-4-methyl-5-iodopyridine (17). A solu-
tion of N-iodosuccinimide (15 g, 66 mmol) in dry DMF (125
mL) was added dropwise to a solution of 16 (8.3 g, 63 mmol)
in DMF (125 mL) at 0 °C. When addition was complete, the
solution was allowed to come to room temperature and stirred
overnight. The volume was reduced in half by rotary evapora-
tion (vacuum pump), the dark-brown solution poured slowly
into 3 M NaOH (1 L), and the precipitate collected. Silica gel
flash chromatography (9:1 CH2Cl2-EtOAc) furnished a beige
solid (5.6 g, 34%): mp 203-204 °C (i-PrOH); MS calcd m/z
2-Bromo-3-cyano-4-methyl-5-(2,5-dimethoxybenzyl)py-
ridine (20). A solution of SbBr3 (300 mg, 0.83 mmol) in CH2-
Br2 (2 mL) was added dropwise to a stirred solution of 19 (150
mg, 0.53 mmol) in CH2Br2 (10 mL) at 0 °C under dry argon.
When addition was complete, the brown solution was treated
dropwise with freshly distilled t-BuONO (1 mL). The reaction
mixture was kept at 4-10 °C (internal) for 6 h, then brought
back down to 4 °C, and poured into ice-cold concentrated
NaHCO3 solution. The product was extracted three times into
CH2Cl2, and the combined organic layers were dried (Na2SO4)
and evaporated to dryness. Column chromatography on flash-
grade silica gel using CH2Cl2 as the eluent gave a pale-orange
solid (80 mg, 43%). Recrystallization from cyclohexane gave
an off-white solid: mp 113-115 °C (lit.14 mp 108-110 °C); MS
calcd m/z 347.21 (based on natural distribution of Br isotopes),
found 346.94, 348.93 (M + 1); IR (KBr) ν 2980, 2850, 2400,
(28) Cobas, J. C.; Sardina, F. J. Concept. Magn. Reson. 2003, 19A,
80.
(29) The once-distilled product was pale-greenish in color. The twice-
distilled product was initially colorless, became straw-colored on
standing overnight at room temperature, and turned orange after being
kept at 4 °C under argon for a week. Freshly distilled product should
be used.
1590, 1560, 1510, 1430, 1230, 1050, 1020, 880, 800, 720 cm-1
;
1H NMR (DMSO-d6) δ 2.46 (s, 3H, CH3), 3.64 (s, 3H, OCH3),
3.70 (s, 3H, OCH3), 3.91 (s, 2H, CH2), 6.61 (d, 1H, J ) 2.8 Hz,
J. Org. Chem, Vol. 70, No. 4, 2005 1367

Chan and Rosowsky
H6′), 6.76 (dd, 1H, J ) 2.8 Hz, 8.8 Hz, H4′), 6.90 (d, 1H, J ) 8.8
Hz, H3′), 8.25 (s, 1H, pyridine H6). Anal. Calcd for C16H15-
BrN2O2: C, 55.35; H, 4.35; N, 8.07. Found: C, 55.14; H, 4.15;
N, 7.95.
mp >310 °C); MS calcd m/z 326.37, found 326.04 (M + 1); IR
(KBr) ν 3480, 3340, 3120, 1650, 1590, 1550, 1500, 1500, 1460,
1230, 1050, 830 cm-1; 1H NMR (DMSO-d6) δ 2.52 (s, H, CH3),
3.59 (s, 3H, OCH3), 3.73 (s, 3H, OCH3), 3.87 (s, 2H, CH2), 6.15
(br s, 2H, NH2, exchangeable with D2O), 6.39 (d, 1H, J ) 3.0
Hz, H6′), 6.70 (dd, 1H, J ) 3.0 Hz, 8.8 Hz, H4′), 6.87 (d, 1H, J
) 3.0 Hz, 8.8 Hz, H3′, with an underlying br s, 2H, NH2,
exchangeable with D2O), 8.32 (s, 1H, pyridine H6); 13C NMR
(DMSO-d6) δ 17.9 (CH3), 30.3 (CH2), 55.2 (CH3), 55.8 (CH3),
105.4 (C), 111.0 (CH), 111.4 (CH), 115.9 (CH, 126.5 (C), 129.2
(C), 151.0 (C), 153.0 (C), 156.0 (CH), 161.3 (C), 161.6 (C), 164.0
(C). Anal. Calcd for C17H19N5O2‚H2O: C, 59.46; H, 6.16; N,
20.40. Found: C, 59.66; H, 6.04; N, 20.09.
2,4-Diamino-6-(2,5-dimethoxybenzyl)-5-methylpyrido-
[2,3-d]pyrimidine (PTX, 1). Sodium metal (46 mg, 2 mmol)
was dissolved in absolute MeOH (2 mL), the solution was
chilled, guanidine hydrochloride (191 mg, 2 mmol) was added,
the mixture was stirred for 15 min, the fine white precipitate
of NaCl was suction filtered, and the filtrate was evaporated
to dryness under reduced pressure. To the residue were then
added 20 (138 mg, 0.4 mmol) and dry pyridine (1.8 mL), and
the mixture was refluxed for 3 h. After being cooled to room
temperature, the solution was diluted with H2O (30 mL) and
cooled to 0 °C. The beige precipitate was collected by suction
filtration, and another crop was recovered from the filtrate by
extracting it several times with CH2Cl2 and evaporating the
pooled extracts to dryness. Column chromatography (flash-
grade silica gel, 95:5 CHCl3-MeOH) afforded 1 as a white solid
(50 mg, 38%): mp 273-274 °C (lit.1a,b mp 252-254 °C, lit.15,30
Acknowledgment. This work was generously sup-
ported by a grant (RO1-AI-29904) from the National
Institute of Allergy and Infectious Diseases, NIH. We
are grateful to Mr. Gregory Heffron, Department of
Biological Chemistry and Molecular Pharmacology,
Harvard Medical School, for his help in obtaining 13C
NMR spectra.
(30) The observed melting or decomposition temperature of con-
densed diaminopyrimidine derivatives can vary considerably depending
on the drying method, particle size, type of melting point apparatus,
and rate at which the sample is heated. The lower melting point of
our sample of 1 relative to that reported by Troschu¨tz and co-workers15
may be due to a small difference in the method of determination or to
the fact that our material, unlike theirs, was solvated with a molecule
of H2O.
Supporting Information Available: 13C NMR spectra for
piritrexim (1) and the synthetic intermediates 15-20. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JO040268Z
1368 J. Org. Chem., Vol. 70, No. 4, 2005