Large-scale synthesis of a pyrrolo[2,3-d]pyrimidine via Dakin-West reaction and Dimroth rearrangement

Article abstract of DOI:10.1021/op010041v

Pyrrolo[2,3-d]pyrimidines have been widely investigated as pharmaceutically active compounds. In this article, we present a short and efficient synthesis of 4-(3-chlorophenylamino)-5,6-dimethyl-7H-pyrrolo[2,3-d]pyrimidine starting from cheap alanine and malononitrile. To the best of our knowledge, the first application of a Dakin-West reaction on plant scale is demonstrated by elaboration of a modified procedure avoiding uncontrolled release of carbon dioxide. Pyrimidine ring formation is achieved in a simple one-pot reaction which is followed by an equally simple isomerization. The whole synthesis requires neither chromatographies nor extractions, no waste treatment and no special equipment, resulting in a remarkably ecological as well as economical process.

Full text of DOI:10.1021/op010041v

Organic Process Research & Development 2001, 5, 581−586
Large-Scale Synthesis of a Pyrrolo[2,3-d]pyrimidine via Dakin-West Reaction
and Dimroth Rearrangement
Reto W. Fischer* and Marian Misun
NoVartis Pharma AG, Chemical and Analytical DeVelopment, CH-4002 Basel, Switzerland
Abstract:
Pyrrolo[2,3-d]pyrimidines have been widely investigated as
pharmaceutically active compounds. In this article, we present
a short and efficient synthesis of 4-(3-chlorophenylamino)-5,6-
dimethyl-7H-pyrrolo[2,3-d]pyrimidine starting from cheap ala-
nine and malononitrile. To the best of our knowledge, the first
application of a Dakin-West reaction on plant scale is dem-
onstrated by elaboration of a modified procedure avoiding
Figure 1.
uncontrolled release of carbon dioxide. Pyrimidine ring forma-
tion is achieved in a simple one-pot reaction which is followed
by an equally simple isomerization. The whole synthesis requires
neither chromatographies nor extractions, no waste treatment
and no special equipment, resulting in a remarkably ecological
as well as economical process.
(2) The pyrimidine cyclisation step in boiling formic acid
raised safety concerns and resulted in dark coloration of the
product which made cumbersome purification procedures
necessary.
(3) The hydroxy- as well as the chloropyrimidine inter-
mediates had extremely poor solubility that resulted in dilute
reactions and a large excess of phosphorus oxychloride being
used for the chlorination step.
Introduction
Development Synthesis Strategy. Pyrrolo[2,3-d]pyri-
midine derivatives and their synthesis are very abundant in
the literature.2 In general, it is favorable to synthesize the
pyrrole ring first and to subsequently build up the pyrimidine
ring even though the opposite strategy has been applied as
well.3 For the synthesis of 1, we came to the conclusion that
2-amino-3-cyano-4,5-dimethylpyrrole 3 was the intermediate
of choice. Its simple synthesis from 3-amino-2-butanone and
malonodinitrile has already been described more than 30
years ago4 and it is sufficiently stable towards air oxidation
to allow for normal handling in a production facility. From
this pyrrole we intended to modify the pyrimidine ring
formation in such a way that no protection was needed and
that formic acid was eliminated. For a successful application
of this synthesis concept (Scheme 2), we therefore needed
to find solutions to the following two key issues:
(1) the preparation of 3-amino-2-butanone derivative 2
and
Derivatives of pyrrolo[2,3-d]pyrimidines have been ex-
tensively investigated as inhibitors of epidermal growth factor
receptor (EGF-R) protein tyrosine kinase, and their potential
as a treatment for proliferative diseases involving mitogenic
signaling from the EGF-R has been recognized.1 Optimiza-
tion of selectivity and the biological profile of such pyrrol-
opyrimidine derivatives by variation of the substitution
pattern, carried out at our research department, led to
compound CGP 59326 (1) (Figure 1) being promoted to
development status as an antitumor agent. To support further
biological profiling as well as to supply drug substance for
initial clinical trails, large amounts in the 100-kg range of
this heterocyclic compound had to be synthesized.
Research Synthesis. The existing research synthesis
(Scheme 1) was relatively straightforward and, with some
optimization, was regarded as suitable for scale-up in this
early development phase. However, for the following main
reasons, we aimed at the elaboration of an improved,
alternative synthesis to be applied already for the first GMP-
campaign in the pilot plant despite significant time con-
straints:
(2) an alternative procedure for the formation of the
pyrimidine ring
Results and Discussion
(1) The use of a benzyl-protecting group for the pyrrole
amine functionality was mandatory for the outlined synthetic
strategy; however, deprotection turned out to be technically
unfavorable because it required a large excess of aluminum
trichloride which subsequently had to be quenched and
disposed of.
Pyrrole Synthesis. Since 3-amino-2-butanone is not a
stable compound, literature suggests that it may be used in
its stable, N-acetylated form 2 for pyrrole condensation.5
(2) For reviews, see for example: (a) Melik, R. G.; Khachatryan, V. E.;
Gapoyan, A. S. Russ. Chem. ReV. 1985, 54 (3), 262. (b) Amarnath, V.;
Madhav, R. Synthesis 1974, 12, 837. (c) Roth, H. J.; Eger, K. Arch Pharm.
(Weinheim, Ger.) 1975, 308, 252.
(3) See, for example: (a) Sakamoto, T.; Satoh, Ch.; Kondo, Y.; Yamanaka, H.
Chem. Pharm. Bull. 1993, 41, 1 (1), 81. (b) Itoh, T.; Fujii, I.; Tomii, Y.;
Nishimura, H.; Ogura, H.; Mizuno, Y. Heterocycles 1986, 25 (4), 927.
(4) Gewald, K. Z. Chem. 1961, 1, 349.
(1) (a) Traxler, P.; Bold, G.; Lang, M.; Frei, J. PCT Int. Appl. WO 9807726,
1998. (b) Traxler, P.; Bold, G.; Brill, W.; Frei, J. PCT Int. Appl. WO
9702266, 1997. (c) Traxler, P.; Furet, P.; Mett, H.; Buchdunger, E.; Meyer,
T.; Lydon, N J. Pharm. Belg. 1997, 52(2), 88. (d) Traxler, P. Expert Opin.
Ther. Pat. 1997, 7(6), 571 and references therein.
10.1021/op010041v CCC: $20.00 © 2001 American Chemical Society and The Royal Society of Chemistry
Published on Web 10/12/2001
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
581

Scheme 1
ketone (Scheme 3).10 Most experimental procedures for
Dakin-West reactions reported in the literature work more
or less the same way: a mixture of amino acid, anhydride,
and pyridine (typically in the very empirical ratio of about
1:6:11) is heated to reflux, at which time, a relatively sudden,
strong evolution of carbon dioxide appears and fades again
soon thereafter. Any experimenting with this procedure or
conditions usually results in reduced or no yield of the desired
aminoketone product. To the best of our knowledge, there
never was a procedure reported that avoided this uncontrolled
formation of carbon dioxide; as a matter of fact, we did not
succeed in finding any published application of the Dakin-
West reaction on a technical scale.11
After some of our own investigations confirmed the
mechanism outlined above, we concluded that the following
kinetic interpretation would at least partly explain the
peculiarities of the Dakin-West reaction: as we see in
Scheme 3, ring-opening hydrolysis of acylated azlactone
formally requires one equivalent of water. Because there
certainly is no sufficient amount of water in a boiling reaction
mixture of excess anhydride and pyridine, this “water-
equivalent” has to come from the formation of acetic
anhydride from acetic acid. Acetic acid itself is formed in
the first two steps of the reaction sequence, namely N-
acylation and azlactone condensation. Therefore its initial
concentration in the reaction mixture is very low, and hence
azlactone hydrolysis is very slow. Once the mixture is heated
and N-acylation of alanine starts to pick up, acetic acid
concentration increases rapidly and therefore azlactone
hydrolysis is accelerated. The following decarboxylation is
virtually spontaneous. This results in the reaction having a
typical, self-accelerating kinetic, which is exactly what you
do not want to have in a plant, particularly not if equivalent
amounts of gas are formed during a reaction.
Scheme 2
However, 2 is not a commercially available compound.
R-Acylaminoketones of this type are easily accessible from
the corresponding amino acids and anhydrides, a reaction
that was initially described by Dakin and West in 19286 and
therefore is known as the Dakin-West reaction. Although
the Dakin-West reaction is by no means of general use to
any type of amino acid,7 its application to alanine to form
the desired aminoketone 2 is well documented.8 Since
(racemic) alanine is very cheap, this seemed to be an
attractive approach to the synthesis of pyrrole 3. However,
there is one major drawback to the Dakin-West reaction
regarding its scale-up to plant scale: it is accompanied by
an almost sudden and uncontrolled formation of one equiva-
lent of carbon dioxide, which is absolutely prohibitive on
technical scale. Our initial attempts to control formation of
carbon dioxide by the variation of reaction parameters such
as addition of components and control of temperature all
failed and in addition never produced yields which were at
least comparable to the standard protocol. Thus, we decided
to look more closely at the reaction mechanism.
Even though the Dakin-West reaction can hardly be
considered well-known, there is quite a lot of work published
about it, including some very detailed mechanistic investiga-
tions particularly by Steglich.9 It seems to be generally
accepted that the mechanism involves dehydrative formation
of an azlactone (oxazolinone) which is then C-acylated (in
equilibria with O-acylation), undergoes ring-opening hy-
drolysis, and then decarboxylates to form the acylamino
To test this hypothesis, we performed experiments where
we added various amounts of acetic acid to the initial mixture
of acetic anhydride and pyridinesan idea which surprisingly
never has been investigated in the literature. The results were
convincing: one equivalent of acetic acid made it possible
to run the reaction in an addition-controlled mode with
(5) (a) Shvedov, V. I.; Mezentseva, M. V.; Grinev, A. N. Khim. Geterotsikl.
Soedin. 1975, 9, 1217. (b) Wamhoff, H.; Wehling, B. Synthesis 1976, 51.
(6) Dakin, H. D.; West, R. J. Biol. Chem. 1928, 78, 91 and 745.
(7) For a review on the Dakin-West reaction, see: Buchanan, G. L. Chem.
Soc. ReV. 1988, 17, 91.
(8) (a) Wiley: R. H.; Borum, O. H. J. Am. Chem. Soc. 1948, 70, 2005. (b)
King, J. A.; McMillan, F. H. J. Am. Chem. Soc. 1955, 77, 2814.
(9) (a) Cleland, G. H.; Niemann, C. J. Am. Chem. Soc. 1949, 71, 841. (b)
Steglich, W.; Ho¨fle, G. Tetrahedron Lett. 1968, 13, 1619, (c) Steglich, W.;
Ho¨fle, G. Chem. Ber. 1969, 102, 883. (d) Steglich, W.; Ho¨fle, G. Chem.
Ber. 1969, 102, 899. (e) Steglich, W.; Ho¨fle, G.; Prox, A. Chem. Ber. 1972,
105, 1718. (f) Knorr, R. Chem. Ber. 1971, 104, 3633. (g) Allinger, N. L.;
Wang, G. L.; Dewhurst, B. B. J. Org. Chem. 1974, 39, 9 (12), 1730.
(10) This mechanism was originally proposed by Cleland and Niemann (see ref
9a); it only applies to the Dakin-West reaction of primary amino acids.
(11) For a typical lab-scale experiment in milligram or low-gram range, the
amount of carbon dioxide evolved is certainly too small to be of any concern.
Hence, there is no need for an alternative procedure to the Dakin-West
reaction on lab scale.
582
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development

Scheme 3
Scheme 4
carbon dioxide evolution being directly linked to alanine
addition. As an extra benefit, acetic anhydride and pyridine
excess could now be significantly reduced. This modified
procedure was run successfully and with well-controlled
carbon dioxide evolution on a 0.5 kmol scale in a 600-L
reactor, whereby acetic anhydride, acetic acid, and pyridine
were charged into the reactor and heated to ca. 120 °C, and
alanine (45 kg) was then added as a slurry in pyridine over
2-3 h.
In a second step, we tried to improve the reaction by
adapting the above-described findings to the modified
Dakin-West procedure with triethylamine/DMAP introduced
by Steglich in 1969.12 This method requires much smaller
amounts of anhydride and base and is therefore economically
more advantageous. Indeed, the above-described modification
of the classical Dakin-West reaction by the addition of acetic
acid to the starting mixture worked equally well when applied
to the triethylamine/DMAP procedure. Over 0.5 kmol alanine
(50 kg) was successfully converted to acetylaminobutanone
2 in a 400-L reactor in over 90% yield with carbon dioxide
evolution (∼11 000 L!), once again being in a virtually linear
relationship to alanine addition. The crude product solution
was concentrated by evaporation of acetic acid, acetic
anhydride, and triethylamine. This concentrate was diluted
with water and malonodinitrile, and the resulting solution
was added to 30% sodium hydroxide, whereby pyrrole 3 was
formed and precipitated in an extremely fast, exothermic,
but fully addition-controlled reaction.13 Crystalline pyrrole
3 was directly isolated by centrifugation, was washed with
water, and was dried and did not require further purification.14
Neither distillate nor mother liquor required any special
treatment. This procedure for the synthesis of pyrrole 3
(Scheme 4) is very efficient, high-yielding (83-85% over
two steps calculated from alanine) and requires only two
standard reactors.
Pyrimidine Ring Synthesis. Numerous examples are
reported in the literature, whereby 2-amino-3-cyano furans,
pyrans, or other related derivatives are converted to pyrim-
idines particularly by formic acid, formamide, or mixtures
thereof.15 As outlined in the introduction of this contribution,
we intended to avoid this type of conversion. Already back
in 1964, Taylor and Hendess reported a procedure16 where
an unprotected amino-cyano-pyrrole was converted to an
aminopyrrolopyrimidine derivative by trimethyl orthoformate
and ammonia. However, in the more recent literature, this
method is only used for thiophens,17 pyrans,18 pyrazines,19
or other heterocycles lacking a free amino group,20 and
almost exclusively in combination with hydrazine as the
amine component, indicating a somewhat limited scope of
this method, particularly for unprotected pyrroles.
However, we reasoned that in principle, it should be
possible to apply this type of pyrimidine ring formation to
pyrroles under carefully optimized conditions which avoid
side reactions at the pyrrole ring nitrogen. To explore this,
we initially prepared iminoester 4a (Figure 2) from pyrrole
3 and triethyl orthoformate. The crude iminoester was then
treated with 3-chloroaniline. This sequence under various
conditions at best produced moderate yields of 3-chlorophen-
yl-pyrrolopyrimidine 5 in rather unsatisfying quality. There-
fore, the reaction sequence was changed: iminoester 4b was
formed from 3-chloroaniline and triethyl orthoformate under
(15) See, for example: Pichler, H.; Folkers, G.; Roth, H. J.; Eger, K. Liebigs
Ann. Chem. 1986, 9, 1485.
(12) Steglich, W.; Ho¨fle, G. Angew. Chem. 1969, 81(23), 1001.
(13) Safety remark: malonodinitrile decomposes autocatalytically at elevated
temperatures with an adiabatic temperature rise in excess of 1000 °C! The
initiation phase of the decomposition is shortened by bases. For these
reasons, when performing this type of pyrrole condensation reactions on
large scale, malonodinitrile should be the component being added to the
sodium hydroxide solution rather than the opposite way to avoid a possible
run-away reaction.
(16) Taylor, E. C.; Hendess, R. W. J. Am. Chem. Soc. 1964, 86, 951.
(17) See, for example: Shishoo, C. J.; Devani, M. B.; Ullas, G. V.; Ananthan,
S.; Bhadti, V. S. J. Heterocycl. Chem. 1981, 18, 43.
(18) See, for example: Younes, M. I.; Metwally, S. A.; Atta, A. H. Synthesis
1990, 704.
(19) See, for example: Taylor, E. C.; Hartke, K. S. J. Am. Chem. Soc. 1959, 81,
2456.
(14) Bulk substance was stored in barrels which were flushed with nitrogen to
avoid pyrrole oxidation.
(20) See, for example: Birkett, P. R.; Chapleo, C. B.; MacKenzie, G. Synthesis
1991, 152.
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
583

3 to pyrrolopyrimidine 5 therewith. Further, the reaction
could be accelerated by using excess aniline with respect to
orthoacetate. To avoid the side reaction of intermediate
amidine 4d with a second pyrrole 3, the procedure was
changed from batch to semi-batch by adding pyrrole 3 over
a few hours, therefore keeping the free pyrrole concentration
in the reaction mixture low. As an additional benefit, this
procedure allowed for the replacement of expensive pyri-
dinium tosylate catalyst by acetic acid. This optimized and
simplified process was run in the pilot plant on 0.45 kmol
scale. Chloroaniline and triethyl orthoformate were dissolved
in ethanol, and the pH of the solution was adjusted to ca.
5.5. Pyrrole 3 was then added over a few hours at 45-50
°C, precipitated pyrrolopyrimidine 5 was centrifuged off,
washed, and dried as above. The yield in the pilot plant was
over 80% (based on pyrrole 3).
Chlorophenyl-pyrrolopyrimidine 5 now only needed to
be converted to its isomer 1. This isomerization, which is
referred to as Dimroth rearrangement,24 is not truly a
rearrangement but rather a hydrolysis of the pyrimidine ring
at the N3-C2 bond, subsequent rotation of 180° around the
C9-C4 bond, and ring closure. For pyrrolopyrimidines, this
type of isomerisation is very scarcely reported in the
literature,25 with only few more examples for other pyrimi-
dine derivatives.26 With that little precedence, the only thing
which seemed obvious was that water would be required for
the isomerisation to occur. First lab trails centered around
the variation of catalyst (acid, base), cosolvent, and temper-
ature. It turned out, that no catalyst at all was required and
that reaction temperature was the most important parameter.
A solvent mixture containing water, having the appropriate
boiling point and a sufficient solubility for pyrrolopyrimidine
5 had to be elaborated. A mixture of ethylene glycol, ethanol,
and water eventually turned out to be optimal; clean
isomerization could be achieved in over 90% yield by simply
refluxing for a few hours. Upon cooling, the product
precipitates and can once again be isolated by filtration,
washing and drying. This procedure was successfully run in
the pilot plant in 0.3 kmol scale. (Scheme 5).
Figure 2.
acid catalysis21 and isolated as a crude oil. Pyrrole 3 was
then treated with iminoester 4b. Lab experiments revealed
that this procedure was definitely superior particularly after
optimization of solvent, acid catalyst, and pH. The procedure
was eventually run successfully in the pilot plant on 0.3 kmol
scale. Pyrrole 3 was thereby dissolved in absolute ethanol,
crude iminoester 4b was added (slight excess), and the pH
was adjusted to 8 by the addition of pyridinium tosylate
solution in ethanol. After 24 h at room temperature, product
5 had precipitated from the reaction and was centrifuged,
washed with ethanol, and dried.
Our lab experiments had made clear that the mechanism
of this pyrimidine ring formation involves a series of ortho
ester-iminoester-amidine equilibria between aniline, pyr-
role-amine, and ortho ester. Since we neversnot even in the
above pilot campaignsobserved a temperature rise during
the reaction, we reasoned that precipitation of the product
from the reaction mixture and not thermodynamic stability
had to be the ‘trap’ which drives these equilibria to the
product side.22 Hence, it should be possible to run the reaction
under kinetic control without preformation of an iminoester
intermediate; this was potentially beneficial because any
procedure involving heat in the presence of the pyrrole
staring material unavoidably resulted in a more or less dark
brown reaction mixture.
With this hypothesis in mind we initiated lab experiments
which started with all three components, pyrrole 3, 3-chlo-
roaniline, and triethyl orthoformate and which varyied the
ratios of staring materials and particularly the pH. In
agreement with our hypothesis, we found that at slightly
acidic pH the reaction proceeded relatively slowly but
surprisingly cleanly. Our TLC- and NMR experiments
confirmed that this formation of pyrrolopyrimidine 5 was
distinct from the first procedure because only trace amounts
of iminoester 4b were detected in the reaction mixture.
Instead, a new, major intermediate was found and upon
isolation was identified as amidine 4c which now seemed
to be the species reactive towards the formation of the
pyrimidine ring.23 This interpretation was confirmed by
preparative synthesis of amidine 4c and conversion of pyrrole
The drug substance free base was brought in its mesylate
salt form by crystallisation from ethanol/water in the presence
of methylsulfonic acid and final recrystallisation from
acetone/water, which yielded a white powder with high purity
(HPLC assay >99%) that met all internal specifications and
was subsequently released for use in humans.
Conclusions
We have demonstrated the development of a very efficient
synthesis of pyrrolopyrimidine 1 in only four simple steps
from inexpensive starting materials such as racemic alanine,
(21) Glickman, A. S. U.S. Patent 2,684,976, 1954.
(24) For a review on Dimroth rearrangements, see for example: Wahren, M. Z.
Chem. 1969, 9(7), 241.
(25) Zaharan, M. A.; Pedersen, E. B.; Nielsen, C. Monatsh. Chem. 1995, 126(11),
1271.
(26) For example: (a) Brown, D. J.; Jacobsen, N. W. J. Chem. Soc. 1960, 1978.
(b) Taylor, E. C.; Loeffler, P. K. J. Am. Chem. Soc. 1960, 82, 3147. (c)
Ohtsuka, Y. Bull. Chem. Soc. Jpn. 1970, 43, 3909. (d) Antonov, D. M.;
Belen’kii, L. I.; Dudinov, A. A.; Krayushkin, M. M. Khim. Geterotsikl.
Soedin. 1994, 4, 450.
(22) This interpretation was later confirmed by applying the final, optimized
procedure to different anilines: only those which gave pyrrolopyrimidines
poorly soluble in ethanol and which were therefore precipitating from the
reaction mixture could be prepared by this procedure.
(23) NMR investigations on the synthesis of iminoester 4b had shown earlier,
that at lower temperature, amidine 4c was the predominant product in the
mixture, and only by heating above 120 °C and with excess orthoformate,
could iminoester 4b be obtained.
584
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development

Scheme 5
was added, and this mixture was slowly added to 500 g of
aqueous sodium hydroxide solution (30%). The speed of
addition was adjusted so that the reaction temperature did
not exceed 60 °C. The resulting suspension was cooled to 0
°C, filtered, washed with water, and dried in vacum at 55
°C to yield 168 g (83% calculated from alanine) of pyrrole
3: assay >96% (HPLC); mp (DSC) 172 °C (lit.4 163-165
1
°C); H NMR(DMSO): 9.83 (s, 1H), 5.36 (s, 2H), 1.92 (s,
3H), 1.83 (s, 3H); MS (ESI+) 136 (MH+).
3-(3-Chlorophenyl)-5,6-dimethyl-4H-pyrrolo[2,3-d]py-
rimidine-4-imine (5). Triethyl orthoformate (204.5 g, 1.35
mol), absolute ethanol (700 mL), and 3-chloroaniline (222
g, 1.72 mol) were charged into a reaction vessel, and the
pH was adjusted to 5-5.5 by addition of acetic acid (6.65
g, 0.11 mol). The mixture was warmed to 50 °C and stirred
for 1 h. Pyrrole 3 (166 g, 1.23 mol) was added as a solid
over 6 h, maintaining temperature at 45-50 °C. After
completion of the addition, the mixture was kept at 50 °C
for additional 4 h and then for another 8 h at room
temperature. Water (100 mL) was added, the mixture was
cooled to 0 °C, and this temperature was maintained for 30
min. The resulting suspension was filtered, washed with
ethanol/water (4:1), and dried in vacum at 50 °C to yield
261 g (78% calculated from pyrrole 3) of pyrrolopyrimidine
5: assay >98% (HPLC); mp >150 °C (change of crystal
malonodinitrile, chloroaniline, and triethyl orthoformate.27
The procedure, which was successfully run on pilot plant
scale, uses no special equipment. Intermediates 3, 5, and 1
(free base) are all crystallized directly from the reaction
mixtures, whereas 2 is used in situ, making extractions or
other cumbersome purification procedures unnecessary as
well as minimizing the amount of waste produced. These
factors all contribute to remarkably low costs for the drug
substance. In addition, westo the best of our knowledges
reported the first application of a Dakin-West reaction on
technical scale with carbon dioxide evolution being con-
trolled by addition of the amino acid thanks to a new,
modified procedure.
1
modification and dec); H NMR (DMSO): 10.97 (s, 1H),
10.26 (s, 1H), 8.56 (s, 1H), 7.39-7.14 (m, 4H), 2.08 (s, 3H),
1.96 (s, 3H); MS (ESI+) 273 (MH+).
4-(3-Chlorophenylamino)-5,6-dimethyl-7H-pyrrolo-
[2,3-d]pyrimidine (1) Free Base. A suspension of pyrrol-
opyrimidine 5 (300 g, 1.1 mol) in water (750 mL), absolute
ethanol (750 mL), and ethylene glycol (1500 mL) was heated
to 95 °C for 4 h. The mixture was cooled to room
temperature within 90 min. After the temperature was
maintained at room temperature for an additional hour, the
suspension was filtered, washed with water, and dried under
vacuum at 50 °C to yield 277 g (92% based on 5) of
pyrrolopyrimidine 1 (free base): assay >98% (HPLC); mp.
Experimental Section
Experimental details are described only for the final
process, which was successfully run in the pilot plant.
Starting materials, solvents, and reagents were of technical
grade. All reactions were carried out under inert atmosphere
(N2). NMR spectra where measured on a Varian Gemini 200
MHz spectrometer, chemical shifts are reported in ppm and
referenced to the solvent lock signal. HPLC analyses were
performed on a Nucleosil-C18 column (5 µm) with UV-
detection using an aqueous sodium perchlorate buffer
(solvent A)/acetonitrile (solvent B) gradient. HPLC assays
are % w/w except where stated otherwise. Reference samples
were prepared by additional chromatography and recrystal-
lisation of compounds.
1
240-255 °C (change of crystal modification); H NMR
(DMSO): 11.53 (s, 1H), 8.21 (s, 1H), 8.15 (s, 1H), 7.95 (d,
1H), 7.71 (t, 1H), 7.33 (t, 1H), 7.03 (q, 1H), 2.41 (s, 3H),
2.28 (s, 3H); MS (ESI+) 273 (MH+).
4-(3-Chlorophenylamino)-5,6-dimethyl-7H-pyrrolo-
[2,3-d]pyrimidine (1) Mesylate Salt. A mixture of free base
1 (30 g, 0.11 mol) in ethanol 97% (390 mL) and methyl-
sulfonic acid (11.1 g, 0.115 mol) was heated to reflux for
45 min and then cooled to 0 °C within 3 h, whereby
crystallisation occurred. The suspension was stirred at 0 °C
for an additional 90 min and then filtered, washed with cold
ethanol, and dried in vacum at 55 °C to yield 36.5 g (90%)
of mesylate salt of 1:assay >98% (HPLC) (analytical data
see below).
4-(3-Chlorophenylamino)-5,6-dimethyl-7H-pyrrolo-
[2,3-d]pyrimidine (1) Mesylate Salt; Final Drug Substance
Recrystallisation. Mesylate salt 1 (30 g, 0.081 mol) was
dissolved in acetone (204 mL) and water (36 mL) at reflux.
The solution was filtered, and the filtrate was cooled to 30
°C within 3 h, seeding at 50 °C. The suspension was cooled
2-Amino-3-cyano-4,5-dimethylpyrrole (3). Acetic an-
hydride (338.5 g, 3.3 mol), acetic acid (45 g, 0.75 mol),
triethylamine (379.5 g, 3.75 mol) and 4-(dimethylamino)-
pyridine (1.83 g, 0.015 mol) were charged into a reaction
vessel and heated to 50 °C. Racemic alanine (135 g, 1.5 mol)
was added as a solid over 6 h thereby maintaining reaction
tempeature at 45 to 55 °C. After completion of alanine
addition, the reaction mixture was stirred at 50 °C for an
additional 8 h. Acetic anhydride, triethylamine, and acetic
acid were distilled off by vacuum distillation (10-15 mbar),
gradually raising the jacket temperature to a maximum of
100 °C. The residue (GC assay > 95% acylaminoketone
intermediate 2) was cooled to room temperature and diluted
with water (815 mL). Malonodinitrile (94.5 g, 1.425 mol)
(27) Misun, M.; Fischer, R.; Mutz, M. PCT Int. Appl. WO 98/43973, 1998.
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
585

to 0 °C within 1 h, and acetone (500 mL) was added. The
resulting suspension was filtered, washed with cold acetone,
and dried in vacum at 55 °C to yield 26.4 g (88%) of
pyrrolopyrimidine 1 mesylate salt: assay >99% (HPLC);
sum of byproducts <0.5% area (HPLC); mp 215-222 °C;
1H NMR (DMSO): 12.53 (s, 1H), 9.55 (s, 1H), 8.28 (s, 1H),
7.70 (s, 1H), 7.54-7.50 (m, 2H), 7.43-7.38 (m, 1H), 2.38
(s, 3H), 2.34 (s, 3H), 2.32 (s, 3H); MS (ESI+) 273 (MH+).
Acknowledgment
The skillfull work of our technicians Ch. Kehrli, M.
Haller, and L. Kohler is gratefully acknowledged. We thank
Ch. Heuberger and his staff at the safety lab for thermody-
namic investigations and M. Tru¨by and his team from the
pilot plant for technical support with scale up.
Received for review May 18, 2001.
OP010041V
586
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development