Highly selective hydrolysis of chloropyrimidines to pyrimidones in 12 N hydrochloric acid

Article abstract of DOI:10.1021/op060093q

A chromatography-free process for synthesis of 6-piperazinyl-2,4-bis- pyrrolidinylpyrimidine in isomerically pure form is described. The key step is the purification of a crude 6-chloro-2,4-bis-pyrrolidinylpyrimidine/2-chloro-4, 6-bis-pyrrolidinylpyrimidine isomer mixture (generated by reaction of 2,4,6-trichloropyrimidine with pyrrolidine) by a highly selective acid-catalyzed hydrolysis of the 2-chloro isomer to the pyrimidone. The 2-chloro isomer hydrolyzes 350 times faster than the 6-chloro isomer in 6 N HCl and 1750 times faster in 12 N HCl. To put these rate ratios in perspective, the 2-chloro isomer reacts with amines and alkoxides only ～ 10-17 times faster than does the 6-chloro isomer. A mechanistic investigation using methodological tools developed by Bunnett established that the transition state for hydrolysis of the 6-chloro isomer involves two more molecules of water (each acting as a base) than does the transition state for hydrolysis of the 2-chloro isomer. As the concentration of HCl increases from 3 N to 6 N to 12 N, there are fewer unprotonated water molecules. Thus, the transition state that involves the greater number of unprotonated water molecules (6-chloro-2,4-bis- pyrrolidinylpyrimidine) is expected to be increasingly disfavored with increasing acid concentration, as is observed. The optimized process was run successfully on production scale.

Full text of DOI:10.1021/op060093q

Organic Process Research & Development 2006, 10, 921−926
Highly Selective Hydrolysis of Chloropyrimidines to Pyrimidones in 12 N
Hydrochloric Acid
Amphlett G. Padilla and Bruce A. Pearlman*,†
Chemical Research and DeVelopment, 31073-091-201, Pfizer, Inc., 7000 Portage Road,
Kalamazoo, Michigan 49001-0102, U.S.A.
Abstract:
Such a separation problem was encountered in process
research directed toward Tirilazad Mesylate (5), a drug
candidate in development at The Upjohn Company in the
early 1990s for treatment of aneurysmal subarachnoid
hemorrhage in males. The first step in the first generation
process7 involved the reaction of trichloropyrimidine with
pyrrolidine in THF, to generate a 92:8 mixture of 6-chloro-
2,4-bis-pyrrolidinylpyrimidine 3u and 2-chloro-4,6-bis-pyr-
rolidinylpyrimidine 3s8 (Scheme 1). Silica gel chromatog-
raphy was required to separate the undesired isomer (3s).
Sharpless has shown that any impurity that is more
reactive than the major component can be removed to any
arbitrary level by reacting the mixture to arbitrary conversion
but that the impurity must be at least ∼20 times more reactive
than the major component, or else a significant fraction of
the major component will be lost.9 For example, hypotheti-
cally, if a reagent could be found that could react 20 times
faster with 2-chloro-4,6-bis-pyrrolidinylpyrimidine 3s than
with 6-chloro-2,4-bis-pyrrolidinylpyrimidine 3u then, on
treatment of the 8:92 mixture with that reagent, 19% yield
of 3u would be lost in driving the level of 3s down to 0.1%.
A chromatography-free process for synthesis of 6-piperazinyl-
2,4-bis-pyrrolidinylpyrimidine in isomerically pure form is
described. The key step is the purification of a crude 6-chloro-
2,4-bis-pyrrolidinylpyrimidine/2-chloro-4,6-bis-pyrrolidinylpy-
rimidine isomer mixture (generated by reaction of 2,4,6-
trichloropyrimidine with pyrrolidine) by a highly selective acid-
catalyzed hydrolysis of the 2-chloro isomer to the pyrimidone.
The 2-chloro isomer hydrolyzes 350 times faster than the
6-chloro isomer in 6 N HCl and 1750 times faster in 12 N HCl.
To put these rate ratios in perspective, the 2-chloro isomer
reacts with amines and alkoxides only ∼10-17 times faster than
does the 6-chloro isomer. A mechanistic investigation using
methodological tools developed by Bunnett established that the
transition state for hydrolysis of the 6-chloro isomer involves
two more molecules of water (each acting as a base) than does
the transition state for hydrolysis of the 2-chloro isomer. As
the concentration of HCl increases from 3 N to 6 N to 12 N,
there are fewer unprotonated water molecules. Thus, the
transition state that involves the greater number of unproto-
nated water molecules (6-chloro-2,4-bis-pyrrolidinylpyrimidine)
is expected to be increasingly disfavored with increasing acid
concentration, as is observed. The optimized process was run
successfully on production scale.
Results and Discussion
Nucleophilic Reagents. A 3u/3s isomer mixture was
treated with a variety of nitrogen and oxygen nucleophiles,
and the relative rates of reaction (ks/ku) were measured (Table
1). In all cases the rate ratio was less than 20.
Introduction
2,4,6-Trichloropyrimidine (1) is a potentially attractive
starting material for synthesis of aminopyrimidines. However,
it reacts with alkylamines,1 arylamines,2 amide anions,3
guanidine,4 and carbamate anions,5 as well as phenoxide
anions,6 to give mixtures of 2- and 4/6-isomers, resulting in
separation problems.
Aqueous Hydrochloric Acid. In contrast with the insuf-
ficiently high selectivities obtained with nucleophilic re-
agents, it was found that, on stirring a 3u/3s mixture in 20°
Be (32 wt %; 10.2 M) hydrochloric acid (an inexpensive,
commercially available grade), 3s hydrolyzed 230 times
faster than 3u (Scheme 2)!10 Moreover, the hydrolysis
products (pyrimidones 6s and 6u) were cleanly separated
from 3u by extraction of the crude hydrolysis mixture in
Isopar H11/heptane/xylene with pH 2.1 water, as the pyrimi-
dones are stronger bases than the chloropyrimidines. Extrac-
tion with pH 3.4 water removed essentially all the 2-pyrim-
* To
whom
correspondence
should
be
addressed.
E-mail:
bruce.pearlman@spcorp.com.
† Current address: Schering-Plough Research Institute, 1011 Morris Ave.,
U-3-2 2100, Union, NJ 07083, USA.
(1) (a) Schneider, R. U.S. Patent 4,025,515, May 24, 1977. (b) Hoegerle, K.;
Berrer, D. U.S. Patent 4,082,535, Apr. 4, 1978. (c) Delia, T. J.; Stark, D.;
Glenn, S. K. J. Heterocycl. Chem. 1995, 32, 1177.
(2) (a) Koga, M.; Schneller, S. W. J. Heterocycl. Chem. 1992, 29, 1741. (b)
Schomaker, J. M.; Delia, T. J. J. Heterocycl. Chem. 2000, 37, 1457.
(3) Delia, T. J.; Meltsner, B. R.; Schomaker, J. M. J. Heterocycl. Chem. 1999,
36, 1259.
(7) Jacobsen, E. J.; McCall, J. M.; Ayer, D. E.; Van Doornik, F. J.; Palmer, J.
R.; Belonga, K. L.; Braughler, J. M.; Hall, E. D.; Houser, D. J.; Krook, M.
A.; Runge, T. A. J. Med. Chem. 1990, 33, 1145.
(4) Ladd, D. L. J. Heterocycl. Chem. 1982, 19, 917.
(8) The letters u and s denote the unsymmetrical and symmetrical isomer,
respectively.
(9) Martin, V. S.; Woodard, S. S.; Katsuki, T.; Yamada, Y.; Ikeda, M.; Sharpless,
K. B. J. Am. Chem. Soc. 1981, 103, 6237.
(10) A preliminary account of this research has been published: Pearlman, B.
Specialty Chemicals 1998, 18, 94.
(11) Isopar H is a mixture of branched hydrocarbons of approximate boiling
point range 177 to 187 °C.
(5) (a) Zanda, M.; Talaga, P.; Wagner, A.; Mioskowski, C. Tetrahedron Lett.
2000, 41, 1757. (b) Montebugnoli, D.; Bravo, P.; Corradi, E.; Dettori, G.;
Mioskowski, C.; Volonterio, A.; Wagner, A.; Zanda, M. Tetrahedron 2002,
58, 2147. (c) Montebugnoli, D.; Bravo, P.; Brenna, E.; Mioskowski, C.;
Panzeri, W.; Viani, F.; Volonterio, A.; Wagner, A.; Zanda, M. Tetrahedron
2003, 59, 7147.
(6) Delia, T. J.; Nagarajan, A. J. Heterocycl. Chem. 1998, 35, 269.
10.1021/op060093q CCC: $33.50 © 2006 American Chemical Society
Published on Web 07/08/2006
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
921

Scheme 1. First generation process to produce Tirilazad Mesylate
a The two monoadducts 2u and 2s are formed in this ratio on treatment of trichloropyrimidine with 1.9 equiv of pyrrolidine (THF, -6 °C). See Supporting Information
for details.
the transition state of the rate-determining step.12 He
Table 1. Selective destruction of 3s by nucleophilic reagents
introduced three variables (ω, ω*, and φ) to measure the
deviation from linearity, defined by eqs 1-3,
log k + Ho ) ω log aH O + C
(1)
(2)
(3)
2
log k + log[HCl] ) ω* log aH O + C
2
log k + Ho ) φ (Ho + log[HCl]) + C
where Ho is the Hammett acidity function (which measures
the effective acidity of the proton as a function of acid
concentration) and aH O is the activity coefficient of water
2
(which measures the effective basicity/nucleophilicity of
water as a function of acid concentration).
entry
nucleophile
conditions
ks/ku
9.6
17.1
>14.9
14.0
To determine the values of the Bunnett coefficients for
hydrolysis of the two isomeric chloropyrimidines 3u and 3s,
the rates of hydrolysis of 3u and 3s were measured in 3 N,
6 N, and 12 N hydrochloric acid. Results are tabulated in
Table 2.
Values of ω, ω*, and φ were then calculated using eqs
1-3 and the rate data in the first two entries of Table 2.
The results are tabulated in Table 3.
Bunnett found that “high” values of ω, ω*, and φ (defined
as higher than 3.3, -2, and 0.58, respectively) are empirically
associated with hydrolysis reactions in which more than one
water molecule (one acting as a nucleophile and the rest as
bases) are involved in the rate-determining step, that “low”
values (i.e., lower than 0, lower than an unspecified value,
and lower than 0, respectively) are associated with hydrolysis
reactions in which water is not involved in the rate-
1
2
3
4
5
piperazine
pyrrolidine
benzylamine
NaOCH3
pyridine, ∆
138 °C
190 °C
MeOH, ∆
DMSO, 52 °C
KOtBu
13.8
idone 6s along with only a minor amount of the 6-pyrimidone
6u, showing that 6s is slightly more basic than 6u. The
chloropyrimidine 3u was recovered in isomerically pure form
(containing 0.02% 3s) with a loss of only 2.2% yield of 3u.
Mechanistic Experiments. Rates of hydrolysis reactions
in strong aqueous acid (g∼3 N) do not, in general, depend
linearly on either the proton concentration or the water
concentration because, as the proton concentration increases,
the protons become increasingly desolvated, thus becoming
more “acidic,” and the oxygen lone pairs of the water
molecules become increasingly hydrogen bonded, thus
becoming less “basic.” Bunnett has shown that the deviation
from linearity gives information about the role of water in
(12) (a) Bunnett, J. F. J. Am. Chem. Soc. 1961, 83, 4956. (b) Bunnett, J. F. J.
Am. Chem. Soc. 1961, 83, 4968. (c) Bunnett, J. F. J. Am. Chem. Soc. 1961,
83, 4973. (d) Bunnett, J. F. J. Am. Chem. Soc. 1961, 83, 4978. (e) Bunnett,
J. F.; Olsen, F. P. Can. J. Chem. 1966, 44, 1899. (f) Bunnett, J. F.; Olsen,
F. P. Can. J. Chem. 1966, 44, 1917.
922
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

Scheme 2. Selective hydrolysis of 3s in 10 N hydrochloric acid
Table 2. Effect of acid concentration on hydrolysis rate
Table 3. Values of Bunnett coefficients for hydrolysis of
chloropyrimidines
[HCl],
entry M (wt %)
ku,
ks,
log
2
a
b
h-1
h-1
ks/ku
Ho
aH O
1
2
3
12.22
(37.5)
6.21
(20.66)
2.96
(10.33)
9.97 × 10-4 1.74
4.72 × 10-4 0.166
0.65 × 10-4 0.0324 500 -1.05 -0.0533
1750 -4.30 -0.5131
350 -2.19 -0.1752
a The values at which the mechanism crosses over from one in which water
acts as a nucleophile to one in which it acts as both a base and nucleophile.
a Hammett acidity constant. Values are taken from Table II of ref 13 (Ho′
column) rather than from ref 12a because the ref 12a dataset does not include a
value for 12.22 M HCl. The values of Ho for 6 N and 3 N HCl given by the two
datasets agree very closely. b Common logarithm of the activity coefficient of
water; values are taken from ref 14.
Activation Parameters. Further insight into the mech-
anism can be gained by comparing the activation parameters
of the two transition states.16 The difference of the activation
free energies for hydrolysis of the two chloropyrimidines
(∆Gqs - ∆Gqu) can be calculated by the Eyring equation (eq
4).
determining step, and that values in between are associated
with reactions in which one molecule of water is involved
as a nucleophile.12b,f In the case of the hydrolysis of 3u, the
values of ω, ω*, and φ are all “high,” indicating that more
than one water molecule (one acting as a nucleophile and
the rest as bases) are involved in the rate-determining step.
In the case of the hydrolysis of 3s, the values of ω, ω*, and
φ are on the border between the “high” and “middle” ranges,
indicating that the rate-determining step involves one mol-
ecule of water acting as a nucleophile and may or may not
involve additional molecules of water acting as bases.
Bunnett also pointed out that “an extreme interpretation
of ω*-values would be that they represent the number of
water molecules of change of hydration between reactants
and transition states and that ω represents the same quantity
on an adjusted scale.”12c The values of both ω and ω* for
3u are about 2 units higher than those for 3s. Thus, according
to this guideline, the transition state for hydrolysis of 3u
involves two more water molecules (each acting as a base)
than that for hydrolysis of 3s.15
ks
∆Gqs - ∆Gqu ) -RT ln
(4)
( )
ku
1.9872 cal
mol K
) -
(298 K) ln(1750)
(
)
kcal
mol
) -4.4
The activation entropies and enthalpies were not measured
experimentally. However, Bunnett has shown that, on
average, the activation entropy (∆Sq) of a transition state
decreases by approximately 4.1 cal/mol K per ω unit.12c Thus,
a reasonable estimate of the difference of the activation
entropies for hydrolysis of the two chloropyrimidines can
be calculated from eq 5.
(14) values interpolated from table published in: Akerlof, G.; Teare, J. W. J.
Am. Chem. Soc. 1937, 59, 1855.
(13) Arnett, E. M.; Mach, G. W. J. Am. Chem. Soc. 1966, 88, 1177.
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
923

(ωs - ωu)(-4.1 cal)
∆Sqs - ∆Squ )
(5)
mol K
(3.2 - 5.3)(-4.1 cal)
)
)
mol K
8.6 cal
mol K
Figure 1. Proposed transition states for hydrolysis of 3u and
3s, respectively.
The difference in the activation enthalpies (∆Hqs - ∆Huq)
can then be calculated from eq 6.
developed. The key step is the purification of a crude
6-chloro-2,4-bis-pyrrolidinylpyrimidine/2-chloro-4,6-bis-pyr-
rolidinylpyrimidine (3u/3s) isomer mixture (generated by
treatment of 2,4,6-trichloropyrimidine with pyrrolidine) by
a highly selective acid-catalyzed hydrolysis of the 2-chloro
isomer to the pyrimidone, which is accomplished simply by
stirring the mixture in 12 N HCl at rt. The 2-chloro isomer
hydrolyzes completely to the pyrimidone, while the 6-chloro
isomer is essentially inert (ks/ku ) 1750). The process was
run successfully on production scale.
∆Hqs - ∆Hqu ) ∆Gqs - ∆Gqu + T(∆Sqs - ∆Suq)
(6)
kcal
mol
8.6 cal
mol K
) -4.4
+ (298 K)
(
)
(
)
kcal
mol
kcal
mol
) -4.4
+ 2.6
kcal
mol
) -1.8
The conclusion that emerges from this analysis is that 3s
hydrolyzes much faster than 3u because both the activation
entropy and activation enthalpy are higher for hydrolysis of
3u than for hydrolysis of 3s, by ca. 2.6 and 1.8 kcal/mol,
respectively.
Experimental Section
Generation of Chloro-bis-(pyrrolidinyl)pyrimidine Iso-
mer Mixture (3u/3s). To a 1 L flask was added in sequence
sodium carbonate (mw 105.99; 58 g, 0.547 mol, 2.01 equiv),
500 mL heptane, and 2,4,6-trichloropyrimidine (mw 183.42;
50.0 g, 0.273 mol). The slurry was agitated, and 80 mL of
water were added. To the mixture was added pyrrolidine (mw
71.12; d 0.852; 55.0 mL, 46.9 g, 0.659 mol, 2.42 equiv) at
a rate such that the pot temperature remained below 80 °C.
Once the addition was complete, the pot temperature was
adjusted to 80 ( 1 °C and the mixture was stirred until the
reaction was complete by TLC18 (12 h). The mixture was
quenched with 240 mL of water, the pot temperature was
adjusted to g70 °C, the phases separated, and the aqueous
phase was extracted with 50 mL of 70 °C heptane. The
heptane extracts were combined and taken on to the next
step. The ratio of 3u to 3s was 97.09:2.91 by LC.19
Selective Hydrolysis of 3s. The combined heptane
extracts from the previous step were cooled to 22 °C. 20°
Be Hydrochloric acid (mw 36.46; 32 wt %; 10.2 M; d 1.1593;
150 mL, 55.65 g, 1.526 mol, 5.60 equiv) was added, resulting
in a 10-12 °C exotherm. The mixture initially formed two
cloudy liquid layers. On stirring, the upper layer became clear
and colorless, and the lower layer became hazy yellow-
orange. The 3u/3s was exclusively in the lower aqueous
layer. The mixture was stirred at 40-50 °C until the level
of 3s was e0.07 mol % (4-5 h) by LC.20 The normalized
Proposed Transition States. To rationalize the conclu-
sion that ∆Hqu is greater than ∆Hqs, we propose that the
carbon-oxygen bond in the transition state for hydrolysis
of 3u is stronger (and thus shorter) than that in the transition
state for hydrolysis of 3s, as depicted in Figure 1.
The extremely high selectivity in 12 N HCl can be
rationalized in terms of the two proposed transition states.
As the concentration of HCl increases from 3 N to 6 N to
12 N, there are fewer “free” (i.e., unprotonated) water
molecules. Thus, the transition state that involves the greater
number of unprotonated water molecules (3u) is expected
to be increasingly disfavored with increasing acid concentra-
tion, as is observed.
Piperazine Displacement. The crude, isomerically pure
chloropyrimidine 3u was converted into the piperazine
derivative 4u by the procedure outlined in Scheme 3. The
yield of 4u starting from 200 kg trichloropyrimidine was
261 kg (79% yield).17
Conclusions
A chromatography-free process for synthesis of isomeri-
cally pure 6-piperazinyl-2,4-bis-pyrrolidinylpyrimidine was
(15) Since 3s is more polar on silica gel TLC than 3u, the possibility was
considered that 3s may hydrolyze more rapidly than 3u because a higher
percentage of 3s than 3u may be protonated. This possibility was investigated
by a study of the effect of solvent (75% CD3CO2D, 3 N HCl, and 6 N HCl)
on the 13C NMR chemical shift of C2/C4/C6 of the two isomers, based on
the assumption that the change in chemical shift should be greater for the
more basic isomer, because a greater fraction of molecules would be
protonated. The signals for C2/C4/C6 of both isomers shifted upfield
progressively from 75% CD3CO2D to 3 N HCl to 6 N HCl by approximately
the same amount, indicating that roughly similar fractions of 3s and 3u are
protonated at each acid concentration, thus contradicting the premise that a
higher percentage of 3s than 3u is protonated. See Supporting Information
for details.
(16) We thank a referee for the suggestion that the high selectivity could be due
to a difference in activation entropy, which would be expected to be greater
for the transition state requiring the ordering of two extra molecules of water
(3u).
(17) This process has been patented: Pearlman, B. A.; Padilla, A. G. U.S. Patent
5,225,555, July 6, 1994.
(18) TLC procedure: plate, silica gel; eluant, 10% ethyl acetate/cyclohexane;
visualization, UV (short wavelength); 2,4,6-trichloropyrimidine, Rf ) 0.17;
2-chloro-4,6-bis-pyrrolidinylpyrimidine (3s), Rf ) 0.20; 2,4-dichloro-6-
pyrrolidinylpyrimidine (2u), Rf ) 0.28; 6-chloro-2,4-bis-pyrrolidinylpyri-
midine (3u), Rf ) 0.44; 4,6-dichloro-2-pyrrolidinylpyrimidine (2s), Rf )
0.58; 2,4,6-trichloropyrimidine, Rf ) 0.58. The reaction mixture is spotted
directly. The reaction is judged complete when only 3u/3s are present.
(19) LC assay: column, DuPont Zorbax C8, 25 cm × 4.6 mm; temperature,
ambient; flow rate, 2.9 mL/min; detection, 254 nm; mobile phase, 500 mL
of methanol/200 mL of acetonitrile/300 mL of water/0.8 mL of triethylamine/
0.4 mL of acetic acid. Retention times: 2,4,6-trichloropyrimidine, 2.03 min;
2,4-dichloro-6-pyrrolidinylpyrimidine (2u), 2.40 min; 2-chloro-4,6-bis-
pyrrolidinylpyrimidine (3s), 3.18 min.; 4,6-dichloro-2-pyrrolidinylpyrimidine
(2s), 4.76 min; 6-chloro-2,4-bis-pyrrolidinylpyrimidine (3u), 6.52 min; 2,4,6-
tripyrrolidinylpyrimidine, 8.94 min.
924
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development

Scheme 3. Optimized process incorporating highly selective acid-catalyzed chloropyrimidine hydrolysis (production run)
ratio of 3u:3s:6u:6s was 94.872:0.015:2.222:2.891, from
which it was calculated that ks/ku ) 227.5. The mixture was
cooled to 14 °C, and 75 mL of 50% aq. sodium hydroxide
were added at a rate such that the pot temperature remained
e50 °C. The slurry was heated to g70 °C, and 75 mL of
water were added. The pH was adjusted to 1.8 to 2.0 by
addition of 3 mL of 50% sodium hydroxide. The pot
temperature was kept at g65 °C during all the extractions.
The homogeneous layers were separated, and the organic
layer was extracted sequentially with the following: (1) a
solution composed of 50 mL of water and 4 mL of 50%
sodium hydroxide; (2) 50 mL of water. Each of the aqueous
layers was held separately and back-extracted in sequence
with the same 50 mL portion of heptane. The pH 2 extraction
removes 95% of the pyrimidones. The pH >10 extraction
removes residual pyrimidones missed in the pH 2 extraction.
The aqueous layers were discarded. The organic layers were
combined and taken on to the next step.
anh. sodium carbonate (mw 105.99; 28.8 g, 0.272 mol, 1.05
equiv), anh. piperazine (mw 86.14; 132.0 g, 1.532 mol, 5.93
equiv), and 30 mL of octane (to wash down piperazine that
sublimed in condenser) were added. The slurry was heated
at 133-135 °C for 21 h, at which time the reaction was
judged complete by TLC (eluant: 35% ethyl acetate/
cyclohexane; Rf [4u] ) 0.00; Rf [4d] ) 0.45; Rf [3u] ) 0.68)
and LC.21 The normalized ratio of 3u:4u:4d by LC was
0.002:97.1:2.897, from which it was calculated that k4u/k4d
(the rate constant for formation of 4u divided by the rate
constant for formation of 4d) ) 1.56. The slurry was cooled
to 100 °C, and 100 mL of toluene, 250 mL of water, and 15
mL of 50% sodium hydroxide were added. The sodium
hydroxide is added to convert sodium bicarbonate to sodium
carbonate. Sodium carbonate is more soluble than sodium
bicarbonate, thus helping to minimize the volume. The higher
basicity of sodium carbonate is also advantageous in extract-
ing trace amounts of pyrimidones. The pot temperature was
kept at g85 °C during the extractions. The homogeneous
phases were separated: organic layer 1A, aqueous layer 1B.
The aqueous layer (1B) was extracted with 70 mL of toluene/
50 mL of water: organic layer 1C, aqueous layer discarded.
To the organic layer 1A was added 250 mL of water and
concd hydrochloric acid (mw 36.46; 37.6 wt %; d 1.1885;
47 mL, 21.00 g, 0.576 mol, 2.23 equiv). The pot temperature
was adjusted to g85 °C, and the pH was adjusted to 3.8-
4.0 with concd hydrochloric acid. Toluene (150 mL) was
Generation and Isolation of 6-Piperazinyl-2,4-bis-
pyrrolidinylpyrimidine (4u). The combined organic layers
from the previous step (containing theoretically 0.259 mol
of 3u) were concentrated under a vacuum to a low volume.
The vacuum was released, and 90 mL of Isopar H were
added. The concentration was continued at a pot temperature
of 100 °C at 176 mm vacuum. After releasing the vacuum,
(20) LC assay: column, DuPont Zorbax C8, 25 cm × 4.6 mm; temperature,
ambient; flow rate, 2.9 mL/min; detection, 254 nm; mobile phase, 50%A/
20%B/30%C where Reservoir A contains 1 L of methanol/1 mL of
triethylamine/0.5 mL of acetic acid, Reservoir B contains 1 L of acetonitrile,
and Reservoir C contains 1 L of water/1 mL of triethylamine/0.5 mL of
acetic acid. Retention times and response factors (amount/area): 4,6-bis-
pyrrolidinyl-2-pyrimidone (6s), 2.21 min and 1.3456 × 10-7; 4,6-bis-
pyrrolidinyl-2-chloropyrimidine (3s), 3.36 min and 2.1855 × 10-8; 2,4-
bis-pyrrolidinyl-6-pyrimidone (6u), 5.57 min and 5.2339 × 10-8; 6-chloro-
(21) LC assay: column, DuPont Zorbax C8, 25 cm × 4.6 mm; temperature,
ambient; flow rate, 4 mL/min; detection, 230 nm; mobile phase, 400 mL of
acetonitrile/400 mL of methanol/200 mL of water/1 mL of triethylamine/
0.05 mL acetic acid. Retention times and response factors (amount/area):
6-chloro-2,4-bis-pyrrolidinylpyrimidine (3u), 2.11 min and 1.8849 × 10-9
;
6-piperazinyl-2,4-bis-pyrrolidinylpyrimidine (4u), 6.7 min and 1.3757 ×
2,4-bis-pyrrolidinylpyrimidine (3u), 7.04 min and 2.5059 × 10-8
.
10-9; 2:1 adduct (4d), 10.8 min and 7.9302 × 10-10
.
Vol. 10, No. 5, 2006 / Organic Process Research & Development
•
925

added to the thin slurry until all the 2:1 adduct 4d dissolved.
The phases were separated: organic layer 1D, aqueous layer
1E. 1C and 1E were combined, and the pH was adjusted to
3.8-4.0 by the addition of concd hydrochloric acid. The
phases were not separated: 1F. To 1D was added 50 mL of
water, and the pH was adjusted to 3.8-4.0. The organic layer
was discarded, and the aqueous layer was combined with
1F. The phases were separated: organic layer 1G, aqueous
layer 1H. 1G was combined with 50 mL of water, and the
pH was adjusted to 3.8-4.0 with hydrochloric acid. The
organic layer was discarded, and the aqueous layer was
combined with 1H: 1I. The combined aqueous layers (1I)
were cooled to ∼50 °C, and 50% sodium hydroxide (mw
40.00; d 1.53; 30 mL, 22.95 g, 0.574 mol, 2.22 equiv) was
added. The pH of the resulting thick slurry was >10. To the
slurry was added 200 mL of toluene and 100 mL of Isopar
H. The pot temperature was kept at g80 °C during the
extractions. The phases were separated, and the organic layer
was extracted with water (4 × 100 mL). Each aqueous layer
was extracted in sequence with 75 mL of toluene/25 mL of
Isopar H. The aqueous layers were discarded, and the organic
layers were combined and concentrated under a vacuum.
Isopar H (315 mL total) was added intermittently as needed
to keep a stirrable slurry. The distillation was continued to
a pot temperature of 100 °C at 176 mm of vacuum. The pot
volume was ∼425 mL. The slurry was cooled to -15 °C,
and the solids were collected. The cake was washed with
300 mL of -15 °C heptane. The cake was dried in an 80 °C
vacuum oven to give a bone white solid identified as
6-piperazinyl-2,4-bis-pyrrolidinylpyrimidine (4u) in pure
form (100.93 wt % relative to a reference standard),
containing no impurity above 0.1%, by LC comparison with
an authentic standard. Weight: 70.04 g (mw 302.43; 0.232
mol, 85.0% yield overall from 2,4,6-trichloropyrimidine). 13C
NMR (75 MHz, CD2Cl2): δ 164.4 (s); 162.9 (s); 160.5 (s);
72.6 (d); 51.7 (t); 45.9 (t); 46.5 (2C, t); 46.4 (2C, t); 45.4
20.5 h. The mixture was found to contain 2u, 3s, 2s, and 3u
in a 1.95:6.06:1.54:90.45 mole ratio by LC analysis.19 The
mixture was quenched with 200 mL of water, diluted with
20 mL of brine, and extracted with ethyl acetate (3 × 100
mL). The organic extracts were back-extracted with the same
100 mL portion of water and then combined and concentrated
to an oil which crystallized on seeding to solids consisting
of a mixture of three components (Rf ) 0.20, 0.28, and 0.44)
by TLC.18 Weight: 13.8 g. A portion of the solids (12.5 g,
from 0.0494 mol trichloropyrimidine) was flash chromato-
graphed on 560 g silica gel (gradient elution, 2% f 30%
ethyl acetate/cyclohexane). The fractions containing the Rf
) 0.44 component were combined, concentrated, diluted with
heptane, and filtered to give a solid (mp 79-82 °C) identified
as 6-chloro-2,4-bis-pyrrolidinylpyrimidine 3u by its spectral
characteristics. By LC, the material was homogeneous except
for 0.10 M% 3s and 0.52 M% 2s. Weight: 11.51 g (mw
252.75; 0.0455 mol, 92.2% yield). 13C NMR (75 MHz,
CDCl3): δ 161.3 (s); 159.7 (s); 158.8 (s); 90.6 (d); 46.5 (3C,
1
t); 46.2 (t); 25.5 (4C, t). H NMR (300 MHz, CDCl3): δ
1.89-1.94 (8H, mults); 3.50-3.55 (8H, mults); 5.63 (1H,
s). MS (CI, CH4): m/e 253/255 (P+ + H, 100%). Anal. Calcd
for C12H17N435Cl: m/e 252.1142. Found: m/e 252.1154.
The fractions containing the Rf ) 0.20 component were
combined and concentrated to a solid (mp 94-96 °C)
identified as 2-chloro-4,6-bis-pyrrolidinylpyrimidine 3s by
its spectral characteristics. By LC, the material was 99.92
M% pure. Weight: 0.616 g (mw 252.75; 0.00244 mol, 4.9%
yield). 13C NMR (75 MHz, CD2Cl2): δ 161.8 (s); 159.6 (s);
78.9 (d); 46.8 (t); 25.6 (t). 1H NMR (300 MHz, CD2Cl2): δ
1.94 (8H, narrow mult); 3.37 (8H, br s); 4.92 (1H, s). Anal.
Calcd for C12H17N435Cl: m/e 252.1142. Found: m/e 252.1152.
Acknowledgment
The authors thank our colleagues Waltraut M. Knoll for
analytical support and Newell D. Taylor and Oseville M.
Smith for engineering support.
1
(t); 44.8 (t); 25.9 (2C, t); 25.7 (2C, t). H NMR (300 MHz,
CD2Cl2): δ 1.85-1.95 (8H, mults); 2.84 (4H, t, J ) 5.1
Hz); 3.36-3.49 (12H, mults); 4.83 (1H, s). MS (CI, CH4):
m/e 303 (P+ + H, 100%).
Synthesis of Reference Standards of 3u and 3s. A
solution of 2,4,6-trichloropyrimidine (mw 183.42; 10.0 g,
0.0545 mol) in 50 mL of methanol was cooled to -17 °C
and treated dropwise with a solution of pyrrolidine (mw
71.12; d 0.852; 20.0 mL, 17.0 g, 0.240 mol, 4.39 equiv) in
65 mL of methanol at a rate such that the pot temperature
remained e-8 °C (7 min). The mixture was stirred at rt for
Supporting Information Available
Experimental procedures; spectroscopic characterization
data; 13C and 1H NMR spectra for all new compounds. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Received for review April 24, 2006.
OP060093Q
926
•
Vol. 10, No. 5, 2006 / Organic Process Research & Development