A concise synthesis of a novel insulin-like growth factor i receptor (IGF-IR) inhibitor

Article abstract of DOI:10.1021/op700052u

An efficient synthesis of a potent insulin-like growth factor I receptor (IGF-IR) inhibitor AEW541 (1) is described. The key step in the synthesis is the cis-selective reductive animation of cyclobutanone, which sets up the desired 1,3-stereochemistry of the cyclobutane ring. The amino group thus generated is used as a handle to build the pyrrolopyrimidine ring. The final step resulting in 1 is accomplished by alkylation of in situ generated mesylate with azetidine.

Full text of DOI:10.1021/op700052u

Organic Process Research & Development 2007, 11, 825–835
A Concise Synthesis of a Novel Insulin-Like Growth Factor I Receptor (IGF-IR)
Inhibitor†
Joel Slade, Joginder Bajwa, Hui Liu, David Parker, James Vivelo, Guang-Pei Chen, John Calienni, Edwin Villhauer,
Kapa Prasad,* Oljan Repicˇ, and Thomas J. Blacklock
Process Research & DeVelopment, NoVartis Pharmaceuticals Corporation, One Health Plaza, East HanoVer,
New Jersey 07936, U.S.A
Abstract:
graphic purifications and/or high-vacuum distillations. It is a
convergent synthesis that was based upon a published proce-
dure.2 Intermediates A and B were prepared in gram quantities
(Schemes 1 and 2) and combined (Scheme 3). Separation of
the cis/trans mixture was accomplished by preparative HPLC,
and the cis isomer was converted to 1.
It was apparent that a new synthesis was necessary for
preparing larger quantities of the drug substance. After a
thorough analysis, we decided to proceed using a strategy that
would provide a suitably substituted cis-1,3-cyclobutane in a
selective manner and then use this key building block to
construct the remainder of the molecule. This approach would
eliminate the highly inefficient and lengthy chromatographic
separation of the cis/trans isomers, which was a major drawback
in the Discovery synthesis (Scheme 3). The approach chosen
for the preparation for 1 is pictured retrosynthetically in Scheme
4.
As is evident from Scheme 5, we predicted that
reductive amination of an appropriately substituted cy-
clobutanone would favor the formation of the desired cis-
3-substituted aminocyclobutane,3 crucial to the success of
this route. Subsequent alkylation of the amino group would
provide the precursor required for the construction of the
pyrrolopyrimidine ring. Finally, the azetidine function
would be introduced by an alkylation method similar to
that used in the Discovery synthesis. The entire new route
is illustrated in Scheme 5.
An efficient synthesis of a potent insulin-like growth factor I
receptor (IGF-IR) inhibitor AEW541 (1) is described. The key step
in the synthesis is the cis-selective reductive amination of cyclobu-
tanone, which sets up the desired 1,3-stereochemistry of the
cyclobutane ring. The amino group thus generated is used as a
handle to build the pyrrolopyrimidine ring. The final step resulting
in 1 is accomplished by alkylation of in situ generated mesylate
with azetidine.
Introduction
The insulin-like growth factor I receptor (IGF-IR) is a
transmembrane tyrosine kinase receptor expressed in a wide
variety of cell types. Binding of its cognate ligands, IGF-I and
IGF-II, to the extracellular domain of the receptor triggers the
activation of its intracellular tyrosine kinase domain and receptor
autophosphorylation. A series of downstream signaling events
are then initiated via phosphorylation of cellular substrates,
which lead to cell proliferation, growth, and survival. Up-
regulated levels of both the receptor and its ligands have been
observed in a variety of human tumors, particularly in associa-
tion with pathological events such as invasion and metastasis.1
Moreover, IGF-IR has been demonstrated experimentally to be
required for anchorage-independent growth and oncogenic
transformation. A research effort within Novartis resulted in
the discovery of AEW541 (1) which potently inhibits (IC50 0.15
µM) the recombinant as well as the native IGF-IR kinase.
cis-Selective Reductive Amination
For the reductive amination of cyclobutanone 2,4 sodium
triacetoxyborohydride was chosen as the reducing agent on the
basis of its availability, ease of handling, and stability. In
addition, recent studies had indicated that this was the reagent
of choice for this type of transformation.5 The results of our
initial investigation are summarized in Table 1. The reaction
involved adding a solution of 2 and 3 in EtOAc to a slurry of
sodium triacetoxyborohydride at 45–50 °C in ethyl acetate. After
4 h, the crude product was isolated in 94% yield as a mixture
of the cis/trans isomers in a ratio of 4.5:1 (Table 1, entry 1).
The Discovery synthesis of AEW541 (1) involved 18
chemical steps (Schemes 1–3) and approximately 10 chromato-
(2) Widler, L.; Green, J.; Missbach, M.; Šuša, M.; Altmann, E. Bioorg.
Med. Chem. Lett. 2001, 11, 849–852.
* To
whom
correspondence
should
be
addressed.
E-mail:
(3) Subsequent to the completion of our work, a report appeared
delineating a similar approach to a 1,3-cis cyclobutane ring; see: Helal,
C. J.; Kang, Z.; Lucas, J.; Bohall, B. R. Org. Lett. 2004, 6, 1853–
1856.
prasad.kapa@novartis.com.
† Dedicated to Prof. Dr. Dieter Seebach on the occasion of his 70th birthday.
(1) (a) For reviews, see: Yu, H.; Rohan, T. J. Natl. Cancer Inst. 2000,
92, 1472–1489. (b) Khandwala, H. M.; McCutcheon, I. E.; Flyvbjerg,
A.; Friend, K. E. Endocrine ReV. 2000, 21, 215–244. (c) Firstenberger,
G.; Sonn, H-G. Lancet Oncol. 2002, 3, 298–302.
(4) The synthesis of this compound will be reported separately.
(5) Abdel-Magid, A. F.; Carson, K.; Harris, B. D.; Maryanoff, C. A.; Shah,
R. D. J. Org. Chem. 1996, 61, 3849–3862.
10.1021/op700052u CCC: $37.00
Published on Web 08/14/2007
2007 American Chemical Society
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
825

Scheme 1. Discovery synthesis of intermediate A
The reaction was also repeated at lower temperatures in an
attempt to improve the selectivity. When the reaction was
carried out at 20–25 °C (Table 1, entry 2) a cis/trans ratio of
5.4:1 was obtained in 84% yield. Acetonitrile and toluene were
also tried as alternate solvents; however, the cis/trans ratios
remained the same.
To separate the cis/trans isomers in the crude product by
nonchromatographic methods, a large number of salts were
prepared and examined. The best of these turned out to be the
hydrochloride salt 4, which was crystallized from an i-PrOH/
EtOAc/H2O solvent mixture to provide the cis isomer in ∼98%
isomeric purity in an overall yield of ∼60% (Table 1, entry 3).
This procedure was scaled up successfully even at higher
concentrations (Table 1, entries 4–6).
Initially, the reaction was carried out with 2.05 equiv of
ammonium formate in ethanol at 40–45 °C. However, this
reaction was very slow, requiring the addition of a total of 5
equiv of ammonium formate. Under these conditions, a
considerable amount of mono-benzyl product (73%) was formed
even after heating to 70–75 °C (Table 2, entry 3). A major
improvement was achieved when a mixture of the free base of
4, n-pentanol, and a 10% loading (by weight) of 10% Pd/C
was heated to 47 °C and then treated with 5 equiv of ammonium
formate in water (Table 2, entry 4). After 4.5 h, the two-phase
reaction was 84% complete, and after 12 h, it was 100%
complete. The amount of the mono-benzyl intermediate was
reduced to less than 1%. With regard to the isolation of 5, we
found that the HCl salt was soluble in n-pentanol. Therefore,
to maximize the isolated yield the preparation of other salts
was investigated. The succinate salt was chosen based on its
ease of formation and poor solubility in n-pentanol.
With the isolation procedure in hand, the reaction was carried
out using a 5% loading of 10% Pd/C (Table 2, entry 5). The
reaction required heating at 40–45 °C for 24 h, and the product
was isolated in 65% yield as the succinate salt. In the following
experiment (Table 2, entry 6), the recovery was increased by
adding heptane to the solution. Product 5 was isolated in 89%
yield and 99% purity. This procedure was repeated on 121 g
of 4 with a 10% loading of catalyst (Table 2, entry 7). After
10 h, the reaction was complete. Work-up and salt formation
resulted in an 83% yield of 5 with 99% HPLC purity.
Hydrogenolysis
Several sets of conditions were studied for debenzylation
of 4. Hydrogenolysis of 4 with 10% Pd/C in ethyl acetate and
acetic acid as solvent at 40–45 °C afforded a slow reaction
(Table 2, entries 1 and 2).6 Instead of attempting to optimize
these conditions, we decided to investigate the transfer-
hydrogenation method using ammonium formate and 10% Pd/C
catalyst.7 The results are summarized in Table 2.
Selective N-Monoalkylation
We had envisioned carrying out the alkylation step on the
amino alcohol 13, which was easily obtained from 5 by basic
hydrolysis. Following a literature procedure,2 amino alcohol 13
was treated with R-bromoketone 6 in ethanol containing 2 equiv
(6) The cleavage of benzylamines with H2/Pd/C is often very slow; see:
Hartung, W. H.; Simonoff, R. Org. React. 1953, VII, 253.
(7) Prasad, K.; Jiang, X.; Slade, J. S.; Clemens, J.; Repicˇ, O.; Blacklock,
T. J. AdV. Synth. Catal. 2005, 347, 1769–1773.
826
•
Vol. 11, No. 5, 2007 / Organic Process Research & Development

Scheme 2. Discovery synthesis of intermediate B
Scheme 3. Final steps in the Discovery synthesis of 1
of diisopropylethylamine at room temperature. HPLC analysis
of the reaction mixture showed a ratio of 1:2.9 of mono 14 and
dialkyl 15 products (Scheme 6).
In addition, we observed that the desired product 14 was
unstable in solution, as the HPLC peak corresponding to it
diminished with time. The nature of the decomposition products
was not investigated further, although on the basis of its mass
spectral data one of the major components appeared to be a
dimer.
To avoid isolation of this unstable compound and to convert
14 directly into the pyrrole 8, the reaction mixture was added
to a solution of sodium salt of malononitrile at elevated
temperature. The desired product 8 was obtained in 30% yield
after chromatography (Scheme 7).
With the idea of combining the two steps, an attempt
was made to optimize the reaction conditions for this
alkylation (13 + 6 f 14). These results are summarized
in Table 3.
With regard to the solvent effects, reaction in methanol,
ethanol, and THF all afforded poor 14:15 ratios even with
the use of 2 equiv of the amine (Table 3, entries 2–4). In
pyridine, N-methylmorpholine, and triethylamine, reaction
between 6 and the amine solvents was observed (Table
3, entry 7). Better product ratios were obtained when the
reaction was carried out in polar aprotic solvents DMF
and NMP (Table 3, entries 8–16).
We found that potassium carbonate was superior to triethy-
lamine and potassium phosphate as a base in DMF (Table 3,
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
827

Scheme 4. Retrosynthetic analysis for the preparation of 1
was added to 7 followed by the slow addition of aqueous KOH.
After heating the reaction mixture for 1 h at reflux, it was
cooled, and water was slowly added to precipitate product 8 in
85% yield with a purity of 98% (Scheme 9).
Synthesis of the Pyrrolo[2,3-d]Pyrimidine Ring System
The initial effort to synthesize 9 was based on a literature
precedent used by the Discovery group (Scheme 10).2 Starting
from compound 8, four steps and chromatographic purifications
were required to prepare the pyrrolopyrimidine 9. Thus, 8 was
heated in triethyl orthoformate (as solvent) in the presence of
catalytic acetic anhydride at 80 °C for 2 h. After the reaction
was complete, the excess orthoformate (bp 146 °C) was distilled
off under vacuum to give crude 17. This compound was treated
with ammonia/methanol at 21 °C for 16 h to afford amidine
18 in 46% yield after the excess ammonia/methanol was
removed under vacuum, and the residue was purified by
chromatography. Amidine 18 was dissolved in ethanol in the
presence of sodium ethoxide and heated at 80 °C for 1 h to
form pyrrolopyrimidine 19. After removal of solvent, the residue
was hydrolyzed with HCl to give crude 9. After chromatogra-
phy, 9 was obtained as a brown solid in ∼65% yield. The
overall yield of 9 from 8 was only 30%.
entries 8, 10, and 12). However, potassium phosphate was
equivalent to potassium carbonate in NMP (Table 3, entries 11
and 13).
As previously mentioned, compound 14 was unstable and
provided only poor yields of the pyrrole 8 regardless of the
conditions followed for its formation. However, we observed
that the free base of the benzoate 5 reacted in a similar fashion
with 6 and, more importantly, provided a stable HCl salt of the
desired product 7 (Scheme 7). Thus, isolation of the free base
13 was not necessary. The reaction was optimized, and the
results are presented in Table 4.
The major drawback of this reaction was that 2 equiv of the
expensive 5 was needed; the excess amine was used as base to
quench the resulting HBr. Fortunately, the excess 5 could be
recovered in 97% purity and recycled. In addition to acetonitrile,
a number of other solvents including ethyl acetate (Table 4,
entry 2), MTBE (Table 4, entry 3), and propionitrile (Table 4,
entry 5) were used but did not offer any advantage over
acetonitrile. In some cases, a considerable amount of the
dialkylated amine side product 16 was formed. Lowering the
reaction temperature to -40 °C (Table 4, entry 6) helped to
suppress the formation of this compound. Use of an external
base (Table 4, entry 7) instead of excess 5 resulted in a low
yield of 7, and the product was contaminated by the salt of the
external base.
We found that the desired pyrrolo[2,3-d]pyrimidine ring
system could be obtained more efficiently by heating 8 in
formamide (as solvent) in the presence of formic acid at 100
°C.13 The reaction was complete in 2–4 h (Scheme 11).
However, under these conditions, formate 20 was formed in
addition to many side products.
To improve the efficiency of the reaction, the formamide/
formic acid was replaced by formamidine acetate, which had
been employed previously in our group in a similar situation.14
This proved to be a better choice. When 8 was treated with 3.5
equiv of formamidine acetate in ethylene glycol (4 mL/g of 8)
at 135 °C for 3–5 h, 9 was formed as the major product
(89–94%; Scheme 12). An intermediate was observed, which
we believed to be amidine 21, as evidenced by LC–MS. This
intermediate was observed only in small amounts in the
beginning of the reaction as it was quickly converted to 9.
The reaction mixture was dark, probably due to some
polymerization of the formamidine. A Celite filtration during
the aqueous workup and a silica gel filtration of the ethyl acetate
solution of the product were needed to remove most of the dark
impurities. The product was obtained in ∼74% yield.
Treatment of 8 with formamidine hydrochloride also pro-
duced 9. However, formamidine hydrochloride is about 15 times
more expensive than formamidine acetate. The reaction with
formamidine hydrochloride provided no advantage over for-
mamidine acetate with respect to impurity formation and ease
of workup.
(12) (a) Gewald, K. Z. Chem. 1961, 1, 349. (b) Abdalla, G. M.; Sowell,
J. W., Sr. J. Heterocycl. Chem. 1987, 24, 297–301. (c) Chen, T.-C.;
Meade, E. A.; Hinkley, J. M.; Townsend, L. B. Org. Lett. 2004, 6,
2857–2859.
The final optimized conditions, (Table 4, entry 8) afforded
compound 7 in 57% yield, and 97.7% purity (HPLC).
(13) (a) Pichler, H.; Folkers, G.; Roth, H. J.; Eger, K. Liebigs Ann. Chem.
1986, 1485–1505. (b) Dave, C. G.; Shah, P. R.; Upadhyaya, S. P.
Ind. J. Chem. 1988, 27B, 778–780.
Construction of the Pyrrole Ring
The preparation of 2-amino-3-cyanopyrroles is well prece-
dented in the literature.12 A solution of malononitrile in methanol
(14) Prasad, K.; Lee, G. T.; Chaudhary, A.; Girgis, M. J.; Streemke, J. W.;
Repicˇ, O. Org. Process Res. DeV. 2003, 7, 723–732.
828
•
Vol. 11, No. 5, 2007 / Organic Process Research & Development

Scheme 5. New synthesis of 1
Table 1. Reductive amination of 2
entry
reaction conditions
cis/trans (%)
yield (%)
94 (crude)
1
1.05 equiv 2, 1.4 equiv NaBH(OAc)3, added 2/3 solution to
borohydride in EtOAc, 5 h, 45 °C
same as entry 1, 20 °C, 16 h
same as entry 2, solvent changed to i-PrOH and HCl added to
form salt
82/18
2
3
77/13
98/2
85 (crude)
61
4
5
6
same as entry 3, scaled up 4×
same as entry 5, double the concentrationa (to optimize
throughput)
98/2
98/2
98/2
64
70
61
same as entry 3, scaled up 8×
a The rate of this reaction was twice as fast as a result of the increase in concentration.
Azetidine Alkylation
The procedure used by Discovery was investigated. It
involved the formation of the tosylate of 9 followed by
displacement of tosylate group with azetidine 12 to give 1. The
tosylation step was problematic since the tosyl chloride had to
be added in several portions, and the reaction was still
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
829

Table 2. Optimization of deprotection conditions
reaction mixture composition
entry
1
reaction conditions
5
mono benzyl
36
4
yield (%)
catalytic hydrogenation; EtOAc, 10% by wt of Pd/C (10%),
16 h, 25 °C
41
16
2
3
4
5
catalytic hydrogenation; AcOH, 20% by wt of Pd/C (10%),
48 h, 25 °C
transfer hydrogenation; EtOH, 20% by wt of Pd/C (10%),
5 equiv HCOONH4, 36 h, 67 °C
transfer hydrogenation; n-pentanol, 10% by wt of Pd/C (10%),
5 equiv HCOONH4, 12 h, 47 °C
transfer hydrogenation; n-pentanol, 5% by wt of 10% Pd/C
(10%), 5 equiv HCOONH4, 12 h, 47 °C. Aqueous work-up,
succinic acid added to crystallize 5
2
73
1
97
2
15
99
92
8
65
6
7
as entry 5, distilled off H2O at 102 °C, added heptane to
increase yield; needed 24 h for complete conversion
as entry 6, 121 g of 4; used 10% by wt of catalyst, 10 h
99
99
<1
<1
89
83
Scheme 6. Alkylation of amino alcohol 13
incomplete after 36 h at 0 °C. At higher temperatures, extensive
decomposition of tosylate was observed.
usually complete in 6 h at 60 °C. A significant amount of
impurity 24 (5%) was also formed as a result of opening the
azetidine ring of 1 with another molecule of azetidine. The
structure of this impurity was confirmed by mass spectroscopy.
Further studies were carried out to determine the effect of
In contrast, the formation of the mesylate proceeded cleanly
even at 0 to -10 °C in THF (Scheme 13). Initially, the mesylate
was isolated prior to the reaction with excess azetidine.
However, we found that the mesylate was unstable on standing
and decomposed into an elimination product 22 and a tricyclic
product 23. The structures of these by-products were tentatively
assigned on the basis of NMR and LC/MS data.
temperature on the formation of this side product. We found
that at 40 °C the formation of this impurity was minimal,
however, the rate of product formation was also slower. At a
reaction temperature of 50 °C, only 1% of the ring-opened
product was present, and the reaction was complete in 16 h.
This led to a change in the procedure wherein the mesylate
was prepared and used directly without isolation. An excess of
the azetidine free base 12 was employed. The reaction was
830
•
Vol. 11, No. 5, 2007 / Organic Process Research & Development

Scheme 7. Preparation of pyrrole 8
Table 3. Attempted monoalkylation of 13 with 6
entry
ratio of 13:6
base (equiv)
solvent
ratio 14:15
1
2
1:1
2:1
(i-Pr)2NEt (2)
Et3N (2)
NaOH (1)
Na2CO3 (2)
(i-Pr)2NEt (2)
DBU (2)
EtOH
MeOH
MeOH
MeOH
THF
1:2.9
1.5:1
3
2:1
4
2:1
1.1:1
1:1.1
complex mixture
5
2:1
6
1:1
THF
7
2:1
pyridine, NMM, Et3N
8
2:1
K2CO3 (2)
DMF
DMF
DMF
NMP
DMF
NMP
NMP
DMF
NMP
9:1
9
3:1
4.6:1
3.3:1
3.4:1
2.9:1
2.8:1
3:1
10
11
12
13
14
15
16
2:1
Et3N (2)
2:1
K3PO4 (2)
K3PO4 (2)
K2CO3 (2)
K2CO3 (4)
K2CO3 (4)
K2CO3 (4)
2:1
2:1
2:1
1.2:1
1.2:1
1.7:1
1:1.9
a No product was observed due to the reaction of 6 with sodium hydroxide. b Compound 6 was unstable in these solvents.
These conditions were chosen for the final process. It is
interesting to note that no elimination product was formed under
these conditions.
For the isolation of the 1, the product was extracted into
aqueous citric acid and the pH adjusted to 10, at which point
the compound crystallized from solution. The compound could
be further purified by recrystallization from ethyl acetate or
isopropyl acetate. Isopropyl acetate was chosen for better
recovery. In the laboratory, the yields were 59–63%, and the
purity of the 1 was 99%.
with azetidine. The yield for this step was 84%, and the azetidine
was found to contain 2.2% of water by Karl Fischer analysis.
Conclusion
We have synthesized 1 in six linear steps starting from
readily available starting materials. The synthesis is efficient
and has been scaled up to prepare 20 kg of 1. An efficient
method for selective cis reductive amination of 3-substituted
cyclobutanone was described. Construction of the pyrrolopy-
rimidine ring system from the corresponding suitably substituted
pyrrole was simplified. The instability of the pure mesylate of
9 was overcome by using the mesylate in situ for azetidine
alkylation.
Azetidine Preparation
Azetidine free base 12 was obtained from the HCl salt 11
according to a literature procedure.15 This was accomplished
by the slow addition of a concentrated solution of the HCl salt
in water to a concentrated solution of KOH in water at 95–
100 °C, at which point the low boiling azetidine distilled out
of the mixture. During this flash distillation, water formed an
azeotrope with azetidine, and as a result the distillate contained
varying amounts of water (up to 11%). On the basis of our
previous experience, we found that the reaction could tolerate
up to 10% water in the azetidine without having a deleterious
effect on the yield or quality of 1. Subsequently, this procedure
was modified. The new process used more KOH and less water
than employed previously. In addition, the reaction temperature
was kept below 85 °C to minimize the co-distillation of water
Experimental Section
General. Melting points were determined on a Thomas-
Hoover melting point apparatus and are uncorrected. NMR
spectra were recorded on a Bruker AVANCE DPX-300
spectrometer (1H NMR at 300 MHz). Analytical high perfor-
mance liquid chromatography (HPLC) was carried out using a
Waters Alliance 2690 Separations Module, a Waters 996
Photodiode Array Detector (MaxPlot), and a 4.6 mm × 25 cm
Waters C18 Symmetry column. Reactions were carried out
under an atmosphere of nitrogen. Residual water content was
determined by Karl Fischer titration. Unless reported otherwise,
all reaction temperatures refer to the measured temperature of
reaction mixtures not to the cooling or heating bath temperatures.
cis-3-Bis(phenylmethyl)aminocyclobutanemethyl Ben-
zoate Hydrochloride 4. Charge a 500-mL, 4-necked, round-
bottomed flask with 24.8 g (0.121 mol) of 2, 22.5 g (0.115
(15) Causey, D. H.; Mays, R. P.; Shamblee, D. A.; Lo, Y. S. Synth.
Commun. 1988, 18, 205–211.
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
831

Table 4. Monoalkylation of 5
entry
1
reaction conditions
product 7 (% HPLC purity)
yielda (%)
28
2 equiv 5, 1 equiv 6, acetonitrile, 20–25 °C, 1 h;
added EtOAc/HCl
2 equiv 5, 1 equiv 6, EtOAc, –15 to –10 °C, 1 h;
added EtOAc/HCl
2 equiv 5, 1 equiv 6, MTBE, –15 to –10 °C, 1 h;
added EtOAc/HCl
98
93
2
3
47
rate of reaction was slow; large amount of
dialkylated material was observed
89.8; 8% dialkyl
4
5
scaled 2×; added 37% HCl
60
47
2 equiv 5, 1 equiv 6, propionitrile, -15 to –10 °C, 1 h;
added 37% HCl
79; 18% dialkyl
6
7
8
2 equiv 5, 1 equiv 6, acetonitrile, -40 to -35 °C, 3 h;
96; 3% 5
55
added 6 N HCl
1.3 equiv 5, 1 equiv 6, acetonitrile, 3 equiv Hunig’s base,
-40 to -35 °C, 3 h; added 6 N HCl
2 equiv 5, 1 equiv 6, acetonitrile, -16 to -10 °C, stirred 1 h; 97.7
95 (material was mainly HCl salt of Hunig’s base)
57
added 3 N HCl
a Based on 6.
Scheme 8. Alkylation of 5
Scheme 9. Preparation of pyrrole 8
mol) of 3,and 101.5 g (112.5 mL) of ethyl acetate. Stir the
homogeneous reaction mixture for 1 h at 20–25 °C. Charge a
2-L, 4-necked, round-bottomed flask with 34.1 g (0.161 mol)
of sodium triacetoxyborohydride and 203.0 g (225 mL) of ethyl
acetate. Stir for 15 min at room temperature. Add the (2 + 3)
solution over 30–40 min. Stir the heterogeneous reaction
mixture at 20–25 °C for 16 h. Add 426.4 g (410 mL) of 2 N
NaOH over 15 min. Stir for 1 h at 20–25 °C. Caution: H2
evolution. Remove the lower aqueous layer and wash the ethyl
acetate solution of 4 with 2 × 300 g of water. Mark a 1-L round-
bottomed flask at 130 mL. Charge the flask with the ethyl
acetate solution of 4 and concentrate to a volume of 130–150
mL, using atmospheric distillation conditions. Cool the residual
solution to 65 °C and add 94.2 g (120 mL) of 2-propanol. Hold
the solution at 65–70 °C. Add 12.0 g (0.121 mol) of hydro-
chloric acid (37 wt %) over 5–10 min. Cool from 65 °C to
43–45 °C. At 45 oC seed the reaction mixture with 20 mg of 4.
Cool from 43–45 °C to 32–33 °C over 0.5 h. Stir at 32–33 °C
for at least 90 min. Cool to 20–25 °C over 0.5 h. Hold 1 h at
20–25 °C. Isolate the solid by filtration and wash the cake with
78.5 g (100 mL) of 2-propanol. Dry the cake on the funnel for
30 min, then dry in a vacuum oven at 60–65 °C (35–50 mm
Hg) for at least 16 h, to yield 4 as a white solid, 29.9 g (62%
yield), 98.9% cis isomer; 1.1% trans isomer by HPLC analysis;
mp 165 °C; 1H NMR (CDCl3) δ 8.0 (d, J ) 9.0 Hz, 2H), 7.55
(m, 13H), 4.35 (d, J ) 9.0 Hz, 2H), 4.07 (m, 4H), 3.31 (m,
1H), 2.86 (m, 2H), 2.20 (m, 3H); 13C NMR (DMSO, d6) δ
165.61, 133.31, 131.43, 130.17, 129.92, 129.51, 129.28, 129.24,
128.70, 128.63, 128.52, 67.49, 551.6, 54.38, 29.70, 26.40;
MS m/z 386 (M+). Anal. Calcd for C26H28ClNO2: C, 74.00; H,
6.69; N, 3.32; Cl, 8.40. Found: C, 74.08; H, 6.65; N, 3.31; Cl,
8.62.
cis-3-Aminocyclobutanemethyl Benzoate Butanedioate
(1:1 Salt) 5. Charge a 1-L round-bottomed flask with 121.0 g
(0.286 mol) of 4 and 405.5 g (500 mL) of 1-pentanol. Stir the
heterogeneous reaction mixture at 20-25 °C. Add 312.0 g (300
mL) of 1 N NaOH. Stir the two-phase reaction mixture for 15
min at 20-25 °C. Heat to 45-50 °C and stir for 5 min. Remove
the lower aqueous layer and cool the organic layer to 20-25
°C. Charge a 2-L, 4-necked, round-bottomed flask with 24.2 g
of 10% Pd/C and add the 1-pentanol solution of the 4 free-
base. Charge a 250-mL round-bottomed flask with 90.2 g (1.43
mol) of ammonium formate and 150.0 g (150 mL) of water.
Stir for 15 min. Heat the reaction mixture containing the free-
base of 4 and 10% Pd/C to 45–50 °C then add the aqueous
ammonium formate solution over 30-40 min. Stir the hetero-
geneous reaction mixture at 45–50 °C for 16 h (possible H2
evolution). When analysis determines completeness of reaction,
cool to 20–25 °C. Filter and wash the cake with 81.1 g (100
mL) of 1-pentanol and 50.0 g (50 mL) of water. Separate the
lower aqueous layer from the filtrate. Wash the organic layer
with 2 × 100 mL of water. Heat to 45–50 °C and stir for 5
min. Remove the lower aqueous layer. Heat to an internal
temperature of 102 °C and collect ∼100 mL of distillate using
832
•
Vol. 11, No. 5, 2007 / Organic Process Research & Development

Scheme 10. Discovery approach to 9
Scheme 11. Alternate preparation of the pyrrolo[2,3-d]pyrimidine ring system
atmospheric distillation conditions. Cool to an internal temper-
ature 30 °C. Add 33.8 g (0.29 mol) of succinic acid. Heat the
organic layer to 95–100 °C to obtain a homogeneous reaction
mixture. Cool from 95–100 to 20–25 °C over 2 h. Stir at 20–25
°C for at least 90 min. Add 171.0 g (250 mL) of heptane over
30 min and stir for 1.5 h at 20–25 °C. Isolate the solid by
filtration and wash the cake with 112.1 g (150 mL) of 1:1
mixture of heptane and 1-pentanol. Dry the cake on the funnel
for 30 min and then dry in a vacuum oven at 55–60 °C (35–50
mm Hg) for at least 16 h to yield 77.5 g of 5 as a white solid
the two-phase reaction mixture for 10 min. Remove the lower
aqueous layer and wash the upper organic layer with 200 mL
of water. Concentrate the organic phase to a volume of 180–220
mL using a maximum internal temperature of 40–45 °C at
100-150 mm Hg vacuum. Add 314.4 g (400 mL) of acetoni-
trile and concentrate the mixture to a volume of 180–220 mL
using a maximum internal temperature of 40–45 °C and 80–150
mm Hg vacuum. Add 314.4 g (400 mL) of acetonitrile and
concentrate the mixture to a volume of 180–220 mL using a
maximum internal temperature of 40–45 °C and 80–150 mm
Hg vacuum. Charge a 2-L, 4-necked, round-bottomed flask with
the free-base solution of 5 and 216.2 g (275 mL) of acetonitrile.
Stir the reaction mixture for 10 min at 20–
25 °C to obtain a solution. Cool to –16 to –10 °C. Charge a
250-mL round-bottomed flask with 47.3 g (0.155 mol) of 6
and 149.3 g (190 mL) of acetonitrile and stir for 15 min to
obtain a solution. Add the acetonitrile solution of 6 to the flask
containing the solution of 5 over 30–40 min. The temperature
is kept at –16 to –10 °C by controlling the rate of addition. Stir
the homogeneous reaction mixture at –16 to –10 °C for 30–40
min. Add 77 mL (0.23 mol) of 3 N hydrochloric acid over 10
min and stir for 1 h at –16 to –10 °C. Isolate the solid by
filtration and wash the cake with 235.8 g (300 mL) of
1
(80% yield); mp 165–170 °C; H NMR (DMSO, d6) δ 9.96
(br s, 4H), 8.00 (m, 2H), 7.68 (m, 1H), 7.54 (m, 2H), 4.25 (d,
2H), 3.55 (m, 1H), 2.45 (m, 3H), 2.33 (s, 4H), 1.92 (m, 2H);
13C NMR (DMSO, d6) δ 175.08, 165.72, 133.34, 129.63,
129.20, 128.72, 67.39, 41.32, 31.44, 30.70, 27.13; MS m/z 206
(M+). Anal. Calcd for C16H24NO6: C, 59.44; H, 7.48; N, 4.33.
Found: C, 59.46; H, 7.21; N, 4.29.
2-[cis-3-[(Benzoyloxy)methyl]cyclobutyl]amino]-1-[3-(phe-
nylmethoxy)phenylethanone Hydrochloride 7. Charge a 1-L
round-bottomed flask with 101.7 g (0.31 mol) of 5, 451.0 g
(500 mL) of ethyl acetate and 200 g (200 mL) of water. Stir
the heterogeneous reaction mixture at 22 ( 5 °C. Add 54.6 g
(0.68 mol) of 50% NaOH and rinse with 20 mL of water. Stir
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
833

Scheme 12. Formamidine approach to 9
Scheme 13. Azetidine alkylation
1
acetonitrile. Dry the cake on the funnel for 30 min, then dry in
an oven at 20–25 °C (35–50 mm Hg) for at least 16 h, to yield
an 40.3 g of 7 as an off-white solid (56% yield based on 6);
mp 203 °C (dec), 97% HPLC purity (area normalization); 1H
NMR (DMSO, d6) δ 9.50 (br s, 2H), 7.98 (m, 2H), 7.45 (m,
12H), 5.22 (s, 2H), 4.70 (s, 2H), 4.31 (m, 2H), 3.76 (m, 1H),
2.33 (m, 4H); 13C NMR (DMSO, d6) δ 192.12, 165.72, 158.55,
136.57, 134.94, 133.35, 130.22, 129.62, 129.24, 128.73, 128.46,
127.97, 127.78, 121.16, 120.81, 114.02, 69.57, 67.37, 50.09,
47.75, 28.78, 26.78; HRMS calcd for C27H28NO4 (M + 1)
430.1940, found 430.1930.
2-Amino-1-[cis-3-(hydroxymethyl)cyclobutyl]-4-[3-(phe-
nylmethoxy)phenyl]-1H-pyrrole-3-carbonitrile 8. Charge a
nitrogen-flushed 2-L, 4-necked round-bottomed flask with
46.6 g of compound 7. Add a solution of 13.2 g of malononitrile
in 369 g (466 mL) of methanol. Stir the mixture at 22 ( 3 °C.
Add a solution of 16.8 g of potassium hydroxide in 84 mL of
water over a period of 5 min. Heat the reaction mixture to reflux
(72 ( 3 °C) over 20 min and maintain this temperature for 30
min. Cool the reaction to 32 ( 3 °C over 15 min. Add 932 mL
of water over 30 min and stir the resulting suspension at this
temperature for an additional 30 min. Filter the solids through
a polypropylene filter pad and wash the solids twice with (2 ×
100 g) of deionized water (22 ( 3 °C). Dry the solids at 45 °C
and 180 mbar for 18 h to afford 8 as a dark orange crystalline
solid (94% yield); mp 128 °C, 96% HPLC purity (area
normalization); H NMR (DMSO, d6) δ 7.37–7.51 (m, 4H),
7.34 (m, 1H), 7.17–7.29 (m, 3H), 6.89 (s, 1H), 6.85 (dd, J )
8.2 Hz, 2.5 Hz, 1H), 5.91 (s, 2H), 5.11 (s, 2H), 4.60 (s, 1H),
4.43 (quintet, J ) 8.4 Hz, 1H), 3.44 (s, 2H), 2.40 (m, 2H),
1.98-2.18 (m, 3H); 13C NMR (DMSO, d6) δ 158.6,148.4, 137.1,
135.4, 129.6, 128.5, 127.9, 127.8, 120.8, 118.6, 117.6, 112.1,
111.6, 109.5, 69.2, 68.7, 64.1, 44.7, 32.2, 29.9; HRMS calcd
for C23H24N3O2 (M + 1) 374.1869, found 374.1852. Anal.
Calcd for C23H23N3O2: C, 73.98; H, 6.21; N, 11.25. Found: C,
73.48; H, 6.69; N, 11.09.
3-[4-Amino-5-[3-(phenylmethoxy)phenyl]-7H-pyrrolo[2,3-
d]pyrimidin-7-yl]-cis-cyclobutanemethanol 9. Charge a 1-L,
4-necked round-bottomed flask with 74.0 g of 8, 72.2 g of
formamidine acetate, and 296 mL of ethylene glycol. Heat the
mixture to 120 (>3 °C over ∼2 h and maintain this temper-
ature for 3 h. A small quantity of ammonia evolves at above
80 °C as indicated by wet pH paper, and the suspension turns
to a dark solution when the temperature reaches ∼100 °C. Heat
to 135 ( 3 °C over ∼0.5 h and maintain this temperature for
4 h. Charge a 5-L, 4-necked round-bottomed flask with 1.6 L
of ethyl acetate. Add the reaction mixture with stirring. Rinse
with 25 mL (27.8 g) of ethylene glycol followed by 190 mL of
ethyl acetate. Charge 590 g of deionized water over a period
of 20 min and stir for 10 min. Stop stirring. Let the mixture
settle for ∼15 min. Separate the bottom layer (aqueous phase).
Restart the stirring and add 30 g of Hyflo Super Gel. After 5–10
834
•
Vol. 11, No. 5, 2007 / Organic Process Research & Development

min of stirring, filter the suspension using a Buchner funnel.
Wash the filter cake with 190 mL of ethyl acetate. Add the
filtrate back to the 5-L flask. Add 590 mL of deionized water
with stirring. Separate the bottom layer. Wash the organic phase
with 2 × 590 mL of deionized water and 590 mL of saturated
sodium chloride solution. To an 4.4 cm (i.d.) glass column,
charge 20 g of sand to fill in the conical bottom. Charge 52 g
of Silica Gel 60 to make a 5.5-cm bed. Cover the bed with
cotton. Stir the reaction mixture and add 37 g of anhydrous
magnesium sulfate. Stir for 5 min. Pour the batch onto the silica
gel bed. Set the nitrogen pressure at 10–65 mm Hg so that the
filtration time is about 2.5 h (10–14 mL/min). Wash the silica
with 2 × 190-mL portions of ethyl acetate under the same
pressure. Briefly blow the cake dry with nitrogen. Transfer the
solution in three portions into a 2-L, 4-necked round-bottomed
flask. Distill the combined solution at 20–35 °C (20–57 °C bath
temperature) at ∼0.2 bar over ∼2 h until the volume is ∼330
mL. Use 200 mL of tetrahydrofuran to dissolve the precipitated
product when needed and add to the solution for distillation.
Cool the batch to 22 ( 3 °C. Charge 260 mL of n-heptane
over ∼2 h. Stir the mixture for a minimum of 4 h at this
temperature. Filter through a 600-mL Buchner funnel. Wash
the cake with 2 × 50 mL (2 × 34 g) of n-heptane. Discard the
mother liquor and washes. Air-dry the product on the filtration
funnel for ∼1 h. Dry the product in a vacuum oven at ∼50
°C/12 mm Hg for 16 h to afford 79.4 g (68%) of 9 as a greenish
solid; mp 140 °C, HPLC purity 98.7% (area normalization);
1H NMR (CDCl3) δ 8.33 (s, 1H), 7.48–7.35 (m, 6), 7.12 (s,
1H), 7.10–7.08 (m, 2H), 7.02–7.00 (m, 1H), 5.16 (s, 2H),
5.14–5.08 (m, 3H), 3.76 (d, J ) 4.4 Hz, 1H), 2.98 (br s, 1H),
2.70-2.56 (m, 4H), 2.54–2.45 (m, 1H), 1.86 (br s, 1H); 13C NMR
(CDCl3) δ 159.1, 157.0, 151.5, 150.2, 136.8, 136.3, 130.2,
128.7, 128.1, 127.4, 121.4, 121.1, 116.2, 115.0, 113.9, 101.4,
69.9, 65.2, 45.8, 32.2, 29.9. Anal. Calcd for C24H24N4O2: C,
71.98; H, 6.04; N, 13.99. Found: C, 71.69; H, 6.13; N, 13.87.
7-[cis-3-(1-Azetidinylmethyl)cyclobutyl]-5-[3-(phenylmeth-
oxy)phenyl]-7H-pyrrolo[2,3-d]-pyrimidin-4-amine 1. Flush
a 2-L, 4-necked round-bottomed flask with nitrogen and add
50 g (0.125 mol) of compound 9 and 500 mL of dry
tetrahydrofuran. Stir the mixture at a batch temperature of -20
to -30 °C and add 21.8 mL of triethylamine. Add 10.6 mL of
methanesulfonyl chloride while maintaining the temperature at
or below -20 °C. Stir for an additional 30 min and filter to
remove the triethylamine hydrochloride. Wash the filter cake
with 100 mL of dry tetrahydrofuran and combine the wash with
the filtrate. Add the combined filtrates to a 2-L, 4-necked round-
bottomed flask. Begin a very slight nitrogen sweep so as not to
displace azetidine from the reaction mixture. Add 54.5 mL (0.75
mol) of azetidine 12 over 1 to 2 min and heat to 50 °C. Maintain
this temperature for at least 12 h and then cool to 15 °C. Add
640 mL of 10% aqueous citric acid solution over 1–2 min to
pH ∼5.5. There will be an exotherm (15–29 °C). Concentrate
at 100 mbar (jacket T ) 50 °C) to remove tetrahydrofuran.
Extract the concentrate with 2 × 250 mL of ethyl acetate.
Concentrate at 100 mbar (jacket T ) 50 °C) to remove traces
of ethyl acetate. Flush the flask with nitrogen and return the
solution (volume, ∼450 mL). Stir at 30 °C and add 90 mL of
absolute ethanol. The pH is ∼5.4. Adjust the pH by the
dropwise addition of 6 N sodium hydroxide solution to the point
where the solution turns cloudy, usually near pH ) 8 (volume
of added base, ∼85 mL). Add seed crystals of pure 1 and stir
for 30 min or until crystallization is observed. Increase the
stirring rate and continue basification to pH ) 10.0 (total volume
of base, ∼108 mL). Filter the solids then wash with 100 mL of
water. Dry the solids in a vacuum oven at 45 °C and 50 mbar
to afford crude 1 (51.03 g, 93% yield) as a tan-colored solid.
Recrystallization. Charge 75.0 g of crude 1 and 1125 mL
of isopropyl acetate to a 3-L, 4-necked round-bottomed flask.
Heat to reflux (90 °C) and hold for 1 h. Cool to 70 °C and
filter the solution through Filter Cel. Wash the filter cake with
100 mL of isopropyl acetate and combine the filtrates. Con-
centrate the combined filtrates to a volume of ∼500 mL on a
rotary evaporator. Transfer the concentrate (which may contain
some precipitated solid) to a clean 2-L, 4-necked round-
bottomed flask and add fresh isopropyl acetate (250 mL or as
much as needed to reach a volume of 750 mL). Heat the solution
at reflux and hold for 15 min. Cool the solution to 22 °C over
0.5 h and stir for a minimum of 3 h. Filter the resulting solids
and wash twice with a total of 300 mL of isopropyl acetate.
Dry the solids at 45 °C and 50 mbar for 18 h to afford 44 g of
1 as an off-white solid (59% overall); mp 145 °C; HPLC purity
98.7% (area normalization); 1H NMR (CDCl3) δ 8.17 (s, 1H),
7.14–7.39 (m, 6H), 6.93–7.06 (m, 6H), 6.86 (m, 1H), 5.37 (br
s, 2H), 5.05 (m, 1H), 5.00 (s, 2H), 3.08 (t, J ) 7.0 Hz, 4H),
2.57 (m, 2H), 2.41 (d, J ) 6.2 Hz, 2H), 1.83-2.17 (m, 5H); 13C
NMR (CDCl3) δ 159.2, 157.2, 152.0, 150.6, 136.9, 136.5, 130.3,
128.8, 128.1, 127.5, 121.4, 120.0, 116.5, 115.1, 113.9, 101.1,
70.0, 65.8, 55.9, 45.3, 36.0, 27.4, 18.2; HRMS calcd for
C27H29N5O (M + 1) 374.1869, found 374.1852.
Azetidine 12. Charge a 1-L Erlenmeyer flask with 374.2 g
(4.0 mol) of compound (11) and 280 g (280 mL) of water. Stir
at 20 ( 3 °C until the solid is dissolved and a solution results.
Charge a 2-L, 4-necked round-bottomed flask with 230 g (230
mL) of water. Stir and carefully add in portions 459.4 g (6.96
mol) of potassium hydroxide. After the exotherm from KOH
addition subsides, warm the solution to 95 °C. Hold the
temperature at 95 ( 5 °C. Charge the solution of azetidine HCl
(11) in water to the hot KOH / water solution. Control the rate
of addition so that the product steadily but slowly distills. Collect
the product (12) in a cooled receiver. Complete the addition in
3 h. Stir the reaction mixture at 95 ( 5 °C until the product
stops distilling. Store the product 12 cold. Yield: 203.7 g (89%),
bp 61–62 °C.
Received for review March 1, 2007.
OP700052U
Vol. 11, No. 5, 2007 / Organic Process Research & Development
•
835