Concise synthesis of a selective α1-adrenoceptor antagonist

Article abstract of DOI:10.1021/op050122h

An efficient synthesis of an adrenoceptor antagonist has been developed and demonstrated in a pilot plant. A linear synthesis that relied on a catalytic reduction of a rather insoluble nitroaromatic proved to be a viable route. The active pharmaceutical ingredient (API) that contained an amidine functional group was generated from the amino-containing precursor by activation of dimethylacetamide (DMA) with phosphorus oxychloride (POCl₃). The reaction between DMA and POCl₃ was studied using ReactIR and was found to be a fast but not instantaneous reaction. The iminium salt generated from DMA and POCl₃ had acceptable stability to allow for its use on a pilot-plant scale; however, a trend towards decomposition was revealed on the basis of in situ FTIR data. Formation of the complex was evaluated in a reaction calorimeter (RC-1), and the stability of the complex was probed with an Advanced Reactive System Screening Tool (ARSST).

Full text of DOI:10.1021/op050122h

Organic Process Research & Development 2006, 10, 391−397
Full Papers
Concise Synthesis of a Selective r1-Adrenoceptor Antagonist
Terrence J. Connolly,* Michael Matchett, Patrick McGarry, Sunil Sukhtankar, and Jiang Zhu
STS Chemical SerVices, Chemical DeVelopment, Roche Palo Alto LLC, 3431 HillView AVenue,
Palo Alto, California 94304, U.S.A.
Scheme 1
Abstract:
An efficient synthesis of an adrenoceptor antagonist has been
developed and demonstrated in a pilot plant. A linear synthesis
that relied on a catalytic reduction of a rather insoluble
nitroaromatic proved to be a viable route. The active pharma-
ceutical ingredient (API) that contained an amidine functional
group was generated from the amino-containing precursor by
activation of dimethylacetamide (DMA) with phosphorus oxy-
chloride (POCl3). The reaction between DMA and POCl3 was
studied using ReactIR and was found to be a fast but not
instantaneous reaction. The iminium salt generated from DMA
and POCl3 had acceptable stability to allow for its use on a
pilot-plant scale; however, a trend towards decomposition was
revealed on the basis of in situ FTIR data. Formation of the
complex was evaluated in a reaction calorimeter (RC-1), and
the stability of the complex was probed with an Advanced
Reactive System Screening Tool (ARSST).
symptoms included antagonists displaying selectivity for R1A
and R1B adrenoceptor subtypes.5-7 Compound 1 was recently
selected as a potential clinical candidate, and larger quantities
of 1 were required to support preclinical studies.
Results and Discussion
The key step in the discovery synthesis is shown in
Scheme 1. This route relied on generating two late-stage
intermediates and coupling them in the final step to generate
1. Although this approach allowed for significant flexibility
during the discovery process, it was not suitable for the
preparation of large amounts of 1 due to the synthesis and
physical properties of 3 (Scheme 2).
Introduction
Benign prostatic hyperplasia (BPH) is a male urological
disorder characterized anatomically by enlargement of the
prostate gland due to cellular proliferation of prostatic tissue.
The symptoms of BPH can be classified as obstructive
(hesitancy, low flow, high residual volume) or irritative
(frequency, urgency, nocturia, reduced bladder capacity).
Clinical studies have shown R-1 adrenoceptor antagonists
to be effective in relieving both the obstructive and irritative
symptoms associated with BPH.1-3 Although subtypes of the
R-1 adrenergic receptors are known (R1A, R1B, and R1D),4
the alpha-1 blockers used to treat BPH currently are
nonsubtype selective and have the potential to cause car-
diovascular side effects such as postural hypotension and
dizziness. Clinical doses can be titrated against the side
effects although efficacy may be sacrificed.
The synthesis of 3 required several steps to assemble the
amidine functional group in the presence of the benzylic
amine. The sequence started with 4 and involved sequential
protection of the amine with tert-butylcarbonyl anhydride,
catalytic hydrogenation of the nitro group, condensation with
dimethylacetamide dimethylacetal (DMADMA), and removal
of the nitrogen protecting group. The absence of crystalline
or solid intermediates in the sequence used to prepare 3
would require telescoping the API synthesis from 4 to 1.
This route was not viewed as having any potential for a large-
scale synthesis. The expected result of such extensive
processing was a final product of unacceptable quality.
Once 1 was identified as a clinical candidate, a diversity-
driven synthesis was no longer important, and an efficient
route to 1 was essential. Regardless of the route to 1, a
method to prepare 2 was required and is shown in Scheme
3. Quinazolinedione 58 was converted to the dichloride 6
Efforts at Roche Palo Alto to develop new chemical
entities (NCE) for treating irritative and obstructive BPH
* Correspondence should be addressed to this author at current address:
Department of Process Chemistry, Celera, 180 Kimball Way, South San
Francisco, CA, 94080, U.S.A. E-mail: t_con2002@yahoo.com.
(1) Djavan, B.; Marberger, M. Eur. Urol. 1999, 36, 1-13.
(2) Kumar, V. L.; Dewan, S. Int. Urol. Nephrol. 2000, 32, 67-71.
(3) Michelotti, G. A.; Price, D. T.; Schwinn, D. H. Pharmacol Ther. 2000, 88,
281-309.
(4) Hieble, J. P.; Bylund, D.; Clarke, D. E.; Eikenburg, D. C.; Langer, S. Z.;
Lefkowitz, R. J.; Minneman, K. P.; Ruffolo, R. R. Pharmacol. ReV. 1995,
47, 267-270.
(5) Becker, C. K.; Caroon, J. M.; Melville, C. R.; Padilla, F.; Pfister, J. R.;
Zhang, X. WO 02/053558, 2005.
(6) Chin, E.; Cournoyer, R. L.; Keitz, P. F.; Lee, E. K.; Lopez-Apia, F. J.;
Melville, C. R.; Padilla, F.; Weinhardt, K. K. WO 2005005397, 2005.
(7) Connolly, T. J.; Keitz, P. F.; Lee, E. K.; Li, J.; Lopez-Apia, F. J.; McGarry,
P. F.; Melville, C. R.; Nitzan, D.; O’Yang, C.; Padilla, F.; Weinhardt, K.
K. WO 2005005395, 2005.
(8) Connolly, T. J.; McGarry, P.; Sukhtankar, S. Green Chem. 2005, 7, 586-
589 2005.
10.1021/op050122h CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/04/2006
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
391

Scheme 2
Scheme 3a
structure 9 (see Figure 1) which could have resulted from
further reaction of the product (7) with 2. Alternatively, 9
could have formed from 10, which may have formed during
the preparation of 2.13
The low solubility of 7 in 2-propanol/water was a factor
in achieving a high yield in the coupling stage. However, it
was soon discovered that 7 had limited solubility in many
solvents, which complicated the development of a catalytic
hydrogenation sequence. Adding base to the alcohol/water
mixtures used as solvents did not increase the solubility of
7 through ionization. Transfer hydrogenation14 was briefly
explored, and the initial results were promising. Compound
7 was reduced to 8 in ethanol/water in the presence of
palladium on carbon, triethylamine, and formic acid. Product
8 was not completely soluble in the reaction mixture but
could be dissolved by lowering the pH to 1-2 with HCl
when the reduction was complete. The heterogeneous catalyst
was removed by filtration, and product 8 was precipitated
from the filtrate by adding sodium hydroxide until the pH
was neutral. As the reaction was scaled up, the extreme
heterogeneous nature15 of the reaction mixture caused pH
fluctuations that impacted the extent of reduction. Although
addition of more formic acid and base eventually caused
complete conversion to 8, salts sometimes precipitated with
product 8. The unpredictable nature of the reaction and
inconsistent quality of the penultimate product were viewed
as process liabilities, and this route was not developed further.
Eventually, standard catalytic hydrogenation conditions
were developed using a rather unconventional solvent system
of THF (10 L/kg of 7), methanol (5 L/kg of 7), and aqueous
NaOH (1 equiv). Under these conditions, 7 dissolved
completely and was smoothly reduced with hydrogen and a
palladium on carbon catalyst. Hydrogenolysis of the product
was not observed, even under grossly extended reaction
times. When the reduction was complete, the catalyst was
removed by filtration, and the THF was replaced with
methanol. Water was added, and 8 was precipitated by
adjusting the pH to 5-6 with acetic acid. Precipitation of 8
from methanol/water at 50-55 °C resulted in a product with
superior filtration characteristics. The reduction was complete
a Reagents and conditions: a) POCl3, ACN, H2O; b) KOH, THF, HOAc.
Figure 1.
with phosphorus oxychloride (4 equiv) in acetonitrile. The
reaction required 24 h to go to completion, even when
conducted at reflux in acetonitrile. The reaction mixture was
quenched into water, and the product precipitated in a loss-
on-drying (LOD)-corrected yield of 87%.9 Water-wet 6 was
hydrolyzed with KOH in THF using conditions described
previously for a related compound,10 and chloroquinazolinone
2 was generated in near quantitative yield with a purity of
97% by area normalized HPLC (AN HPLC) analysis.
The most direct route from 2 to 1 appeared to be through
coupling of 2 and 4, followed by reduction of the nitro group
and conversion to the amidine (Scheme 4). Coupling of 2
and 4 (1.15 equiv as the hydrochloride salt)11 occurred readily
in 2-propanol in the presence of diisopropylethylamine (2.5
equiv). The reaction was performed in 12 volumes (L/kg of
2) of 2-propanol and was complete within a few hours under
reflux. The mixture remained heterogeneous throughout the
reaction. When the coupling was complete, water (3 L/kg
of 2) was added to dissolve diisopropylethylamine hydro-
chloride. Addition of water while the reaction mixture was
at reflux followed by a gradual cooling period rather than
cooling the mixture and adding water led to better filtration
rates of 7. Product 7 was isolated in 94% yield with a purity
of 97%, and entrainment of diisopropylethylamine hydro-
chloride in the product cake was not observed.12 The only
impurity detected had a molecular mass consistent with
(12) When triethylamine was employed as the base in the coupling reaction of
2 and 4 and the same reaction conditions were used, the product cake did
contain triethylamine hydrochloride. Increasing the water charge was not
examined since throughput was already starting to suffer at 15 L/kg 4.
(13) The presence of 10 in 2 was not verified.
(9) Eleven kilograms of product with an LOD of 38% was isolated. Drying
trials indicated that decomposition of the product would occur when the
water-wet cake was dried at elevated temperatures.
(14) Johnstone, R. A. W.; Wilby, A. H.; Entwistle, I. D. Chem. ReV. 1985, 85,
129-170.
(15) Although all hydrogenations involving a metal catalyst on an insoluble
support are heterogeneous, this mixture was termed “extreme” because
neither the starting material nor the product was entirely soluble in the
reaction mixture. The reduction process involved partial solubility of the
starting material and precipitation of the product.
(10) Connolly, T. J.; Matchett, M.; Sarma, K. Org. Process Res. DeV. 2005, 9,
80-87.
(11) Connolly, T. J.; Constantinescu, A.; Lane, T. S.; Matchett, M.; McGarry,
P.; Paperna, M. Org. Process Res. DeV. 2005, 9, 837-842.
392
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

Scheme 4a
a Reagents and conditions: a) DIPEA, IPA, H2O; b) H2, Pd/C, NaOH, MeOH, THF, HOAc; c)POCl3, DMA, ACN, DCM, IPA.
Figure 2.
within 2-3 h, and 8 was isolated in 90% yield on a 0.5 kg
scale and 96% on a 5.1 kg scale. In both cases, product purity
was excellent (>99%) as judged by AN HPLC.
Several conditions to install the amidine functional group
were investigated. Heating 8 in alcohol or ether solvents with
DMADMA (5 equiv) was investigated and gave less than
ideal results. In addition to forming the desired product 1,
impurities 11, 12, and 13 were formed (Figure 2). Com-
pounds 11 and 12 were generated through O-methylation
by DMADMA whereas 13 was formed via an alternative
reaction pathway from intermediate 14.16 O-Methylation
Figure 3. Conversion of 8 to 1 with POCl3-DMA complex.
presumably occurred through base-mediated tautomization
of the quinazolinone moiety that resulted from the presence
of DMADMA.
Tf2O was used. However, activation of DMA appeared to
be fast in acetonitrile and the resulting complex appeared to
be relatively stable. Parameter-ranging experiments demon-
strated that the reaction required only slightly more than 1.5
equiv of the iminium salt to consume all of the starting
material (Figure 3). However, when the complex (1.5 equiv)
was aged at room temperature for 48 h prior to adding 8,
11% starting material remained, indicating complex instabil-
ity. Repeating this experiment with 2.5 equiv of the complex
led to complete conversion of 8 to 1. Reaction stress tests
with 2.5 equiv of the complex indicated that the process
would be robust to long addition and reaction times, and
the only byproduct formed was 15 (See Figure 2). Extending
the reaction time to 5 h led to 2% (AN HPLC) of 15, and
after 5 days at ambient temperature, only 4% (AN HPLC)
of 15 was formed. When the reaction was complete, it was
quenched by addition of water and NaOH and by separation
We next examined Dossena’s conditions, who has shown
that the amidine functional group can be installed through
activation of dimethylacetamide (DMA) with trifluo-
romethansulfonic anhydride (Tf2O).17 When Tf2O was added
to a methylene chloride solution of DMA and the resulting
mixture was added to 8 in methylene chloride, 1 was the
only product formed and was isolated in 98% yield.
The relatively high cost of Tf2O led us to consider other
activating agents, and phosphorus oxychloride (POCl3) was
found to be an effective replacement.18 Formation of the
iminium salt from DMA with POCl3 seemed to be solvent
dependent, and results were a little more variable than when
(16) Interestingly, compound 13 decomposed to 8 in the presence of TFA. Initial
analysis by HPLC that used TFA to modify the mobile phase seemed to
indicate that the forward reaction had stalled. Analysis of the reaction by
TLC indicated the absence of starting material and the presence of a nonpolar
impurity, later identified as 13.
(17) Sforza, S.; Dossena, A.; Corradini, R.; Virgili, E.; Marchelli, R. Tetrahedron
Lett. 1998, 39, 711-714.
(18) Bredereck, H.; Gompper, R.; Klemm, K.; Rempfer, H. Chem. Ber. 1959,
92, 837.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
393

Figure 4. Stacked FTIR spectra collected during first 30 min of the reaction of POCl3 with DMA.
Figure 5. Reaction profile for addition of POCl3 to DMA in ACN.
of the organic layer. The API (1) was isolated from IPA in
a yield of 98% with a purity of >99.5% (AN HPLC).
React-IR Investigation of the Reaction between POCl3
and DMA. Owing to the extended processing times normally
encountered on scale, it is imperative that addition times be
stressed during the development stage to ensure that the
process will stand up to the inevitably longer addition times
required on scale. In the present case, the final step in the
synthesis of the API required that a reactive iminium species
be generated and added to the penultimate intermediate. The
stress tests discussed earlier indicated that the process would
be robust to long addition times; however, a method to access
the stability of the iminium species was preferred.
DMA, and then POCl3 was added while the temperature was
maintained. Figure 4 shows a stacked plot of the 900-1800
cm-1 spectral region collected during the first 30 min of the
reaction. The initial growths at 1620 and 1400 cm-1 were
due to addition of DMA to the acetonitrile-containing cell.
The addition of POCl3 led to an initial growth at 1300 cm-1,
which then decreased in intensity. This spectrum matched
that of POCl3, and the fact that it grew and then decayed
indicated that the reaction between DMA and POCl3 was
not instantaneous. This growth and decay was more obvious
in the reaction profile (Figure 5). Examination of the profile
also showed that the reaction between DMA and POCl3
required 20-30 min to go to completion, even though the
addition of POCl3 was quite rapid. As POCl3 was added,
the peaks due to DMA decreased in intensity, and a new
species formed as indicated by the peak at 1663 cm-1,
The reaction between DMA and POCl3 was investigated
using in situ FTIR monitoring. Briefly, DMA was added to
a reaction cell that was fitted with a ReactIR DiComp probe.
A few scans were collected to establish the IR spectrum of
394
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

vanced Reactive System Screening Tool (ARSST). For the
RC-1 experiment, 1 equiv of DMA was added over 1 h to a
solution of POCl3 in acetonitrile. The exotherm was im-
mediate and directly proportional to the dose of DMA. The
total heat output was mild at 42 kJ/mol DMA and represents
an adiabatic temperature rise of 86 °C. Addition of DMA to
POCl3 did not show any indication of reagent accumulation,
and the reaction appeared to be over as soon as the addition
of DMA was complete. This is contrary to the results
observed with the React-IR which clearly showed accumula-
tion of POCl3. These differing results are likely due to the
longer addition period used for the RC-1 experiment rather
than due to the different order of addition.
An 8-mL aliquot of the RC-1 contents was placed in an
ARSST bomb which was then charged to 207 psi with
nitrogen. The temperature was increased to 200 °C over 90
min and then returned to ambient temperature. Figure 9
shows the temperature and pressure profile from the ARSST.
Although there were no thermal events, the pressure profile
indicates that a gas was generated. Above 120 °C, the
pressure increase accelerated rather than increasing linearly
with temperature. Additionally, the pressure dropped to 215
psi rather than returning to 207 psi when the contents were
returned to ambient temperature.
Figure 6. IR spectrum of species formed during the reaction
of DMA and POCl3.
Conclusion
A concise route to an R adrenoceptor antagonist that
displayed selectivity for the 1A and 1B subtypes was
developed and demonstrated in a pilot plant. Keys to the
success of the route were (1) the development of catalytic
hydrogenation conditions that allowed a substrate of low
solubility to be reduced and separated from the catalyst and
(2) activation of dimethylacetamide with POCl3 for instal-
lation of the amidine functional group. The reaction between
POCl3 and DMA was shown to be quite fast, but not
instantaneous, when POCl3 was added quickly to DMA. The
reaction required 20-30 min to be complete, and the
generated iminium salt was determined to be sufficiently
stable to warrant its use on a large scale; however, slow
decomposition was observed. Addition of DMA to POCl3
over a 1 h period did not lead to accumulation of DMA.
The reaction between DMA and POCl3 was determined to
be moderately exothermic with an Tad of 86 °C. Our studies
showed that decomposition of the DMA-POCl3 complex
did not result in any significant thermal events. Decomposi-
tion of the complex manifests itself as incomplete conversion
to the amidine, a situation that is easily overcome by using
excess complex for the reaction.
Figure 7. Extended reaction profile.
consistent with an iminium salt.19 The IR spectrum of the
new species is shown in Figure 6.
When the analysis of the reaction was extended to 3 h
(Figure 7), a trend showing a decrease in the concentration
of the iminium species was evident. Since the mixture
remained homogeneous, this decrease in concentration was
not due to precipitation of the complex and appeared
consistent with decomposition of the intermediate. Decom-
position of the complex due to adventitious water was ruled
out since the concentration of DMA did not increase.
The formation and stability of the iminium complex was
evaluated with a reaction calorimeter (RC-1) and an Ad-
Experimental Section
General. All reactions were conducted in glass-lined
reactors under a nitrogen atmosphere. All equipment and
lines were checked for leaks by pressurizing the system with
nitrogen prior to use.
2,4-Dichloro-6,7-dimethoxy-5-methyl-quinazoline (6).
6,7-Dimethoxy-5-methyl-1H-quinazoline-2,4-dione (4.54 kg,
19.2 mol, 1.0 equiv), acetonitrile (28 kg), and phosphorus
oxychloride (11.7 kg, 4 equiv) were combined and adjusted
(19) Simple iminium salts have a characteristic CdN stretching band at 1640-
1700 cm-1 (See Merenyi, R. Iminium Salts in Organic Chemistry, Part 1.
AdVances in Organic Chemistry; John Wiley: New York, 1979; Vol. 9,
Chapter 2, pp 23-106.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
395

Figure 8. RC-1 heat flow for addition of DMA to POCl3 in acetonitrile.
70 °C to afford 4.15 kg of 2 (98% yield) with 94% purity
(4.5% dione 5 present). A recrystallized sample had the
following analytical results: Anal. Calcd For C11H11-
ClN2O3: C, 51.88; H, 4.35; N, 11.00. Found: C, 51.87; H,
4.30; N, 11.00. mp 253-254 °C.
2-[(3-Nitro-benzyl)-methyl-amino]-6,7-dimethoxy-5-
methyl-1H-quinazolin-4-one (7). Chloroquinazolinone 2
(3.59 kg, 14.1 mol, 1.0 equiv), methyl-(3-nitrobenzyl)amine
hydrochloride (3.30 kg, 16.3 mol, 1.15 equiv), and diiso-
propylethylamine (4.71 kg, 36.4 mol, 2.5 equiv) were
combined in 2-propanol (35 kg) and heated at reflux for 24
h. Water (15 L) was added, and the reaction mixture was
cooled to 22 °C over ca. 6 h and then filtered, rinsed with
IPA/water (1:1 v/v), and dried to yield 5.11 kg of 7 (94%
yield) having a purity of 98% (AN HPLC). Anal. Calcd For
C19H20N4O5: C, 59.37; H, 5.24; N, 14.58. Found: C, 59.19;
H, 5.18; N, 14.29; mp 262-264 °C; MS: m/z (ESI): 385
(M•+ + 1).
2-[(3-Amino-benzyl)-methyl-amino]-6,7-dimethoxy-5-
methyl-1H-quinazolin-4-one (8). Compound 7 (5.11 kg,
13.3 mol, 1.0 equiv), palladium on carbon (400 g, 10%
palladium, 50% water-wet), THF (47 kg), MeOH (12 kg),
and NaOH (1.1 kg, 50% solution, 13.8 mol, 1.0 equiv) were
combined and stirred at 22 °C for ca. 30 min. The atmosphere
inside the reactor was converted from nitrogen to hydrogen,21
and the mixture was stirred for 2.5 h, after which time in-
process analysis indicated the reaction was compete. The
hydrogen atmosphere was exchanged with nitrogen, the
catalyst was removed by filtration, and the reactor was rinsed
with methanol (14 kg). The combined filtrates were con-
centrated at atmospheric pressure until the distillation tem-
perature was 63-65 °C. The final volume was adjusted to
20-25 L by adding or distilling methanol, and then water
(22 L) was added. The temperature was adjusted to 50-55
°C, and acetic acid (1.2 kg, 1.5 equiv) was added. The
mixture was adjusted to ca. 22 °C and then filtered, rinsed
Figure 9. Temperature and pressure profile for iminium
species determined with ARSST.
to reflux. After 18 h, in-process analysis indicated the
reaction was complete. The mixture was cooled to ca. 5 °C
and then added to a reactor containing water (50 L) while
the temperature was maintained below 20 °C. Product 6,
which precipitated during the quench, was filtered, rinsed
with water, and left on the filter overnight under vacuum
while dry nitrogen was blown through the cake. The product
was discharged and used as a water-wet cake in the next
step (11.0 kg with LOD ) 58.37%, estimated 4.58 kg of
product (87% yield)).
2-Chloro-6,7-dimethoxy-5-methyl-1H-quinazolin-4-
one (2). Water-wet 6 (11 kg, containing 4.58 kg of 6, 16.6
mol) was suspended in a mixture of THF (35 kg) and water
(32 L). Potassium hydroxide (7.3 kg, 64.8 mol, 3.9 equiv)
was added, and the mixture was stirred at 20-25 °C for 18
h (overnight). The bulk of the THF was distilled under
vacuum and replaced with water (ca. 30 kg). Acetic acid
was added until the pH was 6-7.20 The reactor contents were
adjusted to ca. 80 °C and then cooled to ca. 20 °C, filtered,
washed with water (3 × 30 L), and dried under vacuum at
(21) Caution! This reaction is exothermic. At this scale, a 7 °C temperature
(20) At this scale, 4.1 kg of acetic acid was required.
rise was observed over 40 min.
396
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

with methanol/water (1:1 v/v), and dried in vacuo at 75 °C
to afford 8 (4.54 kg, 96% yield). Purity ) 99.4 AN HPLC.
Anal. Calcd For C19H22N4O3: C, 64.39; H, 6.26; N, 15.81.
Found: C, 63.96; H, 6.26; N, 15.59; mp 271-274 °C; MS:
m/z (ESI): 355 (M•+ + 1).
aqueous layer was extracted with dichloromethane (16 kg).
The combined organic layers were combined, washed twice
with water (40 L), separated, and transferred to a clean vessel.
2-Propanol (32 kg) was added, and the resulting solution
was concentrated at atmospheric pressure with a maximum
jacket temperature of 90 °C until the internal volume was
ca. 40 L. The reactor contents were adjusted to 80 °C and
maintained at that temperature for 2-3 h and then cooled to
15 °C over 10 h. The reactor contents were aged at 15 °C
for 4 h and then filtered, rinsed with 2-propanol (18 kg),
and dried under vacuum at 75 °C until the LOD was <0.5%.
Total product was 4.43 kg (92.5% yield), purity 99.9% (AN
HPLC). Anal. Calcd For C23H29N5O3: C, 65.23; H, 6.90;
N, 16.54. Found: C, 65.27; H, 6.80; N, 16.54; mp 188 °C
(Form B); MS: m/z (ESI): 424 (M•+ + 1).
N′-(3-{[(6,7-Dimethoxy-5-methyl-4-oxo-1,4-dihydro-
quinazolin-2-yl)-methyl-amino]-methyl}-phenyl)-N,N-di-
methyl-acetamidine (1). N,N-Dimethylacetamide (2.47 kg,
28.3 mol, 2.5 equiv) was added to a cooled (ca. 10 °C)
solution of phosphorus oxychloride (4.36 kg, 28.3 mol, 2.5
equiv) and acetonitrile (3.14 kg) with the temperature
maintained below 20 °C.22 The resulting mixture was stirred
at 10-20 °C for ca. 1 h and then added to a cooled (ca. 10
°C) mixture of 8 (4.01 kg, 11.31 mol, 1.0 equiv) in
dichloromethane (53 kg) as the temperature was maintained
below 20 °C. The mixture was stirred at ca. 20 °C for ca. 1
h, after which time in-process analysis indicated the reaction
was complete. The reaction mixture was cooled to ca. 10
°C, then water (60 L) was added, and the pH was adjusted
to 8-9 by the addition of sodium hydroxide (11.35 kg, 50%
aqueous solution).23 The layers were then separated, and the
Acknowledgment
We are grateful to Dr. Yvonne Walbroehl, Tim Lane, Tina
Nguyen, and Larry Hennessy for providing analytical support
for this project. The Palo Alto Pilot Plant staff of Gary
Hedden, Gary Byrge, Jesus Dela Paz, Francis Onyegegbu,
Gaudelio Salom, and Viel Teixeira is also thanked for their
contributions during the pilot plant campaigns.
(22) At this scale, the addition required 30 min with 5 °C aqueous glycol
circulating through the reactor jacket.
(23) It is important that the pH remain above 8 for at least 30 min after the pH
adjustment. The pH had a tendency to drift lower following initial
adjustments into the specified pH range.
Received for review July 15, 2005.
OP050122H
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
397