Scale-up of the synthesis of a pyrimidine derivative directly on solid support

Article abstract of DOI:10.1021/op034032t

The solid-phase synthesis of 4-(2-amino-6-phenylpyrimidin-4-yl)benzamide, a compound obtained through combinatorial chemistry and parallel synthesis, can be scaled up directly on solid support in excellent yields and high purity. By applying highly loaded aminomethyl polystyrene as solid support, a good ratio between the product and the starting resin is achieved. For comparison, the synthesis was also performed in solution. The solid-phase synthesis approach has the advantage that the desired compound is easily and quickly accessible in sufficient quantities for early development demands.

Full text of DOI:10.1021/op034032t

Organic Process Research & Development 2003, 7, 553−558
Scale-Up of the Synthesis of a Pyrimidine Derivative Directly on Solid Support
Mark Meisenbach,* Thomas Allmendinger, and Ching-Pong Mak
NoVartis Pharma AG, Chemical and Analytical DeVelopment, CH-4002 Basel, Switzerland
Abstract:
The solid-phase synthesis of 4-(2-amino-6-phenylpyrimidin-4-
yl)benzamide, a compound obtained through combinatorial
chemistry and parallel synthesis, can be scaled up directly on
solid support in excellent yields and high purity. By applying
highly loaded aminomethyl polystyrene as solid support, a good
ratio between the product and the starting resin is achieved.
For comparison, the synthesis was also performed in solution.
The solid-phase synthesis approach has the advantage that the
desired compound is easily and quickly accessible in sufficient
quantities for early development demands.
Figure 1.
Introduction
Research departments in industry and academia increas-
ingly apply combinatorial chemistry for lead finding and lead
optimisation. One of the practical approaches is solid-phase-
supported parallel synthesis which generates libraries for
high-throughput screening. The number of hits from this
approach may increase dramatically, and consequently,
numerous compounds as potential drug candidates are
entering development. Process chemists are then faced with
the problem of rapidly delivering the first amounts necessary
for early preclinical and clinical studies. If the compounds
were obtained by a solid-phase approach, solution-phase
procedures for their synthesis might not be readily available.
To tackle this problem we were interested in gaining
experience of scaling up the solid-phase syntheses directly,
without investing resources, for the search of solution-phase
alternatives. In addition, solid-phase syntheses may be
inherently time saving by allowing multistep procedures to
be conducted without the need for isolation and purification
of intermediates. Similar ideas have already been expressed
and verified by Raillard1 and are illustrated by additional
examples in the present paper.
Figure 2.
equipment and investment, (4) comparison with the solution-
phase alternative.
Results and Discussion
(1) Scale-Up Phenomena and On-Bead Analytics. The
availability of starting materials is an obvious and self-
explanatory issue during scale-up.3 For the anticipated large-
scale solid-phase synthesis of pyrimidine 1 we had to replace
Rink amide resin B4 which was used by our colleagues in
research, because it is not readily available in larger amounts
(see Figure 2). Therefore, we used the Rink amide acetamido
resin 4, which is well established in peptide amide synthesis5
and easily accessible.
Thus (see Scheme 1), 1.5 equiv of the acetic acid
derivative 3 was coupled to aminomethyl polystyrene (AMPS
2a, 1.54 mmol/g) by standard amide-forming conditions.
After complete acylation as indicated by a negative Kaiser-
test6 for unreacted amino functions, resin 4a was checked
for its loading by well elaborated UV techniques.7 After
cleaving the Fmoc group with 20% diethylamine in DMF,
4-carboxybenzaldehyde 5 was coupled (diisopropyl carbo-
diimide, HOBt). Prolonged reaction time allowed the reduc-
tion of the amount of building block 5 by a factor of 10 as
Starting from R,â-unsaturated ketones A (Figure 1) a
library of different heterocycles was prepared in research.2
From that library we choose the pyrimidine 1 as an example
to study the following items: (1) scale-up phenomena of
different solid-phase reactions and the necessary on-bead
analytics, (2) effect of loading (an equivalent to concentration
in solution-phase chemistry), (3) selection of appropriate
(3) Lee, S.; Robinson, G. Process DeVelopment: Fine Chemicals from Grams
to Kilograms; Oxford University Press: Oxford, 1995.
(4) Rink, H. Tetrahedron Lett. 1987, 28, 3787.
* To whom correspondence should be addressed. Telephone: +41-61-696-
77-85. Fax: +41-61-696-27-11. E-mail: mark.meisenbach@pharma.novartis.com.
(1) (a) Raillard, S. P.; Ji, G.; Mann, A. D.; Baer, T. A. Org. Process Res. DeV.
1999, 3, 177. (b) Raillard, S. P.; Ji, G.; Mann, A. D.; Baer, T. A. U.S. Pat.
Appl. Publ. U.S. 2002/028466, March 7, 2002.
(5) Bernatowicz, M. S.; Daniels, S. B.; Ko¨ster, H. Tetrahedron Lett. 1989, 30,
4645.
(6) Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal. Biochem.
1970, 34, 595.
(7) Meienhofer, J.; Waki, M.; Heimer, E. P.; Lambros T. J.; Makofske R. C.;
Chang, C. D. Int. J. Pep. Protein Res. 1979, 13, 35.
(2) Felder, E. R.; Marzinzik, A. L. J. Org. Chem. 1998, 63, 723.
10.1021/op034032t CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/21/2003
Vol. 7, No. 4, 2003 / Organic Process Research & Development
•
553

Scheme 1
Figure 3.
Scheme 2
byproduct. An online Raman spectroscopic method was
established to monitor the conversion of this Claisen-
Schmidt reaction (see Supporting Information).10
The final reaction could be scaled up without any
problems: chalcone 8a and guanidine (liberated from its
hydrochloride with sodium ethoxide) were heated in dimethyl
acetamide while bubbling air through the mixture to form
pyrimidine 9a. After complete conversion (16 h) the product
was cleaved from the support (20% trifluoroacetic acid in
DCM). Pure 4-(2-amino-6-phenyl-pyrimidin-4-yl)benzamide
1 was obtained as its trifluoroacetate salt upon evaporation
of the filtrate and recrystallisation of the residue from ethanol/
water in 56% overall yield based on the solid-phase attached
4-carboxybenzaldehyde 6a.
(2) Effect of Different Loading. After isolating the
product of a solid-phase synthesis, the support (resin +
linker) is usually discarded as waste, although successful
examples of its reuse in further synthetic cycles are known.11
To reduce both volume of operation and amount of waste,
the loading of the resin, quantified as mmol of functionality
per g, has to be increased. In addition to theoretical
limitations (for polystyrene this is reached when every phenyl
ring is substituted by the linker) there may be practical
boundaries for the use of highly loaded resins in solid-phase-
supported synthesis. For the preparation of 1 we verified the
effect of this resin loading on operability, yield, and costs.
This was also another reason to switch from the ether bonded
Rink-linker B to the acetic acid derivative 4, which can be
assembled from 3 and differently loaded aminomethylated
polystyrenes resins (AMPS, 2). AMPS is commercially
available with loading up to 2.9 mmol/g 2b (NovaBiochem,
Switzerland). By adapting and optimising the procedure of
Adams12 higher-loaded resins were prepared. Thus, poly-
styrene cross-linked with 1% divinylbenzene was reacted
with N-chloromethyl phthalimide 12 in dichloromethane
compared to the original protocol. The structure of the
resulting solid-phase-supported aldehyde 6a was confirmed
by IR spectroscopy8 which showed a strong CdO band at
1710 cm-1 and the characteristic C-H band at 2730 cm-1.
The subsequent Claisen-Schmidt reaction was originally
performed on a 10 µmol scale using 20-fold excess of both
acetophenone and LiOH to achieve complete formation of
the chalcone 8. This result could be verified; however, by
employing the same conditions on a 35 mmol scale, no
conversion was observed even after 22 h, as shown by IR
spectroscopy. By cleaving a resin sample with 20% TFA in
dichloromethane, only p-formylbenzamide 11 was detected
by HPLC. This result may be rationalised by the low
solubility of LiOH in DME under dry/aprotic conditions.
Therefore, a small amount of EtOH was added, which
initiated a fast reaction9 and the formation of the desired
chalcone 8 together with 20% of the Michael adduct 10. This
was confirmed by sample cleavage from the resin and LC-
MS analysis (Figure 3). Short reaction screening resulted in
considerable improvements: using less acetophenone 7 and
LiOH (1.5 and 0.5 equiv, respectively) and a mixed-solvent
system of THF and methanol gave complete conversion
within only 1 h, forming only 5% of Michael adduct as
(10) Dyson, R. M.; Helg, A. G.; Meisenbach, M.; Allmendinger, T. 11th
International Symposium on Pharmaceutical and Biomedical Analysis; Basle
May 14-18 2000; P 007.
(8) Yan, B.; Gremlich, H.-U.; Moss, S.; Coppola, G. M.; Sun, Q.; Liu, L. J.
Comb. Chem. 1999, 1, 46.
(9) Chiu, C.; Tang, Z.; Ellingboe, J. W. J. Comb. Chem. 1999, 1, 73.
(11) Frechet, J. M. J.; Haque K. E. Tetrahedron Lett. 1975, 16, 3055.
(12) Adams, J. H.; Cook, R. M.; Hudson, D.; Jammalamadaka, V.; Lyttle, M.
H.; Songster, M. F. J. Org. Chem. 1998, 63, 3706.
554
•
Vol. 7, No. 4, 2003 / Organic Process Research & Development

Table 1. Solid-phase synthesis of pyrimidine 1-TFA salt starting with resins of different loading
yield of crude
1-TFA salt
overall yield and purity after
recrystallization of 1-TFA salt
loading of AMPS,
scale of synthesis
mass of 9 (g)
in (g)
in (%)
purity (area%)a
in (%)
in (area%)a
2a, 1.54 mmol/g, 40 mmol, 26 g
2b, 2.86 mmol/g, 70 mmol, 24.5 g
2c, 4.48 mmol/g, 100 mmol, 22.3 g
49
61.3
74.9
14.9
17.8
31.8
92
66
82
84
80
86
56
47
55
>98
>97
>98
a RP-HPLC.
using iron (III) chloride as Friedel-Crafts catalyst. If the
reaction is performed at 25 °C, a colourless product 13 is
obtained (Scheme 2) in contrast to dark-yellow resins
obtained when working under reflux. The phthaloyl protec-
tive group was removed by means of aminolysis using
methylamine in water/THF to avoid the common but toxic
hydrazine and dioxane, respectively. The desired loading can
be obtained by adjusting the ratio of PS, chloromethyl
phthalimide, and FeCl3. We synthesised AMPS with up to
4.48 mmol/g (2c), corresponding to two out of three phenyl
rings being aminomethylated. We have not extended the
investigation to the highest possible loading of AMPS. A
resin with up to 7.3 mmol/g (96% of phenyl rings on
polystyrene are substituted) has been described previously,13
but its application has not been reported thus far.
To study the effect of loading we repeated the synthesis
depicted in Scheme 1 using different loaded resins. The
conditions during reactions on higher-loaded resins (2b and
2c) as compared to the initially used 2a did not need any
adjustments. The quality of the product was comparable, and
the ratio of product to solid support was increased. The
results are summarised in Table 1.
(3) Equipment Needed. Simple overhead mechanically
stirred glass vessels immersed in an oil bath can be used to
run on-bead reactions; however, filtration and washing of
the whole mixture may be cumbersome. As pointed out by
Raillard,1 the use of reactors employed for scaling up solid-
phase peptide synthesis is advantageous. These consist
essentially of stirred suction filters with different volumes14
(if not, continuous flow reactors are employed).15 As peptide
coupling and deprotection reactions are highly optimised,
they usually run at room temperature; therefore, these suction
filters do not have heating/cooling jackets. We therefore
designed reactors capable of temperature adjustments by
incorporating sintered filter plates into multiple necked
double-wall reactors with volumes up to 4 L.16 The outer
compartment can be connected to a heating/cooling circuit
(thermostat). For single-wall sintered-bottom reaction vessels
a simple cooling finger was designed. Figure 4 depicts two
such vessels used throughout the investigations.
Figure 4. Reactors with coarse fritted filter plates: (A) 250
mL, equipped with three-necks, jacket, and mechanical glass
stirrer with Teflon paddle and (B) 1.5 L, equipped with jacket,
glass anchor stirrer, and a five-necked clamped top.
supported syntheses on larger scale, chemists need to explore
alternative routes utilising solution chemistry. To evaluate
advantages and disadvantages of both approaches, we
prepared pyrimidine 1 in solution phase as well. For this
relatively small and simple molecule, similar chemistry was
applied (see Scheme 3). Some interesting results arose.
Heating 4-carboxybenzaldehyde 5 in excess thionyl
chloride, which is described to form acid chloride 14,17
actually afforded the trichloro compound 15 instead. There-
fore, the milder Vilsmeier reagent was used followed by
aminolysis to obtain 11. Due to its solubility in water the
variable yield of 11 depended on the workup procedure. The
Aldol condensation with acetophenone under standard condi-
tions gave chalcone 17 containing 10% of the intermediate
16. Its structure was revealed by independent synthesis18 to
compare the NMR data and chromatographic properties.
(4) Comparison with Solution-Phase Chemistry. With-
out the experience and the equipment to perform solid-phase-
(13) Zikos, C. C.; Ferderigos, N. G. Tetrahedron Lett. 1995, 36, 3741.
(14) Edelstein, M.; Scott, P. E.; Sherlund, M.; Hansen, A. L.; Hughes, J. L. Chem.
Eng. Sci. 1986, 41, 617.
(15) Caciagli, V.; Cardinali, F.; Hansicke, A.; Tuchalski, G.; Lombardi, P. J.
Pept. Sci. 1997, 3, 224.
(17) Graffner-Nordberg, M.; Sjo¨din, K.; Tunek, A.; Hallberg, A. Chem. Pharm.
Bull. 1998, 46, 591.
(18) Aldol reaction of silylenol ethers with aldehydes catalysed with bismuth
trichloride according to Ohki, H.; Wada, M.; Akiba, K. Tetrahedron Lett.
1988, 29, 4719.
(16) These are made on request by AMSI-Glas AG, CH-4317 Wegenstetten,
Switzerland.
Vol. 7, No. 4, 2003 / Organic Process Research & Development
•
555

Scheme 3
land), 2.86 mmol/g 200-400 mesh from NovaBiochem
(La¨ufelfingen, Switzerland). Polystyrene 100-400 mesh
from Purolite (Pontyclun, UK) was used for the synthesis
of highly loaded resin. All resins are 1% cross-linked with
divinylbenzene. Melting points are uncorrected. 1H- and 13
C
NMR spectra were recorded on a Bruker AC-300. IS-MS
was carried out on a Varian MAT-711. FT-IR spectra (KBr)
were recorded with a Bruker IFS66. The extinction in UV
was determined on a Hewlett-Packard 3453 spectrophotom-
eter. Thin-layer chromatography was performed on Merck
silica gel 60 F254 plates. RP-HPLC was carried out with an
Agilent 1100 system at 50 °C using a Macherey-Nagel CC
125/4 Nucleosil 100-5 C18 column. Synthesis sequences
were performed in reaction vessels containing a jacket, a
frit Por. 2, and multiple necks. The reactions were carried
out under N2 atmosphere.
Highly Loaded Aminomethylated Polystyrene (2c).
Underivatised 1% cross-linked 100-400 mesh polystyrene
(26 g 0.25 mol) was washed twice with DCM. In 250 mL
of DCM were suspended N-chloromethyl phthalimide (29.35
g, 0.15 mol) and anhydrous ferric chloride (12.17 g, 0.075
mol) for 15 min and then added to the pre-swollen resin.
After stirring the suspension for 1 day at 23 °C, the resin
was filtered and washed with DCM (100 mL), THF/H2O
(4:1, 100 mL), THF (150 mL), MeOH (100 mL), DMF (2
× 100 mL), MeOH (2 × 100 mL), THF (2 × 150 mL). A
sample for IR was withdrawn. Then the resin was suspended
in 500 mL of THF, aqueous methylamine (40%, 167 mL)
was added, and the reaction mixture was stirred for 16 h at
55 °C. The resin was filtered and washed with THF (150
mL), alternating MeOH and DMF (2 × 100 mL), DCM (2
× 150 mL), and MeOH (3 × 100 mL). The colourless
product 2c was dried in vacuo at 45 °C for 16 h, providing
29.3 g of aminomethylated polystyrene. Elemental analysis
showed 6.53% N, corresponding to a loading of 4.66 mmol/
g. Titration with HCl gave 4.48 mmol/g (87.5% yield from
chloromethyl phthalimide), and in the IR-spectrum the
characteristic bands of the phthalylated intermediate at 1772
and 1712 cm-1 had disappeared.
Loading of the Aminomethylated Resins 2 with the
Fmoc-Protected Rink-Linker (3). The aminomethylated
resins were swollen in DMF. To a 0.33 M solution of 1.5
equiv of Fmoc-protected Rink-linker and HOBt in DMF, was
added dropwise 1.8 equiv of DIC over 15 min. This solution
was stirred for an additional 15 min at 23 °C, and then the
solution of the activated linker was added to the drained
resins; the suspension was stirred for at least 16 h until the
Kaiser test was negative. The resins were washed with 5
mL/g portions alternating DMF and MeOH (3 ×), DCM (2
×), and MeOH (3 ×). The colourless products (4a-c) were
dried in vacuo at 45 °C for 16 h. The loading efficiency
was determined via quantitative monitoring of the Fmoc
deprotection by UV spectroscopy2 (Table 2).
Hydroxyketone 16 thus prepared (see Experimental Section)
turned out to be hardly soluble (even in dimethyl sulfoxide),
the reason for its precipitation. This phenomena indicates
the advantage of solid-phase-supported synthesis when
intermediates have only low solubility. Finally the condensa-
tion with guanidine and the subsequent oxidation afforded
pyrimidine 1 in moderate yield.
Conclusions
In summary, we demonstrated that a fast scale-up of the
synthesis of the desired pyrimidine derivative was possible
on the solid support. The research protocols could be used
directly with only minor modifications. This saves time,
which is one of the most important factors in the early stage
of drug development. Highly loaded AMPS resin shifts the
ratio of support/product in the desired direction and provides
higher volume productivity. Thus, it makes this support
attractive for the fast scale-up of small molecules. The
production time of the solid-phase synthesis is comparable
with the synthesis in solution due to the few reaction steps,
but the overall yield (47-56%) exceeds the yield obtained
by solution-phase synthesis (21-34%). Sure enough, process
development would find conditions or routes to overcome
drawbacks of the solution-phase synthesis and to improve
the yield; however, we could show the relative ease of scaling
up synthesis on solid support.
Experimental Section
All reagents and solvents were obtained from commercial
suppliers and used without further purification. The 4-{(R,S)-
1-[1-(9H-fluoren-9-yl)methoxycarbonylamino]-(2′,4′-
dimethoxybenzyl)}phenoxyacetic acid (3) was purchased
from Senn Chemicals (Dielsdorf, Switzerland). The following
aminomethylated polystyrene resins were used: 1.54 mmol/g
100-200 mesh from Senn Chemicals (Dielsdorf, Switzer-
Coupling of the 4-Carboxybenzaldehyde (5) onto the
Solid Support. The Fmoc protecting group was cleaved from
the resin 4a-c with 20% diethylamine in DMF (4 mL/g)
within 2 × 15 min. The resins were washed with 5 mL/g
portions of alternating DMF and MeOH (2 ×) and DMF
556
•
Vol. 7, No. 4, 2003 / Organic Process Research & Development

Table 2: Loading of the AMPS resin with the Rink-linker
1. All products gave the expected results in IS-MS and
identical IR and NMR spectra.
loading
(Z, mmol/g)
IS-MS (m/z) 291 [MH]+.
starting material
product
IR (cm-1) 3421 (N-H), 3320 (N-H), 3050-3170 (C-
H), 1674 (CdO), 1613 (CdC).
1H NMR (DMSO-d6) δ 8.31 (d, 2 H, 8.7 Hz); 8.23 (m,
2H); 8.12 (bs, 1H); 8.02 (d, 2 H, 8.6 Hz); 7.82 (s, 1H); 7.55
(m, 3H); 7.49 (bs, 1H).
32.5 g 2a Z ) 1.54 mmol/g
24.5 g 2b Z ) 2.86 mmol/g
22.3 g 2c Z ) 4.48 mmol/g
59.73 g, 4a
59.08 g, 4b
70.72 g, 4c
0.87
1.17
1.45
13C NMR (DMSO-d6) δ 168.2, 165.6, 164.7, 163.3, 139.6,
137.1, 137, 132, 129.6, 128.7, 128.2, 128, 103.5.
Microanalysis, found (calculated) C: 56.44 (56.45); H:
3.74 (3.79); N: 13.86 (13.92); F: 14.1 (13.87).
4-Formylbenzamide (11). The solution of 14.65 g (0.145
mol) of Vilsmeier reagent in 100 mL of dry THF was cooled
under argon to -5 °C, and 17.0 g (0.113 mol) of 4-carboxy-
benzaldehyde 5 was added at once. The suspension was
warmed to 0 °C and stirred for 16 h at this temperature and
then poured onto 30 mL of cold aqueous ammonia. The
mixture was concentrated by distillation at reduced pressure
and the solid product obtained was filtered, washed with
small amounts of cold water, and dried in vacuo to yield
10.23 g of 4-formylbenzamide. A second crop may be
obtained from the mother liquor by extraction using ethyl
acetate.
(3 ×). 4-Carboxybenzaldehyde 5 (1.3 equiv) and 1.3 equiv
of HOBt were dissolved in DMF (0.23 M solution including
DMF on the swollen resin), and 1.5 equiv of DIC was added
dropwise. After 15 min activation the solution was added to
the resin. The suspension was stirred for at least 16 h until
the Kaiser test6 was negative. The resins were washed with
5 mL/g portions of alternating DMF and MeOH (3 ×), DCM
(2 ×), MeOH (3 ×). The product stemming from resin 4a
was dried in vacuo at 25 °C for 48 h, resulting in 54.96 g of
colourless aldehyde resin 6a (Z ) 0.91 mmol/g according
to weight decrease). Resins 6b-c were directly used after
3 × washing with THF for the next step.
IR (cm-1) 3400 (N-H), 3080-2850 (C-H), 2730 (C-
H, aldehyde), 1710 (CdO, aldehyde), 1660 (CdO, amide),
1610 (CdC).
4-(3-Oxo-3-phenylpropenyl)benzamide (8a-c) via
Claisen-Schmidt Reaction. The aldehyde resins were
treated at 23 °C with a 0.25 M solution of 1.5 equiv of
acetophenone 7 and 0.5 equiv of LiOH in THF/MeOH
(4:1). At various times samples were withdrawn, washed,
and treated with 20% TFA in DCM. The solvent was
removed in vacuo, and the residue was analysed by RP-
HPLC and TLC. (Rf ) 0.56, Rf ) 0.41 Michael-adduct,
EtOAc/MeOH/TEA (9:1.5:0.5)). When the aldehyde had
almost completely disappeared (RP-HPLC < 1%), within
about 1.5 h, the resins were isolated by filtration. The resins
were washed with 5 mL/g portions of THF, HOAc, DCM
(2 ×), DMF (2 ×), and MeOH (3 ×). After taking a sample
for IR, the resins were directly used for the next step.
IR (cm-1) 3400 (N-H), 3070-2849 (C-H), 1660 (CdO,
amide), 1607 (CdC).
1H NMR (300 MHz, DMSO-d6) δ 9.85 (s, 1H); 7.95 (bs,
1H); 7.83 (d, 2H, 7.9 Hz); 7.74 (d, 2H, 7.9 Hz); 7.30 (bs,
1H).
Microanalysis, found (calculated) C: 64.42 (64.31); H:
4.73 (4.69); N: 9.39 (9.37).
1-Phenyl-3-(4-aminocarbonyl-phenyl)-2-propen-1-
one (17). The solution of 4 g (26.8 mmol) of 11 and 4.83 g
(40.2 mmol) of acetophenone 7 in 40 mL of methanol was
treated with 0.47 g of a 15% aqueous solution of potassium
hydroxide (1.24 mmol) and stirred at room-temperature
overnight. The thick suspension obtained was diluted with
methanol (10 mL) and water (50 mL). The solid was isolated
by filtration, washed with methanol/water, and dried in vacuo
to yield 4.2 g (17 mmol, 62%) of 17, containing about 10%
1
of the aldol addition product 16 according to H NMR-
4-(2-Amino-6-phenylpyrimidin-4-yl)benzamide Tri-
fluoroacetate (1-TFA Salt). The chalcone resins 8a-c were
washed three times with DMA. Guanidine hydrochloride (6
equiv) was solubilised in DMA and treated with 6 equiv of
NaOEt in DMA. The precipitate was filtered off, and the
resulting 1 M solution of the free guanidine was added to
the solid-supported chalcone. The reaction mixture was
stirred at 100 °C for 16 h under an air atmosphere. The resins
were washed with 5 mL/g portions of DMA, HOAc, DCM
(2 ×), MeOH (2 ×), DMF (2 ×), DCM (2 ×), MeOH (3
×). The products were dried in vacuo at 40 °C for 48 h
(Table 1).
spectroscopy, HPLC and elemental analysis. The aldol
addition product was prepared independently (see below).
A second crop obtained from the mother liquor was
contaminated by the Michael-adduct 10.
1H NMR (300 MHz, DMSO-d6) δ: 8.12 (d, 2H, 7.5 Hz);
7.97 (d, 1H, 15 Hz); 7.91 (m, 4H); 7.88 (bs, 1H); 7.73 (d,
1H, 15 Hz,); 7.61 (d, 1H, 7.5 Hz); 7.52 (t, 2H, 7.5 Hz); 7.42
(bs, 1H).
4-(2-Amino-6-phenylpyrimidin-4-yl)benzamide (1). The
solution of guanidine hydrochloride (1.53 g, 16 mmol) in
N,N-dimethylacetamide (10 mL) was treated with sodium
ethoxide (1.1 g, 16.1 mmol), stirred for 15 min, and filtered.
The filtrate was added to a solution of 17 (2.0 g, 7.95 mmol)
in DMA (10 mL), and the resulting mixture was heated with
stirring to 100 °C for 18 h. The reaction mixture was cooled
to room temperature, and water (30 mL) was added slowly.
The precipitate was filtered, washed with water, and dried
in vacuo at 45 °C to obtain 1 (1.3 g, 56.4%) as a slightly
yellow solid.
IR (cm-1) 3420 (N-H), 3070-2850 (C-H), 1670 (CdO,
amide), 1610 (CdC)
The final cleavage of the product was performed with
20% TFA in DCM (10 mL/g), the resin was filtered off,
and the resulting solutions were evaporated to dryness
yielding slightly coloured solids, which were recrystallised
from EtOH/H2O (9:1). Yields and purity are shown in Table
Vol. 7, No. 4, 2003 / Organic Process Research & Development
•
557

4-(1-Hydroxy-3-oxo-3-phenyl-propyl)benzamide 16 (Pre-
pared for Structure Elucidation and Comparison). A
mixture of 0.36 g (1.14 mmol) of bismuth trichloride, 10
mL of dichloromethane, 0.58 g (3.9 mmol) of 11 and 2.2
mL of 1-phenyl-1-trimethylsilyloxy ethylene (silylenolether
of acetophenone) are stirred under nitrogen for 22 h at room
temperature. Upon addition of methanol (0.5 mL) and
concentrated aqueous HCl (20 µL) the product precipitates.
Isolation by filtration, washing with dichloromethane and
water, and drying in vacuo afforded 588 mg of 16. 1H NMR
(300 MHz, DMSO-d6) δ: 7.90 (d, 2H, 7.6 Hz); 7.87 (bs,
1H); 7.78 (d, 2H, 7.6 Hz); 7.58 (t, 1H, 7.6 Hz); 7.4-7.5 (m,
4H); 7.23 (bs, 1H); 5.42 (d, 1H, 6 Hz); 5.1-5.2 (m, 1H);
3.38 (dd, 1 H, 16 Hz, 7.6 H); 3.12 (dd, 1H, 16 Hz, 5 Hz).
Supporting Information Available
Description of an online Raman spectroscopic method
established to monitor the conversion of this Claisen-
Schmidt reaction. This material is available free of charge
via the Internet at http://pubs.acs.org
Received for review February 25, 2003.
OP034032T
558
•
Vol. 7, No. 4, 2003 / Organic Process Research & Development