A convenient 'catch and release' synthesis of fused 2-alkylthio-pyrimidinones mediated by polymer-bound BEMP

Article abstract of DOI:10.1016/S0040-4039(03)01188-2

A robust 'catch and release' synthesis of fused 2-alkylthio-3-substituted-pyrimidinones mediated by the polymer-bound base P-BEMP is described. This reengineered synthesis combines the efficiency of the classical synthesis (three steps, three diversity points) with the practical benefits of resin-bound reagents. The solution-phase strategy, reagent compatibility, and the results of a representative 48-member combinatorial library are described and presented herein.

Full text of DOI:10.1016/S0040-4039(03)01188-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 5041–5045
A convenient ‘catch and release’ synthesis of fused
2-alkylthio-pyrimidinones mediated by polymer-bound BEMP
Gregory L. Adams, Todd L. Graybill,* Robert M. Sanchez, Victoria W. Magaard, George Burton
and Ralph A. Rivero
GlaxoSmithKline, Discovery Research, 1250 S. Collegeville Road, Collegeville PA 19426, USA
Received 10 April 2003; accepted 7 May 2003
Abstract—A robust ‘catch and release’ synthesis of fused 2-alkylthio-3-substituted-pyrimidinones mediated by the polymer-bound
base P-BEMP is described. This reengineered synthesis combines the efficiency of the classical synthesis (three steps, three diversity
points) with the practical benefits of resin-bound reagents. The solution-phase strategy, reagent compatibility, and the results of
a representative 48-member combinatorial library are described and presented herein. © 2003 Elsevier Science Ltd. All rights
reserved.
1. Introduction
2. Results and discussion
2.1. P-BEMP enabled synthesis
We recently reported a multi-step ‘catch and release’
synthesis of 3-alkylthio-1,2,4-triazoles mediated by the
polymer-bound base P-BEMP.1 This triazole synthesis
exploits both the efficiency of the classical synthesis
(three synthetic steps; three diversity points) and the
practical benefits of resin-bound reagents (use of excess
reagents, ease of use, automation-friendly).2–5 It quickly
became evident this simple ‘catch and release’ strategy
could be further exploited for synthesis of other hetero-
cycles. Fused pyrimidin-4-ones 9 were considered
attractive targets owing to the short concise synthesis
(three steps; three-point diversity),6,7 readily available
pools of diversity reagents,8 and diverse reported bio-
logical activity.9–11 As fused pyrimidin-4-ones were
poorly represented in our corporate compound collec-
tion, we sought a robust solution-phase approach that
not only eliminated the need for isolation of synthetic
The ‘catch and release’ sequence is described in Scheme
1. Central to this scheme is the strongly basic, nonnu-
cleophilic polymer-bound BEMP (2-tert-butylimino-2-
diethylamino-1,3-dimethyl-perhydro-1,3,2-diazaphos-
phorine on polystyrene, P-BEMP, Fig. 1).13,14 Owing to
these properties, P-BEMP is often the reagent of choice
for deprotonation and N-alkylation of weakly acidic
heterocycles.15,16 In this ‘catch and release’ preparation,
P-BEMP plays a key role in nearly every step of the
sequence.
2.2. The ‘catch’ step
A solution of substituted anthranilic acid ester 1 (1.3
equiv.) and isothiocyanate 2 (1.5–2.0 equiv.) in
intermediates but was also complementary to
a
reported solid-phase method.12 In this letter, we report
a ‘catch and release’ synthesis of 2-alkylthio-3-substi-
tuted-3H-quinazolin-4-ones and thienopyrimidin-4-
ones mediated by the polymer-bound base, P-BEMP.
Application highlights, reagent compatibility, and the
dimethylsulfoxide (DMSO) is heated at 100°C for three
days to effect both addition and subsequent cyclization
to the 2-thioxo-quinazolin-4-one 3. Polymer-bound P-
BEMP (1 equiv.) and DMF17 are then added to cleanly
sequester the acidic heterocycle as ion-pair 4. Owing to
the polymer-bound nature of ion-pair 4, the excess
reagents/products (anthranilic ester, isothiocyanate, and
2-thioxo-quinazolinone 3) are easily removed during
subsequent resin washing steps. Surprisingly, DMSO is
an optimal and convenient solvent for preparation of
intermediates 3, 6 and the subsequent conversion to
ion-pair 4.18,19 Under these conditions, reaction of
isothiocyanate 2 with anthranilic ester 1 is believed to
results of
described herein.
a representative 48-member array are
Keywords: quinazolin-4-one; quinazolinone; thienopyrimidinone; 2-
tert-butylimino-2-diethylamino-1,3-dimethyl-perhydro-1,3,2-diaza-
phosphorine; parallel synthesis; solid-supported reagents.
* Corresponding author. Tel.: 610-917-5870; fax: 610-917-4206;
e-mail: todd l graybill@gsk.com
– –
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)01188-2

5042
G. L. Adams et al. / Tetrahedron Letters 44 (2003) 5041–5045
Scheme 1. ‘Catch and release’ synthesis of fused 2-alkylthio-pyrimidinones 9, 10.
substituted epoxide (i.e. 8a–d) ‘releases’ thioethers 9 or
10 into the solution. Simple filtration and subsequent
resin wash typically provides DMSO solutions of crude
products 9 or 10 in moderate to excellent purity
(LCMS-UV214; 50–100%).21,22
A substoichiometric
amount of electrophile 7 or 8 minimizes product con-
tamination as consumption of electrophile is typically
complete in 2–8 h. The scale of the experiment is
configured such that the resulting volume of DMSO
solution (containing products 9 or 10) is compatible
with the injection volume of the reverse-phase HPLC
purification step.23 While we find it convenient to use
DMSO in the release step during array production,
similar (though more variable) results were obtained
using acetonitrile (ACN) during building block qualifi-
cation experiments (refer to Fig. 3 and discussion
below). For example, a proton NMR spectrum of
representative crude quinazolinone 12a (‘released’ by
treatment of 11 with 1-bromopentane 7a in ACN) is
shown in Figure 2. Generic synthetic protocols for
array production (see representative fused pyrimidin-4-
ones in Fig. 4) and quinazolinone 12a are provided in
Ref. 25.24
Figure 1. Structure of polymer-bound BEMP.
be the rate limiting step of the transformation as the
expected thiourea intermediate (analogous to 6) is not
observed by analytical methods (LCMS, NMR).
Based on the above observations, we were subsequently
surprised to find primarily accumulation of uncyclized
thiourea 6 when thiophene-based aminoesters 5 and
isothiocyanates 2 are reacted as described above
(Scheme 1). The results of additional experiments indi-
cate that (1) consumption of 5 (formation of 6) is, in
most cases, complete after heating for 20 h at 100°C;
(2) heating for longer than 20 h is typically counterpro-
ductive as thioureas 6 tend to slowly degrade rather
than cyclize; (3) addition of P-BEMP (1 equiv.) to a
solution of thiourea 6 rapidly effects both sequestration
and cyclization to ion-pair 4; and (4) excess reagents
and by-products are conveniently removed from ion-
pair 4 by a simple wash of the resin. The preferred
protocol for preparation of ion-pair 4 from thiophene-
based aminoesters 5 is highlighted in Scheme 1.
2.4. Reagent compatibility
Building block qualification experiments were used to
define reagent compatibility for this P-BEMP-mediated
sequence. These experiments (designed to test the com-
patibility of a single class of diversity reagent) were
typically configured for either the Robbins FlexChem™
or Argonaut Quest210™ platforms.26 For example, to
identify suitable electrophiles, ion-paired quinazolin-4-
one 11 was treated in parallel with 0.1 M solutions of
diverse alkyl halides (i.e. 7a–d) or epoxides (i.e. 8a–d)
then allowed to react under conditions projected for
library synthesis. Additional details and representative
results are provided in Figure 3. Several observations
are highlighted here.
2.3. The ‘release’ step
DMSO is also an advantageous solvent for the ‘release’
step.20 As a result, it is convenient and robust to couple
the ‘release’ step with a subsequent high throughput
HPLC purification step for preparation of large arrays.
Room temperature treatment of ion-pair 4 with a
DMSO solution of alkylating agent (i.e. 7a–d) or mono-

G. L. Adams et al. / Tetrahedron Letters 44 (2003) 5041–5045
5043
Figure 2. 400 MHz NMR of representative quinazolinone 12a (crude).
Figure 3. Representative results from qualification experiments—electrophiles 7, 8.
Reagents containing functional groups more acidic
than the fused 2-thioxopyrimidinone nucleus (pKa ꢀ
8.5)27 are not tolerated when using this protocol.
ring substitution with electron-deficient groups (i.e. Cl,
F) rapidly decrease performance of these inputs under
these general conditions. Generally, the yields of prod-
ucts (9, 10) derived from anthranilic esters 1 are supe-
rior to those derived from thienyl-based aminoesters 5.
2.4.1. Electrophiles 7, 8. Based on experiments described
in Figure 3, compatibility trends for alkylating agents
parallel those reported earlier for 3-thio-1,2,4-tri-
azoles.1,12 Most common alkyl iodides and bromides
are compatible whether hindered or not (7a,d). In addi-
tion, many diverse benzyl halides (7b), a-haloesters, and
amides (7c) are successfully employed. For mono-sub-
stituted epoxides such as 8a–d, regiospecific ring open-
ing (SN2 attack at least hindered position) provides the
corresponding 2° alcohol derivatives 13a–d. As exem-
plified in Figure 3, use of epoxides as electrophile input
typically provides crude products of lower purity and
yield compared to those derived from alkyl halides
2.4.3. Isothiocyanates 2. With a few exceptions, compat-
ibility trends for isothiocyanates 2 parallel those
reported earlier for 3-thio-1,2,4-triazoles.1 Many
(un)branched alkyl (i.e. 2a,b,c,e) and substituted benzyl
isothiocyanates (i.e. 2d,f) are compatible with the
sequence (Fig. 4). Aryl isothiocyanates (i.e. 2g,h) are
also compatible but acylisothiocyanates (R-CO-NCS)
are not.
2.4.4. Resin-bound base. P-BEMP (Fluka cat. c 20026,
2.2 mmol/gram) was the only polymer-bound base
investigated during this work.28
2.4.2. Aminoesters 1, 5. Many readily-accessible8
phenyl- and thienyl-based aminoesters (i.e. 1a–c, 5a-c,
see Fig. 4) are compatible. Aminoesters containing
neutral or electron-donating groups (i.e. alkoxy, alkyl,
H) at open ring positions performed best. In contrast,
2.5. Application
Using this reagent compatibility information and these
generic protocols, more than 3000 diverse, yet drug-

5044
G. L. Adams et al. / Tetrahedron Letters 44 (2003) 5041–5045
Figure 4. Results for representative subset of hit identification array.
like,29 fused 2-alkylthio-3-substituted-pyrimidin-4-ones
(9, 10) were prepared for hit identification purposes
(90% registration rate after purification) using the Rob-
bins FlexChem™ platform. The 48 compounds high-
lighted in Figure 4 are a representative subset of this hit
identification array.30 This subset illustrates the trends
mentioned earlier in this report. With little modifica-
tion, this robust ‘catch and release’ strategy should be
amenable to other parallel synthesis platforms as well.
2. Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P.
S.; Leach, A. G.; Longbottom, D. A.; Nesi, M.; Scott, J.
S.; Storer, R. I.; Taylor, S. J. J. Chem. Soc., Perkin Trans.
1 2000, 23, 3815–4195.
3. Bhattacharyya, S. Combinatorial Chem. High Throughput
Screening 2000, 3, 65–92.
4. Ley, S. V.; Baxendale, I. R. Spec. Publ.-R. Soc. Chem.
2001, 266, 9–18.
5. Flynn, D. L.; Devraj, R. V.; Parlow, J. J. Curr. Opin.
Drug Disc. Dev. 1998, 1, 41–50.
6. Freundler, M. P. Bull. Soc. Chim. Fr. 1904, 31, 882–884.
7. McCarty, J. E.; Haines, E. L.; VanderWerf, C. A. J. Am.
Chem. Soc. 1960, 82, 964–966.
3. Conclusion
8. As necessary, commercially-available 2-amino (hetero)-
arylacids were conveniently converted to the correspond-
ing methylester derivatives (1, 5) using the method of
Calestani et al. (J. Chem. Soc., Perkins Trans. 1 1998, 11,
1813–1824). Substoichiometric use of methyl iodide
ensured monomethylation and ease of purification.
9. Abdel Hamid, S. G.; El-Obeid, H. A.; Al-Majed, A. A.;
El-Kashef, H. A.; El-Subbagh, H. I. Med. Chem. Res.
2001, 10, 378–389.
In this letter, we report a robust ‘catch and release’
preparation of 2-alkylthio-3-substituted-3H-quinazolin-
4-ones and thienopyrimidin-4-ones mediated by the
polymer-bound base, P-BEMP. This convenient synthe-
sis combines the efficiency of the classical synthesis
(three steps; three diversity points) with the practical
benefits of resin-bound reagents. We believe this is the
first reported solution-phase parallel synthesis of these
heterocycles. Efforts to extend this ‘catch and release’
paradigm to other heterocycles are in progress and will
be the topic of future reports.
10. Modica, M.; Santagati, M.; Russo, F.; Selvaggini, C.;
Cagnotto, A.; Mennini, T. Eur. J. Med. Chem. 2000, 35,
677–689.
11. Hoffmann, U.; Bsharat, N.; Jira, T. Pharmazie 1996, 51,
664–667.
12. Makino, S.; Suzuki, N.; Nakanishi, E.; Tsuji, T. Tetra-
hedron Lett. 2000, 41, 8333–8337.
References
1. Graybill, T. L.; Thomas, S.; Wang, M. A. Tetrahedron
Lett. 2002, 43, 5305–5309.
13. Schwesinger, R. Chimia 1985, 39, 269–272.

G. L. Adams et al. / Tetrahedron Letters 44 (2003) 5041–5045
5045
14. Schwesinger, R. Nachr. Chem. Tech. Lab. 1990, 38, 1214–
1226.
mmol, 1 equiv.) were added to each reaction then mixed
at room temperature overnight.
15. Xu, W.; Mohan, R.; Morrissey, M. M. Bioorg. Med.
Chem. Lett. 1998, 8, 1089–1092.
16. McComas, W.; Chen, L.; Kim, K. Tetrahedron Lett.
2000, 41, 3573–3576.
17. Addition of DMF enhances resin swelling and creates a
free-flowing suspension.
Formation of 4 from aminoesters 5: DMSO solutions of
aminoester 5 (0.5 M, 0.4 mL, 0.195 mmol, 1.3 equiv.) and
isothiocyanate 2 (0.7 M, 0.32 mL, 0.23 mmol, 1.5 equiv.)
were mixed and heated at 100°C for 16 h. DMF (1 mL)
and P-BEMP (68 mg, 0.15 mmol, 1 equiv.) were added to
each reaction well and mixed at 50°C for 3 h then at
room temperature overnight.
‘Release step’ used in array production: Excess solution
was aspirated to waste. The resin was washed with 1 mL
portions of DMF (3×) then DMSO (1×). A solution of
alkylating agent or epoxide (0.14 M, 0.5 mL, 0.07 mmol,
0.7 equiv.) in DMSO was then added to resin 4. The
suspension was then mixed for 24–48 h at room tempera-
ture. The solution containing fused pyrimidinone 9 or 10
was drained directly into a deepwell 96-well microtiter
plate. The resin was subsequently washed with a 0.2 mL
portion of DMSO. The wash volume was also drained
directly into the deepwell 96-well microtiter plate. DMSO
solutions from this plate (0.7 mL per well) were injected
directly into the preparative HPLC.
18. To our knowledge, DMSO has never been reported as a
preferred solvent for quinazolinone ring formation.
19. Advantages include clean and reliable conversion, excel-
lent solubility of reagents and intermediates, minimal
evaporative solvent loss during long periods of heating,
and compatibility with the ‘release’ and purification steps.
20. In our previous work with P-BEMP sequestered 3-thio-
1,2,4-triazoles, ACN was an optimal solvent when
employing alkyl halides and ethanol was optimal for
nucleophilic epoxide ring opening. To our surprise, use of
ACN for preparation of 9 frequently is problematic
owing to poor product solubility in ACN. Further, epox-
ide ring opening in ethanol provides products 10 with low
purity (<40%). In contrast, S-alkylation of ion-pair 4 in
DMSO is reliable with both alkyl halides and epoxides.
While use of DMSO in the ‘release’ step (for both epox-
ides and alkylating agents) does require introduction of a
purification step, a single, robust, production protocol
(with fewer product solubility issues and high success
rates) is very convenient.
21. Although not emphasized in this report, the correspond-
ing 2-thioxo-2,3-dihydro-1H-quinazolin-4-ones 3 are also
obtained in excellent purity (70–100%) after treatment of
4 with dilute AcOH/ACN solution.
22. Purity data reported here was determined by a C18
reverse phase HPLC column, Keystone Aquasil (1×40
mm) in 10–90% ACN/H2O containing 0.02% TFA (3.6
min gradient) and monitored at 214 nm using a UV
detector and by a SEDEX 75 evaporative light scattering
detector (ELSD) operating at 42°C. LCMS M+H signals
were consistent with expected MW for all reported
products.
‘Release’ step for preparation of 12a: Excess solution was
aspirated to waste. The resin was washed with 1 mL
portions of DMF (3×) then ACN (2×). A solution of
1-bromopentane 7a (0.07 M, 1 mL, 0.07 mmol, 0.7
equiv.) in acetonitrile was then added to the resin and the
suspension was mixed for 16 h at room temperature. The
solution containing quinazolinone 12a was drained into a
tared tube. The resin was then washed with 1 mL por-
tions of warm acetonitrile (5×). All filtrates were collected
in the tube. Evaporation of ACN provided quinazolinone
12a as a clear glass (24 mg, 97% yield based on alkyl
halide). LC/MS [M+H]=339.0; 100% purity (UV214
1
nm). H NMR (400 MHz, d6-DMSO) l: 7.97 (d, J=8.1
Hz, 1H), 7.68 (t, J=7.7 Hz, 1H), 7.43 (d, J=8.1 Hz,
1H), 7.34 (t, J=7.5 Hz, 1H), 7.15 (m, 5H), 5.19 (s, 2H),
3.10 (t, J=7.3 Hz, 2H), 1.54 (m, 2H), 1.19 (m, 4H), 0.72
(t, J=7.1 Hz, 3H).
26. Apogent Discoveries, Hudson, NH 03051; Argonaut
Technologies, Inc., Foster City, CA 94404.
27. Ruelke, H.; Martin, E.; Kuehmstedt, K. K. Pharmazie
1990, 45, 862–863.
28. In our triazole work, the performance of P-BEMP was
23. Purification was carried out using a semipreparative
YMC Combiprep ODS-A reversed-phase column (20
mm×50 mm, particle size S-5 mm, 750 mL injection vol-
ume) via use of a 10–95% gradient of water/acetonitrile (4
min gradient, 25 mL/min flow rate) on a Gilson HPLC
system.
24. The generic protocols reported here were developed to
maximize scope and product diversity. In cases where
specific individual compounds (or a more narrow scope)
are desired, significantly improved purity and yield are
typically achieved by straightforward tuning of the
sequence (variation of temperature/time).
superior (faster/cleaner reactions) to
a resin-bound
guanidine base (P-TBD, Fluka cat. c90603). This is
presumably due to P-BEMP’s increased basicity, lower
nucleophilicity, and better dispersion properties especially
in these polar solvents.
29. Lipinski, C. A. J. Pharm. Tox. Methods 2000, 4, 235–249.
30. When the purified yield data for this manuscript was
collected in our laboratory, it was customary to experi-
ence a significant loss (ꢀ50%) owing to cumulative losses
during purification and the post-synthesis processing
(transfers, fractionation, fraction combination etc.). Fur-
ther, additional yield loss was expected here as the maxi-
mum injection volume allowed by the HPLC purification
system (0.75 mL) limited the volume of DMSO used to
wash resin after product release.
25. Synthetic protocols for quinazolinone 12a and represen-
tative 48-member array:
Formation of 4 from anthranilic acid esters 1: DMSO
solutions of anthranilic ester 1 (0.3 M, 0.43 mL, 0.13
mmol, 1.3 equiv.) and isothiocyanate 2 (0.4 M, 0.38 mL,
0.15 mmol, 1.5 equiv.) were mixed and heated at 100°C
for three days. DMF (1 mL) and P-BEMP (45 mg, 0.1