Solid-phase versus solution synthesis of asymmetrically disubstituted furazano[3,4-b]pyrazines

Article abstract of DOI:10.1016/S0040-4039(02)00913-9

Herein we describe a straightforward solid-phase synthesis directed towards the preparation of families of asymmetrically disubstituted furazano[3,4-b]pyrazines by stepwise displacement of the two chlorine atoms in 5,6-dichlorofurazano[3,4-b]pyrazine by nucleophiles. This synthesis has avoided selectivity problems found in solution chemistry.

Full text of DOI:10.1016/S0040-4039(02)00913-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 4741–4745
Solid-phase versus solution synthesis of asymmetrically
disubstituted furazano[3,4-b]pyrazines
E. Ferna´ndez,a S. Garc´ıa-Ochoa,a S. Huss,a A. Mallo,a J. M. Bueno,a,* F. Micheli,b A. Paio,b
E. Pigab and P. Zarantonellob
aGlaxoSmithKline S.A. P.T.M., Severo Ochoa, 2 E-28760 Tres Cantos, Madrid, Spain
bGlaxoSmithKline Medicine Research Centre, Via Alessandro Fleming, 4-37135 Verona, Italy
Received 6 May 2002; revised 7 May 2002; accepted 13 May 2002
Abstract—Herein we describe a straightforward solid-phase synthesis directed towards the preparation of families of asymmetri-
cally disubstituted furazano[3,4-b]pyrazines by stepwise displacement of the two chlorine atoms in 5,6-dichlorofurazano[3,4-
b]pyrazine by nucleophiles. This synthesis has avoided selectivity problems found in solution chemistry. © 2002 Published by
Elsevier Science Ltd.
The furazano[3,4-b]pyrazine heterocyclic system has
attracted much attention from researchers as an exam-
ple of energy-rich polynitrogenated compounds.1 In
addition, it has been found that compounds structurally
related to furazano[3,4-b]pyrazines have also shown
biological activity in the field of antimicrobials,2 herbi-
cides and plant growth regulators.3
In order to study the antibacterial properties shown by
this family of compounds and to improve physicochem-
ical properties such as its low solubility in water, we
started a systematic chemical program aimed at the
preparation of both symmetrical and asymmetrical 5,6-
disubstituted derivatives. Thus, compounds of general
formula I and II became of great interest to us.
As a result of our High Throughput Screening activities
aimed at finding new antibacterials active against S.
aureus and E. coli, we have found that several 5,6-di-
substituted furazano[3,4-b]pyrazine derivatives such as
1 and 2 have exhibited activity against Gram-positive
bacteria.
Polychloride-substituted nitrogen heterocycles such as
2,4,6-trichloro-1,3,5-triazine allows for the stepwise
nucleophilic substitution of the chlorine atoms in solu-
tion leading to heterocycles bearing different residues at
every position substituted.4 However, in our case sub-
stitution of the first chlorine does not inactivate the
substrate enough to prevent the second substitution
from occurring. For this reason, asymmetric substitu-
tion in solution was a difficult task because it led to
complex mixtures of symmetrical and asymmetrical
products which needed careful chromatographic purifi-
cations. In this way, solution synthesis of compounds I
and II (Scheme 1) was tackled by reaction between
dichloroderivative 3 and 2 equiv. of 3-trifluoromethyl-
aniline (1 equiv. as hydrogen chloride acceptor), fol-
lowed by the introduction of a nitrogen or sulphur
nucleophile according to
a procedure previously
reported.5
Despite the great reactivity showed by dichloroderiva-
tive 3 towards nitrogen nucleophiles, we have observed
that the displacement of just one chlorine atom can be
partially controlled by using weak bases such as the
corresponding aniline as hydrogen chloride acceptors
* Corresponding author. Tel.: +34 91 807.05.37; e-mail:
jmb17723@gsk.com
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)00913-9

4742
E. Ferna´ndez et al. / Tetrahedron Letters 43 (2002) 4741–4745
V which were of therapeutic interest within the antibac-
terial project.
The general solid-phase synthesis of target molecules III
is depicted in Scheme 2. The attachment of our first set
of monomers, anilines, was achieved via reductive ami-
nation carried out on a di- or tri-alkoxyaldehyde loaded
into a solid support, resulting in an acid labile benzyl
linker.6,7 Several conditions were tried including differ-
ent solvents (DCM, DMF, DMF/TMOF), reducing
agents (Me4N(OAc)3BH, NaBH3CN, Na(OAc)3BH),
additives (AcOH, benzotriazole), reaction times and
temperature. Best conditions were obtained with 2 M
Na(OAc)3BH in DMF/1% acetic acid at room tempera-
ture, carrying out the reductive alkylation twice. Yield
evaluation was performed by acylation of a small
amount of resin 7 with acetyl chloride, cleavage with
95% trifluoroacetic acid (TFA) in dichloromethane
Scheme 1. Reagents and conditions: (a) 3-(Trifluoromethyl)-
aniline (2 equiv.), acetonitrile, room temperature; (b) primary
or secondary amines, acetonitrile, room temperature; (c)
sodium thiolate, acetonitrile, room temperature.
and hindered nucleophiles.5 However, although this
strategy decreases the amount of symmetrical deriva-
tives formed, it does not avoid tedious chromato-
graphic purifications and also narrows the scope of
anilines available.
1
(DCM) and H NMR of the residue with 2,5-dimethyl-
furan (DMFu) as internal standard.8 Yields measured
in this way ranged from 55 to 100%. In general, lowest
yields were obtained with those anilines bearing elec-
tron-withdrawing groups such as the m-trifluoromethyl-
aniline and the ones specially hindered such as the
o-chloro-p-bromoaniline.
At this point, we decided to design a new strategy for
the preparation of asymmetrical derivatives based on
the use of solid-phase organic synthesis, which would
allow us to perform the selective attachment of 5,6-
dichlorofurazano[3,4-b]pyrazine to a solid support by
displacement of one chlorine atom, leaving the second
one available for an additional nucleophilic attack with
a variety of nucleophiles. Hopefully, the use of solid
supports not only would avoid the formation of sym-
metrical derivatives but would also allow us to get the
final products in high purity after cleavage. Target
compounds were aimed at general formulas III, IV and
Optimization of the introduction of dichloroderivative
3 was the following step. The highly reactive 5,6-dichlo-
rofurazano[3,4-b]pyrazine 3 was prepared as a stable,
tractable and crystalline solid, in multigram scale and in
very good yield and high purity from glyoxal, hydroxyl-
amine hydrochloride and oxalic acid following the syn-
thetic route depicted in Scheme 3.1,5
Scheme 2. Reagents and conditions: (a) (i) aniline 1 M, DMF, 2% AcOH, 1 h, room temperature; (ii) NaBH(OAc)3, overnight,
room temperature (two couples); (b) 5,6-dichlorofurazano[3,4-b]pyrazine 3 (5 equiv.), DIEA, ACN, room temperature, 7 h; (c)
primary amines 1 M, DIEA, ACN, room temperature, overnight; (d) TFA 25%/DCM, 3 h, room temperature.

E. Ferna´ndez et al. / Tetrahedron Letters 43 (2002) 4741–4745
4743
them were chosen for the production of the above
proposed libraries: p-fluoro, p-bromo, m-trifluoro-
methyl and p-methylaniline. An initial set of more than
200 amines were used to generate a virtual library using
ADEPT tool11 (A Daylight Enumeration and Profiling
Tool) and it was refined by cluster analysis and pre-
dicted solvation energies as a way to improve the
solubility of final compounds. After chemical rehearsal,
80 primary amines from the 116 initial set were selected
including alkyl, benzyl and phenethyl amines as well as
a-aminoacids and other amines with additional funtion-
alities. In general, a wide range of amines worked well
and only the most polar ones failed to react, probably
due to poor solubility in the reaction conditions. Spe-
cifically non-protected a-aminoacids, aminoalcohols
and some heterocyclic amines did not work in our
reaction conditions.
Scheme 3. Reagents and conditions: (a) aq. NaOH, hydroxyl-
amine, 5°C to reflux; (b) aq. KOH, 180°C, steel reactor; (c)
oxalic acid, aq. HCl, reflux; (d) PCl5/POCl3, reflux.
Attachment of 3 to the resin was best achieved with
acetonitrile as solvent at room temperature.9 DCM and
longer reaction times (overnight) were the conditions
initially used but we found problems when using
HiTOPS microtiter plate reaction block. HiTOPS
device keeps solvent and reagents in the fritted micro-
titer plate wells by applying an inert gas pressure at the
bottom of the plate. The nitrogen bubbling causes
dryness of the resin when DCM is used as solvent.
Consequently, reaction remains uncompleted and even-
tually moisture exposure causes some hydrolysis to
produce variable amount of 6-anilino-5-hydroxyfura-
zano[3,4-b]pyrazine. To avoid this problem we finally
used MeCN as solvent and 7 h treatments. Due to the
high reactivity of the second chlorine, yield and purity
were not determined by direct cleavage, but instead,
reaction with isoamylamine (1 M) in the presence of 1
M DIEA in MeCN was carried out and cleavage with
TFA in DCM done afterwards. The low basicity of the
nitrogen attached to the resin made possible to reduce
the amount of trifluoroacetic acid needed for cleavage
to 25%.
All final samples of compounds of general structure III
were analyzed by LC–MS and 10% of products were
weighted for yield evaluation (Table 1).12 97% of
desired compounds were present in the library and
purities depended on the aniline used: compounds bear-
ing p-fluoroaniline have purities between 70 and 95%,
p-bromo have purities between 65 and 85%, m-tri-
fluoromethyl have purities above 50% (impurified with
m-trifluoromethylaniline) and finally, p-methyl deriva-
tives have purities greater than 80%. In summary,
approximately 75% of compounds have purities higher
than 70%. Variable amounts of m-trifluoromethylani-
line found in the final samples might have originated
because the remaining imine did not reduce in the
reductive amination step.
To produce libraries of general structures IV and V we
followed the same approach as in the above case
(Scheme 1), but introduction of the second nucleophile
needed optimization. A wide range of conditions were
tried changing solvent (DCM, acetonitrile, dioxane and
DMF), temperature (from room temperature to 100°C)
and base (DIEA and N,N-diethylaniline). Solvents and
temperature did not have a great influence improving
final yields and purities. However, the use of a weaker
base than DIEA gave much better results, in accor-
dance with the results obtained in solution chemistry.
Best conditions found were the use of acetonitrile as
solvent at room temperature and N,N-diethylaniline as
base. After monomer rehearsal 12 anilines and 17
hydrazines were chosen. In this case m-trifluoromethyl-
aniline was not included in the libraries due to the
moderate yields obtained in the first library. Production
of both libraries was performed in the HiTOPS device
as in the above library and quality control was carried
out on all samples by LC–MS obtaining a medium
purity of 70% for all the compounds.
Four different resins were evaluated (Scheme 2). Two
of them bearing PAL linker (5-(4-formyl-3,5-
dimethoxy-phenyloxy)valeric acid)9 supported either in
polystyrene or TentaGel™ resins, (4-formyl-3-
methoxyphenoxy)methyl in Argopore resin6 and 2-(4-
formyl-3-methoxyphenoxy)ethyl polystyrene.10 PAL
linker based resins were initially used due to their
milder cleavage conditions but no results were obtained
when the reductive amination was carried out probably
due to steric hindrance caused by the two methoxy
groups vicinal to the aldehyde. With the Argopore
derivatized resin, a highly-crosslinked macroporous
polystyrene, desired final compounds were obtained in
moderate yields but the benzyl ether spacer seemed to
be sensitive to the cleavage conditions and by-products
containing this residue were obtained. Best results in
yields and purity were obtained with 2-(4-formyl-3-
methoxyphenoxy)ethyl polystyrene which was then
selected for the generation of the libraries.
In conclusion, we have developed a successful new
solid-phase approach to libraries of asymmetrically di-
substituted furazano[3,4-b]pyrazines that overcome the
problems found in solution chemistry, allowing the
sequential introduction of two different substituents
without the formation of undesired symmetrical deriva-
Once the route of synthesis was optimized the monomer
validation was performed. After reaction assessment
with anilines that already showed good antibacterial
activity in the furazano[3,4-b]pyrazine nucleus, four of

4744
E. Ferna´ndez et al. / Tetrahedron Letters 43 (2002) 4741–4745
Table 1. HPLC purity from selected compounds of libraries III–IV
Compd
R1
R2
HPLC puritya
MSb
III.1
III.2
III.8
p-F
p-F
p-F
p-F
p-F
p-F
p-Br
p-Br
p-Br
p-CH3
p-CH3
p-CH3
p-F
p-F
p-F
p-Br
p-CH3
p-CH3
p-F
1-Phenylethyl
1-Cyclohexyl
91
93
87
93
95
70
75
84
75
72
89
89
88
93
86
60
88
92
82
86
68
80
61
73
92
351
329
397
373
289
425
391
398
444
356
386
351
339
363
380
474
349
334
466
396
398
336
400
378
335
3,4-Dimethoxybenzyl
2,6-Difluoro-benzyl
Propyl
3-(4-Hydroxyphenyl)-1-(methoxycarbonyl)-1-ethyl
Tetrahydrofuran-2-ylmethyl
2-Pyridylmethyl
III.10
III.14
III.27
III.100
III.111
III.129
III.258
III.263
III.296
IV.1
3-Phenyl-1,2-propanediolc
2-(N-Morpholino)ethyl
2-(1H-Indol-3-yl)ethyl
3-(Imidazol-1-yl)propyl
4-Hydroxyphenyl
IV.6
1H-Indazol-5-yl
IV.10
IV.17
IV.26
IV.30
V.4
V.12
V.19
V.20
V.39
4-(Methylcarbamoyl)phenyl
2-(Phenylamino)phenyl
4-Hydroxy-2-methylphenyl
4-Aminophenyld
5-Methoxycarbonyl-4-(trifluoromethyl)pyrimidin-2-yl
4-(Methoxycarbonyl)phenyl
Phenyl
Methyl
2-Quinoxalinyl
4-Carboxylphenyl
2-Pyridyl
p-F
p-Br
p-Br
p-CH3
p-CH3
p-CH3
V.42
V.51
a Determined at 230 nm.
b M+1.
c Obtained from the protected diol.
d Obtained from the Boc-protected amine.
tives and with good yields and purities. By using these
solid-phase synthetic protocols, we have synthesized
320 compounds of general structure III, 36 IV and 51
V. Antibacterial activity was determined and will be
published in due course.
S. Tetrahedron 1999, 55, 2827–2834; (c) Yu, K.-L.;
Civiello, R.; Roberts, D. G. M.; Seiler, S. M.; Meanwell,
N. A. Bioorg. Med. Chem. Lett. 1999, 9, 663–666.
7. Bui, C. T.; Bray, A. M.; Pham, Y.; Campbell, R.; Ercole,
F.; Rasoul, F. A.; Maeyi, N. J. Mol. Diversity 1998, 4,
155–163.
8. Gerritz, S. W.; Sefler, A. M. J. Comb. Chem. 2000, 2,
39–41.
9. Alsina, J.; Yokum, T. S.; Albericio, F.; Barany, G. J.
Acknowledgements
Org. Chem. 1999, 64, 8761–8769.
The authors would like to thank J. M. Viana for the
technical support in the production of libraries, M. T.
Quesada for the preparation of starting materials and
our colleagues of Structural Chemistry for analytical
support.
10. Mazurov, A. Tetrahedron Lett. 2000, 41, 7–10.
11. Leach, A. R.; Bradshaw, J.; Gree, D. V. S.; Hann, M. M.
J. Chem. Inf. Comput. Sci. 1999, 39, 1161–1172.
12. General procedure for the library was as follows: (a)
Aldehyde resin was shaken with a solution of the corre-
sponding aniline (1 M) in dry DMF with 1% acetic acid
for 1 h at room temperature. Sodium triacetoxyborohy-
dride (2 M) was added as solid and the suspension was
shaken overnight at room temperature. Reactants were
filtered and the resin was washed with MeOH (4×), water
(4×), 2% AcOH in water (4×), water (4×), MeOH (4×),
DCM (4×), DMF (4×) and DCM (3×). Reaction was
repeated in the same conditions as above. Loading was
evaluated by acetylation and cleavage with 95% TFA/
DCM of a small amount of resin. Cleaved product was
References
1. Zelenin, A. K.; Trudell, M. L. J. Heterocyclic Chem.
1997, 34, 1057–1060 and references cited therein.
2. Tong, Y. C. US Patent 3,850,929, 1974.
3. Clark M. T.; Gilmore I. J. UK Patent 2,122,492 A, 1984.
4. (a) Desai, V. A.; Mistry, N. V.; Desai, K. R. Proc. Natl.
Acad. Sci., India, Sect. A 1992, 62, 743–745; (b) Prakash,
L.; Tyagi, E.; Shaila; Mital, R. L. Pharmazie 1990, 45,
284.
1
evaluated by H NMR with 2,5-dimethylfurane as inter-
nal standard. (b) Aniline resin 7 was spread in a HiTops
5. Starchenkov, I. B.; Andrianov, V. G. Chem. Heterocyclic
Comp. 1997, 33, 1219–1233.
6. (a) Fivush, A. M.; Willson, T. M. Tetrahedron Lett. 1997,
38, 7151–7154; (b) Ouyang, X.; Tamayo, N.; Kiselyov, A.
plate as
a
suspension in N-methylpyrrolidone/
dichlorobenzene 1:1. Resin was dried, washed with DMF
(4×) and dry acetonitrile (3×), and shaken with a solution
of 3 (5 equiv.) in dry acetonitrile and diisopropylethyl-

E. Ferna´ndez et al. / Tetrahedron Letters 43 (2002) 4741–4745
4745
(4×) and DCM (4×). Resin was dried in vacuo. (d) Resin
was shaken with solution of TFA 25% in
dichloromethane for 3 h at room temperature. Solution was
filtered and resin was washed with EtOH (1×) and DCM
(1×). Filtrates were evaporated in vacuo and products
dissolved in methanol (HPLC grade) to LC–MS analysis.
amine (DIEA) (5 equiv.) for 7 h at room temperature.
Reactants were dried and resin directly used in the next
reaction. (c) Resin 8 was shaken with a solution of amine
(1 M) in dry acetonitrile and diisopropylethylamine (DIEA)
(1 M) overnight at room temperature. Reactants were dried
and resin was washed with DCM (4×), DMF (4×), THF
9
a