Studies on the synthetic compatibility of aryloxime linkers in the solid-phase synthesis of 3-aminobenzisoxazoles

Article abstract of DOI:10.1021/jo991263i

Further exploration of the scope of our solid-phase method for the synthesis of 3-aminobenzisoxazoles (using the Kaiser oxime resin 1) is described. The effects of base, leaving group, and solvent on the nucleophilic aromatic substitution based resin-loading reaction are discussed. Representative aryloxime intermediates were subjected to a variety of acidic conditions commonly used in protecting group removal to establish the acid stability profile of this linker. Regioselectivity was evaluated with various di- and trifluorobenzonitriles, which gave single benzisoxazole products after loading and cyclorelease reactions. Substituent effects observed in the course of the acid stability and regioselectivity studies suggest that the nitrile plays a critical role in the oxime hydrolysis mechanism. Finally, to establish the compatibility of the aryloxime linker with a variety of useful on-resin synthetic transformations, functionalized substrates were loaded onto resin 1, and carbon-nitrogen, carbon-oxygen, and carbon-carbon bond-forming reactions were successfully executed.

Full text of DOI:10.1021/jo991263i

2924
J . Org. Chem. 2000, 65, 2924-2932
Stu d ies on th e Syn th etic Com p a tibility of Ar yloxim e Lin k er s in
th e Solid -P h a se Syn th esis of 3-Am in oben zisoxa zoles
Salvatore D. Lepore† and Michael R. Wiley*
Eli Lilly and Company, Indianapolis, Indiana
Received August 9, 1999
Further exploration of the scope of our solid-phase method for the synthesis of 3-aminobenzisoxazoles
(using the Kaiser oxime resin 1) is described. The effects of base, leaving group, and solvent on the
nucleophilic aromatic substitution based resin-loading reaction are discussed. Representative
aryloxime intermediates were subjected to a variety of acidic conditions commonly used in protecting
group removal to establish the acid stability profile of this linker. Regioselectivity was evaluated
with various di- and trifluorobenzonitriles, which gave single benzisoxazole products after loading
and cyclorelease reactions. Substituent effects observed in the course of the acid stability and
regioselectivity studies suggest that the nitrile plays a critical role in the oxime hydrolysis
mechanism. Finally, to establish the compatibility of the aryloxime linker with a variety of useful
on-resin synthetic transformations, functionalized substrates were loaded onto resin 1, and carbon-
nitrogen, carbon-oxygen, and carbon-carbon bond-forming reactions were successfully executed.
Sch em e 1. Rea ction of 2-F lu or oben zon itr ile w ith
Ka iser -DeGr a d o Resin 1 F ollow ed by Cyclitive
Rem ova l
In tr od u ction
In connection with our efforts to identify ligands for a
variety of biological targets, we have been interested in
the synthesis of 3-aminobenzisoxazoles.1 To take advan-
tage of parallel synthesis techniques,2 we have recently
developed a solid-phase adaptation of the Shutske method3
for the synthesis of these interesting heterocycles.4 In our
initial disclosure, we reported that the potassium salt of
the Kaiser-Degrado resin5 1 (Scheme 1) could react with
2-fluorobenzonitrile to give the nucleophilic aromatic
substitution adduct. This aryl oxime intermediate was
then treated with aqueous acidic conditions to effect a
cyclitive6 removal (or cyclorelease7) of the substrate giving
3-aminobenzisoxazole.
Although acylated derivatives of 1 have been widely
utilized for the solid-phase synthesis of amides and
related structures,8 our method involves the first ap-
plication of the Kaiser oxime resin to SNAr reactions.
Moreover, this linker strategy can be thought of as
“traceless” since the desired heterocyclic product contains
no residual functionality as a result of resin attachment.9
Finally, this approach to the synthesis of 3-aminoben-
zisoxazoles is particularly efficient since the cyclization
reaction also serves as the cleavage step.
In this paper, we report further exploration of this
methodology, including the effect of the base and the
leaving group on the loading reaction, as well as studies
on the stability of the aryl oxime linker under a variety
of acidic conditions commonly used for protecting group
removal. Finally, to establish the viability of this meth-
odology for library generation, functionalized aryl oxime
intermediates have been prepared and subjected to a
variety of organic transformations that are widely used
for the formation of C-N, C-O, and C-C bonds.
† Current address: Department of Chemistry, Florida Atlantic
University, Boca Raton, FL 33431.
(1) Suh has recently incorporated this heterocycle in the design of
potent LTB4 receptor antagonists. Suh, H.; J eong, S.; Han, Y. N.; Lee,
H.; Ryu, J . Bioorg. Med. Chem. Lett. 1997, 7, 389.
(2) Kaldor, S. W.; Siegel, M. G. Curr. Opin. Chem. Biol. 1997, 1,
101. Balkenhohl, F.: von dem Bussche-Hu¨nnefeld, C.; Lansky, A.;
Zechel, C. Angew. Chem., Int. Ed. Engl. 1996, 35, 2288.
(3) Shutske, G. M.; Kapples, K. J . J . Heterocycl. Chem. 1989, 26,
1293. For other examples of solution-phase intramolecular nucleophilic
additions to nitrile by nitrogen, see: Kwok, R.; Pranc, P. J . Org. Chem.
1967, 32, 738. Blicke, F. F.; Zambito, A. J .; Stenseth, R. E. J . Org.
Chem. 1961, 26, 1826.
(4) Lepore, S. D.; Wiley: M. R. J . Org. Chem. 1999, 64, 4547.
(5) Degrado, W. F.; Kaiser, E. T. J . Org. Chem. 1980, 45, 1295.
(6) Peng, G.; Gallop, M. A. Book of Abstracts. 216th ACS National
Meeting; Boston,Aug 23-28, 1998; American Chemical Society: Wash-
ington, DC, 1998.
Resu lts a n d Discu ssion
Loa d in g Rea ction . Nucleophilic aromatic substitu-
tion reactions have been extensively studied in solution.10
However, the extent to which this precedent could be
applied to predict the outcome of these heterogeneous
reactions was not clear given the importance of resin-
swelling properties, site accessibility, etc.11 Therefore, we
studied the effect of the solvent, leaving group, and
(7) Backes, B. J .; Ellman, J . A. Curr. Opin. Chem. Biol. 1997, 1, 86.
(8) O¨ sapay, G.; Profit, A.; Taylor, J . W. Tetrahedron Lett. 1990, 31,
6121. Scialdone, M. A. Tetrahedron Lett. 1996, 37, 8141. Golebiowski,
A.; Klopfenstein, S. Tetrahedron Lett. 1998, 39, 3397. Smith, R. A.;
Bobko, M. A.; Lee, W. Biorg. Med. Chem. Lett. 1998, 8, 2369.
(9) Van Maarseveen, J . H. Comb. Chem. High. Throughput Screen-
ing 1998, 1, 185.
(10) Nudelman, N.; Mancini, P. M. E.; Martinez, R. D.; Vottero, L.
R. J . Chem. Soc., Perkin Trans. 2 1987, 951. Miller, J . Aromatic
Nucleophilic Substitution; Elsevier: Amsterdam, 1968. Bernasconi, C.
F. Acc. Chem. Res. 1978, 11, 147.
10.1021/jo991263i CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/22/2000

Solid-Phase Synthesis of 3-Aminobenzisoxazoles
J . Org. Chem., Vol. 65, No. 10, 2000 2925
Ta ble 1. Effect of Lea vin g Gr ou p , Solven t, a n d
Cou n ter ion on Loa d in g Rea ction s
Ta ble 2. Acid Sta bility P r ofile of th e Ar yl Oxim e Lin k er
% loading yield of 3a
entry
1
X
base
THF
MeCN
DMF
DMSO
F
KOBut
78
62
42
<5
72
<5
<5
<5
13
14
20
<5
28
59
60
63
46
41
<5
66
63
65
67
54
15
KHMDS
NaHMDS
LiHMDS
KOBut
2
3
4
5
NO2
Cl
Br
I
KOBut
<5
KOBut
KOBut
a
Determined by resin weight difference (average of three
experiments)
counterion on the solid-phase loading reaction (Table 1).
In comparing THF, MeCN, DMF, and DMSO using
potassium tert-butoxide (KOBut) to form the oxime anion,
THF gave the best results, providing a 78% loading yield
with 2-fluorobenzonitrile (2a ) and 72% with 2-nitroben-
zonitrile (2b).12 A 10-30% decrease in the loading yield
was observed for both 2a and 2b when DMF or DMSO
was used as the solvent. Loading yields also decreased
dramatically with the use of chloro- (2c), bromo- (2d ),
and iodobenzonitriles (2e).
a
Method A ) 4:1 TFA/5 N HCl, 55 °C. Method B ) 99:1 TFA/
H2O, 55 °C. Method C ) 25% THF/CH2Cl2, rt. Method D ) TFA,
55 °C. Method E ) AcOH/THF/H2O, 55 °C. Method F ) TsOH/
THF/H2O, 55 °C. Isolated yield after chromatography and based
b
on loading. Crude purity based on HPLC analysis.
The effect of the counterion on the loading reaction was
evaluated with 2-fluorobenzonitrile as the substrate in
several reaction solvents (Table 1, entry 1). In THF, the
potassium salt of resin 1 gave the highest loading yields
with KOBut giving slightly better results than potassium
hexamethyldisilazide (KHMDS). Treatment of 1 with
sodium hexamethyldisilazide (NaHMDS) gave a loading
yield of 42%. Interestingly, with THF no loading was
observed with lithium hexamethyldisilazide (LiHMDS)
while in DMF and DMSO loading yields of 46% and 67%
were observed. Other bases such as KOBut, KHMDS, and
NaHMDS also gave reasonable loading yields in DMF
and DMSO. In contrast, loading reactions performed in
acetonitrile gave consistently poor results for each of the
bases mentioned above.
As described in our previous report, a variety of
2-fluorobenzonitriles can be loaded on the Kaiser resin.4
Although the presence of an electron-withdrawing group
facilitates loading at room temperature, substrates bear-
ing either electron-withdrawing groups or electron-
donating groups can be loaded in high yield in about 2 h
at 55 °C. We also found that the loading reaction of
2-fluorobenzonitriles is not particularly sensitive to steric
hindrance around the site of nucleophilic substitution.
Stu d ies on th e Acid Sta bility of th e Ar yl Oxim e
Lin k er . As previously reported, the use of 4:1 TFA/
aqueous 5 N HCl at 55 °C for 2 h (Table 2, method A)
led to an efficient cyclorelease of a variety of substituted
aminobenzisoxazoles. By contrast, we saw that the use
of 99:1 TFA/H2O at 55 °C (Table 2, method B) with
intermediates bearing either inductively neutral or elec-
tron-withdrawing substituents required significantly
longer reaction times to reach completion. For example,
hydrolysis of the resin bearing a bromo substituent para
to the nitrile (Table 2, entry 2) using method B required
4 days to give the corresponding aminobenzisoxazole 4b
in 53% yield. On the other hand, treatment of the
p-methoxybenzonitrile derivative 3c (Table 2, entry 3),
under the same conditions, gave the methoxy-substituted
product 4c in 85% isolated yield after only 2 h.
While the studies summarized above were important
for identifying useful cyclorelease conditions, further
experiments were performed in order to determine the
acid stability profile of the linker under a variety of
conditions commonly used for the removal of acid-
sensitive protecting groups. The methoxy-substituted
resin 3c was selected for this study, since it is the most
acid labile and therefore represents the worst case
scenario. Upon treatment of the methoxy-substituted
resin 3c with 25% anhydrous TFA/CH2Cl2 at rt for 2 h
(Table 2, entry 3, method C), conditions precedented for
on-resin Boc removal,4,5 <5% of the cyclization product
4c was removed from the resin. Even under more forcing
conditions, with 100% anhydrous TFA at 55 °C (Table 2,
entry 3, method D) only 9% of the 3-aminobenzisoxazole
product (>96% purity) was removed from the resin after
2 h. The aryl oxime linker was also stable to a number
of milder aqueous acidic conditions that have been used
for the removal of THP, silyl, and acetal protecting groups
in solution. Thus, treatment of the resin 3c with AcOH/
(11) Yan, B.; Fell J . B.; Gnanasambandam, K. J . Org. Chem. 1996,
61, 7467.
(12) Palermo, M. G. Tetrahedron Lett. 1996, 37, 2885.

2926 J . Org. Chem., Vol. 65, No. 10, 2000
Lepore and Wiley
Sch em e 3. On -Resin Mitsu n obu Rea ction
Sch em e 2. On -Resin Hyd r olysis a n d Am id e Bon d
F or m a tion
prepared by reacting the potassium anion of the Kaiser
resin 1 with 2-fluoro-4-(tert-butyldimethylsilyloxymeth-
yl)benzonitrile. The TBS protecting group was removed
using TBAF in THF. The on-resin alcohol was then
treated with p-chlorophenol, triphenylphosphine, and
diisopropylazodicarboxylate (DIAD) in THF.17 The best
results were observed for reactions times of 3 h. Longer
reaction times generally led to decreased purity in the
crude cyclization product. Cyclitive removal using the
standard conditions then gave aryl ether 9 in a 77% yield
(three steps) and 83% crude purity.
On -Resin Nu cleop h ilic Ar om a tic Su bstitu tion .
Recently, examples of on-resin nucleophilic aromatic
displacement reactions have appeared in the literature.18
To evaluate the compatability of the aryl oxime linker
with this reaction, a resin-attached fluorobenzonitrile
(10) was prepared (Scheme 4). To avoid any regioselec-
tivity issues in the loading reaction, the symmetrical 2,6-
difluorobenzonitrile was chosen for this initial investi-
gation. Resin 10 was then reacted with a variety of
nucleophiles including an alkoxide,19 amine,20 and phen-
ol21 (Table 3). The potassium alkoxide of 4-chloropheneth-
anol (Table 3, entry 1) was prepared by treatment with
2 equiv of potassium tert-butoxide. This salt was then
added to a THF suspension of resin 10 to give the
corresponding product 11a after cyclorelease in a two-
step yield of 50%.
Addition of pyrrolidine to 10 proved more challenging
(Table 3, entry 2). In our initial attempt (THF, 6 h, 55
°C), the desired product 11b was produced in low isolated
yield as a result of incomplete displacement. The major
impurity in this reaction was 4-fluoro-3-aminobenzisox-
azole, the cyclization product of the unreacted starting
material (10). In an attempt to increase the rate of the
SNAr reaction in THF, variations in the reaction time,
number of equivalents of nucleophile, and the tempera-
ture were evaluated. Unfortunately, none of these alter-
nate protocols were found to improve the outcome of the
reaction. However significant improvements were ob-
tained with the use of alternate solvents. Although only
a slight improvement was observed with acetonitrile,22
both DMF and DMSO produced a significant increase in
the reaction rate. After 6 h at 55 °C in DMSO, followed
THF/H2O (3/1/1) at 55 °C for 12 h13 (Table 2, entry 3,
method E) or with TsOH/THF/H2O at 55 °C for 12 h14
(Table 2, entry 3, method F) gave no detectable release
of material from the resin.
Am id e Bon d F or m a tion . Due to the importance of
solid-phase peptide synthesis, amide bond forming reac-
tions have been the most widely performed and most
highly developed solid-phase reactions.15 We have previ-
ously shown that amides can be formed in the presence
of an aryl oxime linker by the reaction of an acid chloride
or an acid anhydride with an aryl oxime-linked benzyl-
amine.4 In the present study, we set out to demonstrate
that an aryl oxime-linked acid could be coupled to an
amine under standard peptide coupling conditions. By
proceeding through the intermediacy of methyl ester 5,
we also sought to demonstrate the compatibility of the
aryl oxime linker with the aqueous basic conditions
required for saponification (Scheme 2). Thus, the potas-
sium anion of the Kaiser resin 1 was coupled with methyl
3-fluoro-4-cyanobenzoate to give resin 5 in a 69% loading
yield. Treatment of resin 5 with LiOH in THF/MeOH/
H2O (3/1/1) at room temperature gave the corresponding
acid (6) and no removal of the substrate from the resin
was observed, demonstrating the stability of the aryl
oxime linker under these conditions. Acid 6 was then
coupled to 4-chlorobenzylamine to give the on-resin
amide. Both the resin weight increase and chlorine
analysis of the intermediate suggested that the coupling
reaction went essentially to completion within 12 h. The
resin was then treated with the standard cyclorelease
conditions to give the desired amide 7 in a three step yield
of 81% (93% crude purity).
P h en olic Mit su n ob u R ea ct ion . On-resin carbon-
oxygen bond formation via the Mitsunobu reaction16 has
been identified as an important tool in combinatorial
chemistry.15 Application of this reaction to an aryl oxime-
linked substrate is shown in Scheme 3. Resin 8 was
(16) Mitsunobu, O.; Yamada, M. Bull. Chem. Soc. J pn. 1967, 40,
2380. Mitsunobu, O. Synthesis 1981, 1.
(17) Krchna´k, V.; Flegelova´, Z.; Weichsel, A. S.; Lebl, M. Tetrahedron
Lett. 1995, 36, 6193. Rano, T. A.; Chapman, K. T. Tetrahedron Lett.
1995, 36, 3789.
(18) Wijkmans, J . C. H. M.; Culshaw, A. J .; Baxter, A. D. Mol.
Diversity 1998, 3, 117.
(19) Plenkiewicz, H.; Dmowski, W. J . Fluorine Chem. 1998, 89, 213.
(20) Morales, G. A.; Corbett, J . W.; DeGrado, W. F. J . Org. Chem.
1998, 63, 1172. Vojkovsky, T.; Weichsel, A.; Patek, M. J . Org. Chem.
1998, 63, 3162. Dankwardt, S. M.; Newman, S. R.; Krstenanski, J . L.
Tetrahedron Lett. 1995, 36, 4923.
(21) Kiselyov, A. S.; Eisenberg, S.; Luo, Y. Tetrahedron Lett. 1999,
40, 2465. Burgess, K.; Lim, D.; Bois-Choussy, M.; Zhu, J . Tetrahedron
Lett. 1997, 38, 3345.
(13) Corey, E. J .; Venkateswarlu, A. J . Am. Chem. Soc. 1972, 94,
6190.
(14) Thomas, E. J .; Williams, A. C. J . Chem. Soc., Chem. Commun.
1987, 992.
(15) Kaldor, S. W.; Siegel, M. G. Comb. Chem. Mol. Diversity Drug
Discovery 1998, 307-335; Gordon, E, M., Kerwin, J . F., J r., Eds.; Wiley-
Liss: New York, 1998.

Solid-Phase Synthesis of 3-Aminobenzisoxazoles
J . Org. Chem., Vol. 65, No. 10, 2000 2927
Ta ble 3. On -Resin Nu cleop h ilic Ar om a tic Su bstitu tion
Sch em e 5. Solu tion -P h a se Mod el of th e Loa d in g
Rea ction of 2,4-Diflu or oben zon itr ile
trile. As illustrated in Scheme 4, this reaction could lead
either to the desired aryloxime intermediate 12 through
displacement of the 2-fluoro group or to an undesired
isomer 13 through displacement of the 4-fluoro group.
Treatment of the unsymmetrical nitrile with the potas-
sium salt of resin 1 gave a 85% loading yield based on
weight. Although the ratio of 12 to 13 was not deter-
mined, the presumed mixture was then hydrolyzed under
the standard cyclization conditions to yield a single
product 14a (as determined by HPLC and NMR) in 74%
isolated yield. The high selectivity observed for the
formation of the desired product could arise from either
of two pathways. On one hand, the resin loading reaction
may be quite selective for the 2-position, giving primarily
12, and then subsequently 14a upon cyclorelease. Alter-
natively, mixtures of the two isomeric aryloxime adducts
could be formed with the hydrolysis of intermediate 12
occurring much more rapidly than 13. In order gain
insight into the relative contributions of these two
pathways, several experiments were performed. Scheme
5 depicts a solution-phase model study for the loading
reaction of 2,4-difluorobenzonitrile. In this experiment,
the potassium salt of acetone oxime was treated with 2,4-
difluorobenzonitrile and showed only a slight preference
for the addition to the 2-position (1.4:1).23 Obviously such
poor selectivity, if produced in the solid-phase loading
reaction, could not account for the >96% purity of
cyclorelease product 14a .
a
Isolated yield after chromatography and based on loading.
b
Crude purity based on HPLC analysis. Major impurity was the
unreacted fluorine product (4-fluoro-3-aminobenzisoxazole).
Sch em e 4. Regioselective Syn th esis of
6-F lu or o-3-a m in oben zisoxa zole
Scheme 6 illustrates a comparison of the relative rates
of hydrolysis for analogous 2- vs 4-substituted aryloximes.
Since the isolation of pure 12 and pure 13 was not
practical, model resins 3a and 15 were prepared from
2-fluorobenzonitrile and 4-fluorobenzonitrile, respec-
tively. As described previously, when resin 3a was
treated with the standard cyclitive removal conditions,
complete conversion to 3-aminobenzisoxazole was ob-
served after 2 h.
IR analysis of the recovered resin after hydrolysis
showed complete disappearance of the nitrile peak, which
had clearly been present in the starting material. Alter-
natively, upon exposure of the isomeric resin 15 to the
same hydrolysis conditions, no organic material was
released from the solid phase. In this experiment, IR
analysis of the recovered resin confirmed that the nitrile
remained intact, apparently unaffected by exposure to
the aqueous acid.
These observations led us to reexamine the reaction
shown in Scheme 4. This time, resin 12 and 13 were
recovered after the cyclorelease reaction was complete
and characterized by IR spectroscopy. The analysis
by the cyclorelease reaction, compound 11b was obtained
in 94% crude purity and subsequently isolated in 80%
yield.
Similar to the pyrrolidine SNAr reaction discussed
above, the nucleophilic addition of phenol to resin 10
(Table 3, entry 3) proceeded at a faster rate in DMSO
and DMF compared with THF. However, even in these
solvents, the addition reaction required 36 h to bring
about high conversion of the phenol adduct. Aryl ether
11c was then obtained in 48% yield (two steps) and 93%
crude purity after cyclitive removal.
To explore the question of regioselectivity, the loading
reaction was next attempted with 2,4-difluorobenzoni-
(22) Studies on the fluorine kinetic isotope effect in SNAr reactions
with aryl fluorides have shown that the use of acetonitrile as the
reaction solvent leads to a different rate-limiting step in the overall
mechanism when compared with THF. Persson, J .; Axelsson, S.;
Matsson, O. J . Am. Chem. Soc. 1996, 118, 20.
(23) For related regioselectivity studies on solution-phase SNAr
reactions, see: (a) Wells, K. M.; Shi, Y. J .; Lynch, J . E.; Humphrey, G.
R.; Volante, R. P.; Reider, P. J . Tetrahedron Lett. 1996, 37, 6439. (b)
Sasajima, K.; Ona, K.; Katsube, J .; Yamamoto, H. Chem. Pharm. Bull.
1978, 26, 2502.

2928 J . Org. Chem., Vol. 65, No. 10, 2000
Lepore and Wiley
Ta ble 4. Regioselective Syn th esis of
Sch em e 6. Com p a r ison of th e Rela tive Ra tes of
Oxim e Hyd r olysis for 2- ver su s 4-Su bstitu ted Ar yl
Oxim es
F lu or o-3-a m in oben zisoxa zoles
revealed that a nitrile peak was still present on the resin,
although the intensity of the nitrile was diminished
relative to that observed prior to hydrolysis. These data
support the hypothesis that the difference in the hydroly-
sis rates of intermediate 12 vs 13 serves as the primary
source of selectivity in this reaction.
Having successfully achieved the selective functional-
ization of 2,4-difluorobenzonitrile, a variety of additional
unsymmetrical polyfluorobenzonitriles were carried
through the same sequence. The results, presented in
Table 4, demonstrate that the isolated yield of the desired
product is significantly diminished in many cases, likely
as a result of poor selectivity for the 2-position in the
loading reaction. However, even with the significant
decrease in isolated yield, HPLC analyses of the crude
reaction mixtures illustrate that the desired product is
selectively released from the solid phase.
Ca r bon -Ca r bon Bon d -F or m in g Rea ction s. Com-
patibility with versatile methods for the formation of
carbon-carbon bonds on-resin is an important measure
of the suitability of any new linker.24 Toward that goal,
our initial efforts were focused on the application of
catalytic palladium coupling chemistry such as the Su-
zuki and Sonogashira reactions. Thus, the aryl bromide
resin 3b was prepared4 and reacted with phenylboronic
acid under a variety of palladium-catalyzed coupling
conditions and then hydrolyzed to give the biaryl product
16a (Table 5). Initially, two sets of conditions that have
previously been reported for on-resin Suzuki reactions
were evaluated. Use of Pd(PPh3)4 with triethylamine/
DMF for 12 h25 (not shown) produced 16a , but led to the
formation of numerous byproducts. Alternatively, the use
of Pd(PPh3)4, with Na2CO3 in 1:1 DME/H2O at reflux for
12 h,26 produced 16a in 25% isolated yield, contaminated
only by the cyclization product of the unreacted aryl
bromide. Due to the superior purity observed with
aqueous carbonate, optimization efforts were focused on
these conditions. Numerous reaction parameters were
carefully varied including cosolvent (DMF, DMSO, and
THF) and the stoichiometry of the catalyst, boronic acid,
a
Isolated yield after chromatography and based on loading.
Crude purity based on HPLC analysis.
and sodium carbonate. Although the crude purities were
similar for the various solvents that were evaluated, in
general, THF27 was found to give the highest rate of
product formation and, therefore, the highest isolated
yields. Optimal results were obtained with 1.3 equiv of
2 N Na2CO3, at 55 °C, over 36 h. The reaction was found
to be rather sensitive to deviation from these conditions,
particularly with respect to the amount of added 2 N Na2-
CO3. However, in control experiments, resin 3b was
treated with varying amounts of 2 N Na2CO3 in THF at
55 °C over 36 h, and no removal of organic material from
the resin was observed. As Table 5 indicates, the ap-
plication of the optimized Suzuki conditions to several
other substrates (Table 5, entries 2-4) provided similar
yields and purities.
Application of the Sonogashira reaction to resin 3b
proved less difficult than the Suzuki coupling. We were
able to couple phenethylacetylene to 3b to give alkyne
17 in a 58% two-step yield using Pd(PPh3)4/CuI in THF
(Scheme 7).28 Again, as in the Suzuki chemistry, signifi-
cantly better results were obtained with THF as the
(24) Hanessian, S.; Xie, F. Tetrahedron Lett. 1998, 39, 737.
(25) Han, Y.; Walker S. D.; Young, R. N. Tetrahedron Lett. 1996,
37, 2703. Ruhland, B.; Bombrun, A.; Gallop, M. A. J . Org. Chem. 1997,
62, 7820.
(26) Frenette, R.; Friesen, R. W. Tetrahedron Lett. 1994, 35, 9177.
(27) Backes, B. J .; Ellman, J . A. J . Am. Chem. Soc. 1994, 116, 11171.

Solid-Phase Synthesis of 3-Aminobenzisoxazoles
J . Org. Chem., Vol. 65, No. 10, 2000 2929
Ta ble 5. On -Resin Su zu k i Rea ction
Sch em e 8. On -Resin Hor n er -Em m on s
Olefin a tion
Sch em e 9. P ossible P a th w a ys for Heter ocycle
F or m a tion
a
Methods A-D all use 4.0 equiv of boronic acid and 5%
noacetate in THF (preformed with n-BuLi at 0 °C) gave
olefin 19 in 52% isolated yield (three steps, based on
loading of resin 1) and 92% purity.30 The major impurity
in this reaction is the cinnamic acid derivative that likely
results from competing ester hydrolysis in the cyclitive
removal step.
Pd(PPh3)4, 36 h. A ) 2.5 equiv of 2 M Na2CO3, DME, reflux. B )
1.0 equiv of 2 M Na2CO3, THF, 55 °C. C ) 1.3 equiv of 2 M Na2CO3,
b
THF, 55 °C. D ) 1.5 equiv of 2 M Na2CO3, THF, 55 °C. Isolated
yield after chromatography and based on loading. Crude purity
based on HPLC analysis. Major impurity in all cases is 6-bromo-
3-aminobenzisoxazole.
Sch em e 7. On -Resin Son oga sh ir a Cou p lin g
Mech a n istic Im p lica tion s. In the process of carrying
out the studies described above, substituent effects were
observed that have interesting implications regarding the
mechanism of the cyclitive removal reaction. Although
there may be numerous mechanistic nuances involved,
two general pathways for cyclorelease are illustrated in
Scheme 9. In path A, the first step involves direct proton-
ation of the oxime nitrogen,31 leading to the simple hydro-
lysis of the oxime from the resin to give a 2-aminooxy-
benzonitrile intermediate. This intermediate then cycli-
zes in a subsequent step to give the 3-aminobenzisoxazole
product. Alternatively, in path B, protonation occurs ini-
tially on the nitrile. The oxime nitrogen then cyclizes onto
the nitrilium ion, thus activating the imine toward hy-
drolysis. Both the effects of substituents para to the ni-
trile (illustrated in Table 2) and relative rates of hydroly-
sis observed for the aryloxime regioisomers (illustrated
in Scheme 6) support the predominance of path B.
Comparing entries 2 and 3 in Table 2 (method B)
reveals a significant difference in the hydrolysis rates for
reaction solvent. The Pd2(dba)3/CuI/Et3N conditions de-
scribed by Moore29 failed to give reasonable yields and
purities.
The acquisition of aldehyde resin 18 (Scheme 8) from
the Suzuki reaction with 3b (Table 5, entry 4) provided
an opportunity to evaluate the compatibility of the
aryloxime linker with the Horner-Emmons olefination.
Thus, treatment of 18 with the anion of trimethylphospho-
(30) Chin, J .; Fell, B.; Shapiro, M. J .; Tomesch, J .; Wareing, J . R.;
Bray, A. M. J . Org. Chem. 1997, 62, 538. Rotella, D. P. J . Am. Chem.
Soc. 1996, 118, 12246.
(28) Kahn, S. I.; Grinstaff, M. W. Tetrahedron lett. 1998, 39, 8031.
(29) Nelson, J . C.; Young, J . K.; Moore, J . S. J . Org. Chem. 1996,
61, 8160.
(31) Egberink, H.; VanHeerden, C. Anal. Chim. Acta 1980, 118, 359.

2930 J . Org. Chem., Vol. 65, No. 10, 2000
Lepore and Wiley
(7 mL) and potassium tert-butoxide (640 µL, 1 M in THF, 0.642
mmol). After being shaken by hand for several minutes, the
resin turned a deep purple color. To this suspension was added
2-fluorobenzonitrile (1.07 mmol, 116 µL). The reaction vessel
was rotated at 55 °C in a Robbins oven for 12 h to give resin
3a followed by cooling for 1 h. The resin was then rinsed with
CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), and
MeOH (4 × 5 mL). The resin was dried in a 35 °C vacuum
oven to a constant weight to give a ∆wt ) 42.1 mg (78% loading
yield). An IR spectrum of this resin shows a nitrile absorption
the bromo-substituted aryloxime relative to the methoxy-
substituted analogue. Since these functional groups are
not formally conjugated to the oxime, it seems unlikely
that inductive effects would cause such a profound dif-
ference on the rate of hydrolysis six bonds away. On the
other hand, these substituents are expected to have a
significant effect on the basicity of the nitrile in the para
position.32 By increasing the basicity of the nitrile, the
p-methoxy group33 should facilitate nitrilium ion forma-
tion and, therefore, facilitate hydrolysis through path B.
The relative reactivity of the aryloxime regioisomers
shown in Scheme 6 also illustrates the importance of
nitrile assistance in the hydrolysis reaction (path B).
Since the effect of the nitrile on the basicity of the oxime
is expected to be similar in the ortho versus para position,
clearly the dramatic difference in the hydrolysis rate of
3a and 15 cannot be rationalized on the basis of path A.
On the other hand, the results are completely consistent
with the operation of path B.
Con clu sion . The studies described above illustrate the
broad utility of this approach to the synthesis of 3-ami-
nobenzisoxazoles. While aqueous acidic conditions were
found that efficiently mediate the cyclitive removal of a
variety of substrates, stability studies demonstrated that
the aryloxime linker is also stable to a range of acidic
conditions widely used for protecting group manipulation.
During the course of these studies, we observed substitu-
ent effects that suggest that the nitrile plays a critical
role in facilitating the oxime hydrolysis reaction. Further,
we have demonstrated that a variety of functionalized
substrates can be loaded onto the Kaiser-DeGrado resin.
These resin-bound intermediates can successfully under-
go carbon-nitrogen, carbon-oxygen, and carbon-carbon
bond-forming reactions without compromising the integ-
rity of the aryloxime linker. These results clearly estab-
lish this method as a viable tool for library generation.
at 2227 cm-1
.
Syn th esis of 6-[(p-Ch lor op h en yl)m eth yla m in oca r bo-
n yl]-3-a m in o-1,2-ben zisoxa zole (7). To p-nitrobenzophenone
oxime polystyrene (Kaiser) resin (500 mg, 1.07 mmol/g, 0.54
mmol) in a tared 25 mL Kontes microfilter funnel were added
THF (7 mL) and potassium tert-butoxide (640 µL, 1 M in THF,
0.642 mmol). After shaking by hand for several minutes, the
resin turned a deep purple color. To this suspension was added
2-fluoro-4-(methoxycarbonyl)benzonitrile (1.07 mmol, 192 mg).
The reaction vessel was rotated at 55 °C in a Robbins oven
for 12 h to give resin 5 followed by cooling for 1 h. The resin
was then rinsed with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL),
H2O (2 × 5 mL), and MeOH (4 × 5 mL). The resin was dried
in a 35 °C vacuum oven for 12 h to give a ∆wt ) 50.0 mg (58%
loading yield). To effect ester hydrolysis, the resin was then
suspended in THF (7 mL) followed by the addition of LiOH
(39 mg, 1.61 mmol) dissolved in MeOH/H2O (1:1, 2 mL) and
rotated at rt for 12 h. The resin was then rinsed with CH2Cl2
(2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), MeOH (2 ×
5 mL), and DMF (2 × 5 mL). The resin was then suspended
in DMF (7 mL), and to this were added p-chlorobenzylamine
(261 µL, 2.14 mmol), 1-hydroxybenzotriazole hydrate (HOBt)
(289 mg, 2.14 mmol), benzotriazole-1-yloxytris(dimethylami-
no)phosphonium hexafluorophosphate (BOP) (947 mg, 2.14
mmol), and diisopropylethylamine (DIPEA) (467 µL, 2.68
mmol). The reaction was allowed to proceed for 12 h at rt. The
resin was then rinsed with CH2Cl2 (2 × 5 mL), MeOH (2 × 5
mL), H2O (2 × 5 mL), and MeOH (4 × 5 mL). The resin was
dried in a 35 °C vacuum oven for 12 h to give a ∆wt ) 84 mg.
TFA (4 mL) and 5 N HClaq (1 mL) were then added to the resin,
and the vessel was rotated for 2 h in a 55 °C Robbins oven.
The TFA/HClaq was collected, and the resin was rinsed with
CH2Cl2 (2 × 5 mL). These washings were combined and
concentrated in vacuo to give the crude product 7 (93% purity
by HPLC), which was purified by radial chromatography on a
2 mm plate eluting with 50% EtOAc/hexanes. Concentration
of the product containing fractions gave pure 6-[(p-chlorophe-
nyl)methylaminocarbonyl]-3-amino-1,2-benzisoxazole (7) (76
mg, three-step yield ) 81% based on the ester loading yield):
HPLC retention time (eluent gradient 10% A to 60% A over
Exp er im en ta l Section
Gen er a l Meth od s. Reagents obtained from commercial
sources were used without further purification. Kaiser oxime
resin 1 was purchased from Novabiochem with a loading
capacity of 1.07 mmol/g. All NMR spectra were recorded on a
400 MHz instrument. Mass spectra were obtained with either
ESI or FAB as the ionization method. Infrared spectra of the
functionalized resins were recorded on an FTIR spectropho-
tometer using the KBr pellet method. All purifications were
carried out by radial chromatography (Chromatotron model
8924, Harrison Research) using 1 mm silica gel plates (Anal-
tech). Crude purities were estimated from integrated peak
areas of HPLC chromatographs with the UV detector monitor-
ing at λ ) 215 nm. Analytical HPLC setup: C18 Vydac column
with solvent gradient A ) acetonitrile (0.1% TFA) and B )
water (0.1% TFA) at 1 mL/min flow rate. Unless otherwise
noted, all HPLC retention times are given for an eluent
gradient of 10% A to 60% A over 40 min. The nomenclature of
3-aminobenzisoxazole compounds is based on the heterocycle
numbering system3 shown below:
1
45 min) ) 24.9 min; H NMR (400 MHz, DMSO-d6) δ 9.19 (t,
J ) 5.6 Hz, 1H), 7.84-7.91 (m, 2H), 7.72 (dd, J ) 8.0, 1.2 Hz,
1H), 7.37-7.31 (m, 4H), 6.50 (bs, 2H), 4.45 (d, J ) 6.0 Hz,
2H); MS (ESI) m/z 300 [35Cl, (M + H)+], 302 [37Cl, (M + H)+];
HRMS calcd for C15H13ClN3O2 302.0689, found 302.0696.
Syn th esis of 6-[(p-Ch lor op h en yl)oxym eth yl]-3-a m in o-
1,2-ben zisoxa zole (9). To p-nitrobenzophenone oxime poly-
styrene (Kaiser) resin (500 mg, 1.07 mmol/g, 0.54 mmol) in a
tared 25 mL Kontes microfilter funnel were added THF (7 mL)
and potassium tert-butoxide (640 µL, 1 M in THF, 0.642 mmol).
After being shaken by hand for several minutes, the resin
turned a deep purple color. To this suspension was added
2-fluoro-4-(tert-butyldimethylsilyloxymethly)benzonitrile (1.07
mmol, 285 mg). The reaction vessel was rotated at 55 °C in a
Robbins oven for 12 h to give resin 8 followed by cooling for 1
h. The resin was then rinsed with CH2Cl2 (2 × 5 mL), MeOH
(2 × 5 mL), H2O (2 × 5 mL), and MeOH (4 × 5 mL). The resin
was dried in a 35 °C vacuum oven for 12 h to give a ∆wt )
71.1 mg (54% loading yield). To effect TBS removal, the resin
was then suspended in THF (6 mL) followed by the addition
of TBAF (562 µL, 1 M in THF, 0.562 mmol) and rotated at rt
Syn th esis of O-2-Ben zon itr ile-p-n itr oben zop h en on e
Oxim e P olystyr en e (3a ). To p-nitrobenzophenone oxime
polystyrene (Kaiser) resin (500 mg, 1.07 mmol/g, 0.54 mmol)
in a tared 25 mL Kontes microfilter funnel were added THF
(33) Schaefer, F. C. In The Chemistry of the Cyano Group; Rap-
poport, Z., Ed.; Interscience Publishers (J ohn Wiley & Sons): London,
1970; p 244.
(32) Colominas, C.; Orozco, M.; Luque, F. J .; Borrell, J . I.; Teixido,
J . J . Org. Chem. 1998, 63, 4947.

Solid-Phase Synthesis of 3-Aminobenzisoxazoles
J . Org. Chem., Vol. 65, No. 10, 2000 2931
for 12 h. The resin was then rinsed with CH2Cl2 (2 × 5 mL),
MeOH (2 × 5 mL), H2O (2 × 5 mL), MeOH (2 × 5 mL), and
CH2Cl2 (2 × 5 mL). The resin was then suspended in CH2Cl2
(7 mL), and to this were added p-chlorophenol (690 mg, 5.35
mmol), triphenylphosphine (700 mg, 2.68 mmol), and diiso-
propylazodicarboxylate (DIAD) (530 µL, 2.68 mmol). The
reaction was allowed to proceed for 1 h at rt. The resin was
then rinsed with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O
(2 × 5 mL), and MeOH (4 × 5 mL). The resin was dried in a
35 °C vacuum oven for 12 h. TFA (4 mL) and 5 N HClaq (1
mL) were then added to the resin, and the vessel was rotated
for 2 h in a 55 °C Robbins oven. The TFA/HClaq was collected,
and the resin was rinsed with CH2Cl2 (2 × 5 mL). These
washings were combined and concentrated in vacuo to give
the crude product 7 (83% purity by HPLC), which was purified
by radial chromatography on a 2 mm plate eluting with 50%
EtOAc/hexanes. Concentration of the product containing frac-
tions gave pure 6-[(p-chlorophenyl)oxymethyl]-3-amino-1,2-
benzisoxazole (9) (61 mg, three-step yield ) 77% based on
loading yield): HPLC retention time (eluent gradient 10% A
to 60% A over 45 min) ) 37.5 min; 1H NMR (400 MHz, DMSO-
d6) δ 7.73 (d, J ) 8.0 Hz, 1H), 7.84-7.91 (m, 2H), 7.72 (dd, J
) 8.0, 1.2 Hz, 1H), 7.44 (s, 1H), 7.30 (d, J ) 8.0 Hz, 1H), 7.25-
7.21 (m, 2H), 6.99-6.95 (m, 2H), 5.17 (bs, 2H), 1.22 (d, J )
6.4 Hz, 2H); MS (ESI) m/z 274 [35Cl, (M + H)+], 276 [37Cl, (M
+ H)+]. Anal. Calcd for C14H11ClN2O2: C, 61.21; H, 4.04; N,
10.2; Cl, 12.91. Found: C, 61.08; H, 4.32; N, 10.19; Cl, 12.47.
Syn th esis of Resin 10. To p-nitrobenzophenone oxime
polystyrene (Kaiser) resin (500 mg, 1.07 mmol/g, 0.54 mmol)
in a tared 25 mL Kontes microfilter funnel were added THF
(7 mL) and potassium tert-butoxide (640 µL, 1 M in THF, 0.642
mmol). After being shaken by hand for several minutes, the
resin turned a deep purple color. To this suspension was added
2,6-difluorobenzonitrile (150 mg, 1.07 mmol). The reaction
vessel was rotated at 55 °C in a Robbins oven for 8 h to give
resin 10 followed by cooling for 1 h. The resin was then rinsed
with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL),
and MeOH (4 × 5 mL). The resin was dried in a 35 °C vacuum
oven for 12 h to give a ∆wt ) 40.1 mg (63% loading yield).
Syn th esis of 4-[2-(p-Ch lor op h en yl)eth oxy]-3-a m in o-
1,2-ben zisoxa zole (11a ). In a separate vial, 2-(p-chlorophe-
nyl)ethanol (128 µL, 1.07 mmol) was dissolved in 1 mL of THF
followed by the addition of potassium tert-butoxide (1.07 mL,
1 M in THF, 1.07 mmol). This alkoxide solution was then
added to a THF suspension (6 mL) of resin 10 (540.1 mg,
assume 0.535 mmol). The reaction vessel was rotated for 12 h
in a 55 °C Robbins oven. The resin was then rinsed with CH2-
Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), and MeOH
(4 × 5 mL). The resin was dried in a 35 °C vacuum oven for 3
h. TFA (4 mL) and 5 N HClaq (1 mL) were then added to the
resin, and the vessel was rotated for 2 h in a 55 °C Robbins
oven. The TFA/HClaq was collected, and the resin was rinsed
with CH2Cl2 (2 × 5 mL). These washings were combined and
concentrated in vacuo to give the crude product 11a (95%
purity by HPLC), which was purified by radial chromatogra-
phy on a 2 mm plate eluting with 50% EtOAc/hexanes.
Concentration of the product containing fractions gave pure
6-[(p-chlorophenyl)oxymethyl]-3-amino-1,2-benzisoxazole (11a )
(66 mg, two-step yield ) 50% based on loading yield): HPLC
retention time (eluent gradient 10% A to 80% A over 45 min)
The TFA/HClaq was collected, and the resin was rinsed with
CH2Cl2 (2 × 5 mL). These washings were combined and
concentrated in vacuo to give the crude product 11b (94%
purity by HPLC), which was purified by radial chromatogra-
phy on a 2 mm plate eluting with 50% EtOAc/hexanes.
Concentration of the product-containing fractions gave pure
4-pyrrolidino-3-amino-1,2-benzisoxazole (11b) (54 mg, two-step
yield ) 80% based on loading yield): HPLC retention time
1
(eluent gradient 5% A to 40% A over 45 min) ) 27.9 min; H
NMR (400 MHz, DMSO-d6) δ 7.27 (t, J ) 8.0 Hz, 1H), 6.84 (d,
J ) 8.0 Hz, 1H), 6.55 (d, J ) 7.6, 1H), 5.62 (bs, 2H), 3.27-
3.20 (m, 4H), 1.92-1.83 (m, 4H); MS (ESI) m/z 204 (M + H)+.
Anal. Calcd for C11H13N3O: C, 65.01; H, 6.45; N, 20.67.
Found: C, 64.46 H, 6.22; N, 20.11.
Syn t h esis of 4-P h en oxy-3-a m in o-1,2-b en zisoxa zole
(11c). To a DMF suspension (7 mL) of resin 10 (540.1 mg,
assume 0.535 mmol) were added phenol (503 mg, 5.35 mmol)
and K2CO3 (738 mg, 5.35 mmol). The reaction vessel was
rotated for 36 h in a 55 °C Robbins oven. The resin was then
rinsed with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 ×
5 mL), and MeOH (4 × 5 mL). The resin was dried in a 35 °C
vacuum oven for 3 h. TFA (4 mL) and 5 N HClaq (1 mL) were
then added to the resin, and the vessel was rotated for 2 h in
a 55 °C Robbins oven. The TFA/HClaq was collected, and the
resin was rinsed with CH2Cl2 (2 × 5 mL). These washings were
combined and concentrated in vacuo to give the crude product
11c (93% purity by HPLC), which was purified by radial
chromatography on a 2 mm plate eluting with 50% EtOAc/
hexanes. Concentration of the product-containing fractions
gave pure 4-phenoxy-3-amino-1,2-benzisoxazole (11c) (38 mg,
two-step yield ) 48% based on loading yield): HPLC retention
time (eluent gradient 10% A to 90% A over 45 min) ) 23.7
min; 1H NMR (400 MHz, DMSO-d6) δ 7.44 (t, J ) 7.6 Hz, 2H),
7.36 (t, J ) 9.6 Hz, 1H), 7.26-7.15 (m, 3H), 7.10 (d, J ) 8.4
Hz, 1H), 6.36 (d, J ) 7.6 Hz, 1H), 6.02 (bs, 2H); MS (ESI) m/z
227 (M + H)+. Anal. Calcd for C13H10N2O2: C, 69.02; H, 4.46;
N, 12.38. Found: C, 69.21 H, 4.06; N, 12.75.
Gen er a l P r oced u r e for th e Syn th esis of F lu or o-3-
a m in o-1,2-ben zisoxa zoles 14a -f. To p-nitrobenzophenone
oxime polystyrene (Kaiser) resin (500 mg, 1.07 mmol/g, 0.54
mmol) in a tared 25 mL Kontes microfilter funnel were added
THF (7 mL) and potassium tert-butoxide (640 µL, 1 M in THF,
0.642 mmol). After being shaken by hand for several minutes,
the resin turned a deep purple color. To this suspension was
added 2,4-difluorobenzonitrile (150 mg, 1.07 mmol). The
reaction vessel was rotated at 55 °C in a Robbins oven for 12
h followed by cooling for 1 h. The resin was then rinsed with
CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), and
MeOH (4 × 5 mL). The resin was dried in a 35 °C vacuum
oven for 12 h to give a ∆wt ) 52.4 mg (90% loading yield). An
IR analysis of the resin shows a nitrile stretching peak at
2230.3 cm-1. TFA (4 mL) and 5 N HClaq (1 mL) were then
added to the resin, and the vessel was rotated for 2 h in a 55
°C Robbins oven. The TFA/HClaq was collected, and the resin
was rinsed with CH2Cl2 (2 × 5 mL). An IR analysis of the resin
after the cyclorelease reaction shows a diminished nitrile
stretching peak. The CH2Cl2 washings were combined and
concentrated in vacuo to give the crude product 14a (>96%
purity by HPLC), which was purified by radial chromatogra-
phy on a 2 mm plate eluting with 50% EtOAc/hexanes.
Concentration of the product containing fractions gave pure
6-flu or o-3-a m in oben zisoxa zole (14a ) (40 mg, two-step yield
) 74% based on loading yield): HPLC retention time (eluent
1
) 31.8 min; H NMR (400 MHz, DMSO-d6) δ 7.39-7.31 (m,
5H), 6.93 (d, J ) 8.4 Hz, 2H), 6.69 (d, J ) 8.4, 1H), 5.72 (bs,
2H), 4.30 (t, J ) 6.4 Hz, 2H), 3.12 (t, J ) 6.4 Hz, 2H); MS
(ESI) m/z 289 [35Cl, (M + H)+], 291 [37Cl, (M + H)+]. Anal.
Calcd for C15H13ClN2O2: C, 62.41; H, 4.54; N, 9.70; Cl, 12.28.
Found: C, 62.08; H, 4.61; N, 9.53; Cl, 12.41.
1
gradient 10% A to 60% A over 45 min) ) 13.3 min; H NMR
(400 MHz, DMSO-d6) δ 7.81 (dd, J ) 8.8, 5.2 Hz, 1H), 7.36
(dd, J ) 9.6, 2.0 Hz, 1H), 7.11 (ddd, J ) 11.2, 8.8, 2.4 1H),
6.44 (bs, 2H); MS (FD) m/z 152 M+. Anal. Calcd for C7H5-
FN2O: C, 55.27; H, 3.31; N, 18.41. Found: C, 55.21 H, 3.27;
N, 18.20.
Syn th esis of 4-P yr r olid in o-3-a m in o-1,2-ben zisoxa zole
(11b). To a DMSO suspension (7 mL) of resin 10 (540.1 mg,
assume 0.535 mmol) was added pyrrolidine (134 µL, 1.61
mmol). The reaction vessel was rotated for 8 h in a 55 °C
Robbins oven. The resin was then rinsed with CH2Cl2 (2 × 5
mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), and MeOH (4 × 5
mL). The resin was dried in a 35 °C vacuum oven for 3 h. TFA
(4 mL) and 5 N HClaq (1 mL) were then added to the resin,
and the vessel was rotated for 2 h in a 55 °C Robbins oven.
7-F lu or o-3-a m in oben zisoxa zole (14b) (30 mg, two-step
yield ) 56% based on loading yield): >96% crude purity;
HPLC retention time (eluent gradient 5% A to 40% A over 45
min) ) 20.1 min; 1H NMR (400 MHz, DMSO-d6) δ 7.63 (dd, J
) 8.4, 1.2 Hz, 1H), 7.40 (ddd, J ) 11.6, 8.4, 0.8 Hz, 1H), 7.21
(dd, J ) 8.0, 4.4 1H), 6.58 (bs, 2H); MS (FD) m/z 152 M+. Anal.

2932 J . Org. Chem., Vol. 65, No. 10, 2000
Lepore and Wiley
Calcd for C7H5FN2O: C, 55.27; H, 3.31; N, 18.41. Found: C,
55.19 H, 3.45; N, 18.36.
Son oga sh ir a Cou p lin g w ith Resin 3b. Resin 3b was
prepared in 66% loading yield from 0.535 mmol of Kaiser oxime
1 as detailed above. The resin was then suspended in THF (7
mL), and to this were added 4-phenylbutyne (280 mg, 2.14
mmol), Et3N (500 µL), CuI (5.1 mg, 0.027 mmol), and Pd(PPh3)4
(31 mg, 0.027 mmol). The vessel was then rotated for 36 h in
a 55 °C Robbins oven followed by rinsing with CH2Cl2 (2 × 5
mL), MeOH (2 × 5 mL), H2O (2 × 5 mL), and MeOH (4 × 5
mL). The resin was dried in a 35 °C vacuum oven for 3 h. TFA
(4 mL) and 5 N HClaq (1 mL) were then added to the resin,
and the vessel was rotated for 2 h in a 55 °C Robbins oven.
The TFA/HClaq was collected, and the resin was rinsed with
CH2Cl2 (2 × 5 mL). These washings were combined and
concentrated in vacuo to give the crude product 17 (93% purity
by HPLC), which was purified by radial chromatography on a
2 mm plate eluting with 50% EtOAc/hexanes. Concentration
of the product containing fractions gave pure 6-(4-p h en ylbu -
tyn yl)-3-a m in o-1,2-ben zisoxa zole (17) (58 mg, two-step
yield ) 58% based on loading yield): HPLC retention time
(eluent gradient 20% A to 85% A over 45 min) ) 26.2 min; 1H
NMR (400 MHz, DMSO-d6) δ 7.73 (d, J ) 8.0 Hz, 1H), 7.38 (s,
1H), 7.29-7.27 (m, 4H), 7.21-7.15 (m, 2H), 6.43 (bs, 2H), 2.84
(t, J ) 7.6 Hz, 2H), 2.70 (t, J ) 7.2 Hz, 2H); MS (ESI) m/z 263
(M + H)+; HRMS calcd for C13H10N2O 263.1184, found 263.1172.
6,7-Diflu or o-3-a m in oben zisoxa zole (14e) (23 mg, two-
step yield ) 27% based on loading yield): 78% crude purity;
HPLC retention time (eluent gradient 5% A to 40% A over 45
1
min) ) 27.6 min; H NMR (400 MHz, DMSO-d6) δ 7.65-7.61
(m, 1H), 7.36-7.30 (m, 1H), 6.62 (bs, 2H); MS (FD) m/z 171
M+; HRMS calcd for C7H4F2N2O 171.0370, found 171.0368.
4,6-Diflu or o-3-a m in oben zisoxa zole (14f) (22 mg, 2 step
yield ) 31% based on loading yield): >96% crude purity;
HPLC retention time (eluent gradient 5% A to 40% A over 45
min) ) 21.1 min; 1H NMR (400 MHz, DMSO-d6) δ 7.31 (dd, J
) 8.8, 1.2 Hz, 1H), 7.10 (dd, J ) 10.0, 2.0 Hz, 1H), 6.34 (bs,
2H); MS (FD) m/z 171 M+. Anal. Calcd for C7H4F2N2O: C,
49.42; H, 2.37; N, 16.47. Found: C, 49.36 H, 2.38; N, 16.39.
Gen er a l P r oced u r e for th e Su zu k i Cou p lin g Rea ction s
w ith Resin 3b. To p-nitrobenzophenone oxime polystyrene
(Kaiser) resin (500 mg, 1.07 mmol/g, 0.54 mmol) in a tared 25
mL Kontes microfilter funnel were added THF (7 mL) and
potassium tert-butoxide (640 µL, 1 M in THF, 0.642 mmol).
After being shaken by hand for several minutes, the resin
turned a deep purple color. To this suspension was added
4-bromo-2-fluorobenzonitrile (214 mg, 1.07 mmol). The reaction
vessel was rotated at 55 °C in a Robbins oven for 12 h to give
resin 3b followed by cooling for 1 h. The resin was then rinsed
with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 × 5 mL),
and MeOH (4 × 5 mL). The resin was dried in a 35 °C vacuum
oven for 12 h to give a ∆wt ) 64 mg (66% loading yield). The
resin was then suspended in THF (7 mL), and to this were
added phenylboronic acid (261 mg, 2.14 mmol), Na2CO3 (348
µL, 2 M in H2O, 2.14 mmol), and Pd(PPh3)4. The vessel was
then rotated for 36 h in a 55 °C Robbins oven followed by
rinsing with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 ×
5 mL), and MeOH (4 × 5 mL). The resin was dried in a 35 °C
vacuum oven for 3 h. TFA (4 mL) and 5 N HClaq (1 mL) were
then added to the resin, and the vessel was rotated for 2 h in
a 55 °C Robbins oven. The TFA/HClaq was collected, and the
resin was rinsed with CH2Cl2 (2 × 5 mL). These washings were
combined and concentrated in vacuo to give the crude product
16a (91% purity by HPLC), which was purified by radial
chromatography on a 2 mm plate eluting with 50% EtOAc/
hexanes. Concentration of the product-containing fractions
gave pure 6-p h en yl-3-a m in o-1,2-ben zisoxa zole (16a ) (46
mg, two-step yield ) 58% based on loading yield): HPLC
retention time (eluent gradient 10% A to 60% A over 45 min)
Hor n er -Em m on s Olefin a tion To Give 19. To p-ni-
trobenzophenone oxime polystyrene (Kaiser) resin (500 mg,
1.07 mmol/g, 0.54 mmol) in a tared 25 mL Kontes microfilter
funnel were added THF (7 mL) and potassium tert-butoxide
(640 µL, 1 M in THF, 0.642 mmol). After being shaken by hand
for several minutes, the resin turned a deep purple color. To
this suspension was added 4-bromo-2-fluorobenzonitrile (214
mg, 1.07 mmol). The reaction vessel was rotated at 55 °C in a
Robbins oven for 12 h to give resin 3b followed by cooling for
1 h. The resin was then rinsed with CH2Cl2 (2 × 5 mL), MeOH
(2 × 5 mL), H2O (2 × 5 mL), and MeOH (4 × 5 mL). The resin
was dried in a 35 °C vacuum oven for 12 h to give a ∆wt ) 64
mg (66% loading yield). The resin was then suspended in THF
(7 mL), and to this were added 4-formylphenylboronic acid (482
mg, 3.21 mmol), Na2CO3 (348 µL, 2 M in H2O, 0.696 mmol),
and Pd(PPh3)4 (31 mg, 0.027 mmol). The vessel was then
rotated for 36 h in a 55 °C Robbins oven followed by rinsing
with CH2Cl2 (2 × 5 mL), MeCN (2 × 5 mL), H2O (2 × 5 mL),
and MeCN (4 × 5 mL) (care was taken to not rinse the resin
with methanol in order to avoid methyl acetal formation). The
resin was dried in a 35 °C vacuum oven for 3 h to give aldehyde
resin 18. In a separate vial, trimethylphosponoacetate (172
µL, 1.07 mmol) was dissolved in THF (4 mL) and cooled to 0
°C. To this vial was added BuLi (602 µL, 0.96 mmol, 1.6 M in
hexanes) dropwise. This THF solution was then added drop-
wise by syringe to a suspension of resin 18 in THF (4 mL),
and the vessel was rotated for 6 h at rt. This was followed by
rinsing with CH2Cl2 (2 × 5 mL), MeOH (2 × 5 mL), H2O (2 ×
5 mL), and MeOH (4 × 5 mL). The resin was then dried in a
35 °C vacuum oven for 3 h. TFA (4 mL) and 5 N HClaq (1 mL)
were then added to the resin, and the vessel was rotated for 2
h in a 55 °C Robbins oven. The TFA/HClaq was collected, and
the resin was rinsed with CH2Cl2 (2 × 5 mL). These washings
were combined and concentrated in vacuo to give the crude
product 19 (92% purity by HPLC), which was purified by radial
chromatography on a 2 mm plate eluting with 50% EtOAc/
hexanes. Concentration of the product con-taining fractions
gave pure 6-[4-(m et h oxyca r b on ylet h ylen yl)p h en yl]-3-
a m in o-1,2-ben zisoxa zole (19) (47 mg, three-step yield ) 52%
based on loading yield of resin 1): HPLC retention time (eluent
1
) 13.3 min; H NMR (400 MHz, DMSO-d6) δ 7.86 (d, J ) 8.4
Hz, 1H), 7.74-7.70 (m, 2H), 7.68 (s, 1H), 7.53 (d, J ) 8.0 Hz,
2H), 7.48-7.43 (m, 2H), 7.40-7.35 (m, 1H), 6.42 (bs, 2H); MS
(ESI) m/z 211 (M + H)+. Anal. Calcd for C13H10N2O: C, 74.27;
H, 4.79; N, 13.32. Found: C, 74.59 H, 4.86; N, 13.36.
6-(3,5-Dich lor oph en yl)-3-am in o-1,2-ben zisoxazole (16b)
(two-step yield ) 51% based on loading yield); 95% crude
purity; HPLC retention time (eluent gradient 20% A to 80%
1
A over 45 min) ) 30.4 min; H NMR (400 MHz, DMSO-d6) δ
7.87 (d, J ) 8.8 Hz, 1H), 7.82 (s, 1H), 7.81 (d, J ) 1.6 Hz, 2H),
7.62-7.58 (m, 2H), 6.46 (bs, 2H); MS (FD) m/z 278 [35Cl + 35Cl,
M+], 280 [35Cl + 37Cl, (M + H)+]. Anal. Calcd for C13H8-
Cl2N2O: C, 55.94; H, 2.89; N, 10.04. Found: C, 55.93 H, 2.60;
N, 9.85.
6-(3,5-Ditr iflu or om eth ylp h en yl)-3-a m in o-1,2-ben zisox-
a zole (16c) (two-step yield ) 45% based on loading yield):
81% crude purity; HPLC retention time (eluent gradient 20%
A to 80% A over 45 min) ) 31.0 min; 1H NMR (400 MHz,
DMSO-d6) δ 8.40 (s, 2H), 8.11 (s, 1H), 7.98 (s, 2H), 7.93 (d, J
) 8.4 Hz, 1H), 7.76-7.71 (m, 2H), 6.49 (bs, 2H); MS (ESI) m/z
347 (M + H)+. Anal. Calcd for C15H8F6N2O: C, 52.04; H, 2.33;
N, 8.09. Found: C, 52.00 H, 2.06; N, 7.99.
6-(4-F or m ylp h en yl)-3-a m in o-1,2-b en zisoxa zole (16d )
(two-step yield ) 43% based on loading yield): 92% crude
purity; HPLC retention time (eluent gradient 5% A to 40% A
over 45 min) ) 37.6 min; 1H NMR (400 MHz, DMSO-d6) δ
10.05 (s, 1H), 7.98 (s, 4H), 7.90 (d, J ) 8.4 Hz, 1H), 7.82 (s,
1H), 7.63 (dd, J ) 8.4, 1.6 Hz, 1H), 6.46 (bs, 2H); MS (ESI)
m/z 239 (M + H)+; HRMS calcd for C14H10N2O2 (M + H)+
239.0821, found 239.1816.
1
gradient 20% A to 85% A over 45 min) ) 24.3 min; H NMR
(400 MHz, DMSO-d6) δ 7.86 (d, J ) 8.0 Hz, 1H), 7.80 (s, 4H),
7.76 (s, 1H), 7.69 (d, J ) 16.4 Hz, 1H), 7.59 (d, J ) 8.0 Hz,
1H), 6.69 (d, J ) 16.4 Hz, 1H), 6.43 (bs, 2H), 3.71 (s, 3H);
HRMS calcd for C17H15N2O3 295.1082, found 295.1088.
Ack n ow led gm en t. The authors would like to thank
Dr. Steven W. Kaldor for many helpful discussions.
J O991263I