Synthetic utility of epoxides for chiral functionalization of isoxazoles

Article abstract of DOI:10.1016/j.tetlet.2008.03.059

The lithio-anion of isoxazole 2 was found to ring open propylene oxide in good yields with complete regioselectivity. Vinylic and benzylic epoxides were utilized as key examples of electrophiles and found to produce a mixture of regioisomeric adducts. Add

Full text of DOI:10.1016/j.tetlet.2008.03.059

Available online at www.sciencedirect.com
Tetrahedron Letters 49 (2008) 3078–3082
Synthetic utility of epoxides for chiral functionalization of isoxazoles
Jared K. Nelson a, Christopher T. Burns a, Miles P. Smith a,
Brendan Twamley a, N. R. Natale a,b,
*
a Department of Chemistry, University of Idaho, Moscow, ID 83844-2343, United States
b NIH-COBRE Center for Structural and Functional Neuroscience, University of Montana, Missoula, MT 59812, United States
Received 1 February 2008; revised 10 March 2008; accepted 11 March 2008
Available online 14 March 2008
Abstract
The lithio-anion of isoxazole 2 was found to ring open propylene oxide in good yields with complete regioselectivity. Vinylic and
benzylic epoxides were utilized as key examples of electrophiles and found to produce a mixture of regioisomeric adducts. Additionally,
the use of chiral epoxides was explored, and absolute configuration was determined by X-ray crystallography to prove that nucleophilic
attack at the benzylic carbon of (R)-styrene oxide proceeds with 100% inversion at the benzylic carbon to afford the (S)-alcohol (4b).
Ó 2008 Elsevier Ltd. All rights reserved.
Epoxides have found vast utility in the synthesis of nat-
ural products and novel molecules with potential therapeu-
tic value.1 With a myriad of methods available for facile
synthesis of chiral epoxides, the chemistry of epoxide ring
opening is of ever expanding importance.2 A recent note-
worthy application is the use of intramolecular epoxide ring
opening using a functionalized isoxazole in Myers’s practi-
cal and versatile tetracycline synthesis.3 Reports4 exist
which demonstrate the reaction of resonance-stabilized
anions with epoxides, exemplified by the poor ability of
lithium enolates of ketones and esters to efficiently react
with epoxides,5 and a few reports have illustrated achiral
or racemic epoxide opening with simple isoxazolyl
anions.6–10 Herein we report the usefulness of a branched
and highly functionalized resonance stabilized lithio alkyl
isoxazole carbanion, which—in fact—reacts stereoselec-
tively with chiral epoxides (Scheme 1).
In our studies of isoxazolyl-1,4-dihydropyridines
(DHPs, 1) of (Fig. 1, Scheme 1), we have found that bio-
isoteric replacement of a 4-aryl by substituted isoxazole
results in robust antagonists of the L-type Ca2+ channel.11
The synthetic route utilizes
a
lateral metalation
N
O
N
O
O
N
El
El
CH3
CH3
CH3
1. BuLi
2. El+
1. Deprotection
CH2
CH2
CH3
CO2Et
EtO2C
O
N
O
N
2. Hantzsch
synthesis
CH3
CH3
CH3
CH3
N
H
CH3
CH3
1
2
Scheme 1. Isoxazolyl DHPs function as robust L-type Ca2+ channels ligands.
*
Corresponding author. Tel.: +1 406 243 4132; fax: +1 406 243 5228.
E-mail address: Nicholas.natale@umontana.edu (N. R. Natale).
0040-4039/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2008.03.059

J. K. Nelson et al. / Tetrahedron Letters 49 (2008) 3078–3082
3079
styrene oxide, 86% combined yield was obtained with a
ratio of 70:30 for 3b:4b. The vinylic epoxide butadiene
monoxide gave a ratio for 3c:4c of 41:59, with a 47% com-
bined yield (no SN20 product was observed). This result
clearly illustrates the competing effects of steric hindrance
and resonance stabilization at the secondary carbon, show-
ing that an aromatic moiety at the a-position results in
greater stabilization of a developing positive charge than
does an a-vinyl moiety.
The oxazoline nitrogen is by far the most basic atom in
the isoxazole–oxazoline system, and we suspect that this
heteroatom could be in part responsible for the observed
reactivity of the C-5 methyl isoxazole–oxazoline.21,22 Pres-
ence of an appropriately spaced basic nitrogen has been
shown to result in an increased rate of chemoselective
deprotonation as exemplified by the use of hemi-labile
carbanolamine ligands,23 and functionally similar systems
containing dipole-stabilized anions through the use of
mono-24 or bi-dentate25 ligands which coordinate lithium.
The crystal structure of an oxazoline-containing benzylic
lithio-anion species nicely illustrates the structure of the
oxazoline-coordinated lithio-benzyl anion,26 suggesting a
prominent role in the increased rate of lithiation.
The scope of this reaction was examined to determine if
epoxides could be used as a chiral pool for the synthesis of
enantio-enriched isoxazolyl alcohols. Lateral metalation
followed by quenching with optically active epoxides
would be expected to give optically active alcohols, but
the degree of conservation of chirality was not necessarily
ensured. Attack of the anion at the terminal carbon atom
would result in no perturbation of the chiral center, and
therefore would be expected to retain its chirality intact.
Attack at the benzylic carbon, however, has been shown
in many systems to proceed by an impure SN2 mecha-
nism,27 tainted with at least small amounts of an SN1 prod-
uct. The isoxazole system succeeded in perfect inversion at
the benzylic carbon to give optically pure (+)-3b in addi-
tion to the other regioisomer, (ꢀ)-4b, in optical purity, as
determined by HPLC analysis. A halogenated derivative
of compound (ꢀ)-4b was synthesized to afford X-ray qual-
ity crystals of (+)-5 (Scheme 2), which were examined by
anomalous scattering. In this manner, the absolute config-
uration at the benzylic carbon was determined to be (S)
(Fig. 2) when the isoxazolyl anion was quenched with
(R)-styrene oxide. This corresponds to inversion at the
benzylic carbon. Optically active propylene oxide was also
examined, and complete stereoselectivity was observed by
HPLC for all desired products formed (3a, 3b, 4b).
Fig. 1. Hypothetical DHP docked into our homology model of the L-type
Ca2+ channel.15 Preparation of this ligand requires stereoselective access
to isoxazole 4.
technique12,13 coupled with the venerable Hantzsch di-
hydropyridine synthesis.14 Lateral metalation is a versatile
method for introducing a wide variety of substituents at the
C-5 position of the isoxazole. Deprotection and Hantzsch
synthesis offers a reliable method for conversion of the
C-4 oxazoline to a DHP (Fig. 1).15
Furthermore, the DHP can be considered a privileged
template,16,17 and renewed interest has been piqued by
the recent observation that DHPs appear to inhibit the p-
glycoprotein transporter (Table 1).18,19
In order to adequately delve into an examination of the
reaction pathway presented, three key classes of epoxides
were examined—aliphatic, vinylic, and benzylic—to deter-
mine what role, if any, the substitution on the epoxide
would play. Lateral metalation of isoxazole 2 followed by
treatment with propylene oxide gave exclusively terminal
attack to afford the 2° alcohol 3a in 66% yield. This reac-
tion proceeded much more efficiently than anticipated, in
light of the reported low degree of reactivity of resonance
stabilized anions toward epoxides.5 We were able to then
show that lateral metalation followed by electrophilic
quenching with two other key classes of epoxides resulted
in the formation of compounds 3b,c and 4b,c in fair to
good yields. Note that the vinylic and benzylic epoxides
resulted in a mixture of regioisomers corresponding to
attack at the most and the least substituted carbons of
the epoxide. This trend has been well noted in other sys-
tems, and is apparently due to competing steric and elec-
tronic influences.20 The more electrophilic (i.e., benzylic)
carbon is less sterically accessible to the anion than is the
less electrophilic (terminal) carbon of the epoxide. For
With the stereochemistry of this reaction determined, it
was desirable to alter reaction conditions in an attempt to
reverse or enhance the trend of regioselectivity found for
the reaction of the isoxazole anion with an epoxide. Styrene
oxide is an inexpensive, thoroughly studied system, and
reliably produces a mixture of regioisomers. Therefore, it
was chosen as a model substrate. We expected that
addition of a Lewis acid should further increase the C–O
bond length at the carbon most able to accommodate a
Table 1
Ratios of products formed from nucleophilic ring opening of unsym-
metrical epoxides
Epoxide
3:4
Yield
Propylene oxide
Styrene oxide
Butadiene monoxide
100:0
70:30
41:59
66
86
47

3080
J. K. Nelson et al. / Tetrahedron Letters 49 (2008) 3078–3082
HO
1. BuLi
O
N
O
N
N
N
O
R
R
CH3
CH3
CH3
CH3
O
+
2.
O
O
O
HO
N
N
R
CH3
CH3
CH3
CH3
CH3
CH3
a. R = Me
b. R = Ph
c. R= -CH=CH2
4. a-c
3. a-c
2
Scheme 2. Electrophilic quenching of isoxazolyl anion with monosubstituted epoxides.
complete reversal in the ratio of products formed (27:73
ratio of 3b:4b, respectively), though the yield was reduced
to 30%. Interestingly, changing the solvent to ether resulted
in the exclusive formation of 4b, albeit in only 10% yield.
Lipshutz showed that the use of Li-2-thienyl CuCN in
combination with a lithio-nucleophile formed a higher-
order cuprate species that selectively and efficiently trans-
fers the nucleophile to epoxides or other electrophiles.28b
However, in our hands the use of Li-2-thienyl CuCN
showed virtually no effect on the addition of the lithio-isox-
azole anion to styrene oxide save for the reduction in the
yield to 45%. Using ether as solvent yielded a similar situ-
ation: The ratio of 3b:4b was very little different from the
reaction sans Lewis Acid in ether, except that again the
yield was dramatically reduced (to 19%). Since starting
material was not recovered, we attribute the reduction in
yield to oligomeric or polymeric baseline TLC material.
The highly Lewis acidic BF3ꢁEt2O was found to produce
a very notable result. When styrene oxide was treated with
BF3ꢁEt2O, and subsequently cannulated into the isoxazole
anion, a mixture of a- and b-attack resulted to form 3b
and 4b, but an additional compound, 6, formed as well.
This, we postulated, resulted from acid-induced rearrange-
ment of the epoxide to phenyl acetaldehyde in situ,
followed by the attack of the isoxazolyl anion on the
aldehyde (Scheme 4).29 It was expected, then, that warming
of the BF3ꢁEt2O/styrene oxide solution prior to reaction
with the anion would induce complete rearrangement to
6. In the event, 6 was the only product isolated from the
reaction when the Lewis acid-activated epoxide was
allowed to react at room temperature prior to addition to
Fig. 2. Molecular structure (thermal displacement 30%) of (S)-(+)-5.
Hydrogen atoms omitted for clarity.
developing positive charge (benzylic carbon), resulting in a
product distribution favoring the production of 4b.28
Unfortunately, however, for cases in which the relative
ratio of 3b:4b was substantially altered, overall yield was
also significantly reduced. A table summarizing the Lewis
acids and reaction conditions used is given in Supplemen-
tary data.
The use of a catalytic amount (2 mol %) of Sm(HMDS)3
resulted in little change in the ratio of products 3b:4b
(66:34), though the overall yield was cut in half to 42%
combined yield.28a A full equivalent of Sm(HMDS)3
resulted in further improvement in distribution to give
Cl
p-ClC6H4N=C=O
N
O
N
O
HO
NH
O
CH3
CH3
Hexanes, heat
4 min
O
O
N
O
N
CH3
CH3
CH3
CH3
(S)-(-)-4b
(S)-(+)-5
Scheme 3. Derivatization of 4b for X-ray crystallography.

J. K. Nelson et al. / Tetrahedron Letters 49 (2008) 3078–3082
3081
OH
1. n-BuLi
3b (34%)
+
N
O
1. n-BuLi
N
O
2. BF3-Et2O
Styrene Oxide
-78oC
2. BF3-Et2O
Styrene Oxide
Room Temp.
CH3
CH3
CH3
4b (7%)
+
O
N
O
N
34%
CH3
CH3
CH3
CH3
6 (2.5%)
2
6
Scheme 4. BF3ꢁEt2O induced rearrangement of styrene oxide to phenyl acetaldehyde is controlled by temperature.
the anion (Scheme 3).30 The result also indicated a higher
yield of 6 than before, though still a paltry 34%. In sum-
mary, Lewis acids had little effect on the outcome of the
reaction of 2 with styrene oxide in most cases. More impor-
tantly, with Lewis acids of variable strengths, we have
demonstrated that the C–O bond elongation can be pushed
too far to actually induce rearrangement of the epoxide.
A plausible explanation for the reactivity observed is
illustrated by the hemi-labile, rigid, and sterically encum-
bered transition state 7, shown in Scheme 5. A dihapto
association (g2), might explain the enhanced rate of reac-
tion of this—now hemi-labile23,31—isoxazole anion system
with epoxides, compared to other highly functionalized
isoxazole carbanions lacking a coordinating nitrogen which
failed, in our hands, to exhibit epoxide reactivity.32 In the
aliphatic case, unfavorable non-bonding interactions with
the geminal methyls would encourage a coordinated
approaching epoxide to twist the less substituted carbon
toward the C-5 anion—effectively limiting access to the
more substituted carbon. Aryl p-coordination has been
postulated in a previous example of a rate enhancing
hemi-labile ligand,33 and such a g2 coordination, as in 7b
and c, now represents a favorable bonding interaction
which suffices to bring the benzylic or allylic epoxide posi-
tion into contact with the C-5 isoxazolyl wherein the anion
can twist34 to attack from the antiperiplanar direction pro-
ducing the inversion observed for (ꢀ)-4b and (S)-(+)-5.
We have presented an exploration into the scope and
limitations of epoxides for use in racemic and chiral pool
lateral metalation of isoxazoles. The isoxazole anion
proved surprisingly apt at epoxide ring opening, providing
a single regioisomer for the aliphatic example, and a
mixture of regioisomers for cases in which the epoxide
contained a,b-unsaturation. Our study of various Lewis
acids to control regiochemical outcome was met with lim-
ited success. The complete inversion of configuration at a
benzylic carbon attests to the potential usefulness of this
method.
Acknowledgments
The authors thank the National Institute of Neurologi-
cal Disease and Stroke for the continuing support of Grant
NS38444, and P20 RR015583. J.K.N. acknowledges IN-
BRE Grant P20 RR16454 and the Malcolm and Carol
Renfrew Scholarship Endowment, University of Idaho.
The Bruker (Siemens) SMART APEX diffraction facility was
established at the University of Idaho with the assistance
of the NSF-EPSCOR program and the M. J. Murdock Chari-
table Trust, Vancouver, WA, USA.
Supplementary data
Experimental section and full characterization of all
compounds, a table detailing Lewis acids and conditions
used, the procedure for X-ray analysis, tables of crystallo-
graphic information files (CIFs), HPLC-CSP traces of chi-
ral compounds, and a table of additional racemic examples
from Refs. 7 and 8. Supplementary data associated with
this article can be found, in the online version, at
doi:10.1016/j.tetlet.2008.03.059.
References and notes
1. Padwa, A.; Murphree, S. S. Prog. Heterocycl. Chem. I 2002, 14, 52–
74.
2. (a) Jacobsen, E. N.; Wu, M. H. In Comprehensive Asymmetric
Catalysis II; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
Springer: Berlin, 1999; Vol. 2, pp 621–677; (b) Katsuki, T.; Martin, V.
S. Org. React. 1996, 48, 1–300; (c) Aggarwal, V. K.; Richardson, J.
Chem Commun. 2003, 2644–2651; (d) Denmark, S. E.; Wu, Z. Synlett
1999, 847–859.
N
N
O
O
CH3
CH3
O
O
3. Charest, M. G.; Lerner, C. D.; Brubaker, J. D.; Siegel, D. R.; Myers,
A. G. Science 2005, 308, 395.
Li
O
Li
O
N
N
R
4. Taylor, S. K. Tetrahedron 2000, 56, 1149–1163.
5. Fairlamb, I. J. S. Annu. Rep. Prog. Chem., Sect. B. 2003, 99, 138–
160.
H3C
H3C
CH3
CH3
6. Brunelle, D. J. Tetrahedron Lett. 1981, 3699–3702.
7. Smith, M. P., Ph.D. Dissertation, University of Idaho, 1993.
8. Burns, C. T.; Natale, N. R. In 49th Northwest Regional ACS
Meeting, Anchorage Alaska, June 15–18, 1994.
7b,c
7
Scheme 5. Hemi-labile rationale for increased reactivity and observed
regioselectivity.

3082
J. K. Nelson et al. / Tetrahedron Letters 49 (2008) 3078–3082
9. Diana, G.; Volkots, D.; Nitz, T. J.; Bailey, T. R.; Long, M. A.;
Vescio, N.; Aldous, S.; Pevear, D. C.; Dutko, F. J.. J. Med. Chem.
1994, 37, 2421–2436.
10. Dannhardt, G.; Kiefer, W.; Lambrecht, G.; Laufer, S.; Mutschler, E.;
Schweiger, J.; Striegel, H. G. Eur. J. Med. Chem. 1995, 30, 839–850.
11. McKenna, J. I.; Schlicksupp, L.; Natale, N. R.; Maryanoff, B. E.;
Flaim, S. F.; Willett, R. D. J. Med. Chem. 1988, 31, 473–476.
12. Natale, N. R.; Niou, C.-S. Tetrahedron Lett. 1984, 3943–3946.
13. Natale, N. R.; Mirzaei, Y. R. Org. Prep. Proced. Int. 1993, 25, 515–
556.
25. Beak, P.; Reitz, D. B. Chem. Rev. 1978, 78, 275–316.
26. Stol, M.; Snelders, D. J. M.; de Pater, J. J. M.; van Klink, G. P. M.;
Kooijman, H.; Spek, A. L.; van Koten, G. Organometallics 2005, 24,
743–749.
27. (a) Lin, B.; Whalen, D. L. J. Org. Chem. 1994, 59, 1638–1641; (b)
Sankawa, U.; Sato, T. Tetrahedron Lett. 1978, 19, 981–984; (c) Biggs,
J.; Chapman, N. B.; Wray, V. J. Chem. Soc. B 1971, 71.
28. (a) Mukerji, I.; Wayda, A. L.; Dabbagh, G.; Bertz, S. H. Angew.
Chem. 1986, 98, 756–757; (b) Lipshutz, B. H.; Moretti, R.; Crow, R.
Org. Synth. 1990, 69, 80–88.
14. Hantzsch, A. Justus Liebigs Ann. Chem. 1882, 215, 1.
15. Zamponi, G. W.; Stotz, S. C.; Staples, R. J.; Andro, T. M.; Nelson, J.
K.; Hulubei, V.; Blumenfeld, A.; Natale, N. R. J. Med. Chem. 2003,
46, 87–96.
29. This effect has been noted elsewhere: (a) House, H. O. J. Am. Chem.
Soc. 1955, 77, 3070. Conversion of styrene oxide to phenyl acetalde-
hyde has been optimized (b) Kim, Jong Dae; Cha, Jin Soon Taehan
Hwahakhoe Chi 1983, 27, 73–75.
16. Mueller, G. In Chemogenomics in Drug Discovery; Kubinyi, H.,
Mueller, G., Eds.; John Wiley, 2004; p 19.
17. Triggle, D. J. Cell. Mol. Neurobiol. 2003, 23, 293–303.
18. Nobili, S.; Landini, I.; Giglioni, B.; Mini, E. Curr. Drug. Targets 2006,
7, 861–879.
30. Unpublished result. No starting material or desired product is
recovered on attempted oxidation with the Dess–Martin periodinane,
whereas the secondary alcohol 3b is readily oxidized in high yields,
see: Nelson, J. K.; Burkhart, D. J.; McKenzie, A.; Natale, N. R.
Synlett 2003, 2213–2215.
19. Voigt, B.; Coburger, C.; Monar, J.; Hilgroth, A. Bioorg. Med. Chem.
2007, 15, 5110–5113.
31. Ramirez, A.; Collum, D. B. J. Am. Chem. Soc. 1999, 121, 11114–
11121.
20. Zuidema, G. D.; Cook, P. L.; Van Zyl, G. J. Am. Chem. Soc. 1953, 75,
294–296.
32. The isoxazole anions cited below did not react with epoxides: (a)
Burkhart, D. J.; Zhou, P.; Blumenfeld, A.; Twamley, B.; Natale, N.
R. Tetrahedron 2001, 57, 8039–8046; (b) Han, X.; Li, C.; Rider, K. C.;
Blumenfeld, A.; Twamley, B.; Natale, N. R. Tetrahedron Lett. 2002,
43, 7673–7677.
21. (a) Natale, N. R.; McKenna, J. I.; Niou, C.-S.; Borth, M.; Hope, H. J.
Org. Chem. 1985, 50, 5660–5666; for an improved procedure for the
preparation of unhindered isoxazolyl–oxazolines, see: (b) Zhou, P.;
Blubaum, J. E.; Burns, C. T.; Natale, N. R. Tetrahedron Lett. 1997,
38, 7019–7020.
33. de Bruin, T. J. M.; Magna, L.; Raybaud, P.; Toulhoat, H.
Organometallics 2003, 22, 3404–3413.
22. Gawley, R. E.; Hart, G. C.; Bartolotti, L. J. J. Org. Chem. 1989, 54,
175–181.
34. Just such a g2 coordination has been observed in the solid state: (a)
Dohmeier, C.; Baum, E.; Ecker, A.; Koppe, R.; Schnockel, H.
Organometallics 1996, 15, 4702–4706; (b) Haslam, E. Shikimic
Acid Metabolism and Metabolites; John Wiley & Sons: New York,
1993.
´
23. Ramirez, A.; Lobkovsky, E.; Collum, D. B. J. Am. Chem. Soc. 2003,
125, 15376–15387.
24. Gawley, R. E.; Hart, G.; Goicoechea-Pappas, M.; Smith, A. L. J.
Org. Chem. 1986, 51, 3076–3078.