(Z) -Trifluoromethyl-Trisubstituted Alkenes or Isoxazolines: Divergent Pathways from the Same Allene

Article abstract of DOI:10.1021/acs.orglett.0c02546

Because of a charge-dipole interaction involving nonbonding electron pairs on fluorine, protonation of trifluoromethyl allenes leads to tri- or tetrasubstituted alkenes with high (Z)-selectivity. Treatment of the same allenes with catalytic Au(I) initiates a reaction cascade that produces isoxazolines in high yield.

Full text of DOI:10.1021/acs.orglett.0c02546

pubs.acs.org/OrgLett
Letter
(Z)-Trifluoromethyl-Trisubstituted Alkenes or Isoxazolines:
Divergent Pathways from the Same Allene
Chaolun Liu, Glenn P. A. Yap, Casey A. Rowland, and Marcus A. Tius*
Cite This: https://dx.doi.org/10.1021/acs.orglett.0c02546
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Because of a charge-dipole interaction involving nonbonding
electron pairs on fluorine, protonation of trifluoromethyl allenes leads to tri- or
tetrasubstituted alkenes with high (Z)-selectivity. Treatment of the same allenes
with catalytic Au(I) initiates a reaction cascade that produces isoxazolines in high
yield.
uring our exploration of a Au(I)-catalyzed Nazarov
cyclization, we observed the rearrangement of enyne 1 in
wet dichloromethane to a single geometrical isomer of 2 that
was suppressed in the presence of 4 Å molecular sieves (eq 1;
AgNTf2 in a variety of solvents led to mixtures of geometrical
isomers of 4. The best Z/E ratio, 4:1, was observed in
dichloromethane and in 1,2-dichloroethane. Both solvents
contained traces of water. We suspected that traces of
triflimide in the reaction mixture catalyzed the isomerization
of 4. Control experiments showed that strong acid led to
isomerization, and although the process was slow in the case of
4, we were concerned that it might be substrate-dependent and
that it would take place during the reaction and/or the workup.
To avoid the problem, we made the fateful decision to
redesign the starting material by replacing the acetoxy group in
3 with the N-hydroxycarbamate in 6a (Scheme 1).8 The N-
hydroxy group was designed to intercept the postulated
cationic reaction intermediate, thereby circumventing the
formation of a ketone. Exposure of 6a to 5 mol % each of
(2-t-BuPhO)3PAuCl and AgNTf2 in wet dichloromethane at
room temperature led to (Z)-7a as the minor reaction product.
Although the ratio varied run to run, the major product was
always the unexpected isoxazoline 8a, and the reaction was
slow (∼42 h). Because charged intermediate(s) were being
formed, we next used a mixed solvent of dichloromethane and
HFIP (90/10 v/v). This led to a much faster reaction (2 h)
from which a 1:5 ratio of 7a/8a was isolated (Scheme 1). We
were unable to convert isolated 7a to 8a under any condition,
suggesting that there were two reaction manifolds proceeding
from 6a.9
D
L = (2-t-BuPhO)3P).1,2 In what follows, we describe the
synthesis of (Z)-trisubstituted trifluoromethyl alkenes or
isoxazolines. Vinyl trifluoromethyl groups are featured in a
number of pharmaceutical products,3 and although the
preparation of (E)-trisubstituted trifluoromethyl alkenes can
be accomplished with good control of the geometry, the
diastereoselective synthesis of the (Z)-isomers is more
challenging.4 Recent noteworthy examples of diastereoselective
syntheses of trisubstituted (Z)-trifluoromethyl alkenes that
make use of Togni’s reagent5,6 or Umemoto’s reagent through
a C−H activation process have been described.7 Although
these methods are effective, they use expensive sources of CF3,
which renders them impractical on scale. The method we
disclose exploits a stereoelectronic effect of fluorine to control
alkene geometry. Trifluoromethyl ketones serve as starting
materials, and the reaction proceeds under mild conditions.
We prepared propargylic acetate 3 from lithium phenyl-
acetylide and trifluoroacetone followed by acetic anhydride (eq
2). Exposure of 3 to 5 mol % each of (2-t-BuPhO)3PAuCl and
Received: July 30, 2020
https://dx.doi.org/10.1021/acs.orglett.0c02546
© XXXX American Chemical Society
Org. Lett. XXXX, XXX, XXX−XXX
A

Organic Letters
pubs.acs.org/OrgLett
Letter
Scheme 1. Preparation of Alkenes and Isoxazolines
As we were unable to detect a reaction intermediate in wet
dichloromethane, we next conducted the experiment in the
presence of 4 Å molecular sieves and lowered the catalyst load
to 2 mol % (Scheme 1). The reaction was followed by
monitoring the disappearance of the CF3 singlet in 6a by 19F
NMR (−82 ppm in CDCl3).9 After 1 h, allene 9a was isolated
in 87% yield.10 After 3 days at room temperature, allene 9a
underwent complete conversion to alkene 7a (major (Z)-
isomer). This suggested that traces of DCl in the NMR solvent
had catalyzed the conversion of 9a to 7a and provided us with
a clue to understanding the mechanism.11 A series of control
experiments confirmed that conversion of 9a to 7a took place
in the absence of Au(I) and was catalyzed by acid.9 No
conversion of 9a to 7a was detected in acid-free chloroform
after 14 h, whereas addition of 1 equiv of 1 M HCl to the
chloroform led to complete conversion to 7a in the same
period of time. As isoxazoline 8a was not detected from the
acid treatment of 9a, it seemed likely that its formation was
catalyzed by Au(I). Indeed, treatment of 9a with 0.5 mol %
each of (2-t-BuPhO)3PAuCl and AgNTf2 in acid-free wet
chloroform at room temperature led to the exclusive formation
of 8a. The results indicate that allene 9a is the common
intermediate to 7a and 8a. Both processes were examined.
A series of allenes 9 were converted to trisubstituted alkenes
7 by exposure to 5 equiv of trifluoroacetic acid in chloroform at
room temperature for 2−5 min (Figure 1).9 The alkene
isomers were easily separable chromatographically in most
cases with the (E)-isomer more mobile than the (Z). The CF3
singlet of the (Z)-isomer was always at higher field in the 19F
NMR in comparison to that of the (E)-isomer. Distinguishing
the (Z)- and (E)-isomers in 7k and in 7n−7p required a
different technique.12 Our stereochemical assignment of
alkenes 7 was based on the positive NOE between the vinyl
hydrogen atom and the protons(s) on the cis-carbon atom in
the major isomer. Single-crystal X-ray crystallography of 7c
confirmed our assignment.9
Figure 1. Trisubstituted alkenes. Conditions: (a) 5 equiv of
CF3CO2H, CHCl3, rt, 2−5 min. The first yield reported is of the
isomeric mixture. The ratio of isomers was determined by integration
of the CF3 signal in the 19F NMR spectra. The second yield refers to
the (Z)-isomer following purification by column chromatography.
aThe minor (E)-isomer was not separated.
Scheme 2. Tetrasubstituted Alkenes
(53%) as the major geometrical isomer. The stereochemistry
of 10 was confirmed by X-ray crystallography.9 Exposure of 9a
to iodine at 40 °C led to diene 11 (65%). Both reactions were
terminated through proton loss from the cyclohexyl methine
carbon, and in both cases, the electrophile and the
trifluoromethyl group were trans. This provides an indication
that the trend observed for protons is general and can be used
for the preparation of tetrasubstituted alkenes and dienes.
To determine whether the (Z)-preference would be
observed for electrophiles other than protons, we performed
the reactions of Scheme 2. Treatment of allene 9a with
mercuric trifluoroacetate followed by brine gave diene 10
B
https://dx.doi.org/10.1021/acs.orglett.0c02546
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
Compound 10 (Figure 2) is the sixth reported structure with
a chloromercury group σ-bonded to an alkenyl group13−17 and
conversion of 6 to isoxazolines 8 took place in a single step.
Treatment of the carbamates with 5 mol % each of (2-t-
BuPhO)3PAuCl and AgNTf2 in chloroform and HFIP (60/40
v/v) at room temperature in the presence of 4 Å molecular
sieves led to isoxazolines in 69−85% yield.
Our results had puzzled us because water in the solvent was
necessary for the conversion of propargyl acetate 3 to enone 4,
yet conducting the reactions of 6a in a number of different
solvents containing varying amounts of water led to isoxazoline
8a as the major product. This made it appear that a protic
medium disfavored the formation of alkene 7a. Only after we
realized that weak and strong acids led to different reaction
outcomes did things fall in place, and we were able to
formulate a mechanistic hypothesis. The (Z)-preference in the
case of alkene 7a cannot be driven by a steric effect as there is a
negligible difference in size between methyl and trifluor-
omethyl groups.18 This strongly suggests that the trifluor-
omethyl group directs the approach of the proton to the allene
9, as indicated in transition state 12. As the proton approaches
the sp-hybridized allene carbon atom, polarization of one of
the nonbonding electron pairs on a fluorine atom stabilizes the
developing positive charge on C1, thereby lowering the energy
of 12. This charge-dipole effect has been described19 and has
been applied very ingeniously by Hoveyda in his catalytic
asymmetric allylation of trifluoromethyl ketones.20,21 This also
rationalizes the attenuated (Z)-selectivity in the cases in which
R1 in 12 is phenyl (e.g., 7k, Figure 1). Delocalizing the
developing positive charge on 12 into the aromatic ring
diminishes the importance of the contribution made by
fluorine that selectively lowers the transition state energy
leading to the (Z)-isomer. The isopropyl group in 7b, 7g, and
7n increases the preference for the (Z)-isomer, presumably for
steric reasons. Although it is not possible to draw broad
conclusions from the small number of examples incorporating
a pentafluoroethyl or heptafluoro-n-propyl group, there do not
appear to be any inconsistencies. For example, the much higher
(Z)-selectivity in the case of 7n compared to that of 7p likely
results from the larger steric demand of the isopropyl group in
7n that destabilizes the (E)-isomer. Comparison of 7l (CF3; Z/
E 93:7), 7p (CF2CF3; Z/E 84:16), and 7o (CF2CF2CF3; Z/E
79:21) shows a steady erosion in (Z)-selectivity as the larger
perfluoroalkyl group destabilizes the (Z)-isomer. Having one
less fluorine atom on the allylic carbon atom with which to
stabilize 12 may also contribute to the erosion of (Z)-
selectivity through a statistical effect. This may also explain the
lower (Z)-selectivity in the case of difluoromethyl compound
7q.
Figure 2. Molecular diagram of 10·1.5(HgCl2): purple, Hg; green, Cl;
yellow, F; blue, N; red, O; gray, C; white, H.
is the first structure with a halomercury moiety bonded to
alkenyl−CF3. The Hg−Cl, 2.3349(13) Å, and the CC,
1.325(7) Å, distances are within the previously reported ranges
of 2.299−2.379 and 1.318−1.391 Å, respectively (Figure 2).
The Hg−C distance, 2.073(5) Å, is similar to that of five
earlier structures, 2.001−2.068 Å.14−17 Two HgCl2 units, one
at an inversion center, were found in the asymmetric unit. The
σ-bonded Hg atom is intramolecularly coordinated to the
enone O atom. The Hg atom of the fully occupied HgCl2 unit
is coordinated to the amide and enone O atoms. In addition,
there are several intermolecular short-contact interactions
between the Hg atoms with neighboring Cl atoms (2.27−3.57
Å).
The isoxazoline synthesis was examined next (Figure 3). As
the rearrangement of the propargylic N-hydroxycarbamates 6
to allenes 9 and the subsequent decarboxylative cyclization to
the isoxazolines are both catalyzed by cationic Au(I),
The process that converts allenes 9 to isoxazolines 8 has not
been described.22,23 Our first clue to understanding the process
came from recognizing that strong acids catalyze the rapid
conversion of allenes to alkenes 7 in the absence of Au(I),
whereas the isoxazolines formed in the presence of Au(I) with
weak acids. Our postulated mechanism is summarized in
Scheme 3. Allenes 9 that are formed from the Au(I)-catalyzed
isomerization of carbamates 6 are likely in equilibrium with 13.
In the absence of acid in an aprotic medium, this equilibrium is
unproductive; however, in the presence of weak acids such as
Figure 3. Isoxazolines. Conditions: (a) 5 mol % of (2-t-
BuPhO)3PAuCl, 5 mol % of AgNTf2, CHCl3/HFIP (60/40 v/v), 4
Å MS, rt, 1−4 h. Yields after purification by column chromatography
on silica gel.
C
https://dx.doi.org/10.1021/acs.orglett.0c02546
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
Casey A. Rowland − Department of Chemistry and
Biochemistry, University of Delaware, Newark, Delaware
19716, United States
Scheme 3. Postulated Mechanism for Isoxazolines
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.orglett.0c02546
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Francisco Lopez-Tapia and Mr. Cody Dickinson
for helpful discussions and the University of Hawaii NMR
facility for assistance with acquiring NMR data. G.P.A.Y.
(University of Delaware) gratefully acknowledges the National
Science Foundation CRIF CHE-1048367.
water or HFIP, proton transfers can take place, leading to 14
that undergoes rearrangement to carbamic acid 15 with loss of
Au(I). Irreversible loss of CO2 leads to allenyl ether 16 that
undergoes Au(I)-catalyzed cyclization to 17 with loss of a
proton. Protiodeauration converts 17 to isoxazoline 8, thereby
completing the catalytic cycle. The isoxazolines are not
detected in the presence of strong acids because the conversion
of 9 to alkenes 7 takes place much more rapidly. This
mechanism also explains why alkenes 7 cannot be converted to
isoxazolines 8.
We have developed a synthesis of the difficult to access (Z)-
trifluoromethyl-trisubstituted alkenes.24 This distinguishes our
reaction from a related Meyer−Schuster reaction that proceeds
under thermodynamic control and leads to (E)-trifluoromethyl
alkenes.25 The selectivity for the observed geometrical isomer
is likely due to the polarization of a nonbonding electron pair
on fluorine that stabilizes the transition state leading to the
(Z)-isomers. There are surely many opportunities to exploit
this effect in synthesis and in catalysis. In the process of
developing the alkene synthesis, we also discovered a Au(I)-
catalyzed decarboxylative rearrangement leading to isoxazo-
lines.
REFERENCES
■
(1) Congmon, J.; Tius, M. A. Eur. J. Org. Chem. 2018, 2018, 2926−
2930.
(2) For recent reviews of the Nazarov cyclization, see: (a) West, F.
G.; Scadeng, O.; Wu, Y.-K.; Fradette, R. J.; Joy, S. Comp. Org. Synth.
2014, 5, 827−866. (b) Wenz, D. R.; Read de Alaniz, J. Eur. J. Org.
Chem. 2015, 2015, 23−37. (c) Shirinian, V. Z.; Yadykov, A. Adv.
Synth. Catal. 2020, 362, 702−723. (d) Vinogradov, M. G.; Turova, O.
V.; Zlotin, S. G. Org. Biomol. Chem. 2017, 15, 8245−8269.
(3) 5-Trifluoromethyluracil (trifluridine): (a) Heidelberger, C.;
Parsons, D.; Remy, D. C. J. Am. Chem. Soc. 1962, 84, 3597−3598.
Fludelone: (b) Rivkin, A.; Yoshimura, F.; Gabarda, A. E.; Chou, T.-
C.; Dong, H.; Tong, W. P.; Danishefsky, S. J. J. Am. Chem. Soc. 2003,
125, 2899−2901. Autoinflammatory agent: (c) Kelly, M.; Kincaid, J.;
Duncton, M.; Sahasrabudhe, K.; Janagani, S.; Ravindra, R.; Wu, G.;
Fang, Y.; Wei, Z. Patent Appl. WO 2006093832, 2006. Panomifene:
(d) Borvendeg, J.; Toldy, L.; Horvath, T.; Abraham, G.; Tory, K.;
Hermann, I.; Kiss, E.; Csanyi, E. Patent Appl. DE 3030802, 1981.
(4) Recent examples of vinyl trifluoromethyl alkene synthesis:
(a) Shang, T.; Zhang, J.; Zhang, Y.; Zhang, F.; Li, X.-S.; Zhu, G. Org.
Lett. 2020, 22, 3667−3672. (b) Zhang, S. L.; Xiao, C. J. Org. Chem.
2018, 83, 10908−10915. (c) Straathof, N. J. W.; Cramer, S. E.;
ASSOCIATED CONTENT
* Supporting Information
■
sı
̈
Hessel, V.; Noel, T. Angew. Chem., Int. Ed. 2016, 55, 15549−15553.
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.orglett.0c02546.
́
(d) Tresse, C.; Schweizer, S.; Bisseret, P.; Lalevee, J.; Evano, G.;
Blanchard, N. Synthesis 2016, 48, 3317−3330. (e) Yamamoto, Y.;
Ohkubo, E.; Shibuya, M. Green Chem. 2016, 18, 4628−4632. (f) Zhu,
N.; Wang, F.; Chen, P.; Ye, J.; Liu, G. Org. Lett. 2015, 17, 3580−3583.
Experimental procedures and NMR spectra (PDF)
Accession Codes
̈
(g) Hafner, A.; Brase, S. Adv. Synth. Catal. 2011, 353, 3044−3048.
(5) Feng, C.; Loh, T.-P. Angew. Chem., Int. Ed. 2013, 52, 12414−
CCDC 2010303−2010304 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
12417.
́
́
(6) Frederic, C. J. M.; Cornil, J.; Vandamme, M.; Dumitrescu, L.;
Tikad, A.; Robiette, R.; Vincent, S. P. Org. Lett. 2018, 20, 6769−6773.
(7) Besset, T.; Cahard, D.; Pannecoucke, X. J. Org. Chem. 2014, 79,
413−418.
(8) Prasanthi, A. V. G.; Begum, S.; Srivastava, H. K.; Tiwari, S. K.;
Singh, R. ACS Catal. 2018, 8, 8369−8375.
AUTHOR INFORMATION
Corresponding Author
■
(9) For details, see the Supporting Information.
(10) The structure of the allene was confirmed by IR (1990 cm−1)
by the CF3 singlet at −66 ppm in the 19F NMR and by the 13C
chemical shift of the sp-hybridized carbon atom (195 ppm in CDCl3).
(11) Because the N-hydroxyl group in allene 9a introduces 1 equiv
of a proton and because of water in the CDCl3, no deuterium was
detected in product 7a.
Marcus A. Tius − Chemistry Department, University of Hawaii
at Manoa, Honolulu, Hawaii 96822, United States;
orcid.org/0000-0002-7092-8993; Email: tius@hawaii.edu
Authors
(12) In the 1H NMR, the vinyl proton of the (Z)-isomer appeared at
lower field than the (E)-isomer.
Chaolun Liu − Chemistry Department, University of Hawaii at
Manoa, Honolulu, Hawaii 96822, United States
Glenn P. A. Yap − Department of Chemistry and Biochemistry,
University of Delaware, Newark, Delaware 19716, United
States
(13) Kuz’mina, L. G.; Bokii, N. G.; Rybinskaya, M. I.; Struchkov, Yu.
T.; Popova, T. V. J. Struct. Chem. 1971, 12, 1026−1031.
(14) Drew, M. G. B.; Mann, J.; Pietrzak, B. J. Chem. Soc., Perkin
Trans. 1 1985, 1049−1053.
D
https://dx.doi.org/10.1021/acs.orglett.0c02546
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
pubs.acs.org/OrgLett
Letter
(15) Sugiura, K.; Kato, A.; Iwasaki, K.; Miyasaka, H.; Yamashita, M.;
Hino, S.; Arnold, D. P. Chem. Commun. 2007, 20, 2046−2047.
(16) Kuz’mina, L. G.; Struchkov, Yu. T.; Dmitriev, P. I.; Shapiro, A.
B. Coord. Chem. 1987, 13, 404−409.
(17) Kuz’mina, L. G.; Porai-Koshits, M. A. Coord. Chem. 1989, 15,
185−190.
(18) The van der Waals radii of CH3 and CF3 are both ∼2 Å:
Wermuth, C. G. The Practice of Medicinal Chemsitry, 2nd ed.;
Academic Press: London, 2003; p 207.
(19) (a) Gooseman, N. E. J.; O’Hagan, D.; Peach, M. J. G.; Slawin,
A. M. Z.; Tozer, D. J.; Young, R. J. Angew. Chem., Int. Ed. 2007, 46,
5904−5908. Reviews: (b) O’Hagan, D. Chem. Soc. Rev. 2008, 37,
308−319. (c) Zimmer, L. E.; Sparr, C.; Gilmour, R. Angew. Chem., Int.
Ed. 2011, 50, 11860−11871.
(20) (a) Lee, K.; Silverio, D. L.; Torker, S.; Robbins, D. W.;
Haeffner, F.; van der Mei, F. W.; Hoveyda, A. H. Nat. Chem. 2016, 8,
768−777. See also: (b) van der Mei, F. F.; Qin, C.; Morrison, R. J.;
Hoveyda, A. H. J. Am. Chem. Soc. 2017, 139, 9053−9065. (c) Fager,
D. C.; Lee, K.; Hoveyda, A. H. J. Am. Chem. Soc. 2019, 141, 16125−
16138. (d) Fager, D. C.; Morrison, R. J.; Hoveyda, A. H. Angew.
Chem., Int. Ed. 2020, 59, 11448−11455.
(21) A reviewer suggested that the stereochemistry of the reaction
may also be explained by electron donation from the distal allene π
into the C−CF3 σ* orbital.
(22) Reviews of 4-isoxazoline synthesis: (a) Pinho e Melo, T. M. V.
D. Eur. J. Org. Chem. 2010, 2010, 3363−3376. (b) Freeman, J. P.
Chem. Rev. 1983, 83, 241−261.
(23) Au(I)-catalyzed syntheses of heterocycles: (a) Patil, N. T.;
Yamamoto, Y. Chem. Rev. 2008, 108, 3395−3442. (b) Varela, I.;
Faustino, H.; Diez, E.; Iglesias-Siguenza, J.; Grande-Carmona, F.;
Fernandez, R.; Lassaletta, J. M.; Mascarenas, J. L.; Lopez, F. ACS
Catal. 2017, 7, 2397−2402. (c) Muratore, M. E.; Konovalov, A. I.;
Armengol-Relats, H.; Echavarren, A. M. Chem. - Eur. J. 2018, 24,
15613−15621. (d) Tian, X.; Song, L.; Rudolph, M.; Rominger, F.;
Hashmi, A. S. K. Org. Lett. 2019, 21, 4327−4330. (e) Wang, Q.;
Hoffmann, S.; Schießl, J.; Rudolph, M.; Rominger, F.; Hashmi, A. S.
K. Eur. J. Org. Chem. 2020, 2020, 2384−2388.
(24) (Z)-Trifluoromethyl allylic alcohols can be prepared in a single
step by reducing the cyclic carbamate function. See the Supporting
Information.
(25) Ramasamy, M.; Lin, H. C.; Kuo, S.-C.; Hsieh, M.-T. Synlett
2019, 30, 356−360.
E
https://dx.doi.org/10.1021/acs.orglett.0c02546
Org. Lett. XXXX, XXX, XXX−XXX