Practical, Broadly Applicable, α-Selective, Z-Selective, Diastereoselective, and Enantioselective Addition of Allylboron Compounds to Mono-, Di-, Tri-, and Polyfluoroalkyl Ketones

Article abstract of DOI:10.1021/jacs.7b05011

A practical method for enantioselective synthesis of fluoroalkyl-substituted Z-homoallylic tertiary alcohols has been developed. Reactions may be performed with ketones containing a polylfluoro-, trifluoro-, difluoro-, and monofluoroalkyl group along with an aryl, a heteroaryl, an alkenyl, an alkynyl, or an alkyl substituent. Readily accessible unsaturated organoboron compounds serve as reagents. Transformations were performed with 0.5-2.5 mol % of a boron-based catalyst, generated in situ from a readily accessible valine-derived aminophenol and a Z- or an E-?-substituted boronic acid pinacol ester. With a Z organoboron reagent, additions to trifluoromethyl and polyfluoroalkyl ketones proceeded in 80-98% yield, 97:3 to >98:2 α:? selectivity, >95:5 Z:E selectivity, and 81:19 to >99:1 enantiomeric ratio. In notable contrast to reactions with unsubstituted allylboronic acid pinacol ester, additions to ketones with a mono- or a difluoromethyl group were highly enantioselective as well. Transformations were similarly efficient and α- and Z-selective when an E-allylboronate compound was used, but enantioselectivities were lower. In certain cases, the opposite enantiomer was favored (up to 4:96 er). With a racemic allylboronate reagent that contains an allylic stereogenic center, additions were exceptionally α-selective, affording products expected from ?-addition of a crotylboron compound, in up to 97% yield, 88:12 diastereomeric ratio, and 94:6 enantiomeric ratio. Utility is highlighted by gram-scale preparation of representative products through transformations that were performed without exclusion of air or moisture and through applications in stereoselective olefin metathesis where Z-alkene substrates are required. Mechanistic investigations aided by computational (DFT) studies and offer insight into different selectivity profiles.

Full text of DOI:10.1021/jacs.7b05011

Article
pubs.acs.org/JACS
Practical, Broadly Applicable, α‑Selective, Z‑Selective,
Diastereoselective, and Enantioselective Addition of Allylboron
Compounds to Mono‑, Di‑, Tri‑, and Polyfluoroalkyl Ketones
Farid W. van der Mei, Changming Qin, Ryan J. Morrison, and Amir H. Hoveyda*
Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02467, United States
S
* Supporting Information
ABSTRACT: A practical method for enantioselective synthesis of fluoroalkyl-substituted Z-homoallylic tertiary alcohols has
been developed. Reactions may be performed with ketones containing a polylfluoro-, trifluoro-, difluoro-, and monofluoroalkyl
group along with an aryl, a heteroaryl, an alkenyl, an alkynyl, or an alkyl substituent. Readily accessible unsaturated organoboron
compounds serve as reagents. Transformations were performed with 0.5−2.5 mol % of a boron-based catalyst, generated in situ
from a readily accessible valine-derived aminophenol and a Z- or an E-γ-substituted boronic acid pinacol ester. With a
Z organoboron reagent, additions to trifluoromethyl and polyfluoroalkyl ketones proceeded in 80−98% yield, 97:3 to >98:2 α:γ
selectivity, >95:5 Z:E selectivity, and 81:19 to >99:1 enantiomeric ratio. In notable contrast to reactions with unsubstituted
allylboronic acid pinacol ester, additions to ketones with a mono- or a difluoromethyl group were highly enantioselective as well.
Transformations were similarly efficient and α- and Z-selective when an E-allylboronate compound was used, but enantio-
selectivities were lower. In certain cases, the opposite enantiomer was favored (up to 4:96 er). With a racemic allylboronate
reagent that contains an allylic stereogenic center, additions were exceptionally α-selective, affording products expected from
γ-addition of a crotylboron compound, in up to 97% yield, 88:12 diastereomeric ratio, and 94:6 enantiomeric ratio. Utility is
highlighted by gram-scale preparation of representative products through transformations that were performed without exclusion
of air or moisture and through applications in stereoselective olefin metathesis where Z-alkene substrates are required. Mechanistic
investigations aided by computational (DFT) studies and offer insight into different selectivity profiles.
as ammonium salt derivatives, promote the addition of
symmetric allyl− and allenyl−B(pin) compounds9 (pin =
pinacolato) as well as silyl-protected propargyl−B(pin)
reagents to trifluoromethyl ketones efficiently and with
high α- and enantioselectivity (up to >98:2 α:γ and 99:1 er;
Scheme 1a).10 A collection of experimental and computational
data suggests that in the favored mode of addition i (Scheme 1)
there is minimization of electron−electron repulsion caused by
the nonbonding electrons of the ketone oxygen and the fluorine
atoms through electrostatic interaction with the catalyst’s
ammonium group. We later showed that with a phosphinoy-
limine and the more Lewis acidic Zn(OMe)2 (vs NaOt-Bu)
reaction of γ-substituted allyl−B(pin) compounds proceeds
more efficiently and with far higher γ selectivity (Scheme 1b).11
Preferential formation of the γ addition products likely arises
from the ability of the Lewis acidic Zn salt to cause 1,3-boryl
shift12 (ii → iii → Z-iv, Scheme 1b) to occur faster than the
1. INTRODUCTION
Small organic molecules bearing a fluoroalkyl-substituted carbon
stereogenic center are central to development of therapeutics,1
agrochemicals,2 and materials.3 Practical, efficient, cost-effective,
and broadly applicable catalytic methods that generate such entities
with high enantioselectivity are much sought after. Although such
processes can deliver valuable and readily modifiable building
blocks in high enantiomeric purity, they remain relatively
uncommon.4−7 The limited number of reported studies entail
additions of allyl units to a small set of trifluoromethyl ketones,
need significant amounts4a,b (at times several equivalents)4a of
costly indium salts, can require days to reach completion,4b
and selectivities do not exceed 90:10 er (enantiomeric ratio).4c
Part of the challenge in developing such transformations is
the comparatively small difference in the size of the carbonyl
substituents and high electrophilicity causing nonselective
background addition. Thus, high enantioselectivity must be
attained by incorporation of well-defined electronic factors.
We have demonstrated that valine-derived chiral organo-
boron catalysts (Scheme 1),8 which probably exist in solution
Received: May 15, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. Relevant Previous Observations
Scheme 2. Goals of this Study
B
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
addition of the initially generated chiral allylboron intermediate
to an imine (i.e., in Scheme 1b: ii → α-addition product).
The above findings raise the following key questions: Would
the 1,3-borotropic shift give rise to a preference for formation
of the γ-addition product (vi via v, Scheme 2), or would the
corresponding α-addition product viii (via vii) be generated
preferentially in the case of more electrophilic trifluoromethyl
ketones? If γ-selective, then would the diastereomeric ratio (dr)
and er be high? If α-selective, would er be high or would the
product contain an E- or a Z-alkene? To what degree would
E:Z selectivity be under catalyst control and/or depend on the
stereochemistry of the organoboron reagent? Would additions
to the less electrophilic difluoro- and monofluoromethyl ketone
analogues occur with much lower enantioselectivity, as they did
when symmetric allyl−B(pin) compounds were used or will the
new method offer broader scope?10
Scheme 3. Initial Examination of Reactivity and Selectivity
Here, we describe the results of studies designed to
address the above questions. We detail the development of a
practical, broadly applicable, highly efficient, α-selective,
Z-selective, and enantioselective additions to fluoroalkyl
ketones. The considerable scope of the catalytic protocol, its
utility in chemical synthesis, the role of the stereochemical
identity of the organoboron reagent, the origin of high α-, Z-,
and enantioselectivity as well as the mechanistic basis for
the observed trends in reactivity and selectivity are presented
below.
a
Reactions carried out under N2 atm. Conversion and α:γ ratios
determined by analysis of 19F NMR spectra of unpurified product
mixtures ( 2%). Yields for purified α-addition products ( 5%). The
er was determined by HPLC analysis ( 1%). Experiments were run in
duplicate or more. See the Supporting Information for details.
2. RESULTS AND DISCUSSION
2.1. Initial Evaluation. We first examined the addition
of organoboron reagent Z-1a to trifluoroacetophenone 2a
(Scheme 3). With ligand ap-1 (ap = aminophenol), reaction
was complete in 12 h at 22 °C, affording 3a in 85% yield, >98:2
α:γ ratio, 83:17 Z:E selectivity, and 98:2 er. The transformation
was much faster at 60 °C (>98% conv, 30 min), and 3a was
isolated in 86% yield and 98:2 er, with only slightly lower α:γ
selectivity (96:4). However, Z selectivity was reduced signifi-
cantly (79:21 Z:E). The transformation with ap-2 was similarly
efficient and selective, which was unexpected since previous
studies had indicated transformations with a carboxylic ester
(vs dialkyl amide) are less efficient and less enantioselective.8
With more sterically demanding triphenylsilyl-substituted
ap-3,6k,10 efficiency (87−88% yield), α selectivity (>98:2),
and er (≥98:2) remained high (Scheme 3), along with a boost
in reaction rate (45 min) and Z selectivity, regardless of the
reaction temperature (≥96:4 vs 74:26−83:17 Z:E with ap-1,2).
would have a similar influence on various selectivity issues
remained to be determined.
3. SCOPE
3.1. Additions to Trifluoromethyl Ketones. With
0.5 mol % ap-3, 1.0 mol % Zn(OMe)2, and 1.2 equiv of
commercially available Z-1a (used as received), additions to
aryl-substituted trifluoromethyl ketones proceeded to com-
pletion at 22 °C in 1.5−12 h (Scheme 4). Products were
isolated in 87−98% yield, ≥98:2 α:γ selectivity, 88:12 to >98:2
Z:E selectivity and 96:4 to >99:1 er. Substrates containing
a sterically congested (e.g., 3b, 3d−f, Scheme 4), electron-
rich (e.g., 3c, 3h, 3o), or electron-deficient (e.g., 3l, 3f) aryl
group were suitable. Dimethylaminophenyl-substituted 3n and
methylthiophenyl-substituted 3p were readily generated with
exceptional efficiency and selectivity.
Heterocyclic substrates can be converted to the correspond-
ing homoallylic alcohols (Scheme 5). Indole- (7), furyl- (8a-b),
thienyl- (9a-b), and pyridyl-substituted (10) homoallylic
alcohols were isolated in up to 98% yield, >98% α selectivity,
98% Z selectivity, and >99:1 er. The lower er in the reactions
with 2-furyl and 2-thienyl ketones compared to those with their
3-substituted analogues are consistent with formerly disclosed
observations regarding allyl−B(pin) and allenyl−B(pin) additions.10
̀
The importance of this finding vis-a-vis utility in synthesis aside,
the clear influence of catalyst structure on Z:E selectivity is
notable (see below for analysis). When NaOt-Bu was used
instead of Zn(OMe)2, there was only 12% conversion after
28 h (22 °C, >98:2 α:γ, 97:3 Z:E, 98:2 er). As discussed
previously,11 the positive effect of the Zn(II) salt on efficiency
is probably tied to its coordination to the catalyst’s amide
carbonyl; this reduces carbonyl−boron coordination, resulting
in higher catalyst activity.
With acetophenone (vs the trifluoromethyl variant) and
ap-3, the γ-addition product was slightly favored (62% 6, eq 1).
This suggests that with a less electrophilic ketone, Lewis acid
catalyzed 1,3-borotropic shift becomes more competitive (see
ii → Z-iv vs ii → vii, Scheme 2). The near-complete erosion
of Z selectivity (60:40 vs 97:3 Z:E; eq 1) indicates that the
presence of the trifluoromethyl group is key to the formation of
the higher energy alkene isomer. The mechanistic origins of this
effect and whether a difluoro- or a monofluoromethyl group
C
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 4. Reactions with Aryl-Substituted Trifluoromethylketones
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product mixtures
( 2%). Yields for purified α-addition products ( 5%). The er was determined by HPLC analysis ( 1%). All experiments run in duplicate or more.
See the Supporting Information for details.
Regarding pyridyl product 10, 2.5 mol % ap-3, 5.0 mol %
Zn(OMe)2, 12 h, and 60 °C were required for >98%
conversion; this might be in part because the corresponding
hydrate was the starting material. The X-ray structure secured
for Z-10 allowed us to establish the absolute stereochemical
identity of the major product.13
Additions to other alkyl-substituted trifluoromethyl ketones
(11a−c, Scheme 6), including cyclohexyl-substituted 11c, pro-
ceeded to >98% conversion with 0.5 mol % ap-3 (one case at
2.5 mol %) at 22 °C in 4 h (Scheme 6). Transformations were
efficient (80−97% yield), generating the α-addition products
with considerable selectivity (97:3 to >99:1 α:γ). Levels of
Z-selectivity were surprisingly variable: Benzyl-substituted 11b
formed in 79:21 Z:E ratio (vs 92:8 and >98:2 for 11a and 11c,
respectively). There was a less dramatic fluctuation in enantio-
selectivity with cyclohexyl-substituted 11c (90:10 er vs 96:4 for
11a and 11b; see below for mechanistic analysis).
Reaction of the α,β-unsaturated trifluoromethyl-substituted
ketones afforded dienes 12a−c (Scheme 6) in 93−97% yield,
>98:2 α:γ selectivity, ≥98:2 Z:E selectivity, and 99:1 er.
Reactions affording 1,5-enynes 13a−c were similarly efficient
and α- and Z-selective (Scheme 6), but enantioselectivities were
notably lower (81:19−85:15 er; see below for rationale).
3.2. Additions to Polyfluoroalkyl Di- and Monofluoro-
methyl Ketones. Polyfluorinated products 14 and 15 were
generated efficiently and with excellent α selectivity, Z sel-
ectivity, and er. The highly enantioselective formation of difluoro-
and monofluoro-substituted 16a−c and 17a−c, in contrast, were
unexpected (96:4−98.5:1.5 er, Scheme 7); this was because
addition of allyl−B(pin) to these same ketones was minimally
enantioselective (Scheme 1a).10 The X-ray structure of
p-bromobenzoate derivative 18 showed that there is no change
in the sense of stereochemical induction (vs the trifluoromethyl
analogs). Moreover, these data show that the number of fluorine
D
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 5. Reactions with Heteroaryl-Substituted Trifluoromethyl Ketones
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product mixtures
( 2%). Yields for purified α-addition products ( 5%). The er was determined by HPLC analysis ( 1%). All experiments run in duplicate or more.
See the Supporting Information for details.
atoms adjacent to the ketone unit has considerable influence on
Z-selectivity of the process: The Z:E ratio descended in the
order of 97:3 (Z-3a, Scheme 3) to 91:9 (Z-16a, Scheme 7) to
78:22 (Z-17a) with three, two, and one fluorine atoms on the
α-alkyl site, respectively. The outcome of the reactions with
di- and monofluoromethyl-substituted ketones constitutes one
of the more distinct advantages of the present approach to
reactions with allyl−B(pin),10 where low er was observed with
substrates containing a smaller number of fluorine atoms (see
below for additional examples and analysis).
compared to that when Z-1a was used (24 h vs 45 min for >98%
conv); nonetheless, the expected homoallylic alcohol was isolated
in 86% yield and with exceptional α:γ and Z:E ratios (>98:2).
Transformations with E-1a and other trifluoromethyl ketones
generated variable selectivity patterns (Scheme 9; compared to
that with Z-1a, Scheme 4). With electron-rich aryl moieties,
enantioselectivity was moderate with preference for the same
enantiomer as Z-1a (i.e., 3c in 85:15 vs >99:1 er). However,
o-tolyl product 3b was formed in 44:56 er (vs >99:1 er with
Z-1a). For o-bromophenyl alcohol 3e, the opposite enantiomer
was slightly favored (24:76 er vs 98:2 er with Z-1a).
Remarkably, in the case of o-trifluoromethyl-substituted 3f,
there was near complete reversal of enantioselectivity (6:94 vs
98:2 er with Z-1a). High er (albeit moderate 85:15) in favor of
the same major enantiomer as Z-1a was observed with m- and
p-substituted products 3l and 3t and alkyl-substituted 11a
(96:4 and 97:3 er, respectively). There were two other
curious sets of data: (1) Z selectivity was high in all instances,
except for o-tolyl-substituted 3b (83:17 vs 96:4 to >98:2 Z:E,
respectively; Scheme 9). (2) There was reversal in the sense
of enantioselectivity in the additions involving trifluoromethyl
acetophenone compared to its di- and monofluoromethyl
analogues (i.e., Z-3a vs Z-16a and Z-17a). Moreover, whereas
addition to o-tolyl-substituted trifluoromethyl ketone led
to the formation of Z-3b in nearly racemic form (44:56 er),
corresponding di- and monofluoromethyl derivatives Z-16b and
Z-17b were generated in 9:91 and 4:96 er, respectively. The basis
of these selectivity trends will be discussed below.
3.3. Reactions with Other Organoboron Reagents.14
Transformations with n-heptyl-substituted Z-1a delivered
products 19a−c efficiently (86−95% yield, Scheme 8); although,
in some cases longer reaction times were needed for complete
conversion (24 h for 19a,b). Z selectivity ranged from 90−94%,
and α selectivities and enantioselectivities remained high (97:3 to
>98:2 and 98:2 to >99:1 er, respectively). With the larger
E-phenyl-substituted and γ,γ′-dimethyl-substituted allyl−B(pin)
compounds (20 and 21, respectively; Scheme 8), products were
isolated in 89−90% yield, but extended times were again
required, especially in the former case (72 h). When the reaction
was performed at 60 °C, there was complete conversion after
6 h, and the product was isolated in 90% yield. However,
selectivity levels were lower (92:8 α:γ, 76:24 Z:E, 80:20 er).
Enantioselectivity was lower in the latter two instances (85:15
and 89:11 er, respectively), and 21 was formed with 82:18 Z:E
selectivity.
3.4. Reactions with (E)-Crotyl−B(pin). The somewhat
diminished er for the reaction leading to prenyl alcohol 21
(Scheme 8) suggested that the corresponding transformation
involving E-crotyl−B(pin) (E-1a, Scheme 9) might be less
enantioselective: Z-3a was indeed generated in 91:9 er under
the aforementioned conditions. The process was less efficient
3.5. E- and Z-Organoboron Compounds React via
Different Intermediates. We began by examining why the
same major enantiomer (in most cases) and Z olefin isomer
(in all instances) were being formed, regardless of whether
Z- or E-1a is used. One explanation could be rapid isomerization
E
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 6. Reactions with Alkyl-, Alkenyl-, and Alkynyl-Substituted Trifluoromethyl Ketones
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product mixtures
( 2%). Yields for purified α-addition products ( 5%). The er was determined by HPLC analysis ( 1%). All experiments run in duplicate or more.
See the Supporting Information for details.
between the organoboron isomers. Nonetheless, spectroscopic
studies indicated that Z- or E-1a do not readily interconvert
(<2% in the presence of 5.0 mol % Zn(OMe)2, 1.3 equiv of
MeOH, toluene-d8, 22 °C, 16 h).15 Moreover, reaction with
rac-22 afforded 23a in 85:15 dr16 (Scheme 10; 84% yield,
>98:2 α:γ, 92:8 er for the anti diastereomer, 91:9 er for the syn
isomer). The stereochemical identity of the aforementioned
products was ascertained by obtaining the X-ray structure
of p-bromobenzoate derivative 24 (Scheme 10).13 These data
indicate that each organoboron isomer reacts by a distinct
pathway; otherwise an equal mixture of syn and anti isomers
would be formed (see below for further analysis).
3.6. Reactions with a Racemic Allylboron Reagent.
Aryl-, heteroaryl-, alkenyl-, and alkyl-substituted trifluoromethyl
ketones may be used in diastereo- and enantioselective reactions
with rac-22 (Scheme 10), although selectivity is moderate in
some cases. This is a notable method because of the low catalyst
loading (0.5 mol %), accessibility of the chiral catalyst, brief
reaction times (1.5−3 h), mild conditions (22 °C), high yields,
and exceptional α selectivity, as well as since to the best of our
knowledge related catalytic enantioselective protocols have not
been reported before. The organoboron compound (rac-22) can
be accessed in a single step;15 with enantiomerically enriched 22,
which can be synthesized by a number of methods,17 con-
siderably higher stereoselectivity should be expected.15
4. UTILITY
The method is amenable to gram-scale operations (Scheme 11).
The application to pyridine-substituted 10 again shows that
a basic pyridine moiety is tolerated. The er with which 3j,
precursor to antiparasitic agent Bravecto18 (presently sold as the
racemate) is formed in er higher than that with the addition of
allyl−B(pin) (96.5:3.5 vs 95:5 er).10 It is especially notable that
the gram-scale reactions shown in Scheme 11 were performed
without exclusion of air and/or moisture.
The Z-homoallylic alcohols are suitable substrates for a variety
of directed, highly diastereoselective transformations.19 Because
the homoallylic ether products contain a Z-alkene, they can be
readily converted to the corresponding Z-alkenyl halide derivatives
through stereoretentive catalytic cross-metathesis reactions.20 The
stereoselective synthesis of alkenyl chloride 27 (Scheme 12) and
its subsequent conversion to indole-containing 29 via boronic
F
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 7. Reactions with Other Fluoroalkyl-Substituted Ketones
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product mixtures
( 2%). Yields for purified α-addition products except for 17b−c ( 5%). The er was determined by HPLC analysis ( 1%). All experiments run in
duplicate or more. See the Supporting Information for details.
enriched tertiary alcohol (Z-3a) with commercially available
unsaturated alcohol 30; the desired product (31) was obtained
after 8 h at room temperature in 89% yield and >98:2 Z:E
selectivity. Here, the presence of the Z-methyl-substituted
alkenes is required for high efficiency, and stereoselectivity is
considerably higher compared to that when a terminal alkenes
is involved.23 Such transformations are valuable because by
a combination of catalytic cross-metathesis and cross-coupling
reactions a convenient route for the synthesis of otherwise
difficult-to-access enantiomerically enriched compounds becomes
feasible.
Scheme 8. Reactions with Other γ-Substituted Allyboron
Reagents
a
5. STEREOCHEMICAL MODELS
5.1. Reactions with Organoboron Compound Z-1a.
On the basis of DFT calculations (ωB97XD),15 the high
enantioselectivity in the transformations with Z-1a is due to
reaction via intermediate ix and complex I in preference to II
(Scheme 13). In II there is electron−electron repulsion between
the halogen atoms of the trifluoromethyl group and the catalyst’s
aryloxide, and a near-eclipsing interaction between the methyl
substituent and the catalyst’s B−O bond. Furthermore,
Coulombic attraction10 between the trifluoromethyl moiety and
the catalyst’s ammonium group is only possible in I.10 Another
competing mode of reaction, represented by III (Scheme 13),
would deliver the same sense of enantioselectivity as I but with an
E-alkene. As illustrated in III, the stabilizing Coulombic attraction
between the CF3 and the ammonium groups would be countered
by two significant repulsive interactions.
Variations in aminophenol structure and the associated changes
in Z selectivity and er provide further insight (Table 1). While
there is notable decrease in Z:E ratios with smaller catalyst
substituents (from 97:3 to 83:17 to 84:16 with ap-3, ap-4,
and ap-5, respectively), there is less significant diminution in er
(from 99:1 to 98:2 to 97:3, respectively). As steric hindrance
a
Reactions carried out under N2 atm. For Z-20, the E-allyl−B(pin)
reagent was used. Conversion, α:γ and Z:E ratios determined by
analysis of 19F NMR spectra of unpurified product mixtures ( 2%).
Yields are for purified α-addition products ( 5%). The er was
determined by HPLC analysis ( 1%). All experiments run in duplicate
or more. See the Supporting Information for details.
acid 28 is a case in point. The Z-1,2-substituted alkene is crucial
to cross-metathesis efficiency.20 Reaction of the corresponding
terminal alkene substrate was inefficient when a monoaryloxide
pyrrolide complex was used (∼40% conv, >98:2 Z:E)21 and did not
afford any product with monoaryloxide chloride complex Mo-1.
Another example (Scheme 12) performed with complex
Ru-122 involves the use of an unprotected enantiomerically
G
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 9. Reactions with Crotylboron Compound E-1a: Substantial Variations in Enantioselectivity
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product mixtures
( 2%). Yields for purified α-addition products ( 5%). The er was determined by HPLC analysis ( 1%). All experiments were run in duplicate or
more. See the Supporting Information for details.
caused by the substrate’s phenyl group and the aminophenol
substituent (G) is diminished in III, more of the E-alkene
isomer is formed. In contrast, repulsion between the ketone’s
trifluoromethyl group, which is smaller than Ph, and the same
catalyst moiety, is not as significant; therefore, the impact on er
is less.
account for the outcome of different types of transformations.26
Such interactions are unlikely to be as strong with the cor-
responding E-alkenes since a nonlinear geometry should translate
to less effective electrostatic attraction.
The lower enantioselectivities (81:19−85:15 er) for enynes
13a−c (Scheme 6) might be attributed to competing
electrostatic attraction between the catalyst’s ammonium
unit and the alkynyl group (IV). The slightly higher er for
more electron-poor p-trifluoromethylphenyl-substituted 13c
(85:15 vs 81:19 er for 13a) provide some support, but together
with nearly the same enantioselectivity for p-methoxyphenyl-
substituted 13b, these probably indicate that the change
in electron density at the C−C triple bond is not significant
enough for a stronger effect. Although interactions involving
a C−C triple bond have been investigated only to a limited
extent,24 related associations between an ammonium moiety and
aryl groups have been examined extensively25 and proposed to
5.2. Reactions with Organoboron Compound E-1a.
Enantioselectivities with E-crotyl−B(pin) are lower because of
the smaller energy difference between V and VI (Scheme 14).
The reason Z selectivity remains high in the reactions with
E-1a is because of two counts of steric and electron−electron
H
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 10. Diastereo- and Enantioselective Additions with
rac-22
Scheme 12. In Combination with Catalytic Z-Selective
a
a
Cross-Metathesis
a
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios
determined by analysis of 19F NMR spectra of unpurified product
mixtures ( 2%). Yields for purified α-addition products ( 5%). The
er was determined by HPLC analysis ( 1%). All experiments run in
duplicate or more. See the Supporting Information for details.
a
Conversion and Z:E ratios determined by analysis of 19F NMR
spectra of unpurified product mixtures ( 2%). Yields are for isolated
and purified α-addition products ( 5%). All experiments were run in
duplicate or more. See the Supporting Information for details.
repulsion in VII, which render it less competitive. Reducing the
size of the catalyst’s aryl substituent (G), as shown by the data
in Table 2, diminishes Z-selectivity to a small degree, suggesting
that the electronic repulsion and steric interaction involving the
a
Scheme 11. Reactions on Gram Scale
a
Reactions carried out without inert atmosphere. Conversion, α:γ and Z:E ratios determined by analysis of 19F NMR spectra of unpurified product
mixtures ( 2%). Yields for purified α-addition products ( 5%). The er was determined by HPLC analysis ( 1%). All experiments run in duplicate
or more. See the Supporting Information for details.
I
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 13. Stereochemical Models Based on DFT Studies with Z-Crotyl−B(pin)
a
DFT calculations were performed at the ωB97XD/DEF2TZVPP//ωB97XD/DEF2SVPtoluene(SMD) level. Free energy values for transition states are
relative to the most favorable alternative. See the Supporting Information for details.
Table 1. Effect of Ligand Structure on Stereoselectivity with
Z-Crotyl−B(pin)
triphenylsilyl unit is less influential, as indicated by the Z:E
selectivity values). Analogous modification of catalyst structure
has little impact on er (entries 1−3, Table 2), since there would
likely be similar lowering of steric repulsion in V and VI. The
preference for V is reflected in the adverse effect of a sizable
ortho aryl substituent on Z:E ratio (e.g., 3e−f, Scheme 4),
which arises from an increase in steric repulsion (vs that with
VII wherein there is a pseudoequatorial aryl moiety).
a
5.3. Reactions with Di- and Monofluoromethyl
Ketones. We have previously shown that electrostatic attraction
with the catalyst’s ammonium site is less influential with di- and
monofluoromethyl ketones.10 When Z-1a was used, products
such as difluoromethyl 16a and monofluoromethyl 17a were
generated in high er, indicating that reaction via VIII is more
favored (vs that with IX, Scheme 15a). This is in contrast to
allyl−B(pin) additions to the same ketones because with
the presence of the methyl substituent steric factors become
more influential such that high er can be attained. However,
Z-selectivity hinges on competition between VIII and X
(Scheme 15a), involving addition to different enantiotopic
faces of the ketone carbonyl, as confirmed through determina-
tion of absolute stereochemistry.15 The lower Z:E ratios thus
time (h);
yield
c
b
b
b
d
entry
G
conv (%)
α:γ
Z:E
(%)
er
1
2
3
SiPh3 (ap-3)
t-Bu (ap-4)
H (ap-5)
1.5; >98
12; >98
48; 45
>98:2
>98:2
>98:2
97:3
87
85
32
99:1
98:2
97:3
83:17
84:16
a
b
Reactions were carried out under N2 atm. Determined through
analysis of 19F NMR spectra of the unpurified product mixtures
c
( 2%). Yields are for isolated and purified α-addition products
( 5%). The er was determined through HPLC analysis ( 1%). See
the Supporting Information for details.
d
pseudo-equatorial methyl group in VII is the stronger factor
(i.e., repulsion between the ketone and Ph group of the
J
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 14. Stereochemical Models Based on DFT Studies with E-Crotyl−B(pin)
a
DFT calculations were performed at the ωB97XD/DEF2TZVPP//ωB97XD/DEF2SVPtoluene(SMD) level. Free energy values for transition states are
relative to the most favorable alternative. See the Supporting Information for details.
Table 2. Effect of Ligand Structure on Stereoselectivity with
E-Crotyl-B(pin)
reflect the lower energy difference between the steric hindrance
of the ketone substituent and the triphenylsilyl group in VIII as
well as the pseudoequatorial methyl group and the catalyst frame-
work in X. The slightly higher Z selectivity for difluoromethyl
ketones (vs the monofluoromethyl derivative) might be the result
of stronger electrostatic attraction between the fluoro-organic
moiety and the ammonium group.
a
The sense of enantioselectivity is reversed with E-1a
(Scheme 15b) because the reactions proceed via XI and XII.
Without electrostatic attraction, XII is favored due to lower
steric repulsion, a distinction more pronounced with a smaller
monofluoroalkyl substituent (22:78 vs 9:91 er for 16a and 17a,
respectively; Scheme 15b). It follows that with a more sterically
demanding aryl substituent, XI becomes less favored, resulting
in increased enantioselectivity to the extent that o-trifluor-
omethyl-substituted 3f may be generated in 6:94 er (Scheme 9).
time (h);
yield
c
b
b
b
d
entry
G
conv (%)
α:γ
Z:E
(%)
er
1
2
3
SiPh3 (ap-3)
t-Bu (ap-4)
H (ap-5)
24; >98
24; >98
24; 71
>98:2
>98:2
>98:2
>98:2
97:3
86
84
60
91:9
93:7
89:11
96:4
a
b
Reactions were carried out under N2 atm. Determined through
6. CONCLUSIONS
analysis of 19F NMR spectra of the unpurified product mixtures ( 2%).
c
d
We put forth methods for efficient and enantioselective addition
of readily accessible Z-γ-substituted boronic acid pinacol ester
compounds to fluoroalkyl-substituted ketones. The approach
Yields are for isolated and purified α-addition products ( 5%). The
er was determined through HPLC analysis ( 1%). See the Supporting
Information for details.
K
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
combination of electronic (e.g., electrostatic attraction between
the catalyst’s ammonium group and the fluoroalkyl groups)
and steric factors (e.g., induced by installation of a sizable
triphenylsilyl group) tertiary homoallylic alcohols that contain
a stereogenic carbon center with a hydroxyl and a fluoroalkyl
substituent may be prepared in high yield, high Z-selectivity,
and high enantiomeric purity. Mechanistic models based on
DFT studies provide a rationale for different selectivity profiles
and should be of value in future studies regarding this class of
enantioselective catalysts (Schemes 13−15, Tables 1−2).
The present work allows facile access to many valuable
enantiomerically enriched organofluorine compounds that
might be used in the preparation of therapeutic candidates,
novel materials, and/or more effective agrochemicals. Design
and development of additional boron-based chiral catalysts
and methods for enantioselective synthesis continue in these
laboratories.
Scheme 15. Influence of Number of Fluorine Atoms on
Stereoselectivity
a
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.7b05011.
Crystallographic information files for Z-10, 18, and 24
(CIF, CIF, CIF)
Experimental details for all reactions and analytic details
for all enantiomerically enriched products. Tables of
electronic and free energies and geometries of computed
structures. (PDF)
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: amir.hoveyda@bc.edu.
a
ORCID
Reactions carried out under N2 atm. Conversion, α:γ and Z:E ratios
determined by analysis of 19F NMR spectra of unpurified product
mixtures ( 2%). Yields are for purified α-addition products ( 5%).
The er was determined by HPLC analysis ( 1%). All experiments run
in duplicate or more. See the Supporting Information for details.
Amir H. Hoveyda: 0000-0002-1470-6456
Author Contributions
F.W.v.d.M. and C.Q. contributed equally to the creation of this
work.
Notes
has considerable range as aryl-, heteroaryl-, alkynyl-, alkenyl-,
and alkyl-substituted ketones with a polyfluoro-, trifluoro-,
difluoro-, and monofluoro-alkyl substituent can be used,
affording products with up to 98% yield and >99:1 er
(Schemes 4−8). Reactions can be catalyzed by 0.5−2.5 mol %
of a complex generated in situ from valine-derived aminophenol
ligand may be carried out at room temperature and are complete
within a few hours. Contrary to reactions with unsubstituted
allyl−B(pin), additions to mono- and difluoroalkyl substituted
ketones are highly enantioselective. A noteworthy aspect of the
transformations is that addition of the initial chiral organoboron
intermediate to a fluoro-ketone is faster than 1,3-boryl shift,
and as a result, α selectivity is often exceptionally high. Equally
notable are the high Z:E ratios (up to >98:2), an attribute that is
largely due to effective catalyst control of stereoselectivity. The
assortment of possibilities for functionalization of the stereo-
chemically defined alkene product, such as kinetically controlled
Z-selective olefin metathesis reactions, add to the value of the
approach.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support was provided by the NIH and the NSF
(GM-57212 and in part GM-59426 and CHE-1362763). We
are grateful to KyungA Lee, Thach T. Nguyen, Ming Joo Koh,
Sebastian Torker, and Chaofan Xu for valuable advice and
support.
REFERENCES
■
(1) Gillis, E. P.; Eastman, K. J.; Hill, M. D.; Donnelly, D. J.;
Meanwell, N. A. J. Med. Chem. 2015, 58, 8315−8359.
(2) Fujiwara, T.; O’Hagan, D. J. Fluorine Chem. 2014, 167, 16−29.
(3) Berger, R.; Resnati, G.; Metrangolo, P.; Weber, E.; Hulliger, J.
Chem. Soc. Rev. 2011, 40, 3496−3508.
(4) (a) Loh, T. P.; Zhou, J.-R.; Li, X.-R. Tetrahedron Lett. 1999, 40,
9333−9336. (b) Zhang, X.; Chen, D.; Liu, X.; Feng, X. J. Org. Chem.
2007, 72, 5227−5233. (c) Haddad, T. D.; Hirayama, L. C.; Taynton,
P.; Singaram, B. Tetrahedron Lett. 2008, 49, 508−511.
(5) For recent reviews on enantioselective synthesis through
additions of allyl groups to ketones and imines and their applications,
́ ́
see (a) Yus, M.; Gonzalez-Gomez, J. C.; Foubelo, F. Chem. Rev. 2011,
111, 7774−7854. For corresponding diastereoselective processes, see
We provide the first examples of efficient, diastereo-, and
enantioselective synthesis of the corresponding γ-addition type
products, where a racemic organoboron reagent adds to fluoro-
ketones with >98% α selectivity (Scheme 12). Thus, by a
L
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(b) Yus, M.; Gonzal
5595−5698.
́
ez-Gom
́
ez, J. C.; Foubelo, F. Chem. Rev. 2013, 113,
diastereo- and enantioselectivity, as indicated by investigations detailed
in the Supporting Information.
(17) For synthesis of allylboron compounds by catalytic
enantioselective methods, see (a) Ito, H.; Ito, S.; Sasaki, Y.;
Matsuura, K.; Sawamura, M. J. Am. Chem. Soc. 2007, 129, 14856−
14857. (b) Guzman-Martinez, A.; Hoveyda, A. H. J. Am. Chem. Soc.
2010, 132, 10634−10637. (c) Ito, H.; Okura, T.; Matsuura, K.;
Sawamura, M. Angew. Chem., Int. Ed. 2010, 49, 560−563. (d) Ito, H.;
Kunii, S.; Sawamura, M. Nat. Chem. 2010, 2, 972−976. (e) Park, J. K.;
Lackey, H. H.; Ondrusek, B. A.; McQuade, D. T. J. Am. Chem. Soc.
2011, 133, 2410−2413. (f) Park, J. K.; McQuade, D. T. Angew. Chem.,
Int. Ed. 2012, 51, 2717−2721. (g) Potter, B.; Edelstein, E. K.; Morken,
J. P. Org. Lett. 2016, 18, 3286−3289.
(6) For catalytic enantioselective additions of allyl moieties to
ketones (without proximal fluorine substituents), see (a) Kii, S.;
Maruoka, K. Chirality 2003, 15, 68−70. (b) Wada, R.; Oisaki, K.;
Kanai, M.; Shibasaki, M. J. Am. Chem. Soc. 2004, 126, 8910−8911.
(c) Kim, J. G.; Waltz, K. M.; Garcia, I. F.; Kwiatkowski, D.; Walsh, P. J.
J. Am. Chem. Soc. 2004, 126, 12580−12585. (d) Wadamoto, M.;
Yamamoto, H. J. Am. Chem. Soc. 2005, 127, 14556−14557.
(e) Thornqvist, V.; Manner, S.; Frejd, T. Tetrahedron: Asymmetry
2006, 17, 410−415. (f) Miller, J. J.; Sigman, M. S. J. Am. Chem. Soc.
2007, 129, 2752−2753. (g) Lou, S.; Moquist, P. N.; Schaus, S. E. J.
Am. Chem. Soc. 2006, 128, 12660−12661. (h) Barnett, D. S.; Moquist,
P. N.; Schaus, S. E. Angew. Chem., Int. Ed. 2009, 48, 8679−8682.
(i) Haddad, T. D.; Hirayama, L. C.; Taynton, P.; Singaram, B.
Tetrahedron Lett. 2008, 49, 508−511. (j) Huang, X.-R.; Chen, C.; Lee,
G.-H.; Peng, S.-M. Adv. Synth. Catal. 2009, 351, 3089−3095. (k) Shi,
S.-L.; Xu, L.-W.; Oisaki, K.; Kanai, M.; Shibasaki, M. J. Am. Chem. Soc.
2010, 132, 6638−6639. (l) Zhang, Y.; Li, N.; Qu, B.; Ma, S.; Lee, H.;
Gonnella, N. C.; Gao, J.; Li, W.; Tan, Z.; Reeves, J. T.; Wang, J.;
Lorenz, J. C.; Li, G.; Reeves, D. C.; Premasiri, A.; Grinberg, N.;
Haddad, N.; Lu, B. Z.; Song, J. J.; Senanayake, C. H. Org. Lett. 2013,
15, 1710−1713. (m) Robbins, D. W.; Lee, K.; Silverio, D. L.; Volkov,
A.; Torker, S.; Hoveyda, A. H. Angew. Chem., Int. Ed. 2016, 55, 9610−
9614.
(7) For examples of catalytic enantioselective additions of C-based
nucleophiles to trifluoromethyl ketones (with no secondary chelating
group), see (a) Zhang, G.-W.; Meng, W.; Ma, H.; Nie, J.; Zhang, W.-
Q.; Ma, J.-A. Angew. Chem., Int. Ed. 2011, 50, 3538−3542. (b) Motoki,
R.; Kanai, M.; Shibasaki, M. Org. Lett. 2007, 9, 2997−3000. (c) Cook,
A. M.; Wolf, C. Angew. Chem., Int. Ed. 2016, 55, 2929−2933.
(d) Noda, H.; Amemiya, F.; Weidner, K.; Kumagai, N.; Shibasaki, M.
Chem. Sci. 2017, 8, 3260−3269. (e) Zheng, Y.; Tan, Y.; Harms, K.;
Marsch, M.; Riedel, R.; Zhang, L.; Meggers, E. J. Am. Chem. Soc. 2017,
139, 4322−4325.
(18) Kaiser, F.; Korber, K.; Von Deyn, W.; Deshmukh, P.; Narine, A.;
̈
Dickhaut, J.; Bandur, N. G.; Langewald, J.; Culbertson, D. L.;
Anspaugh, D. D. International Patent WO 2012/007426 A1, January
19, 2012.
(19) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93,
1307−1370.
(20) Koh, M. J.; Nguyen, T. T.; Lam, J. K.; Torker, S.; Hyvl, J.;
Schrock, R. R.; Hoveyda, A. H. Nature 2017, 542, 80−85.
(21) Koh, M. J.; Nguyen, T. T.; Zhang, H.; Schrock, R. R.; Hoveyda,
A. H. Nature 2016, 531, 459−465.
(22) Koh, M. J.; Khan, R. K. M.; Torker, S.; Yu, M.; Mikus, M. S.;
Hoveyda, A. H. Nature 2015, 517, 181−186.
(23) Xu, C.; Shen, X.; Hoveyda, A. H. unpublished results.
(24) Ramachandran, P. V.; Gong, B.; Teodorovic’, A. V.; Brown, H.
C. Tetrahedron: Asymmetry 1994, 5, 1061−1074.
(25) (a) Dougherty, D. A. Science 1996, 271, 163−168. (b) Ma, J. C.;
Dougherty, D. A. Chem. Rev. 1997, 97, 1303−1324. (c) Mahadevi, A.
S.; Sastry, G. N. Chem. Rev. 2013, 113, 2100−2138. (d) Holland, M.
C.; Metternich, J. B.; Muck-Lichtenfeld, C.; Gilmour, R. Chem.
̈
Commun. 2015, 51, 5322−5325. (e) Kamps, J. J. A. G.; Khan, A.; Choi,
H.; Lesniak, R. K.; Brem, J.; Rydzik, A. M.; McDonough, M. A.;
Schofield, C. J.; Claridge, T. D. W.; Mecinovic,
22, 1270−1276.
(26) (a) Yamada, S. Org. Biomol. Chem. 2007, 5, 2903−2912.
(b) Yao, L.; Aube, J. J. Am. Chem. Soc. 2007, 129, 2766−2767.
́
J. Chem. - Eur. J. 2016,
(8) (a) Silverio, D. L.; Torker, S.; Pilyugina, T.; Vieira, E. M.;
Snapper, M. L.; Haeffner, F.; Hoveyda, A. H. Nature 2013, 494, 216−
221. (b) Wu, H.; Haeffner, F.; Hoveyda, A. H. J. Am. Chem. Soc. 2014,
136, 3780−3783.
́
(c) Knowles, R. R.; Lin, S.; Jacobsen, E. N. J. Am. Chem. Soc. 2010,
132, 5030−5032. (d) Yamada, S.; Fossey, J. S. Org. Biomol. Chem.
2011, 9, 7275−7281. (e) Hori, M.; Sakakura, A.; Ishihara, K. J. Am.
Chem. Soc. 2014, 136, 13198−13201.
(9) For a recent overview regarding applications of allylboron
reagents in chemical synthesis, see Diner, C.; Szabo,
Soc. 2017, 139, 2−14.
́
K. J. Am. Chem.
(10) (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.
(b) Mszar, N. W.; Mikus, M. S.; Torker, S.; Haeffner, F.; Hoveyda, A.
H. Angew. Chem., Int. Ed. 2017, DOI:10.1002/anie.201703844.
(11) van der Mei, F. W.; Miyamoto, H.; Silverio, D. L.; Hoveyda, A.
H. Angew. Chem., Int. Ed. 2016, 55, 4701−4706.
(12) For leading studies regarding the mechanism of 1,3-boryl shift
processes, see (a) Hancock, K. G.; Kramer, J. D. J. Am. Chem. Soc.
1973, 95, 6463−6465. (b) Hancock, K. G.; Kramer, J. D. J. Organomet.
Chem. 1974, 64, C29−C31. (c) Henriksen, U.; Snyder, J. P.; Halgren,
T. A. J. Org. Chem. 1981, 46, 3767−3768. (d) Buhl, M.; Schleyer, P. v.
̈
R.; Ibrahim, M. A.; Clark, T. J. Am. Chem. Soc. 1991, 113, 2466−2471.
(e) Bubnov, Y. N.; Gurskii, M. E.; Gridnev, I. D.; Ignatenko, A. V.;
Ustynyuk, Y. A.; Mstislavsky, V. I. J. Organomet. Chem. 1992, 424,
127−132.
(13) See the Supporting Information for details of X-ray structures.
(14) The organoboron compounds used to prepare 20a−c and 21
were prepared according to previously reported procedures. See
(a) Ely, R. J.; Morken, J. P. J. Am. Chem. Soc. 2010, 132, 2534−2535.
(b) Zhang, P.; Roundtree, I. A.; Morken, J. P. Org. Lett. 2012, 14,
1416−1419. The trisubstituted boronate used to synthesize 22 is
commercially available. See the Supporting Information for further
details..
(15) See the Supporting Information for details.
(16) The reason for the 85:15 dr is that the reaction of R- and S-24
afford the same anti-25a predominantly but with different levels of
M
DOI: 10.1021/jacs.7b05011
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX