Ru-Based Catechothiolate Complexes Bearing an Unsaturated NHC Ligand: Effective Cross-Metathesis Catalysts for Synthesis of (Z)-α,β-Unsaturated Esters, Carboxylic Acids, and Primary, Secondary, and Weinreb Amides

Article abstract of DOI:10.1021/jacs.9b02318

Despite notable progress, olefin metathesis methods for preparation of (Z)-α,β-unsaturated carbonyl compounds, applicable to the synthesis of a large variety of bioactive molecules, remain scarce. Especially desirable are transformations that can be promoted by ruthenium-based catalysts, as such entities would allow direct access to carboxylic esters and amides, or acids (in contrast to molybdenum-or tungsten-based alkylidenes). Here, we detail how, based on the mechanistic insight obtained through computational and experimental studies, a readily accessible ruthenium catechothiolate complex was found that may be used to generate many α,β-unsaturated carbonyl compounds in up to 81% yield and ≥98:2 Z/E ratio. We show that through the use of a complex bearing an unsaturated N-heterocyclic carbene (NHC) ligand, for the first time, products derived from the more electron-deficient esters, acids, and Weinreb amides (vs primary or secondary amides) can be synthesized efficiently and with high stereochemical control. The importance of the new advance to synthesis of bioactive compounds is illustrated through two representative applications: An eight-step, 15% overall yield, and completely Z-selective route leading to an intermediate that may be used in synthesis of stagonolide E (vs 11 steps, 4% overall yield and 91% Z, previously), and a five-step, 25% overall yield sequence to access a precursor to dihydrocompactin (vs 13 steps and 5% overall yield, formerly).

Full text of DOI:10.1021/jacs.9b02318

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Ru-Based Catechothiolate Complexes Bearing an Unsaturated NHC
Ligand: Effective Cross-Metathesis Catalysts for Synthesis of (Z)‑α,β-
Unsaturated Esters, Carboxylic Acids, and Primary, Secondary, and
Weinreb Amides
Zhenxing Liu,†,# Chaofan Xu,†,# Juan del Pozo,† Sebastian Torker,*,†,‡ and Amir H. Hoveyda*,†,‡
†Department of Chemistry, Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02467, United States
‡Supramolecular Science and Engineering Institute, University of Strasbourg, CNRS, 67000 Strasbourg, France
S
* Supporting Information
ABSTRACT: Despite notable progress, olefin metathesis
methods for preparation of (Z)-α,β-unsaturated carbonyl
compounds, applicable to the synthesis of a large variety of
bioactive molecules, remain scarce. Especially desirable are
transformations that can be promoted by ruthenium-based
catalysts, as such entities would allow direct access to carboxylic
esters and amides, or acids (in contrast to molybdenum- or
tungsten-based alkylidenes). Here, we detail how, based on the
mechanistic insight obtained through computational and experimental studies, a readily accessible ruthenium catechothiolate
complex was found that may be used to generate many α,β-unsaturated carbonyl compounds in up to 81% yield and ≥98:2 Z/E
ratio. We show that through the use of a complex bearing an unsaturated N-heterocyclic carbene (NHC) ligand, for the first
time, products derived from the more electron-deficient esters, acids, and Weinreb amides (vs primary or secondary amides)
can be synthesized efficiently and with high stereochemical control. The importance of the new advance to synthesis of
bioactive compounds is illustrated through two representative applications: an eight-step, 15% overall yield, and completely Z-
selective route leading to an intermediate that may be used in synthesis of stagonolide E (vs 11 steps, 4% overall yield and 91%
Z, previously), and a five-step, 25% overall yield sequence to access a precursor to dihydrocompactin (vs 13 steps and 5% overall
yield, formerly).
selective synthesis of bioactive molecules (e.g., motualevic acid
B7 and 6-nor-absicic acid,8 Scheme 1b); Weinreb amides, which
are convenient precursors to aldehydes and ketones (e.g., en
route to dihydrocompactin,9 Scheme 1b), are especially
versatile. These reactions would complement Wittig-type olefin
synthesis or partial hydrogenation of an internal ynonate, which
are at times moderately stereoselective (particularly with
amides).10 Over-reduction would no longer be an issue, and
the starting materials would be more readily available (vs an
alkyne) and/or robust (vs an aldehyde). Strongly basic
conditions (e.g., KH,11 KOt-Bu,12 KHMDS)13 to perform a
reaction or to hydrolyze an ester to an acid, low temperatures
(−78 °C),10,14 and/or stoichiometric amounts of toxic additives
(e.g., 18-crown-6)11 would not be needed.
Only a limited number of (Z)-α,β-unsaturated t-butyl enoates
can be obtained by cross-metathesis (Scheme 1c);15 these
reactions are promoted by Mo-based monoaryloxide pyrrolide
(MAP) complexes, which readily decompose in the presence of
a carboxylic acid, and relinquish their catalytic activity when an
α,β-unsaturated amide is involved (Scheme 1c). Cross-meta-
thesis protocols for synthesis of (Z)-α,β-unsaturated carbonyl
INTRODUCTION
■
An endothermic, often turnover-limiting, step in a catalytic cycle
is typically followed by an exothermic process; as such, any
lowering of the energy barrier for the former event, although
seemingly minute, can lead to notable improvements in
efficiency. The degree to which subtle modifications within a
catalyst structure impact the energetics of a process, including
those that can impede catalyst decomposition, are crucial.
However, implementing such alterations requires a certain
degree of mechanistic appreciation. A recent example in
stereoselective olefin metathesis1 is in regards to catalyst
variations meant to stabilize or circumvent the formation of
unstable methylidene species.2 Another instance relates to the
impact of a pair of chlorine atoms within the catechothiolate
ligand of Ru-1a3 (Scheme 1a); the carbenes derived from Ru-
1b4 are longer-living because of the attenuation in electron
density of the sulfide ligands.5 The emergence of Ru-1c6 was
based on the idea that the syn-to-NHC mcb (NHC, N-
heterocyclic carbene; mcb, metallacyclobutane) can better
accommodate a larger alkene substrate (Scheme 1a).
We recently faced the problem of insufficient catalyst
longevity while considering the development of cross-metathesis
reactions that generate (Z)-α,β-unsaturated carboxylic acids,
esters, or amides, compounds frequently needed for stereo-
Received: March 1, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 1
Scheme 2. Initial Experimental Findings
a
Reactions were carried out under N2 atmosphere. Conversion and
Z/E ratios determined by analysis of 1H or 13C NMR spectra of
unpurified product mixtures ( 2%). Yields for purified products
( 5%). Experiments were run in duplicate or more. See the
Supporting Information for details.
enoate after 8 h at ambient temperature. Although 2a and 3a
were generated in >98:2 Z/E selectivity, extensive catalyst
decomposition was observed along with formation of trisub-
stituted α,β-unsaturated ester byproducts 4a and 4b (Scheme 2;
according to analysis of 1H NMR spectra).18
Mechanistic Analysis. In a reaction that is expected to
generate a (Z)-α,β-unsaturated carboxylate, productive cross-
metathesis reaction likely proceeds via ts1productive, mcbproductive
,
and then ts2productive (Scheme 3). An alternative nonproductive
route might be facilitated by an electron-withdrawing carboxyl
group at Cα in mcbnonproductive because the presence of this
moiety would stabilize the electron density buildup at the carbon
of the Cα−Ru bond.19 Density functional theory (DFT) studies,
carried out on complexes lacking the two Cl substituents for
simplicity, revealed to us that the carboxylic substituent can
lower the barrier to metallacyclobutane distortion20,21 (→
mcbdistorted), reducing the severity of trans influence caused by
the Cα−Ru and Ru−S(cis) bonds (see mcbnonproductive
→
tsdistorted → mcbdistorted; Scheme 3). The CNHC−Ru−Cα angle in
compounds preferentially deliver E isomers (substrate con-
trol).16
a saturated complex thus widens from 91.0° in mcbnonproductive to
18
Herein, we provide the details of an investigation where by
examination of the mechanistic principles, a readily accessible
Ru-based catechothiolate catalyst was identified for efficient and
stereoselective cross-metathesis reactions that afford (Z)-α,β-
unsaturated carboxylic esters, acids, Weinreb amides, and
primary and secondary amides.
115.8° in mcbdistorted
.
DFT studies further indicated that
mcbdistorted can be the gateway to catalyst decomposition, which
may occur by β-hydride elimination,20−22 rendered feasible by
the availability of a vacant ligation site, to generate a
trisubstituted enoate (see mcbdistorted → tsβ‑H elim → π-allyl →
trisub enoate; Scheme 3).
We reasoned that by lowering the electron-donating ability of
the NHC ligand we might be able to diminish the trans influence
in mcbproductive therefore minimizing the significance of the
stabilization caused by a carboxylate unit at Cα of
mcbnonproductive. This led us to consider a Ru catechothiolate
complex with an unsaturated NHC ligand. Our plan was partly
based on acidity (pKa) measurements23 and 31P and 79Se NMR
chemical shifts for NHC-phosphinide24 and NHC-selenium25
adducts, suggesting that the saturated NHCs are stronger σ-
donors and better π-acceptors.26 Moreover, the positive effect of
an unsaturated NHC ligand in a Ru-based catalyst on tacticity of
ring-opening metathesis polymerization of norbornene was
recently documented.27
RESULTS AND DISCUSSION
■
Preliminary Experimental Findings. Initial studies
showed that the available Ru-based complexes would be largely
ineffective. As a reminder, we used disubstituted alkenes as
starting materials because Ru catechothiolate complexes
decompose readily in the presence of most monosubstituted
alkenes (unstable methylidene).2 To address this problem, we
recently introduced a strategy, namely, through in situ capping
of the terminal alkene.2 Furthermore, many (Z)-1,2-disubsti-
tuted alkenes are commercially or readily accessible, and
catalytic cross-metathesis reaction offers a convenient method
for converting such entities to considerably more valuable
alkenes.
Cross-metathesis reaction between (Z)-butenoic benzyl ester
(1a) or carboxylic acid (1b) and (Z)-hex-3-ene (Scheme 2) in
the presence of Ru-1c, suitable for other challenging cases (e.g.,
trisubstituted allylic alcohols and ethers),17 was inefficient. With
5.0 mol % Ru-1c there was 32% consumption of the starting
While we were aware that earlier investigations with dichloro
Ru carbene complexes had shown that catalysts bearing a
saturated NHC are in fact more effective in promoting olefin
metathesis,28−30 we appreciated that a clear picture, vis-a-vis, the
̀
electronic differences between saturated and unsaturated NHC
ligands, has yet to emerge. Although Tolman electronic
B
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 3. DFT Studies Regarding Factors Leading to Catalyst Decomposition and the Byproducts Formed
a
See the Supporting Information for details. ts, transition state; NHC, N-heterocyclic carbene; mcb, metallacyclobutane; β-H elim, β-hydride
elimination.
parameters, which signify combined σ-donor and π-acceptor
properties, are similar for the two NHC types,31 redox potential
values obtained experimentally for Ir-32 and Ru-based33
complexes are inconclusive, indicating that the saturated
variants could be stronger or weaker electron donors compared
to their unsaturated counterparts.
It was for the following reasons that we ultimately decided to
probe the chemistry of Ru-1d (Scheme 4): (a) There are
fundamental differences between reactions of dichloro and
catechothiolate catalysts. Most notably, whereas transforma-
tions proceed via anti-to-NHC metallacyclobutanes in the
earlier system, syn-to-NHC metallacyclobutanes are involved
with bidentate catechothiolates;3 this enhances the impact of
steric pressure induced by an NHC ligand. (b) DFT studies
indicated that the presence of an unsaturated NHC lowers the
barrier to mcbproductive (ΔGrel = 12.4 kcal/mol for ts2productive vs
13.4 kcal/mol for the saturated complex; Scheme 3).
Furthermore, formation of mcbdistorted (ΔGrel = 19.2 kcal/mol
for tsdistorted vs 18.0 kcal/mol for the saturated complex) and β-
hydride elimination (ΔGrel = 15.3 kcal/mol for tsβ‑Helim vs 13.9
kcal/mol for the saturated complex) become energetically more
demanding, which should result in a longer catalyst life span.
These variations may be attributed to differences in trans
influence, which are probably stronger in a complex with a
saturated NHC. Specifically, DFT investigations (Scheme 3)
revealed that the Ru−S(trans) bond might be longer in
mcbproductive with a saturated NHC (2.423 vs 2.418 Å,
respectively), and could thus be higher in energy than its
corresponding unsaturated analogue (ΔGrel = 11.6 vs 11.0 kcal/
mol, respectively).(c) With the more “upwardly” oriented N-aryl
moieties in an unsaturated NHC ligand (see mcb-I′ vs mcb-I,
Scheme 4), there might be less of a driving force toward
Scheme 4
structural distortions that lessen the barrier to mcbdistorted
.
Proof-of-Principle Experiments. Efficiency Profile Is
Substrate-Dependent. To probe the validity of the above
hypotheses, we determined the efficiency of cross-metathesis
reaction between (Z)-hex-3-ene and different (Z)-butenoic acid
derivatives in the presence of Ru-1d versus Ru-1c (Scheme 5;
≥98% Z in all cases). In contrast to the reaction with Ru-1c,
affording 2a in 28% yield (Scheme 5), the transformation with
Ru-1d afforded the same product in 78% yield. Furthermore,
although the transformations with carboxylic acid 1b and
Weinreb amide 1c in the presence of Ru-1c proceeded to 32−
35% conversion (25% yield), (Z)-disubstituted alkene products
3a and 5a were isolated in 68−70% yield when unsaturated
C
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
illustrate the higher activity and longer life span of the
unsaturated complex.
Scheme 5. Cross Methods with Ru Complexes 1c versus 1d
We reasoned that β-hydride elimination should be less
favored if the catalyst derived from Ru-1c and a deuterium-
substituted alkene were to be used, resulting in higher yielding
cross-metathesis reaction. In the event, whereas the reaction
with saturated complex Ru-1c, 1a-d2, and (Z)-hexene-d2 was
more efficient than when 1a was used (47% vs 28% yield,
respectively; Scheme 6), there was minimal change when Ru-1d
was involved (78−81% yield). When the rate of β-hydride
elimination is reduced there is a narrower gap between the
efficiency of processes carried out with Ru-1c and Ru-1d.
X-ray Structures of the Ru Complexes. We then focused
on gathering insight regarding the following questions: (a) Why
are reactions that generate a (Z)-α,β-unsaturated carboxylic
ester, a carboxylic acid, or a Weinreb amide higher yielding with
unsaturated complex Ru-1d (vs Ru-1c)? (b) Why does the same
distinction not extend to transformations of primary and
secondary amides?
a
Reactions were carried out under N2 atmosphere. Conversion and
Z/E ratios were determined by analysis of 1H or 13C NMR spectra of
unpurified product mixtures ( 2%). Yields for purified products
( 5%). Experiments were run in duplicate or more. See the
Supporting Information for details.
We began by obtaining the X-ray crystal structures of Ru-1c
and Ru-1d (Scheme 7). The slightly shorter Ru−CNHC bond
Scheme 7. X-ray Crystal Structures of Ru Complexes 1c
a
versus 1d
complex Ru-1d was used. The reactions of secondary amide 1d
and primary amide 1e efficiently afforded 6a and 7a,
respectively, regardless of the catalyst involved, raising several
intriguing mechanistic questions.
Enoate Byproducts. With the knowledge that reactions are
higher yielding in several key cases with unsaturated complex
Ru-1d, we decided to determine whether catalyst decom-
position is a factor. We investigated the process involving enoate
1a and (Z)-hex-3-ene (Scheme 6), establishing that trisub-
a
Scheme 6. Gauging the Influence of β-Hydride Elimination
a
See the Supporting Information for details.
length in imidazolium-containing Ru-1d (2.055 vs 2.072 Å for
Ru-1c) and the more distorted CNHC−Ru−S(trans) (further
from linearity; 139.6° vs 149° for Ru-1c) are contrary to our
initial expectation, which was based on higher Lewis basicity of,
and stronger trans influence caused by, a saturated NHC ligand.
This apparent anomaly is likely due to crystal packing forces.
While there appears to be π−π interaction34 between the
dithiolate ligand and one of the aryl groups on the NHC in Ru-
1d (i.e., smaller CNHC−Ru−S(trans) angle), the same type of
association seems to be disrupted in Ru-1c by either the
positioning of the 2-fluoro-6-methyl phenyl moiety (Scheme 7),
the isopropoxy group, or the aromatic ring of the benzylidene of
a neighboring molecule in the crystal.18 The Ru−S(trans) bond
length in Ru-1c, as would be predicted (weaker π−π association
and stronger trans influence), is somewhat longer than in Ru-1d
(2.283 vs 2.265 Å).
a
Reactions were carried out under N2 atmosphere. Conversion and
Z/E ratios determined by analysis of 1H or 13C NMR spectra of
unpurified product mixtures ( 2%). Yields for purified products
( 5%). Experiments were run in duplicate or more. See the
Supporting Information for details.
stituted alkenes 4a and 4b are indeed formed as byproducts in
the transformations with Ru-1c and Ru-1d (by GC-MS
analysis18). Moreover, the enoates were generated in distinct
ratios: ∼70:30 for Ru-1c and ∼20:80 for Ru-1d, respectively.
Because 4a arises from reaction of starting enoate 1a while 4b is
derived from product 2a, the aforementioned ratios further
An aspect of these X-ray structures that is less susceptible to
crystal packing forces, but is equally informative, relates to the
N−CNHC−N bond angles.23 The larger corresponding value of
106.5° for saturated complex Ru-1c (vs N−CNHC−N angle =
103.9° for Ru-1d) likely means that the N-aryl substituents are
D
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 8. Examination of the Influence of Enoate Structure on the Rate of Olefin Metathesis
a
1
Reactions were carried out under N2 atmosphere. Conversion and Z/E ratios determined by analysis of H or 13C NMR spectra of unpurified
product mixtures ( 2%). Experiments were run in duplicate or more. See the Supporting Information for details.
a
Scheme 9. Scope of the Method: Synthesis of (Z)-α,β-Unsaturated Carboxylic Esters
a
1
Reactions were carried out under N2 atmosphere. Conversion and Z/E ratios determined by analysis of H or 13C NMR spectra of unpurified
product mixtures ( 2%). Yields for purified products ( 5%). Condition A: 1.0 equiv of monosubstituted olefin, 1.0 mol % Ru-1d, 5.0 equiv of (Z)-
butene, thf, 22 °C, 1 h; 5.0 equiv. (Z)-enoate, 5.0 mol % Ru-1d, 100 Torr, thf, 1 h, ambient pressure, 7 h. Condition B: 1.0 equiv of
monosubstituted olefin, 1.0 mol % Ru-1d, 5.0 equiv of (Z)-butene, thf, 22 °C, 1 h; 5.0 equiv of (Z)-enoate, 4.0 mol % Ru-1d, 100 Torr, thf, 1 h,
ambient pressure, 7 h; 4.0 mol % Ru-1d, 100 Torr, thf, 1 h, ambient pressure, 11 h. Condition C: 1.0 equiv of monosubstituted olefin, 5.0 equiv of
(Z)-enoate, 4.0 mol % Ru-1d, 100 Torr, thf, 1 h, ambient pressure, 7 h; 4.0 mol % Ru-1d, 100 Torr, thf, 1 h, ambient pressure, 11 h. Experiments
were run in duplicate or more. See the Supporting Information for details.
further projected toward the dithiolate ligand, increasing the
steric pressure associated with the formation of a syn-to-NHC
metallacyclobutane. A wider N−CNHC−N angle indicates that in
catalysts derived from Ru-1c, where there is an NHC with
nonbonding electrons in an orbital that has higher p-character,
there is stronger trans influence in the derived metal-
lacyclobutanes. The same electronic factor is in all likelihood
exacerbated by the weaker inductive (electron-withdrawing)
ability of the nitrogen atoms, which are attached to a sp3-
hybridized carbon in Ru-1c (vs sp2-hybridized C in Ru-1d).
Impact of Alkene Structure on Rate of Olefin Meta-
thesis. A possible reason why cross-metathesis with a primary
or a secondary amide (1d and 1e) is efficient, no matter which
Ru catechothiolate complex is used, is because these alkenes are
more electron-rich (vs an α,β-unsaturated ester, acid, or
Weinreb amide), and as a result, catalyst decomposition is
E
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
slower. This hypothesis is in line with the fact that there is
minimal conversion when an enone, an enal, or an alkenyl nitrile,
which are even more electron-deficient, is used.
efficiency was observed. The examples provided in Scheme 10
are illustrative.
a
To probe further, we examined the difference between
homometathesis of alkyl-substituted alkene 8 and its cross-
metathesis with α,β-unsaturated ester 1a (Scheme 8). Although
symmetrical alkene 9 was generated with similar efficiency with
Ru-1c or Ru-1d, cross-metathesis product 10 was formed faster
with the latter unsaturated complex (52% conversion to ∼1:3
9:10 vs 20% conversion to ∼3:1 9:10 with Ru-1c). Control
experiments underscored two more points:18 (a) The above-
mentioned difference in chemoselectivity does not arise from
the ability of Ru-1d to react with 9 and convert it to cross-
metathesis product 10. (b) A Lewis base can diminish the
catalytic activity of a Ru catechothiolate complex.35 Further-
more, a carboxylic ester is more detrimental to the activity of Ru-
1d compared to Ru-1c, probably because of the somewhat
higher Lewis acidity of the former, causing it to coordinate more
readily with the Lewis basic carboxyl group.
We also investigated the relative rates of homometathesis of
alkyl-substituted alkene 8 and its cross-metathesis reaction with
benzyl ester 1a (Scheme 8; by 19F NMR spectroscopy).
Regardless of whether Ru-1c or Ru-1d was used, conversion
of 8 to symmetrical 9 proceeded at similar rates. However,
whereas homometathesis of 8 was slightly more facile than its
cross-metathesis with Ru-1d (khomo/kcross = 1.5; Scheme 8),
Scheme 10. Feasibility of Single-Batch Processes
a
Reactions were carried out under N2 atmosphere. Conversion and
Z/E ratios determined by analysis of 1H or 13C NMR spectra of
unpurified product mixtures ( 2%). Experiments were run in
duplicate or more. See the Supporting Information for details.
(Z)-α,β-Unsaturated Acids. The approach can be used to
synthesize (Z)-α,β-unsaturated carboxylic acids directly (e.g.,
3b−f, Scheme 11). These entities cannot be prepared easily by
Wittig-type olefin synthesis (regardless of stereochemistry) and
access via the corresponding carboxylic esters demands basic or
acidic conditions, which raise functional group compatibility
issues and/or cause olefin isomerization. Despite the higher
catalyst loading (see Condition D, Scheme 11), the ability to
access 3g directly by cross-metathesis can improve the efficiency
of a multistep sequence; this is underscored in the context of a
route leading to naturally occurring antifungal/cytotoxic agent
stagonolide E (Scheme 11). Enantiomerically pure 1,3-diene 11,
obtained in seven steps and 28% overall yield from commercially
available materials, was converted to (Z)-carboxylic acid 3g in
53% yield and >98:2 Z/E selectivity. The eight-step procedure,
delivering 3g in 15% overall yield, compares favorably with the
previously shortest reported 11-step sequence, carried out in the
context of a total synthesis of the natural product,14 affording the
same intermediate in 4% overall yield and as a 91:9 Z/E mixture
of isomers (from a Horner−Wadsworth−Emmons reaction).
(Z)-α,β-Unsaturated Weinreb Amides. We were able to
access a number of different (Z)-α,β-unsaturated N-methoxy-N-
methyl (Weinreb) amides (5b−i, Scheme 12). Unlike when
Mo- or W-based catalysts are involved (see Scheme 1c), the
Lewis basic amide moiety did not adversely impact efficiency.
This set of products are notable because they can be efficiently
transformed to (Z)-enones36 or (Z)-enals,37 which similar to
(Z)-α,β-unsaturated alkenyl nitriles,38 cannot be directly
prepared by the present strategy (<10% conv). The stronger
electron-withdrawing ability of a ketone, an aldehyde, or a nitrile
probably means a greater preference for nonproductive
pathways (see mcbnonproductive, Scheme 3), culminating in
more extensive catalyst decomposition. Attempts to perform
stereoretentive cross-metathesis with other (Z)-α,β-unsaturated
tertiary amides were unsuccessful; this was largely as the result of
the difficulties associated with preparation of the requisite (Z)-
methyl-substituted substrates (e.g., facile isomerization to the E
isomer during conversion of an acid to a dimethyl amide).
there was a greater rate difference when Ru-1c was used ((khomo
/
k
cross = 4.6). These data show that although reactions are faster
with Ru-1d it is with the more electron-deficient substrate that
efficiency of the reactions with the saturated complex (Ru-1c)
suffers significantly. As before, we attribute this to the reduced
trans influence in mcbproductive (NHC−Ru−S(trans)) which
lowers the barrier to productive metathesis, rendering entities
derived from Ru-1d catalytically more active.
(Z)-α,β-Unsaturated Esters. To explore the scope of the
method, we began with (Z)-α,β-unsaturated esters, which were
obtained in 55−76% yield and 96:4 to >98:2 Z/E selectivity
(Scheme 9). This included products bearing a benzyl (2b−k) or
the more hindered tert-butyl ester (2l,m) group. Reactions
proceeded readily in the presence of an aldehyde (2f), a
carboxylic acid (2g), an unprotected indole (2i), or an
unprotected phenol (2k), many of which (except for an
unprotected indole) are functional units that quickly disable
Mo- and W-based as well as certain Ru-based complexes.4
Formation of sterically hindered, α-branched (Z)-alkene 2m is
noteworthy. Benzylic olefins, sterically demanding β-branched
starting materials, emerged as suitable substrates.
We developed several sets of conditions, optimal depending
on the circumstances. As noted previously, unless hindered (i.e.,
α-branched) terminal olefins were used, initial capping with Z-
butene was required.2 Subsequently, with 1.0 mol % Ru-1d and
(Z)-butene (thf, 22 °C, 1 h) and (Z)-butene removal (100
Torr), either one (5.0 mol %, Condition A) or two batches (4.0
mol % each, Condition B) of Ru-1d may be added. As the
syntheses of 2b−d, and 2k indicate, yields were higher when the
Ru complex was added sequentially (Condition B). In the case
of an α-branched alkene (cf. 2m), where homocoupling and
formation of the sensitive Ru-methylidene is slow, there is no
need for capping (Condition C).
It should be noted that cross-metathesis reactions can be
performed by mixing two (Z)-alkene substrates and a single
batch of Ru-1d. However, when the reaction mixture was treated
with two sequential batches of Ru-1d (Scheme 10) higher
F
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 11. Scope of the Method: Direct Synthesis of (Z)-α,β-Unsaturated Carboxylic Esters
a
1
Reactions were carried out under N2 atmosphere. Conversion and Z/E ratios determined by analysis of H or 13C NMR spectra of unpurified
product mixtures ( 2%). Yields for purified products ( 5%). For Conditions A−C, see Scheme 9. Condition D: 1.0 equiv of monosubstituted
olefin, 2.0 mol % Ru-1d, 5.0 equiv of (Z)-butene, thf, 1 h; 5.0 equiv of (Z)-α,β-unsaturated carboxylic acid, 4.0 mol % Ru-1d, 100 Torr, thf, 1 h,
ambient pressure, 7 h; 4.0 mol % Ru-1d, 100 Torr, thf, 1 h, ambient pressure, 11 h. Experiments were run in duplicate or more. See the Supporting
Information for details.14
a
Scheme 12. Scope of the Method: Synthesis of (Z)-α,β-Unsaturated Weinreb Amides
a
1
Reactions were carried out under N2 atmosphere. Conversion and Z/E ratios determined by analysis of H or 13C NMR spectra of unpurified
product mixtures ( 2%). Yields for purified products ( 5%). For Conditions A−C, see Scheme 9. Experiments were run in duplicate or more. See
the Supporting Information for details.9
The utility of this aspect of the approach is highlighted by
stereoselective synthesis of enone 13 (Scheme 12), used
previously as an intermediate in a total synthesis of
dihydrocompactin9 (Scheme 1b). Thus, 12 was first prepared
from commercially available starting materials in two steps and
56% overall yield (96:4 er).18 The monosubstituted alkene was
G
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 13. Scope of the Method: Synthesis of (Z)-α,β-Unsaturated Secondary and Primary Amides
a
1
Reactions were carried out under N2 atmosphere. Conversion and Z/E ratios determined by analysis of H or 13C NMR spectra of unpurified
product mixtures ( 2%). Yields for purified products ( 5%). For Conditions A and B, see Scheme 9. Experiments were run in duplicate or more.
See the Supporting Information for details. Bn, benzyl; PMB, p-methoxybenzyl. Burgess reagent, 1-methoxy-N-triethylammoniosulfonylmetha-
nimidate.
then transformed to amide 5j in 73% yield and >98:2 Z/E
selectivity. Ensuing protection of the two hydroxy groups and
alkylation generated (Z)-α,β-unsaturated ketone 13 in 62%
overall yield (for two steps) without any loss of stereoisomeric
purity (Scheme 12). This five-step route, which delivers 13 in
25% overall yield, is superior to the formerly reported 13-step
sequence that furnishes the same product in just 5% overall yield
(>98:2 Z/E, by partial ynone hydrogenation).
(Z)-α,β-Unsaturated Secondary and Primary Amides. We
prepared a number of different (Z)-α,β-unsaturated secondary
(6b−i, Scheme 13a) or primary amides (7a−c, Scheme 13b);
this class of products were generated in similar yields with either
Ru-1c or Ru-1d. On the basis of the mechanistic principles that
were already discussed, with a less electron-withdrawing
carbonyl-containing substituent, there is less stabilization of
electron density at Cα in an mcbnonproductive (see Scheme 3), and
competitive decomposition is diminished. It is worth noting that
a primary amide may be easily converted to the corresponding
(Z)-unsaturated alkenyl nitrile by treatment with commercially
available Burgess reagent for just 15 min at room temperature
(see 14, Scheme 13b).39,40
As the final note, we have been unable to obtain appreciable
conversion to any of the desired aryl- or heteroaryl-substituted
(Z)-α,β-unsaturated carbonyl products (<5% conversion to
cross-metathesis product). While this issue remains to be
resolved, this latter set of products, unlike those presented
above, can be readily accessed through highly efficient catalytic
cross-coupling41 between a readily accessible and reasonably
robust aryl or heteroaryl boronate, and (Z)-3-iodo-2-propenoic
acid or related derivatives, many of which can also be purchased.
CONCLUSIONS
■
The findings described above offer several new insights
regarding the inner workings of Ru catechothiolate catalysts,
attributes that are distinct from those pertaining to the more
classical Ru dichloro complexes. Perhaps the most important is
that the higher efficiency of reactions with electron-deficient
alkenes when unsaturated complex Ru-1d is used (compared to
Ru-1c) is rooted in not only higher activity of the complex
(more sterically accessible olefin coordination site and weaker
trans influence in the intermediate metallacyclobutane) but also
the fact that the derived nonproductive metallacyclobutane
intermediates decompose less readily (e.g., by β-hydride
elimination). Since a relatively strong trans influence is needed
for triggering the structural distortion that serves as the
preamble to β-hydride elimination, metallacycobutanes arising
from the complex with an unsaturated NHC ligand (Ru-1d) are
less likely to decompose. In the case of the more reactive alkene
substrates, namely, those that do not bear a strongly electron-
withdrawing substituent, there is no extra stabilization of the
nonproductive metallacyclobutane, and as a consequence, Ru
catechothiolate complexes that contain a saturated or an
unsaturated NHC are comparably efficient. In contrast, olefin
metathesis reactions involving a more electron-deficient ester,
carboxylic acid, or Weinreb amide can involve electronically
stabilized metallacyclobutane intermediates, which are non-
productive and can sequester the active complex and be the
cause of catalyst decomposition. The new knowledge here will
be needed for the success of future efforts in catalyst design and
reaction method development.
This study expands the frontiers of olefin metathesis on other
equally important fronts, as it delivers a much-needed addition
to the repertoire of catalytic stereoselective olefin metathesis
H
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
reactions.1 For the first time, various (Z)-α,β-unsaturated esters,
acids, Weinreb amides, and primary and secondary amides can
be prepared directly by kinetically controlled cross-metathesis
reaction, an advance that is likely to have an immediate impact
on chemical synthesis.
(5) Khan, R. K. M.; Torker, S.; Hoveyda, A. H. Reactivity and
selectivity differences between catecholate and catechothiolate Ru
complexes. Implications regarding design of stereoselective olefin
metathesis catalysts. J. Am. Chem. Soc. 2014, 136, 14337−14340.
(6) Ahmed, T. S.; Grubbs, R. H. Fast-initiating, ruthenium-based
catalysts for improved activity in highly E-selective cross metathesis. J.
Am. Chem. Soc. 2017, 139, 1532−1537.
ASSOCIATED CONTENT
* Supporting Information
(7) Keffer, J. L.; Plaza, A.; Bewley, C. A. Motualevic acids A−F,
antimicrobial acids from the sponge Siliquariaspongia sp. Org. Lett.
2009, 11, 1087−1090.
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.9b02318.
(8) Ueno, K.; Araki, Y.; Hirai, N.; Saito, S.; Mizutani, M.; Sakata, K.;
Todoroki, Y. Differences between the structural requirements for ABA
8’-hydroxylase inhibition and for ABA activity. Bioorg. Med. Chem.
2005, 13, 3359−3370.
Experimental details for all reactions and analytic details
for all products (PDF)
(9) Sammakia, T.; Johns, D. M.; Kim, G.; Berliner, M. A. Remote
asymmetric induction in an intramolecular ionic Diels−Alder reaction:
Application to the total synthesis of (+)-dihydrocompactin. J. Am.
Chem. Soc. 2005, 127, 6504−6505.
Crystallographic information files for Ru-1c and Ru-1d
(CIF, CIF)
(10) Ando, K. Preparation of (Z)-α,β-unsaturated amides by using
Horner-Wadsworth-Emmons reagents, (diphenylphosphono)-
acetamides. Synlett 2001, 2001, 1272−1274.
AUTHOR INFORMATION
Corresponding Authors
*E-mail: Torker@unistra.fr (S.T.).
*E-mail: amir.hoveyda@bc.edu or ahoveyda@unistra.fr
(A.H.H.)
■
(11) Fortin, S.; Dupont, F.; Deslongchamps, P. A new bis(2,2,2-
trifluoroethyl)phosphonate for the synthesis of Z-unsaturated N-
methoxy-N-methylamides. J. Org. Chem. 2002, 67, 5437−5439.
(12) Kojima, S.; Kawaguchi, K.; Matsukawa, S.; Uchida, K.; Akiba, K.
Highly Z-selective synthesis of disubstituted α,β-unsaturated cyanides
and amides using 10-P-5 Wittig-type reagents. Chem. Lett. 2002, 31,
170−171.
(13) Wang, Y.; Janjic, J.; Kozmin, S. A. Synthesis of leucascandrolide A
via a spontaneous macrolactolization. J. Am. Chem. Soc. 2002, 124,
13670−13671.
(14) Chatterjee, S.; Ghadigaonkar, S.; Sur, P.; Sharma, A.;
Chattopadhyay, S. A chemoenzymatic synthesis of hept-6-ene-2,5-
diol stereomers: Application to asymmetric synthesis of decarestrictine
L, pyrenophorol, and stagonolide E. J. Org. Chem. 2014, 79, 8067−
8076.
(15) Yu, E.; Johnson, B. M.; Townsend, E. M.; Schrock, R. R.;
Hoveyda, A. H. Synthesis of linear (Z)-α,β-unsaturated esters by
catalytic cross-metathesis. The influence of acetonitrile. Angew. Chem.,
Int. Ed. 2016, 55, 13210−13214.
(16) (a) Chatterjee, A. K.; Morgan, J. P.; Scholl, M.; Grubbs, R. H.
Synthesis of functionalized olefins by cross and ring-closing metathesis.
J. Am. Chem. Soc. 2000, 122, 3783−3784. (b) Choi, T.-L.; Lee, C. W.;
Chatterjee, A. K.; Grubbs, R. H. Olefin metathesis involving ruthenium
enoic carbene complexes. J. Am. Chem. Soc. 2001, 123, 10417−10418.
(c) Forman, G. S.; Tooze, R. P. Improved cross-metathesis of acrylate
esters catalyzed by 2nd generation ruthenium carbene complexes. J.
Organomet. Chem. 2005, 690, 5863−5866.
ORCID
Sebastian Torker: 0000-0002-2892-4411
Amir H. Hoveyda: 0000-0002-1470-6456
Author Contributions
#Z.L. and C.X. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research was supported by the John LaMattina Funds and
Alfonso Martin Escudero Foundation in the form of a graduate
and a postdoctoral fellowship to C. X. and J.d.P. Additional
support was provided by the Vanderslice Endowment Funds.
We are grateful to F. Romiti and M. S. Mikus for helpful
suggestions.
REFERENCES
■
(1) (a) Hoveyda, A. H. Evolution of catalytic enantioselective olefin
metathesis: from ancillary transformation to purveyor of stereochemical
̈
identity. J. Org. Chem. 2014, 79, 4763−4792. (b) Furstner, A. Teaching
metathesis old tricks. Science 2013, 341, 1229713. (c) Hoveyda, A. H.;
Khan, R. K. M.; Torker, S.; Malcolmson, S. J. Catalyst-controlled
stereoselective olefin metathesis. In Handbook of Metathesis; Grubbs, R.
H., Wenzel, A. G., O’Leary, D. J., Khosravi, E., Eds.; Wiley−VCH:
Weinheim, 2014; pp 503−562. (d) Montgomery, T. P.; Ahmed, T. S.;
Grubbs, R. H. Stereoretentive olefin metathesis: an avenue to kinetic
selectivity. Angew. Chem., Int. Ed. 2017, 56, 11024−11036.
(2) (a) Xu, C.; Shen, X.; Hoveyda, A. H. In situ methylene capping: a
general strategy for efficient stereoretentive catalytic olefin metathesis.
The concept, methodological implications, and applications to
synthesis of biologically active compounds. J. Am. Chem. Soc. 2017,
139, 10919−10928. (b) Nguyen, T. T.; Koh, M. H.; Mann, T. J.;
Schrock, R. R.; Hoveyda, A. H. Synthesis of E-and Z-trisubstituted
alkenes by catalytic cross-metathesis. Nature 2017, 552, 347−354.
(3) Khan, R. K. M.; Torker, S.; Hoveyda, A. H. Readily accessible and
easily modifiable Ru-based catalysts for efficient and Z-selective ring-
opening metathesis polymerization and ring-opening/cross-metathesis.
J. Am. Chem. Soc. 2013, 135, 10258−10261.
(17) Xu, C.; Liu, Z.; Torker, S.; Shen, X.; Xu, D.; Hoveyda, A. H.
Synthesis of Z- or E-trisubstituted allylic alcohols and ethers by
kinetically controlled cross-metathesis with a Ru catechothiolate
complex. J. Am. Chem. Soc. 2017, 139, 15640−15643.
(18) See the Supporting Information for details.
(19) Torker, S.; Koh, M. J.; Khan, R. K. M.; Hoveyda, A. H. Regarding
a persisting puzzle in olefin metathesis with Ru complexes: what are
transformations of alkenes with a small substituent Z-selective?
Organometallics 2016, 35, 543−563.
(20) Solans-Monfort, X.; Coperet, C.; Eisenstein, O. Shutting down
secondary reaction pathways: the essential role of the pyrrolyl Ligand in
improving silica supported d0-ML4 alkene metathesis catalysts from
DFT calculations. J. Am. Chem. Soc. 2010, 132, 7750−7757.
̈
(21) Pucino, M.; Inoue, M.; Gordon, C. P.; Schowner, R.; Stohr, L.;
́ ́ ́
Sen, S.; Hegedus, C.; Robe, E.; Toth, F.; Buchmeiser, M. R.; Coperet, C.
̈
Promoting terminal olefin metathesis with a supported cationic
molybdenum imido alkylidene N-heterocyclic carbene catalyst.
Angew. Chem., Int. Ed. 2018, 57, 14566−14569.
(4) Koh, M. J.; Khan, R. K. M.; Torker, S.; Yu, M.; Mikus, M. S.;
Hoveyda, A. H. High-value alcohols and higher-oxidation-state
compounds by catalytic Z-selective cross-metathesis. Nature 2015,
517, 181−186.
(22) (a) Janse van Rensburg, W.; Steynberg, P. J.; Meyer, W. H.; Kirk,
M. M.; Forman, G. S. DFT prediction and experimental observation of
substrate-induced catalyst decomposition in ruthenium-catalyzed
I
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
olefin metathesis. J. Am. Chem. Soc. 2004, 126, 14332−14333.
(40) Although (Z)-alkenyl nitriles can be prepared by catalytic cross-
metathesis, Z:E ratios did not exceed 90:10, until recently. (a) Crowe,
W. E.; Goldberg, D. R. Acrylonitrile cross-metathesis: coaxing olefin
metathesis reactivity from a reluctant substrate. J. Am. Chem. Soc. 1995,
117, 5162−5163. (b) Randl, S.; Gessler, S.; Wakamatsu, H.; Blechert, S.
Highly selective cross-metathesis with acrylonitrile using a phosphine
free Ru-complex. Synlett 2001, 2001, 430−432. (c) Miao, X.; Dixneuf,
P. H.; Fischmeister, C.; Bruneau, C. A green route to nitrogen-
containing groups: the acrylonitrile cross-metathesis and applications
to plant oil derivatives. Green Chem. 2011, 13, 2258−2271. (d) Gawin,
R.; Tracz, A.; Chwalba, M.; Kozakiewicz, A.; Trzaskowski, B.;
Skowerski, K. Cyclic alkyl amino ruthenium complexes − efficient
catalysts for macrocyclization and acrylonitrile cross metathesis. ACS
Catal. 2017, 7, 5443−5449. (e) Ref 38.
(41) (a) Yamada, T.; Watanabe, T.; Beppu, T.; Fukuyama, N.; Torii,
K.; Uozumi, Y. Palladium membrane-installed microchannel devices for
instantaneous Suzuki-Miyaura cross-coupling. Chem. - Eur. J. 2010, 16,
11311−11319. (b) Chemler, S. R.; Trauner, D.; Danishefsky, S. J. The
B-alkyl Suzuki-Miyaura cross-coupling reaction: Development, mech-
anistic study, and applications in natural product synthesis. Angew.
Chem., Int. Ed. 2001, 40, 4544−4568.
̈
(b) Engel, J.; Smit, W.; Foscato, M.; Occhipinti, G.; Tornroos, K. W.;
Jensen, V. R. Loss and reformation of ruthenium alkylidene: connecting
olefin metathesis, catalyst deactivation, regeneration, and isomer-
ization. J. Am. Chem. Soc. 2017, 139, 16609−16619.
(23) Magill, A. M.; Cavell, K. J.; Yates, B. F. Basicity of nucleophilic
carbenes in aqueous and nonaqueous solvents−theoretical predictions.
J. Am. Chem. Soc. 2004, 126, 8717−8724.
(24) Back, O.; Henry-Ellinger, M.; Martin, C. D.; Martin, D.;
Bertrand, G. 31P NMR chemical shifts of carbene−phosphinidine
adducts as an indicator of the π-accepting properties of carbenes.
Angew. Chem., Int. Ed. 2013, 52, 2939−2943.
(25) Liske, A.; Verlinden, K.; Buhl, H.; Schaper, K.; Ganter, C.
Determining the π-acceptor properties of N-heterocyclic carbenes by
measuring the 77Se NMR chemical shifts of their selenium adducts.
Organometallics 2013, 32, 5269−5272.
(26) Nelson, D. J.; Nolan, S. P. Quantifying and understanding the
electronic properties of N-heterocyclic carbenes. Chem. Soc. Rev. 2013,
42, 6723−6753.
́
(27) Dumas, A.; Tarrieu, R.; Vives, T.; Roisnel, T.; Dorcet, V.; Basle,
O.; Mauduit, M. Versatile and highly Z-selective olefin metathesis
ruthenium catalyst based on a readily accessible N-heterocyclic
carbene. ACS Catal. 2018, 8, 3257−3262.
(28) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Synthesis an
activity of a new generation of ruthenium-based olefin metathesis
catalysts coordinated with 1,3-dimesityl-4,5-dihydroimidazol-2-ylidene
ligands. Org. Lett. 1999, 1, 953−956.
(29) Bielawski, C. W.; Grubbs, R. H. Highly efficient ring-opening
metathesis polymerization (ROMP) using new ruthenium catalysts
containing N-heterocyclic carbene ligands. Angew. Chem., Int. Ed. 2000,
39, 2903−2906.
(30) Lummiss, J. A. M.; Higman, C. S.; Fyson, D. L.; McDonald, R.;
Fogg, D. E. The divergent effects of strong NHC donation in catalysis.
Chem. Sci. 2015, 6, 6739−6746.
(31) Dorta, R.; Stevens, E. D.; Scott, N. M.; Costabile, C.; Cavallo, L.;
Hoff, C. D.; Nolan, S. P. Steric and electronic properties of N-
heterocyclic carbenes (NHC): a detailed study on their interaction with
Ni(CO)4. J. Am. Chem. Soc. 2005, 127, 2485−2495.
̈
(32) Leuthaußer, S.; Schwarz, D.; Plenio, H. Tuning the electronic
properties of N-heterocyclic carbenes. Chem. - Eur. J. 2007, 13, 7195−
7203.
(33) Sußner, M.; Plenio, H. π-Face donor properties of N-heterocyclic
̈
carbenes. Chem. Commun. 2005, 5417−5419.
(34) Hunter, C. A.; Lawson, K. R.; Perkins, J.; Urch, C. J. Aromatic
interactions. J. Chem. Soc., Perkin Trans. 2001, 2, 651−669.
(35) Jung, K.; Kim, K.; Sung, J.-C.; Ahmed, T. S.; Hong, S. H.; Grubbs,
R. H.; Choi, T.-L. Toward perfect regiocontrol for β-selective
copolymerization using a Ru-based olefin metathesis catalyst. Macro-
molecules 2018, 51, 4564−4571.
(36) Kojima, S.; Hidaka, T.; Yamakawa, A. Application of organo-
cerium reagents for the efficient conversion of Z-α,β-unsaturated
Weinreb amides to Z-α,β-unsaturated ketones. Chem. Lett. 2005, 34,
470−471.
(37) For examples regarding conversion of a Weinreb amide to an
aldehyde in a complex molecule setting, see: (a) Heckrodt, T. J.;
Mulzer, J. Total synthesis of Elisabethan A: intramolecular Diels−Alder
reaction under biomimetic conditions. J. Am. Chem. Soc. 2003, 125,
4680−4681. (b) Evans, D. A.; Nagorny, P.; McRae, K. J.; Reynolds, D.
J.; Sonntag, L.-S.; Vounatsos, F.; Xu, R. Enantioselective synthesis of
oasomycin A, part I: synthesis of the C1−C12 and C13−C28 subunits.
Angew. Chem., Int. Ed. 2007, 46, 537−540.
(38) For a recent cross-metathesis approach for stereoselective
synthesis of alkenyl nitriles with Mo-based complexes, see Mu, Y.;
Nguyen, T. T.; Koh, M. J.; Schrock, R. R.; Hoveyda, A. H. E- and Z-, di-
and tri-substituted alkenyl nitriles through catalytic cross-metathesis-
Nat. Chem. 2019 DOI: 10.1038/s41557-019-0233-x.
(39) Claremon, D. A.; Phillips, B. T. An efficient chemoselective
synthesis of nitriles from primary amides. Tetrahedron Lett. 1988, 29,
2155−2158.
J
DOI: 10.1021/jacs.9b02318
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX