Aldehydes as alkyl carbanion equivalents for additions to carbonyl compounds

Article abstract of DOI:10.1038/nchem.2677

Nucleophilic addition reactions of organometallic reagents to carbonyl compounds for carbon-carbon bond construction have played a pivotal role in modern chemistry. However, this reaction's reliance on petroleum-derived chemical feedstocks and a stoichiometric quantity of metal have prompted the development of many carbanion equivalents and catalytic metal alternatives. Here, we show that naturally occurring carbonyls can be used as latent alkyl carbanion equivalents for additions to carbonyl compounds, via reductive polarity reversal. Such 'umpolung' reactivity is facilitated by a ruthenium catalyst and diphosphine ligand under mild conditions, delivering synthetically valuable secondary and tertiary alcohols in up to 98% yield. The unique chemoselectivity exhibited by carbonyl-derived carbanion equivalents is demonstrated by their tolerance to protic reaction media and good functional group compatibility. Enantioenriched tertiary alcohols can also be accessed with the aid of chiral ligands, albeit with moderate stereocontrol. Such carbonyl-derived carbanion equivalents are anticipated to find broad utility in chemical bond formation.

Full text of DOI:10.1038/nchem.2677

ARTICLES
PUBLISHED ONLINE: 5 DECEMBER 2016 | DOI: 10.1038/NCHEM.2677
Aldehydes as alkyl carbanion equivalents for
additions to carbonyl compounds
Haining Wang†, Xi-Jie Dai† and Chao-Jun Li
*
Nucleophilic addition reactions of organometallic reagents to carbonyl compounds for carbon–carbon bond construction
have played a pivotal role in modern chemistry. However, this reaction’s reliance on petroleum-derived chemical
feedstocks and a stoichiometric quantity of metal have prompted the development of many carbanion equivalents and
catalytic metal alternatives. Here, we show that naturally occurring carbonyls can be used as latent alkyl carbanion
equivalents for additions to carbonyl compounds, via reductive polarity reversal. Such ‘umpolung’ reactivity is facilitated by
a ruthenium catalyst and diphosphine ligand under mild conditions, delivering synthetically valuable secondary and
tertiary alcohols in up to 98% yield. The unique chemoselectivity exhibited by carbonyl-derived carbanion equivalents is
demonstrated by their tolerance to protic reaction media and good functional group compatibility. Enantioenriched tertiary
alcohols can also be accessed with the aid of chiral ligands, albeit with moderate stereocontrol. Such carbonyl-derived
carbanion equivalents are anticipated to find broad utility in chemical bond formation.
he nucleophilic addition of organometallic reagents to carbo- transition metals15,16. Hoveyda and colleagues have successfully
nyl compounds, to form new carbon–carbon bonds, is a funda- developed copper-catalysed borylative enantioselective additions
T
mental process in contemporary organic synthesis1–3. This simple to carbonyl compounds using olefin-derived nucleophiles17,18
.
alkylation process, complementary to the reduction of carbonyl com- Montgomery, Jamison and co-workers have designed nickel-based
pounds, provides a reliable method for generating a wide array of catalysts for stereoselective aldehyde additions, in which alkynes are
alcohol products. These alcohols are frequently encountered as key employed as carbanion equivalents19–21. To synthesize more sterically
building blocks in the synthesis of complex pharmaceutical drugs and encumbered tertiary alcohols, Buchwald, Liu and colleagues have
biologically active molecules. The discovery of Grignard reagents as car- devised enantioenriched alkyl copper intermediates, synthesized
banion equivalents4 and their subsequent additions to carbonyl com- from olefins, for additions to ketones22. The catalytic generation of
pounds marked a milestone in synthetic chemistry, enabling facile carbanion equivalents from either alkenes or alkynes, elegantly exem-
access to a diverse range of alcohols using preformed organomagnesium plified in these reports, has successfully addressed some of the long-
reagents with high generality, reactivity and easy manipulation5–8. Since standing challenges facing organometallic reagents. Nevertheless, as
then, other organometallic reagents9, such as those based on zinc10, alu- the chemical industry shifts from using petrochemicals to renewable
minium11, copper12 and titanium13, have been sought and used to feedstocks, the synthetic community is increasingly giving attention
achieve better selectivity. However, the preparation of these robust orga- to more sustainable and efficient chemical syntheses23,24
nometallic reagents requires stoichiometric quantities of metal (Fig. 1a). In this context, the development of carbanion equivalents that
.
Despite considerable advances, and the abundance of organo- originate from naturally occurring chemical feedstocks, require
metallic reagents developed for additions to carbonyl compounds, only a catalytic quantity of metal, have improved compatibility
three key challenges have endured. First, the dependence on towards benign protic solvents and various functional groups,
stoichiometric, pre-formed organometallic reagents in carbonyl and generate innocuous by-products would be highly desirable
addition reactions produces copious metal waste. This is particularly for additions to carbonyl compounds. Here, we report such alkyl
problematic for large-scale synthesis, as it complicates synthetic carbanion equivalents, derived from the naturally prevalent carbo-
operations and raises environmental concerns. In addition, nyls with umpolung reactivity25–29, for carbonyl addition reactions
petroleum-derived chemical feedstocks (that is, organic halides) are (Fig. 1c). Very recently, we have pioneered a ruthenium-catalysed
typically used to prepare organometallic reagents. Their paucity in redox system30 for direct primary alcohol deoxygenation31.
nature14 constrains the types of nucleophiles accessible to perform This practical deoxygenation chemistry evolved from our initial
carbonyl addition reactions without prior functionalization. iridium-based system32 and proved to be highly chemo- and regio-
Furthermore, the high nucleophilicity and basicity of most organo- selective in complex molecules such as alkaloids and steroids. The
metallic reagents generally result in poor selectivity, making these proposed mechanism involves the in situ generation of a ruthe-
reagents inferior candidates in late-stage chemical transformations nium-coordinated hydrazone intermediate A, followed by a ruthe-
where highly functionalized molecules are present.
nium-assisted Wolff–Kishner (WK) reduction under relatively
To address these challenges, much effort has been devoted to low-temperature conditions (Fig. 2a). Intriguingly, when benzylic
developing original catalytic and asymmetric methods to produce alcohols were subjected to the same catalytic reaction conditions,
enantioenriched alcohols, whereby π-unsaturated hydrocarbons a trace amount of reductive C−C coupling product—the carbonyl
(alkenes or alkynes) are masked as carbanion equivalents addition product—was observed. On the basis of this serendipitous
(Fig. 1b). Krische and co-workers have pioneered stereoselective discovery, we hypothesized that the coordinately unsaturated ruthe-
coupling reactions between diverse π-unsaturated reactants and nium complex in A might rapidly metallate another carbonyl com-
aldehydes under hydrogenative conditions catalysed by late pound and subsequently rearrange to give intermediate C, via
Department of Chemistry and FQRNT Centre for Green Chemistry and Catalysis, McGill University, 801 Sherbrooke Street West, Montreal, Quebec H3A
0B8, Canada. †These authors contributed equally to this work. e-mail: cj.li@mcgill.ca
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2677
ARTICLES
considered whether naturally occurring carbonyl compounds, via
umpolung chemistry, could serve as sustainable alkyl carbanion
equivalents in carbonyl addition reactions.
a
O
[M]
R3 OH
R4
R2
X
R3 R4
[M]
R1
R1 R2
Metallation
R1 R2
Stoichiometric
Results and discussion
We commenced our studies by evaluating benzaldehyde 1a in the
ruthenium catalytic system developed for our deoxygenation
chemistry31. We were delighted to observe the deoxygenative
homo-coupling product in 39% yield at 120 °C after 4 h (Fig. 2b).
The cross-coupling scenario was further examined between benz-
aldehyde 1a (as a nucleophilic carbonyl partner) and acetophenone
2 (as an electrophilic carbonyl partner). Subsequent screenings
suggested that the reaction was both thermodynamically and kine-
tically favoured, as it was completed within 3 h at temperatures
as low as 45 °C. Next, a revised procedure was developed by
replacing in situ generated hydrazone with preformed hydrazone
derived from 1a. This effort turned out to be beneficial to the
overall reaction as attenuated yields were seen for both deoxygenated
(from 1a) and cross-azine by-product (between hydrazone and 2),
providing alcohol product 3a in much higher yield. We found that
the success of this reaction depended on three critical factors:
ligand, base and additive (Table 1). Among all the catalysts and
ligands tested, our previous catalytic system31, consisting of [Ru
(p-cymene)Cl2]2 and 1,2-bis(dimethylphosphino)ethane (dmpe),
showed the highest reactivity (entries 1–4 versus 9). Base is indispen-
sable for the catalytic cycle. Our attempt without base only produced
trace amounts of product (entry 5). However, strong basicity is not
necessary to drive the reaction towards completion. In most cases, mod-
erate or even weak bases delivered satisfactory yields (entries 6–8),
among which potassium phosphate (K3PO4) proved ideal (entry 9).
Equally important is the fact that caesium fluoride (CsF) serves as an
efficient additive to enhance catalysis (entry 9 versus 10), similar to
early reports on palladium-catalysed cross-coupling reactions34,35. It
should be noted that this enhancement was only observed when
both caesium and fluoride ions were present. Starting from a catalytic
amount, we discovered that more CsF generally led to faster reactions,
with substoichiometric quantities being optimal.
R1, R2, R3, R4 = alkyl, aryl, or H
X = Cl, Br, I, etc.
[M] = Mg, Li, Cu, Al, Zn, Ti, etc.
b
O
H
R2
R2
R3 OH
cat. L*[M]–H
*
R3 R4
R1
*
R1
[M]L*
R4
R1
*
Metal-hydride
addition
R2
R1, R2, R3, R4 = alkyl, aryl, or H
[M] = Ir, Rh, Ru, Cu, Ni, etc.
c
NH2
N
R3 OH
cat. [M]L*
O
O
N2H4
This work
R1
R1 R2
R4
*
R1 R2
R2
Hydrazone
formation
R3 R4
R1, R2, R3, R4 = alkyl, aryl, or H
[M] = Ru
Figure 1 | Synthetic strategies to access secondary and tertiary alcohols by
carbonyl addition reactions. a, Traditional alkylation processes rely on
organometallic reagents (that is, Grignard reagents and organolithium
reagents). They are typically prepared from the metallation of petroleum-
derived organic halides in the presence of stoichiometric quantities of metal.
b, π-unsaturated hydrocarbons have recently been advanced as carbanion
equivalents for additions to carbonyl compounds, facilitated by catalytic
quantities of metal and chiral ligands. These elegant methods provide
enantioenriched alcohol products in excellent stereoselectivity. c, Carbonyls
are masked as latent alkylating reagents via hydrazone formation for
carbonyl addition reactions, assisted by catalytic quantities of metal (this
work). This new alkylation approach can be enantioselective when engaging
chiral ligands. Cat., catalytic; L*, chiral ligands. *, stereogenic centres.
With the optimized conditions in hand, we sought to explore the
scope of nucleophilic carbonyl partners in this new umpolung chem-
istry (Table 2). In general, both electron-deficient and electron-rich
aromatic aldehydes (3a–k) performed well under standard reaction
conditions, with a variety of functional groups remaining untouched
(trifluoromethyls, nitriles, ethers, halogens and amines). Of particular
note is the nitrile group, which is known to inhibit the formation of
Grignard reagents36. Importantly, with encouraging chemoselectivity
Zimmerman–Traxler chair-like transition state B (ref. 33). As a con-
sequence, formation of the reductive C−C coupling product might
be kinetically favoured over the WK-type reduction product upon
N2 extrusion (Fig. 2a). Aligned with this hypothesis, we further
a
O
R3 OH
Cl
Cl
L
L
L
3H
N
R3
H
O
R1
R
P
Cl
Cl
R3 R4
N
Ru
Rearrangement
R4
N
Ru
Ru
P
P
N
P
R2
R1
R2
Addition product
R1
P
P
O
N
R4
R4
R1
H2N
R2
R2
C
A
B
Zimmerman-Traxler TS
[Ru(p-cymene)Cl2]2 (3 mol%)
dmpe (24 mol%), N2H4 H2O (1.0 equiv.)
b
.
O
OH
Ph
39% yield
H
H
Ph
WK-type reduction
KOt-Bu (50 mol%), DMSO (30 μl)
t-BuOH, 120 °C, 4 h
R1 R2
Ph
H
Benzaldehyde (1a)
Reduction product
(0.2 mmol)
Figure 2 | A hypothesis is developed from a serendipitous discovery. a, Mechanistic hypothesis to capture the key ruthenium complex A proposed in our
deoxygenation chemistry31 for C−C formation. In light of the deoxygenated product obtained from WK-type reduction, we questioned whether A could be
intercepted by carbonyl compounds to form new C−C bonds. This alternative pathway would therefore afford the addition product alcohols, via
Zimmerman–Traxler chair-like transition state B. b, Aligned with this hypothesis, benzaldehyde (1a) was initially tested as substrates in the catalytic system
developed for our deoxygenation chemistry. The formation of the corresponding secondary alcohol in 39% yield proves the feasibility of our rationale.
WK, Wolff–Kishner; TS, transition state; dmpe, 1,2-bis(dimethylphosphino)ethane.
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2677
ARTICLES
at the α position were also reactive in the reaction, affording reasonable
quantities of alkylation products (3o, 3p). In particular, the good com-
patibility shown by the 2-pyridyl group in this chemistry is intriguing,
as poor reactivity of the ruthenium catalyst might have been expected
due to its strong chelate effect.
Table1 | A model study on nucleophilic addition of
benzaldehyde to acetophenone.
Acetophenone (2)
[Ru(p-cymene)Cl2]2
NH2
O
.
Me OH
N2H4 H2O
N
Ligand
Ph
Having studied the scope of nucleophilic carbonyl partners, we
turned our attention to electrophilic partners (Table 3, aromatic
and aliphatic ketones, aromatic and aliphatic aldehydes). The general-
ity of this umpolung chemistry was validated by excellent reactivity
seen across a broad spectrum of aromatic and aliphatic carbonyl com-
pounds and by a wide range of functionalities tolerated under reaction
conditions. Specifically, good to excellent yields were obtained for
aromatic ketones (5a–k, 5x, 7a, 7b, 65–93% yield); moderate to excel-
lent yields for aliphatic ketones (5l–t, 5y, 50–88% yield) and moderate
yields for both aromatic and aliphatic aldehydes (5u–w, 48–61%
yield, major by-products are alkenes). Relatively complex substrates
were also accommodated in this transformation, as exemplified by
alkylation of an oxo-steroid compound in good yield (5t, 68%
yield). The reduction in reactivity due to steric bias, previously
observed for nucleophilic partners (1a vs. 1l), was similarly observed
with electrophilic partners (5o versus 5r). It is important to note that
tetrahydrofuran (THF) is replaced by tert-butanol (t-BuOH) as the
reaction solvent when aldehydes are used as electrophilic partners
THF, r.t., 30 min
H
Base, additive
THF, 45 °C, 3 h
Ph
Ph
Ph
H
3a
1a
Entry
Ligand
Base
Additive
3a (%)
1
–
K3PO4
K3PO4
K3PO4
K3PO4
–
K2CO3
Cs2CO3
KOtBu
K3PO4
K3PO4
CsF
CsF
CsF
CsF
CsF
CsF
CsF
CsF
CsF
–
13
2
3
4
5
6
7
8
9
10
dppe
dppp
dppf
dmpe
dmpe
dmpe
dmpe
dmpe
dmpe
78
92
58
3
57
51
82
95
85
Reaction conditions: 1a (25 µl, 0.24 mmol, 1.2 equiv.), N2H4·H2O (13 µl, 0.26 mmol, 1.3 equiv.), THF
(0.1 ml), room temperature (r.t.), 30 min; 2 (23.5 µl, 0.20 mmol, 1.0 equiv.), [Ru(p-cymene)Cl2]2
(0.9 mg, 0.0015 mmol, 0.75 mol%), ligand (0.003 mmol, 1.5 mol%), base (0.05 mmol, 25 mol
%), CsF (15 mg, 0.10 mmol, 50 mol%), 45 °C, 3 h, under N2. Yields were determined by crude 1
NMR using mesitylene as an internal standard.
H
displayed by the nucleophilic carbonyl partners, this chemistry enables (5u−w). This result clearly indicates that the carbonyl-derived carba-
complementary access to highly functionalized Grignard reagents—a nion equivalent is compatible with alcoholic solvents. In line with this
rapidly advancing field over the past decade8,37. In addition to good observation, we questioned whether this carbanion equivalent might
functional-group compatibility, we identified a minor influence on have better chemoselectivity than traditional organometallic reagents.
reaction efficiency with respect to differing arene substituent patterns. Indeed, unique chemoselectivity was demonstrated by tolerance to
Compared to para and meta positions, the chloro substituent at the commonly reactive functional groups. Under the standard conditions
ortho position was the least reactive, probably due to the slightly described, moderate to excellent yields resulted from aromatic
increased steric hindrance (3e–g). In addition, we found that elevated ketones featuring an ester, unprotected tert-alcohol, Weinreb amide
temperatures were essential for aromatic ketones to undergo the trans- and unprotected sec-amide (5x, 5y, 7a, 7b, 50–88% yield). Given
formation, albeit with attenuated yield (3l). We reasoned that this the distinct chemoselectivity displayed by the carbonyl-derived carba-
reduced reactivity was more likely caused by a steric, rather than an nion equivalents, orthogonal reactivity could be readily achieved
electronic difference between aromatic aldehyde 1a and ketone 1l. through further diversification on residual functional groups (that
Aliphatic aldehydes can be used as nucleophilic carbonyl partners in is, ester and Weinreb amide). Moreover, survival of the unprotected
our current system, although they showed low reactivity towards the sec-amide implies potential applications of this carbanion equivalent
formation of alcohols (3m, 3n). Instead, more elimination and in peptide chemistry.
cross-azine products were predominant under such circumstances.
To access more synthetically useful enantioenriched alcohols, we
Aldehydes with heteroaromatic substituents (that is, furyl, pyridyl) decided to probe the asymmetric version of this chemistry. When
Table 2 | Scope with respect to nucleophilic carbonyl partners.
2
[Ru(p-cymene)Cl2]2 (0.75 mol%)
Me OH
R1
NH2
R2
O
.
O
N2H4 H2O
N
dmpe (1.5 mol%)
Ph
R1 R2
1a−p
Me
Ph
R1
THF, r.t., 30 min
K3PO4 (25 mol%), CsF (50 mol%)
THF, 45 °C, 3 h
R2
2
3a−p
MeO
O
Cl
O
O
Me OH
Ph
Me OH
Ph
Me OH
Ph
Me OH
Me OH
Ph
Ph
3a, 94%
3b, 58%
3c, 98%
3d, 86%
3e, 85%
Me OH
Me OH
Ph
NC
F3C
Me OH
Ph
Me OH
Ph
Me OH
Ph
Ph
Cl
Cl
3f, 65%
3g, 94%
3h, 43%
3i, 97%
3j, 95%
Me
Me OH
Ph
Me OH
Ph
Me OH
Ph
Me OH
N
Me OH
Ph
Me
Me OH
Ph
O
Ph
N
Me
3m, 23%†
3n, 20%†
3k, 67%
3o, 75%
3p, 53%
3l, 30%*
Reaction conditions: 1a−1p (0.48 mmol), N2H4 · H2O (0.52 mmol), THF (0.2 ml), r.t., 30 min; 2 (0.40 mmol), [Ru(p-cymene)Cl2]2 (0.75 mol%), dmpe (1.5 mol%), K3PO4 (25 mol%), CsF (50 mol%), 45 °C,
3 h, under N2. Isolated yields were reported (average of two runs). *Reaction was conducted at 80 °C for 20 h. †Using KOtBu (25 mol%) instead of K3PO4; See Supplementary Section II,i for details.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2677
ARTICLES
Table 3 | Scope with respect to electrophilic carbonyl partners.
Method A:
4a−t,x,y;6a,b (1.0 equiv.)
[Ru] (0.75 mol%), dmpe (1.5 mol%)
K3PO4 (25 mol%), CsF (50 mol%)
THF, 45 °C, 3 h
Method A:
.
N2H4 H2O
NH2
R4 OH
R3
5a−y
7a,b
O
O
N
THF, r.t., 30 min
Ph
Ph
H
R3 R4
4a−y
6a,b
Method B:
4u−w (1.0 equiv.)
[Ru] (1.5 mol%), dmpe (12 mol%)
K3PO4 (25 mol%)
Method B:
N2H4 in THF (1.0 M)
r.t., 30 min
Ph
H
1a
DMSO, t-BuOH, 120 °C, 6 h
F3C OH
Me OH
Me OH
HO
Ph OH
Et OH
Ph
Ph
Ph
O
Ph
Ph
Ph
Br
Ph
Ph
5a, 89%*
5b, 80%*
5c, 76%*
5d, 65%*
5e, 93%*
5f: 4-Br, 86%*
5g: 2-Br, 71%*
Me OH
Me OH
Me OH
Ph
HO
Ph
Ph
Me OH
Ph
Ph
Ph
OH
OMe
F
I
5h, 93%*
Me OH
Ph
5i, 70%*
5j, 84%*
5k, 79%*
5l, 88%*
5m, 71%*
HO
Ph
Me OH
Et OH
Et
Me OH
Me OH
Ph
Ph
Ph
Ph
OMe
n-C5H11
5s, 75%*
Me
5n, 85%*
5o, 50%*
5p, 55%*
5q, 56%*
Me OH
5r, 82%*
OH
OH
Ph
OH
Ph
Et
Ph
Ph
n-C7H15
5w, 50%†
Ph
5u, 48%†
COOMe
5x, 88%*
Et
5v, 61%†
Me OH
H
Ph
OMe
HO Me
O
H
Ph OH
OH
H
Ph
N
Ph
Ph
Me
N
Ph
OH
H
H
O
5t, 68%*
5y, 50%*
7a, 77%*
7b, 85%*
*Reaction condition A for ketones: 1a (0.48 mmol), N2H4 · H2O (0.52 mmol), THF (0.2 ml), r.t., 30 min; 4a−t,x,y;6a,b (0.40 mmol), [Ru(p-cymene)Cl2]2 (0.75 mol%), dmpe (1.5 mol%),
K3PO4 (25 mol%), CsF (50 mol%), 45 °C, 3 h, under N2. Isolated yields are reported (average of two runs). See Supplementary Section II,i for details. †Reaction condition B for aldehydes: 1a (0.24 mmol),
anhydrous N2H4 in THF (1.0 M, 0.3 mmol), r.t., 30 min; 4u−w (0.2 mmol), [Ru(p-cymene)Cl2]2 (1.5 mol%), dmpe (12 mol%), K3PO4 (25 mol%), DMSO (30 µl), t-BuOH (100 µl), 120 °C, 6 h, under N2.
Isolated yields are reported (average of two runs). See Supplementary Section II,ii for details.
the chiral ligands (S,S)-Ph-BPE and (S,S)-DPEN were combined with stone towards carbon–carbon bond-forming processes built upon
[Ru(COD)Cl2]n in our model study, (R)-3a was obtained in 52% yield more sustainable chemical feedstocks. With a deeper understanding
with 76:24 enantiomeric ratio (e.r.) (Fig. 3). This preliminary result of this reaction, new strategies using carbonyl-derived carbanion
suggest that chiral ligands are involved in the C−C formation step. equivalents for chemical bond formation beyond carbon–carbon
As a result, high levels of stereocontrol could be feasible with the bonds are likely to emerge.
appropriate chiral ligand framework. Taken together, our findings
show both remarkable chemoselectivity of the current chemistry
and high potential for stereoselectivity.
Methods
Full experimental details and the characterization of compounds are provided in the
Supplementary Information.
Conclusion
General procedure. A flame-dried V-shape microwave reaction vial (10 cm3)
equipped with a magnetic stir bar was charged with [Ru(p-cymene)Cl2]2 (1.8 mg,
0.003 mmol, 0.75 mol%) and K3PO4 (21.2 mg, 0.1 mmol, 25 mol%). The reaction
vial was transferred to the glovebox and charged with dmpe (1.0 µl, 0.006 mmol,
1.5 mol%) and CsF (30 mg, 0.20 mmol, 50 mol%) before being sealed with a rubber
septum. The reaction vial was moved out of the glovebox and sequentially
charged with ketones or aldehydes (0.4 mmol, 1.0 equiv.) and hydrazone solution
(250 µl). The reaction mixture was heated to 45 °C in an oil bath. Following stirring
for 3 h, the reaction mixture was filtered through a plug of silica gel with EtOAc
(2 ml) as eluent, concentrated, and purified by flash chromatography (hexane/ethyl
acetate 90:10 as eluent) to give the corresponding alcohols.
In summary, we have described a catalytic, ruthenium-based, umpo-
lung strategy to use aldehydes as alkyl carbanion equivalents for
additions to carbonyl compounds. The full potential of this novel
chemistry is yet to be revealed, as it is currently limited by accessible
nucleophilic carbonyls, and safety and toxicity issues concerning
hydrazine. Nevertheless, many applications based on the key
characteristics of these new carbonyl-derived carbanion equivalents
can be envisioned. Highlighted features of this chemistry are
naturally prevalent carbonyls as renewable chemical feedstocks, a cat-
alytic amount of both metal and ligand, benign stoichiometric by-
products, great functional-group tolerance with unique chemoselec-
tivity (ester, Weinreb amide, free amide and hydroxyl group) and
Notes. Use of the glovebox is not necessary for this chemistry. Given appropriate
manipulation and storage of sensitive chemical reagents (air-sensitive dmpe and
hygroscopic CsF), the reaction can be successfully performed outside the glovebox.
enantioselective control. We believe that this chemistry is a stepping Specifically, air-sensitive dmpe should be kept under inert gas protection (standard
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.2677
ARTICLES
17. Meng, F.-K., Haeffner, F. & Hoveyda, A. H. Diastereo- and enantioselective
reactions of bis(pinacolato)diboron, 1,3-enynes, and aldehydes catalyzed by an
easily accessible bisphosphine–Cu complex. J. Am. Chem. Soc. 136,
11304−11307 (2014).
18. Meng, F.-K., Jang, H., Jung, B. & Hoveyda, A. H. Cu-catalyzed chemoselective
preparation of 2-(pinacolato)boron-substituted allylcopper complexes and their
in situ site-, diastereo-, and enantioselective additions to aldehydes and ketones.
Angew. Chem. Int. Ed. 52, 5046−5051 (2013).
19. Chaulagain, M. R., Sormunen, G. J. & Montgomery, J. New N-heterocyclic
carbene ligand and its application in asymmetric nickel-catalyzed aldehyde/
alkyne reductive couplings. J. Am. Chem. Soc. 129, 9568−9569 (2007).
20. Jackson, E. P. et al. Mechanistic basis for regioselection and regiodivergence
in nickel-catalyzed reductive couplings. Acc. Chem. Res. 48, 1736−1745 (2015).
21. Miller, K. M., Huang, W.-S. & Jamison, T. F. Catalytic asymmetric reductive
coupling of alkynes and aldehydes: enantioselective synthesis of allylic alcohols
and α-hydroxy ketones. J. Am. Chem. Soc. 125, 3442−3443 (2003).
22. Yang, Y., Perry, I. B., Lu, G., Liu, P. & Buchwald, S. L. Copper-catalyzed
asymmetric addition of olefin-derived nucleophiles to ketones. Science.
353, 144−150 (2016).
2 (1.0 equiv.)
[Ru(COD)Cl2]n (1.5 mol%)
L1 (1.5 mol%)
L2 (1.5 mol%)
NH2
.
Me
OH
Ph
N2H4 H2O
N
O
Ph
K3PO4 (25 mol%)
THF, 50 oC, 4 h
Ph
H
Ph
H
(R)-3a
52% yield
76:24 e.r.
1a
Ph
H2N
Ph
NH2
Ph
Ph
O
P
P
Me
Ph
Ph
Ph
(S,S)-Ph-BPE
(S,S)-DPEN
2
L1
L2
Figure 3 | Preliminary results on enantioselective carbonyl addition. Our
attempt to carry out an enantioselective carbonyl addition was successful by
combining a chiral diphosphine ligand (L1) and diamine ligand (L2) with
ruthenium complex [Ru(COD)Cl2]n, providing enantioenriched tertiary
alcohol 3a in moderate yield and e.r. This result indicates that the
stereoselectivity of new C−C formation (step of the nucleophilic carbonyl
addition) is influenced by chiral ligands. The e.r. value was determined by
chiral high-performance liquid chromatography. Absolute configuration was
assigned as (R) based on the literature report38. See Supplementary
Sections II and III for details. COD, 1,5-cyclooctadiene; (S,S)-Ph-BPE, (+)-1,2-
bis((2S,5S)-2,5-diphenylphospholano)ethane; DPEN,
23. Anastas, P. T. & Warner, J. C. Green Chemistry: Theory and Practice
(Oxford Univ. Press, 2000).
24. Li, C.-J. & Trost, B. M. Green chemistry for chemical synthesis. Proc. Natl Acad.
Sci. USA 105, 13197−13202 (2008).
25. Breslow, R. On the mechanism of thiamine action. IV.1 evidence from studies on
model systems. J. Am. Chem. Soc. 80, 3719−3726 (1958).
26. Seebach, D. Methods of reactivity umpolung. Angew. Chem. Int. Ed. 18,
239−258 (1979).
27. Wu, Y., Hu, L., Li, Z. & Deng, L. Catalytic asymmetric umpolung reactions of
imines. Nature 523, 445−450 (2015).
28. Brehme, R., Enders, D., Fernandez, R. & Lassaletta, J. M. Aldehyde N,N-
dialkylhydrazones as neutral acyl anion equivalents: umpolung of the imine
reactivity. Eur. J. Org. Chem. 2007, 5629−5660 (2007).
diphenylethylenediamine.
29. Flanigan, D. M., Romanov-Michailidis, F., White, N. A. & Rovis, T.
Organocatalytic reactions enabled by N-heterocyclic carbenes. Chem. Rev. 115,
9307−9387 (2015).
Schlenk line procedure), and hygroscopic CsF should be properly stored in the
desiccator before use.
30. Bruneau, C. & Dixneuf, P. H. (eds) Ruthenium Catalysts and Fine Chemistry
(Springer, 2004).
31. Dai, X.-J. & Li, C.-J. En route to a practical primary alcohol deoxygenation.
J. Am. Chem. Soc. 138, 5433−5440 (2016).
32. Huang, J.-L., Dai, X.-J. & Li, C.-J. Iridium-catalyzed direct dehydroxylation of
alcohols. Eur. J. Org. Chem. 2013, 6496−6500 (2013).
33. Zimmerman, H. E. & Traxler, M. D. The stereochemistry of the Ivanov and
Reformatsky reactions. I. J. Am. Chem. Soc. 79, 1920−1923 (1957).
34. Wright, S. W., Hageman, D. L. & McClure, L. D. Fluoride-mediated boronic acid
coupling reactions. J. Org. Chem. 59, 6095−6097 (1994).
Hydrazone solution. A mixture of carbonyls (0.48 mmol, 1.2 equiv.) and hydrazine
monohydrate (26 µl, 0.52 mmol, 64–65 wt%, 1.3 equiv.) in THF (0.2 ml) solution
was stirred for 30 min at room temperature. Before injection of this hydrazone
solution A into the reaction mixture, a small amount of anhydrous Na2SO4 was
added to it.
All new compounds were fully characterized (Supplementary Sections III and VI).
Received 23 May 2016; accepted 18 October 2016;
published online 5 December 2016
35. Littke, A. F. & Fu, G. C. The first general method for Stille cross-couplings of aryl
chlorides. Angew. Chem. Int. Ed. 38, 2411−2413 (1999).
36. Burns, T. P. & Rieke, R. D. Highly reactive magnesium and its application to
organic syntheses. J. Org. Chem. 52, 3674−3680 (1987).
37. Krasovskiy, A. & Knochel, P. A LiCl-mediated Br/Mg exchange reaction for the
preparation of functionalized aryl- and heteroarylmagnesium compounds from
organic bromides. Angew. Chem. Int. Ed. 43, 3333−3336 (2004).
38. Kitanosono, T., Xu, P. & Kobayashi, S. Heterogeneous versus homogeneous
copper(II) catalysis in enantioselective conjugate-addition reactions of boron in
water. Chem. Asian. J. 9, 179−188 (2014).
References
1. Corey, E. J. & Cheng, X.-M. The Logic of Chemical Synthesis (Wiley, 1989).
2. Noyori, R. & Kitamura, M. Enantioselective addition of organometallic reagents
to carbonyl compounds: chirality transfer, multiplication, and amplification.
Angew. Chem. Int. Ed. 30, 49−69 (1991).
3. Kobayashi, S. et al. in Comprehensive Organometallic Chemistry III (eds
Crabtree, R. H. & Mingos, D. M. P.) 403−491 (Elsevier, 2007).
4. Stowell, J. C. Carbanion Equivalents in Organic Synthesis (Wiley, 1979).
5. Kharasch, M. S. & Reinmuth, O. Grignard Reactions of Nonmetallic Substances
(Prentice-Hall, 1954).
6. Wakefield, B. J. Organomagnesium Methods in Organic Chemistry
(Academic, 1995).
Acknowledgements
The authors acknowledge the Canada Research Chair Foundation (to C.J.L.), the CFI,
FQRNT Center for Green Chemistry and Catalysis, NSERC and McGill University for
financial support. The authors thank P. Querard and Z. Huang for their donation of
compound 6a and chiral (S,S)-DPEN ligand, respectively. X.J.D. thanks the chemistry
department for a Heather Munroe-Blum fellowship.
7. Silverman, G. S. & Rakita, P. E. Handbook of Grignard Reagents (CRC, 1996).
8. Knochel, P. et al. Highly functionalized organomagnesium reagents prepared
through halogen–metal exchange. Angew. Chem. Int. Ed. 42, 4302−4320 (2003).
9. Negishi, E.-i. Organometallics in Organic Synthesis (Wiley, 1980).
10. Pu, L. & Yu, H.-B. Catalytic asymmetric organozinc additions to carbonyl
compounds. Chem. Rev. 101, 757−824 (2001).
11. Ashby, E. C. & Laemmle, J. T. Stereochemistry of organometallic compound
addition to ketones. Chem. Rev. 75, 521−546 (1975).
12. Shibasaki, M. & Kanai, M. Asymmetric synthesis of tertiary alcohols and α-
tertiary amines via Cu-catalyzed C−C bond formation to ketones and ketimines.
Chem. Rev. 108, 2853−2873 (2008).
13. Duthaler, R. O. & Hafner, A. Chiral titanium complexes for enantioselective
addition of nucleophiles to carbonyl groups. Chem. Rev. 108, 807−832 (1992).
14. Alonso, F., Beletskaya, I. P. & Yus, M. Metal-mediated reductive
hydrodehalogenation of organic halides. Chem. Rev. 102, 4009−4092 (2002).
15. Jang, H.-Y. & Krische, M. J. Catalytic C−C bond formation via capture of
hydrogenation intermediates. Acc. Chem. Res. 37, 653−661 (2004).
16. Skucas, E., Ngai, M. Y., Komanduri, V. & Krische, M. J. Enantiomerically
enriched allylic alcohols and allylic amines via C−C bond-forming
Author contributions
H.W. and X.-J.D. are co-first authors responsible for this work, regardless of the listed
name order. H.W. discovered the reaction. X.-J.D. conceived the concept. X.-J.D.
and H.W. designed and performed the experiments, and analysed the data. X.-J.D.
and H.W. co-wrote the paper with feedback and guidance from C.-J.L. C.-J.L. directed
the project. All authors discussed the experimental results and commented on
the manuscript.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. CorrespondenceandrequestsformaterialsshouldbeaddressedtoC.J.L.
hydrogenation: asymmetric carbonyl and imine vinylation. Acc. Chem. Res. 40, Competing financial interests
1394−1401 (2007).
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
5
© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.