Oxidative umpolung ?±-alkylation of ketones

Article abstract of DOI:10.1021/ol503384c

We disclose a hypervalent iodine mediated ?±-alkylative umpolung reaction of carbonyl compounds with dialkylzinc as the alkyl source. The reaction is applicable to all common classes of ketones including 1,3-dicarbonyl compounds and regular ketones via their lithium enolates. The ?±-alkylated carbonyl products are formed in up to 93% yield. An ionic mechanism is inferred based on meticulous analysis, NMR studies, trapping and crossover experiments, and computational studies.

Full text of DOI:10.1021/ol503384c

Letter
pubs.acs.org/OrgLett
Oxidative Umpolung α‑Alkylation of Ketones
O. Svetlana Shneider,† Evgeni Pisarevsky,† Peter Fristrup,‡ and Alex M. Szpilman*,†
†Schulich Faculty of Chemistry, Technion-Israel Institute of Technology, 3200008, Haifa, Israel
‡Department of Chemistry, Technical University of Denmark, DK-2800, Kgs. Lyngby, Denmark
S
* Supporting Information
ABSTRACT: We disclose a hypervalent iodine mediated α-
alkylative umpolung reaction of carbonyl compounds with
dialkylzinc as the alkyl source. The reaction is applicable to all
common classes of ketones including 1,3-dicarbonyl compounds
and regular ketones via their lithium enolates. The α-alkylated
carbonyl products are formed in up to 93% yield. An ionic
mechanism is inferred based on meticulous analysis, NMR
studies, trapping and crossover experiments, and computational studies.
Scheme 1. Conceptual Differences between Classical and
Umpolung Alkylation (This Work)
mpolung reactions represent an invaluable tool for
1
Umolecular assembly, providing flexibility, chemoselectiv-
ity, and sometimes enhanced reactivity. An illustration of the
power of the umpolung concept is its application in total
synthesis.2 Because umpolung reactivity is anomalous to classical
chemistry by definition, the development of umpolung chemistry
entails exploration of unchartered chemical reactivity space and,
consequently, new discoveries in this area are rare. Indeed, the
inherent difficulty of achieving polarity reversal means that
umpolung is commonly accomplished by heteroatom sub-
stitution in multistep protocols, such as the classical thioacetal
umpolung of aldehydes.1,3,4
A powerful way of achieving direct umpolung in a single step is
through the use of hypervalent iodine reagents.5 However,
hypervalent iodine mediated C−C bond forming reactions are
limited to introduction of aryl (sp2),6,7 alkenyl (sp2),8 and alkynyl
(sp)9 at carbonyl α-positions.5 In contrast, only perfluorinated
alkyl (sp3) substituents can be introduced through this reactivity
manifold.10 This may be due to the instability of alkyl iodide(III)
compounds.5 All of these reactions use the carbonyl compound
in the classical sense as a nucleophile, and it is the alkenyl, alkynyl,
or aryl moiety that is polarity reversed by the hypervalent iodine
(Scheme 1). The reverse approach, i.e., umpolung of the
carbonyl α-position by hypervalent iodine reagents, has been
used successfully for the introduction of various heteroatoms
including halogen, oxygen, and nitrogen.5 The only example of
C−C bond formation is the iodosylbenzene/HBF4 mediated
allylation of silyl enolethers reported by Calpe and Zefirov11 and
enol dimerization.5 C−C bond formation using alkyl-metal
species has so far not been reported. Yet such a direct single step
oxidative alkylation of carbonyl compounds would significantly
expand the scope of umpolung strategies.4,12
We envisioned achieving umpolung of the ketone enolate into
an α-electrophile by the action of a two-electron hypervalent
iodine oxidant (PhIX2) and subsequent addition of an alkyl
nucleophile. Several challenges presented themselves for the
intended transformation. For example, the carbonyl compound
could undergo oxidative dimerization13,5d or α-functionalization
by attack of other reactive nucleophiles present or formed in the
course of the reaction.5 In addition, the “normal” 1,2-addition of
alkyl-metal reagents to the carbonyl group prior to the oxidative
step had to be precluded. These considerations led us to focus on
trialkyl aluminum and dialkyl zinc reagents.
We chose β-keto ester 1 (Table 1) as our model substrate since
it has a high enol content in toluene and a low tendency to
undergo α-heteroatom substitution and was of synthetic interest,
as we would be forging a difficult to form quaternary carbon
center in the reaction. Experiments with oxidants 3−8 (Figure 1)
led to recovery of unreacted β-keto-ester 1 (see Table S1, entries
1−6). In contrast, reaction of the more reactive 10 with Koser’s
In this paper, we disclose the oxidative umpolung α-alkylation
of carbonyl compounds by Koser’s reagent and dialkylzinc
reagents (Scheme 1). We also report preliminary mechanistic
studies that show that the reaction takes place via hypervalent
iodine mediated oxidative umpolung of the ketone.
Received: November 21, 2014
© XXXX American Chemical Society
A
DOI: 10.1021/ol503384c
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
a
Table 1. Optimization of the Reaction
Scheme 2. Scope of Reaction for 1,3 Bis Carbonyl Substrates
(a) and Simple Ketones (b)
a
b
2 equiv of organometal and oxidant. 1 equiv.
Importantly, the lithium enolate, prepared from the TMS
enolether 18 of tetralone by the action of MeLi, reacted with
Koser’s reagent 9 and dimethyl or diethylzinc to form alkylated
ketones 19a and 19b in 68% and 70% yield, respectively (Scheme
2b). The lithium enolate could also be prepared by the reaction
of tetralone using LDA. Subsequent reaction of the enolate with
Koser’s reagent and diethylzinc afforded 19b in 50% yield. For
comparison, alkylation of tetralone using LDA and ethyl
bromide, methyl iodide, or ethyl iodide proceeds in 41−65%
yield.20
These results show that the reaction has the potential to be a
strategically important umpolung alternative to classical
alkylation chemistry for a broad range of carbonyl compounds.
Significantly, while dialkylation is a common problem under
Figure 1. Oxidants screened during optimization of the reaction.
reagent (9) led to formation of a tosylated product in 80% yield
(see Supporting Information (SI) and Scheme 2a for the
structure of 10).14 It is well-known that Koser’s reagent tosylates
β-keto esters rapidly at ambient temperature.15 We found that
the tosylation reaction could be effectively suppressed when the
reaction was performed below 0 °C. When the reaction of 1 with
trimethylaluminum (2 equiv) and Koser’s reagent (9) (2 equiv)
was carried out at −78 °C in toluene, product 2a (R = Me) was
isolated in 18% yield along with 82% recovered starting material
(Table 1, entry 1). Coordinating solvents inhibit the reaction. By
screening numerous additives, we found that, in the presence of
the desiccant magnesium sulfate, quantitative consumption of
the starting material occurred, affording 2a (R = Me) in 49%
isolated yield (entry 2). No other products were observed by
TLC. The use of molecular sieves led to formation of 2a in 47%
isolated yield (entry 3). Switching to dimethylzinc increased the
isolated yield of the alkylated product 2a (R = Me) to 80% (entry
4). The order of addition reagents is inconsequential for the yield
and progress of the reaction. The reaction of 1 with diethylzinc
afforded the product 2b (R = Et) in 28% yield (entry 5). It is well-
known that the reactivity of dialkylzinc can be increased by the
addition of amino alcohols.16 When ethanolamine (1 equiv) was
added to the reaction, the ethylated product 2b was isolated in
50% yield (entry 6). Other aminoalcohols proved inferior to
ethanolamine. For comparison 2a and 2b have been prepared in
97% and 55% yield respectively using the classical conditions of
base, alkyl iodide and heat.17
To demonstrate the potential of the umpolung alkylation
reaction, we applied it to a broad range of ketones (Scheme 2).
For example, β-keto ester 10 reacted under the standard
conditions to give only the monoalkylated product 11a in 92%
yield (Scheme 2a). Ethylation in the presence of ethanolamine
proceeded to give 11b in 86% yield. Reaction of β-keto amide
(12) led to formation of 13 in 88% yield (Scheme 2a). Both alkyl
substituted 1,3-diketone 14 and aryl substituted diketone 16
(Scheme 2a) gave the alkylated products in 66% (15), 76%
(17a), and 65% (17b) yields, respectively. For comparison 10
has been ethylated under classical conditions using NaH and
ethyl bromide in DMF at 60 °C to afford 11b in 45% yield.18
Synthesis of 17b from 16 using excess ethyl iodide and TBAF at
rt overnight proceeds in 41% yield.19
a
Table 2. Study of Functional Group Tolerance
entry product proc.
X1
X2
R3
R4
yield
1
30a
30b
30c
31a
31b
32
A
B
B
A
B
A
A
B
A
B
A
B
A
A
B
A
A
B
Me
Me
Me
F
H
Et
Me
Et
87%
85%
54%
82%
85%
80%
92%
33%
90%
61%
92%
93%
93%
82%
74%
80%
93%
27%
2
H
Et
3
H
Et
Bu
Me
Et
4
H
Et
5
F
H
Et
6
Cl
H
Et
Me
Me
Et
7
33a
33b
34a
34b
35a
35b
36
NO2
NO2
OMe
OMe
H
H
Et
8
H
Et
9
OMe
OMe
H
Et
Me
Et
10
11
12
13
14
15
16
18
19
Et
allyl
allyl
Me
Et
H
H
H
H
propargyl
Me
Me
Et
37a
37b
38
−
−
−
−
−
−
−
−
Et
Et
cholesteryl
Me
Me
Et
39a
39b
Et
Et
−
−
a
Procedure A: 9 (2 equiv), Me2Zn (2 equiv), MgSO4, toluene −78 °C.
Procedure B: same as A but with 9 (4 equiv), Et2Zn (4 equiv).
B
DOI: 10.1021/ol503384c
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
classical conditions, in all of these experiments (Scheme 2), only
monoalkylated products were observed.
Scheme 3. Mechanisms a−e
We evaluated the functional group tolerance of the reaction for
a number of functionalized β-keto esters (Table 2). The reaction
is compatible with a wide range of alkyl-substituted and aryl-
substituted substrates. Functionalities such as fluoro, chloro,
nitro, and methoxy aryl substituents are tolerated (products 31,
32, 33, and 34, entries 4−10). The reaction is compatible with
alkene and alkyne functionalities as shown by the formation of
products 35, 36, and 38 (Table 2, entries 11−13 and 16).
To provide insight into this novel transformation, we
embarked on a mechanistic study. A careful product analysis
showed that in addition to the desired product, and iodobenzene
40, alkyl tosylates e.g. 41 were also formed (eq 1). Additionally,
d, and e only by the source of the transferred alkyl group. As
shown in Scheme 3, 10 reacts with dialkylzinc to give the NMR
observable O-bound zinc enolate 47 with evolution of methane
gas.14,25 Analogously, reaction of Koser’s reagent (9) with
dimethylzinc leads to methane evolution consistent with the
formation of the corresponding zinc alkoxide 48.
These two species may combine through nucleophilic
substitution of OTS on 48 by 47 to form 49. Direct methyl
transfer may take place from MeZnOTs (red methyl, path a) or
from the MeZnO bound to I(III) (blue methyl path b) to give
11a. In path c transfer of methyl to I(III) would lead to formation
of 50 and a reductive elimination-like step lead to products. In
paths d and e rearrangement of 48 would occur to form an
unstable5,26 hypervalent alkyl iodide(III) species 51 which could
act as a classical electrophile and alkylate 47 (path d).
Alternatively, nucleophilic attack by 47 on the iodine atom of
51 would lead to formation of 52. Iodine(III) enolate 52 differs
from 50 only by the source of the methyl group. I(III) enolate 52
would then undergo a reductive elimination-like step to give the
observed products.
gas evolution (methane or ethane) was observed. Methyl tosylate
was shown to be the product of the reaction of Koser’s reagent 11
with trimethylaluminum or dimethylzinc in the absence of
carbonyl compounds (eq 1). This side reaction leads to loss of
the reactive reagent in a time dependent manner. Alkyl tosylates
acting as alkylating agents in the reaction were unambiguously
ruled out by carrying out the umpolung alkylation of 10 in the
presence of d3-methyl tosylate d3-41.14 Product 11a was isolated
in 93% yield with no detectable trace of d3-methylated product.14
We surmised that the reaction might proceed via a C-21 or O-
bound7 I(III) enolate such as 42 (eq 2). This enolate could react
via a carbene, radical, or ionic pathway.5 A carbene, e.g. 43, can be
generated from α-methylene-1,3-bis-carbonyl compounds via
formation of formal CI iodide(III) species.5,22 However, since
the present reaction proceeds for alkyl-substituted compounds
such as 1, a carbene intermediate is precluded. Iodide(III)
enolate 42 may break down homolytically to give a C-centered
radical i.e. 44. To detect a possible radical intermediate, the
reaction of β-keto ester 10 was carried out in the presence of C60
(45), an effective radical trapping agent (see Scheme S1).23 GC-
HRMS analysis of the reaction mixture showed no trace of C60-
10 adducts.14 Only trace amounts of ethylated C60 46 could be
detected by HRMS. The major C60 component in the reaction
mixture was unreacted C60. The formation of ethyl radicals from
diethylzinc by the action of oxygen has been described by
Feringa.24 The possibility of an alkyl-radical or radical 44
participating in product formation was additionally ruled out by
carrying out the reaction of 10 with dimethylzinc in the presence
of 2 equiv of the powerful hydrogen atom donor, 1,4-
cyclohexadiene (29) (eq 3). This experiment led to formation
of 11a in 85% yield indicating that a radical mechanism is
unlikely.
To distinguish between these disparate possibilities, we carried
out a crossover experiment (Scheme 4). We formed the methyl-
Scheme 4. Crossover Experiments Rule out Paths b, d, and e
zinc-enolate 54 by the reaction of dimethylzinc and β-keto-ester
10. In a separate vessel, Koser’s reagent (9) was reacted with
diethylzinc to give 55. In both vessels, gas evolution could be
clearly observed. The vessels were left at −78 °C for 8 h, then the
content of the flask containing 55 was added to the vessel
containing the enolate 54, and the mixture was stirred for 12 h.
After workup, the formation of a single alkylated product, namely
11a, was observed in 40% yield, along with 60% unreacted
starting material 10 (Experiment 1, Scheme 4). None of the
alternative ethyl product 11b could be detected. To rule out the
possibility of a kinetic effect, the experiment was repeated,
switching zinc reagents. In this experiment (Experiment 2,
The remaining alternative was an ionic mechanism. We
identified five distinct possible mechanistic pathways (paths a−e,
Scheme 3). Paths a and c would be distinguishable from paths b,
C
DOI: 10.1021/ol503384c
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(4) For an elegant three-step one-pot procedure for α-arylation and
alkylation of ketones involving enamine formation using isoxazolidine
(2 equiv) and alkylation with an organoaluminium reagent (2 equiv),
see: Miyoshi, T.; Miyakawa, T.; Ueda, A.; Miyata, O. Angew. Chem. 2011,
123, 958.
(5) (a) Zhdankin, V. V.Hypervalent Iodine Chemistry: Preparation,
Structure, and Synthetic Applications of Polyvalent Iodine Compounds;
Wiley: 2014. (b) Singh, F. V.; Wirth, T. Oxidative Functionalization
with Hypervalent Halides. In Comprehensive Organic Synthesis, 2nd ed.;
Molander, G. A., Knochel, P., Eds.; Elsevier: Oxford; 2014; Vol. 7, p 880.
(c) Zhdankin, V. V.; Stang, P. J. Chem. Rev. 2008, 108, 5299. (d) For a
recent example of TMS-enolether dimerization, C−O and C−N bond
formation through umpolung, see: Mizar, P.; Wirth, T. Angew. Chem.,
Int. Ed. 2014, 53, 5993.
Scheme 4), the ethylated product 11b was isolated in 25% yield
along with 75% starting material. None of the methylated
product (11a) was formed. These experiments show that alkyl
transfer takes place from the alkylzinc originally bound in enolate
47 only and therefore unambiguously rule out paths b, d, and e.
A focused DFT-based computational study of the reaction was
carried out to further elucidate the mechanism. Specifically, we
were interested in whether the proposed mixed phenyl-methyl-
iodine(III) 50 (Scheme 3, path c) would be able to selectively
transfer the methyl group or if the proposed path c instead would
result in formation of the phenyl-substituted product. Based on
earlier computational studies,7,27 we constructed a model system,
consisting of a mixed alkyl-aryl iodine(III) enolate 50 (see SI
Scheme S2).
The energy of the transition state for methyl transfer was 20
kJ/mol higher than the transition state for transfer of the phenyl
moiety, calculated as the difference in Gibbs free energies at 195
K (−78 °C). Employment of implicit solvation (PB-SCRF,
benzene) resulted in a marginally larger energy difference (23 kJ/
mol). As only methyl transfer is observed in the reaction, these
results rule out path c, leaving path a as the only plausible
alternative.
We have designed, developed, and studied a hypervalent
iodine mediated alkylative umpolung reaction of ketones. This
reaction constitutes a strategic alternative to classical alkylation
and an important addition to the umpolung arsenal.
Experimental and computational evidence supports the reaction
proceeding through an ionic mechanism (Scheme 3, path a).
Importantly, it is applicable to a wide range of carbonyl
compounds including normal ketones via their lithium enolates
and 1,3-dicarbonyl compounds. Quaternary C-centers may be
formed at low temperatures and mild conditions. Further studies
including in asymmetric synthesis are underway in our
laboratory.
(6) Recent examples: (a) Allen, A.; MacMillan, D. W. C. J. Am. Chem.
Soc. 2011, 133, 4260. (b) Harvey, J. S.; Simonovich, S. P.; Jamison, C. R.;
MacMillan, D. W. C. J. Am. Chem. Soc. 2011, 133, 13782.
(7) Norrby, P.-O.; Petersen, T. B.; Bielawski, M.; Olofsson, B. Chem.
Eur. J. 2010, 16, 8251.
(8) Skucas, E.; MacMillan, D. W. C. J. Am. Chem. Soc. 2012, 134, 9090.
(9) (a) Fernan
2010, 16, 9457. (b) Fernan
Waser, J. Adv. Synth. Catal. 2013, 355, 163.
́
dez Gonzal
́
ez, D.; Brand, J. P.; Waser, J. Chem.Eur. J.
́
dez Gonzal
́
ez, D.; Brand, J. P.; Mondiere, R.;
̀
(10) (a) Kieltsch, I.; Eisenberger, P.; Togni, A. Angew. Chem., Int. Ed.
2007, 46, 754. (b) Eisenberger, P.; Gischig, S.; Togni, A. Chem.Eur. J.
2006, 12, 2579. (c) Allen, A. E.; MacMillan, D. W. C. J. Am. Chem. Soc.
2010, 132, 4986. (d) Umemoto, T.; Gotoh, Y. Bull. Chem. Soc. Jpn. 1987,
60, 38. (e) Umemoto, T.; Kuriu, Y.; Nakayama, S. Tetrahedron Lett.
1982, 23, 1169.
(11) (a) Zhdankin, V. V.; Mullikin, M.; Tykwinski, R.; Berglund, B.;
Caple, R.; Zefirov, N. S.; Koz’min, A. S. J. Org. Chem. 1989, 54, 2605.
(12) For a radical organocatalytic allylation of ketones induced by
cerium, see: Mastracchio, A.; Warkentin, A. A.; Walji, A. M.; MacMillan,
D. W. C. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 20648.
(13) For a review, see: (a) Liu, C.; Zhang, H.; Shi, W.; Lei, A. Chem.
Rev. 2011, 111, 1780. For an example, see: (b) DeMartino, M. P.; Chen,
K.; Baran, P. S. J. Am. Chem. Soc. 2008, 130, 11546.
(14) See Supporting Information for details.
(15) Koser, G. F.; Relenyi, A. G.; Kalos, A. N.; Rebrovic, L.; Wettach, R.
H. J. Org. Chem. 1982, 47, 2487. For a review on Koser’s reagent, see:
Moriarty, R. M.; Vaid, R. K.; Koser, G. F. Synlett 1990, 365.
(16) Pu, L.; Yu, H.-B. Chem. Rev. 2001, 101, 757.
(17) (a) Winfield, C. J.; Al-Mahrizy, Z.; Gravestock, M.; Bugg, T. D. H.
J. Chem. Soc., Perkin Trans. 1 2000, 3277. (b) Garvey, D. S.; Larosa, G. J.;
Greenwood, J. R.; Frye, L. L.; Quach, T.; Cote, J. B.; Berman, J.
WO2013058809A1, 2013.
(18) He, Z.; Li, H.; Li, Z. J. Org. Chem. 2010, 75, 4636.
(19) Stauffer, S. R.; Coletta, C. J.; Tedesco, R.; Nishiguchi, G.; Carlson,
K.; Sun, J.; Katzenellenbogen, B. S.; Katzenellenbogen, J. A. J. Med.
Chem. 2000, 43, 4934.
ASSOCIATED CONTENT
* Supporting Information
Additional schemes, procedures and characterization data. This
material is available free of charge via the Internet at http://pubs.
acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
■
*E-mail: Szpilman@tx.technion.ac.il.
Notes
(20) (a) Adamczyk, M.; Watt, D. S.; Netzel, D. A. J. Org. Chem. 1984,
49, 4226. (b) Eames, J.; Weerasooriya, N.; Coumbarides, G. S. Eur. J.
Org. Chem. 2002, 181. (c) Takeda, T.; Terada, M. J. Am. Chem. Soc.
2013, 135, 15306.
(21) Moriarty, R. M.; Hu, H.; Gupta, S. C. Tetrahedron Lett. 1981, 22,
1283.
(22) Neilands, O.; Karele, B. Zh. Org. Khim. 1965, 1, 1854.
(b) Neilands, O.; Karele, B. Zh. Org. Khim 1966, 2, 488.
(23) Tzirakis, M. D.; Orfanopoulos, M. Chem. Rev. 2013, 113, 5262.
(24) (a) van der Deen, H.; Kellogg, R. M.; Feringa, B. L. Org. Lett.
́
́ ́ ̌ ́
2000, 2, 1593. (b) Fernandez-Ibanez, M. A.; Macia, B.; Minnaard, A. J.;
Feringa, B. L. Org. Lett. 2008, 10, 4041.
(25) (a) van der Steen, F. H.; Boersma, J.; Spek, A. L.; van Koten, G. J.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
This research was supported by an Israel Science Foundation
Individual Research grant (Grant No. 1419/10).
■
REFERENCES
■
(1) (a) Seebach, D.; Corey, E. J. J. Org. Chem. 1975, 40, 231.
(b) Miyata, O.; Miyoshi, T.; Ueda, M. ARKIVOC 2013, 60.
(2) (a) For a review, see: Izquierdo, J.; Hutson, G. E.; Cohen, D. T.;
Scheidt, K. A. Angew. Chem., Int. Ed. 2012, 51, 11686. For some selected
examples, see: (b) Chen, M. Z.; Gutierrez, O.; Smith, A. B. Angew.
Chem., Int. Ed. 2014, 53, 1279. (c) Lee, K.; Kim, H.; Hong, J. Angew.
Chem., Int. Ed. 2012, 51, 5735.
Organomet. Chem. 1990, 390, C2l. (b) Bolm, C.; Muller, J.; Zehnder, M.;
̈
Neuburger, M. A. Chem.Eur. J. 1995, 1, 212.
(26) Guo, W.; Vallcorba, O.; Vallribera, A.; Shafir, A.; Pleixats, R.; Rius,
J. ChemCatChem 2014, 6, 468.
(3) For organocatalytic formation of acyl anion equivalents, see:
(a) Bugaut, X.; Glorius, F. Chem. Soc. Rev. 2012, 41, 3511. (b) Rovis, T.;
Nolan, S. P. Synlett 2013, 24, 1188.
(27) Malmgren, J.; Santoro, S.; Jalalian, N.; Himo, F.; Olofsson, B.
Chem.Eur. J. 2013, 19, 10334.
D
DOI: 10.1021/ol503384c
Org. Lett. XXXX, XXX, XXX−XXX