Decarboxylative Generation of 2-Azaallyl Anions: 2-Iminoalcohols via a Decarboxylative Erlenmeyer Reaction

Article abstract of DOI:10.1021/acs.orglett.5b00107

Condensation between the tetrabutylammonium salt of 2,2-diphenylglycine and aldehydes results in a decarboxylative Erlenmeyer reaction, affording 1,2-diaryl-2-iminoalcohols as a mixture of diastereomers in good yields. The diastereomeric ratio shifts over time, with the anti diastereomer and the syn oxazolidine tautomer serving as the kinetic and thermodynamic products, respectively. Addition of Lewis acids can catalyze the rates of reaction and product equilibration. The results highlight the stereochemical promiscuity of 1,2-diaryl-2-iminoalcohols in the presence of Lewis acids and Bronsted bases. (Chemical Presented).

Full text of DOI:10.1021/acs.orglett.5b00107

Letter
pubs.acs.org/OrgLett
Decarboxylative Generation of 2‑Azaallyl Anions: 2‑Iminoalcohols via
a Decarboxylative Erlenmeyer Reaction
Shaojian Tang, Jong Yeun Park, Andrew A. Yeagley, Michal Sabat, and Jason J. Chruma*
†
‡
‡
‡
,†
†Key Laboratory of Green Chemistry & Technology, College of Chemistry, Sichuan University, Chengdu, Sichuan 610064, P. R.
China
‡
Department of Chemistry, University of Virginia, Charlottesville, Virginia 22904-4319, United States
*
S Supporting Information
ABSTRACT: Condensation between the tetrabutylammonium salt of 2,2-
diphenylglycine and aldehydes results in a decarboxylative Erlenmeyer reaction,
affording 1,2-diaryl-2-iminoalcohols as a mixture of diastereomers in good yields.
The diastereomeric ratio shifts over time, with the anti diastereomer and the syn
oxazolidine tautomer serving as the kinetic and thermodynamic products,
respectively. Addition of Lewis acids can catalyze the rates of reaction and product
equilibration. The results highlight the stereochemical promiscuity of 1,2-diaryl-2-
iminoalcohols in the presence of Lewis acids and Brønsted bases.
ecarboxylation can serve as a gentle alternative to anion
generation via deprotonation with a strong base.
Scheme 2. Decarboxylative Couplings from Dϕg Salts
D
Transition metal catalyzed decarboxylative alkylations in
particular have arisen as powerful and relatively mild options
1
for nucleophilic carbanion generation and functionalization. In
this arena, we previously reported palladium-catalyzed decar-
2
3
boxylative allylation, interceptive decarboxylative allylation,
4
and decarboxylative benzylation reactions, all of which proceed
via the intermediacy of an 2-azaallyl anion (α-imino anion).
Importantly, we noted that these anionic intermediates,
generated via decarboxylation of 2,2-diphenylglycinate (Dϕg)
imines under ambient reaction conditions, could be intercepted
with electrophilic aldehydes prior to allylation by the π-
2a
allylPd(II) species (Scheme 1). An unresolved question at
Scheme 1. Previous Interception of 2-Azaallyl Anions with
Aldehydes
2a
a Brønsted acid (e.g., 3-nitrobenzoic acid) favored formation of
syn-3. Very recently, Han and Soloshonok reported similar
findings employing homochiral N-tert-butylsulfinyl-3,3,3-
trifluoroacetaldimine as an electrophile for the syn diastereose-
lective synthesis of 1-trifluoromethyl-1,2-diamines (Scheme
8
2
b). They proposed that decarboxylation occurs during or
after C−C bond formation, with the protonated carboxylic acid
TS-1 representing the most plausible transition state. These
findings encouraged us to report herein our own observations
gleaned from the exploration of a related decarboxylative
the time was the importance of the Pd catalyst in the actual
decarboxylation event. Computational studies suggest that the
relatively poor oxophilicity of the π-allylPd(II) cation allows for
5
9
Erlenmeyer reaction between tetrabutylammonium 2,2-diphe-
formation of a solvent-separated ion pair, a necessary step in the
6
nylglycinate (TBA·DϕG) and aromatic aldehydes to form 1,2-
diaryl-2-iminoalcohols (Table 1). Current evidence suggests that
the C−C bond formed in our process is sensitive to reversible
scission under the reaction conditions, resulting in a transition
from the kinetically formed anti diastereomer 4 to the
rate-limiting decarboxylation event. Recently, Zhao provided
additional evidence that palladium is not required for the
decarboxylative generation of α-imino anions from Dϕg imines
7
under ambient conditions. Specifically, his group reported that
lithium Dϕg imine salts (1), stable in MeOH, will decarboxylate
in THF to generate the corresponding 2-azaallyl anions which
are then intercepted by N-tosylarylimines to afford 1,2-diaryl-1,2-
diamines as a mixture of diastereomers (Scheme 2a). Addition of
Received: January 13, 2015
Published: April 17, 2015
©
2015 American Chemical Society
2042
DOI: 10.1021/acs.orglett.5b00107
Org. Lett. 2015, 17, 2042−2045

Organic Letters
Letter
Table 1. Decarboxylative Erlenmeyer Reaction
TBA·DϕG with a variety of aldehydes (2 equiv). There was no
significant reaction between TBA·DϕG and the relatively
electrophilic 4-cyanobenzaldehyde in toluene (Table 1, entry
1
). Simply adding Lewis acids to the reaction mixture, changing
the solvent to dichloromethane, or both resulted in formation of
the expected 2-iminoalcohol as a mixture of diastereomers
(
Table 1, entries 2−4). Molecular sieves were added to the
reaction mixture to reduce the amount of benzylimine 7
generated. As expected, syn 5a (but not anti 4a) exists in
12
equilibrium with its oxazolidine tautomer (6a). Overall, this
transformation represents a decarboxylative variant of the
tandem imine condensation/anion generation/C−C bond
9
formation process originally described by Erlenmeyer. The
procedure was successfully applied to a variety of aryl and
heteroaryl aldehydes (Table 1, entries 5−11). Nitro groups
proved incompatible, resulting in exclusive formation of
benzylimine 7i (Table 1, entry 12). Electron-rich aldehydes
were also incompatible with the reaction conditions, as can be
expected for a reaction in which decarboxylation is the rate-
limiting step (Table 1, entry 13). Acetyl 4-bromosalicylaldehyde
reacted rapidly with TBA·DϕG in PhCH (Table 1, entry 14).
3
From the resulting complex reaction mixture, we were able to
isolate and obtain X-ray single crystal structures for two of the
products, syn diastereomers 8 and 9 (Figure 1). In both products,
an acetyl group migrated to the secondary alcohol from the
1
3,14
nearby phenolic oxygen.
Figure 1. X-ray crystal structures of 8 (top) and 9 (bottom,
cocrystallized with EtOAc).
Further investigation of the Lewis acid mediated trans-
formation revealed that the diastereomeric ratio (dr) was highly
dependent on reaction time. Addition of chiral ligands (e.g.,
15
Jacobsen’s ligand ) to the reaction mixture did not dramatically
alter the final dr values, but their ability to fully solubilize the
Lewis acid catalyst did facilitate reaction monitoring by H NMR
a
b
1
Isolated yield of diastereomeric mixture. Determined from H NMR
1
of a single crude reaction mixture; dr values varied significantly
c
depending on reaction time (vide infra). Only imine 7i isolated.
spectroscopy. Accordingly, we monitored the reaction between
TBA·DϕG and 3-bromo-4-methoxybenzaldehyde (2 equiv) in
CD Cl . As shown in Figure 2, key peaks corresponding to
d
Combined isolated yield of 9 (43%) and 8 (10%).
2
2
thermodynamically preferred oxazolidine 6. These results lend
mechanistic insight into other efforts involving 2-azaallyl
anions.
products 4l−7l were readily distinguished from each other. The
four tetrabutylammonium methyl groups (set at 12H) were used
as an internal standard to determine yields for each product
relative to TBA·DϕG. In the absence of a Lewis acid or ligand,
the reaction proceeded to 61% yield over ∼6 days with a final dr
of roughly 1:1 (Table 2A). At lower conversions, anti 4l
predominated and none of the syn oxazolidine tautomer 6l was
observed. As the reaction progressed, the dr shifted toward unity,
5
,10,21
Initial interest in Dϕg as an amine transfer reagent arose from
reports that it will undergo direct decarboxylative imination of α-
ketoesters in aqueous solutions containing alkylated polyethy-
11
lenimine catalysts. To further investigate the ability of Dϕg
imines to decarboxylate under mild conditions we combined
2
043
DOI: 10.1021/acs.orglett.5b00107
Org. Lett. 2015, 17, 2042−2045

Organic Letters
Letter
There are at least two possible explanations for these
observations: either the crucial C−C bond-forming event is
actively reversible under the reaction conditions, with the anti
diastereomer 4 representing the kinetic product and the syn
oxazolidine tautomer 6 serving as the thermodynamic minimum,
or 4-anti is epimerizing to the more stable 5/6-syn via
deprotonation−reprotonation of either benzylic C−H. In
contrast to the related system described by Han and Solo-
8
shonok, the presence of benzylic imine 7l and its increasing
concentration throughout the reaction strongly suggest that
decarboxylative formation of a 2-azaallyl anion occurs prior to
C−C bond formation. The possibility of an alternative ene-type
reaction mechanism was excluded by the failure of imine 10 and
3
-bromo-4-methoxybenzaldehyde to form the corresponding 2-
1
iminoalcohols 4−6l under conditions otherwise identical to our
Lewis acid catalyzed decarboxylative Erlenmeyer procedure
(Scheme 3).
Figure 2. H NMR for a mixture of 4−7l in CD Cl .
2
2
1
Table 2. Changes in dr over Time As Determined by H NMR
Scheme 3. Attempted Lewis Acid Catalyzed ene-Type
Addition
A. Conditions: No Lewis acid or ligand
time
conversion
4l
5l
6l
anti:syn
Combining pure anti diastereomer 4l with 10 mol % Sc(OTf)3
in CH Cl at 20 °C for 36 h resulted in absolutely no change,
9
1
4
6
1
min
8%
44%
53%
57%
60%
6%
29%
31%
31%
29%
2%
12%
14%
15%
16%
0%
2%
3:1
2.1:1
1.6:1
1.3:1
∼1:1
2
2
8.6 h
2.3 h
6.3 h
38.8 h
indicating that Lewis acid alone is not sufficient to affect the
6%
16
observed equilibration to the syn diastereomer 5/6l. Depro-
9%
tonation of benzhydrilimine 10 (2 equiv) with either KHMDS or
n-BuLi followed by addition of the resulting 2-azaallyl anion to a
solution of 4l-anti and 10 mol % Sc(OTf) in CH Cl resulted in
14%
B. Conditions: Sc(OTf) (16 mol %), (R,R)-Jacobsen’s ligand (14 mol %)
3
3
2
2
time
conversion
4l
5l
6l
anti:syn
formation of 3-bromo-4-methoxybenzaldehyde at the expense of
18
2
6
2
4
9
1
.5 h
38%
49%
64%
72%
72%
77%
23%
27%
22%
19%
13%
13%
9%
12%
15%
16%
17%
18%
3%
6%
1.9:1
1.5:1
1:1.5
1:2.2
1:3.8
1:4.1
the starting 2-iminoalcohol 4l (Scheme 4).
h
7.3 h
5.5 h
3.5 h
67.5 h
19%
26%
32%
35%
Scheme 4. Decomposition of 4l-anti with 2-Azaallyl Anions
C. Conditions: (R,R)-Jacobsen’s ligand (20 mol %), no Lewis acid
time
conversion
4l
5l
6l
anti:syn
3
2
4
7
9
1
0 min
3.5 h
7.8 h
1.5 h
5.5 h
67.8 h
8%
38%
46%
48%
47%
51%
6%
21%
19%
16%
13%
11%
2%
10%
11%
12%
12%
13%
0%
5%
3:1
1.4:1
1:1.3
1:1.8
1:2.3
1:3.3
1
Importantly, the syn isomer 5/6l was not detected in the H
16
NMR spectra of the concentrated crude reaction mixture. This
13%
16%
18%
23%
suggests that (1) the hydroxylic, and not benzylic, proton of 4l is
more sensitive to deprotonation and (2) the potassium and
lithium 2-azaallyl anion salts are not as reactive as the “naked”
tetrabutylammonium salt. This observation implies that, for our
method, the 2-azaallyl anion intermediate can either attack the
aldehyde to form the products or deprotonate the hydroxylic
(
and not benzylic) proton of the products resulting in C−C bond
with the concentration of 6l increasing steadily. When the
cleavage, thus setting up an equilibrating process (Scheme 5).
The presence of Lewis or mild Brønsted acids (“M” in Scheme 5)
could accelerate the rate of this equilibration by increasing the
reaction was conducted in the presence of Sc(OTf) and (R,R)-
3
Jacobsen’s ligand the overall yield as well as the rates at which the
oxazolidine 6l formed and the dr switched in favor of the syn
isomer all increased dramatically (Table 2B). The dr at
equilibrium was significantly shifted in favor of the syn products,
with 6l becoming the predominant species within 2 days. A
Scheme 5. 2-Azaallyl Anion-Mediated Equilibration
similar trend was seen using Sc(OTf) alone (no ligand), but
3
poor peak resolution in this circumstance precluded accurate
1
6
peak integration. Remarkably, using Jacobsen’s ligand alone,
without any additional Lewis acid, had similar impacts on the dr
over time, but the overall conversion was significantly lower
1
7
(
Table 2C).
2
044
DOI: 10.1021/acs.orglett.5b00107
Org. Lett. 2015, 17, 2042−2045

Organic Letters
Letter
acidity of the hydroxylic proton via complexation to the imine
and hydroxyl moieties in 4−6. The exact mode of C−C bond
formation is not clear, as we are currently unable to discount the
possibility that the oxazolidine tautomer 6 may also be a product
of a Lewis acid catalyzed [3 + 2] cycloaddition between the
(3) Yeagley, A. A.; Lowder, M. A.; Chruma, J. J. Org. Lett. 2009, 11,
022−4025.
4
(
4) Fields, W. H.; Chruma, J. J. Org. Lett. 2010, 12, 316−319.
(5) In complementary studies, Tunge demonstrated that palladium
was critical for the decarboxylation of potassium N-(diphenyl-
methylene)phenylalanate at 100 °C: Burger, E. C.; Tunge, J. A. J. Am.
Chem. Soc. 2006, 128, 10002.
19
aldehyde and the corresponding azomethine ylide (Scheme 6).
(6) Li, Z.; Jiang, Y.-Y.; Yeagley, A. A.; Bour, J. P.; Liu, L.; Chruma, J. J.;
Scheme 6. Possible [3 + 2] Cycloaddition
Fu, Y. Chem.Eur. J. 2012, 18, 14527.
(
7) Liu, X.; Gao, A.; Ding, L.; Xu, J.; Zhao, B. Org. Lett. 2014, 16, 2118.
8) Wu, L.; Xie, C.; Mei, H.; Dai, Y.; Han, J.; Soloshonok, V. A.; Pa, Y. J.
(
Org. Chem. 2015, 80, 3187.
(
(
9) Erlenmeyer, E., Jr. Annalen 1899, 307, 70.
10) (a) Zhu, Y.; Buchwald, S. L. J. Am. Chem. Soc. 2014, 136, 4500.
(b) Li, M.; Berritt, S.; Walsh, P. J. Org. Lett. 2014, 16, 4312. (c) Li, M.;
Yucel, B.; Adrio, J.; Bellomo, A.; Walsh, P. J. Chem. Sci. 2014, 5, 2383.
(d) Niwa, T.; Suehiro, T.; Yorimitsu, H.; Oshima, K. Tetrahedron 2009,
65, 5125. (e) Niwa, T.; Yormitsu, H.; Oshima, K. Org. Lett. 2008, 10,
4689. (f) Pearson, W. H. Pure Appl. Chem. 2002, 74, 1339.
The import of these discoveries extends beyond simply
providing a new strategy for constructing 2-iminoalcohols. For
example, the observed reversibility of the key C−C bond-
forming step should serve as a clear warning about the relative
stereochemical instability of 2-iminoalcohols (particularly those
with anion stabilizing substituents) when combined with both
Lewis acids and Brønsted bases! This is most relevant to
procedures in which 1,2-diaryl-2-iminoalcohols are used as chiral
(g) O’Donnell, M. J.; Bennett, W. D.; Bruder, W. A.; Jacobsen, W. N.;
Knuth, K.; LeClef, B.; Polt, R. L.; Bordwell, F. G.; Mrozack, S. R.; Cripe,
T. A. J. Am. Chem. Soc. 1988, 110, 8520. (h) Kauffmann, T.; Berg, H.;
Ko
11) Liu, L.; Zhou, W.; Chruma, J.; Breslow, R. J. Am. Chem. Soc. 2004,
26, 8136.
12) (a) Polt, R.; Peterson, M. A.; DeYoung, L. J. Org. Chem. 1992, 57,
469. (b) Polt, R.; Sames, D.; Chruma, J. J. Org. Chem. 1999, 64, 6147.
(13) 4-Bromosalicylaldehyde also reacts rapidly with TBA·Dϕg,
̈
ppelmann, E.; Kuhlmann, D. Chem. Ber. 1997, 110, 2659.
(
1
(
5
2
0
ligands for Lewis acids. In summary, we reported a
decarboxylative variant of the Erlenmeyer reaction for the
construction of 1,2-diaryl-2-iminoalcohols. Our results provide
further validation that the decarboxylative formation of 2-azaallyl
anions from amino acid Schiff bases can proceed readily in the
absence of a transition metal catalyst. Most importantly, the
process is reversible and Lewis acids can accelerate equilibration
toward the thermodynamically favored syn oxazolidine tautomer.
The reactivity of semistabilized 2-azaallyl anions is an area of
indicating that an intramolecular proton transfer can replace the acyl
transfer step.
(14) Condensation between the potassium salt of Dϕg and 3,5-di(tert-
butyl)salicylaldehyde in EtOH forms an imine that is remarkably
resistant to decarboxylation: (a) Waser, J.; Nambu, H.; Carreira, E. M. J.
Am. Chem. Soc. 2005, 127, 8294. (b) Waser, J.; Gaspar, B.; Nambu, H.;
Carreira, E. M. J. Am. Chem. Soc. 2006, 128, 11693.
5
,7−10,21
significant interest.
The mechanistic insights gained from
(
15) Jacobsen, E. N.; Zhang, W.; Muci, A. R.; Ecker, J. R.; Deng, Li J.
Am. Chem. Soc. 1991, 113, 7063.
16) See Supporting Information for representative H NMR spectra
of all mechanistic investigations.
our studies are readily applicable to some of these other reaction
manifolds.
1
(
ASSOCIATED CONTENT
Supporting Information
(17) When Ph-pybox, which lacks the phenolic OH groups, was used,
the results were similar to those observed with no Lewis acid/ligand
■
*
S
(
Table 2A). Peak overlap between protons on the ligand and the
products prevented accurate peak integration, however.
18) Significantly more aldehyde was generated when n-BuLi was used
versus the less basic KHMDS.
19) For related [3 + 2] cycloadditions of azomethine ylides with
nonalkene dipolarophiles, see: (a) Mloston, G.; Linden, A.;
General reaction procedures, compound characterization,
transition state model, and CIF files for 8 and 9. This material
is available free of charge via the Internet at http://pubs.acs.org.
(
(
AUTHOR INFORMATION
Corresponding Author
́
■
Heimgertner, H. Helv. Chim. Acta 1998, 81, 558. (b) Gebert, A.;
Heimgartner, H. Helv. Chim. Acta 2002, 85, 2073. (c) Gebert, A.;
*E-mail: chruma@scu.edu.cn.
Linden, A.; Mloston
(d) Komatsu, M.; Okada, H.; Yokoi, S.; Minakata, S. Tetrahedron Lett.
003, 44, 1603. (e) Aldous, D. J.; Drew, M. G. B.; Draffin, W. N.;
Hamelin, E. M.-N.; Harwood, L. M.; Thurairatnam, S. Synthesis 2005,
271. (f) Draffin, W. N.; Harwood, L. M. Synlett 2006, 857.
20) For selected examples, see: (a) Colak, M.; Demirel, N.
́
, G.; Heimgartner, H. Heterocycles 2002, 56, 393.
Notes
2
The authors declare no competing financial interest.
3
(
ACKNOWLEDGMENTS
■
Financial support for this work was provided by the UVA
Institute on Aging, the Thomas F. & Kate Miller Jeffress
Memorial Trust (J-808), and the NSFC (21372159). Compound
characterization (NMR, HRMS, IR) performed by the
Comprehensive Specialized Laboratory Training Platform,
College of Chemistry, Sichuan University.
Tetrahderon: Asymmetry 2008, 19, 635. (b) Zhang, H.-L.; Jiang, F.;
Zhang, X.-M.; Cui, X.; Gong, L.-Z.; Mi, A.-Q.; Jiang, Y.-Z.; Wu, Y.-D.
Chem.Eur. J. 2004, 10, 1481. (c) Yao, S.; Meng, J.-C.; Siuzdak, G.;
Finn, M. G. J. Org. Chem. 2003, 68, 2540. (d) Tian, Q.; Jiang, C.; Li, Y.;
Jiang, C.; You, T. J. Mol. Catal. A 2004, 219, 315. (e) Liao, L.-X.; Wang,
Z.-M.; Zhou, W.-S. Tetrahderon: Asymmetry 1997, 8, 1951.
(21) (a) Ding, L.; Chen, J.; Hu, Y.; Xu, J.; Gong, X.; Xu, D.; Zhao, B.; Li,
H. Org. Lett. 2014, 16, 720. (b) Xu, J.; Chen, J.; Yang, Q.; Ding, L.; Liu,
X.; Xu, D.; Zhao, B. Adv. Synth. Catal. 2014, 356, 3219.
REFERENCES
■
(
1) For a recent review, see: Weaver, J. D.; Recio, A., III; Grenning, A.
J.; Tunge, J. A. Chem. Rev. 2011, 111, 1846.
(
2) (a) Yeagley, A. A.; Chruma, J. J. Org. Lett. 2007, 9, 2879. (b) Fields,
W. H.; Khan, A. K.; Sabat, M.; Chruma, J. J. Org. Lett. 2008, 10, 5131.
c) Qian, X.; Ji, P.; He, C.; Zirimwabagabo, J.-O.; Archibald, M. M.;
Yeagley, A. A.; Chruma, J. J. Org. Lett. 2014, 16, 5228.
(
2
045
DOI: 10.1021/acs.orglett.5b00107
Org. Lett. 2015, 17, 2042−2045