Mechanistic investigations of the ZnCl2-mediated tandem Mukaiyama aldol lactonization: Evidence for asynchronous, concerted transition states and discovery of 2-oxopyridyl ketene acetal variants

Article abstract of DOI:10.1021/ja209163w

The ZnCl₂-mediated tandem Mukaiyama aldol lactonization (TMAL) reaction of aldehydes and thiopyridyl ketene acetals provides a versatile, highly diastereoselective approach to trans-1,2-disubstituted β-lactones. Mechanistic and theoretical studies described herein demonstrate that both the efficiency of this process and the high diastereoselectivity are highly dependent upon the type of ketene acetal employed but independent of ketene acetal geometry. Significantly, we propose a novel and distinct mechanistic pathway for the ZnCl₂-mediated TMAL process versus other Mukaiyama aldol reactions based on our experimental evidence to date and further supported by calculations (B3LYP/BSI). Contrary to the commonly invoked mechanistic extremes of [2+2] cycloaddition and aldol lactonization mechanisms, investigations of the TMAL process suggest a concerted but asynchronous transition state between aldehydes and thiopyridyl ketene acetals. These calculations support a boat-like transition state that differs from commonly invoked Mukaiyama "open" or Zimmerman-Traxler "chair-like" transition-state models. Furthermore, experimental studies support the beneficial effect of pre-coordination between ZnCl₂ and thiopyridyl ketene acetals prior to aldehyde addition for optimal reaction rates. Our previously proposed, silylated β-lactone intermediate that led to successful TMAL-based cascade sequences is also supported by the described calculations and ancillary experiments. These findings suggested that a similar TMAL process leading to β-lactones would be possible with an oxopyridyl ketene acetal, and this was confirmed experimentally, leading to a novel TMAL process that proceeds with efficiency comparable to that of the thiopyridyl system.

Full text of DOI:10.1021/ja209163w

Article
pubs.acs.org/JACS
Mechanistic Investigations of the ZnCl2-Mediated Tandem Mukaiyama
Aldol Lactonization: Evidence for Asynchronous, Concerted Transition
States and Discovery of 2-Oxopyridyl Ketene Acetal Variants
Cunxiang Zhao,†,⊥ T. Andrew Mitchell,*,†,‡ Ravikrishna Vallakati,† Lisa M. Perez,*,# and Daniel Romo*,†
́
#
†Department of Chemistry and Laboratory for Molecular Simulation, Texas A&M University, P.O. Box 30012, College Station,
Texas 77842-3012, United States
‡Department of Chemistry, Illinois State University, Campus Box 4160, Normal, Illinois 61790-4160, United States
S
* Supporting Information
ABSTRACT: The ZnCl2-mediated tandem Mukaiyama aldol
lactonization (TMAL) reaction of aldehydes and thiopyridyl
ketene acetals provides a versatile, highly diastereoselective
approach to trans-1,2-disubstituted β-lactones. Mechanistic and
theoretical studies described herein demonstrate that both the
efficiency of this process and the high diastereoselectivity are
highly dependent upon the type of ketene acetal employed
but independent of ketene acetal geometry. Significantly, we
propose a novel and distinct mechanistic pathway for the ZnCl2-mediated TMAL process versus other Mukaiyama aldol reactions
based on our experimental evidence to date and further supported by calculations (B3LYP/BSI). Contrary to the commonly invoked
mechanistic extremes of [2+2] cycloaddition and aldol lactonization mechanisms, investigations of the TMAL process suggest a
concerted but asynchronous transition state between aldehydes and thiopyridyl ketene acetals. These calculations support a boat-like
transition state that differs from commonly invoked Mukaiyama “open” or Zimmerman−Traxler “chair-like” transition-state models.
Furthermore, experimental studies support the beneficial effect of pre-coordination between ZnCl2 and thiopyridyl ketene acetals prior
to aldehyde addition for optimal reaction rates. Our previously proposed, silylated β-lactone intermediate that led to successful TMAL-
based cascade sequences is also supported by the described calculations and ancillary experiments. These findings suggested that a
similar TMAL process leading to β-lactones would be possible with an oxopyridyl ketene acetal, and this was confirmed experimentally,
leading to a novel TMAL process that proceeds with efficiency comparable to that of the thiopyridyl system.
Scheme 1. Synthesis of cis- and trans-β-Lactones Promoted
by SnCl4 and ZnCl2
INTRODUCTION
■
Due to the versatility of the β-lactone (2-oxetanone) nucleus,
several groups have recently pursued the development of
stereoselective routes to these strained heterocycles and their
application to natural product synthesis.1 Two of the most
widely employed strategies to construct these heterocycles
directly from aldehydes are [2+2] cycloadditions with ketenes
or ketene equivalents2 and tandem aldol lactonizations with
enolates or enolate equivalents.3 Although Staudinger first
introduced the synthesis of β-lactones via the [2+2] cyclo-
addition in 1911,4 it was not until 1982 that Wynberg achieved
an effective asymmetric synthesis of β-lactones proceeding
through a proposed aldol-lactonization process.5 Utilizing
ketenes and catalytic amounts of Cinchona alkaloids, presumed
ammonium enolates were generated and provided β-lactones
when reacted with highly electrophilic aldehydes, such as
chloral. Since this groundbreaking achievement, stereoselective
routes to β-lactones have increased dramatically.1 Building on a
single example described by Hirai,6 we have extensively deve-
loped the tandem Mukaiyama aldol lactonization (TMAL)
reaction of aldehydes 1 and thiopyridyl ketene acetals 2 with
and others10 have utilized the TMAL process to access both
racemic and optically active β-lactones with high diastereose-
lectivity. For example, the TMAL process was utilized in the
total synthesis of (−)-panclicin D,8 tetrahydrolipstatin/orlistat
and derivatives,11 okinonellin B,12 brefeldin A,13 and belactosin
C.14 The TMAL reaction was also recently employed to access
activity-based probes premised on the tetrahydrolipstatin core
to determine off-targets of this potential anticancer agent15
which is known to inhibit the thioesterase domain of mam-
malian fatty acid synthase.16
7
8
both SnCl4 and ZnCl2 for the diastereoselective synthesis of
Received: September 29, 2011
Published: December 30, 2011
either cis- or trans-β-lactones 3, respectively (Scheme 1). We9
© 2011 American Chemical Society
3084
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
The Mukaiyama aldol reaction17 is a venerable, benchmark
reaction for the formation of C−C bonds and typically pro-
ceeds through an open transition state in contrast to the cyclic
Zimmerman-Traxler transition states typical of aldol reactions
involving metalated enolates (Scheme 2).18 In contrast, we now
dialkyl effects21) was observed unlike related lactonization pro-
cesses that deliver β-lactones,22 (2) only stoichiometric quan-
tities of ZnCl2 and SnCl4 efficiently promote the reaction
and no detectable quantities of typical aldol products (i.e.,
β-hydroxy and/or β-silyloxy thiopyridyl esters) were observed
with these Lewis acids, (3) aliphatic aldehydes and p-nitro-
benzaldehyde were reactive, but ketones were unreactive, (4)
with respect to ketene acetal geometry, stereoconvergence was
observed with aliphatic, achiral aldehydes leading to observed
high diastereoselectivity for trans-β-lactones with the exception
of p-nitrobenzaldehyde (vide infra). The described studies led
to optimization of the TMAL process through the identification
of the silylated β-lactone intermediate (E)-5 that could also
be intercepted to access novel tandem or cascade processes
including a route to β-chlorosilyl esters7 and a cascade, three-
component synthesis of tetrahydrofurans (THFs)23 and tetra-
hydropyrans (THPs).24 Finally, we describe a new TMAL
reaction employing an oxopyridyl ketene acetal, which was
predicted to react in a similar manner based on B3LYP/BSI
calculations (see Figure 1b). This was indeed verified experi-
mentally to deliver trans-β-lactones in comparable yields to
sulfur analogues.
Scheme 2. Commonly Invoked Transition-State
Arrangements for Mukaiyama Aldol Reactions (Open
Transition State, TS) and Metalated Enolate Aldols
(Zimmerman−Traxler, Chairlike TS)
describe detailed experimental and computational studies that
support a unique transition-state arrangement for the ZnCl2-
mediated tandem Mukaiyama aldol-lactonization process that
directly delivers β-lactones. In particular, a novel asynchronous,
concerted pathway for the ZnCl2-mediated TMAL process
proceeding through transition state (E)-4 involving tetrahedral
zinc(II) and a functional, silylated β-lactone (E)-5 intermediate
are supported by B3LYP/BSI calculations19 of the TMAL pro-
cess (Figure 1a). Several features of the TMAL reaction that
RESULTS AND DISCUSSION
■
Optimization of the TMAL Process: The Indispensable
Pyridyl Moiety25 for Efficient β-Lactone Formation and
High Diastereoselectivity. During early optimization studies
of the TMAL process with hydrocinnamaldehyde (1a) and
ketene acetals 2, 7, and 8, we discovered that the formation of
β-lactones, as opposed to typical aldol adducts, was highly
dependent on the ketene acetal and Lewis acid employed in
terms of both product distribution and diastereoselectivity
(Table 1). Use of phenoxy ketene acetal 7a (E/Z, 7:1) with
BF3·OEt2, TiCl4, and MgCl2 delivered primarily aldol adduct
10a (entries 1, 2) or led to no reaction (entry 3).26 Application
of MgBr2·OEt2 at reduced temperature (−40 °C) to avoid
known reactions of β-lactones with this Lewis acid (e.g.,
dyotropic rearrangements27), delivered β-lactone 3a in 23%
yield and excellent diastereoselectivity (>19:1) favoring the
trans-β-lactone, but accompanied by significant quantities of
aldol adducts 9a/10a (entry 4). ZnCl2 also delivered appre-
ciable yields (20%) of β-lactone 3a, but with poor diastereo-
selectivity (∼2:1) along with greater quantities of aldol adduct
9a (entry 5).
When the tert-butyldimethylsilyl (TBS) thiophenyl ketene
acetal 8a (E/Z, 20:1) was utilized with ZnCl2, β-lactone 3a
was isolated in 16% yield along with significant quantities of
aldol adducts 9b (40%) and thioacetal byproduct 11b (17%,
entry 6). β-Lactone 3a was not isolated when triethylsilyl
(TES) thiophenyl ketene acetal 8b (E/Z, 1:7) was employed
with TiCl4, rather only aldol adduct 10c was isolated in 57%
yield (entry 7). Although the yield of β-lactone 3a increased
(32−35%) when TES (E)- or (Z)-thiophenyl ketene acetals
8b were employed with ZnCl2, diastereoselectivity was poor
(∼2:1), and this reaction led to comparable combined yields of
β-silyloxy thiophenyl ester 9c and thioacetal 11c (35−42%)
(entries 8, 9). As previously described,8,9 the use of TES
thiopyridyl ketene acetals 2b led to optimal yields of β-lactone
3a (70%) as a single diastereomer (dr, >19:1) with no aldol
Figure 1. Unique features of the ZnCl2-mediated TMAL process.
B3LYP/BSI optimized structures showing asynchronous concerted
transition states for (a) thiopyridyl ketene acetal (E)-4 and (b) a new
oxopyridyl ketene acetal (E)-6 and a versatile, silylated β-lactone
intermediate (E)-5 with a thiopyridyl dichlorozincate counterion.
1
adducts detected by H NMR analysis of the crude reaction
mixture and only trace quantities of thioacetal 11d (entry 10).8a
In contrast, reduced yields but high diastereoselectivity were
obtained with TBS thiopyridyl ketene acetals 2a (E/Z, 20:1,
were uncovered in the course of these studies20 include: (1)
lack of dependence on α-substitution (i.e., gem-dialkyl or vic-
3085
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Table 1. Dependence of the TMAL Process on Ketene Acetal and Lewis Acid: Product Distribution and Diastereoselectivity
a
b
a
a
a
entry
X
SiR3
E:Z ratio
Lewis acid
3a % yield
3a dr
9a−c % yield
10a−c % yield
11a−c % yield
d
c
g
c
c
1
OPh
OPh
OPh
OPh
OPh
SPh
SPh
SPh
SPh
SPy
TMS
TMS
TMS
TMS
TMS
TBS
TES
TES
TES
TES
TES
TBS
7:1
7:1
7:1
7:1
7:1
20:1
1:7
9:1
1:7
7:1
1:4
20:1
BF3·OEt2
TiCl4
<5
NA
<5
62
61
<5
e
c
g
c
c
2
<5
NA
<5
<5
c
g
c
c
c
3
MgCl2
<5
NA
<5
<5
<5
f
c
4
MgBr2·OEt2
ZnCl2
23
20
16
>19:1
2.2:1
1:1
9
18
<5
c
c
5
28
40
<5
<5
c
6
ZnCl2
<5
17
<5
21
26
e
c
g
c
7
TiCl4
<5
NA
<5
57
c
8
ZnCl2
ZnCl2
ZnCl2
ZnCl2
ZnCl2
35
32
70
70
1:1.7
2.0:1
>19:1
>19:1
14
16
<5
c
9
<5
c
c
c
10
11
12
<5
<5
<5
c
c
c
SPy
<5
<5
<5
c
c
c
SPy
57
>19:1
<5
<5
<5
a
b
1
Isolated yield after column chromatography unless otherwise noted. Estimated by H NMR (200 MHz) analysis of the crude reaction mixture.
c
d
e
1
Not detected by H NMR (200 MHz) analysis of the crude reaction mixture. Reaction performed at 0 °C. Reaction performed at −78→0 °C.
f
g
h
Reaction performed at −40 °C. Not applicable. See ref 7. TMS = trimethylsilyl; TES = triethylsilyl; TBS = tert-butyldimethylsilyl.
entry 12, vide infra).8b,28 Other nitrogen heterocycles (i.e.,
3-methylimidazole and 3-methylpyrimidine vs 2-thiopyridyl)
were also studied and led to inferior yields of β-lactones.8b
Taken together, these results point to the critical importance
of the combination of a 2-thiopyridyl ketene acetal and ZnCl2
as Lewis acid29 for optimal yields of trans-β-lactones via the
TMAL process.
Scheme 3. Stereodivergent TMAL Process with
p-Nitrobenzaldehyde (1b) and (E)/(Z)-Ketene Acetals 2b
The TMAL process is stereoconvergent with respect to
ketene acetal geometry with all aliphatic aldehyde substrates
studied to date but not with p-nitrobenzaldehyde. Other aro-
matic aldehydes (e.g., benzaldehyde) lead to β-lactones with
benzylic C−O bonds that are labile under TMAL conditions.
In Hirai’s single reported example of this process leading to
β-lactone 3b, only the cis-diastereomer was isolated (23%)
in reaction of (E)-propionyl ketene acetal Me-2a with
p-nitrobenzaldehyde.6 We reproduced this result with the non-
methylated thiopyridyl ketene acetal (E)-2b (7:1, E/Z ratio)
with slightly improved yield (36%) to give high diastereose-
lectivity for the cis-β-lactone 3b and the relative stereochemistry
was confirmed by X-ray analysis.8 In contrast, use of (Z)-
2-thiopyridyl propionyl ketene acetal (4:1, Z/E ratio) with
p-nitrobenzaldehyde gave exclusively trans-β-lactone 3b
(dr >19:1), albeit in only 16% yield (Scheme 3).30 The low
yields observed in these reactions make it difficult to draw
definitive conclusions from these results and ascertain the exact
nature of the observed, apparent stereospecificity; however,
the observed stereodivergence with respect to ketene acetal
geometry with p-nitrobenzaldehyde is in stark contrast to ali-
phatic aldehydes but is consistent with our proposed transition
states for the TMAL process (vide infra).
moiety (Scheme 4). When p-nitrobenzaldehyde 1b was treated
with thiophenyl ketene acetal (E)-8a under typical TMAL
conditions, the expected β-lactone 3b was isolated in 16%
yield. However, additional byproducts included thioacetal 11e
(22% yield) and a silyl ester byproduct 12 obtained as a mix-
ture of diastereomers. Desilylation occurred during flash col-
umn chromatography and the resulting carboxylic acid was
treated with diazomethane to aid purification delivering the
β-thiophenyl methyl ester 13 in 37% yield over two steps.
Related byproducts have never been observed with thiopyridyl
ketene acetals, suggesting a reduced nucleophilicity or availabi-
lity of the thiopyridyl group in comparison to the thiophenyl
group, which is likely due to enhanced stability of the biden-
tate 2-thiopyridine·ZnCl2 complex compared to a monodentate
thiophenyl group.
Evidence for a Six-Membered, Zinc(II)−Ketene Acetal
Complex in the TMAL Reaction. The stereoconvergent dia-
stereoselectivity and optimal yields of β-lactone 3a when thio-
pyridyl ketene acetal 2b was utilized (cf. Table 1, entries 10, 11)
indicates that the nitrogen of the thiopyridyl ring must perform
a pivotal function in the TMAL process. One rationale is an
An additional experiment further demonstrated the superi-
ority of the thiopyridyl moiety compared to the thiophenyl
3086
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Scheme 4. TMAL Byproducts with Thiophenyl Ketene Acetal 8a
inductive effect of the nitrogen, but this appears insufficient to
explain the sharp difference in diastereoselectivity between
these ketene acetals (i.e., thiophenyl 8a versus thiopyridyl 2a)
and does not correlate with the previously described lack of
reactivity when more electron-withdrawing heterocycles were
utilized in place of thiopyridyl.8b Another explanation invokes
Zn(II)−nitrogen coordination leading to a high degree of rigid-
ity in proposed transition-state arrangements. A related cyclic
transition state was proposed by Choo and Suh for Ti(IV)-
mediated Mukaiyama aldol reactions with thiopyridyl silylke-
tene acetals.31 In this case, bidentate Lewis acid coordination
was invoked between Ti(IV) and both the pyridine nitrogen
and a silylether oxygen leading to exclusive formation of the
syn-aldols independent of the geometry of the ketene acetal
(stereoconvergent). However, in the present reaction employ-
ing ZnCl2, which is also capable of bidentate chelation, the
trans-β-lactone derived from an anti-aldol process is the major
product obtained for all aliphatic aldehydes studied.
shifts. The observed chemical shifts for the complex formed
between (E)-2a and ZnCl2 are most consistent with formation
of a six-membered chelate (E)-2a·ZnCl2 versus a four-
membered chelate (Figure 2). In particular, the predicted
chemical shifts of Hc and He for the six-membered chelate are
significantly closer to the experimental data. Notably, the
calculations for the four-membered chelate (E)-2a·ZnCl2* pre-
dict an upfield shift for He compared to (E)-2a, whereas for the
six-membered chelate a downfield shift for He is predicted. To
further corroborate the energetically favored formation of the six-
membered chelate (E)-2a·ZnCl2 versus the four-membered
chelate, calculations found this complex to be lower in energy
by 1.6 kcal/mol. While it is commonly assumed that silyl
ethers are unable to coordinate for effective Cram-type chelation
during additions to α-silyloxy carbonyls,34 recent studies by Walsh
with alkyl and vinylzinc reagents are consistent with chelation-
controlled addition to α-silyloxy aldehydes and ketones.35
Additional studies support the existence and importance of
forming the pre-complexed thiopyridyl ketene acetal (E)-
2a·ZnCl2 followed by subsequent addition of the aldehyde.
Hereafter, these conditions are referred to as pre-coordination
conditions. Despite the low solubility of ZnCl2 in CH2Cl2,36
we found that mixing an equimolar amount of ZnCl2 and
thiopyridyl ketene acetal (E)-2a in CH2Cl2 resulted in a nearly
homogeneous solution after 10 h (∼0.07 M) while the (Z)-
ketene acetal 2a required only 5 h to reach homogeneity. In
contrast, mixing an equimolar amount of ZnCl2 and thiophenyl
ketene acetal (E)-8a or pyridine in CH2Cl2 did not generate a
homogeneous solution at a similar concentration. These effects
were substantiated, as described above, by significant shifts of
the protons of the thiopyridyl ring in the 1H NMR (500 MHz)
spectrum of a 1:1 mixture of thiopyridyl ketene acetal (E)-2a
and ZnCl2 (Figure 2a) in comparison to pure ketene acetal (E)-
2a (Figure 2b, in CDCl3).37 Furthermore, the reaction of pre-
coordinated (E)-2b·ZnCl2 and hydrocinnamaldehyde 1a in
While four-membered Zn−S complexes were invoked in our
originally proposed transition-state arrangements23a and
supported by X-ray crystal structures of related Zn complexes,32
we were interested in analyzing the structure of the initial
ketene acetal·ZnCl2 complexes, both experimentally and com-
putationally,19 prior to addition of aldehyde. The H NMR
1
(500 MHz, CDCl3) chemical shifts of both uncomplexed
and ZnCl2-complexed thiopyridyl ketene acetals, (E)-2a and
(E)-2a·ZnCl2 (1.0 equiv of 2a, 1.0 equiv of ZnCl2, homo-
geneous solution in CH2Cl2), respectively, were determined in
CDCl3 following solvent exchange (Table 2 and Figure 2).33 In
1
Table 2. Experimental and Calculated H NMR Shifts of
(E)-2a and (E)-2a·ZnCl2
1
CH2Cl2 was monitored by aliquot H NMR analysis (follow-
ing solvent exchange to CDCl3) and showed the production
of β-lactone 3a with simultaneous production of TESCl
(Scheme 5); however, the β-lactones were generated with
greatly diminished diastereoselectivity compared to normal
conditions (cf. Table 1, entry 10). In contrast, diminished dia-
stereoselectivity was not observed employing pre-coordination
conditions with (Z)-2b·ZnCl2 compared to normal conditions
(cf. Table 1, entry 11).
The impact of pre-coordination conditions was further de-
monstrated during the TMAL reaction of 7-trimethylsilyl-
5-heptenal 1c38 with (Z)-2b·ZnCl2 pointing to formation of
a kinetically stable Zn(II)−thiopyridyl ketene acetal complex
under these conditions (Scheme 6). Use of ketene acetal (Z)-2b
under normal TMAL conditions (i.e., without pre-complexation
to ZnCl2) gave only a trace amount of β-lactone 3c while the
major product resulted from facile, Lewis acid-mediated, intra-
molecular cyclization of the allylsilane to the pendant aldehyde
leading to vinyl cyclopentane 14. We predicted that precoordi-
nation of ZnCl2 with ketene acetal (Z)-2b would lead to
(E)-2a
a
(E)-2a·ZnCl2
ac
(E)-2a·ZnCl2*
b
b
b
,
H
expt
5.46
calcd
expt
5.56
calcd
calcd
Ha
Hb
Hc
Hd
He
5.66
6.94
7.47
7.53
8.59
6.26
7.25
7.29
7.69
6.48
7.39
8.06
7.88
8.37
6.99−7.02
7.33−7.35
7.54−7.57
8.42−8.44
7.32−7.35
7.63−7.64
7.83−7.87
8.85−8.89
8.82
a
b
1H NMR (500 MHz, CDCl3) data. Calculated H NMR shifts by
B3LYP/BSII//B3LYP/BSI computational studies. (E)-2a (1.0 equiv)
and ZnCl2 (1.0 equiv) in CDCl3.
1
c
addition, multiple conformations of TBS-silyl ketene acetals (E)-2a,
(E)-2a·ZnCl2 (six-membered chelate), (E)-2a·ZnCl2*(four-
membered chelate) were optimized and analyzed (B3LYP/
BSII//B3LYP/BSI) to identify low energy conforma-
1
tions for each and provide calculated H NMR chemical
3087
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
1
Figure 2. B3LYP/BSI-optimized structures accompanied by H NMR spectra (500 MHz, CDCl3) of thiopyridyl ketene acetal (E)-2a (a) with 1.0
equiv of ZnCl2 and (b) without ZnCl2. The minor peaks in the aromatic region of (a) are presumably due to the corresponding ZnCl2-complexed
thiopyridyl ester from desilylation of ketene acetal (E)-2a.
Scheme 5. Detrimental Effect of Pre-coordination
Conditions on the Reaction of Hydrocinnamaldehyde
1a and (E) Geometrical Isomer of Ketene Acetal 2b
Scheme 7. Beneficial Effect of Pre-coordination
Conditions for the TMAL Process with 4-tert-
Butyldimethylsilyloxybutanal 1d
Scheme 6. A TMAL Process Rendered Competitive
with an Intramolecuar Allylsilane Cyclization under
Pre-coordination Conditions
from a facile intramolecular cyclization. When pre-coordination
conditions were utilized, the β-lactone 3d was isolated in good
yield with <5% yield of lactol 15 once again pointing to the
stability of the (Z)-2b·ZnCl2 complex and the utility of these
conditions.
Evidence for a Silylated β-Lactone Intermediate in the
TMAL Process. We previously proposed the intermediacy of a
silylated β-lactone intermediate 5 (cf. Figure 1) in the TMAL
process.8 Tetrahydrofuran 16, isolated under normal conditions
with ketene acetal (Z)-2a following mild base treatment to
cleave the corresponding silyl ester,8a is likely generated by
cyclization of the pendant silyloxy moiety with the silylated
β-lactone 5 generated during the TMAL process (Scheme 8).41
This result sheds some light on the lack of β-lactone 3d
observed under “normal” conditions (cf. Scheme 7) since the
small quantity of β-lactone 3d that does form is most likely
opened to afford THF 16, likely aided by the action of uncom-
plexed ZnCl2. We previously reported that the size of the silyl
group of thiopyridyl ketene acetals 2 significantly impacted
isolated yields of β-lactones8b with less bulky silyl groups
(i.e., TES and TBS) providing higher yields versus larger silyl
groups (i.e., TIPS) (Table 3). The major byproducts isolated with
reduced Lewis acidity and facilitate access to a productive
TMAL process. Indeed, while cyclopentane 14 remained
the major product (48%), pre-coordination conditions gave
β-lactone 3c in 17% yield with the typically observed high
diastereoselectivity (dr >19:1) for the trans-β-lactone. Thus, pre-
coordination conditions enabled an intermolecular TMAL
process (1c + (Z)-2b·ZnCl2 → 3c) to be competitive with a
facile Lewis acid-mediated, intramolecular cyclization (1c → 14).
A more practical example of the utility of pre-coordination
conditions was observed during the TMAL reaction of 4-tert-
butyldimethylsilyloxybutanal 1d39 (Scheme 7). When this
aldehyde was treated with thiopyridyl ketene acetal (Z)-2b
under normal TMAL conditions, none of the β-lactone 3d was
generated and the major product was lactol 1540 again derived
3088
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Scheme 9. Cascade Process toward the Synthesis of
Tetrahydrofurans and Tetrahydropyrans Premised on the
Silylated β-Lactone Intermediate
Scheme 8. Isolation of Tetrahydrofuran Byproduct 16 from
4-tert-Butyldimethylsiloxybutanal 1d and Ketene Acetal
(Z)-2a Indicative of a Silylated β-Lactone Intermediate 5
Table 3. Effect of the Silyl Group of Thiopyridyl Ketene
Acetals 2c−f on the TMAL Reaction
Mechanistic Proposal for the ZnCl2-Mediated TMAL
Reaction of Achiral Aldehydes and Thiopyridyl Ketene
Acetals Garnered from B3LYP/BSI Calculations: Asyn-
chronous Concerted, Boatlike Transition-State Arrange-
ments en Route to Functional Silylated β-Lactone Inter-
mediates. Computational studies (B3LYP/BSI) were performed
to locate possible intermediates and transition states for the
various steps proposed for the TMAL mechanism using
both isomeric ketene acetals, (E)-2 and (Z)-2, leading to
thiopyridyl·ZnCl complex 24, trans-β-lactone 3, and Me3SiCl
(Scheme 10). To simplify the calculations, a trimethylsilyl
entry
SiR3 (2)
3e % yield
17 % yield
a
b
1
2
3
4
TES (2c)
66
<5 (17a)
a
b
TBS (2d)
TIPS (2e)
TBDPS (2f)
53
8
(17b)
Scheme 10. Hypothetical TMAL Process Studied by B3LYP/
BSI Calculations
a
b
20
40 (17c)
c
a
<5
56 (17d)
a
b
Refers to isolated yields. Estimated based on crude 1H NMR
c
analysis. Not detected based on crude 1H NMR (300 MHz) analysis.
bulky silyl groups were β-chlorosilyl esters 17a−d, reasonably
derived from intermolecular cleavage of the silylated β-lactone
intermediate 5 by chloride ion. This process was previ-
ously optimized for the synthesis of a trans-β-chlorosilyl ester
using a SnCl4-mediated TMAL reaction proposed to proceed
through a silylated cis-β-lactone.13 In a similar manner, use of a
TBDPS-ketene acetal in the ZnCl2-mediated TMAL process
leads to near exclusive formation of a β-chlorosilylester 17d
(Table 3, entry 4).23a Thus, a partioning between desilylation
leading to a β-lactone product (pathway a) or chloride cleavage
leading to a β-chlorosilyl ester (pathway b) that depends on the
steric size of the silyl group provides further evidence for the
silylated β-lactone intermediate 5.
The discovery of pre-coordination conditions (cf. (E)-2·ZnCl2)
and silylated β-lactone intermediates in the TMAL process
was exploited by the development of a novel cascade, three-
component synthesis of tetrahydrofurans23 and tetrahydropyr-
ans24 from ketoaldehydes via the TMAL process (Scheme 9).
This cascade sequence capitalized on the expected reactivity at
the C4 carbon of the silylated β-lactone intermediate 19a
toward substitution by a pendant ketone23b and a derived cyclic
oxocarbenium 19b that could undergo subsequent nucleophilic
addition by silane nucleophiles. This three-component coup-
ling involving a ketoaldehyde, a thiopyridyl ketene acetal, and
a silane nucleophile was applied to a concise synthesis of the
tetrahydrofuran fragment of colopsinol B.23a
group was used rather than the typical experimentally employed
triethyl or TBS groups for the ketene acetal. Structural details
of both E and Z pathways can be found in Figures 3 and 4,
respectively, and the overall energy profile of the TMAL path-
ways for both ketene acetal geometries is shown in Figure 5.
Note that only the lowest energy conformations (E Conf1 and
Z Conf2) for the (E)- and (Z)-ketene acetal reaction profiles
are shown.42
Following complexation of both the thiopyridyl ketene
acetal (E)-2 and aldehyde 1 to ZnCl2 which are calculated to
be energetically favorable (Figure 5), computational studies
indicate that formation of oxetane (E)-22 involves a concerted
but asynchronous formation of both the C−O and C−C
bond via transition state (E)-4 (i.e., (E)-21-22),43 with a ΔH⧧
(ΔG⧧) barrier of 23.48 kcal/mol (30.57 kcal/mol) and
an imaginary mode of 187.2i cm−1. Formation of the oxetane
(E)-22 is slightly endothermic (endergonic) being 0.15 kcal/mol
(7.63 kcal/mol) higher in enthalpy (free energy) than
the aldehyde−ZnCl2−thiopyridyl ketene acetal complex (E)-
21. After formation of oxetane (E)-22, the Zn−O (formerly
aldehyde O) bond is broken via transition state (E)-22-23 to
form intermediate (E)-23 with a ΔH⧧ (ΔG⧧) barrier of 6.97
(6.34) kcal/mol and a small imaginary mode of 13.2i cm−1. From
3089
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Figure 5. Reaction profile for the TMAL reaction with calculated
(B3LYP/BSI) Gibbs free energy values (ΔG).
bond breaks, increasing from 1.92 to 3.96 Å, with a ΔH⧧ (ΔG⧧)
barrier of 5.56 (7.47) kcal/mol to form the silylated β-lactone
intermediate (E)-5. The transition state, (E)-23-5, has an
imaginary mode of 95.2i cm−1. During scission of the S−C
bond, the Zn−S bond distance decreases from 3.01 Å in (E)-23
to 2.46 Å in (E)-5. The last step in the mechanism involves
chloride ion addition to silicon giving a pentacoordinate silicate
and ultimately providing trans-β-lactone 3, Zn(II)-SPy complex 24,
and Me3SiCl with a ΔH⧧ (ΔG⧧) barrier of 15.69 (16.91) kcal/mol.
The transition state for desilylation, (E)-5-products, has an
imaginary mode of 147.5i cm−1. The β-lactone 3 and
byproducts are 3.20 (24.65) kcal/mol downhill in energy
from (E)-5. In summary, the rate-limiting step for the reaction
of (E)-21 to the products (trans-3, 24, and Me3SiCl) is
calculated to be the first step ((E)-21 to (E)-22) involving a
concerted, asynchronous formation of the oxetane with a ΔH⧧
(ΔG⧧) barrier of 23.48 (30.57) kcal/mol and the ΔHrxn (ΔGrxn)
from (E)-21 to the products is calculated to be exothermic
(exergonic) by 3.20 (24.65) kcal/mol.
Figure 3. Calculated structures (B3LYP/BSI) for intermediates and
transition states of the TMAL process for the (E)-ketene acetal 2
pathway leading to trans-β-lactone 3.
The TMAL process for the (Z)-ketene acetal 21 follows
a reaction path nearly identical to that for the (E)-ketene
acetal 21. The most notable difference between these two reac-
tion pathways is the rate-limiting step, (Z)-21 → (Z)-22, which
is 4.41 (4.61) kcal/mol lower in enthalpy (free energy) for the
(Z)- compared to the (E)-ketene acetal. These computational
studies are consistent with faster reaction rates qualitatively ob-
served for the (Z)- vs (E)-ketene acetals since the former
avoids development of an unfavorable, eclipsing interaction
between the R1 group of the aldehyde and the R2 group of the
ketene acetal in the transition state (vide infra).
To determine if a larger R3 group (silicon substituent) would
have a significant effect on the rate-limiting step, the first step
of the reaction mechanism involving (E)-ketene acetal 21 was
calculated with R1 = R2 = Me and R3 = t-BuMe2 and the
activation energy decreased by less than 3 kcal/mol compared
to the TMS ketene acetal. To investigate the importance of the
sulfur atom, the rate-limiting step was recalculated with oxygen
replacing sulfur, an oxopyridyl ketene acetal, and with carbon
replacing sulfur with the previously employed substrates (R1 =
R2 = Me; R3 = Me3). The activation barriers for the oxo and
carbon ketene acetals were calculated to be within 4 kcal/mol
of the sulfur-containing model. These results lead us to
Figure 4. Calculated structures (B3LYP/BSI) for intermediates and
transition states for the TMAL process of of the (Z)-ketene acetal 2
pathway leading to trans-β-lactone 3.
(E)-22 to (E)-23 the Zn−O bond distance increases from 2.20
to 3.87 Å. In addition to the Zn−O bond breaking, the Zn−S
bond distance decreases from 3.46 to 3.01 Å. Inter-
mediate (E)-23 is 7.41 (3.63) kcal/mol higher in energy com-
pared to (E)-22.44 After cleavage of the Zn−O bond, the S−C
3090
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Scheme 11. Overall Mechanistic Proposal for the ZnCl2-Mediated TMAL Reaction of Achiral Aldehydes and Both (E)- and
a
(Z)-Thiopyridyl Ketene Acetals
a
To simplify comparison, enantiomeric structures are shown for the (Z) pathway.
conclude that the sulfur atom was not absolutely critical to the
success of the TMAL process.
and complete stereodivergent diastereoselectivity was ob-
served with p-nitrobenzaldehyde, likely due to the high energy
barrier required for syn coordination, to provide cis-β-lactone
(cf. Scheme 3).6 An asynchronous concerted, boat-like transition
state is proposed leading to formation of oxetanes (E)- and
(Z)-22. We suggest that cleavage of the Zn−O bond followed
by S−C bond cleavage releases ring strain providing a driving
force leading to diastereomeric silylated β-lactone intermediates
(E)- and (Z)-5. Stable bidentate chelation with S and N,32
between the liberated zinc and the thiopyridyl prevents release
from the thiopyridyl dichlorozincate and instead leads to chlo-
ride induced desilylation to deliver β-lactone 3 and thiopyridyl
zinc monochloride 24. Consideration of bond strengths (Zn−S,
50.0 kcal/mol; Zn−Cl, 54.7 kcal/mol)49 and results with
thiophenyl ketene acetals 8a,b (cf. Table 1, entries 6−9, and
Scheme 4) lend further evidence for the stability of Zn(II)−SPy
complexes leading to the bidentate dichlorozincate intermedi-
ate (E)-5. In contrast, when thiophenyl ketene acetals 8a,b
were utilized, no β-chlorosilyl ester byproducts (cf. Table 3)
were observed as with the use of thiopyridyl ketene acetals but
rather β-thiophenyl silylesters (cf. Scheme 4) were produced
demonstrating a Zn(II) affinity trend of PhS− > Cl− > PyS−
pointing to the importance of the thiopyridyl moiety for the
highest yield of β-lactones. Formation of the bidentate chelate
24 also prevents generation of a silyl pyridyl sulfide (PySSiR3)
that may be responsible for silicon-catalyzed aldol products as
observed in the TMAL reaction of thiophenyl ketene acetals.50
We propose that tetrahydrofurans (THFs), tetrahydropyrans
(THPs), and β-chlorosilyl esters (cf. Table 3, Schemes 8 and 9)
are all derived from competitive, invertive alkyl C−O ring
scission of the silylated β-lactone intermediate (E)- and (Z)-5
as opposed to intermolecular desilylation by chloride ion which
delivers trans-β-lactone 3. This result also correlates to the find-
ing that more readily cleaved, less bulky silyl groups (e.g.,
The proposed overall mechanism for the TMAL process
based on the combination of experimental and computational
data collected to date is presented in Scheme 11. Initially, two
distinct tetrahedral zinc(II)−thiopyridyl ketene acetal com-
plexes (E)- and (Z)-2·ZnCl2, involving coordination of ZnCl2
to both the pyridine nitrogen and silylether oxygen of the
ketene acetal, are supported by both calculations and 1H NMR
studies presented herein (cf. Table 2 and Figure 2). The lowest
energy conformations of these complexes, as determined by
calculations, are in agreement with pinwheel conformations
of ketene acetals previously described by Wilcox for similar
O-alkyl silylketene acetals.45 Upon addition of aldehyde 1, ex-
change of the silyl ether oxygen for the aldehyde occurs to form
diastereomeric tetrahedral zinc(II) complexes (E)- and (Z)-21.
Four-coordinate zinc(II) complexes with tetrahedral geo-
metry are the most common46 and several pseudorotation and
dissociation/re-coordination pathways that are not discussed
or shown may be necessary in order to access the productive
complexes (E)- and (Z)-21.47 There is a distinct energetic
difference between these two diastereomeric transition-state
arrangements leading to the asynchronous, concerted C−C and
C−O bond-forming events that generate the oxetane inter-
mediates (E)- and (Z)-22. An unfavorable syn coordination of
the aldehyde to Zn(II) is required for (E)-21 but not (Z)-21 to
access the reactive conformation for oxetane formation. MP2/
6-31G* calculations suggest that anti coordination of BF3·
acetaldehyde is favored over syn coordination by 1.22 kcal/mol
and an anti-BF3·benzaldehyde complex was estimated to be
5.31 kcal/mol more stable than the corresponding syn com-
plex.48 In line with these reports, we found that (Z)-2b pro-
vided higher diastereoselectivity than (E)-2b with hydrocinna-
maldehyde under precoordination conditions (cf. Scheme 5)
3091
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
TES or TBS) provide higher yields of β-lactone 3 (cf. Table 3).
In contrast, larger silyl groups (i.e., TIPS or TBDPS) divert
nucleophilic attack of chloride ion to the β-carbon of the
silylated β-lactone intermediate 5, leading to higher yields of
β-chloro silylesters. The stereochemical outcome of the C−O
cleavage leading to THFs and THPs23,24 (cf. Scheme 9) and
characterization of the isolable tert-butyldiphenylsilyl ester 14d
(cf. Table 3, entry 4) supports an SN2-like, invertive alkyl C−O
ring scission. Finally, these mechanistic proposals are consistent
with our finding that the ZnCl2-mediated TMAL reaction with
thiopyridyl ketene acetals 2 is stoichiometric while the same
reaction employing thiophenyl ketene acetals 8 is catalytic in
ZnCl2.
Development of an Oxopyridyl Ketene Acetal-Based
TMAL Process. The predictive power of a mechanistic pro-
posal is its greatest value. Our described experiments and calcu-
lations leading to the described mechanistic model for the
TMAL process, suggested that an oxopyridyl ketene acetal
(e.g., (E)-25) would also participate in a TMAL process since
the sulfur atom of thiopyridyl ketene acetals does not coordi-
nate to zinc in the rate limiting-step. This prediction was indeed
corroborated by experiment. As seen in Scheme 12, reasonable
the TMAL reaction and suggests that it is responsible for the
high efficiency and diastereoselectivity of this process. This
critical interaction was further supported by calculations that
show strong evidence for an asynchronous concerted, boat-like
transition-state arrangement. This mechanistic proposal
provides plausible rationalizations for the unique features of
the TMAL reaction including the absence of a gem-dialkyl effect
and the fact that typical aldol products are not produced.
Calculations and further experiments support the existence of a
silylated β-lactone intermediate, which was previously proposed
and utilized in novel multi-component processes involving the
TMAL process.23a,24 Finally, as a demonstration of the utility
of the calculations and mechanistic proposals described herein,
an oxo version of the TMAL process was proposed and, as
predicted, was found to provide similar yields and diastereo-
selectivity to the thiopyridyl analogue.
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental details and characterization data for (Z)-thiopyridyl
ketene acetal 2b, β-lactones 3c-d, thiophenyl ketene acetal 8a,
aldol adducts 9a,b and 10a,b, thioacetal 11b, β-thiophenyl
methyl ester 13, oxopyridyl ester S1, and oxopyridyl ketene
acetal 25; NMR studies; computational details; and ancillary
calculations. This material is available free of charge via the
Internet at http://pubs.acs.org.
Scheme 12. Synthesis of β-Lactone 3a from Oxopyridyl
Ketene Acetal (E)-25 and Comparison of Transition States
for Thio- versus Oxopyridyl Ketene Acetals
AUTHOR INFORMATION
Corresponding Author
romo@tamu.edu; mouse@tamu.edu; mitchell@ilstu.edu
■
Present Addresses
⊥Sundia MedTech Co., Ltd., 332 Adiseng Rd., Zhangjiang Hi-
Tech Park, Shanghai 201203
ACKNOWLEDGMENTS
■
Support of this work by the NSF (CHE 0809747) and the
Robert A. Welch Foundation (A-1280) is gratefully acknowl-
edged. We thank Prof. Dan Singleton (Texas A&M), Dr. Hong
Woon Yang (Array Biopharma), and Dr. Yingcai Wang
(Amgen) for helpful discussions and Mr. Alfred Tuley for
resynthesis of thiopyridyl ketene acetal (E)-2a. We thank
the Supercomputing Facility at Texas A&M University for
providing computer time and NSF (CHE-0541587). Mass
spectral analyses were obtained at the Texas A&M Center for
Chemical Characterization and Analysis on instruments acquired
by generous funding from the NSF (CHE-8705697) and the
TAMU Board of Regents Research Program. The NSF (CHE-
0416260, TAMU; CHE-0722385, ISU) also provided funds for
purchase of NMR instrumentation.
yield (57%) and high diastereoselectivity was obtained using
oxopyridyl ketene acetals (E)/(Z)-25 (4:1 ratio) with hydro-
cinnamaldehyde (cf. Table 1, entry 10; 70% yield with thio-
pyridyl ketene acetal) under similar reaction conditions. The
reaction is also stereoconvergent with respect to ketene acetal
geometry as expected. The slightly decreased yield compared to
the thiopyridyl ketene acetal (E)-2b suggests that the sulfur
atom may play some role toward maximum conversion. Despite
the large difference in bond lengths between the C−S and
C−O bonds (1.80 vs 1.40 Å; 1.74 vs 1.33 Å), a similar transition-
state arrangement is observed for both reactions (Scheme 12, only
(E)-ketene acetal transition state shown). This variation
in bond lengths is accommodated by a more acute C−S−C
bond angle, which leads to nearly identical bond lengths for
the forming C−C bond (1.80 vs 1.83 Å) in the two transi-
tion states.
REFERENCES
■
(1) For reviews describing β-lactones in the context of natural
products, see: (a) Wang, Y.; Tennyson, R. L.; Romo, D. Heterocycles
2004, 64, 605−658. (b) Lowe, C.; Vederas, J. C. Org. Prep. Proced. Int.
1995, 27, 305−346. For reviews describing stereoselective β-lactone
synthesis and their utility, see: (c) Yang, H. W.; Romo, D. Tetrahedron
1999, 55, 6403−6434. (d) Pommier, A.; Pons, J.-M. Synthesis 1995,
729−744. (e) Pommier, A.; Pons, J.-M. Synthesis 1993, 441−449. For
selected recent methods for β-lactone synthesis, see: (f) Merlic, C. A.;
Doroh, B. C. J. Org. Chem. 2003, 68, 6056−6059. (g) Nelson, S. G.;
Zhu, C.; Shen, X. J. Am. Chem. Soc. 2004, 126, 14−15. (h) Wilson, J. E.;
Fu, G. C. Angew. Chem., Int. Ed. 2004, 43, 6358−6360. (i) Getzle, Y.
D. Y. L.; Kundnani, V.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem.
CONCLUSIONS
■
The tandem Mukaiyama aldol lactonization (TMAL) reaction
of aldehydes and thiopyridyl ketene acetals provides a highly
diastereoselective approach to a variety of trans-1,2-disubsti-
tuted β-lactones. Through both experimental and theoretical
studies, we demonstrated that coordination of the nitrogen
of the thiopyridyl group with Zn(II) plays a crucial role in
3092
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
Soc. 2004, 126, 6842−6843. (j) Zhu, C.; Shen, X.; Nelson, S. G.
J. Am. Chem. Soc. 2004, 126, 5352−5253. (k) Calter, M. A.; Tretyak,
O. A.; Flaschenriem, C. Org. Lett. 2005, 7, 1809−1812. (l) Oh, S.;
Cortez, G. S.; Romo, D. J. Org. Chem. 2005, 70, 2835−2838.
(m) Henry-Riyad, H.; Lee, C.; Purohit, V. C.; Romo, D. Org. Lett.
2006, 8, 4363−4366. (n) Reddy, L. R.; Corey, E. J. Org. Lett. 2006, 8,
1717−1719. (o) Cho, S. W.; Romo, D. Org. Lett. 2007, 9, 1537−
1540. (p) Ma, G.; Nguyen, H.; Romo, D. Org. Lett. 2007, 9, 2143−
2146. (q) Kull, T.; Peters, R. Angew. Chem., Int. Ed. 2008, 47, 5461−
5464. (r) He, L.; Lv, H.; Zhang, Y. R.; Ye, S. J. Org. Chem. 2008, 73,
8101−8103. (s) Wang, X.-N.; Shao, P.-L.; Lv, H.; Ye, S. Org. Lett.
2009, 11, 4029−4031. (t) Chidara, S.; Lin, Y.-M. Synlett 2009, 1675−
1679. (u) Mondal, M.; Ibrahim, A. A.; Wheeler, K. A.; Kerrigan, N. J.
Org. Lett. 2010, 12, 1664−1667. (v) Kull, T.; Cabrera, J.; Peters, R.
Chem.Eur. J. 2010, 16, 9132−9139. (w) Aronica, L. A.; Mazzoni,
C.; Caporusso, A. M. Tetrahedron 2010, 66, 265−273. (x) Morris, K.
A.; Arendt, K.; Oh, S.; Romo, D. Org. Lett. 2010, 12, 3764−3767.
(y) Leverett, C. A.; Purohit, V. C.; Romo, D. Angew. Chem., Int. Ed.
2010, 49, 9479−9483.
2010, 132, 12871−12873. (i) Saadi, J.; Akakura, M.; Yamamoto, H.
J. Am. Chem. Soc. 2011, 133, 14248−14251.
(18) (a) Zimmerman, H. E.; Traxler, M. D. J. Am. Chem. Soc. 1957,
79, 1920−1923. (b) Palomo, C.; Oiarbide, M.; Garcia, J. M. Chem.
Eur. J. 2002, 8, 36−44. (c) Geary, L. M.; Hultin, P. G. Tetrahedron:
Asymmetry 2009, 20, 131−173.
(19) Computational details can be found in the Supporting
Information.
(20) Some of the experimental studies described herein were
described previously: Zhao, C. Ph. D. Thesis, Texas A&M University,
2001.
(21) (a) Allinger, N. L.; Zalkow, V. J. Am. Chem. Soc. 1960, 25, 701−
704. (b) Parrill, A. L.; Dalata, D. P. Tetrahedron Lett. 1994, 35, 7319−
7322. (c) For a review, see: Jung, M. E. Synlett 1999, 843−846.
(22) (a) Danheiser, R. L.; Nowick, J. S.; Lee, J. H.; Miller, R. F.;
Huboux, A. H. Org. Synth. 1996, 73, 61−72. (b) Wedler, C.; Schick, H.
Org. Synth. 1998, 75, 116−123. (c) Wedler, C.; Ludwig, R.; Schick, H.
Pure Appl. Chem. 1997, 69, 605−608.
(23) (a) Mitchell, T. A.; Zhao, C.; Romo, D. Angew. Chem., Int. Ed.
2008, 47, 5026−5029. For related stepwise processes, see: (b) Mead, K. T.;
Pillai, S. K. Tetrahedron Lett. 1993, 34, 6997−7000. (c) Mitchell, T. A.;
Romo, D. J. Org. Chem. 2007, 72, 9053−9059.
(24) Mitchell, T. A.; Zhao, C.; Romo, D. J. Org. Chem. 2008, 73,
9544−9551.
(25) 2-Mercaptopyridine has served as an invaluable scaffold for
many reactions. For a review, see: Schmidt, B.; Kuhn, C. J. Prakt.
Chem. 1999, 341, 114−120.
(26) More thiophilic Lewis acids, such as HgCl2, CdCl2, and CuCl2,
were also studied for their ability to promote a TMAL reaction with
thiopyridyl ketene acetals. HgCl2 and CuCl2 only led to desilylation of
the ketene acetal to the corresponding thiopyridyl ester along with
recovered aldehyde and CdCl2 gave no reaction. In light of these
failures, we reason that the accessible geometries of the metal
complexes must also play an important role for a successful TMAL
reaction. For a discussion of the affinity of transition metals toward
sulfur ligands, see: Kuehn, C. G.; Isied, S. S. Prog. Inorg. Chem. 1980,
27, 153−218.
(27) For a recent review of dyotropic rearrangements including that
of β-lactones, see: Fernandez, I.; Cossio, F. P.; Sierra, M. A. Chem. Rev.
2009, 109, 6687−6711.
(28) We considered that a zinc enolate could be formed in the
TMAL by transmetalation of the thiopyridyl ketene acetal, but
independent generation of the zinc enolate of thiopyridyl propionate
gave only aldol products with no detectable β-lactone products, see:
Abdel-Magid, A.; Pridgen, L. N.; Eggleston, D. S.; Lantos, I. J. Am.
Chem. Soc. 1986, 108, 4595−4602.
(29) A significant temperature dependence was observed for the
ZnCl2-mediated TMAL process since no reaction occurred below
23 °C.
(30) The diminished yield in this example is presumably due to facile
decarboxylation to deliver the alkene and corresponds to the yield
observed by Hirai (see ref 6).
(31) (a) Suh, K.-H.; Choo, D.-J. Tetrahedron Lett. 1995, 36, 6109−
6112. (b) Suh, K.-H.; Choo, D.-J. Bull. Kor. Chem. Soc. 1996, 17, 674−
676.
(2) (a) Tidwell, T. T. Angew. Chem., Int. Ed. 2005, 44, 5778−5785.
(b) Tidwell, T. T. Eur. J. Org. Chem. 2006, 563−576.
(3) Purohit, V. C.; Matla, A. S.; Romo, D. Heterocycles 2008, 76,
949−979.
(4) Staudinger, H.; Bereza, S. Ann. 1911, 380, 243−277.
(5) Wynberg, H.; Staring, E. G. J. J. Am. Chem. Soc. 1982, 104, 166−
168.
(6) Hirai, K.; Homma, H.; Mikoshiba, I. Heterocycles 1994, 38, 281−
282.
(7) Wang, Y.; Zhao, C.; Romo, D. Org. Lett. 1999, 1, 1197−1999.
(8) (a) Yang, H. W.; Romo, D. J. Org. Chem. 1997, 62, 4−5.
(b) Yang, H. W.; Zhao, C.; Romo, D. Tetrahedron 1997, 53, 16471−
16488.
(9) (a) Yang, H. W.; Romo, D. J. Org. Chem. 1998, 63, 1344−1347.
(b) Romo, D.; Harrison, P. H. M.; Jenkins, S. I.; Riddoch, R. W.;
Park, K.; Yang, H. W.; Zhao, C.; Wright, G. D. Bioorg. Med. Chem. 1998,
6, 1255−1272.
(10) (a) Dollinger, L. M.; Howell, A. R. J. Org. Chem. 1998, 63,
6782−6783. (b) Yin, J.; Yang, X. B.; Chen, Z. X.; Zhang, Y. H. Chin.
Chem. Lett. 2005, 16, 1448−1450. (c) Fora less successful use of the
TMAL process, see: Trost, B. M.; Papillon, J. P. N; Nussbaumer, T.
J. Am. Chem. Soc. 2005, 127, 17921−17937.
(11) (a) Ma, G.; Zancanella, M.; Oyola, Y.; Richardson, R. D.; Smith,
J. W.; Romo, D. Org. Lett. 2006, 8, 4497−4500. (b) Zhang, W.;
Richardson, R. D.; Chamni, S.; Smith, J. W.; Romo, D. Bioorg. Med.
Chem. Lett. 2008, 28, 2491−2494. (c) Richardson, R. D.; Ma, G.;
Oyola, Y.; Zancanella, M. A.; Knowles, L. M.; Cieplak, P.; Romo, D.;
Smith, J. W. J. Med. Chem. 2008, 51, 5285−5296.
(12) Schmitz, W. D.; Messerschmidt, N. B.; Romo, D. J. Org. Chem.
1998, 63, 2058−2059.
(13) Wang, Y.; Romo, D. Org. Lett. 2002, 4, 3231−3234.
(14) Cho, S. W.; Romo, D. Org. Lett. 2007, 9, 1537−1540.
(15) Yang, P.-Y.; Liu, K.; Ngai, M. H.; Lear, M. J.; Wenk, M. R.; Yao,
S. Q. J. Am. Chem. Soc. 2010, 132, 656−666.
(16) (a) Kridel, S. J.; Axelrod, F.; Rozenkrantz, N.; Smith, J. W.
Cancer Res. 2004, 64, 2070. (b) Knowles, L. M.; Axelrod, F.; Browne,
C. D.; Smith, J. W. J. Biol. Chem. 2004, 279, 30540. (c) Purohit, V. C.;
Richardson, R. D.; Smith, J. W.; Romo, D. J. Org. Chem. 2006, 71,
4549−4558.
(17) (a) Mukaiyama, T.; Narasaka, K.; Banno, K. Chem. Lett. 1973,
1011−1014. (b) Mukaiyama, T.; Banno, T.; Narasaka, K. J. Am. Chem.
Soc. 1974, 96, 7503−7509. (c) Heathcock, C. H. Science 1981, 214,
395−400. (d) Mukaiyama, T. Aldrichimica Acta 1996, 29, 59−76. For
selected recent papers describing the continued utility of the
Mukaiyama aldol reaction, see: (e) Bolla, M. L.; Patterson, B.;
Rychnovsky, S. D. J. Am. Chem. Soc. 2005, 127, 16044−16045.
(f) Cheon, C. H.; Yamamoto, H. Org. Lett. 2010, 12, 2476−2479.
(g) Alam, J.; Keller, T. H.; Loh, T.-P. J. Am. Chem. Soc. 2010, 132,
9546−9548. (h) Mei, Y.; Dissanayake, P.; Allen, M. J. J. Am. Chem. Soc.
(32) (a) Bonamico, M.; Dessy, G.; Fares, V.; Scaramuzza, L. J. Chem.
Soc., Dalton Trans. 1972, 32, 2515−2517. (b) Ashworth, C. C.; Bailey,
N. A.; Johnson, M.; McCleverty, H. A.; Morrison, N.; Tabbiner, B.
J. Chem. Soc., Chem. Commun. 1976, 743−744. (c) Becker, B.;
Radacki, K.; Wojnowski, W. J. Organomet. Chem. 1996, 521, 39−49.
(d) Fraser, K. A.; Harding, M. M. Acta Crystallogr. 1967, 22, 75−80.
(33) The TBS ketene acetal (E)-2a was utilized in NMR studies due
to facile desilylation of the corresponding TES ketene acetal (E)-2b,
which complicated spectral interpretation.
(34) While it is well known that sp3 silyl ether oxygens are
significantly less Lewis basic than the corresponding sp3 alkyl ether
oxygens, far less information is available for sp2 silyl ether oxygen
atoms. Regarding the former, see: (a) Shambayati, S.; Blake, J. F.;
Wierschke, S. G.; Jorgensen, W. L.; Schreiber, S. L. J. Am. Chem. Soc.
3093
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094

Journal of the American Chemical Society
Article
1990, 112, 697−703. (b) Keck, G. E.; Castillino, S. Tetrahedron Lett.
1987, 28, 281−284. (c) Keck, G. E.; Boden, E. P. Tetrahedron Lett.
1984, 25, 265−268.
(35) (a) Stanton, G. R.; Johnson, C. N.; Walsh, P. J. J. Am. Chem. Soc.
2010, 132, 4399−4408. (b) Stanton, G. R.; Koz, G.; Walsh, P. J. J. Am.
Chem. Soc. 2011, 133, 7969−7976.
(36) Although we have found no published data for the solubility of
ZnCl2 in CH2Cl2, we estimate, on the basis of our experiments, that it
has a solubility of <0.002 M.
(37) The downfield shift of the pyridine ring is not likely caused by a
change in solvent polarity due to the low solubility of ZnCl2 in CHCl3.
(38) Tan, T. S.; Mather, A. N.; Procter, G.; Davidson, A. H. J. Chem.
Soc., Chem. Commun. 1984, 585−586.
(39) Ikeda, Y.; Ukai, J.; Ikeda, N.; Yamamoto, H. Tetrahedron Lett.
1984, 25, 5177−5180.
(40) The yield of the volatile lactol 15 was not determined because
most of it was lost upon aqueous workup and silica gel purification.
For related reactions, see: Denmark, S. E.; Henke, B. R.; Weber, E.
J. Am. Chem. Soc. 1987, 109, 2512−2514 and references cited therein.
(41) The stereochemical assignment follows from analysis of 1H
chemical shifts and comparison to those reported in the literature for
the α-methyl groups of syn (δ 1.1) and anti (δ 1.2) diastereomeric
tetrahydrofuran methyl esters, see: (a) Mead, K. T.; Yang, H.-L.
Tetrahedron Lett. 1989, 30, 6829−6832. (b) Gouzolules, F. H.;
Whitney, R. A. Tetrahedron Lett. 1980, 26, 3441−3445. (c) Bartlett, P. A.;
Meadows, J. D.; Ottow, E. J. Am. Chem. Soc. 1984, 106, 5304−5311.
(42) The energies of E Conf2 and Z Conf1 are provided in the
Supporting Information.
(43) Transition states are numbered on the basis of structural
numbers for the starting material and product for each step; e.g., 22-23
refers to the transition state for conversion of 22 to 23.
(44) The intermediate (E)-22 is higher in energy than the Zn−O
bond-breaking transition state, (E)-22-23, on the enthalpic surface
because they are very close in energy and this transition state has a
smaller zero-point energy than (E)-22. On the E(0 K) potential energy
surface, the Zn−O bond-breaking transition state is 0.09 kcal/mol
higher in energy than (E)-22.
(45) (a) Wilcox, C. S.; Babston, R. E. J. Org. Chem. 1984, 49, 1451−
1453. (b) Babston, R. E.; Lynch, V.; Wilcox, C. S. Tetrahedron Lett.
1989, 30, 447−450. (c) For rationalization of diastereoselectivity in
Mukaiyama aldol reactions based on conformational analysis of
thioketene acetals, see: Annunziata, R.; Molteni, V.; Raimondi, L.
Magn. Reson. Chem. 1996, 34, 858−864.
(46) (a) Lipscomb, W. N.; Strater, N. Chem. Rev. 1996, 96, 2375−
2433. (b) Maret, W.; Li, Y. Chem. Rev. 2009, 109, 4682−4707.
(47) For a discussion of pseudorotation and dissociation/re-
coordination pathways in the TMAL, see ref 20.
(48) Gung, B. W.; Wolf, M. A. J. Org. Chem. 1992, 57, 1370−1375.
(49) Kerr, J. A. In CRC Handbook of Chemistry and Physics, Lide, D.
R., Ed.; CRC Press: New York, 1998−1999, pp 9-51−9-62.
(50) Carreira, E. M.; Singer, R. A. Tetrahedron Lett. 1994, 35, 4323−
4326.
3094
dx.doi.org/10.1021/ja209163w | J. Am. Chem.Soc. 2012, 134, 3084−3094