Triflic acid-catalyzed additions of 2-alkoxycarbonyl allylboronates to aldehydes. Study of scope and mechanistic investigation of the reaction stereochemistry

Article abstract of DOI:10.1021/jo062151b

(Chemical Equation Presented) The substrate scope and the effect of substrate on the observed inversion of stereoselectivity in the triflic acid-catalyzed allylboration reaction between 2-alkoxycarbonyl allylboronates and aldehydes are presented. A mechan

Full text of DOI:10.1021/jo062151b

Triflic Acid-Catalyzed Additions of 2-Alkoxycarbonyl Allylboronates
to Aldehydes. Study of Scope and Mechanistic Investigation of the
Reaction Stereochemistry
Tim G. Elford, Yuichiro Arimura, Siu Hong Yu, and Dennis G. Hall*
Department of Chemistry, UniVersity of Alberta, Edmonton, Alberta, Canada, T6G 2G2
dennis.hall@ualberta.ca
ReceiVed October 17, 2006
The substrate scope and the effect of substrate on the observed inversion of stereoselectivity in the triflic
acid-catalyzed allylboration reaction between 2-alkoxycarbonyl allylboronates and aldehydes are presented.
A mechanistic investigation is described so as to confirm the involvement of a carbocation intermediate
as the source of stereochemical inversion. This methodology allows a facile access to â,γ-disubstituted
five-membered ring lactones with an exo-methylene at the R-position.
Introduction
conditions susceptible to promote undesired processes such as
oligomerization or protodeboronation.3 Indeed, we found that
triflic acid is particularly efficient at catalyzing the additions
of very deactivated 2-alkoxycarbonyl allylboronates to alde-
hydes. This new variant proceeds with ease at a temperature
more than 100 °C lower than the corresponding uncatalyzed
reactions.2 The novel triflic acid-catalyzed procedure was applied
to the synthesis of all four diastereomers of eupomatilone-6,2 a
new natural product member of a structurally intriguing class
of lignans isolated from the indigenous Australian shrub
Eupomatia bennettii.7 In this synthesis, the E-configured allyl-
boronate E-1 was found, unexpectedly, to provide the trans
lactone product 3 upon addition to aldehyde 2 (eq 1). Indeed,
according to the expected six-membered transition structure 4
(cf., eq 2), the trans stereochemistry of product 3 is opposite to
that expected from the allylboronate’s E geometry. In fact, the
thermal uncatalyzed addition between E-1 and 2 (eq 2) afforded
the expected cis-configured lactone 6 (after an acid-promoted
lactonization on open intermediate 5). The fact that the related
Z-configured allylboronate Z-7 (the ethyl ester analogue of Z-1)
provided the expected trans lactone 3 under triflic acid catalysis2
(eq 3) suggested that isomerization had occurred somewhere
along the reaction pathway between E-1 and 3 (eq 1). In this
In the past few years, several reports have highlighted the
efficiency of simple Brønsted acids as catalysts that can facilitate
difficult organic reactions and expand their substrate scope.1
Strong organic acids like triflic acid and triflimide are remark-
ably effective in promoting various polar additions and cy-
cloadditions and are very convenient to use. These strong
Brønsted acids are soluble in a wide range of organic solvents,
and they can be prepared and employed in an anhydrous form.
Our laboratory has recently described the first examples on the
use of strong Brønsted acids to catalyze additions of allylic
boronates to aldehydes.2 Although strong Lewis acids had
previously been reported to catalyze these additions,3-6 it is quite
surprising that allylic boronates can tolerate strong protic
* Corresponding author. Tel.: (780) 492-3141. Fax: (780) 492-8231.
(1) Recent examples: (a) Williams, A. L.; Johnston, J. N. J. Am. Chem.
Soc. 2004, 126, 1612-1613. (b) Zhang, L.; Kozmin, S. A. J. Am. Chem.
Soc. 2004, 126, 10204-10205. (c) Mahoney, J. M.; Smith, C. R.; Johnston,
J. N. J. Am. Chem. Soc. 2005, 127, 1354-1355. (d) Zhang, Y.; Hsung, R.
P.; Zhang, X.; Huang, J.; Slafer, B. W.; Davis, A. Org. Lett. 2005, 7, 1047-
1050. (e) Nakashima, D.; Yamamoto, H. Org. Lett. 2005, 7, 1251-1253.
(2) Yu, S. H.; Ferguson, M. J.; McDonald, R.; Hall, D. G. J. Am. Chem.
Soc. 2005, 127, 12808-1280.
(3) Hall, D. G. Boronic Acids-Preparation, Applications in Organic
Synthesis and Medicine; Hall, D. G., Ed.; Wiley-VCH: Weinheim, 2005;
Ch. 1.
(4) Kennedy, J. W. J.; Hall, D. G. J. Am. Chem. Soc. 2002, 124, 11586-
11587.
(6) Ishiyama, T.; Ahiko, T.-a.; Miyaura, N. J. Am. Chem. Soc. 2002,
124, 12414-12415.
(7) (a) Carroll, A. R.; Taylor, W. C. Aust. J. Chem. 1991, 44, 1615-
1626. (b) Carroll, A. R.; Taylor, W. C. Aust. J. Chem. 1991, 44, 1705-
1714.
(5) Kennedy, J. W. J.; Hall, D. G. J. Org. Chem. 2004, 69, 4412-4428.
10.1021/jo062151b CCC: $37.00 © 2007 American Chemical Society
Published on Web 01/20/2007
1276
J. Org. Chem. 2007, 72, 1276-1284

2-Alkoxycarbonyl Allylboronates to Aldehydes
TABLE 1. Optimization of a Model Brønsted Acid-Catalyzed
Reaction between Deactivated Allylboronate 8 and Benzaldehyde to
Give Lactone Product 9a
Entry equiv of PhCHO
catalyst
none
CF3CO2H toluene, rt, 24 h
Tf2NH
TfOH
TfOH
Sc(OTf)3
TfOH
conditions
yield (%)b
1
2
3
4
5
6
7
8
0.9
0.9
0.9
0.9
0.9
0.9
1.5
2.0
toluene, rt, 24 h
<5
77
99
99
78
<5
96
99
toluene, rt, 24 h
toluene, rt, 24 h
toluene, 0 °C, 16 h
toluene, 0 °C, 16 h
toluene, 0 °C, 16 h
toluene, 0 °C, 16 h
TfOH
a Standard conditions. Reaction scale: approximately 0.4 mmol of
allylboronate, 1.0 M solution. Entries 1-4: R ) Et and entries 5-8: R )
methyl. b Isolated yields.
allylboronates leads to R-exo-methylene-γ-lactones.4,5,12-15 These
deactivated allylboronates, however, add very slowly under ther-
mal conditions and require reaction temperatures over 100 °C.
To optimize product yield in the addition of prototype allylbo-
ronate 8 to benzaldehyde (Table 1), we investigated the use of
Lewis acids4 and strong Brønsted acids.2 All Lewis acids tested,
including Sc(OTf)3 (entry 6), gave a slower reaction, and the
superiority of bis(triflamide) and triflic acid in providing higher
conversions was quickly evident. Also noteworthy is the
concomitant lactonization of the intermediate hydroxy ester,
which is most likely promoted by the acid catalyst. Because of
the easier handling and measurement of liquid reagents, triflic
acid was chosen for further optimization. The TfOH-catalyzed
reaction was investigated at a lower temperature in view of using
sensitive aldehydes. Further fine-tuning of reaction parameters
and substrate stoichiometry led to the conditions of entry 8.
2. Scope of Substrate. Thermal versus Triflic Acid-
Catalyzed Conditions. Once the conditions for the triflic acid-
catalyzed allylboration reaction had been optimized, we next
turned our attention to determining the substrate scope for the
aldehyde. Our interest in this investigation was 3-fold. First,
do all aldehydes display the reversal of cis/trans selectivity that
was observed in the synthesis of 3 (cf. eq 1), or is there some
governing factor that allows for only certain trans-lactones to
be formed? Second, what types of aldehydes are amenable as
substrates for this reaction so as to be a general reaction? Third,
how do triflic acid-catalyzed reactions compare with the
corresponding thermal (uncatalyzed) variant? To address these
issues, various model aldehydes were chosen and reacted with
paper, we describe an investigation of the origin of the
unexpected stereochemistry in the triflic acid-catalyzed additions
of 2-alkoxycarbonyl allylic boronates. We also present a study
on the scope of substrates for this reaction as well as a full
comparison with the thermal, uncatalyzed conditions. These
R-exo-methylene-γ-lactone products are of interest because they
could be used as building blocks for the synthesis of several
natural products. The stereoselective synthesis of such building
blocks is quite significant8,9 and could lead to a wide variety of
natural product derivatives to be screened for their biological
properties. Natural products containing this R-exo-methylene-
γ-lactone core are known to possess significant biological
activity.8 As well, derivatives of R-exo-methylene-γ-lactones
have shown significant antitumor properties.10 The biological
importance of this class of compounds is well-documented due
to their cytotoxic, allergenic, anti-inflammatory, phytotoxic, and
antimicrobial properties.11 Hence, there is a great need for a
stereospecific route to this class of compounds that would allow
for chemical diversity to be easily introduced along the synthetic
pathway.
(8) Hoffmann, H. M. R.; Rabe, J. Angew. Chem., Int. Ed. Engl. 1985,
24, 110.
(9) Bandichhor, R.; Nosse, B.; Reiser, O. Top. Curr. Chem. 2005, 243,
43-72.
(10) Matsuda, F.; Ohsaki, M.; Yamada, K.; Terashima, S. Bull. Chem.
Soc. Jpn. 1988, 61, 2123-2131.
(11) Janecki, T.; Blaszcyk, E.; Studzian, K.; Janecka, A.; Krajewska,
U.; Rozalski, M. J. Med. Chem. 2005, 48, 3516-3521.
(12) Chataigner, I.; Lebreton, J.; Zammatio, F.; Villie´ras, J. Tetrahedron
Lett. 1997, 38, 3719-3722.
(13) Kennedy, J. W. J.; Hall, D. G. J. Am. Chem. Soc. 2002, 124, 898-
899.
Results and Discussion
(14) Ramachandran, P. V.; Pratihar, D.; Biswas, D.; Srivastava, A.;
Reddy, M. V. R. Org. Lett. 2004, 6, 481-484.
(15) Kabalka, G. W.; Venkataiah, B.; Dong, G. J. Org. Chem. 2004, 69,
5807-5809.
1. Development of Optimal Protic Acid-Catalyzed Condi-
tions. The allylation of aldehydes with 2-alkoxycarbonyl
J. Org. Chem, Vol. 72, No. 4, 2007 1277

Elford et al.
TABLE 2. Reaction of E-1 with Various Aldehydes under
Thermal or TfOH-Catalyzed Conditions
TABLE 3. Reaction of Z-1 with Various Aldehydes under
Thermal or TfOH-Catalyzed Conditions
entry
aldehyde
conditionsa productb d.r.c yield (%)d
entry
aldehyde
conditionsa productb d.r.c yield (%)d
1
2
3
4
5
6
7
8
9
p-NO2C6H4CHO
p-NO2C6H4CHO
p-BrC6H4CHO
p-BrC6H4CHO
p-MeOC6H4CHO
p-MeOC6H4CHO
o-MeC6H4CHO
o-MeC6H4CHO
p-MeC6H4CHO
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
10a
10a
11b
11a
12b
12a/12b
13b
13a
14b
14a
15b
15a
16b
16a
17a
17a
18a
18a
19a
19a
>19:1
>19:1
>19:1
>19:1
>19:1
1:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
>19:1
53
90
62
79
62
62
49
80
59
48
58
67
79
92
63
51
46
59
11
66
1
2
3
4
5
6
7
8
9
p-NO2C6H4CHO
p-NO2C6H4CHO
p-BrC6H4CHO
p-BrC6H4CHO
p-MeOC6H4CHO
p-MeOC6H4CHO
o-MeC6H4CHO
o-MeC6H4CHO
p-MeC6H4CHO
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
A
B
10b
10b
11b
11b
12b
12b
13b
13b
14b
14b
15b
15b
16b
16b
17b
17b
18b
18b
19b
19b
>19:1
3.6:1
>19:1
3.1:1
>19:1
5.7:1
>19:1
6.1:1
>19:1
>19:1
>19:1
>19:1
>19:1
5.3:1
2.7:1
5.2:1
>19:1
2.5:1
52
61
59
53
40
57
49
49
91
71
78
66
26
58
46
62
44
40
71
47
10 p-MeC6H4CHO
11 C6H5CHO
12 C6H5CHO
13 E-CH3CHdCHCHO
14 E-CH3CHdCHCHO
15 H11C5CtCCHO
16 H11C5CtCCHO
17 PhCH2CH2CHO
18 PhCH2CH2CHO
19 c-C6H11CHO
10 p-MeC6H4CHO
11 C6H5CHO
12 C6H5CHO
13 E-CH3CHdCHCHO
14 E-CH3CHdCHCHO
15 H11C5CtCCHO
16 H11C5CtCCHO
17 PhCH2CH2CHO
18 PhCH2CH2CHO
19 c-C6H11CHO
20 c-C6H11CHO
>19:1
2.9:1
20 c-C6H11CHO
a Conditions used: (A) TfOH, 0 °C for 16 h or (B) 110 °C for 72 h
followed by p-TSA at room temperature overnight. b Relative configuration
was determined by cycle NOE experiments (see Supporting Information).
c The d.r. were determined from 1H NMR spectra of crude products. d Yield
is isolated yield after flash chromatography. p-TSA ) para-toluenesulfonic
acid, TfOH ) triflic acid, d.r. ) diastereomeric ratio and pin )
OCMe2CMe2O.
a Conditions used: (A) TfOH, 0 °C for 16 h or (B) 110 °C for 72 h
followed by p-TSA at room temperature overnight. b Relative configuration
was determined by cycle NOE experiments (see Supporting Information).
c The d.r. were determined from 1H NMR spectra of crude products. d Yield
is isolated yield after flash chromatography. d.r. ) diastereomeric ratio and
pin ) OCMe2CMe2O.
both E-1 and Z-1 under thermal and acid-catalyzed conditions.
The results from these reactions are summarized in Tables 2
and 3.
as major diastereomer, some of the allylborations using E-1 did
in fact give the opposite selectivity and produced the trans-
lactone as well. However, this is not always the case. Some of
the aldehydes did produce the initially expected cis-lactone when
reacted with E-1. Thus, it is apparent that not all aldehydes are
prone to give the reverse selectivity as previously observed when
using E-1 and aldehyde 2 in the synthesis of 3 (eq 1). From the
electronic nature of these aldehydes, it would seem that
aldehydes with a group (R) capable of stabilizing a carbocation
at the 5-position (γ) in the resulting lactones (entries 3, 5, 7, 9,
11, and 13 in Table 2) give unexpected results (i.e., trans-
lactones), while others that are less effective at stabilizing a
carbocation (entries 1, 15, 17, and 19 in Table 2) lead to the
expected cis-lactones when subjected to the triflic acid-catalyzed
conditions. For example, the p-nitrophenyl group (entry 1 in
Table 2) would destabilize the benzylic carbocation formed at
the γ-position, and it does lead to the observed cis-lactone being
formed exclusively.
For all new lactones that were synthesized, cycle NOE
experiments were used to determine the relative stereochemistry
between the methyl group at the 4-position and the R group at
the 5-position (see Supporting Information for a listing of
values). Because of the puckering of the five-membered
lactones, the same correlations could sometimes be seen in both
isomers. When this was the case, the size of the correlation had
to be compared. These were very different in magnitude and
allowed for definitive assignments of the relative stereochemistry
(see Supporting Information for comparison of values). Typical
correlations can be seen in Figure 1. As one can see, protons
From these results, a comparison between the isomeric
allylboronates E-1 and Z-1 can be made. For both of these
allylboronates, there was no clear trend as to which one gave
better yields for a given aldehyde. While some reactions gave
high yields of lactones from E-1 (entries 2, 8, and 14 in Table
2), others were not nearly as successful and gave lower yields.
However, the same situation occurred with Z-1, where some
aldehydes gave significantly higher yields than others. It is also
useful to note that, while most of the aldehydes used were
aromatic or straight-chain aliphatic ones, branched aldehydes
are also suitable substrates for this reaction. Specifically, the
use of cyclohexanecarboxaldehyde (entries 19 and 20 in Tables
2 and 3) as well as o-tolualdehyde (entries 7 and 8 in Tables 2
and 3) demonstrated that more sterically hindered aldehydes
are suitable for use in the triflic acid-catalyzed allylboration
reaction. R,â-Unsaturated aldehydes are also suitable substrates
as evidenced by the successful addition of both E-1 and Z-1
onto crotonaldehyde and 2-octynal (entries 13-16 in Tables 2
and 3). One reaction was attempted using E-1 with cinnamal-
dehyde under triflic acid-catalyzed conditions. No product was
isolated, and a nearly quantitative amount of starting materials
(E-1 and aldehyde) was recovered. Likewise, model ketones
did not react even at room temperature.
More interesting is the trend that is observed when one looks
at the relative stereochemistry of these lactone products. While
all reactions with Z-1 led to the corresponding trans-lactones
1278 J. Org. Chem., Vol. 72, No. 4, 2007

2-Alkoxycarbonyl Allylboronates to Aldehydes
From our work on the substrate scope, we immediately
noticed that only allylboronate E-1 gave an inversion of the
expected stereochemistry in the lactone products, while products
obtained using Z-1 did not show this reversal. Thus, there must
be an intermediate, possibly carbocation A in Figure 2, which
when E-1 is used allows for a conformational change in the
molecule (A to B, Figure 2) and subsequent formation of the
trans-lactone. With the reasonable assumption that the trans-
lactone is the kinetically favored product, the same intermediate
could also be present when Z-1 is used but still lead to the
expected product stereochemistry. Also from the substrate scope
(Tables 2 and 3), if one takes a close look at the electronic
nature of the aldehydes used, those with a group (R) that would
stabilize a carbocation on the resulting adduct are the ones that
are observed to give the unexpected stereochemistry. Any
aldehydes that contained electron-withdrawing groups or could
not stablize a nearby carbocation (e.g., alkyl and propargyl) did
not display the reversal of stereochemistry. From these observa-
tions, we strongly hinted that the lactonization must be going
through a carbocation (i.e., SN1) mechanism, but conclusive
proof needed to be found. A recent follow-up of our original
paper2 by another group16 concludes that an SN1 mechanism is
operative for lactonizations under strong Lewis acid conditions
but offers no evidence for it nor addresses several key issues.
FIGURE 1. Typical cycle NOE correlations.
present on the R group of the lactone (i.e., the substituent
originating from the aldehyde) were also useful in obtaining
NOE data to allow for stereochemical assignments.
It is also interesting to look at the proton NMR data for these
lactones and compare some of the chemical shifts between the
different cis and trans isomers. The proton at the 4-position was,
in all cases studied, further upfield in the trans-lactone than
the corresponding proton in the cis-lactone. Furthermore, the
proton at the 5-position was also further upfield in the trans-
lactone than the corresponding proton in the cis-lactone. These
results, summarized in the Supporting Information, further
confirm the relative stereochemical assignments.
Once this pattern had been established, the relative stereo-
chemistry of subsequently formed lactones was assigned based
on the relative position of these peaks in the H NMR. Cycle
1
NOE experiments were eventually performed and did in fact
confirm all of our assignments. This chemical shift pattern may
prove to be a useful way of assigning relative stereochemistry
in 5-membered exo-methylene lactones that are of similar
structure without the need to run time-consuming NOE experi-
ments. With regard to the current study on the triflic acid-
catalyzed allylboration reaction, we next turned our attention
toward investigating the mechanism of this reaction in an attempt
to understand why allylboronate E-1 leads to trans-lactones
11b-16b and to determine the nature of the observed reversal
of stereoselectivity.
3. Mechanistic Investigation into the Origin of Isomer-
ization. Several questions remained in understanding the
mechanism of the triflic acid-catalyzed reaction, such as whether
a carbocation was being formed or not, how the triflic acid was
turning over, and what the intermediate leading to the observed
reversal of stereoselectivity was. As first reported by our group,2
and further suggested by others,16 we initially suspected that
the open borate intermediate must be forming a carbocation,
which would then be trapped by the nearby ester group to form
the observed five-membered lactones (Figure 2).
3.1. Control Experiments to Address the Possibility of
Lactone Isomerization. On the basis of the acidic conditions
used in this allylboration reaction, we were initially aware that
it could be the lactones themselves that were isomerizing via a
reversible SN1 mechanism. Because of the eclipsing bulky
substituents in the cis-lactones, the trans-lactones are likely to
be thermodynamically favorable. Thus, in the presence of triflic
acid, it might be possible that the initially formed cis-lactone
product was opening up to form a carbocation, which was
reorientating itself by C-C bond rotation (i.e., A to B in Figure
2) to alleviate the steric strain of the nearby bulky groups, and
then closing back up to form the observed trans-lactones. Under
this hypothesis, aldehydes with a group (R) that could stabilize
the carbocation would undergo this inversion of stereoselectivity.
To assess whether it was the lactone products isomerizing
instead of the open borate intermediate (Figure 2), we studied
three different lactones. We started with the cis-lactones that
FIGURE 2. Early mechanistic hypotheses and questions that required investigation.
J. Org. Chem, Vol. 72, No. 4, 2007 1279

Elford et al.
using 10% 18O-enriched water. The oxygen-18 labeled aldehydes
were confirmed by high-resolution mass spectrometry (3-4%
18O incorporation, see Supporting Information for details). These
aldehydes were subjected to the normal triflic acid-catalyzed
allylboration reaction conditions, and the lactones were purified
in the typical fashion. Once isolated, they were analyzed by
mass spectrometry to determine if there were elevated levels
of oxygen-18. The normal isotopic distribution of oxygen
includes 0.20% of naturally occurring oxygen-18. On the basis
of the results from the mass spectrometry analysis, the two
lactones coming from 4-nitrobenzaldehyde and 4-bromoben-
zaldehyde showed a level of oxygen-18 that was not distin-
guishable from the naturally occurring level. This experiment
was performed twice, with the same conclusion that the oxygen
from the aldehyde was not present in the final isolated lactone
in each case. These results further supported our hypothesis that
the triflic acid-catalyzed allylboration reaction proceeds through
a carbocation intermediate where the aldehyde oxygen is
eventually lost via the ionization process. In a control experi-
ment, the oxygen-18 labeled p-nitrobenzaldehyde was also
subjected to the thermal allylboration conditions with E-1. In
this lactone product the oxygen-18 labeling could be detected
at a level similar to the starting aldehyde.
3.3. Isolation of Open Intermediates and Investigation of
Their Isomerization. To further confirm our findings, we next
turned our investigation toward the exact nature of the open
intermediate leading to a carbocation. To this end, we set out
to isolate the open form of the allylboration product. Previous
work in our group had found that a decrease in the temperature
for the thermal allylboration reaction could prevent lactonization
from occurring.5 Using this information, we synthesized the
corresponding acyclic methyl esters of three different aldehydes,
4-nitrobenzaldehyde (20), 4-bromobenzaldehyde (21), and
4-methoxybenzaldehyde (22). These model substrates were
reacted under thermal allylboration conditions with both E-1
and Z-1 in an attempt to obtain both the cis and the trans acyclic
methyl esters. Four of the methyl esters were isolated and
purified (Figure 3).
TABLE 4. Attempts of Isomerization of cis-Lactones to
trans-Lactones as Possible Source for Reversed Stereochemistry
Entry
initial lactone
ratio of lactone products (a/b)a
1
2
3
10a R ) NO2
11a R ) Br
12a R ) OMe
100:0
74:26
5:95
1
a Ratio of products determined by H NMR of crude reaction mixture.
had the p-nitro phenyl group (10a), the p-bromo phenyl group
(11a), and the p-methoxy phenyl group (12a). Each of these
cis-lactones was subjected to the same conditions used in the
catalyzed allylboration reaction, that is, 0 °C for 16 h with 10
1
mol% triflic acid. H NMR spectra of crude products were
obtained after the typical workup, and the ratio of cis/trans
isomers was measured (Table 4).
When 10a was subjected to these conditions, there was no
isomerization observed. This can be rationalized because of the
severe electron-withdrawing nature of the p-nitrophenyl group.
The resulting benzylic carbocation would be highly destabilized
by any buildup of charge, and the lactone would immediately
reform before any rotation/isomerization could occur. Thus, none
of trans-lactone 10b is observed. The complete opposite is true
for 12a. Here, the p-methoxyphenyl group stabilizes a benzylic
carbocation, and this allows for the cis isomer to completely
isomerize to the more stable trans-lactone. These two results
are consistent to those obtained in Table 2. However, the result
from entry 2 in Table 4 provides new information that rules
out the notion that isomerization of the lactone products is at
the origin of the inverted stereochemistry observed in many
cases. Even though there is partial isomerization of lactone 11a
to 11b observed under the reaction conditions, it is clearly not
enough to explain the full isomerization observed previously
for the allylboration/lactonization leading to 11b in Table 2
(entry 3). Hence, lactone isomerization cannot be the sole
mechanism affording this observed reversal of stereochemistry,
and the main origin of this isomerization must occur prior to
lactone formation.
3.2. Study of Isotopic Labeling for Tracking Down the
Aldehyde Oxygen. We next decided to use oxygen-18 labeling
of aldehydes to probe the location of the aldehyde oxygen
throughout the reaction with respect to the suspected interme-
diacy of a carbocation in the isomerization event. If the
aldehydes were labeled with oxygen-18 and then reacted with
E-1, it would be possible to determine if the oxygen from the
aldehyde ended up being incorporated into the lactones. We
chose 4-nitrobenzaldehyde and 4-bromobenzaldehyde so as to
make a good comparison between extremes of electronic
properties. These two different oxygen-18 labeled aldehydes
were made according to a previously reported literature method17
FIGURE 3. Synthesis of cis- and trans-acyclic methyl hydroxy esters.
All three aldehydes reacted with E-1 to give the corresponding
cis-methyl ester products. The assignments of stereochemistry
were done on the basis of coupling constants and comparison
of chemical shifts in the proton NMR between the methyl esters
and their corresponding lactones. This stereochemical outcome
is also consistent with standard allylboration chemistry based
on a closed six-membered transition state. When the three alde-
hydes were each reacted with Z-1, only the 4-nitrobenzaldehyde
(16) Ramachandran, P. V.; Pratihar, D.; Biswas, D. Org. Lett. 2006, 8,
3877-3879.
(17) Baraniak, J.; Wojciech, J. S. J. Chem. Soc., Perkin Trans. 1 1987,
8, 1645-1656.
1280 J. Org. Chem., Vol. 72, No. 4, 2007

2-Alkoxycarbonyl Allylboronates to Aldehydes
TABLE 5. Lactonization of Acyclic Methyl Esters under Triflic
Acid Conditions
SCHEME 1. Synthesis of Borate Intermediate and
Lactonization under TfOH Conditions
Entry
methyl ester
product
ratio of productsa
1
2
3
4
20a R ) NO2
21a R ) Br
22a R ) OMe
20b R ) NO2
10a/10b
11a/11b
12a/12b
10a/10b
90:1
66:1
16:84
0:100
1
a Ratio determined by H NMR analysis of crude reaction mixture.
trans-lactone as the major product with only a trace amount of
the cis-lactone present.
As a side note, this surprising observation that the free alcohol
intermediate 21a leads to one isomer while the corresponding
borate intermediate 23 leads to the other isomer brings to light
a selectivity issue with triflic acid activation. As discussed
earlier, in the triflic acid-catalyzed allylboration reaction, the
triflic acid activates the borate and causes carbocation formation,
which then allows for intramolecular trapping via the ester either
before any bond rotation can occur (A, Figure 2, leading to the
expected cis-lactones) or after bond rotation (B, Figure 2, leading
to the observed reversal of stereochemical outcome in the form
of the trans-lactone). However, if the borate intermediate is
trapped and hydrolyzed to the corresponding alcohol, and then
subjected to the triflic acid conditions, a different stereochemical
outcome is observed. To explain this dichotomy, a mechanism
that does not include carbocation formation must be considered
to explain the lactonization that provides retention of stereo-
chemistry. In these instances, the triflic acid must be preferen-
tially activating the carboxyester in the presence of the free
alcohol (Figure 4). The activated ester is then attacked intramo-
resulted in any acyclic methyl ester product being isolated. For
the other two cases, only the corresponding lactones were iso-
lated after workup. Interestingly, after 20b was left on the bench-
top for 2 weeks, an 1H NMR experiment was re-run, and
considerable conversion of 20b to lactone 10b was observed.
Therefore, it was assumed that 21b and 22b were not isolable
due to the fact that they were probably cyclizing spontaneously
under the reaction conditions or at room temperature to their
corresponding lactones. The isolated acyclic methyl esters were
subjected to triflic acid catalysis at 0 °C for 16 h (standard
catalyzed conditions). All underwent lactonization to their
corresponding lactones, and these results are summarized in
Table 5.
As can be seen from these results, 20a cyclizes to give mainly
the cis-lactone, and 20b cyclizes to give the trans-lactone, which
is in agreement with our previous results (entry 1, Tables 2 and
3). Furthermore, 22a cyclizes to give mainly the trans-lactone
(entry 3, Table 5), which is also in agreement with our previous
result (entry 5, Table 2). Specifically, if a carbocation can be
stabilized, it will give the trans-lactone as the major product.
If no appreciable stabilization is possible, then the cis-lactone
will be observed as the major product (entry 1, Table 5).
Whereas these first two results were quite reassuring, we noted
that the result in entry 2, however, was completely opposite to
what was expected. The p-bromophenyl group can stabilize a
carbocation through resonance, so the trans-lactone was ex-
pected as the major product based on our previous result using
allylboronate E-1 and p-bromobenzaldehyde under triflic acid
catalysis (entry 3, Table 2). Instead, the cis isomer was observed
as the major product from 21a. It should be realized, however,
that the free alcohols that we isolated and then subjected to the
catalyzed allylboration conditions are not the true intermediates
that would be present in the typical allylboration/lactonization
reactions. The free alcohols may form in a serendipitous manner
depending on the reaction conditions, but the anticipated
intermediate would contain a pinacol borate unit at the benzylic
position, not a free hydroxyl group. These borate intermediates
may be more prone to acid-catalyzed isomerization. At this
point, we went back and made the borate analogue of 21a, 23,
via a standard thermal allylboration procedure (Scheme 1).
FIGURE 4. Cyclization of open methyl ester intermediates containing
a free alcohol.
lecularly by the nearby alcohol, which leads to the eventual
elimination of a molecule of methanol. This substitution-type
mechanism would explain the retention of stereochemistry with
certain aldehyde substrates. Another alternative, which is more
consistent with the absence of O18 labeling when using
p-nitrobenzaldehyde (vide supra), supports the notion that a
carbocation is also formed in these cases but that its extreme
reactivity precludes a bond rotation prior to attack by the ester.
3.4. Proposal of a Full Mechanistic Cycle and Issue of Tri-
flic Acid Turnover. With all of this information now in hand,
the reaction stereochemistry has been clarified. The reversal of
1
An H NMR analysis was performed on the crude reaction
mixture before any workup was done to confirm the presence
of the borate group on the molecule. Once this had been
confirmed, the borate intermediate 23 was subjected to the triflic
acid-catalyzed conditions. To our satisfaction, 1H NMR analysis
of the crude product after workup showed the presence of the
J. Org. Chem, Vol. 72, No. 4, 2007 1281

Elford et al.
FIGURE 5. Triflic acid-catalyzed allylboration/lactonization mechanism.
observed stereochemistry originates from the formation of a car-
bocation from the borate intermediate, which is followed by
bond rotation to a more favorable conformer and subsequent
trapping by the nearby ester group to give the trans-lactone
(Figure 5). For aldehydes that do not stabilize a carbocation,
this catalytic cycle is still valid. The difference for these alde-
hydes is that once the intermediate carbocation forms, it would
be attacked immediately by the nearby ester to form the lactone
before any bond rotation can occur. As protonation/ionization
could require a conformation where both the borate and the car-
boxyester are held in close proximity to allow an effective proton
transfer, the carbocation would be generated in the unfavorable
arrangement (with syn R and Me groups) that leads to the cis-
lactone instead of the trans-lactone as shown in Figure 5.
Triflic acid plays numerous functions in the postulated
multistep mechanism of Figure 5. First, it is expected to catalyze
the allylboration reaction by the same electrophilic boronate
activation mechanism demonstrated before for the Lewis acid-
catalyzed variant.18 As seen from this mechanistic cycle, the
possible catalyst turnover requires the formation of one molecule
of water from pinBOH and another equivalent of triflic acid.
This water molecule would then act as a nucleophile on the
methoxy moiety of the alkoxyoxonium intermediate, leading
to the lactone product and a molecule of methanol. One full
equivalent of the triflic acid is also regenerated at this point via
elimination of pinBOMe, making the entire cycle catalytic in
TfOH. The overall reaction for the byproducts is shown in eq
4. As already mentioned, the molecule of water produced is
the absence of oxygen-18 labeling in the experiments described
in section 3.2.
It is important to note that the issue of catalyst turnover remains
uncertain and that there are other options for this catalytic cycle.
Instead of a molecule of water, it could be the triflate anion
itself that attacks the oxonium’s methoxy group to form the
final lactone products. This mechanism would lead to the
formation of MeOTf, which would be a dead-end as it would
cause irreversible consumption of the triflic acid. The initial
catalyst, triflic acid, would not be regenerated, and the reaction
would require another species to perpetuate the cycle. In this
regard, it is possible that pinBOTf could take over in the
catalytic cycle and act as a Lewis acid to promote further
allylboration/lactonization reactions to occur. At this point, it
is not immediately evident which catalytic cycle is occurring.
The triflate anion is a very poor nucleophile but still might be
capable of undergoing attack on the oxonium’s alkoxy group.
A borate can be observed in the crude 11B NMR at ∼22 ppm,
but this could correspond to either pinBOTf or pinBOMe. To
gain more information about this proposed catalytic cycle, we
went one step further and synthesized the isopropyl allylboronate
E-26 analogue of E-1 and subjected it to the triflic acid
allylboration conditions with two different aldehydes (eq 5).
pinBOH + MeOH TfOH8 pinBOMe + H2O
(4)
consumed in the reaction, and the only byproduct of the reaction
would be pinBOMe. It should be noted that the attack of water
on the oxonium carbon (as opposed to the methoxy carbon)
would lead to the same outcome, but it is not consistent with
(18) Rauniyar, V.; Hall, D. G. J. Am. Chem. Soc. 2004, 126, 4518-
4519.
1282 J. Org. Chem., Vol. 72, No. 4, 2007

2-Alkoxycarbonyl Allylboronates to Aldehydes
Quite surprisingly, both aldehydes reacted with E-26 to give
the expected corresponding lactones 10a and 11b as products.
As can be seen, however, there were considerable amounts
of the open allylboration products that had failed to undergo
lactonization. For the reaction between E-26 and p-bromoben-
zaldehyde, 28a, 11a, and 11b were obtained in a ratio of 1.3:
1:2.8. The trans-lactone 11b was still the major product, and
the presence of 28a can be rationalized by the intermediate
borate being much slower to undergo lactonization as compared
to the methyl ester analogue (Figure 2). The presence of 11a
could be due to direct attack of the hydroxyborate intermediate
on the ester. This process would be slow but would explain
how some of the cis-lactone 11a is formed. For the reaction
between E-26 and p-nitrobenzaldehyde, 27a and 10a were
obtained in a ratio of 3.2:1. This result was surprising, but if
one considers the electronic nature of the aldehyde (electron
poor), carbocation formation is less favorable and slower. Thus,
most of the borate intermediate would not undergo carbocation
formation and subsequent lactonization, and would be quenched
as the cis-alcohol during workup. Regardless of the obtained
mixtures, the presence of 11b provides some interesting insight
into the process of lactonization. Attack of a nucleophile (either
H2O or TfO-) on the isopropoxy group is much slower as
compared to the methoxy group, and this would allow for the
previously slower process of hydroxyborate attack on the ester
to become much more important and lead to the presence of
the cis-lactone 11a. The same rational holds true for the
p-nitrobenzaldehyde example; however, it is inconsequential
since the cis product is the expected one regardless of which
process is faster. In both cases, lactonization is slower, which
is in agreement with a mechanism involving nucleophilic attack
on the oxonium’s alkoxy substituent by either water or the
triflate anion. Overall, although triflic acid is essential to initiate
this allylboration/lactonization process, the issue of detailed
catalyst turnover remains speculative.
Experimental Procedures
Allylboronates E-1, Z-1, and E-26 were synthesized according
to previous literature procedures.2
General Procedure for the Synthesis of Lactones under
TfOH-Catalyzed Conditions using E-1 or Z-1 (10-19). A solu-
tion of E- or Z-allylboronate 1 (100 mg, 0.42 mmol) and aldehyde
(0.83 mmol) in toluene (1 mL) at 0 °C was treated with TfOH (4
µL, 0.04 mmol) and stirred at 0 °C under an Ar atmosphere for 16
h. The mixture was then diluted with NH4Cl(aq)/NH4OH (9:1 v/v,
10 mL) and extracted with Et2O (3 × 20 mL). The combined
extracts were washed with brine (2 × 20 mL), dried with anhydrous
Na2SO4, filtered, and concentrated. Crude products were then
purified by flash chromatography to yield the corresponding lactone.
cis-4-Methyl-3-methylene-5-(4-nitrophenyl)-dihydro-furan-2-
one (10a). Flash chromatography (20% EtOAc/hexanes) yielded
the product as a yellow solid in 53% yield. H NMR (400 MHz,
1
CDCl3): δ 8.27-8.25 (m, 2H), 7.41-7.38 (m, 2H), 6.38 (d, 1H, J
) 2.5 Hz), 5.70 (d, 1H, J ) 8.1 Hz), 5.66 (d, 1H, J ) 2.4 Hz),
3.57-3.47 (m, 1H), 0.80 (d, 3H, J ) 7.2 Hz); 13C NMR (100 MHz,
CDCl3): δ 169.7, 147.9, 143.6, 135.2, 126.9, 123.8, 122.9, 80.7,
38.7, 15.8; IR (CH2Cl2 cast film, cm-1): 3082, 2973, 2933, 1770,
1521, 1349; HRMS (EI, m/z) Calcd for C12H11NO4: 233.06880.
Found: 233.06846. Elem. Anal. (%) Calcd for C12H11NO4: C,
61.80; H, 4.75; N, 6.01. Found: C, 61.99; H, 4.97; N, 5.74.
trans-4-Methyl-3-methylene-5-(4-nitrophenyl)-dihydro-furan-
2-one (10b). Flash chromatography (20% EtOAc/hexanes) yielded
the product as a pale yellow solid in 52% yield. H NMR (400
1
MHz, CDCl3): δ 8.30-8.26 (m, 2H), 7.57-7.52 (m, 2H), 6.38 (d,
1H, J ) 3.3 Hz), 5.65 (d, 1H, J ) 2.7 Hz), 5.02 (d, 1H, J ) 7.6),
2.96-2.86 (m, 1H), 1.40 (d, 3H, J ) 6.8 Hz); 13C NMR (100 MHz,
CDCl3): δ 169.3, 145.7, 139.3, 126.5, 124.2, 122.2, 94.4, 84.3,
43.5, 16.2; IR (microscope, cm-1): 3510, 3114, 2974, 2936, 1773,
1519, 1352, 1266, 1145; HRMS (EI, m/z) Calcd for C12H11NO4:
233.06880. Found: 233.06893. Elem. Anal. (%) Calcd for C12H11-
NO4: C, 61.84; H, 4.76; N, 6.01. Found: C, 61.91; H, 4.83; N,
5.99.
General Procedure for the Synthesis of Lactones under
Thermal Conditions Followed by p-Toluenesulfonic Acid Mono-
hydrate using E-1 or Z-1 (10-19). A solution of E- or Z-
allylboronate 1 (100 mg, 0.42 mmol) and aldehyde (0.46 mmol) in
toluene (0.5 mL) was heated to 110 °C in a high-pressure vessel
under an Ar atmosphere for 72 h. p-TSA•H2O (230 mg, 1.2 mmol)
was then added, and the mixture was stirred overnight at room
temperature. The reaction was quenched with NaHCO3 (aq) (20
mL) and extracted with Et2O (3 × 20 mL). The combined extracts
were washed with brine (20 mL), dried over anhydrous Na2SO4,
filtered, and concentrated. Flash chromatography gave the corre-
sponding lactone.
General Procedure for the Synthesis of Butyric Acid Methyl
Esters (20-22). A solution of the corresponding E- or Z-
allylboronate (100 mg, 0.42 mmol) and aldehyde (0.46 mmol) in
toluene (0.5 mL) was heated at 95 °C under an Ar atmosphere for
42 h. The reaction was allowed to cool to room temperature, and
the solvent was removed. Flash chromatography gave the corre-
sponding methyl ester.
cis-4-Hydroxy-3-methyl-2-methylene-4-(4-nitrophenyl)-bu-
tyric Acid Methyl Ester (20a). Flash chromatography (20%
EtOAc/hexanes) yielded the product as a yellow oil in 69% yield.
(Note: the corresponding trans-lactone was isolated in 6% yield.)
1H NMR (400 MHz, CDCl3): δ 8.21-8.16 (m, 2H), 7.56-7.51
(m, 2H), 6.31 (d, 1H, J ) 0.8 Hz), 5.59 (t, 1H, J ) 0.9 Hz), 5.00
(t, 1H, J ) 3.4 Hz), 3.80 (s, 3H), 3.13 (ddddd, 1H, J ) 7.1, 7.1,
7.1, 3.4, 1.0 Hz), 2.75, (d, 1H, J ) 1.0 Hz), 0.98 (d, 3H, J ) 7.2
Hz); 13C NMR (100 MHz, CDCl3): δ 168.1, 150.1, 147.2, 141.8,
127.1, 127.0, 123.3, 74.6, 52.3, 42.9, 12.0; IR (cast film microscope,
cm-1): 3508, 2952, 1712, 1520, 1348; HRMS (EI, m/z) Calcd for
C12H12NO4 [M - OCH3]+: 234.07663. Found: 234.07614.
Conclusion
In summary, we have described a study of the substrate scope
for the triflic acid-catalyzed allylboration/lactonization reaction
and identified the presence of a unique reversal in observed
stereochemistry in many of the R-exo-methylene-γ-lactone
products. The nature of the aldehyde substrate is determinant
for the stereochemistry of the lactone products. We went on to
investigate the mechanism of this reaction process and confirmed
our previous suspicions that the lactonization was proceeding
via a carbocation-promoted mechanism. We showed that lactone
epimerization can occur in most cases, however, not in
substantially large enough amounts to account for the observed
diastereomeric ratios in the triflic acid-catalyzed allylboration
reaction. Furthermore, oxygen-18 labeling was used to track
the aldehyde oxygen throughout the reaction sequence and
indicated that none of the aldehyde oxygen was present in the
final lactone products. The main mechanism of this triflic acid-
catalyzed allylboration reaction involves the formation of a
carbocation intermediate from the initially formed open borate
product. This event is followed by trapping of the carbocation
by the neighboring ester, either before or after bond rotation
occurs, which leads to the observed diastereoselectivities in the
lactone products. Mechanistic possibilities to explain catalyst
turnover were discussed, and control experiments support a cycle
involving a nucleophilic attack on the ester’s alkoxy substi-
tuent.
J. Org. Chem, Vol. 72, No. 4, 2007 1283

Elford et al.
Acknowledgment. This work was funded by the Natural
Sciences and Engineering Research Council of Canada (NSERC),
by the University of Alberta, and by an AstraZeneca Award in
Chemistry to D.G.H. T.G.E. thanks the University of Alberta
Faculty of Graduate Studies and Research and the Alberta
Ingenuity Fund for graduate scholarships. We thank Glen Bigam
and Mark Miskolzie for discussions involving NMR analyses
and Dr. Hugo Lachance for helpful discussions on oxygen
labeling and lactonization experiments.
Supporting Information Available: Full experimental details,
spectral data, and copies of 1H and 13C NMR for all new compounds
10-12, 14, 16-22, and 26. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO062151B
1284 J. Org. Chem., Vol. 72, No. 4, 2007