Template Macrolactonization of Trichloroethyl Ester Derivatives Catalyzed by Potassium Salts

Full text of DOI:10.1021/jo970942v

J . Org. Chem. 1998, 63, 2715-2718
2715
Alkali metal binding studies and molecular modeling
indicate that macrolides such as 1 and 2 are well-suited
to bind potassium ions;4 thus, one might expect alkali
metal salts to facilitate macrocyclization via a substrate
templating effect. As precedent, the templating effect of
alkali metal hydroxides in the formation of crown ethers
is well-established,11 and high yields in macrolactoniza-
tions of cesium ω-halocarboxylates are thought to be the
result of the templating ability of cesium.12 However,
there have been no reports of the templating effect of
alkali metal ions on the lactonization of ω-hydroxy acid
derivatives.5 We wish to report here that a variety of
alkali metal salts promote macrolactonizations to give 1
or 2 under very mild conditions.
Tem p la te Ma cr ola cton iza tion of
Tr ich lor oeth yl Ester Der iva tives Ca ta lyzed
by P ota ssiu m Sa lts
Steven D. Burke,* Todd S. McDermott, and
Christopher J . O’Donnell
Department of Chemistry, University of
WisconsinsMadison, Madison, Wisconsin 53706
Received May 28, 1997
In tr od u ction
We have initiated a program involving the synthesis
and study of a variety of unnatural hydropyran-contain-
ing macrocycles1-4 for applications to alkali metal ion
sequestration and transport.5-7 We have recently re-
ported syntheses,1,2 solid-state structures,3,4 variable-
temperature NMR behavior, and alkali metal binding
affinity4 of a number of hydropyran-containing cyclic
oligolides. Macrocycles 1 and 2 are examples of this class
of ionophores, both available from the corresponding
ω-hydroxy acids.
Resu lts a n d Discu ssion
During the synthesis of the second-generation trisami-
nomethyl-appended macrolide 1, a serendipitous result
revealed the ability of alkali metal salts to promote
macrolactonization. Acyclic trimer alcohol 3a (R1 ) CH2-
NHBoc, R2 ) Me, R3dCH2CCl3) was allowed to stand in
methylene chloride over a small amount of aqueous K2-
CO3 during workup. After several hours, the sample had
been converted to material that was identified as mac-
rocycle 1, isolated in 83% yield, along with a small,
undetermined amount of the bicyclic lactone 4a (R1
CH2NHBoc, R2 ) Me, eq 1).
)
The methods we have used in the past for formation
of the 18-membered ring employed high dilution condi-
tions, with carboxylate activation using a carbodiimide-
DMAP protocol (Keck-Steglich coupling)8,9 or involve-
ment of the 2,4,6-trichlorobenzoyl mixed anhydride
(Yamaguchi coupling).10 Although these macrolacton-
ization methods are generally successful, they are com-
plicated by the formation of side products, the necessity
to activate the carboxylic acid group, and the requirement
of high dilution or syringe-pump methods.
Although the inadvertent conditions for this first
cyclization were difficult to reproduce, further investiga-
tion was indicated. Ester 3a (R1 ) CH2NHBoc, R2 ) Me,
R3 ) CH2CCl3) was thus subjected to a variety of cycli-
zation conditions as summarized in Table 1. Treatment
of the trichloroethyl ester 3a with excess solid alkali
metal salt (used without prior purification or drying) was
found to give the most reliable results. A survey of
common solvents indicated that tetrahydrofuran (THF)
is best suited for this reaction (Table 1, entries 1-4). The
cyclizations gave decreasing yields of 1 as the concentra-
tion was increased (Table 1, entries 4-7), and signifi-
cantly more side products were observed. It is notewor-
thy that the use of the smaller alkali metal cations,
Na2CO3 and Li2CO3 (Table 1, entries 8 and 9), gave no
detectable cyclization product, while the use of the larger
(1) Burke, S. D.; O’Donnell, C. J .; Hans, J . J .; Moon, C. W.; Ng, R.
A.; Adkins, T. W.; Packard, G. K. Tetrahedron Lett. 1997, 38, 2593-
2596.
(2) Burke, S. D.; Heap, C. R.; Porter, W. J .; Song, Y. Tetrahedron
Lett. 1996, 37, 343-346.
(3) Burke, S. D.; O’Donnell, C. J .; Porter, W. J .; Song, Y.; Powell, D.
R.; Hayashi, R. K. J . Chem. Soc., Chem. Commun., submitted for
publication.
(4) Burke, S. D.; O’Donnell, C. J .; Porter, W. J .; Song, Y. J . Am.
Chem. Soc. 1995, 117, 12649-12650.
(5) Dietrich, B.; Viout, P.; Lehn, J .-M. Macrocyclic Chemistry:
Aspects of Organic and Inorganic Supramolecular Chemistry; VCH:
Weinheim, Germany, 1993.
(6) Cram, D. J . Angew. Chem., Int. Ed. Engl. 1988, 27, 1009-1112.
(7) Schneider, H.-J . Angew. Chem., Int. Ed. Engl. 1991, 30, 1417-
1436.
(8) Boden, E. P.; Keck, G. E. J . Org. Chem. 1985, 50, 2394-2395.
(9) Neises, B.; Steglich, W. Angew. Chem., Int. Ed. Engl. 1978, 17,
522-524.
(10) Inanaga, J .; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi, M.
Bull. Chem. Soc. J pn. 1979, 52, 1989-1993.
(11) Ercolani, G.; Mandolini, L.; Masci, B. J . Am. Chem. Soc. 1981,
103, 2780-2782 and references therein.
(12) Kruizinga, W. H.; Kellogg, R. M. J . Am. Chem. Soc. 1981, 103,
5183-5189.
S0022-3263(97)00942-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/19/1998

2716 J . Org. Chem., Vol. 63, No. 8, 1998
Notes
Ta ble 1. Cycliza tion of 3a (R1 ) CH2NHBoc, R2 ) Me,
R3 ) CH2CCl3) To Give Ma cr ocycle 1a
Ta ble 2. Alk a li Meta l Bin d in g Affin ity of Tr ich lor oeth yl
Ester 3a (R1 ) CH2NHBoc, R2 ) Me, R3 ) CH2CCl3) a n d
Ma cr olid e 1
concn
(M)
time
(d)
% yield
entry
saltb
solventc
of 1d,e
Ka (M-1)a, -∆G°b (kcal/mol)
f
f
1
2
3
4
5
6
7
8
9
10
11
12
13
14
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
K2CO3
CH2Cl2
PhH
0.002
0.002
0.002
0.002
0.005
0.01
5
1
1
1
2
1
1
4
4
8 h
3
3 h
4
3
74 (10)
64 (11)
38 (62)
88 (12)
69 (7)i
60 (10)j
24 (26)j
0k
hostc
Li+
Na+
K+
Cs+
3a
1
3.9 × 104; 6.2 2.4 × 104; 6.0 5.5 × 104; 6.4 1.2 × 104; 5.5
CH3CNg
THFh
THF
3.5 × 104; 6.2 7.1 × 104; 6.6 2.2 × 106; 8.6 1.5 × 104; 5.7
a
b
Association constant. Free energy of association. c There was
no evidence of macrolide formation upon exposure of 3a to alkali
metal picrate salts.
THF
THF
THF
THF
THF
THF
THF
THF
0.1
f
Li2CO3
0.002
0.002
0.002
0.002
0.002
0.002
0.002
Na2CO3
Cs2CO3
KOAc
KI
0k
52 (20)
47 (0)
65 (0)
0k
Ta ble 3. Cycliza tion of Ester s 3a -d (R1 ) CH2NHBoc,
R2 ) Me, R3 ) Tr ich lor oeth yl, Me, Allyl, Ben zyl)
% yielda
NaI
time
(d)
KBPh4
THF
0l
entry
R3
1
4
5
a
b
All reactions were carried out at 30 °C. 10 equiv of solid salt
was used without prior purification, unless otherwise noted.
c Solvents were dried in the normal way unless otherwise noted.
1
2
3
4
CH2CCl3 (3a )
Me (3b)
allyl (3c)
1
88
18
39
37
12
50
28
16
2b
5
20
17
17
Isolated yield of macrocycle 1. e Numbers in parentheses indicate
d
benzyl (3d )
5
isolated yield of bicyclic lactone 4a . f 20 equiv of solid salt used.
g
h
i
a
b
Used without prior drying. Tetrahydrofuran. Seco acid ob-
Isolated yields. 24 h at 35 °C, 24 h at reflux.
served by TLC. j Trichloroethyl ester hydrolysis products isolated.
k
l
Starting material only. No macrocycle present by TLC after 3
days.
As was the case in earlier picrate binding studies,4
macrolide 1 selectively binds potassium over the other
alkali metal cations tested. The trichloroethyl ester 3a
has a relatively weak affinity for and little selectivity
between alkali metal cations. Not all complexes between
3a and metal cations are likely to have structures that
place the ester and hydroxyl groups in proximity. Tem-
plated transesterification via a productive binding con-
formation between 3a and K+ is most likely because of
the substantial preference (>2 kcal/mol) for K+‚1 over
the other M+‚1 complexes. In addition, the product
macrocycle will more effectively solubilize K+ salts over
those of Li+, Na+,, or Cs+. In general, potassium salts
gave the highest yields of macrolide product (Table 1).
We next studied the effect of different esters on the
cyclization reaction. A series of esters 3b-d (R1 ) CH2-
NHBoc, R2 ) Me, R3 ) Me, allyl, benzyl) were prepared
and exposed to the optimized cyclization conditions for
trichloroethyl ester 3a (eq 2 and Table 3). Esters 3b-d
alkali metal cation, Cs2CO3 (Table 1, entry 10), led to
decreased yield. Finally, the effect of the anion was
briefly investigated. The use of the less basic KOAc and
KI (Table 1, entries 11 and 12) resulted in diminished
yield of macrocycle. However, it should be noted that
none of the bicyclic lactone side product 4a was detected
in these cases. The use of NaI (Table 1, entry 13)
resulted in complete recovery of starting material, and
the use of a potassium salt with a nonbasic counterion,
KBPh4, resulted in no cyclization after 3 days (Table 1,
entry 14). It should be noted that, with the exception of
NaI, all of the salts used are essentially insoluble in THF
and are thus present in solution in catalytic amounts.
Some additional comment is necessary about the effect
of concentration on the cyclization reaction. A cyclization
that truly proceeds by the templating effect of a cation
would be expected to tolerate higher concentrations.13
This is clearly not the case here (Table 1, entries 4-7).
As the concentration of the reaction increases, the
isolated yield of macrocycle 1 decreases, due principally
to competing formation of bicyclic lactone 4a or hydrolysis
of the trichloroethyl ester. The low solubility of potas-
sium carbonate in THF ensures a low and constant
concentration relative to that of the substrate in these
entries. The side reaction leading to 4a , involving base-
induced intramolecular transacylation, would be acceler-
ated by increased substrate concentration, whereas the
templated cyclization would not. It is noteworthy that
increased substrate concentration does not lead to higher
order oligomer formation, further supporting the asser-
tion that a template effect is operative.13
were less reactive, requiring higher temperatures or
longer reaction times for complete consumption of start-
ing material. In addition to macrolide 1 and bicyclic
lactone 4a , the monomer hydroxy esters 5b-d (R1 ) CH2-
NHBoc, R2 ) Me) were isolated from the reaction
mixtures. A competition experiment between 3a (R3 )
CH2CCl3) and 3b (R3 ) Me) in the presence of phenan-
threne as an internal standard demonstrated the differ-
ent rates of cyclization for these esters. Treatment of the
mixture with K2CO3 in THF at 30 °C for 39 h resulted in
complete conversion of 3a to macrocycle 1, while leaving
Another issue raised by the data in Table 1 is that of
cation selectivity. Alkali metal picrate extraction stud-
ies4,14,15 were carried out to determine the cation binding
affinity of hydroxy ester 3a and macrolide 1 (Table 2).
(13) Galli, C.; Mandolini, L. J . Chem. Soc., Chem. Commun. 1982,
251-253.
1
3b unchanged by H NMR analysis (85% recovery after
purification).16
(14) Moore, S. S.; Tarnowski, T. L.; Newcomb, M.; Cram, D. J . J .
Am. Chem. Soc. 1977, 99, 6398-6405.
The possibility exists that the alkali metal cations are
simply acting as Lewis acids and thus promoting trans-
(15) Koenig, K. E.; Lein, G. M.; Stuckler, P.; Kaneda, T.; Cram, D.
J . J . Am. Chem. Soc. 1979, 101, 3553-3566.

Notes
J . Org. Chem., Vol. 63, No. 8, 1998 2717
Ta ble 4. Cycliza tion of 3e (R1 ) R2 ) H, R3 ) CH2CCl3)
Ta ble 5. Alk a li Meta l Bin d in g Affin ity
to Ma cr ocycle 2 (Eq 3)
Ka (M-1)a; -∆G°b (kcal/mol)
salt (10 equiv)
hostc
Na+
K+
3e
8 2
(3)
THF (0.002 M), rt
R1 ) R2 ) H,
3e
2
4.9 × 104; 6.4
3.6 × 104; 6.2
7.2 × 105; 8.0
2.0 × 108; 11.3
4.6 × 104; 6.3
R3 ) CH2CCl3
cis-dicyclohexano-
18-crown-6d
time
(h)
yieldb
(%)
salta
a
b
Association constant. Free energy of association. c There was
entry
no evidence of macrolide formation upon exposure of 3e to
potassium or sodium picrate. See ref 15.
1
2
3
4
5
6
7
8
9
K2CO3
Na2CO3
Cs2CO3
KOAc
KI
2
36
1
1
2
10 min
18
24
24
24
25
22
90
73c
93
96
81
90
10e
0e
d
Ta ble 6. Cycliza tion of 3e (R1 ) R2 ) H, R3 ) CH2CCl3)
w ith a Ca ta lytic Am ou n t of Aceta te Sa lt
NaId
-
M
+
AcO
LiId
KBF4
NaBF4
KBPh4
NaBPh4
KPF6
3e
THF, rt 8 2
(4)
R1 ) R2 ) H,
0e
R3 ) CH2CCl3
10
11
12
0e
d
0e,f
0e
concn
(M)
18-C-6a
(equiv)
timeb
(h)
% yield
entry
M+ (equiv)
of 2c
a
b
Used without prior purification or drying. Isolated yield of
1
K (0.5)
K (0.5)
K (0.05)
Bu4N (0.5)
Bu4N (0.5)
0.002
0.002
0.05e
0.002
0.002
1
24
1
8
18
100
97
macrocycle 2. c Starting material and bicyclic lactone 4e (R1 ) R2
2
1
d
) H) were observed by crude NMR. Completely soluble in THF.
3d
4
87
e Unreacted starting material was recovered. f Unidentified de-
composition product was noted by TLC.
tracef
tracef
5
1
a
b
cis-Dicyclohexano-18-crown-6. The reactions were allowed to
esterification, as has been demonstrated for a different
system.18 If the cation’s only function was as a Lewis
acid and the templating effect was unnecessary, inter-
as well as intramolecular transesterifications should be
facilitated.13 However, attempts to promote intermolecu-
lar transesterifications under similar conditions, even in
the presence of excess methanol, were unsuccessful.
Formation of the unsubstituted oligolide 2 (R1 ) R2 )
H) from the corresponding trichloroethyl ester 3e (R1 )
R2 ) H, R3 ) CH2CCl3) was then investigated, as sum-
marized in Table 4 (eq 3). The optimal conditions for
macrocycle 1 also gave an excellent yield of macrocycle
2 (Table 4, entry 1). The reaction time necessary for
consumption of starting material was considerably shorter
for the unsubstituted case. The use of the smaller sodium
cation, Na2CO3, resulted in a sluggish reaction, giving
73% of 2 after 36 h, along with bicyclic lactone 4e (R1 )
R2 ) H) and unreacted starting material (Table 4, entry
2). The larger cesium cation, Cs2CO3, showed reactivity
similar to that of K2CO3 (Table 4, entry 3). The use of
the less basic KOAc (Table 4, entry 4) gave fast, clean
conversion to the desired macrolide. The use of alkali
metal iodides (Table 4, entries 5-7) gave unexpected
results. KI behaved similarly to potassium salts in earlier
entries, giving macrocycle 2 in good yield after about 2
h, and NaI gave a 90% yield of macrolide 2 in 10 min.
Our prior experience was that sodium salts were less
effective than the corresponding potassium salts. How-
ever, this discrepancy may be partially explained (see
below) by the fact that NaI is completely soluble in THF
at the concentration of the reaction, where KI is much
less soluble. Solubility of the alkali metal salt is not
sufficient to guarantee success, as evidenced by the fact
that LiI results in only 10% conversion to macrocycle
after 18 h and NaBPh4 results in no reaction after 25 h
(Table 4, entry 11). A series of potassium and sodium
proceed until starting material was gone by TLC. c Isolated yield
of macrolide 2. This experiment was carried out on 1.2 mmol
d
(0.68 g) of trimer alcohol 3e. e This was the highest concentration
at which trimer alcohol 3e was completely soluble in THF. f NMR
analysis of the crude reaction mixture showed a small amount of
macrolide 2 along with bicyclic lactone 4e and unidentified
degradation products.
salts with nonbasic counterions (Table 4, entries 8-12)
routinely returned starting material after 24 h. The
reaction conditions that typically gave the highest yield
of macrocycle 2 in the shortest amount of time with a
simple filtration workup were excess KOAc in THF at
room temperature for about 1 h (Table 4, entry 4).
Table 5 lists the sodium and potassium picrate binding
affinities of unsubstituted ester 3e (R1 ) R2 ) H, R3 )
CH2CCl3) and macrolide 2 and the potassium picrate
binding affinity of cis-dicyclohexano-18-crown-6.15 The
lack of substitution in 2 removes one of the major
conformational control elements present in macrolide 1,
resulting in a less preorganized binding array.4 The
lower overall binding affinity for K+ and selectivity for
K+/Na+ (compare Tables 2 and 5) is a reflection of the
increased flexibility of macrolide 2 relative to macrolide
1.4 Presumably, this increased flexibility applies as well
to the substrate 3e, allowing productive template com-
plexes with a range of cation sizes.
On the basis of the optimized conditions from Table 4,
the experiments presented in Table 6 (eq 4) provide
informative data regarding the cyclization reaction.
First, the cyclization proceeds to completion with subs-
toichiometric amounts of the potassium salt (Table 6,
entries 1 and 3), indicating that the macrocycle, while
selective for binding potassium (see Table 5), is not a
sufficiently strong binder to effectively remove it from
the reaction mixture. In the presence of cis-dicyclohex-
ano-18-crown-6, a more effective potassium binder,15 the
reaction is substantially inhibited (Table 6, entry 2). The
cyclization reaction was found to tolerate higher concen-
tration and less catalyst (Table 6, entry 3); e.g., 680 mg
of trimer alcohol 3e was converted to macrocycle 2 in 87%
yield using 6 mg of KOAc in 24 mL of THF. Replacement
(16) The details of the competition experiment are available with
the Supporting Information for this paper.
(17) Koskikallio, J . In The Chemistry of Carboxylic Acids and Esters;
Patai, S., Ed.; Interscience: New York, 1969; pp 103-136.
(18) Pregel, M. J .; Dunn, E. J .; Nagelkerke, R.; Thatcher, G. R. J .;
Bruncel, E. Chem. Soc. Rev. 1995, 24, 449-455.

2718 J . Org. Chem., Vol. 63, No. 8, 1998
Notes
of K+ with a nontemplating cation (Table 6, entries 4 and
5) provided only a small amount of macrocycle 2, along
with bicyclic lactone 4e and other degradation products.
To determine whether the polydentate coordinating
ability of the substrates 3a and 3e is necessary for
successful cyclization, trichloroethyl 17-hydroxyhepta-
decanoate (6) was prepared19and subjected to several
lactonization attempts. In all cases, NMR analysis of the
crude reaction mixtures showed only starting material
after heating for extended periods (eq 5). These experi-
ments indicate that the trichloroethyl ester group is not
sufficiently activated to under go an intra- or intermo-
lecular transesterification reaction under these condi-
tions.
spectra were recorded at 75 MHz. Where indicated, distortion-
less enhancement by polarization transfer (DEPT) was used to
assign carbon resonances as CH3, CH2, CH, or C. Chemical
shifts are reported in ppm relative to Me4Si (δ 0.00). Elemental
analyses were performed by Desert Analytics Laboratories
(Phoenix, AZ).
Gen er a l P r oced u r e for Ma cr ola cton iza tion . To a solu-
tion of acyclic trimer alcohol 3 in THF (0.002 M) was added the
alkali metal salt (10 equiv). The reaction flask was flushed with
N2 and capped, and the mixture was stirred at the indicated
temperature. The progress of the reaction was followed by TLC
analysis until the starting material was consumed. The mixture
was filtered and the solid washed with ether or CH2Cl2 and
concentrated. The crude reaction mixture was purified by silica
gel chromatography.
Ma cr olid e 1: Rf 0.45 (5% MeOH in CH2Cl2); 1H NMR (CD3-
OD, 250 MHz, 85 °C) δ 6.08 (ddd, 3H, J ) 10.3, 5.7, 2.4 Hz),
5.71 (br d, 3H, J ) 10.1 Hz), 5.35-5.29 (m, 3H), 4.55-4.52 (m,
3H), 4.41 (d, 3H, J ) 3.4 Hz), 3.66-3.55 (m, 3H), 3.40 (dd, 3H,
J ) 14.5, 2.9 Hz), 2.58-2.48 (m, 3H), 1.42 (s, 27H), 1.05 (d, 9H,
J ) 7.0 Hz); 13C NMR (CD3OD, 75 MHz) δ 171.1 (C), 158.4 (C),
134.0 (CH), 124.6 (CH), 80.6 (C), 77.8 (CH), 77.5 (CH), 75.4 (CH),
39.0 (CH2), 33.3(CH), 28.8(CH3 × 3), 15.3(CH3); IR (thin film)
3408 (br), 2976, 1741, 1709, 1506, 1367, 1286, 1170 cm-1; MS
(FAB) m/e (relative intensity, assignment) 982.3 (60, M + Cs+),
888.4 (90, M + K+), 872.4 (75, M + Na+), 850.4 (10, M + H+),
750.4 (100, M + H+ - C4H8 - CO2), 694.3 (30, M + H+ - 2C4H8
Su m m a r y a n d Con clu sion
We have described macrolactone formation from a
hydroxy ester utilizing the templating effect of alkali
metal cations to facilitate an intramolecular transesteri-
fication reaction. In most cases, the rate-limiting step
of a transesterification is the addition of the alcohol or
alkoxide fragment to the carbonyl carbon, followed by fast
breakdown of the resulting tetrahedral intermediate.17
The data presented herein suggest that the conversions
of 3a to 1 and 3e to 2 are M+-templated cyclizations
wherein the collapse of the tetrahedral intermediate is
the slow step.20 With the proximity of the hydroxyl and
ester moieties enforced and the associated entropic costs
paid by the template effect, the formation of the tetra-
hedral intermediate would be facilitated for all substrates
3a -d . The rate of the productive collapse of this
intermediate would be dependent upon the departing
alkoxide nucleofuge, thus the superiority of the trichlo-
roethyl esters. Macrolactonization of ester 3e was suc-
cessful with a catalytic amount of KOAc and was
inhibited by the addition of cis-dicyclohexano-18-crown-
6. Unfunctionalized hydroxy trichloroethyl ester 6,
incapable of being templated, was completely inert to the
cyclization conditions. All of these observations are
consistent with a templated cyclization mechanism.
- CO2); [R]22 +218.0 (c)3.9, CH2Cl2).
D
Bicyclic la cton e 4a (R1 ) CH2NHBoc, R2 ) Me): mp 110-
1
112 °C; Rf 0.26 (25% EtOAc in hexanes); H NMR (CDCl3, 300
MHz) δ 5.89 (br d, 1H, J ) 10.3 Hz), 5.72 (ddd, 1H, J ) 10.3,
3.7, 2.6 Hz), 5.00-4.95 (m, 1H), 4.72 (ddd, 1H, J ) 8.1, 4.0, 4.0
Hz), 4.43 (d, 1H, J ) 5.9 Hz), 4.35-4.33 (m, 1H), 3.58-3.49 (m,
1H), 2.97-2.82 (m, 2H), 1.40 (s, 9H), 1.04 (d, 3H, J ) 7.7 Hz);
13C NMR (CDCl3, 75 MHz) δ 166.5 (C), 155.8 (C), 134.6 (CH),
120.9 (CH), 81.5 (CH), 80.0 (C), 73.6 (CH), 66.3 (CH), 41.3 (CH2),
32.7 (CH), 28.3 (CH3), 15.2 (CH3); IR (thin film) 3256 (br), 2978,
1745, 1711, 1514, 1367, 1229, 1170 cm-1; MS (FAB) m/e (relative
intensity, assignment) 306.2 (100, M + Na+), 250.1 (40, M +
Na+ - C4H8), 228.2 (45, M + H+ - C4H8); [R]22 +70.6 (c ) 0.6,
D
CH2Cl2).
1
Ma cr olid e 2: mp 80-87 °C; Rf 0.25 (2% MeOH/CH2Cl2); H
NMR (300 MHz, CDCl3) δ 6.02-5.94 (m, 1H), 5.58-5.51 (m, 1H),
4.60-4.52 (m, 2H), 4.25 (dd, J ) 10.7, 3.7 Hz, 1H), 3.95 (dd,
J ) 12.1, 9.2 Hz, 1H), 2.47-2.20 (m, 2H); 13C NMR (75 MHz,
CDCl3) δ 170.8 (C), 127.0 (CH), 124.8 (CH), 73.6 (CH), 72.5 (CH),
66.6 (CH2), 27.7 (CH2); IR (neat) 2947, 1734, 1298, 1184, 1099,
727 cm-1; MS (FAB) m/e 420.1 (M + H+), 443.1 (M + Na+), 460.1
(M+K+); [R]20 +181° (c ) 1.53, CHCl3). Anal. Calcd for
D
C21H24O9: C, 60.00; H, 5.75. Found: C, 59.48; H, 5.95.
Exp er im en ta l Section
Ack n ow led gm en t. This research was supported by
NIH Grant No. GM 28321 (S.D.B.), NIH Chemistry/
Biology Interface Training Grant GM 08505 (C.J .O.),
and NIH Postdoctoral Fellowship GM 17009-01A1
(T.S.M.). The University of Wisconsin NMR facility
receives financial support from NSF (CHE-9208463) and
NIH (1S10 RR0 8389-D1).
Gen er a l Meth od s. Unless otherwise noted, all reagents
were purchased from commercial suppliers and used without
further purification. Tetrahydrofuran (THF) and benzene were
distilled from sodium/benzophenone immediately prior to use.
Methylene chloride (CH2Cl2) was distilled from CaH2 im-
mediately prior to use. Silica gel chromatography was performed
according to the method of Still.21 Proton nuclear magnetic
resonance (1H NMR) spectra were recorded at 300 MHz.
Chemical shifts are reported in parts per million (ppm, δ)
relative to Me4Si (δ 0.00), and coupling constants (J ) are reported
in hertz (Hz). Carbon-13 nuclear magnetic resonance (13C NMR)
Su p p or tin g In for m a tion Ava ila ble: Copies of 1H and 13C
NMR spectra of compounds 1, 2, 3a , 3e, and 4e; details of the
syntheses of ester derivatives 3b-d (Table 3); and details of
competition experiment between 3a and 3b are presented (18
pages). This material is contained in libraries on microfiche,
immediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
(19) Richard, M. A.; Deutch, J .; Whitesides, G. M. J . Am. Chem.
Soc. 1978, 100, 6613-6625.
(20) For an example of transesterification where the second mecha-
nistic step, also dependent upon the pKa of the departing group, is
rate-limiting; see: Breslow, R.; Chung, S. Tetrahedron Lett. 1990, 31,
631-634.
(21) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.
J O970942V