Stereopure 1,3-butadiene-2-carboxylates and their conversion into non-racemic α-alkylidenebutyrolactone natural products by asymmetric dihydroxylation

Article abstract of DOI:10.1016/S0040-4039(01)00598-6

Dienoic esters 1 with the four possible permutations of the C=C configurations were prepared in two steps via non-stereoselective aldol additions followed by stereospecific β-eliminations. Sharpless dihydroxylations of these esters yielded natural and unnatural α-alkylidene-β-hydroxybutyrolactones 2. Among these were synthetic dihydromahubanolide B (cis,Z-2a), isodihydromahubanolide B (cis,E-2a) and, for the first time, litsenolide D₁ (ent-trans,Z-2b) and the enantiomer trans,E-2b of litsenolide D₂. Competitively, dihydroxyesters 10 were formed.

Full text of DOI:10.1016/S0040-4039(01)00598-6

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 3967–3971
Stereopure 1,3-butadiene-2-carboxylates and their conversion into
non-racemic a-alkylidenebutyrolactone natural products by
asymmetric dihydroxylation
Christian Harcken and Reinhard Bru¨ckner*
Institut fu¨r Organische Chemie und Biochemie, Albert-Ludwigs-Universita¨t, Albertstraße 21, D-79104 Freiburg, Germany
Received 26 February 2001; accepted 2 April 2001
Abstract—Dienoic esters 1 with the four possible permutations of the CꢀC configurations were prepared in two steps via
non-stereoselective aldol additions followed by stereospecific b-eliminations. Sharpless dihydroxylations of these esters yielded
natural and unnatural a-alkylidene-b-hydroxybutyrolactones 2. Among these were synthetic dihydromahubanolide B (cis,Z-2a),
isodihydromahubanolide B (cis,E-2a) and, for the first time, litsenolide D1 (ent-trans,Z-2b) and the enantiomer trans,E-2b of
litsenolide D2. Competitively, dihydroxyesters 10 were formed. © 2001 Elsevier Science Ltd. All rights reserved.
Naturally occurring a-alkylidene-b-hydroxylactones 2
occur with all conceivable substitution patterns cis,E-2,
R
R
cis,Z-2, trans,E-2 and trans,Z-2 (Scheme 1). They com-
prise (−)-isodihydromahubanolide B (cis,E-2a), (−)-
trans,E-1a-c
trans,Z-1a-c
CO2Me
CO2Me
dihydromahubanolide
B (cis,Z-2a), (+)-isodihydro-
AD mix-α
AD mix-α
mahubanolide A (trans,E-2a), (−)-litsenolide D2 (ent-
trans,E-2b), (−)-litsenolide C2 (ent-trans,E-2c), (+)-
dihydromahubanolide A (trans,Z-2a), (−)-litsenolide D1
(ent-trans,Z-2b), (−)-litsenolide C1 (ent-trans,Z-2c), and
analogous compounds containing CꢀC or CꢀC bonds
in the side-chain R. The (iso)dihydromahubanolides
stem from the wood of Clinostemon mahuba in Brazil1
and the litsenolides from the leaves of Litsea japonica in
Japan.2
OH
R
OH
R
cis,Z-2a-c
cis,E-2a-c
O
O
O
O
inversion of configuration?
We have shown that the asymmetric dihydroxylation
(‘AD’)3 of trans-configurated b,g-unsaturated esters
leads, via cis-configurated b-hydroxy-g-lactones,4 to
structurally diverse, optically active g-lactones.5 We
now studied whether the a,b%,b,g-unsaturated esters 1
are dihydroxylated at the CbꢀCg bond in a similar
fashion (Scheme 1). This could furnish the four
diastereomers of type-2 lactones in optically active form
directly or after Mitsunobu inversion of the allylic OH
group. Both in the trans- and the cis-configurated esters
1 the intended site of attack of the AD reagent was a
disubstituted CꢀC bond (which is—differently substi-
OH
OH
R
R
trans,E-2a-c
trans,Z-2a-c
O
O
O
O
AD mix-β
AD mix-β
R
R
cis,Z-1a-c
cis,E-1a-c
CO2Me
CO2Me
Keywords: asymmetric synthesis; elimination reactions; esters;
hydroxylation; lactones.
* Corresponding author. E-mail: reinhard.brueckner@organik.
chemie.uni-freiburg.de
a: R = nC15H31; b: R = nC11H23; c: R = nC13H27
Scheme 1.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)00598-6

3968
C. Harcken, R. Bru¨ckner / Tetrahedron Letters 42 (2001) 3967–3971
at the acceptor-free, i.e. CbꢀCg, bond.6 Fortunately, this
12
is not forcedly so: Two syntheses of racemic litsenolide
C2 (trans,E-2c) via the OsO4/NMO-mediated dihydroxy-
lation of the dienoic ester cis,E-7 (Scheme 2) relied
upon the same CbꢀCg selectivity which we required.7
a)
b)
OH
CO2Et
CO2Et
3
cis,syn-4
+
The first objective of this study8 was to obtain the
isomerically pure a,b%,b,g-unsaturated esters trans,E-1
and trans,Z-1 (Scheme 3).9 As the first step, the lithium
enolate of ester trans-8 was hydroxyalkylated with hex-
adecanal. The 55:45 mixture of the trans,syn- and
trans,anti-diastereomers of the resulting b-hydroxyester
9a was separated by flash chromatography on silica
gel,10 the minor diastereomer preceding the major. Sim-
ilarly, the aldol addition between lithiated ester trans-8
and dodecanal furnished the b-hydroxyesters trans,syn-
9b and trans,anti-9b as a readily separable 66:34 mix-
ture. Being unaware of comprehensive 1H NMR
analyses of all diastereomers of an unsaturated hydroxy-
cis,anti-4
S
S
12
12
β´
β
α
c1)
c2)
γ
Li2(NC)Cu
Li2(NC)Cu
12
CO2Et
CO2Et
cis,E-7
5
6
OH
12
d)
rac-litsenolide C2
(trans,E-2c)
3
ester like compound 9b, pertinent l1-H and JH,H values
O
are collected in Table 1.11
O
Scheme 2. (a) LDA, HMPA/THF, −78°C; RCHO.7b (b)
MsCl, NEt3, CH2Cl2; KH, THF, 0“20°C; 92%.7b (c1) HCꢀ
CCO2Et; (c2) cis-BrHCꢀCHMe, Pd(PPh3)4 (cat.).7a (d) OsO4
(20 mol%), NMO, t-BuOH/acetone/H2O, 20°C, 1 day; HCl, 4
h; 66%.7b
Dehydrations of b-hydroxyesters can be syn- or anti-
selective but mixed mechanisms or subsequent isomer-
izations
can
interfere
with
obtaining
good
stereocontrol. This is a notorious problem when the
Z-configurated, less stable a,b-unsaturated ester is to be
formed—no matter whether by a syn-elimination from
a syn-configurated b-hydroxyester12 or by an anti-elimi-
nation from its anti-isomer.13 We achieved highly anti-
selective dehydrations of all hydroxyesters trans,syn-
tuted—usually dihydroxylated with 90–99% ee in the
trans-series and with 20–80% ee in the cis-series3).
However, AD reactions of esters akin to 1 may occur at
the acceptor-substituted, i.e. CaꢀCb%, bond rather than
trans-8
CO2Me
n
β
n
β
a)
α
α
trans,syn-9a,b
trans,E-1a,b
β´
trans,anti-9a,b
trans,Z-1a,b
β´
OH
OH
γ
γ
CO2Me
b)
CO2Me
c)
n
β
β
β´
β´
n
γ
γ
CO2Me
f) or g)
CO2Me
d) or e)
a: n = 14; b: n = 10
OH
OH
OH
OH
n
+
+
n
HO CO2Mne
HO CO2Mne
O
O
O
O
cis,E-2a,b
trans,anti-10a,b (or enantiomer)
trans,syn-10a,b (or enantiomer)
cis,Z-2a,b
Scheme 3. (a) LDA (1.1 equiv.), THF, −78°C, 30 min; RCHO (0.95 equiv.), “0°C, 30 min; 47% trans,syn-9a, 38% trans,anti-9a;
57% trans,syn-9b, 29% trans,anti-9b. (b) PPh3 (2.0 equiv.), DEAD (2.0 equiv.), THF, −40“20°C, 5 h; a: 91% (E/Z 98:2); b: 86%
(E/Z 98:2). (c) Same as (b); a: 92% (Z/E>98:2); b: 82% (Z/E 98:2). (d) AD-mix a™ (1.4 g/mmol), MeSO2NH2 (1.0 equiv.),
t-BuOH/H2O 1:1, 0°C, 6 days; 21% cis,E-2a (45% b.o.r.s.m.; 78% ee), 10% trans,anti-10a (21% b.o.r.s.m.), 53% trans,E-1a; 28%
cis,E-2b, 7% trans,anti-10b. (e) K3Fe(CN)6 (3.0 equiv.), K2CO3 (3.0 equiv.), (DHQ)2PHAL (10 mol%), K2OsO3(OH)2 (2.0 mol%),
MeSO2NH2 (1.0 equiv.), t-BuOH/H2O (1:1), 0°C, 24 h; 28% cis,E-2a (68% ee), 8% trans,anti-10a. (f) Same as (d); 4 days; 49%
cis,Z-2a, 29% trans,syn-10a; 65% cis,Z-2b, 11% trans,syn-10b. (g) Same as (e); 24% cis,Z-2a, 18% trans,syn-10a.

C. Harcken, R. Bru¨ckner / Tetrahedron Letters 42 (2001) 3967–3971
Table 1. Selected H NMR data (CDCl3, 300 MHz) of aldol adducts 9b
3969
1
Compound
lg-H
lb-H
la-H
lb%-H
Jg,b
Jb,a
Ja,b
trans,syn-9b
trans,anti-9b
cis,syn-9b
5.67
5.65
5.82
5.71
5.55
5.41
5.57
5.38
3.01
3.05
3.43
3.43
3.86
3.78
3.93
3.83
15.4
15.5
10.7
10.8
9.0
9.0
10.0
9.1
4.7
8.3
4.4
9.1
cis,anti-9b
and trans,anti-9a,b by treatment with excess PPh3/
diethyl azodicarboxylate (THF, −40°C“rt; Scheme
3):14 b-hydroxyesters trans,syn-9a,b yielded dienoic
esters trans,E-1a,b (86–91%, E/Z 98:2) while b-hydroxy-
esters trans,anti-9a,b provided esters trans,Z-1a,b (82–
92%, Z/E]98:2). The configuration of the newly
formed CꢀC bond was deduced from l3-H=6.60 (Z) or
5.78 (E).
cis,E-2a was 78%, according to chiral HPLC, when
formed under standard AD conditions [21% yield; 45%
based on recovered starting material (‘b.o.r.s.m.’)] or
68% using the modified AD procedure (28% yield).
From the specific rotation [h]D=−38.2 for a 76% ee
90:10 mixture of hydroxylactones cis,Z-/cis,E-2a16 and
the specific rotation [h]D=−93.3 of natural cis,E-2a,1b
one deduces [h]D:(−38.2+0.1×0.76×93.3)/(0.9×0.76)=
−45.5 for cis,Z-2a.17 Our specimen of cis,Z-2a pos-
sessed [h]D=−30 so that one infers ee:(−30)/(−45.5)=
66%.
Dienoic esters 1a,b were dihydroxylated using AD-mix
a™ and dienoic esters 1a additionally by using 10
mol% (DHQ)2PHAL (instead of 1 mol% in AD-mix
a™) and 2.0 mol% K2OsO3(OH)2 (instead of 0.2 mol%
in AD-mix a™) (Scheme 3). Such an ‘improved
procedure’15 had increased the ee of the AD of another
substrate containing a methyl-substituted CꢀC bond
from 80 to 94%.5e The a-alkylidene-b-hydroxylactones
2a,b resulted in 21–65% yields. Competing AD of the
CaꢀCb bond furnished dihydroxyesters 10a,b in 7–18%
yield. Sparing solubility of the highly lipophilic sub-
strates 1 in the polar reaction mixture contributed to
the sub-optimal yields; for example, we re-isolated ester
trans,E-1a in up to 53% yield. The ee of hydroxylactone
Overall, optically active isodihydromahubanolide B
(cis,E-2a) and dihydromahubanolide B (cis,Z-2a) were
obtained after only three synthetic steps. This compares
well with previous syntheses which had led to a cis,Z-/
cis,E-2a mixture (separable only by preparative TLC;
14 steps from dimethyl
cis,Z-2a (13 steps from
L
-tartrate18), to both cis,E- and
D
-glucose bisacetonide19), to a
90:10 cis,Z-/cis,E-2a mixture (76% ee; seven steps from
1-octadecyne16) or to a 75:25 cis,E-/cis,Z-2a mixture
(eight steps from ethyl S-lactate20).
CO2Me
trans-11
10
10
β
β
a)
α
α
γ
γ
β´
β´
cis,syn-9b
cis,E-1b
OH
OH cis,anti-9b
CO2Me
b)
CO2Me
c)
10
β
β
β´
γ
γ
β´
cis,Z-1b
10
CO2Me
d) or e)
CO2Me
f) or g)
OH
OH
OH
OH
10
+
+
10
10
10
O
HO CO2Me
O
HO CO2Me
O
O
ent-trans,Z-2b
trans,E-2b
cis,anti-10b (or enantiomer)
cis,syn-10b (or enantiomer)
Scheme 4. (a) LDA (1.23 equiv.), THF/DMPU 8:1, −78°C, 30 min; dodecanal (1.4 equiv.), “0°C, 1 h; 49% cis,syn-9b, 38%
cis,anti-9b. (b) PPh3 (2.0 equiv.), DEAD (2.0 equiv.), THF, −20“20°C, 4 h; 81% (E/Z>99:1, cis/trans 97:3). (c) Same as (b); 83%
(Z/E>99:1, cis/trans 95:5). (d) AD-mix a™ (1.4 g/mmol), MeSO2NH2 (1.0 equiv.), t-BuOH/H2O 1:1, 5°C, 7 days; 22% trans,E-2b
(47% b.o.r.s.m.; 28% ee), 15% cis,anti-10b (32% b.o.r.s.m.), 53% cis,E-1b. (e) K3Fe(CN)6 (3.0 equiv.), K2CO3 (3.0 equiv.),
(DHQ)2PHAL (10 mol%), K2OsO3(OH)2 (2.0 mol%), MeSO2NH2 (1.0 equiv.), t-BuOH/H2O (1:1), 0°C, 24 h; 15% trans,E-2b
(36% ee), 14% cis,anti-10b. (f) Same as (d); 28% trans,Z-2b (43% b.o.r.s.m.; 16% ee), 15% cis,syn-10b (23% b.o.r.s.m.), 34%
cis,Z-1b. (g) Same as (e); 30% trans,Z-2b (16% ee), 6% cis,syn-10b.

3970
C. Harcken, R. Bru¨ckner / Tetrahedron Letters 42 (2001) 3967–3971
The OH group of lactone cis,E-2b stayed inert under
Mitsunobu conditions21 even after prolonged stirring (2
days) at elevated temperature (80°C). Treatment with
PBu3, tetramethyl azodicarboxamide and PhCO2H22
furnished 47% of a 87:13 E/Z-mixture of the 1,2-elimi-
nation product, i.e. a-dodecylidene-g-methyl-D4-
butenolide. Attempted SN reactions with the mesylate
derived from cis,E-2b gave mixtures of the S%N-product
(KO2, DMSO, rt, 12 h23) and the 1,4-elimination
product (KOAc, 18-crown-6, DMF, 0°C, 8 h24). Lac-
tone cis,Z-2b resisted attempts of configurational inver-
sion, too.
859.
3. Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B.
Chem. Rev. 1994, 94, 2483–2547.
4. Precedents: (a) Wang, Z.-M.; Zhang, X.-L.; Sharpless, K.
B.; Sinha, S. C.; Sinha-Bagchi, A.; Keinan, E. Tetra-
hedron Lett. 1992, 33, 6407–6410; (b) Miyazaki, Y.;
Hotta, H.; Sato, F. Tetrahedron Lett. 1994, 35, 4389–
4392.
5. (a) Harcken, C.; Bru¨ckner, R. Angew. Chem. 1997, 109,
2866–2868; Angew. Chem., Int. Ed. Engl. 1997, 36, 2750–
2752; (b) Harcken, C.; Bru¨ckner, R.; Rank, E. Chem.
Eur. J. 1998, 4, 2342–2352 (corrigendum ibid. 2390); (c)
Berkenbusch, T.; Bru¨ckner, R. Tetrahedron 1998, 54,
11461–11470; (d) Berkenbusch, T.; Bru¨ckner, R. Tetra-
hedron 1998, 54, 11471–11480; (e) Harcken, C.; Bru¨ckner,
R. New J. Chem. 2001, 40–54; (f) Harcken, C.; Bru¨ckner,
R. Synlett 2001, 718–721.
Circumventing this inertia, a-alkylidene-b-hydroxylac-
tones trans,E-2b and trans,Z-2b were synthesized from
ester trans-11 by the strategy of Scheme 3, i.e. by an
aldol addition (1H NMR data: Table 1) followed by
stereospecific b-eliminations and standard or modified
ADs (Scheme 4). Even the PPh3/DEAD-mediated dehy-
dration delivering the hindered dienoic ester cis,Z-1b
was highly anti-selective (“>99:1 Z/E- and 95:5 cis/
trans-ratios). The AD reactions of cis,E- and cis,Z-1b
produced hydroxylactones trans,E- and ent-trans,Z-2b
in only 15–22 and 28–30% yield, respectively. This
seems to be caused by lacking solubility—53% dienoic
ester cis,E-1b and 34% cis,Z-1b were re-isolated—and
by competing ADs of the CaꢀCb bond—6–15% dihy-
droxyesters 10a,b were found. According to chiral
HPLC, the ee of lactone trans,E-2b was 28–36% and
the ee of ent-trans,Z-2b 16%.25 The discrepancy
between the formation of (+)-litsenolide D2 (trans,E-2b)
from cis,E-1b and AD-mix a™ and Sharpless’
mnemonic guideline for the side-selectivity of this reac-
6. AD of the CaꢀCb% bond: Nicolaou, K. C.; Yue, E. W.; La
Greca, S.; Nadin, A.; Yang, Z.; Leresche, J. E.; Tsuri, T.;
Naniwa, Y.; De Riccardis, F. Chem. Eur. J. 1995, 1,
467–494.
7. (a) Li, Y.; Chen, Z.-X.; Huang, J.-M.; Huang, J.-X.; Xu,
Z.-H. Gaodeng Xuexiao Huaxue Xuebao 1997, 18, 914–
916 (Chem. Abs. 1997, 127, 220509); (b) Kende, A. S.;
Toder, B. H. J. Org. Chem. 1982, 47, 167–169.
8. Harcken, C. Doktorarbeit; Universita¨t Go¨ttingen: Ger-
many, 2000.
9. All new compounds gave satisfactory 1H NMR and IR
spectra as well as correct combustion analyses.
10. Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43,
2923–2925.
11. In the analogously configurated hydroxyesters 9 (n=0)
Ja,b% decreases in the same order cis,anti-9 (9.2 Hz)>
trans,anti-9 (8.4 Hz)>trans,syn-9 (5.2 Hz)>cis,syn-9 (4.8
Hz) as in our hydroxyesters 9a,b: Galatsis, P.; Millan, S.
D.; Nechala, P.; Ferguson, G. J. Org. Chem. 1994, 59,
6643–6651.
1
tion should be noted.3 By their H NMR spectra and
the signs of their specific rotations, lactones ent-
trans,Z-2b (levorotatory) and trans,E-2b (dextrorota-
tory) were identical with natural (−)-litsenolide D1 and
the mirror image of natural (−)-litsenolide D2, respec-
tively. These compounds were thus prepared in non-
racemic form for the first time, the straightforwardness
of our three-step route being attractive in view of
step-requirements between 8 and 13 of the previous
syntheses.26
12. Complete Z-selectivity: Corey, E. J.; Letavic, M. A. J.
Am. Chem. Soc. 1995, 117, 9616–9617 (corrigendum ibid.
12017).
13. Incomplete Z-selectivity: Knight, S. D.; Overman, L. E.;
Pairaudeau, G. J. Am. Chem. Soc. 1995, 117, 5776–5788.
14. Procedure: (a) Bradbury, R. H.; Walker, K. A. M. J.
Org. Chem. 1983, 48, 1741–1750; b-eliminations with
PPh3/DEAD from our group: (b) von der Ohe, F.;
Bru¨ckner, R. Tetrahedron Lett. 1998, 39, 1909–1910; (c)
von der Ohe, F.; Bru¨ckner, R. New J. Chem. 2000,
659–669; (d) Ref. 5e.
Acknowledgements
15. (a) Bennani, Y. L.; Sharpless, K. B. Tetrahedron Lett.
1993, 34, 2079–2082; (b) Blundell, P.; Ganguly, A. K.;
Girijavallabhan, V. M. Synlett 1994, 263–265.
16. Adam, W.; Klug, P. Synthesis 1994, 567–572.
17. No [h]D for natural cis,Z-2a was reported in Ref. 1b.
Ref. 18 gives [h]D=−37 for synthetic cis,Z-2a. Based on
that value our synthetic cis,Z-2a has ee:(−30)/(−37)=
81%.
This work was supported by the Deutsche Forschungs-
gemeinschaft. We are indebted to Dr. Olivier Lohse
(Novartis Pharma AG, Basel) for the determination of
the ee’s by chiral HPLC.
References
18. Tamaka, A.; Yamashita, K. Chem. Lett. 1981, 319–322.
19. Wood, W. W.; Watson, G. M. J. Chem. Soc., Perkin
Trans. 1 1987, 2681–2688 (steps 1–6) and J. Chem. Soc.,
Perkin Trans. 1 1990, 3201–3203 (steps 7–13).
20. Adam, W.; Renze, J.; Wirth, T. J. Org. Chem. 1998, 63,
226–227.
1. (a) Martinez V., J. C.; Yoshida, M.; Gottlieb, O. R. Tet-
rahedron Lett. 1979, 1021–1024; (b) Martinez V., J. C.;
Yoshida, M.; Gottlieb, O. R. Phytochemistry 1981, 20,
459–464.
2. (a) Takeda, K.; Sakurawi, K.; Ishii, H. Tetrahedron 1972,
28, 3757–3766; (b) Tanaka, H.; Nakamura, T.; Ichino,
K.; Ito, K.; Tanaka, T. Phytochemistry 1990, 29, 857–
21. Reviews: (a) Mitsunobu, O. Synthesis 1981, 1–28; (b)
Hughes, D. L. Org. React. 1992, 42, 335–656.

C. Harcken, R. Bru¨ckner / Tetrahedron Letters 42 (2001) 3967–3971
3971
22. Method: Tsunoda, T.; Yamamiya, Y.; Kawamura, Y.;
Itoˆ, S. Tetrahedron Lett. 1995, 36, 2529–2530.
23. Method: (a) San Filippo, J.; Chern, C.-I.; Valentine, J. S.
J. Org. Chem. 1975, 40, 1678–1680; (b) Corey, E. J.;
Nicolaou, K. C.; Shibasaki, M.; Machida, Y.; Shiner, C.
S. Tetrahedron Lett. 1975, 37, 3183–3186.
analogously configurated but C2-elongated and/or more
unsaturated litsenolides A2 (Ref. 2a), B2 (Ref. 2a), C2
(Ref. 2a), and E2 (Ref. 2b), show −40.4<[h]D<−45.2. The
specific rotation [h]D=−7.3 measured for synthetic ent-
trans,Z-2b (16% ee according to HPLC) corresponds to
[h]D=−46 for enantiopure material. This deviates from
the value [h]D=−8.4 of the natural product (Ref. 2b),
too.
24. Procedure: Dener, J. M.; Hart, D. J.; Ramesh, S. J. Org.
Chem. 1988, 53, 6022–6030.
25. The specific rotation [h]D=13.8 measured for synthetic
trans,E-2b (28% ee according to HPLC) implies [h]D=
49.2 for the pure enantiomer. The absolute value dis-
agrees from [h]D=−21.6 reported for the natural product
(Ref. 2b). The latter must be questioned since the
26. (a) Chen, S.-Y.; Joullie´, M. M. Tetrahedron Lett. 1983,
24, 5027–5030; (b) Wakabayashi, S.; Ogawa, H.; Ueno,
N.; Kunieda, N.; Mandai, T.; Nokami, J. Chem. Lett.
1987, 875–878; (c) Wood, W. W.; Watson, G. M. J.
Chem. Soc., Perkin Trans. 1 1987, 2681–2688.
.