Improvements of the total synthesis of monocillin I and radicicol via Miyaura-Suzuki couplings

Article abstract of DOI:10.1016/S0040-4039(02)00714-1

The palladium-catalyzed coupling of the vinyldisiamylborane, formed in situ, affords the isocoumarins which can be used further for the synthesis of radicicol and related macrolides, in 72% global overall yield from the alkyne. Advantages over related vinylboronates or vinyltin couplings are discussed in the case of the present examples.

Full text of DOI:10.1016/S0040-4039(02)00714-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 4003–4007
Improvements of the total synthesis of monocillin I and
radicicol via Miyaura–Suzuki couplings
Isabelle Tichkowsky and Robert Lett*
Unite´ Mixte CNRS-AVENTIS Pharma (UMR 26), 102, route de Noisy, 93235 Romainville, France
Received 5 April 2002; accepted 11 April 2002
Abstract—The palladium-catalyzed coupling of the vinyldisiamylborane, formed in situ, affords the isocoumarins which can be
used further for the synthesis of radicicol and related macrolides, in 72% global overall yield from the alkyne. Advantages over
related vinylboronates or vinyltin couplings are discussed in the case of the present examples. © 2002 Elsevier Science Ltd. All
rights reserved.
The convergent approach we developed for the first
total synthesis of monocillin I and radicicol involved
the palladium-catalyzed coupling of a functionalized
enantiopure vinylstannane and an appropriate
chloromethylisocoumarin as a key-step.1–3 Our scheme
was quite flexible for studying related Miyaura–Suzuki
couplings of vinyldialkylboranes 3 or vinyldialkyl-
boronates 4,4 according to Scheme 1.
presence of the epoxide and of the propargylic OMPM
ether which should also diminish—by its inductive
effect—the nucleophilic character of the triple bond; a
situation which to our knowledge had no precedent.
Moreover, it was particularly challenging to study
Miyaura–Suzuki couplings here, since we were aware of
the high sensitivity of the coupling product to mild
basic conditions: when trying to find suitable conditions
for the isocoumarin cleavage, even at rt, dilute NaOH
(0.01 M) or dilute NaHCO3 in H2O/DME (1/1) imme-
diately only induced degradation and the formation of
numerous products due mostly to the highly activated
methylene at the 3% position.1 On the other hand, there
was no appreciable degradation in strictly anhydrous
conditions with K2CO3 in THF or DME (reflux).
Boron a priori had the following advantages over tin
chemistry: regio- and stereospecificity of the hydrobora-
tion of terminal alkynes, much cheaper reagents, much
easier purification of the products and avoiding tin
contaminants. However, it was necessary to be able to
achieve the efficient hydroboration of the alkyne, in the
Scheme 1.
Keywords: macrolides; isocoumarins; boron and compounds; Suzuki reaction.
* Corresponding author. Tel.: 33-1 49 91 52 80; fax: 33-1 49 91 38 00; e-mail: robert.lett@aventis.com
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00714-1

4004
I. Tichkowsky, R. Lett / Tetrahedron Letters 43 (2002) 4003–4007
in quasi-stoichiometric conditions, affording the cou-
pling product in quite comparable yields, either with
vinyldicyclohexylboranes or the derived boronates
(after in situ oxidation by Me3NO). The best solvents
were found to be benzene, DME, 1,4-dioxane or AcOEt
in a 90/10 mixture with H2O.2
Therefore, in the case of the coupling of halides with
vinylic boron derivatives, the conditions developed just
before we began this work by Suzuki and co-workers,
using K2CO3 or K3PO4 as a base in THF or dioxane5
instead of NaOH or NaOR (R=Me, Et) which were
almost generally employed before, appeared to us to be
more promising for preserving the structure of the
coupling product. Since a base is required for the
Suzuki couplings4 and we checked that in the present
examples,2 an other important problem was to keep the
phenol OTBS protective groups, especially the para
4-OTBS which was optimal for the further completion
of the synthesis.1
Influence of the nature of the palladium catalyst was
studied for optimizing the coupling of the vinyl-dicyclo-
hexylboronate, generated in situ from the alkyne 5,
with the bromomethylisocoumarin 11 and some results
obtained in the conditions specified in Scheme 3 are
given in Table 1.
The required alkynes 5 and 10 were prepared from the
same enantiopure epoxyaldehyde 6 we used in our first
approach1 (Scheme 2). The condensation of 6 and 7, in
THF/hexane (3/2) (−35°C to 0°C) afforded the two
diastereoisomers, epimeric at 6%, in 93% yield (in a ratio
2/1); the reaction of 7 in the presence of CeCl3 (1
equiv.) was not stereoselective (89% yield, 1/1 ratio). As
we showed in the preceding communication, quite for-
tunately the two adducts can be used for the synthesis
of monocillin I and radicicol.3 Consequently, the whole
work described herein was achieved starting with the
2/1 mixture of adducts 8 and careful examination of the
isolated synthetic intermediates showed there was no
significant variation of this ratio throughout the whole
sequence. Model studies on simpler acetylenic epoxides
showed that catecholborane led to competitive epoxide
opening and that either dicyclohexylborane or disiamyl-
borane (generated in situ in THF from the olefin—
freshly redistilled over CaH2—and the BH3·Me2S
complex c10 M) were quite efficient for achieving the
regio- and stereospecific hydroboration of the terminal
alkyne with no epoxide opening. Hence, after prelimi-
nary optimization, one-pot couplings could be achieved
For the same one-pot coupling of the vinyldicyclohexyl-
boronate, using 1.5 mol% Pd2(dba)3 as the source of
palladium and in the presence of 4 equiv. of phosphine/
Pd, yields of the coupling product obtained in the
presence of BHT (or not) are given in Table 2. Con-
cerning the ligand,6 the whole data show that in this
case there is no direct correlation between the steric or
electronic effects and the efficiency of the coupling.
Ligands as different as PPh3, P(o-tolyl)3, (3-
MeOC6H4)3P, AsPh3, and dppf here give quite com-
Table 1. Pd catalytic system (3 mol%) and yield of 12
(Scheme 3)
Pd(PPh3)4
55–75%
48%
57%
56%
62%
62%
46%
56%
56%
PdCl2(PPh3)2+2DIBAH
Pd(OAc)2+2PPh3
Pd(OAc)2+4PPh3
PdCl2(CH3CN)2+2PPh3
PdCl2(PPh3)2
1/2Pd(dba)3+4PPh3
PdCl2dppf
[PhPd(mOH)(PPh3)2]
Scheme 2. Reagents and conditions: (a) 7 (1.5 equiv.) from trimethylsilylacetylene (2.0 equiv.) and n-BuLi (1.5 equiv.) in
THF/hexane (2/1), −60°C, then addition of 6 in THF at −35°C, THF/hexane (3/2), 2 h at −35°C and 12 h at 0°C; (b) K2CO3 (1
equiv.), MeOH, rt, 30 min; (c) BuLi (1 equiv.)/THF/hexane (1/1), −78°C, then MPMCl (1.5 equiv.), HMPA, 48 h, rt; (d) TBAF
1 M in THF (1.1 equiv.), rt, overnight.
Scheme 3.

I. Tichkowsky, R. Lett / Tetrahedron Letters 43 (2002) 4003–4007
4005
Table 2. Influence of the ligand, with 10 mol% BHT or
without BHT, on the yield of 12 with 3 mol% Pd catalyst
(Scheme 3)
only 16 was produced; use of a weaker base also
afforded only 16 (Scheme 4). No solution to this prob-
lem was found with the use of the bromomethylisocou-
marin 15.
Ligand
Yield
Yield
Time
(no BHT)
(10% BHT)
The role of BHT8 was shown to be quite complex and
not only related to the eventual quench of radicals
formed in the presence of traces of oxygen or due to the
autoxidation of organoboranes: addition of BHT some-
times improved the coupling or had no effect; in fact,
the coupling in the presence of other quenching agents
of oxygenated radicals such as duroquinone or
galvinoxyl (0.1 equiv.) occurred in lower yield than the
same coupling with no BHT or even with no argon
atmosphere (at solvent reflux); on the other hand,
addition of 10 mol% p-dinitrobenzene or 1 equiv.
nitrobenzene had no significant effect on the yield of
the coupling.2 We observed however that addition of
BHT clearly diminished the formation of 13,2 14 always
being a minor by-product (less than 3%); compounds
which were also isolated in some couplings of the
analogous vinyl-tributylstannanes.1a Thus, for the cou-
pling corresponding to Scheme 3 with Pd2(dba)3 and
AsPh3 (Table 2), 13 was isolated in 11% and 20–25%
yield, respectively, in the presence of 10 mol% BHT and
in the absence of BHT, when consistently 12 was
obtained in 50% and 30–35% yield. On the other hand,
in the case of the vinyl-dicyclohexylboronates generated
in situ, the yield of the desired coupling product was
sometimes limited in some reaction conditions by the
competitive formation of 17, resulting not from a radi-
cal coupling but from a palladium-catalyzed homocou-
pling of the vinylboron component,9 as clearly shown
by the high stereoselectivity of its formation since only
the E,E-diene was formed (1H NMR at 400 MHz, and
13C); moreover we showed that 17 was formed in the
same reaction conditions, in the absence of the isocou-
marin, and that the formation of 17 was not dimin-
ished, but even significantly enhanced with 10 mol%
PhMe2P
Ph2MeP
Ph3P
13%
Trace
12 h
12 h
60–90 min
50 min
30 min
1 h
1 h
1 h
5–10 min
2 h
12 h
46%
40%
33%
35%
35%
44%
36%
51%
(4-ClC6H4)3P
(4-MeOC6H4)3P
(4-Me2NC6H4)3P
(3-ClC6H4)3P
(3-MeOC6H4)3P
(2,4,6-(MeO)3C6H2)3P
(2-MeC6H4)3P
Ph2PCH2CH2PPh2
(2-Furyl)3P
35%
50%
18%
31%
25%
31%
50%
12 h
12 h
12 h
15 min
a
(3-SO3C6H4)3PNa3
(PhO)3P
Ph3As
36%
a In this case benzene/H2O 75/25.
parable yields, showing that there is no real rate deter-
mining step in the catalytic cycle, that is no great
energy differences between the catalytic species involved
in the different steps. Addition of CuI (1 equiv.)7 in the
presence of 3 mol% Pd(PPh3)4, keeping otherwise the
same conditions as in Scheme 3, led to a faster cou-
pling, but to unstable catalytic species, thus affording
12 in 15% yield after 2 h with other by-products.
Quite comparable yields of coupling product were
obtained with either the bromomethylisocoumarin 11
or the chloromethylisocoumarin 2, as we already
noticed for the couplings of the related vinylstan-
nanes.1–3 However, full deprotection of the TBS protec-
tive groups of the phenols occurred in the same
conditions with the chloromethyl–isocoumarin 2, and
Scheme 4.

4006
I. Tichkowsky, R. Lett / Tetrahedron Letters 43 (2002) 4003–4007
BHT (Scheme 5). These same experiments also showed
that protonolysis in the reaction conditions was not
important.
boronates; thus, after 30 min in the conditions
described in Scheme 6, we could isolate the desired
isocoumarins 19 and 20 in, respectively, 58 and 14%
yield (72% global yield), in which the para-phenol was
still protected as a TBS ether; the mixture of 19 and 20
as such could be used further in the synthesis, since the
next two-step sequence for the isocoumarin cleavage
(DIBAH reduction, NaClO2 oxidation) leads to a spe-
cific total deprotection of the ortho-phenol TBS ether
and that this free phenol is required for favoring the
macrolactonization in Mitsunobu reaction conditions.1–3
In order to try to diminish that homocoupling reaction,
we decided to examine more hindered boranes for
making the oxidative addition of Pd° more difficult
than on the previous vinyldicyclohexylboronates. Tak-
ing also in account the results we obtained in prelimi-
nary studies on simpler models for related (10%-desoxy)
vinyldicyclohexyl-boranes and -boronates (10%-desoxy),2
we thus studied related couplings with vinyldisiamyl-
boron derivatives. After some optimization, the cou-
plings described in Scheme 6 could be achieved and, in
those experiments, only traces of 17 were observed.
Quite interestingly, the one-pot coupling could be suc-
cessfully achieved on the deprotected alcohol with the
vinyldisiamylborane generated in situ from 10; this was
fortunate since oxidation by Me3NO of the correspond-
ing vinyldicyclohexylborane proved to be troublesome
and the desired coupling product was then not
obtained.2 Therefore, we could avoid the difficult prob-
lem of the specific deprotection of the 10%-OTBS ether
in the presence of the phenol OTBS ethers which would
have been required otherwise for the completion of the
synthesis, moreover for finding conditions preserving
the structure of the coupling product. During our
study, in the present examples, we observed that in fact
the couplings of those vinyldisiamylboranes are also
faster than those of the related vinyldicyclohexyl-
In conclusion, for the examples studied herein, we
showed that vinyldisiamylboranes have the following
advantages over the corresponding vinyldicyclohexyl-
boranes or boronates: minimization of the formation of
the diene 17 resulting from a palladium-catalyzed
homocoupling and hence higher yields of the desired
product in quasi-stoichiometric conditions, phenol TBS
protective groups can be preserved, suppression of the
oxidation step of the borane into the boronate by
Me3NO (which moreover here allows to achieve the
coupling with the vinyldisiamylborane derived from the
free 10%-OH alkyne 10 in a high yield). Hence, the
modification described herein affords the desired iso-
coumarins for further completion of the synthesis in
45% overall yield from the same epoxy-aldehyde 6,
instead of 33% overall yield via the sequence developed
previously with tin chemistry1–3 (uncorrected yields).
Scheme 5.
Scheme 6.

I. Tichkowsky, R. Lett / Tetrahedron Letters 43 (2002) 4003–4007
4007
Acknowledgements
2. Tichkowsky, I. Ph.D. Thesis, Paris VI University,
1996.
We thank the staff of the Analytical Department of the
Research Center at Romainville, Roussel Uclaf and the
CNRS for a PhD grant to I. Tichkowsky, and the
Direction des Recherches Chimiques of Roussel-Uclaf
for support of this work.
3. Tichkowsky, I.; Lett, R. Tetrahedron Lett. 2002, 43,
3997–4001.
4. Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457–
2483.
5. (a) Abe, S.; Miyaura, N.; Suzuki, A. Bull. Chem. Soc.
Jpn. 1992, 65, 2863–2865; (b) Ishiyama, T.; Abe, S.;
Miyaura, N.; Suzuki, A. Chem. Lett. 1992, 691–694.
6. Farina, V.; Krishnan, B. J. Am. Chem. Soc. 1991, 113,
9585–9595.
References
7. Farina, V.; Kapadia, S.; Krishnan, B.; Wang, C.; Liebe-
skind, L. S. J. Org. Chem. 1994, 59, 5905–5911.
8. (a) Farina, V.; Baker, S. R.; Benigni, D. A.; Hauck, S.
I.; Sapino, C. J. Org. Chem. 1990, 55, 5833–5847; (b)
Farina, V.; Krishnan, B.; Marshall, D. R.; Roth, G. P.
J. Org. Chem. 1993, 58, 5434–5444.
1. (a) Lampilas, M. Ph.D. Thesis, Paris VI University,
1991; (b) Lampilas, M.; Lett, R. Tetrahedron Lett. 1992,
33, 773–776; (c) Lampilas, M.; Lett, R. Tetrahedron
Lett. 1992, 33, 777–780; (d) Lett, R.; Lampilas, M.;
Tichkowsky, I. In Recent Progress in the Chemical Syn-
thesis of Antibiotics and Related Microbial Products;
Lukacs, G., Ed.; Springer Verlag, 1993; Vol. 2, pp. 99–
120.
9. Miyaura, N.; Suzuki, A. Main Group Met. Chem. 1987,
10, 295–300.