Mechanistic studies on the O-directed free-radical hydrostannation of disubstituted acetylenes with Ph3SnH and Et3B and on the lodination of allylically oxygenated α-triphenylstannylalkenes

Article abstract of DOI:10.1021/ol051937d

(Chemical Equation Presented) The free-radical hydrostannation of 1 with Ph₃SnH and catalytic Et₃B in PhMe has been mechanistically probed. At high Ph₃SnH concentrations, the O-directed hydrostannation pathway dominates, a

Full text of DOI:10.1021/ol051937d

ORGANIC
LETTERS
2005
Vol. 7, No. 24
5377-5380
Mechanistic Studies on the O-Directed
Free-Radical Hydrostannation of
Disubstituted Acetylenes with Ph3SnH
and Et3B and on the Iodination of
Allylically Oxygenated
r
-Triphenylstannylalkenes
Paschalis Dimopoulos, Jonathan George, Derek A. Tocher,
Soraya Manaviazar, and Karl J. Hale*
The Christopher Ingold Laboratories, The Chemistry Department,
UniVersity College London, 20 Gordon Street, London WC1H 0AJ, United Kingdom
k.j.hale@ucl.ac.uk
Received August 10, 2005
ABSTRACT
The free-radical hydrostannation of 1 with Ph3SnH and catalytic Et3B in PhMe has been mechanistically probed. At high Ph3SnH concentrations,
the O-directed hydrostannation pathway dominates, and 2 is formed with good selectivity (ca. 11.1:1). Substantially lower stannane/substrate
concentrations increase the amount of tandem 5-exo-trig cyclization product 3 that is observed.
In this final paper of the series,1 we shall discuss the
mechanism of the O-directed free-radical hydrostannation
reaction in light of recent experiments that we have
performed with the propargylic alcohols 1 and 4 (Figure 1).
was significant in such systems, it should be quantifiable
with the reporter molecules 1 and 4, as these would generate
R-vinyl radical intermediates with a capacity to cyclize via
the 5-exo-trig pathway2 (see 15 in Scheme 2). A reliable
One aspect of the hydrostannation mechanism that we were
particularly keen to investigate was whether propargylically
oxygenated alkyl acetylenes would undergo significant, but
reversible, triphenylstannyl radical addition to the â-acety-
lenic carbon when subjected to the standard Ph3SnH/cat. Et3B
hydrostannation conditions. We reasoned that if â-addition
(1) For part 1, see: (a) Dimopoulos, P.; Athlan, A.; George, J.;
Manaviazar, S.; Walters, M.; Lazarides, L.; Aliev, A. E.; Hale, K. J. Org.
Lett. 2005, 7, 5369. For part 2, see: (b) Dimopoulos, P.; Athlan, A.;
Manaviazar, S.; Hale, K. J. Org. Lett. 2005, 7, 5373.
Figure 1. Free-radical hydrostannation probes 1 and 4.
10.1021/ol051937d CCC: $30.25
© 2005 American Chemical Society
Published on Web 10/26/2005

Scheme 1. Routes Used To Access Probes 1 and 4
Scheme 2. Concentration Effects in the Free Radical
Hydrostannation of 1 and 4 with Ph3SnH and Catalytic Et3B
estimate of the amounts of cyclization vs hydrostannation
products that were being generated from 1 and 4 at different
stannane and substrate concentrations might potentially
clarify the role of the propargylic O-atom in the O-directed
free radical hydrostannation mechanism, and simultaneously
give important new insights into the operational workings
of these reactions.
A detailed study of 1 and 4 might also allow experimental
conditions to be defined for performing the O-directed
hydrostannation process on propargyloxy substrates with a
substituted 3-butenyl grouping, without the occurrence of
5-exo-trig cyclization. The ability to control the hydrostan-
nation outcome in such systems would not only enhance the
overall utility of this reaction in organic synthesis, it would
also potentially allow it to be used in the total synthesis of
molecules such as allopumiliotoxin 334A,3 A83586C,4 stipi-
amide,5 and halichomycin.6
With these considerations in mind, we synthesized the
probe molecules 1 and 4 from the known aldol adduct 57 by
the routes shown in Scheme 1 (see the Supporting Informa-
tion for further details and literature references). Initially,
we subjected the propargylic alcohols 1 and 4 to our standard
free radical hydrostannation conditions (i.e., 1.5 equiv of Ph3-
SnH/0.1 equiv of Et3B, and PhMe [c of 1 and 4 ) 0.1 M] at
rt for 16 and 4 h, respectively). In the case of 1 (Scheme 2,
entry 1), compounds 2/3 were obtained in 67% yield and
3.5:1 ratio, alongside a tiny amount of 18. For alkynol 4
(Scheme 2, entry 5), a comparable ratio of the products 16/
17 (3.7:1) was obtained in almost identical yield (66%)
alongside a trace amount of 19. Similar results (4.5:1 of 2:3)
were observed when 1 was reacted with 4 equiv of Ph3SnH
and 0.1 equiv of Et3B in PhMe (c of 1 ) 0.1 M) for 5.5 h
at rt (entry 2).
Importantly, however, when 1 was subjected to free-radical
hydrostannation under much more concentrated conditions
(i.e., 6 equiv of Ph3SnH and 0.1 equiv of Et3B in PhMe (c
of 1 ) 1 M) at rt for 5.5 h), the ratio of 2/3 changed
substantially to 11.1:1 (entry 3). Significantly, the method
had now become preparatiVely useful for the synthesis of
reasonably pure R-triphenylstannyl alkenes with a substituted
3-butenyl grouping.
(2) Padwa, A.; Rashatasakhon, P.; Daut Ozdemir, A.; Willis, J. J. Org.
Chem. 2005, 70, 519.
By way of contrast, when the same hydrostannation was
performed under fairly dilute conditions, (i.e., 3 equiv of Ph3-
SnH/0.1 equiv of Et3B in PhMe [c of 1 in PhMe ) 0.01 M]
at rt for 24 h) a 1.7:1:0.6 ratio of 3/2/1 was obtained (Scheme
2, entry 4). Interestingly, compound 3 was now the pre-
dominant reaction product.
(3) (a) For recent synthetic work and references on the pumiliotoxins
see: Tang, X.-Q.; Montgomery, J. J. Am. Chem. Soc. 2000, 122, 6950. (b)
See also: Franklin, A.; Overman, L. E. Chem ReV. 1996, 96, 505.
(4) For our asymmetric total synthesis of the antitumor antibiotic
A83586C, see: Hale, K. J.; Cai, J. Chem. Commun. 1997, 2913.
(5) Stipiamide total synthesis: Andrus, M. B.; Lepore, S. D.; Turner, T.
M., J. Am. Chem. Soc. 1997, 119, 12159.
Given that high stannane concentrations markedly favored
addition of the Ph3Sn radical to the R-acetylenic carbon of
(6) Halichomycin synthetic studies: Hale, K. J.; Dimopoulos, P.; Cheung,
M. L. F.; Steed, J. W.; Levett, P. Org. Lett. 2002, 4, 897.
5378
Org. Lett., Vol. 7, No. 24, 2005

1, while low stannane concentrations marginally favored
attack on the more electron-rich â-acetylenic carbon, this
strongly pointed to high stannane concentrations promoting
coordination between the stannane and the R-directing
propargyl O-atom.
Scheme 3. Some of the Mechanistic Processes Operating in
the O-Directed Free-Radical Hydrostannation Mechanism
Since, under the “standard” hydrostannation conditions,
most propargylically oxygenated alkyl acetylenes generally
react with much higher levels of R/â regioselectivity than
do 1 and 4 (see Scheme 2, entries 1 and 5), this naturally
raised questions as to why these two alkynes should appear
to deviate so markedly. It is our belief that alkynes 1 and 4
do not behave differently from other propargylically oxygen-
ated alkyl acetylenes, with respect to the regiochemistry of
Ph3Sn• addition, under the standard conditions. Rather, we
contend that R- and â-triphenylstannylvinyl radical inter-
mediates have a very significant tendency to rapidly dis-
sociate back into either the starting alkyne and the Ph3Sn
radical, or an O-complexed Ph3Sn radical (see intermediate
22 in Scheme 3), and that an extremely complex and
reversible equilibrium8,9 is operational at many stages of these
reactions which, ultimately, favors accumulation of the
O-directed R-triphenylstannyl alkene products.
In the case of 1 and 4, because the 5-exo-trig vinyl radical
cyclization of 15 (R ) TBS or TBDPS) is quite facile
(Scheme 2), it is possible to detect and intercept this set
of fairly short-lived, and rapidly inverting, regioisomeric
â-triphenylstannylvinyl radicals, before they can fully revert.
The reporter molecules 1 and 4 thus enabled us to confirm
that Ph3Sn radicals are almost certainly adding to both
acetylenic carbons of most propargylically oxygenated alkyl-
acetylenes, under the standard experimental conditions we
use, but with a distinct preference for the R-carbon.
While it is possible to invoke that a reversible O-directed
equilibrium, between propargyloxystannyl radicals and in-
termediary vinyl radicals, is solely responsible for the
excellent ratios of R/â-triphenylstannylalkene regioisomers
that are observed in most reactions, we believe that such a
mechanistic picture would not only be highly simplistic, it
would also be rather illusory. Past work by Utimoto and
Oshima8 has demonstrated that a significantly enriched 9:1
mixture of (Z):(E)-1-triphenylstannyl-1-octene isomers can
be readily converted into a 4:1 (E):(Z) mixture by the action
of 0.2 equiv of Ph3SnH and 0.1 equiv of Et3B in hexane at
rt for 12 h. Their results,8 and those of others,10 thus indicate
that vinylstannane isomerization is almost certainly occurring
in most (if not all) alkyne free-radical hydrostannation
reactions.
that we believe are contributing to the excellent levels of
regio- and stereocontrol that are typically observed. In our
newly proposed scheme, the combination of moderately
enhanced Lewis acidity and greatly magnified steric effects
in the stannane, as well as multiple reversible additions and
eliminations of the complexed and uncomplexed triphenyl-
stannyl radical, are all suggested to contribute to the final
successful hydrostannation outcome.
We postulate that, under the “standard” hydrostannation
conditions, complexation of the stannane is significant, and
that H-atom abstraction occurs more readily from the
stannane of complex 21 than from Ph3SnH itself.11,12 We also
contend that 22 preferentially delivers its O-coordinated Ph3-
Sn radical to the R-alkyne carbon to give a mixture of rapidly
interconverting R-triphenylstannylvinyl radicals 23 and 24,
in which the preferred and more populated radical conformer
is 24 due to adverse allylic (A1,3) strain effects destabilizing
23. Although one can easily imagine that 24 will benefit from
In light of this combined data, we have now put forward
a mechanistic scheme (Scheme 3) for the O-directed free-
radical hydrostannation reaction of propargyloxy alkyl
acetylenes with Ph3SnH, which summarizes the main factors
(7) Hale, K. J.; Bhatia, G. S.; Peak, S. A.; Manaviazar, S. Tetrahedron
Lett. 1993, 34, 5343.
(8) Taniguchi, M.; Nozaki, K.; Miura, K.; Oshima, K.; Utimoto, K. Bull.
Chem. Soc. Jpn. 1992, 65, 349.
(9) Marco-Contelles, J.; Mainetti, E.; Fensterbank, L.; Malacria, M. Eur.
J. Org. Chem. 2003, 1759.
(10) Leusink, A. J.; Budding, H. A.; Drenth, W. J. Organomet. Chem.
1968, 11, 541.
Org. Lett., Vol. 7, No. 24, 2005
5379

a hyperconjugative interaction between the sp2 σ-radical and
the back lobe of the C-Sn σ-bond, recent ab initio
calculations on comparable â-silyl substituted vinyl radicals
suggest that â-silyl hyperconjugative stabilization is negli-
gible in such systems.13 By analogy, therefore, it is probably
correct to assume that hyperconjugative stabilization of 24
will also be minimal. It is now proposed that the exceedingly
bulky Ph3SnH preferentially donates its H-atom to the vinyl
radical conformer 24 rather than 23, since this will lead to
a far less sterically crowded transition state than the
alternative; this would, of course, preferentially deliver
vinylstannane 25.
Figure 2. Vinylstannanes 32-36.
It is further proposed that the majority of 26 that is formed
will have a very strong tendency to be converted into 25 by
the reversible addition of Ph3Sn radicals to the â-olefinic
carbon of 26.14 In this regard, molecular models of the
transition state for â-addition suggest that it will be far less
sterically crowded than that for R-addition; the â-addition
pathway would also produce a much less sterically encum-
bered tertiary radical 27 that would be doubly stabilized by
the R- and â-stannyl groups. By way of contrast, stannyl
radical addition to the R-carbon of 26 would lead to a much
more severely crowded transition state, and afford a less
stable (but, nevertheless, hyperconjugatively stabilized)
secondary carbon radical. Once generated, radical 27 has the
choice of either dissociating back into 26 or undergoing
immediate C-C bond rotation to place the two bulky Ph3-
Sn substituents as far apart as possible, as in radical rotamer
28; further bond rotation in the direction of rotamer 29 would
then allow the â-C-Sn bond to hyperconjugatively stabilize
the adjacent radical center. Elimination at this stage would
produce 25 and place the R group cis to the R-Ph3Sn group.
In our view, this mechanistic picture most satisfactorily
explains all of the observations made so far on this
remarkable reaction.
The angles subtended at the metal fell within the range
101.75° to 125.27°, with both extremes being found within
the crystal structure of 32. The nonbonded Sn‚‚‚O distances
generally fell within a relatively narrow range (3.09-3.44
Å). The only two exceptions were the structures for 33 and
34, where one molecule in the asymmetric unit had a
“normal” short contact, while the second independent
molecule adopted a different conformation with a much
greater Sn‚‚‚O separation (4.41 Å). Although most of these
distances are less than the sum of the van der Waals radii of
tin and oxygen (3.57 Å) they are much greater than a typical
Sn-O covalent bond distance (2.20-2.25 Å). Since there
are no significant distortions to the tetrahedral geometry of
Sn, we have concluded that, in the solid at least, there is no
evidence for valence expansion at the metal nor for a
consequential selective weakening of any of the aryl C-Sn
bonds within compounds 32-36. This behavior contrasts
with (Z)-disubstituted â-triarylstannylated allyl alcohols
where the central Sn atom is always distorted trigonal
bipyramidal with the apical aryl C-Sn bonds being signifi-
cantly weakened and more readily cleaved by I2. The lack
of internal Sn‚‚‚O coordination in allylically oxygenated
R-triphenylstannylalkenes means that they behave like
normal vinyltrialkylstannanes with respect to the metal-
halogen exchange reaction; in other words, they are converted
to vinyl iodides with retention of olefin geometry.
At this juncture, we would now like to change direction
and comment on the reason why allylically oxygenated R-
triphenylstannylalkenes react successfully with I2, while
allylically oxygenated (Z)-â-triphenylstannylalkenes do not.
To gain some insights into this behavior, we examined the
X-ray crystal structures of 32-36 (see the Supporting
Information). The five studies gave rise to 10 independent
determinations of geometry at the four-coordinate Sn atoms.
In closing, we believe that the mechanistic and structural
insights that we have provided will aid chemists wishing to
apply our methodology soon.
Acknowledgment. We thank the EPSRC (Project Grants
GR/N20959/01 and GR/S27733/01), the University of Lon-
don Central Research Fund, Novartis (Basel/USA), Merck
(Harlow), and Pfizer (Sandwich) for their generous financial
support and Profs. B.P. Roberts and A.G. Davies, FRS, of
UCL for helpful discussions.
(11) The significantly enhanced reactivity of trigonal bipyramidal,
heteroatom-complexed, tin hydrides towards Sn-H homolysis has already
been documented in the literature for the organostannatrane hydride, Me2N-
(CH2)3(Me)2SnH. The latter stannane does not even require any external
initiator to be added for it to mediate its free-radical reduction of
3-iodobenzoic acid to benzoic acid in H2O at rt. For more details, see: Han,
X.; Hartmann, G. A.; Brazzale, A.; Gaston, R. D. Tetrahedron Lett. 2001,
42, 5837.
(12) In light of this, one would expect that complexation of Ph3SnH with
the propargylic O-atom of 20 would likewise cause a significant elongation
and weakening of the Sn-H bond in the resulting complex 21; this would
clearly facilitate H-atom abstraction from its stannyl component as is being
proposed. It is also very reasonable to assume that O-coordinated triphenyltin
radicals such as 22 will have extra longevity than uncoordinated Ph3Sn
radicals due to magnified steric hindrance around the radical center; again,
this should favor the R-mode of addition.
Supporting Information Available: Full experimental
procedures and detailed spectral data, 500 MHz 1H and 125
MHz 13C spectra, and HRMS spectra for all new compounds
are provided, along with X-ray crystallographic data and
product ratio determinations in the hydrostannation reactions.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(13) Lalitha, S.; Chandrasekhar, J. Proc. Indian Acad. Sci. (Chem. Sci.)
1994, 106, 259.
(14) This isomerization will be especially favorable when R1 and R2 are
both alkyl groups due to significant A1, strain being present within 26.
OL051937D
3
5380
Org. Lett., Vol. 7, No. 24, 2005