Communications
Scheme 3. RRCM with allylmalonate as the relay activator. HG2=sec-
ond-generation Hoveyda–Grubbs initiator; [Ru=CH(o-
isopropoxyPh)(Cl)2(H2IMes)]. SM=starting material.
Scheme 4. Mechanism for the relay stage of RRCM. [Ru*]=
Ru(Cl)2(H2IMes)Ln.
polyenes 8, which contain the relay subunits A–C (Scheme 2),
the rate of the critical relay event (a/a’) was too slow to
compete with the undesired macrocyclizations. This led us to
study the allylmalonate ester derivatives 9a and 9b
(Scheme 3). Although we had used allylmalonate ester
derivatives in our initial demonstrations of RRCM,[2] and
dimethyl diallylmalonate itself has served as an important test
substrate for many aspects of olefin metathesis,[4] this
structural subunit had not yet been explored as an expendable
relay-activator moiety. Treatment of 9a, which contains a
terminal n-butyl group on C8’, with HG2 in hot methylene
chloride resulted in the formation of the two RRCM products
10a and 10b in a 4:1 ratio (Scheme 3, tabular inset).
Pleasingly, the allylmalonate ester 9b, which contains a
moieties in the metal-ligand sphere reduces the avidity of
Ru–alkene binding. This scenario would result in a faster rate
of decomplexation of the cyclopentene from 11d relative to
the analogous events in the possible reactions of 8a–e. The
practical advantage of the allylmalonate relay activator has
been demonstrated here and we believe that this subunit
could potentially have wider utility as a tool to combat
macrocyclization[7,8] or truncation[2] processes that have
plagued the application of the RRCM strategy.[2]
We performed additional studies to investigate the ease of
ejection of the relay alkene from the ruthenium coordination
sphere. The compound 12 produced the hindered Z-alkene
14[2,9,10] along with several by-products upon treatment with
the first-generation Grubbs precatalyst (Scheme 5). These by-
products were isolated and shown to be the macrocyclic
dienes (E)-13, (Z)-13, and 15. Interestingly, 15 is a constitu-
tional isomer of 13 in which the allylic methyl substituent has
moved from C9 to C7. We suggest that diene 15 was produced
through the reinsertion process indicated by 16a–d. Wenzel
and Grubbs have reported the dynamic nature of an alkene
=
C8 C8’ terminal alkene and is therefore a direct analogue
of 8a–e, cyclized to give 10b, in an 88% yield.[5] Surprisingly,
=
we did not detect the symmetrical C8 C8 dimer what would
arise from the cross-metathesis of two molecules of 10 in
either of these experiments.
There are two possible explanations for the greater
efficiency of the RRCM with the allylmalonate ester deriv-
atives 9a–b (Scheme 3) compared with the substrates 8a–e
(Scheme 2). First,
a gem-dimethyl-like effect (Thorpe–
Ingold) should accelerate the initial stage of the relay event
for the substrates 9a–b (kon for 11a!11b, Scheme 4). Second,
ejection of the dimethyl cyclopentene-1,1-dicarboxylate
moiety from the coordination sphere of the ruthenium
center (11d!11e, Scheme 4) should be faster than the
corresponding decomplexation of the relay by-products for
the substrates 8a–e. This second point deserves further
explanation.
The reversible nature of several important steps in olefin
metathesis, including dissociation of the product alkene from
the metal center, has been examined in detail by Piers and co-
workers.[6] We suggest that the rate-determining step in some
RRCM reactions is the decomplexation of the relay alkene, in
this case a cyclopentene derivative from 11d (i.e., koff for 11d,
=
or an associative counterpart involving the C16 C16’
Scheme 5. Rearranged macrocylic diene 15 that was formed during the
RRCM reaction of compound 12. G1=first-generation Grubbs initia-
tor; [Ru=CHPh(Cl)2(PCy3)2].
alkene).[4] Here we suggest that the steric repulsion between
the bulky geminal dicarboxylate groups and the other
2142
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 2141 –2143