Rhodium carbenoid-initiated Claisen rearrangement: scope and mechanistic observations.

Article abstract of DOI:10.1021/ol990697x

[formula: see text] It has been shown that alpha-diazoketones react with allylic alcohols in the presence of Rh(II) catalysts to furnish intermediate enols which subsequently undergo Claisen rearrangement to alpha-hydroxyketones. Herein we report (1) studies into the mechanism of this transformation which establish that Claisen rearrangement is neither rhodium- nor acid-catalyzed but a reaction intrinsic to the intermediate enols that proceeds at a rate governed by enol substituents (R3, R4, R5) and (2) the reaction of alpha-diazoketones with propargylic alcohols and preliminary investigations into its scope and mechanism.

Full text of DOI:10.1021/ol990697x

ORGANIC
LETTERS
1999
Vol. 1, No. 3
371-374
Rhodium Carbenoid-Initiated Claisen
Rearrangement: Scope and Mechanistic
Observations
John L. Wood* and George A. Moniz
Sterling Chemistry Laboratory, Yale UniVersity, New HaVen, Connecticut 06520-8107
john.wood@yale.edu
Received May 28, 1999
ABSTRACT
It has been shown that r-diazoketones react with allylic alcohols in the presence of Rh(II) catalysts to furnish intermediate enols which
subsequently undergo Claisen rearrangement to r-hydroxyketones. Herein we report (1) studies into the mechanism of this transformation
which establish that Claisen rearrangement is neither rhodium- nor acid-catalyzed but a reaction intrinsic to the intermediate enols that
proceeds at a rate governed by enol substituents (R , R , R ) and (2) the reaction of r-diazoketones with propargylic alcohols and preliminary
3
4
5
investigations into its scope and mechanism.
Earlier this year we reported that R-keto rhodium carbenoids
react with allylic alcohols to stereoselectively furnish R-hy-
droxy ketones.1 In that study, a combination of experiments
performed with spectroscopic monitoring demonstrated that
the reaction proceeds via initial formation of a Z-enol which
subsequently undergoes Claisen rearrangement faster than
ketonization (e.g., 1 f 2 f 3, Table 1).2,3 In addition, the
reaction was found to be general with regard to both diazo
and alcohol substrate. Intrigued by the ease with which the
derived enols (e.g., 2) undergo rearrangement, we decided
to extended our investigations by exploring the kinetics of
the Claisen chemistry. As herein reported, these latter efforts
clearly demonstrate the reactive nature of the enol intermedi-
ates and have led to further experiments which extend the
scope of this process.
To facilitate the planned kinetic studies, we sought a series
of enols that would undergo rearrangement over a broad
temperature range. We thus became interested in a report
describing the preparation and crystallographic analysis of
6, a stable enol that is produced by reacting a methanol
solution of azibenzil (5) with H2SO4.4 Hoping to extend our
Rh(II) conditions to this system, we treated a CH2Cl2 solution
of 5 (1.0 equiv) and CH3OH (1.1 equiv) with Rh2(TFA)4
(0.01 equiv) and were delighted to observe the rapid and
selective formation of 6 (Scheme 1).5,6 Importantly, similar
(4) McGarrity, J. F.; Pinkerton, A. A.; Schwartzenbach, D.; Flack, H.
D. Angew. Chem., Int. Ed. Engl. 1983, 22, 405.
(5) As reported by McGarrity (ref 4), 6 can be isolated via crystallization
or converted to several derivatives. To confirm the stability of the enol
geometry under the conditions of derivatization, we prepared the corre-
sponding acetate (ii) and established the structure by single-crystal X-ray
analysis (see Supporting Information).
(1) Wood, J. L.; Moniz, G. A.; Pflum, D. A.; Stoltz, B. M.; Holubec, A.
A.; Dietrich, H.-J. J. Am. Chem. Soc. 1999, 121, 1748.
(2) The highly selective formation of a Z-enol suggests the reaction
proceeds via intramolecular proton transfer to oxygen (i.e., i f ii).
(6) A study of the tautomerization of this enol has recently been reported,
see: Jefferson, E. A.; Kresge, A. J.; Wu, Z. Can. J. Chem. 1998, 76, 1284.
(7) (a) Our interest in this possibility derived from the known Lewis
acidity of some Rh(II) catalysts, see: Doyle, M. P. Chem. ReV. 1986, 86,
919. For a leading reference describing Lewis acid promotion of the Claisen
rearrangement, see: Lutz, R. P. Chem. ReV. 1984, 84, 205.
(3) For recent reviews on the chemistry of enols, see: (a) Hart, H. Chem.
ReV. 1979, 79, 515. (b) Kresge, A. J. CHEMTECH 1986, 16, 250. (c) Capon,
B.; Guo, B.; Kwok, F. C.; Siddhanta, A. K.; Zucco, C. Acc. Chem. Res.
1988, 21, 135. (d) Rappoport, Z.; Biali, S. E. Acc. Chem. Res. 1988, 21,
442. (e) Kresge, A. J. Acc. Chem. Res. 1990, 23, 43.
10.1021/ol990697x CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/07/1999

Table 1. Reaction of Doubly Stabilized R-Diazoketones with
Table 2. Influence of Reaction Conditions on the
Rearrangement of Allyloxy Enol 8 in CD2Cl2 at 298 K
Various Substituted Allylic Alcohols1
activation for the Claisen rearrangement of enols 13 and 14
to be 1-2 kcal/mol lower than those of the parent allyl vinyl
ether (12, Scheme 2).9 Variable temperature kinetic experi-
Scheme 2
ments performed with enol 8 revealed a ∆Hq of 20.6 kcal/
mol, corresponding to a ∆∆Hq of -4.8 kcal/mol between 8
and 12 (see Figure 1). This difference is substantially larger
conditions using either Rh2(OAc)4 or Rh2(TFA)4 were found
to promote the coupling of 5 with a range of allylic alcohols.
With primary allylic alcohols, both catalysts afforded a single
1
enol (e.g., 7 and 9, observed by H NMR). Decomposition
of 5 with Rh2(OAc)4 in the presence of 3-buten-2-ol afforded
enol 8 as a single isomer whereas catalysis of the same
reaction with Rh2(TFA)4 results in a 7:1 mixture of isomers.
As with 6, derivatization of 7-9 is possible and in the case
of 7 the corresponding triflate 10 proved amenable to single-
crystal X-ray analysis (see Supporting Information).
Scheme 1
Having accessed enols 6-9, we began investigating the
possibility that rhodium was promoting the facile Claisen
chemistry.7 For these studies we used 8 as a representative
substrate and observed first-order kinetics over a range of
5-90% conversion. In addition we found the rate to be
independent of the ligand on rhodium and unaffected by
either catalyst concentration or the presence of Proton Sponge
(1.0 equiv, Table 2).8 Thus, the propensity of allyloxy enols
to undergo Claisen chemistry appears to be intrinsic and not
the result of conditions present during their formation.
Qualitatively, these observations are consistent with a
computational study by Houk which predicts enthalpies of
Figure 1. Arrhenius plot for allyloxy enol 8.
Org. Lett., Vol. 1, No. 3, 1999
372

than the computed ∆∆Hq values for 13 and 14, a discrepancy
we attributed to the presence of rate-enhancing substituents
in 8. As illustrated in Table 3, this latter notion is supported
by the kinetic behavior of 7-9 which clearly illustrates the
significant influence of substituents on the rate of Claisen
rearrangement. In accord with other reports,10 we found that
alkyl substituents at C4 enhance the Claisen rate relative to
the parent allylic system 7, whereas substrates containing
analogous subtituents at C5 (i.e., 9) rearrange more slowly.
With regard to the substituents on the enol-baring carbons,
the kinetics of 15 indicate that the aromatic group at C1 does
not offer any significant acceleration relative to an alkyl
group at this position; however, the rearrangement rate was
dramatically retarded when electron donation from the enol
hydroxyl was reduced by derivatization as the corresponding
trifluoroacetate 11.11 In addition to illustrating the importance
of substituent effects, these investigations serve to explain
the different Claisen/OH-insertion ratios observed with some
substrate combinations.1 This product distribution derives
from a delicate balance between ketonization and rearrange-
ment that is clearly reflected by the data in Table 1 where
one sees that ketonization can become competitive when the
effects on enol stability. In our studies to date we had utilized
R-diazoketones that furnish enols conjugated to either an aryl
or carbonyl moiety. To determine if enol conjugation was a
prerequisite for the success of this reaction, we extended our
investigation to include R-diazoketones 16a and 16b. As
shown in Table 4, 16a and 16b are both excellent substrates
Table 4. Reaction of Monostabilized R-Diazoketones with
Various Substituted Allylic Alcohols13
Table 3. Substituent Effects on the Rearrangement Rate of
Allyloxy Enols in C6D6 at 313 K
that combine with a variety of allylic alcohols to furnish the
corresponding R-hydroxy ketones with chirality transfer
equivalent to their conjugated counterparts. In addition, the
successful use of 16b serves to illustrate that competing
intramolecular events such as carbonyl ylide formation will
not inhibit conversion to the enol.
Intrigued by both the intrinsic reactivity of the allyloxy
enols and the generality of the reaction with regard to both
diazo and alcohol substrate, we have been considering
applications of other similarly derived reactive enols.14 To
this end we explored the coupling of 16a and 18 with several
propargylic alcohols. As shown by the data presented in
Table 5, the intrinsic enol reactivity is again manifest and
the expected allenic R-hydroxyketones are produced in good
yields under very mild conditions.15
Claisen rearrangement is slowed by allylic alcohol substrates
either lacking accelerating (entry a) or containing retarding
(entry c and d) substituents.12,13
Having established the importance of allylic alcohol
substitution, attention was turned to the issue of substituent
(8) The experiments with Proton Sponge indicated that protonation of
the incipient enol is not promoting the Claisen rearrangement. Experiments
with other amine bases such as Et3N resulted in rapid tautomerization of
the intermediate enol to the R-allyloxy ketone (i.e., the formal O-H insertion
product).
(9) (a) Yoo, H. Y.; Houk, K. N. J. Am. Chem. Soc. 1997, 119, 2877. (b)
We thank Professor Houk for providing unpublished results regarding the
Z-enol 14.
(10) Ziegler, F. E.; Chem. ReV. 1988, 88, 1423. Wipf, P. In Compre-
hensiVe Organic Syntheses; Trost, B. M., Fleming, I., Eds.; Pergamon
Press: New York, 1991; Vol. 5, p 827 and references therein.
(11) A similar effect was noted by Koreeda in the anionic variant of
this reaction, see: Koreeda, M.; Luengo, J. I. J. Am. Chem. Soc. 1985,
107, 5572.
(12) The delicate balance between ketonization and rearrangement is
further illustrated by a deuterium isotope study wherein allyl alcohol OD
was found to combine with 1 and funish a 6:1 ratio of Claisen/OD insertion
products.
In contrast to our experiences with allylic alcohols, the
reaction course with propargylic substrates was found to be
(13) The structure assigned to each new compound is in accord with its
infrared and high-field 1H (500 MHz) and 13C (125 MHz) spectra, as well
as appropriate parent ion identification by high-resolution mass spectrometry.
(14) For a recent application of reactive enols in synthesis, see: Wood,
J. L.; Holubec, A. A.; Stoltz, B. M.; Weiss, M. M.; Dixon, J. A.; Doan, B.
D.; Shamji, M. F.; Chen, J. M.; Heffron, T. P. J. Am. Chem. Soc. In Press.
(15) For a review of pericyclic reactions of acetylenic compounds, see:
Viola, A.; Collins, J. J.; Filipp, N. Tetrahedron 1981, 37, 3765.
(16) For a similar application, see the preceding Letter in this issue (Jung,
M. E.; Pontillo, J. Org. Lett. 1999, 1, 367). We thank Professor Jung for
sharing this information prior to publication.
Org. Lett., Vol. 1, No. 3, 1999
373

that treatment of diazoketone 18 and 3-butyn-2-ol with 1
mol % of Rh2(OAc)4 results in clean and rapid formation of
an enol (19) (Table 6) which, upon continued monitoring,
undergoes conversion to both the [3,3] and [2,3] rearrange-
ment products 20 and 21, respectively (t1/2 for [2,3] ) 210
min at 40 °C).17 In contrast, when 1 mol % of Rh2(TFA)4 is
used the substrates combined to form the [2,3]-rearranged
product (21) before a single spectrum could be acquired.
Upon slowing this reaction by using only 0.1 mol % of Rh2-
(TFA)4, one again observes enol 19 which, upon continued
monitoring, undergoes rapid [2,3] rearrangement to 21 in a
(pseudo) first-order sense (t1/2 ) 5.4 min, 25 °C) and to the
exclusion of 20. Importantly, a side-by-side experiment was
performed wherein two equimolar solutions of 18 and
3-butyn-2-ol in C6D6 at 25 °C were each treated with 1 mol
% of Rh2(OAc)4 to effect diazo decomposition. Once
complete (as evidenced by decolorization and cessation of
N2(g) release, about 1 min), 1 mol % of Rh2(TFA)4 was
added to one mixture. Analysis of aliquots after 5 min (1H
NMR) revealed complete conversion of 19 to 21 in the Rh2-
(TFA)4-treated case and less than 5% conversion in the
control. Taken together, these observations clearly support
a mechanism wherein enol 19 bifurcates between rearrange-
ment pathways and only the [2,3]-path is dramatically
facilitated by Rh(II).
In summary, we have established that the coupling of
R-diazoketones with allylic alcohols under Rh(II) catalysis
furnishes allyloxy enols which, due to the electron-donating
hydroxyl group, are intrinsically predisposed to undergo
Claisen rearrangement. Importantly, this predisposition en-
ables the Claisen chemistry to compete with ketonization, a
delicate reactivity balance which can be influenced by
substituent effects. An additional level of difficulty arises
with enols derived from propargylic alcohols since these
intermediates undergo [2,3] rearrangement in competition
with the two aforementioned processes. Fortunately, proper
choice of reaction conditions enables selection among the
possible pathways. Further efforts to manipulate R-ketocar-
benoid-derived enols in a variety of synthetically useful ways
are ongoing.
Table 5. Reaction of R-Diazoketones with Propargylic
Alcohols13
highly dependent upon conditions. Specifically, as indicated
in Table 6, the reaction conditions influence a competition
between paths leading to either [3,3] or [2,3] rearrangement.16
Of particular note is the dramatic difference between
reactions run with the acetate and trifluoroacetate ligands
(cf. entries 1 and 5), an observation which suggests the
Table 6. Effect of Reaction Conditions on [2,3]/[3,3] Ratio
Observed with R-Diazoketone 18 and 3-Butyn-2-ol
Acknowledgment. We are pleased to acknowledge the
support of this work by the NIH (1RO1GM54131) and
Bristol-Myers Squibb, Eli Lilly, Glaxo-Wellcome, Yaman-
ouchi, and Zeneca through their Faculty Awards Programs.
J.L.W. is a fellow of the Alfred P. Sloan Foundation. G.A.M.
thanks Bristol-Myers Squibb for a graduate fellowship. In
addition, we acknowledge Susan DeGala for assistance in
securing the X-ray analyses.
Supporting Information Available: Experimental and
spectral data pertaining to all new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
involvement of rhodium in the [2,3] process and stands in
stark contrast to the reactivity of the allyloxy enols (see Table
2). Attempting to rationalize this difference in behavior, we
explored the possibility that the [2,3]- and [3,3]-derived
products might not be linked by a common enol intermediate.
OL990697X
(17) Neither [3,3]-product 20 nor [2,3]-product 21 undergoes a [1,2] shift
upon either silica gel chromatography or treatment with 5 mol % of Rh2-
(OAc)4 in refluxing benzene for 3 h.
1
In the event, monitoring experiments by H NMR revealed
374
Org. Lett., Vol. 1, No. 3, 1999