Organic Letters
Letter
a fundamental role in the stereodetermining step (again no in
situ formed oxaziridines 2 were observed in any of these
experiments). Other N-protecting groups turned out to be not
well-suited,20 and we thus used 4a for our further testing. Among
different solvents, dry MTBE was the best (entry 12), and
analogously to our previous oxaziridine report,8 higher dilutions
were slightly beneficial (entry 13), which can be rationalized by
the aggregation tendency of the catalysts at higher concen-
trations.8 We also investigated the influence of catalyst loading
and reaction temperature, but no further improvement was
possible.20 The optimized conditions turned out to be rather
robust, allowing for the same selectivity and high yield on a 1
mmol 1a scale, as well (entry 13). Other oxidants were tested, as
well, but neither of them matched the performance of H2O2.20
Finally, testing different catalyst counteranions showed that
oxidizable anions are absolutely crucial (entries 13−15).
Whereas bromide catalyst A2 performed almost as good as
iodide A1 (entries 13 and 14), non-oxidizable anions did not
allow for any noteworthy catalysis (entry 15 shows one example,
but others were tested with the same outcome).
experiments in the presence of radical scavengers such as
2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) or dibutylhy-
droxytoluene (BHT) point toward a radical-free pathway, as
neither of these two additives influenced the reaction.
Concerning the oxidation of I−, higher oxidation states like
−
−
IO3 or even IO4 can most likely be ruled out under these
neutral conditions.24 Among the lower oxidation state species
(I2, I3 , IO−, and IO2 ), our observations suggest that I2 itself is
not the catalytically relevant species, as I2 in combination with 4a
did not allow for any α-hydroxylation, whereas addition of H2O2
to this mixture allowed for product formation. Investigations by
Ishihara and others clearly showed that IO− is most likely the
catalytically relevant species for quaternary ammonium iodides
under oxidative conditions.10,11 Formation of hypoiodite may
also explain why this reaction proceeds well under base-free
oxidative conditions when I− is present: HIO has a pKa of 10.4,25
and thus IO− may not only be relevant for the oxidation but also
can play a role in the deprotonation of β-ketoesters 1.26 It should,
−
−
−
however, be noted that the formation and catalytic activity of I3
or even the unstable IO2− cannot yet be perfectly ruled out; for
example, some experiments with stoichiometric or catalytic
amounts of Bu4NI3 or in situ formed Bu4NIO11 in the presence
of H2O2 and 4a gave formation of racemic 3a, and in situ
formation of IO2− may also be possible under these conditions.27
Accordingly, a plausible and very general mechanistic scenario
is shown in Scheme 2B. H2O2 first oxidizes the ammonium
iodide to a hypervalent species I. This species then reacts with
the imine 4 to a yet unknown O-transfer species such as
compounds II (other options are feasible, as well!). The catalyst
then controls the enantioselective reaction of this species, with
the β-ketoester 1 giving 3 combined with a release of the imine 4
again. In addition, the reduced catalyst species Ired will be
reoxidized with H2O2 again, closing the proposed catalytic cycle.
With an optimized procedure for the asymmetric α-
hydroxylation of the parent substrate 1a at hand, we next
investigated the application scope of this protocol (Scheme 3).
Testing different ester groups first, we noticed that bulky esters
As mentioned before, other catalysts were less selective and
less active.20 Especially noteworthy are the results obtained with
Maruoka’s ammonium salts B4 and achiral PTCs (Scheme 2A).
Scheme 2. Influence of Ammonium Salt Properties and
Proposed Mechanistic Scenario
a
Scheme 3. Application Scope
In contrast to ammonium bromide A2 (entry 14, Table 1), the
commercially available ammonium bromide B2 was found to be
catalytically relatively incompetent, whereas the analogous
ammonium iodide B1 (derivatives thereof were used by Ishihara
in their oxidative approaches11,12) allowed for some product
formation, albeit with low enantioselectivity (Scheme 2A).
Using tetrabutylammonium iodide (C1, TBAI) or bromide (C2,
TBAB) as achiral PTCs, the conversion was also measurably
slower compared to that with catalysts A. This striking difference
between catalyst class A and ammonium salts B or C most likely
can be attributed to the H-bonding motif of catalysts A, but the
exact mode of activation remains yet speculative.
Nevertheless, all of our observations clearly substantiate a
mechanism where oxidation of the counteranion (i.e., iodide)
and formation of an activated O-transfer agent by reaction of the
oxidized halide species with the imine occurs. Control
a
b
All reactions were carried out using 0.1 mmol β-ketoester 1.20 Fast
c
decarboxylation of the ester under the reaction conditions. Limited
conversion of around 50% after 20 h.
C
Org. Lett. XXXX, XXX, XXX−XXX