Angewandte
Communications
Chemie
reactive a,b-unsaturated amides developed by Sharpless[18]
DDQ delivered indoloquinone 11. Surprisingly, the ensuing
quinone reduction and hydroquinone protection failed[17]
under conditions (H2, Pd/C then TBSOTf[8b] or TMSOTf[21]
and co-workers efficiently provided a diol intermediate.
Subsequent straightforward diol cleavage, ketone reduction,
and TBS alcohol deprotection yielded compound (À)-7, while
the other diastereomer was not observed. Its relative config-
uration was established by 2D NMR spectroscopy.[13] The
newly generated C2 alcohol was also converted into the
corresponding Mosher esters,[19] which indicated that the C2
center is S-configured.[13] These results confirmed that no
epimerization had occurred during these transformations.
In our proposed synthetic scheme (Figure 2), the A ring
has to be decorated with two OTBS groups, such as those in 1,
and we surmised that the transformation of the diol into the
azido alcohol should better be conducted after this manipu-
lation. Thus a 2-(tert-butoxycarbonyl)ethylidene (Boc-ethyl-
idene) protecting group was chosen to replace the existing
TBS groups given its mild installation and basic deprotection
conditions.[20] More importantly, this protecting group would
not transform the C1 alcohol into a good leaving group, which
might otherwise render this stereocenter fragile during future
transformations considering the reductive activation mecha-
nism of the mitomycins.[1] Accordingly, amide (À)-7 was
reduced with a borane tetrahydrofuran complex, and the
TBS-protected diol (À)-2 was then unmasked with tetra-n-
butylammonium fluoride (Scheme 2). The Boc-ethylidene
protecting group was installed onto the free diol under
reported conditions[20] to afford a diastereomeric mixture of 8.
As the acetal chiral center is inconsequential, 8 was directly
debenzylated by catalytic hydrogenolysis to afford phenol 9.
Whereas salcomine-catalyzed phenol oxidation with molec-
ular oxygen led to decomposition of 9,[17] the desired
oxidation to quinone 10 was successful with Frꢀmyꢁs salt
under argon atmosphere. Further dehydrogenation with
)
reported for structurally similar substrates for unknown
reasons. Fortunately, sequential reduction of the quinone
with sodium dithionite under biphasic conditions (dichloro-
methane/water) followed by TBS protection delivered 12 in
high yield. The Boc-ethylidene protecting group was subse-
quently removed according to reported procedures.[20] The
resulting free diol (+)-13 was activated through the formation
of a cyclic sulfite, which was opened by immediate treatment
with sodium azide to provide azido alcohol (À)-1 in 96% ee
with a diastereoselectivity of about 8:1. The diastereoselec-
tivity increased to > 50:1 when the diol was transformed into
a thiocarbonate[22] for activation before the sodium azide
treatment.
In Jimenezꢁs 14 step racemic synthesis of mitomycin K,[8]
this compound was next mesylated and oxidized[23] to
introduce the C9a hydroxy group in a completely diastereo-
selective manner (Scheme 3). Subsequent methylation of the
Scheme 3. Enantioselective synthesis of (+)-mitomycin K according to
Jimenez’s route. Reagents and conditions: a) MsCl, Et3N, DCM, 08C
to RT, 90%; b) DMDO, AcOH, acetone, À30 to 08C, 47%; c) NaH,
Me2SO4, THF, 08C, 76%; d) Ph3P, Et3N, THF, H2O, RT; e) MeOTf,
pyridine, DCM, 08C, 54% over 2 steps; f) TMSCH2Li, THF, À108C;
g) PCC, DCM, 08C, 43% over 2 steps, 97% ee. DMDO=dimethyldiox-
irane, Ms=methanesulfonyl, PCC=pyridinium chlorochromate.
C9a hydroxy group,[23] azide reduction, spontaneous aziridi-
nation, Peterson olefination, and oxidation with PCC had
furnished racemic mitomycin K. We followed the same route
to obtain enantioenriched (+)-mitomycin K in seven steps
from azido alcohol (À)-1 with 97% ee. The C2 chiral center of
the mitomycins has been reported to be derived from the C2
chiral center of d-glucosamine, presumably without any
epimerization during biosynthesis.[2d] Thus S absolute config-
uration would be expected for all mitomycins, including
mitomycin K. This hypothesis has previously been confirmed
by X-ray crystal-structure analysis of mitomycin A–C.[24] In
our current study, the absolute configuration of the C2 posi-
tion in (À)-1 was determined to be R[13] by Mosher ester
analysis.[19] Subsequent alcohol activation and intramolecular
SN2 attack should in theory give rise to S configuration at the
C2 position within the aziridine moiety, which is the same as
that in naturally occurring mitomycin K. Therefore, we have
completed the first enantioselective synthesis of mitomycin K
in 33 overall steps from commercially available starting
materials.[25]
Scheme 2. Enantioselective synthesis of (À)-1. Reagents and condi-
tions: a) BH3·THF, THF, reflux, 99%; b) TBAF, THF, RT, 69%; c) tert-
butyl propiolate, DMAP, MeCN, RT, quantitative; d) H2, Pd/C, EtOH,
RT, 87%; e) Frꢁmy’s salt, NaHCO3, acetone, H2O, argon atmosphere,
RT, 83%, 92% BRSM; f) DDQ, MeOH, RT, 86%; g) Na2S2O4, H2O,
DCM, RT; then TBSCl, imidazole, DMAP, 3 ꢀ M.S., DCM, RT, 71%,
85% BRSM; h) n-BuLi, pyrrolidine, THF, À78 to 08C; then RT, 85%;
i) SO(im)2, THF, À108C; then NaN3, acetone, H2O, RT, ca. 8:1 d.r.,
78%, 96% ee or CSCl2, DMAP, THF, À788C to RT; then NaN3,
acetone, H2O, RT, 75%, >50:1 d.r. BRSM=based on recovered
starting material, DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone,
im=imidazolate, TBAF=tetra-n-butylammonium fluoride.
Angew. Chem. Int. Ed. 2017, 56, 1 – 5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!