Organic Letters
Letter
the desired ring fusion thus reducing the selectivity of the
cyclization.
and the overall connectivity of the hymenin framework was
confirmed by X-ray crystallography. Dibromination of the
pyrrole was achieved by exposure to 2.2 equiv of NBS at low
temperature to deliver 31, again confirmed by X-ray
crystallography (Scheme 4). This was a more challenging
transformation as bromination also occurred on the imidazole
ring (C5) leading to 32 and thus the reaction had to be
monitored carefully by NMR spectroscopy to maximize the
yield of the dibromide ∼50%. Bromination at C5 on the
imidazole moiety began to impinge as the degree of conversion
increased and thus reaction was terminated once over
bromination was observed. The monobromide 28 was
deprotected by acid-catalyzed hydrolysis of the sulfonyl urea
to provide the corresponding free imidazole 29 (Scheme 4).
Completion of the synthesis of 7 simply required conversion of
the azide to the amine and removal of the O-methyl group.
Finding a suitable reductant to effect both conversions while
retaining the bromide(s) was challenging. As expected, Zn/
HOAc was effective in the required reductions (cf. 25 → 27,
Scheme 4), but also resulted in reductive debromination. At
this point, we explored using a two-step reduction sequence by
converting the azide to the amine by treatment with Lindlar
catalyst and hydrogen which delivered the corresponding
amines 30 (and 33) in accordance with prior experi-
ence.18,42−45 SmI2, which we have employed in related but
less advanced intermediates to cleave an N−O bond
reductively was investigated.39 However, upon reaction of
either 30 or 33 to SmI2, reductive cleavage was accompanied
by partial reductive debromination, which in the case of 33
delivered debromohymenin (7) in good yield. Harran and co-
workers have observed similar reductive debromination on
pyrroles during their synthesis of advanced axinellamine
derivatives.46 At this point, we sought reagents with different
mechanisms for N−O bond cleavage, whereupon the use of
Mo(CO)6 in wet acetonitrile for cleavage of isoxazolidines was
identified as a possibility.47 Gratifyingly, treatment of 30 under
these conditions resulted in a clean reduction of the N−O
bond which upon purification provided 2-debromohymenin
(7). We also found that reaction of the azide-containing
precursor 29 with four equivalents of Mo(CO)6 delivered 7 in
good yield and thus telescoping the last two steps (Scheme
4).48 Dibrominated intermediate 33 was subjected to the same
sequence of reactions, unfortunately clean hymenin (6) was
not obtained from this sequence giving rise to inseparable
mixtures of products (Scheme 4).
The known propargyl amine derivative 1740 was coupled
with 4-iodoimidazole 1641 through a Sonogashira reaction
(Scheme 2). Exposure of 18 to TFA removed the carbonate
moiety and the resulting N-methoxyamine was acylated with
the pyrrolecarbonyl chloride 19, affording pyrrole amide 20.
Treatment of the pyrrole carboxamide 20 with AuCl3 in
dioxane led to the formation of two pyrroloazepinones 21 and
22 (X-ray), of which the major product 21 (2,3-fusion) had
the correct orientation of the ring fusion for application to the
synthesis of hymenin and stevensine. In addition to the
cyclization products, a small amount of the trans acylated
derivative 23 (X-ray) and ca. 10−20% of unreacted starting
material 20 were recovered. Our initial plan was to employ the
nonrearranged adduct en route to stevensine via bromination
of the pyrrole ring; however, exposure of 21 to NBS resulted in
competitive bromination of the azepinone double bond and
thus these advanced intermediates were better configured for a
pursuit of hymenin.
The major 2,3-isomer 21 was subjected to catalytic
hydrogenation to deliver the saturated congener, which upon
lithiation with LDA (3.9 equiv) and exposure of the resulting
organolithium to TsN3 resulted in the formation of the 2-azido
derivative 24 (Scheme 3).18,42−45 Deprotection of the
Scheme 3. Elaboration of the Pyrroloazepinone Core
In summary, we have developed a convenient 9-step
synthesis of pyrroloazepinone-containing natural product 2-
debromohymenin (7) from a commercially available iodoimi-
dazole via a key gold-catalyzed hydroarylation. Critical to the
success of this chemistry was the use of an N-OMe group as a
protecting group to facilitate the selectivity of the pyrrole ring
fusion. Elaboration of the pyrroloazepinone through reduction,
C2-azidation and bromination delivered key late stage
intermediates. Chemoselective reduction of the C2-azide and
reductive cleavage of the N-methoxy group with Mo(CO)6
delivered the debromohymenin congener in good yield. This
synthesis is longer than the previously reported approach from
the Horne group, but it is potentially more flexible and avoids
the use of extremely long reaction times (2 × 7 days). Efforts
are ongoing to find an alternative endgame solution to afford
higher brominated congener, hymenin as well as addressing the
issues of a stevensine synthesis.
sulfonylurea afforded the free imidazole 25 (confirmed by X-
reduction with Zn/AcOH afforded desbromohymenin (27)
in good overall yield.7,19 Through careful control of the
stoichiometry and the reaction time (for 26 Zn = 10 equiv, 1 h;
for 27 Zn = 25 equiv, 6 h), the azide can be reduced
chemoselectively to afford 26. Attempts were made to effect
bromination of 27 to generate hymenin (or debromohymenin)
with little success, presumably due to competitive reaction of
the imidazole at C5; therefore, bromination was explored
earlier in the sequence.
Gratifyingly, monobromination of 24 occurred readily at C4
on the pyrrole using 1.0 equiv of NBS in THF/H2O at −60 °C
affording 28 in 80% (Scheme 4); the location of the bromine
C
Org. Lett. XXXX, XXX, XXX−XXX