Organic Letters
Letter
Scheme 2. Method and Scope, Isolated Yields
Figure 1. Benzophenothiazine applications.
coupling partners.2 Eventually, we settled for an oxidative
procedurebased onan(enantiopure)Jacobsen-type Cr(III) Lewis
acid catalyst,8 which we adapted from a recently reported and
inspiring oxidative coupling of 2-naphthylamines by Kozlowski
and Paniak.9 A few adjustments in terms of temperature and
solvents allowed us to access a very diverse number of scaffolds in
good yields (Scheme 2). In some cases, toluene was found to be a
superior solvent. In some other cases it was hexafluoroisopro-
panol (HFIP).
We then explored the dehydrogenative substrate scope with a
number of important phenols10 and indoles.11 Several phenols
were well accommodated in the cross-dehydrogenative coupling
to benzophenothiazine, affording C−C biaryl coupling products
with yields up to 85% (structures [2−10]). Likewise, C3−H
indoles delivered the expected C3−C cross dehydrogenative
coupling products with yields up to 68% (compounds [11−16,
22]). Importantly, when the C3 position is blocked by a
functional group, thus preventing re-aromatization, the cross-
dehydrogenative coupling reaction still occurs. However, an
additional subsequent C2−N bond-forming cyclization event
then takes place, yielding unique highly functionalized
paddlewheel-like scaffolds (structures [17−21]) with yields up
to 79%. Standard COSY, NOESY, HSQC, and HMBC NMR
techniques easily confirmed the structural assignments (see the
SI). Moreover, interestingly, some small enantiomeric excess
could be observed for structure [2] (ee = 13%). Decreasing the
reaction temperature, however, did not significantly improve the
ee (20%), associated with a decrease in conversion. Some of the
coupling products did not display any ee at all ([19] and [23], ee
= 0%). Our efforts at optimizing the enantioselectivity of this
method have failed so far, but are continuing in our laboratory.
Mechanistically, the reaction is assumed to proceed in a radical
fashion. Indeed, while the addition 1.5 equiv of TEMPO did not
significantly alter the reaction, BHT as well as α-methylstyrene
completely suppressed the formation of product [2] (see the SI).
In the latter cases, some benzophenothiazine homocoupling
product could then be detected. Finally, the facts that 2-
acetylphenothiazine and Nonhebel’s phenol (3,5-di-tert-butyl-
phenol)12 arecompetent N-and O-coupling partners (structures
[23−24]), respectively, are also in line with a radical mechanism.
It should, moreover, be noted that ordinary phenothiazine
(PTZH) also converted well in those reaction conditions with
3,4,5-trimethylphenol. The latter substrate, however, delivered
only the classical dehydrogenative C−N bond-coupled product
[25] (86%, Scheme 3). In contrast, the C−C bond-coupled
product [25′] could not be detected. Thus, the herein described
Cr(III) catalyzed method merely accelerates the pre-existing
intrinsic tendency of benzophenothiazine to couple in a C−C
bond forming fashion.13 The latter propensity for C−C bond
formation would be due to the benzo unit affecting the radical
a
Toluene was replaced with HFIP.
behavior of benzophenothiazine compared to that of PTZH1
(see Scheme 3, proposed mechanism).
In summary, we investigated the cross-dehydrogenative
reactivity of benzophenothiazine with indoles and phenols and
found it to be at odds with that of ordinary phenothiazines. In
particular, benzophenothiazine strongly favors C−C over C−N
cross-dehydrogenative couplings. Moreover, benzophenothia-
zine easily furnishes paddlewheel-like structures with indoles,
thus yielding interesting scaffolds in the chemistry of
benzophenothiazine-derived materials.
B
Org. Lett. XXXX, XXX, XXX−XXX
Bub, Christina L.
