Communication
group, or a pre-purified cationic metal catalyst. Our transform-
able N-nitroso group enables RhIII-catalyzed synthesis of target
molecule 3a in 96% isolated yield at 508C. An optimal catalyt-
ic system includes 2.5 mol% [RhCl2Cp*]2/10 mol% (Cp*=
1,2,3,4,5-pentamethylcyclopentadienyl) AgSbF6 as the catalyst
precursor, 10 mol% NaOAc as the base and proton shuttle,
and dichloroethane as the solvent. The reaction can be scaled
up to 10 mmol (~2.5 g 3a) level without sacrificing the prod-
uct yield.
group (1p). For
a
meta-substituted N-nitrosoaniline, the
degree of regioselectivity is governed by the substituent
(1q–u). The reaction proceeds in a regiocontrolled manner for
the two examined disubstituted N-nitrosoanilines (1v, 1w);
however, a diverged site selectivity is displayed. The critical di-
recting role of N-nitroso group is demonstrated by the ab-
sence of reaction between N-methylaniline and 2.
Mechanistic experiments indicate the intermediacy of a five-
membered rhodacycle[7] in the catalytic cycle and CÀH activa-
tion as the turnover-limiting step (kinetic isotope effect value
of 2.0). A catalytic mechanism involving N-nitroso-directed
electrophilic CÀH activation/ortho-rhodation (with the assis-
tance of OAcÀ), migratory insertion of the aldehyde carbonyl
group, and proto-demetalation for product release and catalyst
turnover, is therefore proposed.
An investigation of the reactivity of N-nitrosoaniline deriva-
tives identifies exclusive N-nitroso-directed ortho selectivity
and compatibility of a broad range of substitution patterns
(1a–w, Table 1). The existence of syn and anti isomers[7] in the
CÀH functionalization products reported herein due to thermo-
dynamically restricted NÀN bond rotation does not affect their
synthetic utility because further elaboration of the transform-
able N-nitroso group is typically mandated and allows for the
generation of identical synthetic targets. The yield is observed
to be inversely correlated with the steric bulkiness of the N-
substituent (1a–d). The ortho-substituted N-nitrosoanilines
(1e–i) generally exhibit a lower reactivity than the para-substi-
tuted counterparts (1j, 1l–o). The reaction is favored for an
electron-donating group (1j, 1l) over an electron-withdrawing
With CÀH functionalization products 3a–w in hand, we then
set out to explore the N=O bond as a synthetic handle for the
creation of a heterocycle scaffold. The [2+2] cycloaddition re-
activity of the C-nitroso N=O bond toward the ketene,[8] cou-
pled with ketene generation from an a proton-bearing ester[12]
and remarkable stability of NÀN bond[10] under basic condi-
tions, highlights a hypothetically viable mechanistic course.
The use of tBuOK (2 equiv) in dichloromethane proved effec-
tive in delivering 1-methyl-1H-indazol-3-ol (4a, as confirmed by
single-crystal analysis, Table 2) from 3a at RT, through a formal
[2+2] cycloaddition/fragmentation process. The synthetic po-
tential of such a distinct enabling methodology is illustrated
by a broad substrate scope (3a–w, Table 2), regardless of ste-
rics and electronics. This represents a significant advantage
over previously documented methodologies that involve either
the preinstallation of synthetically demanding groups (e.g.,
halogens)[9b–h] or imposition of undesired auxiliaries (aryl sub-
stituents at N2)[9i,11a] on the products.
Table 1. Substrate scope for RhIII-catalyzed N-nitroso-directed CÀH addi-
tion to ethyl 2-oxoacetate.[a,b]
Initial attempted 1H NMR spectroscopic characterization of
the 3a to 4a reaction course directly under basic conditions
provides no informative signals, as a result of the low solubility
of ionic species involved in the transformation. Acidification
with HOAc (with inevitable presence of adventitious H2O), a re-
agent that 3a does not react with (Scheme 1, middle left, top
panel), allows the conversion process to be monitored
(Scheme 1, top panel) and, together with other pertinent ob-
servations, suggests a ketene as a key intermediate for the
[2+2] cycloaddition/fragmentation process: 1) the transforma-
tion features a fast detachment of the EtOÀ fragment from de-
1
protonated 3a (no H NMR spectroscopic signals observed for
3a after merely 20 min; detachment of EtOÀ, as confirmed by
1
a change in the H NMR spectroscopic chemical shift and split-
ting pattern of its methylene protons), presumably to generate
ketene 3a-K (as revealed by the observation of its nucleophilic
addition-afforded hydration product,[12a] 3a-det), and a slow
conversion of 3a-K to 3a-CA-FO (decrease of 3a-det and in-
crease of 4a with the elapse of time, but with their combined
amount approximately constant and equivalent to that of EtOÀ
(Scheme 1, bottom left, top panel); consistent with the initial
production of 3a-K and EtOÀ and subsequent exclusive trans-
formation of 3a-K to 3a-CA-FO). 2) Only a trace amount of 4a
can be generated from 3a-det (ruling out hydrolyzed species,
[a] Reaction conditions: N-nitrosamine 1a–w (0.4 mmol), ethyl 2-oxoace-
tate 2 (0.8 mmol), dichloroethane (DCE; 2 mL). [b] Isolated yields. [c] Iso-
mers due to restricted NÀN bond rotation; syn: N-alkyl (except for 3d, N-
phenyl) cis to nitroso oxygen atom; anti: N-alkyl (except for 3d, N-
phenyl) trans to the nitroso oxygen atom. [d] Further splitting of NMR
spectroscopic signals observed for anti-like isomers due to, presumably,
additional restricted bond rotation. [e] syn isomer.
&
&
Chem. Eur. J. 2014, 20, 1 – 6
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!