Organic Letters
Letter
more favorable than those with electron-withdrawing groups
(3g, 3j, and 3k vs 3h and 3i). Notably, alkyl-substituted alkyne 1l
took part in this reaction smoothly to give 3l in a yield of 54%.
Second, N-aryl piperidines 2 with different substituents attached
to the phenyl rings were also tested. Delightfully, functional
groups ranging from electron-deficient groups (fluoro and
bromo) to electron-rich groups (methyl and methoxy) were
suitable to this transformation, delivering target products 3m−
3s in yields of 28−82% without showing an obvious electronic
effect. However, no intended product was formed when 1-(o-
tolyl)piperidine was used as the substrate, probably due to the
steric effect. In addition, this reaction was compatible with 1-
biphenylpiperidine for the production of 3t in a yield of 80%.
Third, the suitability of different N-aryl pyrrolidines was
explored. It was found that all of them were amenable to this
cascade reaction, albeit giving corresponding products 3u−3y in
somewhat lower yields. The structure of 3g was confirmed by X-
ray single-crystal diffraction analysis (see the Supporting
Information). In addition, byproducts such as naphthyl chain
amines (4) and/or 2-(2-oxo-2-phenylethyl)benzaldehyde de-
rivatives (5) were isolated from some of the reactions for the
construction of 3. The NMR spectra and isolated yields of these
byproducts are included in the Supporting Information.
Remarkably, besides piperidines, N-phenyl pyrazine (2o) with
two nitrogen atoms could undergo this transformation to furnish
3z in a yield of 26% (Scheme 3). However, when secondary
cyclic amines, such as piperidine and pyrrolidine instead of
tertiary amines, were tried, the target 1,2,3,4-tetrahydrobenzo-
[g]quinoline derivatives were not found, and amide products 3a′
and 3b′ were obtained in yields of 31% and 22%, respectively
(Scheme 4).
14% and 18%, respectively, but pyrrolidine afforded 3b′ in a
yield of 16% (Scheme 6).
Scheme 6. Synthesis of 3a′, 3b′, and 3″
To elucidate the reaction pathway, a series of control
experiments were conducted. First, 2 equiv of radical inhibitor
butylated hydroxytoluene (BHT) was added to the standard
reaction systems for the preparation of 3a and 4a. It turned out
that the yields of 3a and 4a were not affected obviously,
excluding the radical mechanism [Scheme 7, (1) and (2)].
Scheme 7. Control Experiments
Next, we turned our attention to the substrate scope for the
synthesis of 4 (Scheme 5). First, a number of 1-arylpyridines 2
with various substituents, such as fluoro, chloro, bromo, cyano,
ester, methoxy, and methyl groups, attached at the para, meta, or
ortho position of the phenyl ring were found to be tolerated,
giving the desired products 4a−4k in yields of 21−75%. Among
them, electron-rich or halogen-substituted N-arylpyridines did
not affect the efficiency of the reaction obviously (4b−4d, 4g,
and 4h vs 4a), but 1-(4-cyanophenyl)piperidine and 1-[4-
(methoxycarbonyl)phenyl]piperidine, which were substituted
with strong electron-withdrawing group, gave the desired
products (4e and 4f) in somewhat lower yields. Due to steric
hindrance, 2 with a methyl unit at the ortho position of the
phenyl ring transformed into 4k in a low yield of 21%. Notably,
this reaction was compatible with 1-biphenylpiperidine and 1-
naphthylpiperidine, delivering 4l and 4m in yields of 46% and
28%, respectively. Second, a range of 2-alkynylbenzaldehydes 1
were studied (4n−4s). The results showed that forms of 1 with
electron-rich groups on the phenyl rings were more favorable
than those with electron-deficient groups (4o and 4p vs 4n).
Moreover, 1l, attached with an alkyl-substituted alkyne unit, was
suitable for this reaction to afford 4t in a yield of 41%. Third, 1-
aryl-substituted pyrrolidines were tested in this transformation,
affording products 4u and 4v in yields of 40% and 24%,
respectively. In addition, byproducts such as 1,2,3,4-
tetrahydrobenzo[g]quinoline derivatives (3) and/or 2-(2-oxo-
2-phenylethyl)benzaldehyde derivatives (5) were isolated from
some of the reactions for the construction of 4. The NMR
spectra and isolated yields of these byproducts are included in
the Supporting Information. Interestingly, when piperidine was
used as the substrate, fused N-heterocycle benzo[g]quinolin-10-
yl(phenyl)methanone (3″) and 3a′ were delivered in yields of
Second, 2-(2-oxo-2-phenylethyl)benzaldehyde (5a), which
could be isolated from the reaction mixture for the formation
of 4a, was treated with 1a under the standard conditions for the
construction of 3a and 4a. Consequently, 3a was not obtained
and 4a was delivered in a yield of 49%. It showed that 5 may be
the possible intermediate for 4a but not for 3a [Scheme 7, (3)
and (4)]. Third, 18O labeling experiments with H218O for the
formation of 3a and 4a were carried out [Scheme 7, (5) and
(6)], from which [16O]-3a versus [18O]-3a and [16O]-4a versus
[18O]-4a were given in a ratio of 1:0.8 and 0.2:1, respectively, as
determined by HRMS analysis. Fourth, the reactions with [18O]-
1a (0.6:1 16O:18O) were also tested. As a result, the
corresponding products [16O]-3a versus [18O]-3a and [16O]-
4a versus [18O]-4a were obtained in ratios of 0.7:1 and 8.5:1,
respectively [Scheme 7, (7) and (8)]. Fifth, when the reactions
for the synthesis of 3a and 4a were conducted under N2 instead
of air, 3a and 4a were still obtained in yields of 59% and 55%,
respectively [Scheme 7, (9) and (10)]. Given that H2O might be
C
Org. Lett. XXXX, XXX, XXX−XXX