J. Wang et al.
Table 1. Substrate scope of azides.[a]
bered-ring ketones, we reduced the catalyst loading from
20 mol% to 10 mol%. The reaction still afforded an 89%
yield in 30 h (Table 2, entry 7). It is noteworthy that acyclic
ketones also gave appreciable reaction yields (Table 2, en-
tries 9 and 10). For example, the asymmetric acyclic ketone
2-butanone 2j gave a 75% yield of 3aj in 24 h (Table 2,
entry 9), and no other regioisomers were observed. We
deduce that the thermodynamically controlled enamine for-
mation of 2-butanone 2j affects the regioselectivity of this
reaction. The symmetric acyclic ketone 3-pentanone 2k
gave a 65% yield of 3ak in 20 h. Additionally, a phenyl-
ring-fused cyclohexanone 2l also reacted at room tempera-
ture (Table 2, entry 11) to afford 3al, albeit in a slightly
lower yield (57%, 10 h) due to some decomposition of this
product under the reaction conditions. Besides the above
types of ketones 2a–l, we also examined some a-functional-
ized ketones (Table 2, entries 12–16). Gratifyingly, b-ketoni-
trile 2m, 1,3-diketone 2n, and a-ester ketone 2o reacted ef-
ficiently, giving high product yields (Table 2, entries 12–14,
90–95%). Moreover, alkyl azide 1m could be reacted with
active b-ketonitrile 2m to afford an appreciable synthetic
yield of 3mm (Table 2, entry 16, 80%, 24 h). In contrast to
our previously reported results for reactions catalyzed by an
acyclic secondary amine catalyst (diethylamine V),[3] we
found that the cyclic secondary amine (pyrrolidine II) could
also efficiently catalyze such transformations under the opti-
mized reaction conditions. The regioselectivities of products
3aa[14] and 3ac[15] were determined by single-crystal X-ray
diffraction analysis.
To further extend the scope in terms of azides, we pro-
ceeded to examine tosyl azide 1n. As shown in Equa-
tion (3), the reaction with this substrate was completely dif-
ferent. In this case, pyrrolidine II served as a starting materi-
al, providing easy access to the biologically interesting N-
tosyl amidine 3na[16] in 85% yield. We deduce that when the
triazoline ring bears an electron-withdrawing group at the
N1-position, it is very labile. Thus, the initially formed tria-
zoline intermediate decomposes immediately to produce
amidine 3na through rearrangement with the loss of N2. Sur-
prisingly, l-proline showed no activity in this reaction. How-
ever, b-tetralone 2l reacted with tosyl azide 1n in the pres-
ence of l-proline to efficiently produce the desired triazole
compound 3nl (Equation (4), 85%, 5 h, RT). In this case, no
amidine product was formed. This may possibly be attribut-
ed to the steric hindrance of the carboxylic group, which
would inhibit the potential rearrangement and prohibit the
formation of the crowded amidine. Subsequent attempts to
subject two linear aldehydes (propionaldehyde 2r and 2-
phenylacetaldehyde 2s) to the reaction were unsuccessful.
The failure to obtain the desired triazoles might be attribut-
ed to a competitive self-aldol reaction.
Entry
1
R
Product[b]
Entry
7
R
Product[b]
3aa; 84%
24 h
3ga; 75%
12 h
3ba; 88%; 16 h
3ha; 62%
24 h
2
3
8
9
(78%; 30 h)[c]
3ca; 85%
14 h
3ia; 45%
18 h
3da; 72%
12 h
3ja; 80%
20 h
4
10
3ea; 72%
16 h
3ka; 67%
10 h
5
6
11
12
3 fa; 68%
18 h
3la; 60%
16 h
[a] Reaction conditions: 1a–i (0.25 mmol, 1.0 equiv), 2a (0.5 mmol,
2.0 equiv), DMSO (0.5 mL), 20 mol% catalyst II at 808C. [b] Yield of
isolated product. [c] 10 mol% of catalyst.
(Table 1, entries 2–4 and 10), or electron-neutral (Table 1,
entry 1) groups on the phenyl ring of the azide did not
affect the reaction. In some cases, a steric interaction ap-
peared to have a negative effect on the reaction yield. For
example, the presence of an ester group at the o-position of
the phenyl ring led to a lower yield (Table 1, entry 9, 45%).
Notably, a naphthalene ring was also tolerated in this reac-
tem (Table 1, entry 12, 60%). If a lower catalyst loading
(10 mol%) was employed, the reaction still afforded a good
yield in a reasonable time (Table 1, 78%, 30 h). However,
the reaction dramatically slowed down (<5% yield, 24 h)
when an alkyl (ethyl group) azide was applied to replace the
phenyl azide. Slow conversion of starting material was de-
tected by 1H NMR analysis.
Having investigated the reactivity of azides, we then eval-
uated carbonyls (Table 2). To our delight, this method can
tolerate a broad range of unmodified, commercially avail-
able carbonyl compounds. Cyclic six-member ring based ke-
tones gave good to high yields (Table 2, entries 1–6). This
six-member ring can bear alkyl, ester, dialkyl, or ketal
groups at the b- or g-positions with respect to the ketone
function (Table 2, entries 1–5). Furthermore, the six-mem-
bered ring can also incorporate a heteroatom (Table 2,
entry 6). Interestingly, seven- or eight-membered-ring cyclic
ketones also efficiently provided high yields under the stan-
dard conditions, thus demonstrating that the ring size of the
ketone had little effect on this cycloaddition (Table 2, en-
tries 7 and 8, 94% and 85%, respectively). To verify the
high reaction efficiency seen with seven- or eight-mem-
To demonstrate the synthetic utility of this methodology,
we applied it to the synthesis of a CB1 cannabinoid receptor
antagonist (Scheme 2). The pyrrolidine II-catalyzed Huisgen
[3+2] cycloaddition between azide 1o and b-keto ester 2q
under optimized conditions (see Figure 2 and the Supporting
Information) furnished the intermediate 3oq in 98% yield
6090
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6088 – 6093