Angewandte
Research Articles
Chemie
Table 1: Screening and optimization.[a]
importance in synthetic chemistry and biology.[9] Most of the
À
synthetic strategies rely on late-strategy S S bond formation
À
due to its facile nature. Strategically, direct sulfuration via C
S bond formation provides a straightforward and powerful
approach for the synthesis of disulfides in terms of synthetic
modularity and structural versatility.[10] However, the fragile
À
S S bond poses serious challenges toward this end. There is
only one report about catalytic asymmetric disulfuration via
[11]
À
C S formation, though their counterpart enantioselective
sulfuration have been extensively reported.[12] In this case, the
reaction was limited in terms of scopes and stereoselectvity.
Recently, Xian,[13] Jiang,[14] Xu[15] and Wang[16] have developed
an elegant “mask” strategy for disulfuration. The identifica-
tion of a triaryl borane Lewis pair is the key for the reaction
with masked disulfide precursors (Scheme 1). Inspired by
Jiangꢀs work,[14] we challenged our chiral Lewis trio in
asymmetric direct disulfuration reaction with b-ketocarbonyls
and branched aldehydes. An enamine-based disulfuration was
developed with high chemo- and enantio- selectivity. The
Entry
Variation from standard conditions
Yield [%][b]
ee [%][c]
1
2
none
With 1a/TfOH
92
2
95
61
3
4
5
1b
1c
1d
trace
45
11
–
66
60
6
1e
5
–
7
8
9
BPh3
17
31
52
20
8
90
90
84
N.R.
35
trace
56
52
12
78
39
59
rac
rac
92
93
93
–
80
–
96
92
91
B(p-MeC6H4-)3
B(p-FC6H4-)3
DBU
DABCO
5b
10
11
12
13
14
15
16
17
18
19
20
À
reaction features both cyclic and acyclic quaternary C S
formation catalyzed by synergistically working bulky Lewis
trio (BLT).
5c
5d
No B(C6F5)3
No 5a
No 1a
r.t.
Ethyl acetate
MeCN
Results and Discussion
The reaction between acetoacetate 2b and electrophilic
polysulfurating reagents 3a was first selected as the model
reaction. Initially, the typical amine–Bronsted acid conjugate
such as 1a/TfOH was found to be virtually inactive (Table 1,
entry 2). When switching to bulky Lewis pair/trio system, we
quickly identified an optimal trio involving 1a/B(C6F5)3/5a,
delivering the desired disulfide product 4ab in 92% yield and
95% ee (Table 1, entry 1, see Supporting Information for
details on screening and optimization). An investigation of
different primary tertiary diamines was conducted with
[a] Reaction were performed at 358C in CHCl3 (0.2 mL) with 2b
(0.12 mmol), 3a (0.10 mmol), 1a/B(C6F5)3 (20 mol%) and 1,2,2,6,6-
Pentamethyl-4-piperidinol 5a (0.2 equiv) for 32 h at Ar atmosphere.
[b] Determined by 1H NMR analysis using 1, 3, 5-trimethoxybenzene as
an internal standard. [c] Determined by HPLC analysis.
try 12), suggesting the free alcohol of 5a has little impact on
catalysis. The use of inorganic base such as NaHCO3 showed
variations on the tertiary amino moiety (Table 1, entries 3– no activity at all (not shown). Optimization of the reaction
6). The smaller dialkyl groups such as dimethyl (1b), diethyl
(1c) or di-n-butyl (1d) showed lower activity and enantiose-
lectivity (entries 4 and 5), or even no activity (entry 3). On the
other hand, further increasing the bulkiness with di-(3-
pentanyl) amino-appended 1e led to nearly complete loss of
activity (entry 6). These results indicated the bulkiness of the
tertiary amine is critical for both activity and stereoselectivity
and a delicate balance of steric hindrance is required for
reactivity as known with FLP catalysis.[4] Different triaryl
boranes have also been examined, and the fully fluorinated
triaryl borane B(C6F5)3 is the most effective in terms of both
yield and enantioselectivity (Table 1, entries 7–9).
A survey of the second Lewis base led to the identification
of 2,2,6,6-tetramethyl-piperidine as the critical third compo-
nent (Table 1, entries 10–14). Such a bulky amine has been
previously utilized in FLPs.[3] The bulkiness of this base is
critical as the use of typical amine such as DBU and DABCO
gave poor activity with no enantioselectivity at all (Table 1,
entries 10 and 11). The install of N-methyl (5c vs. 5d, entry 13
vs. 14) or 4-hydroxyl (5a vs. 5c, entry 1 vs. entry 13) showed
slight improvement on yield and enantioselectivity. The 4-
methoxyl analogue 5b worked equally well (Table 1, en-
conditions showed that the reaction at room temperature was
slightly slower with incomplete conversion (Table 1, en-
try 18). The reaction also worked in other solvents such as
ethyl acetate or acetonitrile, but chloroform gave the best
productivity (entries 19 and 20 vs. entry 1).
In control experiments, the reaction did not proceed at all
in the absence of borane Lewis acid (entry 15). The two-
component Lewis pairs 1a/B(C6F5)3 or 5a/ B(C6F5)3 were also
examined, showing rather lower activity and poor enantiose-
lectivity (with 1a ) (Table 1, entries 16 and 17). This result,
together with the observation that no enamine formation was
detected with only 1a/B(C6F5)3, suggested that the reaction
with only Lewis pair 1a/B(C6F5)3 may proceed via an enol-
based pathway instead of the well-controlled enamine path-
way. Similar enol process was known.[10]
With the optimized conditions in hand, we next examined
the scopes of this BLT catalysis (Scheme 2). As shown in
Scheme 2, a variety of b-keto esters, including methyl, benzyl,
ethyl, iso-propyl, allyl and n-butyl ester all gave the desired
products with excellent ee (90–98% ee) and good yields (up to
94%) (Scheme 2, entries 1–7). The alteration of a-substitu-
ents on acetoacetates have also been examined, and a-ethyl
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 10971 –10976