A. Suzuki, K. Takagi, K. Sato et al.
Tetrahedron Letters 75 (2021) 153179
Furthermore, PyNTP gave the highest selectivity for thioamide for-
mation (Table 1, entry 4).
Next, the O/S selectivity in the nucleophilic attack of thiocar-
boxylic acid 1 to the condensing reagent was evaluated by measur-
ing the ratio of the resulting phosphine oxide and phosphine
sulfide using 31P NMR spectroscopy (Table 1, P@S (%)) using the
method reported by Hoeg-Jensen et al. [9] Although PyNTP
afforded higher selectivity than PyBOP, the O/S selectivity in the
nucleophilic attack of thiocarboxylic acid 1 to PyNTP was lower
than that of the analogous reaction with PyBOP (Table 1, entries
3 vs. 4). Therefore, we proposed a plausible mechanism for this
reaction as shown in Fig. 2. When the oxygen atom of the thiocar-
boxylic acid reacts with PyNTP, an intermediate 5 is formed. Then
3-nitro-1,2,4-triazole attacks to the intermediate 5 to give a nitro-
triazolide intermediate 6. Although the nucleophilic attack of the
amine to 6 can be expected to lead to the formation of the thioa-
mide 3, the reaction of 6 with the thiocarboxylic acid affords acid
anhydride intermediates 7a and 7b. There is also a possibility that
the intermediate 5 reacts with the thiocarboxylic acid and then the
anhydrides 7a and 7b are formed. The mixed anhydride 7b is a pos-
sible acylating reagent of the amine.
Fig. 1. Plausible mechanism for the formation of thioamide bonds using phospho-
nium-type condensing reagents.
Next, the effect of the reaction solvents was investigated using
PyNTP (Table 1, entries 4–7). The reaction in CHCl3 improved the
selectivity (Table 1, entry 5). In contrast, the condensation reaction
proceeded with low selectivity (Table 1, entries 6 and 7) in polar
solvents such as acetonitrile and N,N-dimethylformamide (DMF).
According to the P@S values of entries 6 and 7 (Table 1), it can
be assumed that the nucleophilicity of the thiocarboxylic acid is
higher in a polar solvent owing to the solvation of the counter
cation, which results in a low chemoselective nucleophile attack
to the phosphorus center. These results indicate that CHCl3 is the
most suitable solvent for the condensation reaction.
First, the conditions for the condensation reaction were investi-
gated using Fmoc-thioglycine 1 and benzylamine 2. Compound 1
was synthesized from Fmoc-Gly-OH following a method reported
previously [14] and stored as a potassium salt, because thiocar-
boxylic acids are known to react with atmospheric oxygen to form
disulfides [15,16]. Prior to the condensation reactions, the potas-
sium salt of the thiocarboxylic acid was transformed into a free
acid by its treatment with an aqueous solution of KHSO4. Various
condensing reagents were examined by reacting equimolar
amounts of thiocarboxylic acid 1 (0.1 M) and amine 2 in the pres-
ence of 4 equivalents of a condensing reagent in CH2Cl2 at room
temperature (rt) for 18 h. 4 equivalents of condensing reagents
were used to minimize the effect of hydrolysis of condensing
reagent by water on the reaction outcome [17]. Then, the mixture
was analyzed by 31P NMR to establish the ratio between phosphine
oxide and phosphine sulfide, and reverse-phase high-performance
liquid chromatography (RP-HPLC) was performed to determine the
ratio between thioamide 3 and amide 4. The results are shown in
Scheme 1 and Table 1 (entries 1–4).
When the thiocarboxylic acid and the amine were condensed in
the presence of DCC or HBTU, which are used commonly for the
synthesis of oligopeptides, the formation of the amide was signifi-
cant, and the thioamide was hardly formed (Table 1, entries 1 and
2). In contrast, the use of benzotriazol-l-yl-oxy-tris(pyrrolidin-1-
yl)phosphonium hexafluorophosphate (PyBOP), a phosphonium-
type condensing reagent having 1-hydroxybenzotriazole as a
leaving group, afforded the desired thioamides (Table 1, entry 3).
As 3-nitro-1,2,4-triazole is a good leaving group, the active
intermediate 6 is susceptible to the nucleophilic attack of thiocar-
boxylic acid 1. Therefore, to modulate the reactivity of the active
intermediate 6, a series of nucleophilic catalysts was investigated.
Chlorotris(pyrrolidin-1-yl)phosphonium
hexafluorophosphate
(PyCloP) was selected as a condensing reagent for the investigation
of the nucleophilic catalysts because its leaving group is a chloride
ion, which is not an effective nucleophilic catalyst. Equimolar
amounts of thiocarboxylic acid 1 (0.05 M) and amine 2 were
allowed to react with 4 equivalents of PyCloP and 8 equivalents
of a nucleophilic catalyst in CHCl3 at rt for 18 h. When the reaction
was conducted in the presence of a triazole derivative as a nucle-
ophilic catalyst, 13 equivalents of base was used; otherwise, 5
equivalents was used. Then, the mixture was analyzed by RP-HPLC
to determine the ratio between thioamide 3 and amide 4. The
results are shown in Scheme 2 and Table 2.
The reaction using PyCloP as a condensing reagent in the
absence of an additive resulted in lower selectivity than the reac-
tion with PyNTP and a higher P@S% value (Table 2, entries 1 vs.
2). On the other hand, in the presence of one of the triazoles gave
lower P@S% values than in the absence of triazoles. (Table 2, entries
2 vs. 3–6). These results indicated that the triazoles were acted as
nucleophilic catalysts. The acidity of the nucleophilic catalysts
evaluated follows the order 8a > 8b > 8c > 8d, according to Ham-
mett’s rule [18]. When 8 equivalents of 3-nitro-1,2,4-triazole (8a)
was added as a nucleophilic catalyst, the reaction selectivity
increased (Table 2, entry 3). Furthermore, the addition of 3-
cyano-1,2,4-triazole (8b), which is less acidic than 8a, improved
the selectivity (thioamide/amide = 77:23) (Table 2, entry 4),
whereas nucleophilic catalysts with lower acidity, such as
methyl-1,2,4-triazole-3-carboxylate (8c) and 3-chloro-1,2,4-tria-
zole (8d), gave low selectivities (Table 2, entries 5 and 6). From
these results, 8b was selected as a suitable nucleophilic catalyst
for the condensation reaction.
Scheme 1. Investigation of condensing reagents and solvents.
2
Suzuki, Ayaka
