Organic Letters
Letter
a
Table 1. Comparative Study between Vibrating Ball Milling and Solution Synthesis of Z-Ala-Phg-Ile-OMe
b
b
entry
reagents
temperature (°C)
time, t (min)
yield (%)
purity (%)
LDL (%)
1
2
3
4
5
6
7
8
EDC·HCl/Oxyma
EDC·HCl/HOBt·H2O
EDC·HCl/HOAt
DIC/HOAt
DIC/Oxyma
HATU/Et3N
33 (33)
34 (34)
34 (34)
30 (31)
31 (31)
34 (34)
33 (33)
10 (30)
10 (30)
10 (30)
10 (40)
10 (40)
10 (60)
10 (20)
30
93 (88)
90 (90)
88 (90)
n.d. (n.d.)
n.d. (n.d.)
85 (88)
86 (82)
96
>99 (32)
70 (48)
95 (59)
67 (39)
46 (<10)
88 (58)
71 (55)
>99
<1 (9)
25 (35)
<1 (26)
17 (33)
c
c
c
<1 (n.d.)
1 (<1)
2 (9)
<1
HBTU/Et3N
EDC·HCl/Oxyma
d
c
n.d.
a
Milling reactions were performed in a 5 mL PTFE jar with three stainless steel balls (5 mm diameter) at 25 Hz. The total mass of reactants was 50
b
c
d
mg. The values in parentheses correspond to solution reactions. Determined by HPLC. n.d. = not determined. Reaction performed on a total
mass of 250 mg of reactants in a 15 mL PTFE jar with one stainless steel ball (10 mm diameter) at 25 Hz for 30 min using 1.7 equiv of EDC·HCl.
EtOAc was used as a liquid additive instead of DMF.
stopped, the temperature of the milled material reached 33
°C.32 For the sake of comparison, the same temperature was
applied to the synthesis of Z-Ala-Phg-Ile-OMe under classical
magnetic stirring conditions, while using the minimal amount
of DMF enabling proper mixing (η = 20 μL/mg). Under these
experimental conditions, Z-Ala-Phg-Ile-OMe was produced in
88% yield, 32% purity, and 9% of the LDL epimer (Table 1,
entry 1; see values given in parentheses). These results clearly
showed that liquid-assisted ball milling could minimize
epimerization of a highly epimerization-prone C-term activated
peptide fragment. In addition, the coupling was complete after
10 min of ball milling, while 30 min were required under
classical solution conditions. This latter observation could be
explained by the higher concentration of the reaction mixture
in the ball mill, compared to that in solution. Other coupling
reagents and additives were screened to establish the broadness
of our observations. When couplings were performed in
solution, the quantity of DMF was kept as low as possible
while enabling proper mixing under classical magnetic stirring
(11 μL/mg < η < 26 μL/mg). When Oxyma was replaced with
HOBt·H2O, ball milling conditions led to 25% of the LDL
epimer and 70% purity, while 35% of the same epimer and 48%
purity were obtained for the synthesis in solution (Table 1,
entry 2). The use of HOAt gave better results than HOBt·
H2O: only 1% of the LDL epimer was produced in the ball-
milling conditions while 26% LDL epimer were obtained after
synthesis in solution (Table 1, entry 3). Using DIC with HOAt
under ball milling conditions furnished the tripeptide with 67%
purity along with 17% of the LDL epimer, while synthesis in
solution led to lower purity (39%) and higher epimerization
(33%) (Table 1, entry 4). These results obtained with DIC
were not improved when replacing HOAt with Oxyma (Table
1, entry 5). Concerning the stand-alone coupling reagents
HBTU and HATU, their use induced low epimerization levels
both by ball milling and in solution (Table 1, entries 6 and 7).
Yet, the HPLC purity profiles of the products obtained by ball
milling were much better than when synthesized in solution.
The comparison between the ball milling and solution-phase
processes clearly showed the superiority of the mechanochem-
ical approach over the classical synthesis in solution, since
lower epimerization rates and better purity profiles were
obtained with almost all tested coupling agents and additives,
while yields remained comparable. In addition, performing the
coupling with the best coupling agent and additive (EDC·
HCl/Oxyma) on a larger scale (250 mg total mass instead of
50 mg) with larger reactor and ball (15 mL reactor with one
10-mm-diameter ball) did not affect the yield (96%) or
epimerization rate (<1% LDL; Table 1, entry 8) (see the
tions of jar and balls).
We next applied the best ball-milling experimental
conditions (EDC·HCl/Oxyma) to the synthesis of other
challenging peptides. Here, EtOAc was used instead of DMF
to improve the overall environmental impact of our ball-milling
conditions. Thus, ball milling Z-Ala-Phg-OH with HCl·H-Phe-
OMe produced the tripeptide Z-Ala-Phg-Phe-OMe in 89%
yield with full conservation of the stereogenic centers (Table 2,
entry 1). When Z-Ala-D-Phg-OH was used, the tripeptide Z-
Ala-D-Phg-Phe-OMe was isolated with a similar yield of 92%
and excellent purity (Table 2, entry 2). Cysteine, which is
another epimerization-prone amino acid,9,33 was not racemized
during the synthesis of Z-Ala-Cys(Bn)-Ala-OMe and Z-Ala-
Cys(Bn)-Phe-OMe under our conditions. In both cases,
tripeptides were obtained in excellent yields (>94%) and
diastereomeric excesses (>99%) (Table 2, entries 3 and 4).
Like isoleucine, valine is a β-branched amino acid known to be
more difficult for coupling than most other amino acids.34 Yet,
the reaction between Z-Phe-Val-OH and HCl·H-Cys(Bn)-
OMe led to the desired tripeptide, Z-Phe-Val-Cys(Bn)-OMe,
in almost quantitative yield after 30 min of ball milling (Table
2, entry 5). Importantly, no trace of the LDL epimer could be
detected by HPLC analysis.
Similar results were obtained starting with the Z-Phe-D-Val-
OH (Table 2, entry 6). When HCl·H-Cys(Bn)-OMe was
replaced with HCl·H-Ser(tBu)-OtBu, excellent yields were
obtained, again with no epimerization (Table 2, entries 7 and
632
Org. Lett. 2021, 23, 631−635