4600
M. Pigeaux et al. / Tetrahedron Letters 57 (2016) 4599–4603
Table 1
Synthesis of precursor 2a and temperature dependence of EDP
4a
1. (Boc)2O
2. NaH, BnBr
QD-thiourea
(1 equiv)
CO2H
H
N
EtO2C
CO2Et
CO2Et
Boc
Ph
Ph
CO2Et
NH2
N
3. KOH
T °C
THF, 72 h,
Boc
H
H
1
2a
(S)-3a
CF3
N
OMe
S
4a
H
QD-thiourea
F3C
N
N
H
H
N
Entry
Substrate
T (°C)
Yield (%)a
er (%)b
1
2
3
4
5
2a
2a
2a
2a
2a
0
25
30
40
60
20
40
68
94
95
69:31
85:15
84:16
82:18
75:25
a
Isolated yield. Starting material is the only other compound detected in the reaction for yields lower than 90%.
er values were determined by chiral HPLC analysis.
b
good compromise between reaction time and enantioselectivity for
the synthesis of protected phenylalanine 3a. The reaction time was
fixed at 72 h, QD-thiourea 4a (1 equiv) was employed as the cata-
lyst and THF as solvent, as these conditions regularly gave the best
outcomes in our earlier studies.6 The outcomes from that series of
experiments are summarised in Table 1.
Notably, the yield continued to rise as the temperature was
increased but enantioselectivity peaked at 25 °C (entry 2), implying
that this was the temperature of isoinversion for the EDP reaction.7
Our discovery of an isoinversion temperature has important impli-
cations as these typically arise in reactions where two enantiomers
are formed through the same mechanism. Thus, the concentration
of the R and S isomers is proportional to the rate constants for pro-
tonation of the two enantiofaces of the enolate, kR and kS. The tem-
perature of isoinversion corresponds to the point at which the
dominance of enthalpy over entropy switches.
Our attention next turned to the organocatalysts. In particular,
we were drawn to a report by Rawal et al. featuring hybrid squara-
mide–cinchona alkaloid catalysts.9a,b Although studies on such sys-
tems were limited,9c,d they seemed to function in a manner akin to
the thiourea-based catalysts we had been using, making them
worthy of study in this context. Moreover, the groups of Song and
Soos10 had each reported that the slow rate of reactivity of thiourea
and squaramide organocatalysts was due, at least in part, to self-
aggregation and that squaramides bearing two cinchona alkaloid
residues often performed better.10c Thus, in our search for a
universal catalyst for the EDP reaction, we decided to prepare six
squaramide–cinchona alkaloid hybrid catalysts, 5a–c and 6a–c,
and tested in our model EDP reaction using hemimalonate 2c.
Again, all reactions were conducted at 30 °C in THF for 72 h using
the catalysts in stoichiometric and substoichiometric amounts.
The results attained are summarised in Table 3.
From this study we were able to prepare phenylalanine 3a with
an er of 82:18 in 94% yield when the reaction was conducted at
40 °C (entry 4). It should be emphasised that the isolated yields
quoted for all of the EDP reactions described in this article reflected
the level of conversion, as in all cases starting material and catalyst
were the only other compounds found in the product mixtures.
Our attention now turned to the influence of the N-protecting on
the course of the reaction. DFT calculations had indicated that a
strong H-bond might form between the carbonyl of N-protecting
group and the thiourea in the catalyst.8a If true, we reasoned that
the EDP reaction might perform better with substrates bearing
other protecting groups.
Compared to thiourea 4a, the QD-squaramide hybrid catalyst 5a
gave a higher yield but reduced enantioselectivity in favour of the
product (S)-3c (Table 3, entry 1). The related bis-QD-squaramide
hybrid 6a performed similarly in respect of yield but showed a
significant improvement in respect of enantioselectivity (Table 3,
entry 2). The pseudo-enantiomer 5b performed worse, affording
the (R)-3c in moderate yield and comparable er (entry 3). Notably,
the related bis-QN-analogue 6b performed much better, giving (R)-
3c in 89% yield with an er of 80:20 (entry 4). Their performance
was in stark contrast to that exhibited by the cinchonidine series
(CD), where the bis-alkaloid hybrids performed worse than the
mono-alkaloid hybrids (entries 5 and 6).
To that end a series of hemimalonates with Cbz, Ac, CHO and
o-NO2C6H4 N-protection, 2b–2e, were prepared and tested in the
EDP reaction with thiourea catalyst 4a or its pseudoenantiomer
4b8b (Table 2). For this study the reaction time was fixed at 72 h
and temperature was fixed at 30 °C.
The Cbz protected precursor 2b gave the corresponding amino
acid with an er of 84/16 but its conversion was slow (entry 3).
Indeed, it exhibited the same trends as the Boc protected precursor
2a in respect of the influence of temperature on selectivity (not
shown). Acetate 2c gave the best enantioselectivity in this series
(entry 4).5a Faster conversion rates were given by the formyl and
o-nitrobenzoyl analogues, 2d and 2e, but these came at the
expense of er (entries 5–8). The results support our hypothesis
that H-bonding between the carbonyl of N-protecting group and
the thiourea in the catalyst is a critical interaction.
Buoyed by these encouraging results, we next examined the
performance of each catalyst when used at substoichiometric
levels. Notably, using 20 mol % of each hybrid in the EDP reaction,
bis-QD-squaramide 6a again emerged as the best organocatalyst in
terms of yield and enantioselectivity (Table 3, entry 8). The others,
while affording good selectivity in some cases (e.g., 5b, entry 9; 5c,
entry 11), exhibited very slow conversion rates as evidenced by the
poor yields attained for 3c following work-up. With some promis-
ing new catalysts identified, our attention now turned to an assess-
ment of their universality.
To that end, hemimalonates 7–9 were each synthesised and
their performance assessed in the newly established EDP protocol
using QD-thiourea 4a and bis-QD-squaramide 6a as organocata-
lysts. For comparison, the performance of QN-squaramide 5b was
also tested in some cases.
Pigeaux, Morgane
