CHEMCATCHEM
COMMUNICATIONS
In a first set of experiments, we screened a panel of ADHs in
the one-pot oxidation of the model substrate 3-phenylpen-
tane-1,5-diol (1a; Table 1) to prepare the corresponding enan-
tioenriched d-valerolactone 3a (Table 1). Of those ADHs, horse
liver alcohol dehydrogenase (HLADH) showed the highest ac-
tivity (see the Supporting Information for details) and therefore
it was selected for the optimisation of the biocatalytic process.
Formation of the valerolactone occurs by desymmetrization
of 1a, which is oxidized and generates an unstable hydroxy al-
dehyde that cyclises to afford the corresponding hemiacetal
(2a). The subsequent reoxidation of 2a gave access to the
stable lactone 3a.[10] On the basis of these initial results and to
use catalytic amounts of nicotinamide adenine dinucleotide
(NAD+) cofactor, we then attempted the enantioselective oxi-
dation of diol 1a via incubation with HLADH with flavin mono-
nucleotide as a recycling system.[4a] The reaction was moni-
tored by using GC, and a slow oxidation of diol 1a was ob-
served. Under the above-mentioned conditions, a mixture of
hemiacetal 2a and lactone 3a was observed in a 1:1.8 ratio,
which yielded the lactone 3a with a modest 48% ee after 72 h
(Table 1, entry 1). Notably, previously reported ee value for the
desymmetrisation of 1a was 21% ee.[4a] The biotransformation
proceeded with stereoselectivity towards the pro-S hydroxy
group, and preferentially, after reoxidation to the (S)-lactone.
To increase the conversion values towards the lactone, we
explored other recycling systems and reaction conditions. First-
ly, the addition of acetaldehyde to the reaction mixture was at-
tempted. Unfortunately, no reaction was observed, probably
owing to inhibition issues. On the basis of this unsuccessful
result with the coupled substrate approach, l-lactate dehydro-
genase (LDH) was used. This enzyme catalyses the reduction of
pyruvate to lactate with the concomitant oxidation of nicotin-
amide adenine dinucleotide to regenerate NAD+.[11] This regen-
eration system led to the total conversion of the process,
which yielded the (S)-lactone in moderate ee after 72 h
(Table 1, entry 2). The experimental results indicate that the
enantioselection in the 1a!2a step is the most important
contribution. The ee of the final lactone 3a remains unaltered
if the conversion values are between 50 and 100% (see the
Supporting Information for detailed time course study).
provement in the enantioselectivity was observed. The best re-
sults in terms of conversion and ee were obtained with THF
(Table 1, entries 8–10), and a 74% ee was observed in the pres-
ence of 7% of THF. However, the yield of 3a decreased consid-
erably in this case. Percentages of cosolvent higher than 15%
inhibited the oxidation reaction (data not shown). Solvent en-
gineering studies demonstrated that with acetonitrile as the
cosolvent, conversions were slightly lower than those obtained
with THF (Table 1, entries 11 and 12). Moreover, the addition of
a non-miscible cosolvent such as hexane enabled the addition
of higher percentages of cosolvent, which gave conversions
and ee values similar to those obtained in the aqueous
medium (data not shown).
Table 2. Effect of the substituent position and cosolvent on the enantio-
selective preparation of d-valerolactones 3a–g.
Entry[a]
R
THF [%]
c [%][b]
(S)-3a–g ee [%][c]
1
2
3
4
5
6
7
8
C6H5 (a)
C6H5 (a)
–
7
–
7
–
7
–
7
–
7
–
7
–
7
100 (1:9)
61 (1:2.5)
92 (1:17)
55 (1:3)
72 (1:2.6)
39 (2:1)
52
74
60
68
65
76
40
53
52
75
92
97
50
55
4-MeO-C6H4 (b)
4-MeO-C6H4 (b)
4-F-C6H4 (c)
4-F-C6H4 (c)
3-MeO-C6H4 (d)
3-MeO-C6H4 (d)
3-F-C6H4 (e)
3-F-C6H4 (e)
2-MeO-C6H4 (f)
2-MeO-C6H4 (f)
2-F-C6H4 (g)
2-F-C6H4 (g)
82 (1:9)
43 (1:1.8)
71 (1:3.5)
31 (1:0.8)
62 (1:15)
21 (1:3)
9
10
11
12
13
14
76 (1:12)
32 (1:2)
[a] Reaction conditions: substrates 1a–g (25 mm), NAD+ (0.84 mm), pyru-
vate (68 mm), LDH (2 U) and of HLADH (2 U) in Tris buffer (pH 7.5) at
308C for 72 h; [b] Conversion values were determined from GC analysis,
considering the formation of the corresponding hemiacetal and lactone,
and their ratio is given in parentheses; [c] The ee values of lactones were
determined from HPLC analysis.
At this point, different parameters were investigated to in-
crease the ee of this biotransformation. As expected, a change
in the pH of the reaction drastically affected the yield of the
process (Table 1, entries 3 and 4) without an increase in the ee
and a decrease in the 2a/3a ratio. Subsequently, the effect of
the temperature was studied, and although an increase in the
temperature favoured the formation of the thermodynamically
stable lactone 3a (notably, hemiacetal 2a was not detected at
378C; entry 5), the conversion of the process decreased drasti-
cally. Yet, if the reaction was performed at lower temperatures
(20 and 48C; entries 6 and 7, respectively), only lower conver-
sions were observed, indicating that both pH and temperature
could not be tuned to achieve better ee values.
We also extended this methodology to other 3-arylpentane-
1,5-diols 1b–g[13] (Table 2) to study the effect of the pattern
substitution in the aromatic ring on the lactonisation reaction.
From the results in Table 2, it becomes clear that the introduc-
tion of a group in the aromatic ring at the para position (1b
and c) led to a slight increase in the ee (Table 2, entries 3 and
5), which could owe to additional interactions in the active site
of HLADH. Nevertheless, the conversions for the oxidation of
these substrates decreased.
Motivated by the fact that oxidoreductases can work in or-
ganic solvents with changes in their properties,[12] a percentage
of cosolvent with respect to the buffer solution was added to
the reaction mixture (Table 1, entries 8–12). A significant im-
We then turned our attention to diols 1d–g possessing sub-
stituents closer to the prochiral centre (meta and ortho posi-
tions). We can observe that 3-(m-methoxyphenyl)pentanediol
(1d) reacts slower than the corresponding para-substituted
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 977 – 980 978