Formate dehydrogenase - A biocatalyst with novel applications in organic chemistry

Article abstract of DOI:10.1039/c1ob06064c

In the field of industrial biocatalysis, formate dehydrogenase (FDH) is well established, in particular for its broad application in cofactor regeneration. Further applications have been limited by the enzyme's narrow range of substrates. These restrictio

Full text of DOI:10.1039/c1ob06064c

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 7941
www.rsc.org/obc
PAPER
Formate dehydrogenase - a biocatalyst with novel applications in organic
chemistry
Peter Fr o¨ hlich, Kristin Albert and Martin Bertau*
Received 1st July 2011, Accepted 5th September 2011
DOI: 10.1039/c1ob06064c
In the field of industrial biocatalysis, formate dehydrogenase (FDH) is well established, in particular for
its broad application in cofactor regeneration. Further applications have been limited by the enzyme’s
narrow range of substrates. These restrictions have been overcome now by the finding, that the enzyme
is capable of selectively cleaving formic acid esters to the respective alcohol. Five homologous alkyl
formates and phenyl formate as an aromatic ester were converted quantitatively by FDH from Candida
boidinii in a batch reaction within 3 to 5 h. The substrates were turned irreversibly into carbon dioxide
and the respective alcohol through hydride abstraction from the formyl group with full conversion. The
mechanism shows parallels to hydrolysis reactions of the AAC1-type. K -values and reactions rates of
M
the tested formic acid esters display a tendency to higher conversion rates with increasing chain length.
FDH emerged to be a superior deformylation catalyst compared to hydrolases as well as classical
catalysts, as was shown by the selective deformylation of 1-acetoxy-4-formoxy butane (92%) and
1
,3-bis(3-formoxypropyl)tetramethyldisiloxane. In particular its capability to distinguish between
formic acid esters and non-formic acid esters renders the method particularly suitable for protective
group chemistry. Furthermore the completeness of deformylation allows for converting substrates
highly incompatible with aqueous media like siloxanes within a few hours.
Introduction
There exists a series of different methodologies to synthesise
formic acid esters starting from primary and secondary alcohols,
phenols or bromides (Table 1). Their deprotection can be achieved
by chemical or enzymatic methods, what in the latter case is usually
Formate dehydrogenase (FDH, EC1.2.1.2) is well-established as
a redox enzyme for the regeneration of NADH in cofactor-
dependent biocatalysis. In industrial processes the enzyme is
involved in the large-scale production of tert-L-leucine for
instance. A significant advantage arising from FDH-supported
processes is the circumstance of the oxidation product CO
being chemically inert, while the thermodynamic equilibrium is
completely shifted to the product through release of the gas.
In fact it is irreversibility which renders this reaction indispens-
able in industrial biocatalysis. Typically enzymatic reactions are
characterised by being reversible, thus in most instances only
incomplete conversion can be achieved. As a consequence space-
time yields are unfavourable and the processes are economically
less attractive. This holds true particularly for hydrolase mediated
reactions where considerable effort is undertaken in order to reach
full conversion as this is done for instance, in deracemisation
reactions with vinyl acetate. The analogy of forming a gaseous
product which irreversibly drives the reaction to full conversion
prompted us to investigate the FDH as a potential candidate for
the irreversible cleavage of formic acid esters–thus an oxidative
alternative to classical enzyme-driven ester hydrolysis.
1,2
6,7
done by hydrolase action on the substrate. This approach allows
3,4
mild reaction conditions, but suffers from the abovementioned
drawbacks of equilibrium-mediated re-esterification. Moreover,
hydrolases will not differentiate between acylated residues in
general and formylated moieties in special, which means that a
selective deprotection of differently masked functional groups is
not possible as shown in Fig. 1. Deprotection by chemical methods
2
5
8
has been thoroughly investigated in organic chemistry as well.
Institute of Technical Chemistry, Department of Chemistry and Physics,
Freiberg University of Mining and Technology, 09596, Freiberg, Germany.
E-mail: martin.bertau@chemie.tu-freiberg.de; Fax: +49 3731 39-2324;
Tel: +49 3731 39-2384
Fig. 1 Chemoselectivity of hydrolase vs. FDH-mediated ester cleavage.
Org. Biomol. Chem., 2011, 9, 7941–7950 | 7941
This journal is © The Royal Society of Chemistry 2011

View Article Online
Table 1 Methods for formylation of selected substrates
Functionality
Substrate
Conditions/ Catalyst
Cyanuric chloride
Product
Ref.
9,10
Primary alcohols
Trichloro-1,3,5-triazine
Primary bromides
a
a
a
11,12
Secondary bromides
3 4
DMF , DIPEA as base PPh /CBr in methyl/ethyl formate
a
DMF = N,N-Dimethylformamide, DIPEA = N,N-Diisopropylethylamine, PPh
3
/CBr
4
Triphenylphosphine/Tetrabromomethane
1
8
However, these approaches suffer from rather harsh reaction con-
ditions and poor selectivity, thus displaying a lack of methodology
in particular for the modification of pre-functionalised complex
substrates as e.g. in pharmaceutical synthesis. It is for these reasons
why FDH appeared as a promising candidate to overcome these
shortcomings.
medium in the presence of FDH. However this reaction has been
attributed to hydrolase impurities of the FDH used. In addition
there exists no further information on autohydrolysis. Thus, from
literature no reliable information could be derived whether or not
FDH is capable of cleaving formic acid esters.
For mechanistic considerations we initially discussed a working
hypothesis based on undissociated formic acid, as this is usually
done in hydrogen transfer catalysis. To be precise, the FDH-
reaction yields a proton and a hydride ion under evolvement of
The catalytic mechanism of hydride transfer from formate to
+
NAD is well examined and lots of theoretical and practical
13–15
investigations have been done thus far.
If hydride abstraction
22
were realisable with the neutral, protonated species, there would be
no reason why the enzyme should not be capable of also cleaving
carbon dioxide.
+
According to this initial postulate, hydride transfer to NAD is
supported by concomitant nucleophilic attack of water at the acid
proton. This concept can easily be translated into the cleavage of
formic acid esters, where the attack of water takes place at the C-1
atom of the alkyl- or aryl residue. This analogy can be expressed
formic acid esters. This oxidation reaction would yield CO and the
2
respective alcohol. It would almost indistinguishably resemble of
a conventional hydrolase-mediated ester cleavage accompanied by
a FDH-mediated shift of the equilibrium to the product through
23,24
CO
2
release. Yet, in this case no differentiation between formyl
as given in Fig. 2. From protein structure data
the alignment
and other acyl groups would be observable (Fig. 1).
of a formic acid ester and its stabilisation through the positively
charged side chains of asparagine and arginine is conceivable to
proceed accordingly to the binding of formate which is the natural
substrate. The alkyl- or aryl residue of the substrate would find
Whether or not the concept holds true what can be derived
from only scarce literature reports which between themselves
diverge in their information on substrate specificity of FDH.
An activity towards esters and thioesters of formic acid has
been discussed for enzyme preparations from bacteria (e.g. Pseu-
sufficient space in the substrate binding channel. Consequently, the
+
reaction would yield NADH/H , CO
(Fig. 2).
2
and the respective alcohol
1
6
17
domonas sp. 101, Achromobacter parvulus ) and from higher
18
plants (e.g. pea seeds of Pisum sativum ). An activity towards the
thioester S-formylglutathione (GSF) which has been described
for the methylotrophic yeasts Hansenula polymorpha and Pichia
pinus as well as Candida boidinii and Kloeckera sp. 2201 has
been attributed to S-formylglutathione hydrolase.
19
20
21
22,23
In fact the
abovementioned enzymatic activity of pea seed extracts towards
formic acid derivatives was shown to result from exactly this
18
hydrolase which is able to hydrolyse GSF even in the absence
of a cofactor. Any activity whatsoever of FDH towards formic
acid esters has not been described thus far, with the exception
of one single report on ethyl formate being hydrolysed in aqueous
Fig. 2 Initial working hypothesis on the catalytic mechanism of hydride
transfer from formic acid and formic acid esters, respectively, to NAD
+
.
7
942 | Org. Biomol. Chem., 2011, 9, 7941–7950
This journal is © The Royal Society of Chemistry 2011

View Article Online
Hence, the FDH-catalysed dehydrogenation of formic acid
esters would turn out to be a potent and selective method for
the irreversible deprotection of formyl protected alkyl and aryl
alcohols. The proposed mechanism, an initial working hypothesis
only, applies for formic acid esters only and is not suitable for other
organic acid derivatives. It was for this reason why we undertook a
detailed study on the FDH-mediated cleavage of formic acid esters.
Table 2 Conversions of formates 1 – 6 by FDH-catalysed cleavage after
a reaction time of 3 h. The data are corrected by the amount of water
assisted hydrolysis of formates
R
Conversion in (%)
1
2
3
4
5
6
Methyl
Ethyl
n-Propyl
n-Butyl
Phenyl
23 ± 1,5
26 ± 4,7
39 ± 3,0
41 ± 3,7
86 ± 4,4
59 ± 2,4
Results and discussion
Formic acid ester cleavage by FDH
1-Acetoxy
In order to validate the concept of FDH catalysed cleavage of
formic acid esters, initial studies were undertaken with 2 as an
exemplary substrate. The chromatogram in Fig. 3 shows the
progress of the reaction within four hours. As can be seen, the
concept holds true: there is a clear consumption of 2 and ethanol
is being produced. What can be retrieved from Fig. 3 in addition
is the missing of formic acid even in trace amounts. From these
results can be taken that there is an oxidative but not hydrolytic
cleavage of the ester bond.
After the concept of FDH-mediated cleavage of formic acid
esters had been proven true with the example of 2, also the
homologues 1 and 3–6 were investigated. Since the initial working
hypothesis discussed a potential nucleophilic attack of water at
the carbinol C, submission of an aryl formate to FDH action
would provide crucial data for elucidating the mechanism of FDH-
mediated cleavage of formates.
in Fig. 4F. FDH was shown to catalyse the cleavage of the formate
group with high specificity while the acetate group remained
unaffected. Hydrolysis contributed to total conversion in an extent
observed for 1–4. Thus even a rather complex disubstituted ester,
1
-acetoxy-4-formoxy butane (6), was shown to be a substrate for
FDH. This result gives rise to a new perspective on the role of FDH
in protection chemistry of alcohols to catalyse the deprotection
step of formate esters under mild conditions. Needless to mention
that potential hydrolase impurity interfering with the FDH-
mediated reaction had been ruled out to occur (vide infra).
The results of the FDH-catalysed quantitative ester cleavage
are shown in Fig. 4. As reaction products an alcohol and CO
2
The kinetic data K
M
and vmax were determined for all alkyl
are evolved. As can be seen from Table 2 higher conversion rates
were revealed with increasing chain length. The methyl ester was
converted nearly half as slow as the butyl ester 4. This implies a
higher affinity of long-chain formates within the catalytic centre
of FDH. The solvolysis of 1–6 by water was almost linear in a
range between 0 . . . 50% and fast hydrolysis was observed for 5.
The hydrolysis rate of 5 by water at neutral pH value leads to
an exponential decrease of the content to 14% after 180 min.
This was in evident contrast to the linear hydrolysis rate by
aliphatic formates. Among the tested formate esters the mixed one
formates 1–3 as demonstrated in Table 3.
Table 3 Kinetic constants for FDH-mediated formate ester cleavage
Substrate
vmax (U/mL)
M
K (mM)
Sodium formate
49,7 ± 1,8
10,3 ± 0,7
12,8 ± 0,5
15,9 ± 1,4
21,7 ± 0,60
4,4 ± 0,48
2,9 ± 0,28
4,1 ± 0,31
4,7 ± 0,41
1,7 ± 0,06
1
2
3
5
6
displayed the fastest reaction within the first 60 min as depicted
Fig. 3 FDH-catalysed cleavage of 2.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7941–7950 | 7943

View Article Online
Fig. 4 Cleavage of aliphatic and aromatic formic acid esters 1–6 by FDH. The dashed lines illustrate the hydrolysis rate of formic acid esters by water.
What can be retrieved from Table 3 is the small difference
between the vmax-values of “natural” substrate sodium formate
and the non-classical ester substrates which differ by no more
than a factor of ~2 . . . 3. However, it was noticeable that the
maximum reaction rate for phenyl formate 5 was twice as high
secondary stabilisation effects within the active site of the enzyme
inferred by the aromaticity of the phenyl ring.
The active site is characterised by several hydrophobic clusters,
among which there are two aromatic amino acids (Phe98, His332)
with which aromatic substrates may undergo p–p–interactions.
There is further stabilisation by hydrophobic interactions with
as that for methyl formate 1. And the K
lower than that of the natural substrate sodium formate. The K
value determined for the latter is in line with literature sources
M
-value of 5 is even
24
M
-
Pro97 and Val309 (Fig. 5). These secondary interactions may
be suited to promote the rate of oxidative cleavage of aromatic
formates. On the other hand Pro97 and Val309 are not in direct
proximity to the far smaller alkyl substrates 1–4 which therefore
experience only weaker stabilising effects. Moreover, it appears
13,20,25–27
which report the value to range from 1,7 to 16 mM.
experimentally determined K -values of the alkyl esters (Table 3)
imply a correlation between chain length, i.e. hydrophobicity, and
. It is for this reason why a higher K -value for 5 would have
The
M
K
M
M
that the chain length dependent increasing K -values are the
M
been only too plausible, yet the value is the lowest one among all.
This reaction behaviour is proposed to be understood in terms of
outcome of stabilising effects within the active centre. From a
mechanistic point of view it cannot be excluded that also leaving
7
944 | Org. Biomol. Chem., 2011, 9, 7941–7950
This journal is © The Royal Society of Chemistry 2011

View Article Online
Table 4 Degree of hydrolysis of formic acid esters by water
Formate
Degree of hydrolysis in (%)
1
2
3
4
5
6
15,2 ± 0,40
40,7 ± 0,72
19,0 ± 0,38
23,6 ± 0,80
88,7 ± 10,6
46,6 ± 1,50
formic acid or formate detectable in presence of FDH. On the other
hand the instantaneous in situ removal of formic acid released from
the ester through the enzymatic action of FDH on first sight may
appear sufficient to explain the observed effects (Fig. 4). However
this approach is not in line with the time dependent product
formation, and it is by no means suitable to explain the formation
of mixed carbonates (vide infra). It was these observations that
prompted us to check for underlying hydrolase activity.
Potential contribution of competing enzymatic activity
Fig. 5 Hydrophobic areas around the active site of FDH to stabilise
aromatic and long-chain formic acid esters as demonstrated for 5. Arrows
illustrate p–p–interactions and dashed lines represent hydrogen bond
linkages.
In order to ensure that the observed hydrolysis was not the
outcome of hydrolase impurities acting on ester bonds (see
40
Fig. 6), a differentiation assay with p-nitrophenyl acetate was
performed. Any hydrolase present in the enzyme preparation
would result in a conversion of the substrate to p-nitrophenol. The
differentiation assay was performed with and without addition
of a protease inhibitor cocktail, whereas in both cases blanks
were determined in analogous manner with buffer instead of
enzyme solution. No hydrolase activity was detected in any of
the samples, so any impurities could be excluded. Adulterant
group properties contribute to the susceptibility of a formic acid
ester towards cleavage by FDH. These issues require further
clarification which is matter of forthcoming studies.
Non-enzymatic cleavage
The enzymatic cleavage of formates is accompanied by non-
enzymatic solvolysis through water, resulting in a promotion
of reaction rate and conversion, respectively (Table 4). Non-
enzymatic hydrolysis of alkyl formates contributes up to 14% of
the conversion after 240 min, while more than 35% of 5 were
hydrolysed already after 70 min. (Fig. 4). This effect is interpreted
as an outcome of the good leaving group properties of phenolate,
thus rendering 5 a good substrate for nucleophilic attack of water
+
effects through non-enzymatic reactions of NAD with phosphate
29
ions become effective only at concentrations ≥1,5 M for which
reason no such effects were effective to a considerable extent
in the 0,1 M phosphate buffer used. There were no indications
for a non-enzymatic reduction of the cofactor. Consequently any
increase in absorption detected by photometric measurements was
attributable to FDH-activity.
(
pH = 7.5). Formic acid generated by non-enzymatic hydrolysis
Our hypothesis of FDH playing the active part in cleaving
carbonic acid esters experienced further corroboration by the
was deprotonated by the buffer medium in order to exclude
side effects caused by acid catalysis. Under the conditions of
FDH-mediated cleavage solvolytically produced formic acid was
converted into CO instantly and did therefore not exert catalytic
effects on formic acid ester hydrolysis. In fact there was no free
28
observation of no homologous aliphatic ester (C
2
. . . C ) being
6
cleaved through the action of FDH. In fact formic acid esters as
the sole species bearing an abstractable hydride-H are the only
entities suitable for enzymatic attack by FDH.
2
Fig. 6 Formic acid ester cleavage by FDH and hydrolase.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7941–7950 | 7945

View Article Online
Table 5 Alcohol tolerance of FDH
initially discussed and illustrated in Fig. 2, we now discuss another
mechanism which is well suited to explain the experimental
observations made. The proposed mechanism bears analogies to
an AAC1-type of ester hydrolysis as introduced by Ingold (Fig. 7,
pathway 2).
Alcohol
IC50 (vol.-%)
33
Methanol
Ethanol
n-Propanol
Isopropanol
n-Butanol
Phenol
10,2 ± 0,9
4,8 ± 0,3
2,5 ± 0,1
4,7 ± 0,3
1,4 ± 0,1
0,32 ± 0,01
15,6 ± 1,0
One of the most important features of the AAC1-mechanism
is the involvement of carbocations doubly bound to oxygen (a,
a¢, c). The latter is known to withdraw electron density (-I-
effect) and to undergo resonance with carbocations (+M-effect)
what is a common motive in a multitude of reactions and
constitutes a motive well-known from biochemical reactions, e.g.
1
-Acetoxy-butane-4-ol
Effects of released alcohol on enzyme activity
34
in glycogen metabolism. As can be seen in Fig. 7, the electrophilic
carbocation c reacts with water to form a protonated carboxylic
acid, here formic acid (d). The deprotonation of the latter yields
the organic acid (e) in its neutral, undissociated form.
The alcohols released from their formic acid esters through the
action of FDH were tested for potential effects on FDH-CB
activity in the range of 0 . . . 45 vol.-% (Table 5). For all tested
species a notable tolerance was observed i.e. there is no significant
enzyme deactivation to be expected at typical working contents
of 0.50 . . . 1.29% (v/v). The only exception is made by phenol
which is known as an effective protein denaturing agent and
This underlying principle was supposed to apply to the FDH-
catalysed oxidative ester cleavage as well (Fig. 7, pathway 1).
Hydride abstraction through the enzyme would yield a resonance
stabilised cationic entity (A, A¢) which is turned into a carbonic
acid monoester (B) in analogy to the AAC1-pathway (Fig. 7,
pathway 2). What distinguishes the FDH-catalysed reaction
from the AAC1-pathway is the special spatial arrangement of
substituents at the carbonyl moiety. A transitional five-membered
ring formation, including the non-binding orbitals of ester oxygen,
favours the transfer of acid hydrogen from the formyl residue
to the latter. Thereby an electronically unstable betaine (C) is
produced which decomposes under release of carbon dioxide
and the respective alcohol (D). The process is favoured both
entropically and electronically by extrusion of a gaseous neutral
compound.
According to this new mechanistic model, three compounds
are to be expected in the reaction mixture. Firstly, the carbonic
acid monoester (B) has to be detectable, at least in trace amounts.
Dialkyl/diaryl carbonate (E) is formed by reaction of the carbo-
cation (A, A¢) with alcohol (D) which is more nucleophilic than
water, and finally isopropyl alkyl/aryl carbonate (F), where the
solvent isopropanol traps the carbocation (see Fig. 8).
30,31
potential inhibitor for dehydrogenases.
Here concentrations
exceeding 0.89% (v/v) resulted in significant loss of activity, and
the enzyme was fully inactive at c > 0.9% (v/v). On first sight
these results may appear to contradict the fast cleavage rates
observed for 5 (Table 2, entry 5). For the reasons discussed
above, 5 exhibits excellent substrate properties through stabilising
interactions with hydrophobic amino acid side chains in the active
site of the enzyme. On the other hand the released carbinol,
phenol, is a potent enzyme deactivator. Taken together enzyme
deactivation through the action of phenol which becomes more
effective with increasing phenol concentration competes with fast
substrate conversion. Obviously the protein deactivating effect
does not become significantly effective within the time frame of
ester cleavage.
Mechanistic considerations
While the mechanism of FDH-mediated formate oxidation is well
32
explored, little is known about the mechanism of FDH-mediated
ester cleavage. We therefore aimed at elucidating the process of the
latter on a molecular level.
Fig. 9 represents the gas chromatogram of FDH-catalysed
cleavage of 1 in which a peak at 4,3 min was determined by MS
analysis corresponding to the mixed carbonate methyl isopropyl
carbonate with characteristic fragments at m/z = 59. At the same
time, the intermediary formation of formic acid or formate had
Because experimental data obtained in the course of these
investigations did not provide support for the working hypothesis
Fig. 7 Pathway 1 - proposed mechanism for FDH-catalysed cleavage of alkyl and aryl formates, pathway 2 - AAC 1-type of ester hydrolysis.
7
946 | Org. Biomol. Chem., 2011, 9, 7941–7950
This journal is © The Royal Society of Chemistry 2011

View Article Online
+
Fig. 10 Regeneration of NAD by ADH-catalysed reduction of acetone
to isopropanol.
Fig. 8 Formation of species B, E and F confirms the proposed mechanis-
tic model.
is the product of the regeneration system, was shown to not
interfere with the FDH-mediated ester cleavage. The same applies
for the sacrificial substrate acetone on FDH. It cannot be excluded
to be excluded. Furthermore, release of carbon dioxide had to be
demonstrated.
+
however that in the presence of a dehydrogenase and NAD free
alcohol groups of the product are subject to enzymatic oxidation.
On the contrary no such process has been observed to take
place with ADH-CB so far. It was shown in these investigations
that up to 40 vol.-% of isopropanol and 9 vol.-% acetone were
tolerated in which a significant decrease at a concentration of
Application of FDH mediated ester cleavage in preparative and
protective group chemistry
In the course of the reaction intermediate carbonic acid mo-
noesters of type B are formed. These species are relatively
unstable, and by releasing carbon dioxide, the reaction equilibrium
is irreversibly shifted towards the free alcohol. This feature is
of invaluable use for enzyme-catalysed hydrolytic reactions, in
particular where these proceed within a time frame that allows the
equilibrium between hydrolysis and re-esterification to establish.
Under these conditions full conversion cannot be achieved, and
the yields obtained will be low. This applies particularly to
low substrate concentrations. Phase interfacial phenomena, poor
substrate solubilities or low functionalisation densities are possible
drawbacks that can be overcome by using the FDH reaction.
However, for all technical applications of FDH a cost-effective
cofactor regeneration system is a critical aspect. A regeneration
process comprising NADH-dependent reduction of acetone by
alcohol dehydrogenase from Candida boidinii (ADH-CB) is well-
≥
19 vol.-% for isopropanol was observed. These findings are in
line with those obtained for the effects of the released alcohols
on FDH-activity (vide supra). In our experiments FDH-mediated
deprotection succeeded on the milligram-scale when acetone was
added to the reaction mixture in order to have the reaction proceed
with catalytic instead of stoichiometric amounts of NADH.
Consequently this reaction appears interesting wherever for-
mate acts as a protective group. In the following we wish to report
two examples of where the transfer of the abovementioned results
in FDH-mediated deformylation brought the breakthrough.
1. Release of formate protected alcohol moieties in functional
organosiloxanes. The deprotection of formate protected 1,3-
bis(3-formoxypropyl)tetramethyldisiloxane 7 constitutes a major
challenge since the deformylation is impeded by low substrate
concentrations, phase interfacial phenomena, poor substrate
35,36
established for industrial use (see Fig. 10).
Isopropanol which
Fig. 9 GC-MS of reaction mixture from FDH-catalysed cleavage of 1. The chromatogram above shows a peak at 4,3 min which turned out to be from
methyl isopropyl carbonate as depicted in the mass spectrum below.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7941–7950 | 7947

View Article Online
Fig. 11 FDH-catalysed deprotection of 7 to 8 proceeds over 56% within 4 h. In contrast, conversion does not exceed 50% after 96 h with a conventional
hydrolase.
solubilities as well as low functionalisation densities. In conven-
tional hydrolysis, lipase from Candida antarctica has provided
the best results so far. Anyway, conversion did not exceed 43%
Experimental
General
37
within 96 h. In contrast conversion reached 56% after only 4 h
using FDH from Candida boidinii, and the reaction yielded the
quantitatively deformylated siloxane 8 after 17 h (see Fig. 11).
Again, as pointed out above, potential contributions of hy-
drolase activity, in concert with FDH-mediated oxidation of
intermediately released formic acid, were shown to not interfere
with the FDH-mediated deformylation of 7.
Formic acid esters, FDH from Candida boidinii and protease
inhibitor cocktail were purchased from Sigma Aldrich (St. Louis,
+
MO, USA). Lipase Novozym 388 and NAD were kindly provided
by Julich Chiral Solutions GmbH–a Codexis company. All other
reagents and solvents were in reagent grade and used as received
without further purification.
From the viewpoint of a technical application it has to be noted
here that terminally functionalised a,w-hydroxyalkylpolysiloxanes
are produced from the respective formoxy protected entities what
is unavoidable in order to prevent the polysiloxane backbone from
being degraded through intramolecular attack of the terminal
carbinol groups (“back-biting”) (Fig. 12).
Instruments
All reactions were monitored by gas chromatography (GC) and
1
H-NMR. Quantitative GC analyses were conducted using a
Clarus 500 gas chromatograph from PerkinElmer Inc. (Waltham,
MA, USA). Mechanistic studies were performed using an IT
4
00 GC-MS-system (ion trap) equipped with a FactorFour
column coupled to a MS 210 ion trap mass analyser, all from
Varian, Inc. (Palo Alto, CA, USA). Optical measurements were
conducted on a V630 Bio UV/Vis spectrometer from Jasco GmbH
(
Groß-Umstadt, Germany). Reaction vessels were handled in a
thermomixer (Eppendorf AG, Germany).
Fig. 12 Intramolecular degradation of hydroxyalkyl substituted
siloxanes.
Synthesis of 1-acetoxy-4-formoxy butane (6)
In a two necked flask equipped with a dropping funnel 0,75
mol 1,4-butanediol were mixed with 0,50 mol pyridine at 0 C
Consequently, the use of formylated hydroxyalkyl entities in
concert with FDH-mediated deformylation is an economic way
thus far for the production of terminally hydroxyalkyl function-
alised polysiloxanes.
◦
and 0,5 mol acetic anhydride were added dropwise in order to
keep temperature constant. After stirring the solution for 12 h
at room temperature brine was added. After phase separation
the organic layer was washed neutral and dried over sodium
sulfate. After distillation at 11 mbar a clear solution with a
2
.
1-Acetoxy-butane-4-ol. Another application arises from
the property of FDH to excellently differentiate between C
1
-
◦
carbonic acids (formic acid) and higher homologues. Since hydro-
lases release whatever carbonic acid from the protected alcohol,
mixed esters of the type m-acyloxy-n-formoxy-alkane can easily be
converted into the respective acyloxy carbinol by FDH-mediated
boiling range of 76 . . . 84 C was isolated. After removal of
impurities by column chromatography on silica gel (Merck AG)
with chloroform/methanol (95/5) as eluent 47,6 g 1-acetoxy-
1
butane-4-ol were isolated (0,36 mol, 48%). H-NMR(500 MHz,
release of CO
2
.
CDCl
3
): d (ppm) 1,57 (m, 2H), 1,67 (m, 2H), 2,0 (s, 3H), 2,45 (s,
1
3
This feature shall be exemplified with 1-acetoxy-butane-4-
ol. This monoester-monocarbinol species is easily obtained by
deformylation of its disubstituted progenitor 6 in single step
with >99,5% conversion and yield by FDH (Fig. 4). The option
to selectively modify a distinct carbinol moiety is of particular
interest wherever there is no stereochemical distinction between
two chemically equivalent alcohol groups (what would allow
application of the so-called ‘meso-trick’). Both a hydrolase and a
chemical catalyst like acid or base are incapable of differentiating
between these two ester entities and will entirely hydrolyse those.
That implies the need for the FDH to differentiate them and left
the higher carbonic acid ester unreacted.
H), 3,60 (t, 2H), 4,04 (t, 2H). C-NMR (125 MHz, CDCl
(ppm) 171,35; 64,34; 62,11; 29,02; 25,05; 20,95. n
3
): d
2
0
D
= 1,4352.
In the formylation step 27 mmol copper(II) nitrate trihydrate
was suspended in 18 mmol 1-acetoxy butane-4-ol and 36 mL
formic acid ethyl ester and heated under reflux for 3,5 h. The
solution was filtered to remove the catalyst and saturated with
brine. After separation the aqueous phase was extracted three
times with chloroform. The combined organic phases were dried
over sodium sulfate and volatile components were removed in
vacuo. The isolated yield was 71% in relation to 1-acetoxy butane-
4-ol with a purity >97% (GC). H-NMR(500 MHz, CDCl
(ppm) 1,71 (m, 2H), 1,74 (m, 2H), 2,02 (s, 3H), 4,06 (t, 2H), 4,17
38
1
3
): d
7
948 | Org. Biomol. Chem., 2011, 9, 7941–7950
This journal is © The Royal Society of Chemistry 2011

View Article Online
1
3
(
t, 2H), 8,03 (s, H). C-NMR (125 MHz, CDCl
3
): d (ppm) 170,7,
the solubility of formate esters 60 mL of isopropanol were added
before the reaction was started by adding 50 mL of diluted FDH.
+
1
6
1
60,8, 63,5, 63,1, 24,9, 20,6. MS (EI, 70 eV: m/z (%) = 161 (MH ,
%), 115 (58%), 101 (100%), 71 (30%), 54 (48%). Boiling point:
The reaction rate and K
M
-values were determined by Michaelis–
◦
20
15 C (10 Torr). n
D
= 1,4357.
Menten kinetic.
Exclusion of hydrolase-mediated side reactions
Extent of conversion
In order to ensure that the observed cleavage of esters 1–6 was
Enzymatic conversion of formic acid esters 1–6 was measured
by GC on a Carbowax (FS-CW) column (50 m ¥ 0.32 mm ¥
not the outcome of hydrolase impurities acting on ester bonds, a
40
differentiation assay with p-nitrophenyl acetate was performed.
Any hydrolase present in the enzyme preparation would result
in a conversion of the substrate to p-nitrophenol which can be
detected by photometric absorption measurement at 410 nm.
The reaction mixture was prepared by mixing 50 mM sodium
phosphate buffer (800 mL) at pH 7.5 and 10 mM p-nitrophenyl
acetate in isopropanol (100 mL). 100 mL of the diluted enzyme
solution (see above) were added and the amount of p-nitrophenol
released was determined by immediate photometric measurement
0
.5 mm) from CS Chromatographie Service GmbH (Langerwehe,
Germany). The pressure of the carrier gas H
temperatures of injector and detector were 250 C and 260 C,
respectively. The GC oven temperature was maintained at 50 C
for 2 min, programmed to 90 C at 4,5 K min , then to 180 C
at 40 K min-1 and maintained at 180 C for 2 min. Standard
calibration curves for all formic acid esters were plotted within
a range of 1–500 mM. Mechanistic analyses were done by GC-
MS measurements on a GC-431 with mass spectrometer MS-210
2
◦
was 0.8 bar,
◦
◦
◦
-1
◦
◦
◦
at 410 nm. The assay was performed at 30 C. The differentiation
(
ion trap, Varian Inc.) equipped with a FactorFour VF-WAXms
assay was performed with and without addition of a protease
inhibitor cocktail, whereas in both cases blanks were determined
in an analogous manner with buffer instead of enzyme solution.
capillary column (30 m ¥ 0.25 mm ¥ 0.25 mm) from Varian, Inc.
(
Palo Alto, CA, USA). Helium was used as carrier gas with a flow
◦
of 1 mL/min. The GC oven temperature was maintained at 50 C
for 5 min, programmed to 110 C at 10 K min , then to 200 C at
◦
-1
◦
0 K min-1 and maintained at 200 C for 3 min.
◦
Quantitative measurement of formic acid ester cleavage
4
For each formic acid ester 1–6, a 500 mM stock solution in
isopropanol was prepared. The reaction mixture for quantitative
measurement of formic acid ester cleavage contained 940 mL of 100
Standard activity assay for FDH
For activity assays a standard reaction mixture was prepared by
mixing 100 mM potassium phosphate buffer at pH 7.5 (1,0 mL)
with 1 M sodium formate in distilled water (250 mL) and 10
mM potassium phosphate buffer at pH 7.5, 60 mL of isopropanol,
+
2
50 mL of 10 mM NAD -solution (see above) and 250 mL formic
+
acid ester stock solution in order to provide substrate and cofactor
in equimolar amounts. The mixture was agitated for 10 min in a
mM NAD in 100 mM potassium phosphate buffer (250 mL).
The enzyme solution was prepared by diluting the commercial
FDH stock solution in a ratio of 1 : 10 with 100 mM potassium
phosphate buffer at pH 7.5. The standard reaction mixture was
◦
thermo mixer at 30 C and 1000 rpm. The reaction was initiated
by addition of the diluted enzyme solution (see above). In order to
determine extent of conversion 200 mL samples were taken which
were added 20 mL of n-hexanol (500 mM) as internal standard and
diluted up to 1 mL with isopropanol. Degree of conversion was
immediately determined by gas chromatography. For regeneration
◦
then kept at 30 C in a thermomixer for 10 min. Then 50 mL of the
cold enzyme solution (5 U/mL) were added to give a total reaction
volume of 1550 mL. Enzymatic activity was determined after 1
min reaction time by photometric measurements of absorption at
+
of NAD 50 mL (5 U/mL) alcohol dehydrogenase from Candida
3
40 nm. Linearisation and calculation of the Michaelis–Menten-
constant K and the maximum reaction rate vmax were performed
according to the method of Hanes.
boidinii (ADH-CB) and 250 mL of 550 mM acetone were added to
the reaction mixture.
M
39
Conclusion
Influence of temperature and alcohol content on FDH activity
Does FDH catalyse hydrolytic reactions? Sure, but not in a
conventional way. Beyond doubt the enzyme acts oxidising,
The influence of temperature was determined according to the
◦
◦
activity assay as described ranging from 10 C . . . 60 C over a
period of 1 . . . 5 h. Tolerance towards the cosolvent isopropanol
and the alcohols released from 1–6 through the action of FDH was
determined by replacing phosphate buffer by these compounds
within the range of 0 . . . 50% (v/v) in accordance with the activity
assay described above.
although the reaction pathway bears analogies with the AAC
1
hydrolysis mechanism. Providing a final answer to mechanistic
questions will certainly require further mechanistic studies, but
water is definitely involved. In any case, the role of formate
dehydrogenase in biocatalysis has to be reassessed. The enzyme
is not only an invaluable tool in cofactor regeneration, but its
catalytic activity serves well to deprotect alcohol groups esterified
with formic acid. Thereby the reaction proceeds quantitatively,
irreversibly and under mild conditions. With all experiments,
no detrimental effects of the reaction products, which are the
respective alcohols, were observed.
Activity of FDH towards formic acid esters
The activity of FDH at different concentrations of formic acid
esters 1–6 in the range of 0,1–500 mM was analysed spectrophoto-
metrically at 340 nm for 1–10 min. For these purposes formate ester
was dissolved in isopropanol up to a concentration of 500 mM.
In view of the considerable preparative value of the FDH-
method in protective group chemistry, immobilisation experiments
will serve to assign our present findings to a broader application.
2
50 mL of formate ester were mixed with 940 mL sodium phosphate
+
buffer (pH 7.5) and 250 mL of a 10 mM NAD solution. To raise
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7941–7950 | 7949

View Article Online
As can be seen from the results of conversion measurement,
substrate structures determine the kinetic properties of FDH-
mediated formic acid ester cleavage.
16 V. Popov and V. Lamzin, Biochem. J., 1994, 301, 625–643.
1
7 A. Egorov, V. Tishkov, T. Avilova and V. Popov, Biochem. Biophys. Res.
Commun., 1982, 104, 1–5.
1
1
8 L. Uotila and M. Koivusalo, Arch. Biochem. Biophys., 1979, 196, 33–45.
9 J. van Dijken, G. Oostra-Demkes, R. Otto and W. Harder, Arch.
Microbiol., 1976, 111, 77–83.
2
2
2
2
Acknowledgements
0 I. Neben, H. Sahm and M.-R. Kula, Biochim. Biophys. Acta - Enzymol.,
1
980, 614, 81–91.
Financial support by the Deutsche Bundesstiftung Umwelt (con-
tract #13166-32) is gratefully acknowledged. Siloxanes were kindly
provided by Wacker Chemie AG, Consortium f u¨ r elektrochemis-
1 N. Kato, C. Sakazawa, T. Nishizawa, Y. Tani and H. Yamada, Biochim.
Biophys. Acta, Enzymol., 1980, 611, 323–332.
2 T. Matsuda, R. Yamanaka and K. Nakamura, Tetrahedron: Asymme-
try, 2009, 20, 513–557.
+
che Industrie (Munich, Germany). ADH-CB and NAD were
3 R. Castillo, M. Oliva, S. Marti and V. Moliner, J. Phys. Chem. B, 2008,
112, 10012–10022.
kindly provided by Julich Chiral Solutions GmbH–a Codexis
company (J u¨ lich, Germany).
24 V. S. Lamzin, Z. Dauter, V. Popov, E. Harutyunyan and K. Wilson, J.
Mol. Biol., 1994, 236, 759–785.
2
5 H. Slusarczyk, S. Felber, M.-R. Kula and M. Pohl, Eur. J. Biochem.,
2
000, 267, 1280–1289.
References
2
6 M. Ansorge-Schumacher, H. Slusarczyk, J. Sch u¨ mers and D. Hirtz,
1
A. Egorov, T. Avilova, M. Dikov, V. Popov, Y. Rodionov and I. Berezin,
Eur. J. Biochem., 1979, 99, 569–576.
V. Lamzin, A. Aleshin, B. Strokopytov, M. Yukhnevich, V. Popov, E.
Harutyunyan and K. Wilson, Eur. J. Biochem., 1992, 206, 441–452.
V. Tishkov and V. Popov, Biomol. Eng., 2006, 23, 89–110.
FEBS J., 2006, 273, 3938–3945.
27 N. Kato, H. Sahm and F. Wagner, Biochim. Biophys. Acta - Enzymol.,
1979, 566, 12–20.
2
28 EP Pat., 5998, 1979.
3
4
29 S. Alivisatos, F. Ungar and G. Abraham, Nature, 1964, 203, 973–
A. Bommarius, M. Schwarm, K. Stingl, M. Kottenhahn, K. Huth-
macher and K. Drauz, Tetrahedron: Asymmetry, 1995, 6, 2851–2888.
W. Hummel and M.-R. Kula, Eur. J. Biochem., 1989, 184, 1–13.
C. Simons, J. van Leeuwen, R. Stemmer, I. Arends, T. Maschmeyer,
R. A. Sheldon and U. Hanefeld, J. Mol. Catal. B: Enzym., 2008, 54,
975.
30 S. Alunni, A. Cipiciani, G. Fioroni and L. Ottavi, Arch. Biochem.
Biophys., 2003, 412, 170–175.
31 M. Stockdale and M. Selwyn, Eur. J. Biochem., 1971, 21, 416–423.
32 A. Serov, A. Popov and V. Tishkov, Dokl. Biochem. Biophys., 2002, 382,
26–405.
5
6
6
7–71.
P. Savary, Biochim. Biophys. Acta, Lipids Lipid Metab., 1972, 270, 463–
71.
7
33 C. Ingold, in Structure and Mechanism in Organic ChemistryCornell
4
University Press, Ithaca, New York, 1969.
8
9
C. Reese and J. Stewart, Tetrahedron Lett., 1968, 9, 4273–4276.
34 J. Berg, J. Tymoczko, L. Stryer, in BiochemistryW. H. Freeman & Co
Ltd, New York, 2006.
L. De Luca, G. Giacomelli and A. Porcheddu, J. Org. Chem., 2002, 15,
5
153–5155.
35 J. Peters, T. Zelinski, T. Minuth and M.-R. Kula, Tetrahedron:
1
1
0 S. Koeller and J.-P. Lellouche, Tetrahedron Lett., 1999, 40, 7043–7046.
1 M. Dakanali, G. Tsikalas, H. Krautscheid and H. Katerinopoulos,
Tetrahedron Lett., 2008, 49, 1648–1651.
2 H. Hagiwara, K. Morohashi, H. Sakai, T. Suzuki and M. Ando,
Tetrahedron, 1998, 54, 5845–5852.
3 Y. Wu and K. Houk, J. Am. Chem. Soc., 1987, 109, 2226–2227.
4 Y.-D. Wu and K. Houk, J. Am. Chem. Soc., 1987, 109, 906–908.
5 B. Schiott, Y.-J. Zheng and T. Bruice, J. Am. Chem. Soc., 1998, 120,
Asymmetry, 1993, 4, 1683–1692.
36 T. Stillger, M. B o¨ nitz, M. Villela Filho and A. Liese, Chem. Ing. Tech.,
2002, 74, 1035–1039.
1
37 P. Fr o¨ hlich and M. Bertau, Chem. Ing. Tech., 2010, 82, 51–63.
38 G. Gamalevich, B. Morozov, A. Vlasyuk and E. Serebryakov, Tetrahe-
dron, 1999, 55, 3665–3674.
39 C. Hanes, Biochem. J., 1932, 26, 1406–1421.
40 V. Khalameyzer, I. Fischer, U. Bornscheuer and J. Altenbuchner, J.
Appl. Environ. Microbiol., 1999, 65, 477–482.
1
1
1
7
192–7200.
7
950 | Org. Biomol. Chem., 2011, 9, 7941–7950
This journal is © The Royal Society of Chemistry 2011