Engineered phenylalanine dehydrogenase in organic solvents: Homogeneous and biphasic enzymatic reactions

Article abstract of DOI:10.1039/b510816k

The use of engineered phenylalanine dehydrogenase N145A supported on Celite for the reductive amination of phenylpyruvic acid in homogeneous and biphasic aqueous-organic solvents is reported. The results indicate that the immobilised biocatalyst is remarkably robust, even in the presence of high concentrations of polar or non-polar organic solvents such as acetone, methanol, n-hexane, toluene and methylene chloride. Cofactor regeneration with alcohol dehydrogenase from Saccharomyces cerevisiae and ethanol was successfully explored. Application to the non-natural poorly water-soluble 2-oxo acid p-NO ₂-phenylpyruvic acid was successfully performed, resulting in the biocatalytic synthesis of p-NO₂-phenylalanine. In all cases 100% stereoselectivity for the production of the amino acid was retained. The Royal Society of Chemistry 2005.

Full text of DOI:10.1039/b510816k

View Article Online / Journal Homepage / Table of Contents for this issue
A R T I C L E
Engineered phenylalanine dehydrogenase in organic solvents:
homogeneous and biphasic enzymatic reactions
a
b
a
a
a
Gianfranco Cainelli,* Paul C. Engel,* Paola Galletti, Daria Giacomini, Andrea Gualandi
b
and Francesca Paradisi
Department of Chemistry “G. Ciamician”, University of Bologna, Via Selmi 2, I-40126,
a
Bologna, Italy. E-mail: gianfranco.cainelli@unibo.it; Fax: +39.015.2099456;
Tel: +39.051.2099510
Department of Biochemistry, Conway Institute of Biomolecular and Biomedical Research,
b
University College Dublin, Belfield, Dublin, 4, Ireland. E-mail: paul.engel@ucd.ie;
Fax: (+353) 1 283 7211; Tel: (+353) 1 716 6764
Received 1st August 2005, Accepted 10th October 2005
First published as an Advance Article on the web 14th November 2005
The use of engineered phenylalanine dehydrogenase N145A supported on Celite for the reductive amination of
phenylpyruvic acid in homogeneous and biphasic aqueous–organic solvents is reported. The results indicate that the
immobilised biocatalyst is remarkably robust, even in the presence of high concentrations of polar or non-polar
organic solvents such as acetone, methanol, n-hexane, toluene and methylene chloride. Cofactor regeneration with
alcohol dehydrogenase from Saccharomyces cerevisiae and ethanol was successfully explored. Application to the
non-natural poorly water-soluble 2-oxo acid p-NO
2
-phenylpyruvic acid was successfully performed, resulting in the
biocatalytic synthesis of p-NO
2
-phenylalanine. In all cases 100% stereoselectivity for the production of the amino
acid was retained.
Introduction
of the enzyme, allowing the new biocatalysts to accept non-
natural substrates (2-oxo-acids), yielding the corresponding L-
Enzymatic synthesis in non-aqueous media has emerged as an
attractive alternative to chemical synthesis. Numerous strategies
have been developed for carrying out enzymatic reactions in
organic media including multiphase aqueous–organic media,
11,12
amino acids with total retention of the enantioselectivity.
Among the different mutations, the N145A PheDH, in which
asparagines at position 145 is replaced by alanine, proved
to be the most versatile and efficient biocatalyst with several
synthetic 2-oxo-acids. Unfortunately, substituted phenylpyruvic
1
reverse micelles and monophasic media.
Most synthetic uses of enzymes to date, including those in
non-aqueous solvents, have involved hydrolases such as lipases.
The effects of organic media are particularly straightforward
in such cases, especially in terms of the stereoselectivity. It is
known, in fact, that organic solvents can alter the enzymatic
selectivity of hydrolases thus allowing a modulation of the
stereoselectivity control through the solvent. Several examples
of solvent-controlled enantioselection have been reported in
resolution of racemates with lipases. Furthermore, in the
presence of organic solvents lipases show an increased stability
to temperature, an improved control of unwanted side-reactions
acids (e.g. p-MeO-, p-NO
2
- and p-CF -phenylpyruvic acid) have
3
dramatically diminished solubility in water (10- to 20- fold
relative to phenylpyruvic acid), casting a substantial doubt on
the possibility of scaling-up those reactions.
2,3
To overcome this problem, performing the enzymatic reduc-
tive amination in the presence of organic solvents would be
the ideal solution. Toleration of the organic solvents by the
biocatalyst can be judged in two senses. First of all, there is the
basic question of whether the enzyme catalyst is active under
these conditions. Secondly, even where there is a substantial
level of catalytic activity, there is also the issue of longer-term
stability. To our knowledge, even though there have been several
4
5
(
hydrolysis) and the possibility of using highly hydrophobic
substrates insoluble in aqueous phases.
reports of enantioselective reduction of ketones to alcohols, via
In the production of enantiomerically pure compounds, redox
enzymes have a significant advantage over lipases because
they may permit substrate conversion of up to 100%. In this
field many alcohol dehydrogenases display high stereoselectivity
and a broad substrate range. However, dehydrogenases have
been used less often in organic solvents, the main drawback
13,14
alcohol dehydrogenases for example,
there are no previous
studies on amino acid dehydrogenases in organic solvents, and
here we report our latest results on the use of the N145A mutant
immobilised by deposition on Celite with regard to stability,
efficiency and yields in mono- and biphasic systems in presence
of organic solvents. The commercially available phenylpyruvic
acid is used as the substrate in all the screening reactions
6
7
being their limited stability in non-aqueous solvents and the
requirement for a co-factor which is sparingly soluble and
unstable in some organic solvents (e.g. nicotinamide adenine
dinucleotide species are almost insoluble in many lipophilic
(
Scheme 1).
8
media). Notwithstanding these limitations, interesting contri-
butions have recently appeared in the literature on alcohol
9
dehydrogenases.
Phenylalanine dehydrogenase (PheDH) interconverts phenyl-
alanine and phenylpyruvate plus ammonia via a NAD -
Scheme 1
+
dependent oxidoreduction. The wild-type enzyme from Bacillus
sphaericus accepts various non-natural amino substrates in place
of phenylalanine. We have recently reported a successful study
based on this specific phenylalanine dehydrogenase in which
single point mutations greatly affected the substrate specificity
Finally, we report the practical application of our system in
the synthesis of a more hydrophobic non-natural amino acid,
10
p-NO -phenylalanine, starting from the poorly water-soluble p-
2
NO -phenylpyruvic acid.
2
4
3 1 6
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 3 1 6 – 4 3 2 0
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 5

View Article Online
Results and discussion
Enzyme immobilisation
The PheDH used in this work is not commercially available
15
but cultured and purified as previously reported, and usually
◦
stored as an ammonium sulfate precipitate at 4 C. For synthetic
application, however, a much more practical way to handle the
biocatalyst is to have it in a solid form.
Free enzyme powder continues to be used even in nearly
anhydrous media, but immobilization offers an enhanced sta-
16
bilization of enzymes in organic solvents. We have tried two
different preparations: lyophilised and immobilised enzyme. The
lyophilised PheDH lost almost 10% activity with respect to the
15
free enzyme under standard reaction conditions, while when
ꢀ
R
the purified enzyme was immobilised by deposition on Celite
17
5
21 as reported by Persson et al. the retention of activity was
◦
excellent. The supported enzyme stored at 4 C appeared to
be stable for several months. Also, when the freshly prepared
material was redissolved in buffer and centrifuged to remove
the Celite and other solid debris, over 90% of the enzyme
could be recovered in a fully active state. Although in our case
extensive experiments showed that the unsupported enzyme is
extremely stable (results not shown), the augmented volume of
the immobilised biocatalyst simplified the preparation of the
reaction mixture with a more accurate quantification.
Fig. 1
was somewhat slower than in purely aqueous solution, but
still reached almost complete conversion within 2 h. In 50%
acetone, the extent of conversion was much less and the results
suggest that after 2 h the catalytic activity was lost. In methanol,
conversion did not exceed 90%, but nevertheless even in 50%
solvent reaction was clearly proceeding steadily well beyond the
second hour.
In the presence of large amounts of organic co-solvents, KCl
Homogeneous organic co-solvents
and NH
KCl was omitted, and NH
4
Cl precipitated, so that for the main aqueous protocol
OAc replaced NH Cl with good
We first examined organic solvents which form homogeneous
solutions with the aqueous buffer. In the absence of a recycling
system for the cofactor, an excess of NADH, with respect to the
substrate, is necessary to push the reaction to completion. Reac-
tion time-courses were therefore followed in homogeneous co-
solvent systems with THF, acetonitrile, acetone and methanol as
4
4
results in terms of recovery of L-phenylalanine (Table 2).
KCl is usually added to the reaction mixture to improve
the stability of the free enzyme; in this case it seems that
the supported enzyme shows excellent stability and optimum
efficiency even without the salt. However, when pure organic
solvents (such as ethyl ether or hexane), previously saturated
for 24 h with a 50 mM Tris solution (pH 8.5), were used, the
reaction did not proceed at all with or without KCl or with
10, 50, and 90% (v/v) components of the liquid phase. Aliquot
samples were filtered and directly analysed by HPLC using a
chiral column. This allowed a check on both the yield and the
enantiopurity of the product. In all cases the enantioselectivity
of the enzyme afforded exclusively L-phenylalanine and no traces
of D-phenylalanine were ever detected. Results, reported in Fig. 1
and Table 1, are expressed as % conversion. The reactions with
various amount of co-solvents were compared with reactions
in purely aqueous buffered solution, which were efficient and
quantitative in 1 h.
or without NH
4
OAc replacing NH Cl. Finally, replacing am-
4
monium chloride with hydrochlorides of organic amines, (such
as methylamine hydrochloride or benzylamine hydrochloride,
which are soluble in organic solvents) failed to support any
reaction.
Homogeneous organic co-solvents with cofactor regeneration
In THF, at only 10% the reaction proceeded more slowly
than in Tris buffer alone, but negligible reaction was seen
with 50 or 90% THF, which forms a biphasic system at these
higher concentrations. The enzyme was also found to be very
sensitive to the purity of this solvent: reactions performed in
THF other than freshly distilled failed to yield any product.
The breakdown products of THF, such as butyraldehyde, 4-
hydroxybutyraldehyde and several others as reported in the
literature seem to be responsible for destroying the enzyme
activity. When the THF is freshly distilled the reaction proceeds
up to 99.5% conversion.
The reaction in acetonitrile was slow, but nevertheless pro-
ceeded even in 50% organic solvent. In 10% acetone, reaction
Since the design of economically useful enzymatic synthesis
requires cofactor regeneration, the possibility of recycling
Table 2 Conversion (%) of phenylpyruvic acid into L-phenylalanine
using Celite-supported enzyme in aqueous Tris buffer solution (pH 8.5),
no co-solvents and different ammonia source
Time
5 min
Tris–NH
4
Cl
Tris–NH
4
OAc–KCl
4
Tris–NH OAc
18
1
51.7
60.8
92.0
97.0
66.4
71.2
82.8
94.1
45.7
75.1
93.2
97.6
1
2
3
h
h
h
Table 1 Conversion (%) of phenylpyruvic acid into L-phenylalanine using Celite-supported enzyme in a homogeneous system of aqueous Tris buffer
solution (pH 8.5) and different amounts of co-solvents
THF THF THF CH
3
CN CH
3
CN CH
3
aCN Acetone
10%
Acetone
50%
Acetone
90%
MeOH MeOH MeOH
a
a
a
Time/min Tris
10%
50%
90%
10%
50%
90%
10%
50%
90%
1
6
20
5
0
80.3
100
100
100
19.2
45.9
64.2
99.5
—
—
—
—
—
—
—
—
24.0
28.7
28.4
30.9
12.3
18.7
23.6
28.2
—
—
—
—
32.9
88.1
99.7
93.0
6.8
11.6
16.7
16.4
—
—
—
—
34.5
81.7
79.9
87.1
18.5
47.2
65.3
85.2
—
—
—
—
1
1
440
a
With a 90 : 10 mixture vs. Tris, the conversion reached 1% after 1 week.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 3 1 6 – 4 3 2 0
4 3 1 7

View Article Online
Fig. 2
with alcohol dehydrogenase from Saccharomyces cerevisiae and
ethanol was explored (Fig. 2).
Fig. 3
Under these circumstances, despite the sub-stoichiometric
(
20%) NADH concentration relative to phenylpyruvate, virtu-
ether. Strict enantioselectivity is retained in all cases as only L-
phenylalanine was detected in the crude reaction mixture.
The trial of immiscible solvents was extended by reversing the
ratio and using 80% of organic phase with only 20% aqueous.
The results show that conversion rates were faster, reaching 50–
ally complete conversion to L-phenylalanine was again accom-
plished in the aqueous system, showing that the recycling system
with ethanol and alcohol dehydrogenase is effective. In general,
the results in the presence of organic solvents are not dissimilar
from those of Table 1 (NADH in excess and no recycling) and in
fact show somewhat higher conversion in most cases (Table 3).
Fig. 2 also includes results for ethanol, with near complete
conversion in 10% ethanol and almost 70% conversion even
in 50% ethanol.
6
0% conversion in most cases in the first hour, possibly because
of the higher coenzyme concentrations in the 20% aqueous phase
where the reaction takes place, or more efficient inter-phase
transfer of the substrate. With the possible exception of toluene,
high final extents of conversion were achieved even with these
4
: 1 organic–aqueous systems.
Biphasic organic co-solvents
A number of solvents immiscible with water were tested in
biphasic systems in a ratio of 20% organic phase added to
Biphasic system with co-factor regeneration
8
0% aqueous reaction mixture. These included tert-butylmethyl
ether, dichloromethane, toluene, n-hexane, and diethyl ether
as members of different classes: oxygenated and halogenated
solvents, aromatic and aliphatic hydrocarbons (Fig. 3).
In experiments to test whether the enzymatic recycling system
for the cofactor would also tolerate the organic solvents, in the
case of 20 : 80 organic–aqueous mixtures all incubations gave
a high yield of product, the lowest being 93% after 24 h with
hexane (Fig. 4).
On the other hand, when the organic : aqueous ratio was
reversed, the yields in the biphasic systems after 24 h were in
all cases less than 90% (Table 5). Comparison with the results
in Table 4 suggests that the poorer result here probably reflects
sensitivity of the recycling system (alcohol dehydrogenase) to
the high concentrations of organic solvent.
The organic and the aqueous phases were separately analysed.
The former contained phenylpyruvic acid and a small amount
of NADH while the L-phenylalanine was in the aqueous phase
with most of the NADH. Although reaction was slower under
these conditions (50–60% after 6 h), the enzymatic system
appears to tolerate the presence of an immiscible organic solvent
well, with over 90% conversion in every case (Table 4) and
virtual completion with methylene dichloride and t-butyl methyl
Table 3 Conversion (%) of phenylpyruvic acid into L-phenylalanine using Celite-supported enzyme in homogeneous system of aqueous Tris buffer
solution (pH 8.5) with different co-solvents and co-factor regeneration
THF
10%
THF
50%
CH
10%
3
CN
CH
50%
3
CN
Acetone
10%
Acetone
50%
MeOH
10%
MeOH
50%
EtOH
10%
EtOH
50%
Time
Tris
1
2
5 min
47.9
58.5
95.6
98.4
29.5
39.8
41.3
45.8
22.9
32.6
35.2
37.6
30.0
36.3
40.9
38.7
28.4
33.0
37.1
40.5
67.8
92.1
98.7
98.6
11.5
12.4
18.0
19.4
84.6
98.6
91.5
98.5
32.2
37.3
40.3
45.7
32.2
39.0
48.7
95.1
31.0
27.6
36.7
67.6
1
2
h
h
4 h
Table 4 Conversion (%) of phenylpyruvic acid into L-phenylalanine using Celite-supported enzyme in biphasic systems of aqueous Tris buffer
solution (pH 8.5), 20 and 80% of different co-solvents
Hexane
20%
Hexane
80%
Toluene
20%
Toluene
80%
Et
20%
2
O
Et
80%
2
O
tBuOMe
20%
tBuOMe
80%
CH
20%
2
Cl
2
2 2
CH Cl
80%
Time
1
5 min
12.6
27
38.1
54.2
93.1
20.4
39.5
45.6
57.9
95.5
23.6
37
44.1
51.3
95
32.4
82.5
88.4
86.5
86.7
7.7
26.5
55.6
63.1
90.9
23
18.7
40.8
42.6
50.5
99.7
34.7
79
93.4
94.4
98.8
20.2
41.7
44.5
50.9
97.9
42.2
84
85.9
88.5
95.2
1
2
6
h
h
h
46.5
84.6
83.4
97.3
2
4 h
4
3 1 8
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 3 1 6 – 4 3 2 0

View Article Online
Table 5 Conversion (%) of phenylpyruvic acid into L-phenylalanine with co-factor regeneration using Celite-supported enzyme in biphasic systems
of aqueous Tris buffer solution (pH 8.5), 20 and 80% of different co-solvents
Hexane
20%
Hexane
80%
Toluene
20%
Toluene
80%
Et
20%
2
O
Et
80%
2
O
tBuOMe
20%
tBuOMe
80%
CH
20%
2
Cl
2
2 2
CH Cl
80%
Time
Tris
1
2
5 min
47.9
58.5
95.6
98.4
18.5
28.1
34.3
93.3
18.4
23.6
35.9
89.9
24.5
30.1
31.9
96.7
38.7
50.7
46.7
70.6
20.3
33.7
37.1
96.9
48.4
54.1
61.9
71.3
18.9
33.7
46.1
98.1
25.7
51.1
81.7
85.1
18.6
27.5
28.1
96.5
25.2
38.5
49.1
85.5
1
2
h
h
4 h
(89.3%) in reaction A (5% EtOH) and 16 mM (67%) in reaction
B (5% EtOH and 10% MeOH). Injections repeated after 43 h did
not show increased yields and the apparent shortfall from 100%
may be attributed to the separation of a portion of the total
product as precipitate. Analysis of the crude reaction mixture
1
by H-NMR did not show any traces of the starting material in
reaction A, confirming total conversion, while 5–10% unreacted
oxo-acid was still detectable in the spectrum of reaction B.
Conclusion
Overall, the results indicate that the immobilised biocatalyst is
remarkably robust, even in the presence of high concentrations
of non-polar organic solvents, offering a good prospect for
its use even with poorly water-soluble substrates. It has been
12,19
shown elsewhere
that the range of substrate specificity
Fig. 4
of phenylalanine dehydrogenase may be greatly extended by
site-directed mutagenesis. However, the utility of such novel
biocatalysts could, in principle, be limited by the insolubility
of some of the target substrates. It is therefore of considerable
practical importance that in the immobilised state, as used here,
phenylalanine dehydrogenase will tolerate concentrations of
organic solvent in homogeneous systems that will greatly extend
the solubility of such substrates. In fact, we have successfully
exemplified how the synthesis of the non-natural amino acid
Application to a non-natural oxo-acid: synthesis of
p-NO -phenylalanine
2
As mentioned above, our mutated PheDH is efficient with
several non-natural 2-oxoacids. However, there were aqueous
solubility concerns in many cases. To further explore the
improved performance of the supported N145A PheDH in
organic solvents, several experiments were carried out to convert
p-NO -phenylalanine can be scaled-up in the presence of
2
the commercially available p-NO -phenylpyruvic acid into the
2
organic solvent: our results show that it is possible to work
at much higher concentrations, with correspondingly higher
yields. This favourable outcome is reinforced by the results
obtained with biphasic systems where the enzyme remains in the
predominantly aqueous phase and the organic phase serves as a
reservoir for the substrate. As expected, in no case was the strict
enantioselectivity of the catalyst compromised. Observation of
the reactions over several hours confirms that the catalytic
capacity of the enzyme persists, although there are indications
of declining activity in some cases.
corresponding amino acid. The solubility of such substrates in
fact, is greatly improved by the addition of organic solvents.†
Two different reactions were compared: (A) the reductive
amination of 10 mg p-NO -phenylpyruvic acid in 3 mL 50 mM
Tris buffer (pH 8.5) with 5% EtOH and (B) of 15 mg of the same
substrate in 3 mL 50 mM Tris buffer (pH 8.5) with 5% EtOH and
2
1
0% MeOH (same reaction conditions as reported for cofactor
regeneration in homogeneous systems). Owing to the formation
of a precipitate, samples were centrifuged and supernatant and
precipitate were analysed separately. The product concentration
in the supernatant is reported in Table 6.
In both cases, the precipitate was redissolved in fresh buffer
and it was confirmed by HPLC to be the amino acid product.
Experimental
Materials
Concentration of p-NO
2
-phenylalanine in solution was 14.2 mM
All organic solvents, L-phenylalanine and phenylpyruvic acid
ꢀ
R
Table 6 Conversion of p-NO
2
-phenylpyruvic acid into p-NO
2
-L-
were analytical grade purchased from Sigma-Aldrich, Celite
phenylalanine using Celite-supported enzyme in a homogeneous system
of aqueous Tris buffer solution (pH 8.5), 10% MeOH with co-factor
regeneration. Values are expressed as % of the total expected conversion
and also as product concentration (mM)
521 was purchased from Sigma-Aldrich, NADH was ob-
tained from Roche Diagnostic, and Saccharomyces cerevisiae
alcohol dehydrogenase from Sigma. p-NO
was purchased from Frinton Laboratories, Inc., while p-NO
2
-phenylpyruvic acid
2
-
phenylalanine was obtained from Bachem. Mutant pheny-
lalanine dehydrogenase N145A was prepared as described
elsewhere.
Time/h
Tris
MeOH 10%
15
1
2
4
8
4
3
89.3% (14.2 mM)
89.3% (14.2 mM)
89.3% (14.2 mM)
67% (16.0 mM)
67% (16.0 mM)
67% (16.0 mM)
Immobilisation of the enzyme
−
1
The purified enzyme (0.6 mg mL ) was taken directly from
the eluate of the Procion Red P3BN chromatography column
in potassium phosphate buffer (20 mM), pH 7.9, containing
−
1
†
(
In pure Tris buffer we estimated a maximum solubility of 1.8 mg mL
ca. 9 mM). Simply by adding 5% EtOH necessary to use the recycling
15
−
1
NaCl (0.5 M), without precipitating with ammonium sulfate.
mL of this protein solution was added to Celite (1 mg). The
system, the solubility improved by almost 2 fold (3.3 mg mL , ca.
6 mM). MeOH was selected among the miscible organic solvents for
this experiment at a 10% ratio. Under these conditions the solubility of
2
1
preparation was dried for approx. 18 h under vacuum and then
−
1
◦
the substrate was increased to 5 mg mL (ca. 24 mM).
stored at 4 C.
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 3 1 6 – 4 3 2 0
4 3 1 9

View Article Online
Reactions in homogeneous co-solvents
above. As the reactions proceeded, the formation of an insoluble
product was observed. In a different study (unpublished results),
For these experiments a freshly prepared stock reaction mixture
we observed that p-NO -phenylalanine was extremely insoluble
2
was used consisting of KCl (200 mM), NH
NADH (2 mM) and phenylpyruvic acid (1 mM) in Tris
50 mM, pH 8.5). To 3 mL of this solution different amounts
of homogeneous organic co-solvents: THF, CH CN, acetone
4
Cl (800 mM),
in aqueous medium (2 mM) and that in presence of 10% MeOH
we could achieve a concentration of 10 mM. Thus, it may be
deduced that, in the present experiment, because of the high
concentration of substrate, part of the product precipitated.
After 24 h, samples of the reactions were centrifuged and
supernatants were injected into the HPLC column to assess
enantiopurity, and product concentration. The precipitated
product was also analysed by HPLC. The crude reaction mixture
(
3
and MeOH were added. The reactions were initiated with
the enzyme-impregnated Celite (3 mg) and reaction mixtures
were magnetically stirred in capped vials at room temperature.
When cofactor regeneration was employed, the reaction mixture
differed from that above in the addition of ethanol (5%) and
EDTA (1 mM) and also the use of a lower NADH concentration
1
was then evaporated and analysed by H-MNR.
(
0.2 mM). When reaction was initiated as above, yeast alcohol
Quantitative product analysis
dehydrogenase (1 mg) was also added.
Formation of L-phenylalanine and p-NO
monitored by HPLC analysis on a chiral column CHIROBI-
OTIC T (250 × 4.6 mm, eluent H O–EtOH 1 : 1, flow rate
2
-phenylalanine was
Reactions in biphasic systems
The freshly prepared stock reaction mixture was as for the
experiments of Table 1. In the first set of experiments (20%)
organic solvent (0.75 mL) (tBuOMe, CH
2
−
1
1
mL min , l = 208 nm). At different times, aliquot samples were
filtered on a PTFE filter (0.20 m), diluted and directly injected.
Calibration curves obtained with commercial L-phenylalanine
2
2
Cl , toluene, hexane
or diethyl ether) was added to this solution (3 mL). To minimize
changes in the phase volumes, the organic solvent was previously
saturated with the Tris solution before use. The reactions were
initiated as before with immobilised enzyme and the mixtures
were efficiently magnetically stirred in capped vials at room
temperature. In a second set of experiments the organic : aqueous
ratio was reversed. To maintain the same total amounts of
reactants in the combined volumes of both phases, the stock
solution in Tris (50 mM, pH 8.5) was modified as follows: KCl
and p-NO -phenylalanine were used for quantitative analysis.
2
Acknowledgements
This work was supported by MURST (ex-60% and FIRB),
University of Bologna (Funds for selected topics) and in part by
a grant from the Advanced Technology Research Programme of
Enterprise Ireland.
(
1600 mM), NH Cl (3200 mM), phenylpyruvic acid (4 mM)
4
and NADH (8 mM): (80%) of the organic solvent (6 mL) and
the supported enzyme (6 mg) were added to the stock solution
References
1
2
A. Klibanov, Nature, 2001, 409, 241–246.
G. Cainelli, V. De Matteis, P. Galletti, D. Giacomini and P. Orioli,
Chem. Commun., 2000, 23, 2351–2352.
G. Cainelli, P. Galletti, D. Giacomini, A. Gualandi and A.
Quintavalla, Helv. Chim. Acta, 2003, 86, 3548–3559.
(
1.5 mL).
The recycling system for cofactor was also tested as follows:
the supported enzyme (3 mg) and yeast alcohol dehydrogenase
3
(
(
1 mg) were added to the aqueous Tris buffer stock solution
3 mL), followed by (20%) of tBuOMe, CH Cl , toluene or
2
2
4 G. Carrea and S. Riva, Angew. Chem., Int. Ed., 2000, 112, 2226–2254.
5 A. Ghanem and H. Y. Aboul-Enein, Tetrahedron: Asymmetry, 2004,
diethyl ether (0.75 mL), and in this case the stock solution,
freshly prepared in Tris buffer (50 mM, pH 8.5), contained: KCl
1
5, 3331–3351.
6
7
8
W. Hummel, Trends Biotechnol., 1999, 17, 487–492.
(
200 mM), NH Cl (800 mM), phenylpyruvic acid (1 mM),
4
A. M. Klibanov, Curr. Opin. Biotechnol., 2003, 14, 427–431.
D. Metrangolo-Ru `ı z de Teminio, W. Hartmeier and M. B.
Ansorge-Schumacher, Enzyme Microb. Technol., 2005, 36, 3–9.
NADH (0.2 mM), EDTA (1 mM) and 5% EtOH.
Moving to the 80 : 20 organic: aqueous ratio we changed
the stock solution as follows: freshly prepared stock solution
in Tris buffer (50 mM, pH 8.5) contained KCl (1600 mM),
9 H. Gr o¨ ger, W. Hummel, S. Buchholz, K. Drauz, T. Van Nguyen, C.
Rollmann, H. Husken and K. Abokitse, Org. Lett., 2003, 5, 173–176.
0 Y. Asano, A. Nakazawa and K. Endo, J. Biol. Chem., 1987, 21,
1
1
1
NH
4
Cl (3200 mM), phenylpyruvic acid (4 mM), NADH
1
0346–10354.
(
0.8 mM), EDTA (4 mM) and 5% EtOH. To the Tris stock
1 S. Y. K. Seah, K. L. Britton, D. W. Rice, Y. Asano and P. C. Engel,
Biochemistry, 2002, 41, 11390.
2 P. Busca, F. Paradisi, E. Moynihan, A. R. Maguire and P. C. Engel,
Org. Biomol. Chem., 2004, 2, 2684–2691.
solution (1.5 mL) the supported enzyme (6 mg), alcohol
dehydrogenase from Saccharomyces cerevisiae (2 mg) and (80%)
of one of the following solvents tBuOMe, CH
2
2
Cl , toluene, and
diethyl ether (6 mL) were added.
13 J. Grunwald, B. Wirz, M. P. Scollar and A. M. Klibanov, J. Am.
Chem. Soc., 1986, 108, 6732–6734.
1
1
4 J. B. Jones and T. Takemura, Can. J. Chem., 1984, 62, 77–80.
5 Y. K. Seah, K. L. Britton, P. J. Baker, D. W. Rice, Y. Asano and P. C.
Engel, FEBS Lett., 1995, 370, 93–96.
Synthesis of p-NO -phenylalanine
2
The buffers for these reactions were prepared as described above
for the reaction with homogeneous co-solvent and cofactor
regeneration. The substrate concentration differed in the two
experiments as follows: when no co-solvent other than 5% EtOH
was added to the reaction mixture, p-NO -phenylpyruvic acid
was 16 mM, while when 10% MeOH is present, the substrate
concentration was 24 mM. Reactions were started as described
1
1
1
6 M. Persson, E. Wehtje and P. Adlercreutz, ChemBioChem, 2002, 3,
5
66–571.
7 M. Persson, E. Wehtje and P. Adlercreutz, Biotechnol. Lett., 2000,
2
2, 1571–1575.
2
8 T. Morikawa and K. Yoshida, Kagaku to Kogyo, 1963, 37, 107.
19 P. C. Engel, A. R. Maguire, F. Paradisi, O. McCrohan, D. Giacomini,
P. Galletti, Ir. Pat. S2004/0547, 2004.
4
3 2 0
O r g . B i o m o l . C h e m . , 2 0 0 5 , 3 , 4 3 1 6 – 4 3 2 0