The first fluorogenic assay for detecting a Baeyer-Villigerase activity in microbial cells

Article abstract of DOI:10.1039/b306687h

The first fluorogenic assay allowing for detection of microbial enzymes able to perform Baeyer-Villiger oxidation is described. This is based on the use of 4-oxopentyl umbelliferyl ether 1 as a fluorogenic substrate. When Baeyer-Villigerases active against this test ketone are present in the selected whole cells, 1 is transformed into 3-hydroxypropyl umbelliferyl ether 3, which, in a subsequent step, releases the fluorescent product umbelliferone. Different microorganisms, known to be endowed with Baeyer-Villigerase activity, were assayed.

Full text of DOI:10.1039/b306687h

View Article Online / Journal Homepage / Table of Contents for this issue
The first fluorogenic assay for detecting a Baeyer–Villigerase
activity in microbial cells
María C. Gutiérrez, Arthur Sleegers, Helen D. Simpson, Véronique Alphand and
Roland Furstoss*
Groupe Biocatalyse et Chimie Fine UMR-CNRS 6111, Université de la Méditerranée,
Faculté des Sciences de Luminy, case 901, 163 Avenue de Luminy, 13288 Marseille, Cedex 9,
France. E-mail: furstoss@luminy.univ-mrs.fr
Received 16th June 2003, Accepted 14th July 2003
First published as an Advance Article on the web 16th September 2003
The first fluorogenic assay allowing for detection of microbial enzymes able to perform Baeyer–Villiger
oxidation is described. This is based on the use of 4-oxopentyl umbelliferyl ether 1 as a fluorogenic substrate.
When Baeyer–Villigerases active against this test ketone are present in the selected whole cells, 1 is transformed
into 3-hydroxypropyl umbelliferyl ether 3, which, in a subsequent step, releases the fluorescent product
umbelliferone. Different microorganisms, known to be endowed with Baeyer–Villigerase activity, were
assayed.
assay, keeping in mind that such an assay should also be (a)
Introduction
reliable, (b) enzyme specific and (c) fast and easy to use.
A process involving a fluorogenic test substrate should meet
these requirements best. Elegant applications of such a strategy
have been documented previously, for example by Reymond
and coworkers for detecting alcohol dehydrogenase,10 as well as
lipase or esterase, aldolase and epoxide hydrolase activity.11 In
this paper, we describe the first fluorimetric test allowing for
the detection of Baeyer–Villigerase activity in microbial whole-
cells.
The use of a biotransformation strategy for achieving fine
organic synthesis is nowadays recognized as an indisputable
fact.1 This is due to both the fantastic potential of enzymes as
(chiral) catalysts and, also, to the fact that such approaches can
be regarded as attractive “green chemistry” alternatives to
chemical reactants or catalysts. In this context, the main issue is
obviously the selection of an appropriate enzyme. Recently,
several High Throughput Screening (HTS) techniques have
been elaborated, which allow the very rapid screening of
libraries of thousands of potential enzyme sources within
collections of wild type microorganisms or libraries of clones
obtained via molecular biology techniques.2,3 Obviously, the
crucial key point of such screening approaches lies in the
availability of an appropriate assay allowing selection of
the proper biocatalyst.
We have been interested for several years in studying
asymmetric Baeyer–Villiger (BV) oxidation (a type of reaction
which still proves to be very difficult, if not impossible, using
conventional chemistry)4 by way of a biocatalytic approach.5–8
To the best of our knowledge, a first (colorimetric) assay, which
only applies to monocyclic ketones and is limited to the milli-
molar concentration range of test substrate, has been recently
described for the detection of BV monooxygenase (BVMO)
activity.9 Due to the intrinsic limitations of this first assay, we
decided to try to set up a more sensitive Baeyer–Villigerase
Results and discussion
Assay principle
Our strategy was based on the use of ketone 1 as a fluorogenic
test substrate. The principle of the procedure is illustrated in
Scheme 1. It consists of a two-step process involving: step 1: the
biocatalysed Baeyer–Villiger oxidation of 1 by a putative native
Baeyer–Villigerase present in the cell (followed by the spon-
taneous, or enzyme catalysed, hydrolysis of the thus primarily
formed acetate 2 into the alcohol intermediate 3); step 2: oxi-
dation of this alcohol into the aryloxyaldehyde 4,12 which, upon
spontaneous β-elimination, is known to lead to the fluorescent
umbelliferone (5). This second step could be performed by
using an oxidative enzyme, i.e. an alcohol dehydrogenase in the
presence of added NADH and of bovine serum albumin (BSA)
Scheme 1 Principle of the fluorogenic assay: (a) enzymatic BV oxidation; (b) spontaneous or enzymatic hydrolysis; (c) enzymatic (HLADH–
NADϩ) or chemical (TEMPO–NaOCl) oxidation; (d) β-elimination.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
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 3
3500

View Article Online
as previously described by Reymond.10 However, due to the
rather slow reaction rate of this second step, we also were inter-
ested in exploring the possibility of performing this oxidation
using a chemical oxidant, an approach which we hoped would
make this second step more efficient, easier and cheaper to
carry out on a large scale (i.e. for directed evolution approaches
implying thousands of strains for example). We here describe
the results we have obtained along these lines.
9871.20,21 Analytical (or small preparative scale) biotransform-
ations of 1 using these different bacteria were performed and
monitored by HPLC. Our results indicate that 1 was efficiently
oxidised by most of these strains. Thus, (a) Acinetobacter sp.
TD63 afforded a 90% isolated yield of alcohol 3 after 16 hours;
(b) Pseudomonas sp. NCIMB 9872 produced 3 with an excellent
analytical yield (90%) after 7 hours (after 24 hours this alcohol
concentration decreased slightly); (c) Arthrobacter sp. M5
performed total oxidation of 1 after 4 hours, leading to an 80%
analytical yield of 3. (As in the case of Pseudomonas sp., the
concentration of 3 decreased with time and the product dis-
appeared within 24 h). In some cases, traces of the primarily
formed ester intermediate 2 were also detected indicating
that, as expected, spontaneous hydrolysis was indeed
occurring. However, 1 was not oxidised by the wild type strain
of A. calcoaceticus NCIMB 9871, nor by the recombinant E.
coli TOP10 [pQR239], indicating that, most probably, 1 is not a
substrate for this specific enzyme. Thus, 85% of 1 (analytical
HPLC) could be detected after 24 hours biotransformation in
spite of the fact that, for both strains, the Baeyer–Villigerase
activity was checked to be present against bicyclo[3.2.0]hept-
2-en-6-one,14 a well known substrate of CHMO.
Preliminary experiments
In order to define the most appropriate experimental con-
ditions, we decided to check separately each one of the two
steps of the proposed procedure. This necessitated first the
synthesis of the test ketone 1, as well as the preparation of
an authentic sample of the (putative) alcohol intermediate 3.
We then studied separately: (a) the biotransformation of 1
using different bacterial strains; (b) the optimisation of the
parameters involved in the oxidation of 3. The test ketone 1, i.e.
7-(4-oxopentoxy)-2H-benzopyran-2-one, was synthesised from
umbelliferone (5) as described in Scheme 2. Compound 3, i.e. 7-
(3-hydroxypropoxy)-2H-1-benzopyran-2-one, was synthesised
by condensation of 5 with tetrahydropyranyl 3-bromopropyl
ether, as described in Scheme 3.
Step 2: oxidation of alcohol 3
As far as the second step of this assay is concerned, two possi-
bilities were envisaged: (a) the use of an enzymatic approach
(as already described11 for the detection of other types of
enzymatic reactions); (b) the application of
oxidation process.
a chemical
Enzymatic oxidation of 3. The possibility of achieving the
oxidation of 3 using an alcohol dehydrogenase was explored.
Different commercial enzymes (including yeast alcohol
dehydrogenases (YAD), horse liver alcohol dehydrogenase
(HLADH) as well as the Lactobacillus kefir alcohol dehydro-
genase) were tested. In all cases but one (the L. kefir enzyme), a
time-dependent appearance of fluorescence was observed upon
oxidation of 3. HLADH proved to be the most efficient. This
was further optimised and it appeared that special attention
had to be paid to the buffer composition and pH. Two types of
buffers, i.e. a 20 mM borate buffer as well as a 100 mM glycine–
NaOH buffer, were assayed with the different enzymes. The best
results were always found in the latter buffer. Furthermore,
experiments conducted in the glycine–NaOH buffer showed
that appearance of fluorescence, i.e. formation of umbelliferone
5, occurred within about 1.5 hours at pH 7.7, but was roughly
four times faster at pH 9.6. However, the maximum
fluorescence intensity obtained at pH 9.6 was not higher than
50% of the theoretical maximum (Fig. 1). Since we have verified
that ketone 1 was perfectly stable under these experimental
conditions, this discrepancy can be due to either incomplete
oxidation of 3 (as confirmed by TLC analysis) and/or to notice-
able decomposition of 5. Test experiments using commercial
5 indeed indicated a decrease of fluorescence of about 25%
per hour under these experimental conditions (Fig. 1). Inter-
estingly, however, this decrease slowed down and became
negligible after about 2 hours.
Scheme 2 Synthesis of test compound 1. (a) NaH, DMF, 110 ЊC, 12 h,
N2, 75%; (b) Amberlite IR-120; acetone, r.t., 81%.
Scheme 3 Synthesis of alcohol 3. (a) NaH, DMF, 70 ЊC, 12 h, N2,
70%; (b) p-TsOH, MeOH–H2O, r.t., 64%.
Step 1: biotransformation of 1 using different bacterial strains
The main drawback of this type of assay is the possibility that
the test ketone would not be a substrate for the concerned
enzymes. In our case, this point deserved special attention
because only a few linear ketones have been described as
Baeyer–Villigerase substrates.5,7,13 Biotransformation of 1 (at a
typical 1 mM substrate concentration) was therefore checked
with a set of different bacteria previously used for enzymatic
BV oxidation.14–16 These included the wild type strains: Acineto-
bacter calcoaceticus NCIMB 9871,17 Acinetobacter sp. TD63,
Pseudomonas sp. NCIMB 987218 and Arthrobacter sp. M5,19 as
well as a recombinant E. coli TOP10 [pQR239] host recently
constructed to overexpress the well known cyclohexanone
monooxygenase (CHMO) from A. calcoaceticus NCIMB
Chemical oxidation of 3. As pointed out above, we also were
interested in exploring the possibility of performing this second
step (oxidation of 3) by using a chemical reagent instead of an
enzyme. An improvement of this step would be of general
application to the type of assays described by Reymond
et al., whatever enzymatic activity was looked for, i.e. lipase,
esterase... The problem was, therefore, to find a chemical
reactant allowing the oxidation of alcohol 3 into aldehyde 4 in
aqueous solution. Moreover, it was necessary to find a method
where no over-oxidation of
4 (into the corresponding
carboxylic acid) would occur, since this acid itself does not
undergo β-elimination.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3501

View Article Online
solution of TEMPO in CH2Cl2 and of a (3%) aqueous NaOCl
solution. Test ketone 1 appeared to be stable in these con-
ditions. Using these experimental conditions, chemical oxid-
ation of 3 appeared to be noticeably faster and more efficient
than the enzymatic process. This led to a nearly 3 times more
intense fluorescent response, this being reached after only
30 minutes as compared to 90 minutes for the enzymatic
process. However, as in the above described enzymatic
approach, again only 50% of the theoretical fluorescence
intensity was reached.
Assay validation
In order to validate this assay, the biotransformation experi-
ments on 1 performed using the five described strains and
checked by analytical HPLC as described above, were analysed
using the following typical procedure. Biotransformations using
wild type bacteria were carried out directly in the culture
medium, after the appropriate growth period. In the case of
the recombinant strain, harvested cells were centrifuged and
re-suspended in the same volume of phosphate buffer (50 mM,
pH 7.5); together with 0.05% of glycerol. Ketone 1 was dis-
solved in the minimum volume of ethanol and was added to
each biotransformation medium to obtain a final concentration
of 1 mM. Aliquots (1 mL) were withdrawn at different time
intervals, centrifuged and frozen. They were thawed at room
temperature just before testing and placed into a (polypropyl-
ene) 96 well plate (5 µl per well). All samples were tested by
performing in parallel both enzymatic and chemical oxidation
procedures. The results obtained over a 1.5 hour period are
presented in Fig. 3. As can be seen, both types of method-
ologies, enzymatic or chemical based oxidation, afforded com-
parable qualitative results. However, as a general feature, chem-
ical oxidation led to a faster and more intense response. The
fluorescence intensity increased in all cases up to a maximum,
which however was always lower than the value obtained from
the control experiment achieved using a 1 mM (50% aqueous
MeCN) solution of 3 in the absence of cells. This might be due
to adsorption of the produced alcohol inside the cell, which
would protect it from oxidation. We have found that addition of
inactivated cells into the standard alcohol 3 solution prior to
addition of the oxidant indeed led to a noticeable decrease of
the fluorescence signal intensity. As can be seen in Fig. 3, all
samples of 1 biotransformed by Acinetobacter sp. TD63 and
Pseudomonas sp. NCIMB 9872 produced intense fluorescent
responses due to efficient biooxidation of 1 by these strains.
Fluorescence appearance, as well as decay, was in perfect
agreement with the results obtained by HPLC analysis. Thus,
samples obtained upon biotransformation of 1 by Arthrobacter
sp. led to a signal whose intensity decreased after 3 hours and
totally disappeared after 24 hours. Blank experiments con-
ducted using the non-induced cells (lacking this enzymatic
activity) did not exhibit a noticeable fluorescent signal. As
expected from HPLC analysis, the recombinant E. coli TOP10
[pQR239] bacteria overexpressing the well known CHMO
enzyme, did not lead to any noticeable fluorescent response, due
to the fact that 1 is not oxidised by this enzyme. This result
interestingly illustrates the potential bottleneck of any such
screening assay based on the use of a specific test compound
which, as a general feature, can lead to so called “false negative
answers” if the test compound is not recognised as a substrate
by a specific enzyme. In principle, another limitation of such an
assay could of course be the fact that other types of enzymes
would lead to “false positive responses”. Although this hypo-
thesis cannot be totally eliminated, this seems highly unlikely
due to the two-step nature of this assay, where appearance of
fluorescence (in step 2) necessitates prior formation (in step 1)
of a fluorogenic (and stable) intermediate, i.e. the alcohol 3.
Fluorescence detected in the first step is not considered as a
positive result. In any case, further more detailed analysis is
Fig. 1 Fluorescent signal observed upon enzymatic oxidation of
alcohol 3. Fluorescence signal (arbitrary units) observed at λex 360 20
nm and λem 460 20 nm during enzymatic oxidation of alcohol 3 (10
µM) to umbelliferone 5 in a 100 mM glycine–NaOH buffer. HLADH
(5 µg mLϪ1), NADϩ (1 mM) and BSA (2 mg mLϪ1): ꢀ pH 9.6, ᭝ pH
7.7. Fluorescence signal (arbitrary units) observed, using the same
conditions, with 10 µM umbelliferone (5): ᭿ pH 9.6.
Albeit numerous reactants enabling oxidation of a primary
alcohol are available, methods allowing the performance of
such a reaction in the aqueous phase are surprisingly scarce.
Preliminary experiments led us to select the TEMPO–NaOCl
oxidative combination.22 Thus, a bicarbonate buffer (0.5 M, pH
9.2) solution of 3 containing BSA (1.3 mg mLϪ1) was distrib-
uted in a 96 well plate and both a methylene chloride solution
of TEMPO and a 3% aqueous sodium hypochlorite solution
were added to each well. We thus could observe very fast
appearance of fluorescence. This is due to oxidation of 3
followed by spontaneous β-elimination of the intermediate
formed (aryloxyaldehyde 4), thus leading to formation of
umbelliferone 5. Again, optimisation of this process indicated a
noticeable influence of the pH value, as well as of the ionic
strength, on the reaction rate and signal intensity. Various
aqueous (pH 9.2) bicarbonate buffer solutions of different ionic
strength were tested, indicating that the reaction rate increased
with ionic strength. Oxidation conducted in a bicarbonate
buffer of 550 mM concentration showed the highest rate
(Fig. 2), whereas use of higher amounts of oxidant (or co-
oxidant) did not produce any significant increase of fluor-
escence. The best experimental conditions were: 550 mM
bicarbonate buffer (pH 9.2), 1.3 mg mLϪ1 of BSA, 1.3 mM
Fig. 2 Fluorescent signal observed upon chemical oxidation of
alcohol 3. Fluorescence signal (arbitrary units) observed upon chemical
oxidation of alcohol 3 (0.05 mM), using TEMPO (CH2Cl2)–NaClO in
bicarbonate buffers (pH 9.2) of different ionic strength: ᭹ 550 mM, ᭜
165 mM, ᭿ 55 mM. Enzymatic oxidation of alcohol intermediate
3 (0.05 mM), ∆ in a 100 mM glycine–NaOH buffer, pH 9.6, HLADH–
NADϩ.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3502

View Article Online
Fig. 3 On line observation of fluorescence appearance in a 96 well plate. Fluorescence signal (arbitrary units) observed: Lines A–D: upon enzymatic
oxidation. Lines E–F: upon chemical oxidation. Columns 1, 3, 5, 7: blank experiments. Columns 2, 4, 6, 8: test experiments.
obviously needed once positive answers are obtained in any
such assay.
development of high throughput screening (HTS) strategies
aimed at detecting novel Baeyer–Villigerase activities in large
collections of microorganisms.
Conclusion
Experimental
We have described in this work the first fluorimetric assay
allowing detection of Baeyer–Villigerase activity displayed by
whole bacterial cells. This two-step assay is based on the use of
the test ketone 1 which we have selected and synthesised. In this
context, we have set up a new methodology which allows the
replacement of the previously described enzymatic method-
ology, based on the use of an alcohol dehydrogenase (HLADH)
catalysed oxidation, by a chemical oxidation methodology,
thus leading to an about three times faster and more intense
response. We have shown that, when applied to bacterial strains
known to be able to perform the Baeyer–Villiger oxidation of
the test ketone 1, i.e.: Acinetobacter sp. TD63, Pseudomonas sp.
NCIMB 9872 and Arthrobacter sp. M5, it was possible to detect
this enzymatic activity. Concentrations as low as 50 µM of test
alcohol 3 proved to be sufficient for qualitative detection. The
use of cheap chemical reactant and co-oxidant (aqueous
NaOCl solution) is an additional advantage of this method.
This (first) fluorogenic assay clearly could be of value in the
General procedures and materials
Gas chromatography analyses were performed with a Shimadzu
GC-14A chromatograph equipped with an Optima 5 fused silica
capillary column (diameter 0.5 µm, length 30 m) (Macherey-
Nagel GmbH & Co., Duren, Germany) with helium as
the carrier gas. HPLC analyses were achieved in a Shimadzu
chromatograph equipped with a UV-detector (λ = 310 nm) and
a Hypersil DBS C18 (250 × 4.6 mm, 5 µ) column using
acetonitrile–H2O (50%, 0.5 mL minϪ1) as eluent. Fluorescence
measurements were performed in individual HTS multiwell
storage plates, each containing 96 U bottomed wells made of
polypropylene (Costar and Falcon) with a Cytofluor Series
4000 (Perseptive Biosystems, filters λex 360 20 nm, λem 460
20 nm) and a Fluoroskan Ascent FL (Labsystems type 374,
filters λex 360
20 nm, λem 460
20 nm) fluorescence-plate
reader. Melting points were determined on a Büchi apparatus
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3503

View Article Online
1
and are uncorrected. H and 13C NMR spectra were recorded
(2 H, t, J 7.0, CO-CH2), 4.04 (2 H, t, J 6.0, CH2O), 6.25 (1 H, d,
J 9.5, H-3), 6.80 (1 H, d, J 8.3, H-6), 6.84 (1 H, s, H-8), 7.38
(1 H, d, J 8.3, H-5), 7.65 (1 H, d, J 9.5, H-4); δC(62.5 MHz;
CDCl3; Me4Si): 208, 162, 161, 156, 144, 129, 113, 113, 112, 101,
68, 40, 30, 23.
on a Bruker AC 250 spectrometer in CDCl3 solutions. Chemical
shifts (δ) are quoted in ppm with Me4Si as reference and
J values are in Hz. Microorganisms were cultured in a 2 L
fermentor (Setric).
The overexpressing strain E. coli TOP10 [pQR 239], the wild
type strains Acinetobacter sp. TD63 and Arthrobacter sp. M5
were generous gifts from Professor J. Ward (University College,
London), Professor P. W. Trudgill (University College of
Wales) and Professor D. B. Janssen (University of Groningen),
respectively. The wild type strains Acinetobacter calcoaceticus
NCIMB 9871 and Pseudomonas sp. NCIMB 9872 were
obtained from the National Collection of Industrial and
Marine Bacteria (UK). Stock cultures were grown on nutrient
agar at 30 ЊC and stored at 4 ЊC, except the recombinant strain,
kept at Ϫ20 ЊC in a 40% glycerol solution.
The alcohol dehydrogenases from Lactobacillus kefir (05643),
HLADH (05645) and YADH (05640), NADϩ, BSA, 5-chloro-
2-pentanone, 3-[(tetrahydro-2H-pyran-2-yl)oxy]propyl bromide
and umbelliferone were purchased from Fluka, baker’s yeast
alcohol dehydrogenase (A-7011) and 2,2,6,6-tetramethyl-1-
piperidinyloxyl free radical (TEMPO) from Sigma-Aldrich.
Solvents were distilled prior to use. Reactions with NaH were
carried out under nitrogen atmosphere.
7-{3-[(Tetrahydro-2H-pyran-2-yl)oxy]propoxy}-2H-1-benzo-
pyran-2-one (8). A mixture of 7-hydroxy-2H-1-benzopyran-2-
one (5) (2.92 g, 18 mmol), NaH (0.6 g, 25 mmol, 60%
suspension in oil) and 3-[(tetrahydro-2H-pyran-2-yl)oxy]propyl
bromide (4.0 g, 18 mmol) was heated in 10 mL anhydrous DMF
under nitrogen atmosphere for 12 h at 70 ЊC. Residual metal
hydride was then hydrolysed with a few drops of water. The
mixture was diluted with 100 mL of AcOEt and the organic
layer was washed with 3 × 10 mL of 10% NaOH aq. solution
and 3 × 10 mL of brine. The organic phase was then dried over
anhydrous Na2SO4 and the solvent was removed at reduced
pressure. Chromatography on a silica gel column yielded
4.0 g (70%) of compound 8: colourless oil; νmax (CCl4)/cmϪ1
:
2920, 1730, 1600, 1110, 1020; δH(250 MHz; CDCl3; Me4Si): 1.8–
1.5 (6 H, m, CH2-cycle), 2.10 (2 H, m, CH2), 3.60 (2 H, m,
OCH2-cycle), 3.90 (2 H, m, CH2O), 4.60 (1 H, br s, O2CH ), 6.25
(1 H, d, J 9, H-3), 6.80 (1 H, d, J 8, H-6), 6.85 (1 H, s, H-8), 7.37
(1 H, d, J 8, H-5), 7.65 (1 H, d, J 9, H-4).
Syntheses
7-(3-Hydroxypropoxy)-2H-1-benzopyran-2-one (3). 3.0 g (10
mmol) of 7-{3-[(tetrahydro-2H-pyran-2-yl)oxy]propoxy}-2H-
1-benzopyran-2-one (8) and 190 mg (1 mmol) of p-toluene-
sulfonic acid monohydrate were dissolved in 40 mL of MeOH–
H2O (95 : 5). The mixture was stirred at room temperature for
12 h. 0.8 mL of triethylamine were then added to neutralise the
mixture and the solvent was evaporated under reduced pres-
sure. The residue was diluted with 100 mL of AcOEt and the
organic layer was washed with 3 × 10 mL of 10% NaOH aq.
solution and also with 3 × 10 mL of brine. The organic phase
was dried over anhydrous Na2SO4, then filtrated and evapor-
ated. Recrystallisation afforded 1.4 g (64%) of compound 3.
Colorless solid, mp 79 ЊC (toluene); νmax (CCl4)/cmϪ1: 3400 (bb),
2920, 1730, 1670, 1600, 1340, 1100; δH(250 MHz; CDCl3;
Me4Si): 1.74 (1 H, br t, OH ), 2.09 (2 H, m, CH2), 3.87 (2 H, m,
CH2-OH), 4.18 (2 H, t, J 6, CH2-O), 6.25 (1 H, d, J 9, H-3), 6.84
(1 H, d, J 8, H-6), 6.86 (1 H, s, H-8), 7.36 (1 H, d, J 8, H-5), 7.64
(1 H, d, J 9, H-4); δC(62.5 MHz; CDCl3; Me4Si): 162, 161, 156,
144, 129, 113, 113, 112, 101, 66, 60, 32.
2-(3-Chloropropyl)-2-methyl-1,3-dioxolane (6). A mixture of
5-chloro-2-pentanone (19.6 g, 158 mmol), ethylene glycol (49.7
g, 800 mmol) and p-toluenesulfonic acid monohydrate (300 mg,
1.6 mmol) was heated in 500 mL of toluene under reflux with a
Dean–Stark trap for 24 hours. The mixture was then washed
with 10% aq. NaHCO3 solution (3 × 30 mL), followed by brine
(3 × 30 mL). The organic phase was dried over anhydrous
Na2SO4. After filtration the solvent was removed under reduced
pressure and the residual liquid was distilled to give 24 g (92%)
of compound 6: colorless liquid; δH(250 MHz; CDCl3; Me4Si):
1.35 (3 H, s, CH3), 1.85 (4 H, m, CH2-CH2), 3.60 (2 H, t, J 7,
CH2Cl), 3.95 (4 H, m, OCH2CH2O).
7-[(4,4-Ethylenedioxy)pentoxy]-2H-1-benzopyran-2-one (7).
A mixture of 7-hydroxy-2H-1-benzopyran-2-one (5) (8.4 g, 52
mmol), NaH (1.7 g, 70 mmol, 60% suspension in oil) and 2-(3-
chloropropyl)-2-methyl-1,3-dioxolane (6) (8.5 g, 52 mmol) was
heated in 50 mL anhydrous DMF for 12 h at 110 ЊC under
nitrogen atmosphere. The residual NaH was then destroyed by
a few drops of water and the mixture was evaporated at reduced
pressure. The residue was diluted with 200 mL AcOEt and was
washed with 10% NaOH solution (3 × 20 mL) and then with
brine (3 × 20 mL). The organic phase was dried over anhydrous
Na2SO4, then filtrated and evaporated. A crude solid was
obtained and the recrystallisation gives 11.3 g (75%) of
compound 7: colourless solid, mp 86 ЊC (from MeOH); (Found:
C 65.85; H 6.23. Calc. for C16H18O5: C 66.23; H 6.20%); νmax
(CCl4)/cmϪ1: 2980, 1740, 1600, 1275, 1120, 830; δH(250 MHz;
CDCl3; Me4Si): 1.38 (3 H, s, CH3), 1.90 (4 H, m, CH2-CH2),
3.98 (4 H, m, OCH2-CH2O), 4.05 (2 H, t, J 6.0, CH2O), 6.25
(1 H, d, J 9.5, H-3), 6.80 (1 H, d, J 8.3, H-6), 6.84 (1 H, s, H-8),
7.38 (1 H, d, J 8.3, H-5), 7.65 (1 H, d, J 9.5, H-4).
7-(3-Acetoxypropoxy)-2H-1-benzopyran-2-one (2). A mixture
of 7-(3-hydroxypropoxy)-2H-1-benzopyran-2-one (3, 50 mg,
0.23 mmol), acetic anhydride (0.5 mL), glacial acetic acid (2
mL) and p-toluenesulfonic acid monohydrate (6 mg) was stirred
for 16 h at 60 ЊC. The reaction mixture was then evaporated to
dryness and the crude product was chromatographed on a silica
gel column to obtain 40 mg (67%) of compound 2. Colourless
oil; δH(250 MHz; CDCl3; Me4Si): 2.10 (3 H, s, COCH3), 2.18
(2 H, q, J 6, CH2), 4.10 (2 H, t, J 6, CH2-O), 4.25 (2 H, t, J 6,
CH2-OCO), 6.25 (1 H, d, J 9, H-3), 6.80 (1 H, d, J 8, H-6), 6.85
(1 H, s, H-8), 7.37 (1 H, d, J 8, H-5), 7.64 (1 H, d, J 9, H-4);
δC(62.5 MHz; CDCl3; Me4Si): 170, 162, 161, 156, 144, 129, 113,
113, 112, 101, 66, 60, 28, 21.
Microbial culture conditions
7-(4-Oxopentoxy)-2H-1-benzopyran-2-one (1). 6.6
g (23
mmol) of 7-[(4,4-ethylenedioxy)pentoxy]-2H-1-benzopyran-2-
one (7) and 28 g of acid Amberlite IR 120 were stirred in 200
mL of acetone at room temperature for 12 h. The mixture was
then filtered and the liquid phase was evaporated at reduced
pressure. Recrystallisation of the residue yields 4.6 g (81%) of
compound 1: white solid, mp 79 ЊC (toluene); (Found: C 68.35;
H 5.82. Calc. for C14H14O4: C 68.31; H 5.69%); νmax (CCl4)/
cmϪ1: 2980, 1740, 1710, 1600, 1275, 1120, 830; δH(250 MHz;
CDCl3; Me4Si): 2.10 (2 H, m, CH2), 2.20 (3 H, s, CH3), 2.69
Recombinant Escherichia coli. E. coli TOP10 [pQR 239]
contains a pBAD plasmid into which the CHMO gene from
Acinetobacter calcoaceticus NCIMB 9871 has been cloned. The
expression of the CHMO gene was induced by -(ϩ)-arabinose.
The recombinant bacterium was cultivated essentially as
described previously.14,20 The medium (pH 6.8) contained 10 g
L
L
Ϫ1 each of glycerol, yeast extract, peptone, NaCl, and 100 mg
Ϫ1 of ampicillin. A 2 L fermentor was filled with 1 L of culture
medium and was inoculated with 50 mL of a 12 h pre-culture at
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3504

View Article Online
37 ЊC. The fermentor was stirred at 500 rpm with an air flow of
16 L hϪ1. -(ϩ)-arabinose (0.05% w/v) was added to induce the
Calibration experiments. In order to ascertain the accuracy
of our calculation of umbelliferone 5 concentration in the
selected experimental conditions, we verified that the fluor-
escence intensity could be converted into concentration values
according to a calibration curve.
enzymatic activity into the exponential growth phase (A600
:
2–3). After a further 4 h growth period the cells were harvested
by centrifugation at 4 ЊC and used for biotransformation.
Standard solutions, 0.04, 0.1, 0.2, 0.4, 1, 2 and 5 mM, of
umbelliferone (5) were prepared from the 10 mM stock solution
of 5 in 50% aq. MeCN. The fluorescent responses were analysed
in both the test media. Thus it appeared in both media that, up
to about 20 µM, the relation was linear. However, higher
concentration values led to a non-proportional increase of
fluorescence signals due to inherent quenching. Thus,
determination of the accurate concentration of 5 became
increasingly imprecise at concentrations between 20 and 100
µM.
Wild type strains: A. calcoaceticus NCIMB 9871, Acineto-
bacter sp. TD63, Arthrobacter sp. M5 and Pseudomonas sp.
NCIMB 9872. The culture media were a 1 L typical minimal
mineral medium (4 g Na2HPO4, 2 g KH2PO4, 3 g (NH4)2SO4,
0.1 g CaCl2, 0.01 g FeSO4, 1 mL trace elements solution and
0.2 g yeast extract, pH 7) supplemented with 1.5 g of 1,2-
cyclohexanediol (Acinetobacter strains) or 1.5 g of cyclo-
pentanol (Pseudomonas sp.) or 0.5 g of acetophenone (Arthro-
bacter sp.) as the only carbon source. After inoculation with a
6–24 h preculture, cells were grown for 15–20 h at 30 ЊC.
Typical procedure for the enzymatic assay (100 ꢀL). 90 µL of
NADϩ–BSA solution and 5 µL of HLADH solution (0.1 mg
mLϪ1) were added in this order into each well containing: 5 µL
of 1 mM alcohol 3 standard solution (control experiments) or 5
µL of an aliquot of the biotransformation media (test experi-
ments). In blank experiments each well also contained 5 µL of
an aliquot of the biotransformation media, but no addition of
the HLADH solution was performed and 5 µL of 100 mM
glycine–NaOH buffer solution were added in its place. The 96
well plate was incubated (25 ЊC, 1200 rpm), and fluorescence
output was measured. After 90 min, the maximum fluorescence
was reached.
Biotransformation conditions
Biotransformations using wild type bacteria were carried out
directly in the culture medium, after a growth period. In the
case of E. coli TOP10 [pQR 239], harvested cells were cen-
trifuged and re-suspended in the same volume of phosphate
buffer (50 mM, pH 7.5). An addition of glycerol (0.05%) was
also necessary for biotransformations.14
Analytical experiments. The biotransformations were
routinely performed on a 50 mL scale in 250 mL Erlenmeyer
flasks in an agitated water bath maintained at 30 ЊC. Ketone 1
was pre-dissolved in the minimum volume of ethanol and was
added to the biotransformation medium to obtain a final
concentration 1 mM. 1 mL aliquots of the biotransformation
media were taken at the different biocatalysis times and cells
were removed by centrifugation. The biooxidations were
monitored by HPLC analysis of samplings. After addition of
methanol, the supernatant was filtrated over a 0.45 µm Teflon
filter. Quantitative evaluation of ketone and alcohol concen-
tration was determined using a calibration curve.
Typical procedure for the chemical assay (100 ꢀL). 45 µL of
1.3 mM TEMPO solution, then 45 µL of 550 mM bicarbonate
buffer (containing BSA, 1.3 mg mLϪ1) and 5 µL of 3% aq.
NaOCl solution were added in this order into each individual
well containing: 5 µL of 1 mM alcohol 3 standard solution
(control experiments) or 5 µL of an aliquot of the biotransfor-
mation media (test experiments). In blank experiments the wells
also contained 5 µL of an aliquot of the biotransformation
medium, but the oxidant and co-oxidant solutions were not
added, so 45 µL of CH2Cl2 and 5 µL of distilled water were
added in replacement. The multi-well plate was agitated (1200
rpm, 25 ЊC) and fluorescence was measured. After 30 min, the
chemical reaction was finished and the maximum fluorescent
response was reached.
Preparative scale experiments. Preparative scale biotrans-
formations were achieved in a 2 L fermentor using experimental
conditions identical to those used for the culture. Each 0.5 g of
ketone 1, dissolved in a minimum volume of EtOH, was added
into the whole cell media of Acinetobacter sp. TD63 and A.
calcoaceticus NCIMB 9871, respectively. After 16 h, the
biotransformation with Acinetobacter sp. TD63 cells was
stopped and continually extracted using AcOEt. The extracts
were dried over anhydrous Na2SO4 and the solvent removed
under reduced pressure to yield 0.4 g (90%) of alcohol 3. On the
other hand, after 24 h no oxidation of test ketone 1 was
observed for the biotransformation by A. calcoaceticus NCIMB
9871.
Acknowledgements
This work was partly funded by the Project BIO4-CT98-0267
and QLK3-CT2001-00403 as part of the European Community
Framework Programmes. We would like to express our thanks
to Professor J. Ward, Professor P. W. Trudgill and Professor
D. B. Janssen for having provided us with the bacterial strains.
One of the authors (M.C.G.) would like to express her gratitude
to the Foundation Ramón Areces for providing a postdoctoral
fellowship to carry out the research work.
Fluorescence measurements
Standard solutions, 1 mM, of test ketone 1, alcohol 3 and
umbelliferone 5 were diluted from the 10 mM stock solutions
in 50% aq. MeCN. Enzymes were also diluted from the com-
mercial solutions in the different buffers to prepare different
concentration solutions. 100 mM borate buffer, pH 8.8; 100
mM glycine–NaOH buffers, pH 9.6 and pH 7.7; 55 mM, 165
mM and 550 mM bicarbonate buffers, pH 9.2; and 3% NaClO
aq. solution were prepared. BSA and NADϩ were diluted into
the borate and glycine–NaOH buffers to a final concentration
2 mg mLϪ1 and 0.7 mg mLϪ1, respectively, in the enzymatic
assays. BSA was also diluted into the bicarbonate buffers to a
final concentration 1.3 mg mLϪ1 and 1.3 mM TEMPO solution
in dichloromethane was prepared in the chemical assays. The
200 or 100 µL assays were followed in individual “U” bottomed
wells of polypropylene 96 well plates, with a fluorescence plate-
reader (filters λex 360 20 nm, λem 460 20 nm) at 1200 rpm.
References
1 See, for instance: D. C. Demirjian, P. C. Shah, F. Morís-Varas, in
Biocatalysis: from Discovery to Application, Ed. W. C. Fessner,
Springer, Berlin, 1999, vol. 200, pp. 1–29.
2 M. Olsen, B. Iverson and G. Georgiou, Curr. Opin. Biotechnol.,
2000, 11, 331–337.
3 D. Wahler and J. L. Reymond, Curr. Opin. Biotechnol., 2001, 12,
535–544.
4 For a recent publication in chemical asymmetric Baeyer–Villiger
oxidation see, for example: A. Watanabe, T. Uchida, K. Ito and
T. Katsuki, Tetrahedron Lett., 2002, 43, 4481–4485 and references
therein.
5 V. Alphand, R. Furstoss, in Enzyme Catalysis in Organic Synthesis:
a Comprehensive Handbook, Ed. K. Danz, H. Waldmann, VCH,
Weinheim, 1995, vol. 2, pp. 745–772.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3505

View Article Online
6 A. Willetts, Trends Biotechnol., 1997, 15, 55–62.
7 M. D. Mihovilovic, B. Müller and P. Stanetty, Eur. J. Org. Chem.,
2002, 22, 3711–3730.
8 V. Alphand, G. Carrea, R. Wohlgemuth, R. Furstoss and
J. M. Woodley, Trends Biotechnol., 2003, 21, 318–323.
9 A. B. Watts, J. Beecher, C. S. Whitcher and J. A. Littlechild, Biocatal.
Biotransfor., 2002, 20, 209–214.
14 H. D. Simpson, V. Alphand and R. Furstoss, J. Mol. Catal. B:
Enzym., 2001, 16, 101–108.
15 V. Alphand and R. Furstoss, J. Org. Chem., 1992, 57, 1306–1309.
16 M. C. Gutiérrez, V. Alphand and R. Furstoss, J. Mol. Catal. B:
Enzym., 2003, 21, 231–238.
17 N. A. Donoghue and P. W. Trudgill, Eur. J. Biochem., 1975, 60, 1–7.
18 M. Griffin and P. W. Trudgill, Biochem. J., 1972, 129, 595–603.
19 J. Havel and W. Reineke, Appl. Environ. Microbiol., 1993, 59,
2706–2712.
10 G. Klein and J. L. Reymond, Bioorg. Med. Chem. Lett., 1998, 8,
1113–1116.
11 D. Wahler and J. L. Reymond, Curr. Opin. Chem. Biol., 2001, 5,
20 S. D. Doig, L. M. O’Sullivan, S. Patel, J. M. Ward and J. M. Woodley,
Enzyme Microb. Technol., 2001, 28, 265–274.
152–158.
12 R. Pérez-Carlón, N. Jourdain and J. L. Reymond, Chem. Eur. J.,
2000, 6, 4154–4162.
13 N. M. Kamerbeek, A. J. J. Olsthoorn, M. W. Fraaije and
D. B. Janssen, Appl. Environ. Microbiol., 2003, 69, 419–426.
21 S. D. Doig, H. Simpson, V. Alphand, R. Furstoss and J. M. Woodley,
Enzyme Microb. Technol., 2003, 32, 347–355.
22 A. E. J. de Nooy, A. C. Besemer and H. van Bekkum, Synthesis,
1996, 1153–1174.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 5 0 0 – 3 5 0 6
3506