Biochemical peculiarities of benzaldehyde lyase from Pseudomonas fluorescens Biovar I in the dependency on pH and cosolvent concentration

Article abstract of DOI:10.1016/j.bioorg.2009.03.001

Benzaldehyde lyase from Pseudomonas fluorescens (BAL, EC 4.1.2.38) is a versatile catalyst for stereoselective carboligations. Nevertheless, rather inconsistent data about its biochemical properties are reported in literature. In this study, the dependenc

Full text of DOI:10.1016/j.bioorg.2009.03.001

Bioorganic Chemistry 37 (2009) 84–89
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Biochemical peculiarities of benzaldehyde lyase from Pseudomonas fluorescens
Biovar I in the dependency on pH and cosolvent concentration
T. Schmidt a, M. Zavrel b, A. Spieß b, M.B. Ansorge-Schumacher a,c,
*
a Chair of Biotechnology, RWTH Aachen University, Worringerweg 1, D-52074 Aachen, Germany
b Aachener Verfahrenstechnik, RWTH Aachen University, Worringerweg 1, D-52074 Aachen, Germany
c Institute of Chemistry, Department of Enzyme Technology, Technical University of Berlin, Str. des 17. Juni 124, D-10623 Berlin, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 December 2008
Available online 17 March 2009
Benzaldehyde lyase from Pseudomonas fluorescens (BAL, EC 4.1.2.38) is a versatile catalyst for stereoselec-
tive carboligations. Nevertheless, rather inconsistent data about its biochemical properties are reported
in literature. In this study, the dependency of BAL activity on ionic strength, pH, and concentration of
DMSO was for the first time systematically investigated and interpreted. It was found that the activity
of BAL strongly depends on all three parameters, and a correlation exists between the dependency on
pH and DMSO concentration. This correlation could be explained by an interaction of DMSO with an ionic
amino acid in the catalytic site. A model-based analysis indicated that the pKa of this residue shifts to the
alkaline milieu upon addition of DMSO. Consequently, the optimum pH also shifts to alkaline values
when DMSO is present. Potentiometric experiments confirmed that the pKa can most probably be attrib-
uted to Glu50 which governs the activity increase of BAL on the acidic limb of its pH-activity profile. With
these findings, the apparently contradicting data from literature become comprehensible and optimal
reaction conditions for synthesis can easily be deduced.
Keywords:
Benzaldehyde lyase
Carboligation
DMSO
Ionic strength
pH profile
Glu50
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
(DMSO), which is often employed with BAL to enhance substrate
solubility [8,9,11–13] has been reported by Domínguez de María
Benzaldehyde lyase from Pseudomonas fluorescens Biovar I
(BAL1, EC 4.1.2.38) is a synthetically valuable biocatalyst for the gen-
eration of enantiomerically pure building blocks. It catalyses the
stereoselective carboligation of various substituted aromatic and ali-
phatic aldehydes to the corresponding hydroxy ketones [1–4] by
using one of the aldehydes as electron donor and the other as elec-
tron acceptor [5]. The reaction probably proceeds at the covalently
bound cofactor thiamine diphosphate (ThDP) [2], and was assigned
to be a reversible Bi–Uni mechanism according to Cleland [6].
Interestingly, reported data about technically relevant biochem-
ical properties of BAL considerably vary or even contradict each
other. Thus, no distinct pH optimum of enzyme activity can be
identified, since values ranging from pH 6.5 and 7.5 [7] over 8.0
[8] and 9.0 [9] to 9.5 [10] have been observed. Additionally, an
activity increase upon addition of the cosolvent dimethyl sulfoxide
et al. [10], whilst Janzen found that the presence of DMSO de-
creased BAL activity by almost 90% [14]. In contrast, both authors
determined a stabilisation of the biocatalyst by the solvent. In par-
ticular the considerably lower pH optimum reported by Nemeria
et al. [7] might be attributed to differences in the determination
of BAL activity, which in contrast to all other studies [8–10] in-
volved a coupled enzyme assay monitoring benzoin cleavage in-
stead of direct quantification of benzoin formation. However, a
distinct explanation for the findings has not been offered to date.
In this study, a systematic theoretical and experimental investi-
gation of the interactive influences of pH, ionic strength, and DMSO
concentration on the activity of BAL was performed. A model-
based mathematical analysis of dissociation constants was used
to elucidate effects of these parameters on the behaviour of the
two catalytically important amino acids Glu50 and His29, and
potentiometric experiments were applied to evaluate the impact
of the two residues on the biochemical properties of BAL. Results
are validated against available data on pH-dependency, structure,
and kinetics of the enzyme.
* Corresponding author. Address: Institute of Chemistry, Department of Enzyme
Technology, Technical University of Berlin, Str. des 17. Juni 124, D-10623 Berlin,
Germany. Fax: +49 30 314 22261.
E-mail addresses: th.schmidt@biotec.rwth-aachen.de (T. Schmidt), michael.zav-
rel@avt.rwth-aachen.de (M. Zavrel), antje.spiess@avt.rwth-aachen.de (A. Spieß),
m.ansorge@chem.tu-berlin.de (M.B. Ansorge-Schumacher).
Abbreviations used: BAL, benzaldehyde lyase; DMBA, 3,5-dimethoxy benzalde-
hyde; DMSO, dimethyl sulfoxide; ThDP, thiamin diphosphate; TMB, (R)-3,30,5,50-
tetramethoxy benzoin.
2. Materials and methods
1
All chemicals were purchased from Sigma–Aldrich (Deisenho-
fen, Germany) or Fluka (Neu-Ulm, Germany) and used as obtained.
0045-2068/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2009.03.001

T. Schmidt et al. / Bioorganic Chemistry 37 (2009) 84–89
85
Nomenclature
Symbols
k
wavelength (nm)
a
c
CI
d
model parameter (–)
pH
pHopt
pKa
t
pH value (–)
concentration (mM)
pH optimum (–)
dissociation constant (–)
time (s)
model parameter (mMꢀ0.5
optical pathlength (cm)
)
DMBA
DMSO
e
Ext
concentration of DMBA (mM)
concentration of DMSO (v/v)
extinction coefficient (mMꢀ1 cmꢀ1
extinction (–)
TMB
kcat
kcat,opt
z
concentration of TMB (mM)
turnover number (sꢀ1
)
)
turnover number at pH optimum (sꢀ1
charge (–)
)
I
ionic strength (mM)
i
chemical species (–)
Benzaldehyde lyase from P. fluorescens Biovar I (BAL) was prepared
from Escherichia coli SG13009prep4 [pBAL-his6] according to Iding
et al. [15], lyophilised and stored at ꢀ20°C.
DMSO. The ionic strength was adjusted to 100 mM with KCl. The
solution was titrated from both the acidic and alkaline milieu, i.e.
the initial pH value was set to 2.5 and 12 with 5 M HCl and 5 M
KOH, respectively. The solution was titrated to a final pH value of
12 and 2.5, respectively, using 100 mM KOH and HCl, respectively.
The consumption of KOH and HCl, and the corresponding pH value
were recorded. The obtained data were smoothed and subsequently
differentiated with thesoftware Matlab2006R (TheMathWorksInc.,
Natick, USA). Dissociation constants were determined from the min-
ima of the pH value against a dpH/dV plot.
Enzyme activity was determined for (R)-3,30,5,50-tetramethoxy
benzoin (TMB) formation from 3,5-dimethoxy benzaldehyde
(DMBA) (Fig. 1) using cofactor concentrations of 0.25 mM MgCl2
and 0.25 mM ThDP. KPi at a pH of 5.0 to 11.5 and an ionic strength
of 25–500 mM was used as buffer, and 5–40% (v/v) DMSO were
added as cosolvent. The turnover number was defined as the TMB
concentration that was formed per second related to the number
of BAL active sites in the assay. The molecular concentration of active
sites was calculatedby dividingthe weightof purified enzymeby the
molecular weight of one subunit (59,800 Da) [8].
3. Results and discussion
Kinetic measurements were conducted at 25 °C and an initial
substrate concentration of 5 mM. The reaction was initiated by
Literature survey: Published data on the pH optimum (pHopt) of
BAL activity towards synthesis reveal remarkable discrepancies of
up to three units. However, upon a closer look to the experimental
conditions underlying the published data it becomes obvious that
various types of buffer, different buffer strengths and diverse DMSO
concentrations were used for the respective studies (Table 1).
To date, little is known about the effects of buffer type and
DMSO concentration on the optimal pH of enzyme catalysed reac-
tions. The use of different buffer strengths, however, can certainly
affect pH-activity via concomitant changes in the ionic strength
[18,19]. For KPi buffer this can briefly be explained by Eq. (3) (Hen-
derson–Hasselbalch) and Eq. (4).
adding 8.3 ꢁ 10ꢀ2–3.3 ꢁ 10ꢀ1
lM BAL and recorded over a time
period of ca. 120 s at 325 nm on an UV–Vis spectrometer (Varian,
Cary 50, Palo Alto, USA). Maximum conversion within this time
was 2%. Initial rates were calculated from the slope of the extinc-
tion (Ext) against time according to Lambert–Beer’s law (Eq. (1)).
Extinction coefficients of DMBA (
e
DMBA) and TMB (
e
TMB) were
2.434 mMꢀ1 cmꢀ1 and 2.260 mMꢀ1 cmꢀ1, respectively.
Ext ¼ ðeDMBA ꢂ DMBA þ eTMB ꢂ TMBÞ ꢂ d
ð1Þ
The reaction was balanced according to Eq. (2).
DMBA0 ¼ DMBAðtÞ þ 2 ꢂ TMBðtÞ
cðHPO24ꢀÞ
ð2Þ
pH ¼ pK0a þ
ð3Þ
ð4Þ
cðH2POꢀÞ
In these equations, d denotes the optical pathlength. DMBA0 is
the initial DMBA concentration and DMBA(t) and TMB(t) are DMBA
and TMB concentration measured at time point t.
Determination of pH was accomplished with a conventional pH
electrode (Digital-pH-Meter 646, Knick, Berlin, Germany). This was
calibrated with 50 mM potassiumhydrogenphthalate and 50 mM
KPi buffer according to Taylor [16]. The measured values were cor-
rected according to Mukerjee and Ostrow [17] when the DMSO
concentration in the aqueous solution exceeded 30%.
4
X
I ¼ 0:5 ꢂ
ci ꢂ z2i ¼ 3 ꢂ cðK2HPO4Þ þ cðKH2PO4Þ
i
Table 1
Literature data on pH optima of BAL activity towards synthesis.
pHopt
Buffer
Type DMSO (%, v/v)
Reference
9.5
9.0
6.5–7.5
8.0
50 mM KPi
10 mM KPi
50 mM acetic acid/MES/Tris
50 mM KPi
30
30
17
0
[10]
[9]
[7]
Titration curves were determined at 25 °C using an auto-titrator
(Titro-line, Fa. Schott, Mainz, Germany). 5 mM
L-glutamate or L-his-
[8]
tidine was dissolved in deionised water with or without 30% (v/v)
Fig. 1. BAL-catalysed carboligation of DMBA to TMB.

86
T. Schmidt et al. / Bioorganic Chemistry 37 (2009) 84–89
In these equations, pK0a denotes the apparent dissociation constant
of the buffer, c the molar concentration of dissociated buffer species,
zi their charge in solution, and I the ionic strength. Conversely, these
relations also imply that changes in the ratio of differently charged
dissociated buffer species cause a drift of ionic strength. As a conse-
quence, the ionic strength also changes when the pH is varied at
constant buffer strength like in the studies cited above [7–10]. Nev-
ertheless, no data considering the concrete influences of ionic
strength on BAL activity is available from published literature.
In the following, the dependency of synthetic BAL activity on io-
nic strength, and additionally the influence of DMSO concentration
on pH-dependency were thoroughly studied. As a model reaction,
the symmetrical carboligation of 3,5-dimethoxy benzaldehyde
(DMBA) yielding 3,30,5,50-tetramethoxy benzoin (TMB) was used
(Fig. 1) instead of the condensation of unmodified benzaldehyde
yielding benzoin reported in former studies [8–10]. This was done
for a better solubility of substrates and products in aqueous media
and enabled a very precise determination of concentrations by
fluorescence spectroscopy. An impact on the comparability of re-
sults with reported data [8–10] must not be expected since the
mechanism of benzoin condensation will not be altered by a sub-
strate change.
nic strengths in the studies with Eqs. (3), (4), and (6) (Debye–Hüc-
kel equation), and subsequent estimation of the influence of the
obtained values on BAL activity using Eq. (5). The pKa value of
phosphate buffer in a 30% DMSO solution was set to 7.7 according
to Taylor [16], cofactor concentrations were considered negligible
for ionic strengths. Results of this procedure are summarised in Ta-
ble 2. The so-determined activity fluctuations can be regarded as
the impact of ionic strengths on published data formerly attributed
to an influence of pH. In the two investigated buffers (KPi at differ-
ent concentrations, see Table 2), this impact ranges up to 15% and
30%, respectively.
!
pffiffi
0:5114 ꢂ
I
pK0a ¼ pKa þ ð2 ꢂ za ꢀ 1Þ ꢂ
ꢀ 0:1 ꢂ I
ð6Þ
pffiffi
1 þ
I
In Eq. (6), pKa denotes the thermodynamic dissociation constant of
KPi and za is the charge on the conjugate acid species.
Surprisingly, despite this clear influence of the ionic strength on
the measured activity of BAL, cross effects on the pH-dependency
of this enzyme [7–10] were only neglectable. This was demon-
strated by a comparison of an activity profile of BAL measured at
varying pH, constant ionic strength (100 mM), and constant DMSO
concentration (30% v/v) to activity profiles derived from tagging
this curve with the probable cross effects of the ionic strength in
the studies of Domínguez de María et al. [10] and Stillger et al.
[9] (Fig. 3).
Eq. (7) was fitted to each profile in order to determine the
respective pH optima as proposed by others [7,22–25]. The result-
ing values were 9.41 in 50 mM KPi [cf. 10], 9.44 in 10 mM KPi [cf.
9], and 9.46 ( 0.07) for the curve measured in this study. The devi-
ations were insignificant and not nearly in the range of the re-
ported differences in pH optima.
Influence of ionic strength: Investigations of synthetic BAL activ-
ity in KPi buffer (pH 8.5, 30% v/v DMSO) revealed that the ionic
strength of the buffer system has a very strong reciprocal effect
on the activity of BAL, i.e. in a range between 25 mM and
500 mM, highest initial rates were observed at lowest ionic
strength, and activity decreased with increasing ionic strength
(Fig. 2). This was calculated with Eq. (5).
Iꢀpffiffiffiffiffiffi
p
ffi
kcat
kopt;I
Iopt
Þ
¼ 10C ꢂð
ð5Þ
I
Subsequently, a recalculation of the pH optimum from the
uncorrected raw data of Domínguez de María et al. [10] revealed
that the discrepancy between the values reported by this author,
Stillger et al. [9] and our own findings (Fig. 3) is most probably
due to differences in the evaluation of experimental results. The
pH optimum found by application of Eq. (7) according to Buchholz
et al. [26] and Marangoni [24] to the data of Domínguez de María
et al. [10] was 9.27 0.19. This is statistically almost equivalent to
the value of 9.0 reported by Stillger et al. [9] and complies well
with the pH optimum of 9.46 ( 0.07) determined in this study
(see previous paragraph). A similar processing of the data reported
by Stillger et al. [9] was impossible, because the number of avail-
able data points was too small. An explanation for the residual dif-
ferences in the pH optima of Stillger et al. [9] and Domínguez de
María et al. [10] can probably be found in the different determina-
tion of pH in both studies. While Stillger et al. [9] determined pH
before addition of DMSO (30% v/v), Domínguez de María et al.
[10] used the mixture of buffer/DMSO DMSO (30% v/v) for pH
In this equation, CI is a model parameter and kopt,I the turnover
number at the optimal ionic strength Iopt (25 mM), as formerly pro-
posed [19–21]. Eq. (5) was fitted to the experimental data in order
to estimate the slope of the linearised curve (CI). The distinct value
of CI was ꢀ0.0296 0.0016 mMꢀ0.5
.
Based on this finding it can be assumed that the activity of BAL
described in literature is cross-influenced by a drift of ionic
strength connected with the use of the respective buffer systems
at different pH values. This was confirmed by calculation of the io-
Table 2
Cross influences of changes in the ionic strength at different pH on activity
determinations in 50 mM KPi [cf. 10] and 10 mM KPi [cf. 9]. Ionic strength was
determined according to Eqs. (3), (4).
pH
50 mM KPi
10 mM KPi
Cross influence
Cross influence
I (mM)
(%)
I (mM)
(%)
5.0
6.0
7.0
8.0
9.0
10.0
11.0
50.3
52.6
71.2
122.9
146.4
149.6
150.0
0
1
9
24
29
30
30
10.1
10.5
14.2
24.6
29.3
29.9
30.0
0
1
4
11
14
15
15
Fig. 2. Influence of ionic strength on the synthetic activity of BAL in KPi (pH8.5, 30%
v/v DMSO) calculated according to Eq. (5). Ionic strength was determined according
to Eqs. (3) and (4).

T. Schmidt et al. / Bioorganic Chemistry 37 (2009) 84–89
87
et al. [8] performed their study without DMSO. For solubility
reasons, it was not possible to do this in our study. An extrapo-
lation of the observed relationship between pH optimum and
DMSO concentration to 0% DMSO gave 8.60 0.05. This is a sta-
tistically relevant difference to the pH optimum calculated from
the raw data of Janzen et al. [8] (pHopt: 8.24 0.24), but defi-
nitely more in the range of this value than the value of Domín-
guez de María et al. [10] and Stillger et al. [9]. Considering the
usual variations of kinetic measurements performed in different
labs, it can be concluded that the discrepancies in pH-depen-
dency of BAL reported by these authors can now be attributed
to differences in the cosolvent concentration during analysis. In
contrast, a final explanation for the extraordinary low pH opti-
mum observed by Nemeria et al. [7] cannot be offered here. It
may result from using a coupled activity assay determining ben-
zoin cleavage instead of formation as stated in the introductory
part of this manuscript, indicate cross effects related to the com-
plex buffer system in the experiments, or simply result from an
inaccurate assignment of pH in aqueous solutions of DMSO as
described in other studies [16,17]. These found that the mea-
sured or apparent pH of an aqueous solution can be increased
upon addition of DMSO. This increase may either be attributed
to a real change of proton activity or merely result from a dis-
turbance of pH determination by the organic solvent. In phos-
phate buffers the pH increases with increasing amounts of
DMSO because of an increasing dissociation constant. The appar-
ent pH resembles the real pH up to about 30% (v/v) DMSO in
buffer. Above this threshold real pH and apparent pH increas-
ingly differ. A calibration procedure to rule out mistakes in pH
determination in the presence of DMSO is well known in litera-
ture [16,17], but it remains unclear wether Nemeria et al. [7]
conformed to it in their study.
Fig. 3. Upper graph: synthetic activity of BAL at various pH values, constant ionic
strength (100 mM) and 30% (v/v) DMSO (d). In silico experiments with constant
buffer strength of 50 mM (s) [cf. 10]. Fit of Eq. (7) to experiments to determine
▬,
▬
pHopt
( ). Lower graph: development of ionic strength related to the upper graph.
Relation of DMSO-dependency and amino acid residues: A calcula-
tion of the two pKa values, pKa1 and pKa2, describing the acidic
(pKa1) and alkaline (pKa2) limb of the DMSO-dependent pH-activity
profiles (Fig. 4) shows that the observed DMSO-dependent change
of pHopt is accompanied by an increase in pKa1 whereas pKa2 re-
mains almost constant (Table 3). Considering that pHopt comprises
the mean value of pKa1 and pKa2 [24,26], this indicates that the
dependency of pHopt is almost exclusively governed by pKa1. Such
a dependency of pKa values on DMSO can frequently be found for
acid–base moieties [27], and has recently been reported for amino
acid side-chains in proteins [28]. It can be attributed to the forma-
tion of strong hydrogen bonds between the alkaline oxygen of
DMSO and acidic protons, whereupon the physico-chemical char-
acteristics of surrounding water may be severely affected [29].
With this finding, and considering that according to Buchholz
et al. and Marangoni pKa1 and pKa2 may be assigned to ionisable
side-chains in the active site of an enzyme [24,26], an investigation
on the relation between DMSO-dependency of BAL activity and
structural features of the catalyst was performed. Examination of
the active site of BAL for appropriate acid–base side-chains exclu-
sively revealed two residues, namely glutamate at position 50
(Glu50) and histidine at position 29 (His29), as candidates for pos-
sible interactions with DMSO. It has been suggested for His29 that
the protonation state of this amino acid residue might play an
important role in catalysis [30]. A similar role can be assumed
for Glu50, as this complies with a highly conserved residue in all
ThDP-dependent enzymes [31] which is responsible for the essen-
tial binding of the cofactor [32]. As this binding requires the depro-
tonation of Glu50 [7] both residues can be considered to govern the
changes in BAL activity in the acidic limb of the pH-activity profile.
The ability of each residue to interact with DMSO was investi-
determination. This must have caused a shift of the apparent pH as
described in detail in the following section.
In summary, the findings clearly demonstrate that ionic
strength, though different in the reported experiments, cannot be
the reason for the peculiar pH-dependency observed for BAL cata-
lysed reactions. This is in accordance with the fact that even upon
recalculation of the pH optimum from the raw data of Janzen et al.
[8] (Table 1) by application of Eq. (7) (pHopt: 8.24 0.24), a clear
discrepancy to the value of Domínguez de María et al. [10] can
be observed (pHopt: 9.27 0.19), although the same buffer was ap-
plied (50 mM KPi).
kcat
kopt;pH
a
1ꢀpH
¼
ð7Þ
1 þ 10pKa
þ 10pHꢀpKa
2
In Eq. (7), kcat and kopt,pH denote the turnover numbers at investi-
gated and optimal pH, respectively, a is a model parameter without
physical meaning, and pKa1 and pKa2 mediate the acidic and alkaline
limb in the pH-activity profiles.
Influence of DMSO: Records of pH-activity profiles at different
DMSO concentrations and constant ionic strength (100 mM) re-
vealed a strong influence of the cosolvent on the activity of
BAL, and indicated a cross effect on the optimal pH with a shift
of profiles to the alkaline milieu (Fig. 4). Subsequent initial rate
analysis at pHopt gave the highest activity between 20% and 30%
DMSO (v/v) with a peak at 25%, which is in accordance with the
activating effect of DMSO observed by Domínguez de María et al.
[10].
Calculation of pHopt by application of Eq. (7) showed a clear
increase of this parameter with increasing concentrations of
DMSO (Table 3). The calculated values are in good agreement
with the recalculated or published pH optima of Domínguez de
María et al. [10] and Stillger et al. [9], respectively, who both
used a DMSO concentration of 30% (v/v) for their studies. Janzen
gated by potentiometric titrations of isolated
and -glutamate ( -glu). In the absence of DMSO, the pKa values
of the amino acids were 4.31 for -Glu and 6.14 for -His (Fig. 5)
L-histidine (L-his)
L
L
L
L

88
T. Schmidt et al. / Bioorganic Chemistry 37 (2009) 84–89
Fig. 4. Upper graphs: pH-activity profiles of BAL at different DMSO concentrations. Lower graph: optimal pH value for carboligation activity and the respective activity as a
function of the DMSO content.
Table 3
profiles of BAL (Table 3). In contrast, the pKa of L-Glu increased
Influence of DMSO on pHopt, pKa1 and pKa2. Values were calculated using Eq. (7).
by 0.52 units under the same conditions, resembling the increase
of pKa1 in the presence of 30% (v/v) DMSO. This suggests that in
BAL, Glu50 is affected by DMSO and governs the acidic limb of
the pH-activity profile, whereas His29 is without importance. This
finding could be challenged with the established knowledge that
free amino acids frequently exhibit other pKa values and behave
differently from amino acids in a protein structure [24]. Likewise
the catalytic cycle of BAL contains several charged reaction inter-
mediates and DMSO may interact with them or charged residues
on the enzyme surface which indirectly influence catalysis. How-
ever, it is in line with the sound proposition of Jordan and cowork-
ers that the activity increase of BAL accompanying the
deprotonation of Glu50 can thermodynamically be described by
the pKa of this amino acid residue [7,33], and can therefore be
strongly supported.
(%)
pHopt
pKa1
pKa2
5
8.70 0,05
8.98 0,05
9.04 0,04
9.19 0,05
9.22 0,06
9.46 0,07
9.65 0,05
9.79 0,12
6.81 0,04
7.08 0,04
7.04 0,03
7.44 0,04
7.40 0,04
7.99 0,04
8.14 0,04
8.85 0,10
10.58 0.06
10.88 0.05
11.04 0.04
10.93 0.06
11.04 0.07
10.92 0.09
11.16 0.06
10.73 0.13
10
15
20
25
30
35
40
and thus were in good agreement with the published values of 4.3
and 6.0, respectively [23]. In the presence of 30% (v/v) DMSO, the
pKa of
L-His slightly decreased, which does not correspond to the
behaviour of pKa1 calculated from DMSO-dependent pH-activity

T. Schmidt et al. / Bioorganic Chemistry 37 (2009) 84–89
89
Fig. 5. Progression of the potentiometric titration of L-Glu and L-His without (d) and with (s) 30% (v/v) DMSO.
[6] M. Zavrel, T. Schmidt, C. Michalik, M. Ansorge-Schumacher, W. Marquardt, J.
Büchs, A.C. Spiess, Biotechnol. Bioeng. 101 (2008) 27–38.
[7] N. Nemeria, L. Korotchkina, M.J. McLeish, G.L. Kenyon, M.S. Patel, F. Jordan,
Biochemistry 46 (2007) 10739–10744.
[8] E. Janzen, M. Muller, D. Kolter-Jung, M.M. Kneen, M.J. McLeish, M. Pohl, Bioorg.
Chem. 34 (2006) 345–361.
[9] T. Stillger, M. Pohl, C. Wandrey, A. Liese, Org. Process Res. Dev. 10 (2006) 1172–
1177.
[10] P. Domínguez de María, T. Stillger, M. Pohl, S. Wallert, K. Drauz, H. Grögerm, H.
Trauthwein, A. Liese, J. Mol. Catal. B: Enzym. 38 (2006) 43–47.
[11] A.S. Demir, Ö. Sesenoglu, P. Dunkelmann, M. Muller, Org. Lett. 5 (2003) 2047–
2050.
4. Conclusions
Reported biochemical peculiarities of benzaldehyde lyase from
Pseudomonas fluorescens Biovar I. (BAL, EC 4.1.2.38) regarding pH
optimum and dependence on DMSO can be explained by a direct
interaction of DMSO with the amino acid side chain Glu50 in the
catalytic site of the biocatalyst. This finding is important since it al-
lows a better understanding of the catalytic mechanism of this en-
zyme, and consequently
a deduction of optimal reaction
[12] A.S. Demir, Ö. Sesenoglu, E. Eren, B. Hosrik, M. Pohl, E. Janzen, D. Kolter, R.
Feldmann, P. Dünkelmann, M. Müller, Adv. Synth. Catal. 344 (2002)
96–103.
[13] T. Hischer, D. Gocke, M. Fernández, P. Hoyos, A.R. Alcántara, J.V. Sinisterra,
W. Hartmeier, M.B. Ansorge-Schumacher, Tetrahedron 61 (2005)
7378–7383.
conditions. Moreover, as Glu50 is an amino acid residue which is
highly conserved in the whole class of ThDP-dependent enzymes,
e.g. pyruvate decarboxylase [31], it can be expected that the ob-
served relation between enzyme activity, pH-dependency, and
DMSO concentration will also apply to these. Thus, the findings
provide a novel argument against the use of DMSO as cosolvent
in the application of a whole subclass of synthetically valuable bio-
catalysts, additional to the known complication of downstream
processing caused by DMSO. It will have to be further investigated
whether similar effects occur with alternative water miscible sol-
vents required to enhance water solubility of substrates and prod-
ucts in the enzyme catalysed reactions.
[14] E. Janzen, Biochemische Charakterisierung und die Untersuchung von
Struktur-Funktionsbeziehungen,
2002.
Heinrich-Heine-Universität,
Düsseldorf,
[15] H. Iding, P. Siegert, K. Mesch, M. Pohl, Biochim. Biophys. Acta, Protein Struct.
Mol. Enzymol. 1385 (1998) 307–322.
[16] M.J. Taylor, J. Chem. Eng. Data 24 (1979) 230–233.
[17] P. Mukerjee, J.D. Ostrow, Tetrahedron Lett. 39 (1998) 423–426.
[18] K. Laidler, P. Bunting, The Chemical Kinetics of Enzyme Action, Clarendon
Press, Oxford, 1973.
[19] B.E. Dale, D.H. White, Enzyme Microb. Technol. 5 (1983) 227–229.
[20] K. Hiyama, S. Okada, J. Biochem. 82 (1977) 429–436.
[21] G. Kossekova, B. Atanasov, R. Bolli, A. Azzi, Biochem. J. 262 (1989) 591–596.
[22] Z. Yang, D. Zacherl, A.J. Russell, J. Am. Chem. Soc. 115 (1993) 12251–12257.
[23] P.A. Frey, A.D. Hegeman, Enzymatic Reaction Mechanisms, Oxford University
Press, New York, 2007.
[24] A. Marangoni, Enzyme Kinetics – A Modern Approach, John Wiley & Sons, Inc,
Hoboken, New Jersey, 2003.
[25] K. Taylor, Enzyme Kinetics and Mechanisms, Kluwer Academic Publishers,
New York, 2002.
[26] K. Buchholz, V. Kasche, U. Bornscheuer, Biocatalysts and Enzyme Technology,
VCH-Wiley, Weinheim, 2005.
Acknowledgments
We gratefully acknowledge the financial support by the Deut-
sche Forschungsgemeinschaft DFG via the collaborative research
center 540 (Model-based experimental analysis of kinetic phenom-
ena in fluid multi-phase reactive systems) and GRK 1166 (Biocatal-
ysis using non-conventional media – BioNoCo).
[27] F. Bordwell, Acc. Chem. Res. 21 (1988) 456–463.
[28] J. Kruusma, A. Rhodes, R. Bhatia, J. Williams, A. Benham, R. Kataky, J. Solution
Chem. 36 (2007) 517–529.
[29] G.J. Safford, P.C. Schaffer, P.S. Leung, G.F. Doebbler, W.B. George, E.F.X. Lyden, J.
Chem. Phys. 50 (1969) 2140–2159.
References
[1] A.S. Demir, H. Hamamci, Ö. Sesenoglu, F. Aydogan, D. Capanoglu, R.
Neslihanoglu, Tetrahedron: Asymmetry 12 (2001) 1953–1956.
[2] A.S. Demir, M. Pohl, E. Janzen, M. Müller, J. Chem. Soc., Perkin Trans. 1 (2001)
633–635.
[30] T.G. Mosbacher, M. Mueller, G.E. Schulz, Febs J. 272 (2005) 6067–6076.
[31] F. Jordan, Nat. Prod. Rep. 2 (2003) 184–201.
[3] M. Pohl, B. Lingen, M. Müller, Chem. Eur. J. 8 (2002) 5289–5295.
[4] H. Iding, T. Dünnwald, L. Greiner, A. Liese, M. Müller, P. Siegert, J. Grötzinger,
A.S. Demir, M. Pohl, Chem. Eur. J. 6 (2000) 1483–1495.
[5] P. Dünkelmann, D. Kolter-Jung, A. Nitsche, A.S. Demir, P. Siegert, B. Lingen, M.
Baumann, M. Pohl, M. Muller, J. Am. Chem. Soc. 124 (2002) 12084–12085.
[32] A. Schellenberg, Biochim. Biophys. Acta – Protein Struct. Mol. Enzymol. 1385
(1998) 177–186.
[33] S. Chakraborty, N. Nemeria, A. Yep, M.J. McLeish, G.L. Kenyon, F. Jordan,
Biochemistry 47 (2008) 3800–3809.