Functional modeling of alcoholdehydrogenase with pyrazolylborate-zinc complexes

Article abstract of DOI:10.1016/S0020-1693(00)00390-X

Unlike free isopropanol, the complex Tp^Cum,MeZn-OCHMe₂ (Tp^Cum,Me=tris(3-cumenyl-5-methylpyrazolyl)borate) is oxidized by N-benzylnicotinamide chloride to acetone and Tp^Cum,MeZn-Cl in small yield. The stronger oxidant 2,3-dichloro-5,6-dicyanobenzoquinone oxidizes Tp^Cum,MeZn-phenolates with o-hydroxymethyl substituents to the corresponding zinc phenolates with o-formyl substituents. Benzaldehydes with electronegative substituents (F, NO₂, CF₃) are reduced by N-benzyldihydronicotinamide in the presence of Tp^Cum,MeZn-OH to the corresponding benzyl alcohols which are isolated as Tp^Cum,MeZn-OCH₂Ar. A kinetic analysis has shown that the reduction of the aldehydes alone is roughly two times slower than that in the presence of the zinc complex.

Full text of DOI:10.1016/S0020-1693(00)00390-X

www.elsevier.nl/locate/ica
Inorganica Chimica Acta 314 (2001) 58–62
Functional modeling of alcoholdehydrogenase with
pyrazolylborate–zinc complexes
R. Walz, H. Vahrenkamp *
Institut f u¨ r Anorganische und Analytische Chemie der Uni6ersit a¨ t Freiburg, Albertstrasse 21, D-79104 Freiburg, Germany
Received 11 August 2000; accepted 22 November 2000
Abstract
Cum,Me
Cum,Me
Unlike free isopropanol, the complex Tp
Zn–OCHMe (Tp
=tris(3-cumenyl-5-methylpyrazolyl)borate) is oxidized
Zn–Cl in small yield. The stronger oxidant 2,3-dichloro-5,6-di-
Zn–phenolates with o-hydroxymethyl substituents to the corresponding zinc phenolates
with o-formyl substituents. Benzaldehydes with electronegative substituents (F, NO , CF ) are reduced by N-benzyldihydronicoti-
2
Cum,Me
by N-benzylnicotinamide chloride to acetone and Tp
Cum,Me
cyanobenzoquinone oxidizes Tp
2
3
Cum,Me
Cum,Me
namide in the presence of Tp
Zn–OH to the corresponding benzyl alcohols which are isolated as Tp
Zn–OCH Ar. A
2
kinetic analysis has shown that the reduction of the aldehydes alone is roughly two times slower than that in the presence of the
zinc complex. © 2001 Elsevier Science B.V. All rights reserved.
Keywords: Zinc complexes; Alcoholdehydrogenase; Pyrazolylborate ligands; Enzyme modeling
1
. Introduction
correct structural models were obtained only very re-
cently [5,6], and the number of ligand systems for
structural modeling is actually quite small [7–9]. Sec-
ondly functional modeling was initially limited to the
observation that hydride reductions of aldehydic func-
tions in the vicinity of the N atoms of heterocycles
proceed faster in the presence of zinc salts [10–13],
which points to activation by N,O chelation.
There are just a few examples of functional ADH
modeling by zinc complexes. The simplest approach
has involved catalysis of the Meerwein–Ponndorf reac-
tion (i.e. hydride exchange between electron-rich alco-
hols and electron-poor aldehydes, which is best
catalyzed by aluminium alkoxides) by zinc complexes of
tridentate ligands [14,15]. Metal mediated oxidations of
electron-rich alcohols by N-alkyl–nicotinamide salts
have been observed [14,16]. The only metal complex
with the ‘correct’ ligand set (a N,S,S ligand with cobalt
instead of zinc) to be active in a ketone reduction by
N-benzyl–dihydronicotinamide was described by Kel-
logg [8].
Among the group of alcoholdehydrogenase enzymes,
horse liver alcoholdehydrogenase (HLADH [1,2]),
which converts both alcohols to aldehydes and alde-
hydes to alcohols, and yeast alcoholdehydrogenase
(
YADH [3,4]) which converts aldehydes to alcohols are
the most well-known examples. They perform their
redox reactions using zinc in the active center, which is
noteworthy considering the inability of the zinc ion to
change its oxidation state. Accordingly, the function of
the zinc ion rests in its ability to activate the substrates
(
alkoxide, aldehyde) by Zn–O coordination for hydride
+
transfer to NAD or from NADH.
The structural and functional modeling of these en-
zymes by coordination compounds of zinc has turned
out to be difficult. Firstly the NSS (histidine, cysteinate,
cysteinate) attachment of zinc to the protein asks for
modeling with amine-bis(thiolate) ligands which as a
1
rule form oligomeric zinc complexes . For this reason
*
Corresponding author. Tel.: +49-761-203 6120; fax: +49-761-
Our own contributions to ADH modeling were
firstly centered on gaining knowledge on the basic
coordination chemistry of zinc with alcohols [17] and
aldehydes [18]. Thereby, we learned how to make the
203 6001.
E-mail address: vahrenka@uni-freiburg.de (H. Vahrenkamp).
1
For references on structural modeling and model ligands for
ADH see Ref. [5].
0020-1693/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0020-1693(00)00390-X

R. Walz, H. Vahrenkamp / Inorganica Chimica Acta 314 (2001) 58–62
59
first fully characterized structural models of both the
alcohol and the aldehyde adduct of the enzyme [5]. At
the same time advances in our zinc complex chemistry
with encapsulating pyrazolylborate (Tp) ligands allowed
us to prepare TpZn complexes containing alkoxides and
aldehydic functions attached to zinc [19–21]. We thereby
had established a set of isolated and characterized
compounds, which model the oxidized and the reduced
states of the ADH enzyme–substrate complex. This
motivated us to attempt hydride transfer interconver-
+
sions between these species brought about by NAD and
NADH substitutes. This paper describes our results for
the oxidation of alcohols as well as for the reduction of
aldehydes.
After refluxing an equimolar solution of 1b and
(BNA)Cl in isopropanol for 20 h the H NMR spectrum
1
indicated that the only pyrazolylborate complex in solu-
1
tion was the chloride 1c. A weak H NMR signal at 2.08
ppm assignable to acetone had appeared. Referencing the
intensity of this signal against that of the pyrazole’s
methyl signal gave an estimated yield of acetone of
2. Results and discussion
1
0–20%. After workup for the uncharged organic com-
2.1. Oxidation of isopropanol
1
ponents BNAH and (BNA) were identified by their H
2
NMR and UV spectra [24]. Quantitative assessment by
H NMR referencing against a known amount of
+
The biological oxidant in ADH reactions is NAD . A
1
convenient substitute for it in model studies is the
benzyl-substituted nicotinamide BNA , being applied as
the chloride. Hydride transfer reduces it to BNAH; when
single-electron (i.e. radical) steps are involved it may also
be reduced to (BNA)2.
(
(
BNA)Cl gave yields of 14% of BNAH and 1% of
BNA)2.
+
These observations prove that ADH modeling with
our zinc alkoxides is possible, albeit quite inefficiently.
Yet our product yields surpass those of previously
reported ‘catalytic’ systems [14]. Part of the inefficiency
of our system may be due to the excessive water
sensitivity of 1b. Traces of water convert it to the
Cum,Me
hydroxide Tp
Zn–OH which does not catalyze the
oxidation of isopropanol. The observation that the
chloride 1c is the only TpZn complex remaining gives
no information in this respect. The free chloride in
+
solution will form 1c either from ‘naked’ TpZn re-
maining after elimination of acetone or from TpZn–OH
by hydroxide substitution [25]. The predominant forma-
tion of BNAH as the reduction product points to hydride
+
transfer as the standard mechanism of BNA
action. However, the best way of explaining the forma-
tion of the trace product (BNA)2 is single electron
transfer. Thus mechanistic questions related to these
redox processes must remain unanswered at the present
time.
We verified that neat ethanol or isopropanol are not
oxidized by (BNA)Cl. Our mode of activating the alcohol
corresponded to the proposed activation in the enzyme:
isopropanol was attached to zinc as the alkoxide in
complex 1b. Pyrazolylborate–zinc–alkoxide complexes
like 1b are extremely sensitive to moisture, being hy-
drolyzed to TpZn–OH [15,19,20,22]. For this reason
complex 1b [23] which is accessible by Parkin’s method
2.2. Oxidation of 2-hydroxymethyl phenolates
A different approach of testing the efficiency of TpZn
units in supporting alcohol oxidations was to fix the
alcoholic functions in the vicinity of zinc without attach-
ing them to the metal. We used this approach for
2-hydroxymethyl phenolates, knowing that the alcoholic
functions in complexes 2a and 2c are not coordinated to
zinc but the aldehydic functions in 2b and 2d are [21].
This gave hope that a small thermodynamic advantage
[15] from the hydride 1a and isopropanol was generated
in situ by dissolving 1a in anhydrous isopropanol imme-
diately prior to the oxidation reaction.
+
would favor the oxidations of 2a and 2c by BNA .

60
R. Walz, H. Vahrenkamp / Inorganica Chimica Acta 314 (2001) 58–62
It turned out that (BNA)Cl is not strong enough an
oxidant to react with 2a and 2c. It took the rather
strong oxidant dichlorodicyanobenzoquinone (DDQ)
[
26,27] to produce the aldehydes 2b and 2d in moderate
1
yields. H NMR was again used to identify and quan-
tify the formation of 2b, 2d and the hydroquinone
DDHQ. These reactions cannot be related to enzyme
modeling. But they demonstrate the use of the TpZn
unit as a protecting group. The free phenols of 2a and
2
c are converted by DDQ to dark intractable masses.
The reductions in dichloromethane involved the com-
plete consumption of BNAH as evidenced by the com-
plete disappearance of its UV absorption at 350 nm
within 1–10 h at room temperature. The resulting
Coordination to zinc keeps the phenolic systems out of
reach of the strong oxidant, and in principle the free
phenols of 2b and 2d can be retrieved by hydrolytic
cleavage [20].
benzylate complexes 3a–c were isolated in 29, 19 and
1
1
8% yield, respectively. The H NMR spectra showed
that considerable amounts of the unreacted aldehydes
were still present after the reactions. Thus, it must be
concluded that the reactions are not clean and lead to
side products which are unidentified as yet.
In order to quantify the catalytic effect of the zinc
complex kinetic measurements were performed for the
2
.3. Reduction of aromatic aldehydes
The reduction of aromatic aldehydes and ketones,
which are not too electron-rich, by NADH model
compounds like BNAH is a thermodynamically feasible
process which was found to be catalyzed by zinc salts
and complexes in certain cases [8–13]. We now per-
formed such reductions using BNAH as the reductant
and Tp
enhancing the reaction rate the zinc complex provided a
means of identifying the resulting benzyl alcohols as
their zinc complexes 3 which result from condensation
with the Zn–OH complex and which we had character-
ized previously [20].
reduction of pentafluorobenzaldehyde by BNAH in the
Cum,Me
presence and in the absence of Tp
Zn–OH. The
kinetic traces displayed in Fig. 1 were obtained from
chloroform solutions by recording the intensity of the
Cum,Me
Zn–OH as the ‘catalyst’. In addition to
3
50 nm UV absorption of BNAH at 50°C. Pseudo-first
order conditions were established by using concentra-
−1
Cum,Me
tions of 1.0 mmol l
for C F CHO and Tp
Zn–
6
5
−1
OH and of 0.01 mmol l
for BNAH, following the
procedure of Ohno et al. [12]. Under these conditions
the reactions had a half-life of 2–3 min.

R. Walz, H. Vahrenkamp / Inorganica Chimica Acta 314 (2001) 58–62
61
+
In order to achieve alcohol oxidations by BNA and
aldehyde reductions by BNAH the reaction conditions
had to be quite unlike those in the enzymes. Non-
aqueous conditions had to be applied, the driving force
for the oxidation of isopropanol had to be provided by
making it extremely electron-rich (the zinc–isopropox-
ide being accessible only from the zinc hydride), the
driving force for the reductions of the aldehydes had to
be provided by making them very electron-poor. Actu-
ally a modeling of ADH reactions by using aliphatic
alcohols or aldehydes in aqueous media has not been
reported as yet.
Thus, the challenge to model HLADH and YADH
remains, but also the question whether the proposed
mechanisms for the enzymatic reactions are correct.
Even advanced coordination chemistry has not over-
come the problems that monomolecular zinc alkoxides
are extremely water-sensitive and that aliphatic alde-
hydes are reluctant to interact with zinc ions. We will
invest further efforts to refine the model systems and to
improve their model reactions.
Fig. 1. Kinetic traces for C F CHO reduction. I(t) on the ordinate is
the intensity of the 350 nm UV absorption.
6
5
3
. Experimental
As observed previously [10,12] a straightforward ki-
netic analysis is not possible for this reaction system. In
the literature a nonlinear influence of the zinc concen-
tration, possible association of BNAH and possible
reaction alternatives involving single electron transfer
have been invoked to explain this [10,12]. In our case
the fact that the reactions are not clean points in a
similar reaction. Nevertheless the rate enhancement by
the zinc complex is obvious. Judging from the initial
rates the catalyzed reaction is 1.5–2 times faster. Con-
structing a log plot based on the assumption of a clean
3.1. General procedures
All reactions were carried out in carefully degassed
and dehydratet solvents under an inert atmosphere in
the dark. Zinc containing starting materials were pre-
pared as described in the given references, as were
(
BNA)Cl and BNAH [24]. All other organic reagents
were obtained commercially. Reaction products were
1
19
identified by their H NMR and, if possible, F NMR
1
spectra. Yields were determined by H NMR spec-
troscopy in CDCl measuring either the benzyl CH
3
2
−
4
signals of the BNA or benzylalcohol derivatives, the
first-order reaction yields rate constants of 1.1×10
−
4
−1
CHO resonances of the aldehydes or the CH(pyrazole)
and 0.4×10
s
for the catalyzed and the uncata-
Cum,Me
signals of the Tp
derivatives. Intensity referencing
lyzed reactions, respectively, corresponding to a rate
enhancement by a factor of 2.7. Rate enhancements by
this small order of magnitude have been reported be-
fore for such hydrogenations of aldehydes [10–13]. It
follows that our zinc complex is as viable, yet as
mediocre, as a catalyst as the previously reported zinc
compounds.
was done either against other components in the reac-
tion mixtures or against added aliquots of the unre-
acted starting materials. UV–Vis spectra and kinetic
data were recorded with a JASCO V-570 instrument.
3.2. Reactions
3
.2.1. Oxidation of isopropanol
A total of 200 mg (0.30 mmol) of 1a were dissolved
2.4. Conclusions
in 20 ml of isopropanol and refluxed for 30 min. A total
of 70 mg (0.30 mmol) of (BNA)Cl were added and the
The work in this paper has made humble contribu-
1
tions to the understanding of zinc-mediated oxidations
mixture refluxed for 20 h after which time the H NMR
+
1
of alcohols by NAD and reductions of aldehydes by
resonances of (BNA)Cl had disappeared and a
H
NADH. Unlike in previous publications on this topic
all oxidation or reduction products were isolated or
identified by their spectra. In each case the involvement
of zinc complexes of the reactants or products was
certified, but the effects of zinc ‘catalysis’ in terms of
rate enhancements or product yields were small.
NMR resonance at 2.08 ppm, assigned to acetone, had
appeared. 20 ml of water were added and the solution
extracted twice with 20 ml of dichloromethane each.
The extract was evaporated to dryness and the yellow
residue picked up in CDCl for identification and quan-
3
1
tification by H NMR and UV spectroscopy.

62
R. Walz, H. Vahrenkamp / Inorganica Chimica Acta 314 (2001) 58–62
3.2.2. Oxidations of 2a and 2c
bach and H. Brombacher for starting materials and
A total of 200 mg (0.25 mmol) of 2a or 210 mg (0.25
help with the kinetic analysis.
mmol) of 2c in 20 ml of dichloromethane were treated
with 57 mg (0.25 mmol) of DDQ in 50 ml of acetoni-
trile. The red solutions were refluxed for 20 h and then
References
evaporated to dryness. The product yields determined
1
[
1] H. Eklund, B. Nordstr o¨ m, E. Zeppezauer, G. S o¨ derlund, I.
Ohlsson, T. Boiwe, B.O. S o¨ derberg, O. Tapia, C.I. Br a¨ nd e´ n, A.
Akeson, J. Mol. Biol. 102 (1976) 27.
by H NMR in CDCl were 25% for 2b and 28% for 2d.
3
3
.2.3. Aldehyde reductions
A total of 200 mg (0.29 mmol) of Tp
[
2] S. Ramaswamy, H. Eklund, B.V. Plapp, Biochemistry 33 (1994)
5230.
Cum,Me
Zn–OH
were dissolved in 50 ml of dichloromethane. 0.29 mmol
[3] F.M. Dickinson, Eur. J. Biochem. 41 (1974) 31.
[
[
4] A.J. Ganzhorn, B.V. Plapp, J. Biol. Chem. 263 (1988) 5446.
5] B. M u¨ ller, A. Schneider, M. Tesmer, H. Vahrenkamp, Inorg.
Chem. 38 (1999) 1900.
of the aldehyde (57 mg of C F CHO, 44 mg of p-NO –
6
5
2
C H CHO, 50 mg of p-CF –C H CHO) were added,
6
4
3
6
4
followed by 58 mg (0.29 mmol) of BNAH. After stir-
ring for 20 h the mixtures were evaporated to dryness
[
6] C. Kimblin, T. Hascall, G. Parkin, Inorg. Chem. 36 (1997) 5680.
[7] M. Bochmann, G.C. Bwembya, R. Grinter, A.K. Powell, K.J.
Webb, M.B. Hursthouse, K.M. Abdul Malik, M.A. Mazid,
Inorg. Chem. 33 (1994) 2290.
and the residues picked up in CDCl for product iden-
3
tification and yield determination. The benzylate com-
[
8] (a) B. Kaptein, L. Wang-Griffin, G. Barf, R.M. Kellogg, J.
Chem. Soc., Chem. Commun. (1987) 1457. (b) B. Kaptein, G.
Barf, R.M. Kellogg, F. van Bolhuis, J. Org. Chem. 55 (1990)
1890.
plexes 3a–c were crystallized from CH Cl /CH CN as
2
2
3
described [19]. The same reactions were performed
Cum,Me
without adding Tp
Zn–OH. After 20 h the
product distributions and the yields of the benzyl alco-
[9] N.J. Curtis, R.S. Brown, Can. J. Chem. 59 (1981) 65.
10] D.J. Creighton, J. Hajdu, D.S. Sigman, J. Am. Chem. Soc. 98
(1976) 4619.
11] M. Hughes, R.H. Prince, J. Inorg. Nucl. Chem. 40 (1978) 703.
12] A. Ohno, S. Yasui, R.A. Gase, S. Oka, V.K. Pandit, Bioorg.
Chem. 9 (1980) 199.
[
hols were very similar to those in the presence of
Cum,Me
Tp
Zn–OH
[
[
3.3. Kinetic measurements
[
[
[
13] F.J.F. Engbersen, A. Koudijs, H.C. van der Plas, J. Org. Chem.
55 (1990) 3547.
14] E. Kimura, M. Shionoya, A. Hoshino, T. Ikeda, Y. Yamada, J.
Am. Chem. Soc. 114 (1992) 10134.
−
1
Stock solutions of C F CHO (10 mmol
l
),
6
5
−1
Cum,Me
Tp
Zn–OH (10 mmol l ) and BNAH (0.1 mmol
) in chloroform were combined and diluted such
−
1
l
15] C. Bergquist, G. Parkin, Inorg. Chem. 38 (1999) 422.
that the concentrations for the measurement were ap-
[16] A. Shirra, C.J. Suckling, Tetrahedron Lett. 38 (1975) 3323.
[17] C. Sudbrake, H. Vahrenkamp, Eur. J. Inorg. Chem. (1999) 2009.
[18] (a) B. M u¨ ller, H. Vahrenkamp, Eur. J. Inorg. Chem. (1999) 117.
−
1
proximately 1.0 mmol l
complex and 0.01 mmol l
for the aldehyde and zinc
for BNAH. The mixture
−
1
(
(
b) B. M u¨ ller, H. Vahrenkamp, Eur. J. Inorg. Chem. (1999) 129.
c) B. M u¨ ller, H. Vahrenkamp, Eur. J. Inorg. Chem. (1999), 137.
was thermostated to 50°C in the UV spectrometer, and
the absorption of the band at 350 nm was recorded at
[
19] M. Ruf, F.A. Schell, R. Walz, H. Vahrenkamp, Chem. Ber./Re-
4
s intervals. The measurements were repeated three
cueil 130 (1997) 101.
Cum,Me
times both with and without Tp
value of I was taken after 10 min, i.e. approximately
Zn–OH. The
[20] R. Walz, K. Weis, M. Ruf, H. Vahrenkamp, Chem. Ber./Recueil
30 (1997) 975.
21] R. Walz, M. Ruf, H. Vahrenkamp, Eur. J. Inorg. Chem. (2000)
39.
1
8
[
5t1/2. The pseudo first-order rate constants were derived
1
from the equation ln[1−(I −I )]= −kobst.
t
ꢀ
[22] R. Alsfasser, M. Ruf, S. Trofimenko, H. Vahrenkamp, Chem.
Ber. 126 (1993) 703.
[
[
[
[
[
23] H. Brombacher, H. Vahrenkamp, Unpublished results.
24] K. Wallenfels, M. Gellerich, Chem. Ber. 92 (1959) 1406.
25] M. Ruf, H. Vahrenkamp, Inorg. Chem. 35 (1996) 6571.
26] P. M u¨ ller, J. Rocek, J. Am. Chem. Soc. 94 (1972) 2716.
27] S. Fukuzumi, N. Nishizawa, T. Takanaka, J. Chem. Soc., Perkin
Trans. 2 (1985) 371.
Acknowledgements
This work was supported by the Deutsche
Forschungsgemeinschaft. We are indebted to M. Rom-
.