Temperature dependent electrochemical investigations of molybdenum and tungsten oxobisdithiolene complexes

Article abstract of DOI:10.1039/b414853c

To achieve a better understanding why thermophilic and hyperthermophilic organisms use tungsten instead of molybdenum within the active sites of their molybdopterin dependent oxidases, electrochemical investigations of model complexes for the active sites

Full text of DOI:10.1039/b414853c

View Article Online / Journal Homepage / Table of Contents for this issue
F U L L P A P E R
Temperature dependent electrochemical investigations of
molybdenum and tungsten oxobisdithiolene complexes†
Carola Schulzke*
Institut fu¨r Anorganische Chemie der Universita¨t Go¨ttingen, 37077 Go¨ttingen, Germany.
E-mail: carola.schulzke@chem.uni-goettingen.de; Fax: ++49 551 393373;
Tel: ++49 551 393323
Received 24th September 2004, Accepted 10th December 2004
First published as an Advance Article on the web 13th January 2005
To achieve a better understanding why thermophilic and hyperthermophilic organisms use tungsten instead of
molybdenum within the active sites of their molybdopterin dependent oxidases, electrochemical investigations of
model complexes for the active sites of enzymes belonging to the DMSO reductase (molybdenum) and the aldehyde
oxidoreductase (tungsten) family have been undertaken. Cyclic voltammetry and differential pulse voltammetry of
four pairs of molybdenum and tungsten oxobisdithiolene compounds show huge differences in the response of their
redox potentials to rising or decreasing temperatures, depending on the substituents at the dithiolene group. The
mnt2− compounds (1a, 1b) respond with decreasing redox potentials E1/2 to rising temperatures whereas all other
compounds show positive gradients dE/dT. In every case the values for the gradients for the tungsten compounds are
greater than those for the molybdenum compounds. Six of the investigated compounds are known in the literature
and two compounds were newly synthesized. These two new compounds include the pyrane subunit of the native
molybdopterin ligand and should therefore be even better models for the active site of the molybdopterin containing
enzymes. The molybdenum/tungsten pair with these new ligands shows a remarkably small difference for the redox
potentials of the transition MIV↔ MV of only 30 mV at 25 ◦C and the reversion of the usual order with higher
potentials for the molybdenum than the tungsten compound at a temperature of 70 ◦C; a temperature that is in the
range where usually tungsten containing enzymes instead of molybdenum containing ones are found.
tasks regarding the substrate and the direction of the above
Introduction
reaction 1 and also some structural differences (see Fig. 1).9,10
Some of the most interesting forms of life are those, which
prosper in an environment that is life-hostile from a human
The reason for two different metals (molybdenum vs. tung-
sten) at the active sites is of great interest as its understanding
point of view: the so-called Extremophiles. These organisms
might give some insight into the catalytic processes of these
prefer habitats with high salt concentration, high or low pH
enzymes. It is discussed in the literature that because the
value, high pressure or extreme temperature. The study of these
thermophilic bacteria and hyperthermophilic archaebacteria
organisms and their extremozymes constantly gains significance,
belong to the oldest forms of life on this planet, the distribution
since they are interesting not only from the scientific, but also
of molybdenum and tungsten containing enzymes reflect the
from the industrial point of view.1
Thermophilic and hyperthermophilic microorganisms live in
evolution of these enzymes.11 Meaning: under the conditions
of the early planet earth, that were presumably hot and sulfur
or next to hot and sulfur rich submarine vents or springs
rich, the molybdopterin enzymes were developed with tungsten
on the surface. In contrast to most though not all of the
utilized for their active sites and with the change of these
mesophilic organisms these unusual life forms use tungsten
conditions descending organisms switched from tungsten to
instead of molybdenum at the active sites of their molybdopterin
molybdenum. But the question whether the change of metal
containing oxidases which catalyze hydroxylation and oxo-
happened due to environmental conditions because of supply-,
transfer reactions that are two electron redox reactions (see
stability- or redox potential-reasons is still unsolved.
reaction 1).
The assumption that the molybdenum and tungsten enzymes
R + H2O ↔ RO + 2 H+ + 2 e−
(1)
in general simply catalyze different reactions can be refuted
by two observations. First: by the existence of a molybde-
num dependent formate dehydrogenase from Escherichia coli12
and also a tungsten dependent formate dehydrogenase from
Chlostridium thermoaceticum (optimum growth temperature
In Fig. 1 the different groups of molybdopterin containing
enzymes following the Hille classification are shown. Despite
its name molybdopterin2 contains no molybdenum but is an
organic compound also referred to as pyranopterindithiolate,3
pterindithiolene4 or pterin-ene-dithiolate (see Fig. 1).5
This ligand is believed to function not only as a modulator
of the metal ion redox potential5,6 but also to couple the metal
into superexchange pathways for facilitating electron transfer
regeneration of the active sites during the catalytical process.7,8
The molybdopterin dependent enzymes in general are found in
almost any known organism where they can have quite different
T
opt is 55 ◦C).13,14 And second: the Methanobacterium thermoau-
totrophicum (Topt = 65 ◦C) expresses a molybdenum dependent
formylmethanofurane dehydrogenase as well as a tungsten
dependent formylmethanofurane DH depending on the environ-
mental conditions.15–18 For in vitro experiments these conditions
were determined by the supply of only molybdenum or only
tungsten. Interestingly, the two metals are not incorporated into
the same enzyme but rather into two different though similar
enzymes that are genetically separated and are expressed one or
the other in response to the environment.19,20
Strictly inorganic arguments for the supply-determined
change of metal follow the observation that in a sulfur and sulfide
rich surrounding, molybdenum has the tendency to form only
slightly soluble MoS2 while tungsten under the same conditions
† Electronic supplementary information (ESI) available: Fig. S1: Tem-
perature dependent plot of E1/2 vs. Fc/Fc+ of compound 4b, sample 1.
Typical course of the electrochemical measurements; indicated by the
arrows. See http://www.rsc.org/suppdata/dt/b4/b414853c/
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
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0
7 1 3
©

View Article Online
Fig. 1 Oxidized active sites of the molybdopterin dependent enzymes confirmed by EXAFS or X-ray structural analysis. The organization into
families follows that of Hille9 for molybdenum and that of Johnson et al.10 for tungsten. For others than the aldehyde oxidoreductase family no
structures are available for the tungsten enzymes.3
2−
2−
favours WO2S2 and WOS3 ions, which are soluble and can
be easily utilized by the organisms.21,22
3-Bromoflavanone,25
Na3K[WO2(CN)4]·6H2O,26
Na3K-
[MoO2(CN)4]·6H2O,27 [Et4N]2[MoO(mnt)2]28 ([Et4N]+ was
used as cation instead of [Bu4N]+), [Et4N]2[WO(mnt)2],22
[Et4N]2[MoO(bdt)2],29 [Et4N]2[WO(bdt)2],30 [Et4N]2[MoO(S2-
C2Me2)2]31 and [Et4N]2[WO(S2C2Me2)2]32 were prepared ac-
cording to literature procedures.
Arguments for the stability determined change rely on inves-
tigations of the LMCT energies of MoS4 and WS42−. They
2−
show that the ground state of the tungsten species is more
stabilized than that of the molybdenum species, because the
transfer of electron density from the sulfur to the metal requires
more energy for the tungsten compound.23 Furthermore model
complexes for the active sites of the molybdopterin enzymes with
tungsten show a higher thermal stability than their molybdenum
counterparts due to stronger p–p-interactions between sulfur
and metal.21 Under the extreme conditions in the environment
of the thermophilic and hyperthermophilic organisms this could
be of some advantage. On the other hand, one has to take into
account that the stability of an enzyme is usually determined
by its protein structure while the active site is shielded very
effectively from the environment.
Syntheses
3-(NN-Diethyldithiocarbamate)-flavanone I. 3-Bromoflava-
none (10.72 mmol; 3.25 g) dissolved in diethylether (80 mL) was
added dropwise over a period of 15 min to a refluxing solution
of sodium N,N-diethyldithiocarbamate (11.76 mmol; 2.65 g) in
ethanol (10 mL). After the addition was complete the mixture
was refluxed at 50 ◦C for another 10 min. The solvents were
removed in vacuo and the residue extracted with water (70 mL)
and dichloromethane (70 mL). The water layer was washed
3 times with CH2Cl2 (10 mL). The combined organic layers
were washed 3 times with brine (10 mL), dried over MgSO4
and the solvent removed in vacuo. The residue was recrystallized
from ethanol giving a pale yellow crystalline solid. Yield: 2.45 g
(6.59 mmol; 62%).
Elemental analysis: calcd. for C20H21NO2S2: C: 64.66%, H:
5.70%, N: 3.77%; expt: C: 65.18%, H: 6.21%, N: 3.85%; IR
(KBr) [cm−1]: 3061 (m, mAr H), 3033 (m, mAr H), 2980 (m, mCHaliph.),
2934 (m, mCHaliph.), 1689 (vs, mCO), 765 (vs), 698 (s), 315 (vs); H-
NMR (CD2Cl2, 400 MHz) [ppm]: d = 1.25 (d), 3.72, 5.71 (d),
5.80 (d), 7.02 (s), 7.29 (s), 7.40–7.51 (m), 7.80(s).
Kinetic investigations of the oxidation of benzoin by
[MO2(bdt)2]2− (M Mo, W) at different temperatures revealed
=
that the reactivity improvement at higher temperatures was
greater for the tungsten compound than for the molybdenum
compound.24 Although within the temperature range of the
experiment (RT → 100 ◦C) the molybdenum compounds
reactivity was always better than that of the tungsten compound,
this is an indication for a reactivity related change of metal at the
enzymes active sites. Also based on a reactivity related change
of metal is the suggestion that the tungsten enzymes might be
better suited for higher temperatures than their molybdenum
counterparts due to a different shift of their redox potentials
with rising temperature. To test this assumption the temperature
dependence of the redox properties of pairs of analogous
known and new molybdenum and tungsten oxobisdithiolene
compounds that model the active sites of the enzymes of
the DMSO reductase family (molybdenum) or the aldehyde
oxidoreductase family (tungsten) (see Fig. 1) were investigated
by cyclic voltammetry or differential pulse voltammetry.
–
–
1
2-(N,N-Diethylamino)-4,5-flavanyl-1,3-dithiolium hydrogen-
sulfate II. Sulfuric acid (2.5 mL; 95%) was added dropwise
over a period of 20 min to vigorously stirred solid I (5.32 mmol;
1.977 g) cooled by a water/ice-bath to 0 ◦C. ◦After the addition
was complete the mixture was◦heated to 60 C for a period of
10 min and cooled again to 0 C before 150 mL of ether were
added. It was stirred vigorously for 30 min giving a greenish
gumlike solid. The solvent was decanted and this procedure
repeated twice. Acetone (40 mL) was added to the now honeylike
residue and the mixture again stirred vigorously. The solvent
was decanted and this procedure repeated six times until giving
a microcrystalline white solid that was isolated by filtration,
washed 3 times with acetone (10 mL) and dried in vacuo. Yield:
1.44 g (3.19 mmol, 60%).
Experimental
Materials
All chemical reagents were purchased from Acros and used
without further purification. Water and ethanol were deaerated
with N2. Solvents (p.a.) for the organic reactions were used as
received. All reactions were carried out in an atmosphere of N2
using standard Schlenk line techniques, and only the isolation
procedures for the organic reactions were carried out in air.
Elemental analysis: calcd. for C20H21NO5S3: C: 53.19%, H:
4.69%; N: 3.10%; expt: C: 51.79%, H: 4.68%; N: 2.84%; IR
(KBr) [cm−1]: 3057 (m, mAr H), 3017 (m, mAr H), 2980 (m, mCHaliph.),
–
–
2924 (m, mCHaliph.), 1584 (m), 1540 (s), 1223 (vs), 1195 (vs),
7 1 4
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0

View Article Online
1
703 (m), 590 (s), 467 (s); H-NMR (CD2Cl2, 400 MHz) [ppm]:
Other physical measurements
d = 1.1–1.4 (m), 3.55–3.95 (m), 5.25 (s), 6.65–7.70 (m).
Elemental analyses were determined with a Heraeus CHN–O-
Rapid elemental analyzer.
[PPh4]2[MoIVO(fdt)2]·EtOH 4a. A solution of II (1.10 mmol;
0.495 g) in ethanol (8 mL) was added to a solution of
Na3K[MoO2(CN)4]·6 H2O (0.55 mmol; 0.267 g) and NaOH
(4.42 mmol; 0.177 g) in water (8 mL). The reaction mixture
was heated to 65 ◦C for 3 h. After cooling the reaction mixture
Infrared spectra were recorded as KBr pellets on a Matt-
son Genesis FT-MIR-spectrophotometer ATI from 4000 to
400 cm−1 and as RbI pellets on a Bruker FIR-spectrophotometer
IFS 66 from 500 to 200 cm−1. Absorption spectra were obtained
on a Varian Cary 5 spectrophotometer as 10−4 M solutions in
CH3CN (Uvasol).
◦
to 30 C a solution of [PPh4]Cl (1.1 mmol; 0.431g) in ethanol
1H-NMR spectra were recorded on a Bruker NMR-
spectrometer AM 400 and on a Bruker Advance DRX 500 and
referenced to the chemical shifts of the deuterated solvent.
(8 mL) was added dropwise giving a red–brown precipitate which
was isolated by filtration, washed with water (3 times; 4 mL) and
dried in vacuo. Yield: 378.6 mg (0.275 mmol, 50%).
Elemental analysis: calcd. for C80H64O4P2S4Mo: C: 69.86%,
H: 4.69%; expt: C: 69.19%, H: 4.62%; IR (KBr) [cm−1]: 3054
Results and discussion
(m, mAr H), 3021 (m, mAr H), 1598 (m), 1585 (m), 1483 (m), 1436
–
–
Choice and syntheses of compounds
(s, mP Ph), 1107 (s), 1029 (m), 1017 (shoulder, mMo O), 996 (m),
=
–
722 (s, dAr H), 689 (s, dAr H), 527 (vs); F-IR (RbI) [cm−1]: 392 (w,
–
–
The four pairs of analogous molybdenum and tungsten oxobis-
dithiolene complexes that have been synthesized (see Fig. 2) were
chosen for the different strength of non-innocence character
of their dithiolene ligands; a characteristic that is also part
of the natural molybdopterin leading to the consideration of
the metal and the coordinated molybdopterin as one redox
center.33,34 This is because non-innocence means that the ligand
itself is redox active (see Fig. 3) and the division of electrons
between metal centre and ligand is flexible (Fig. 3). Mesomeric
distribution of the intermediate unpaired electron as well as
electron withdrawing groups at the dithiolene enhance its non-
innocence character and hence the covalency of the metal sulfur
bond.35,36 In this group of ligand systems the non-innocence
decreases therefore from mnt2− over bdt2− to S2C2Me22− with the
fdt2− ligand with the highest mesomeric potential presumably in
the range of the mnt2− ligand.
mMo S), 375 (vw, mMo S); UV-vis (CH3CN solution, 10−4 M) [nm]:
–
–
1
557, 367, 274, 269; H-NMR (CD3CN, 500 MHz) [ppm]: d =
1.32 (s), 2.14 (m), 5.38 (s), 7.70 (m), 7.75 (m), 7.93 (t).
[PPh4]2[WIVO(fdt)2]·EtOH 4b. A solution of II (1.99 mmol;
0.90 g) in ethanol (16 mL) was added to a solution of
Na3K[WO2(CN)4]·6 H2O (1.01 mmol; 0.58 g) and NaOH
(13.75 mmol; 0.55 g) in water (16 mL). The reaction mixture
was heated to 65 ◦C for 3 h. After cooling the reaction mixture
◦
to 30 C a solution of [PPh4]Cl (2.14 mmol; 0.85 g) in ethanol
(16 mL) was added dropwise giving a dark brown precipitate
which was isolated by filtration, washed with water (3 times;
8 mL) and dried in vacuo. Yield: 184.20 mg (0.048 mmol, 12%).
Elemental analysis: calcd. for C80H64O4P2S4W: C: 65.66%, H:
4.41%; expt: C: 65.08%, H: 4.80%; IR (KBr) [cm−1]: 3053 (m,
mAr H), 3019 (m, mAr H), 1600 (m), 1585 (m), 1483 (m), 1436 (s,
–
–
–
–
mP Ph), 1107 (s), 1029 (m), 995 (m, mW O), 722 (s, dAr H), 689 (s,
=
–
dAr H), 526 (vs); F-IR (RbI) [cm−1]: 392 (w, mW S), 363 (w, mW S),
–
–
348 (vw, mW S); UV-vis (CH3CN solution, 10−4 M) [nm]: 560,
–
1
449, 360, 274, 269; H-NMR (CD3CN, 500 MHz) [ppm]: d =
1.27 (s), 2.10 (m), 5.39 (s), 7.67 (m), 7.72 (m), 7.90 (t).
Electrochemical measurements
Electrochemical measurements were performed with an AUTO-
LAB PGSTAT12 potentiostat/galvanostat with a glassy carbon
working electrode, a platinum wire reference electrode and a
platinum wire auxiliary electrode in acetonitrile solutions within
a glovebox under argon atmosphere. Supporting electrolyte was
0.1 M Bu4NPF6 used as received from Fluka. Acetonitrile for
the electrochemistry was predried over CaCl2, dried by reflux
over CaH2 for two days and distilled onto freshly regenerated
molecular sieve A3. Referenciation was done internally using
the ferrocene/ferrocenium couple. Temperature during the
measurements was controlled with a Julabo refrigerated/heating
circulator FP50-MV. The solutions have been prepared within
a glove box by dissolving 10−2 mol of the Bu4NPF6 electrolyte
and exactly 10−4 mol of the compounds in the cases of 1ab, 2ab,
3ab and due to poor solubility 0.5 × 10−5 mol in the cases
Fig. 2 Synthesized dithiolene complexes. (a) The complexes 4a and 4b
were synthesized as a mixture of cis- and trans-isomers; here only the
trans-isomer is shown.
of 4a and 4b in 100 mL of acetonitrile resulting in 10−3
M
solutions and 0.5 × 10−4 M solutions respectively. For each
measurement 6 mL of these solutions were transferred to the
double wall electrochemistry cell. The temperature of the cell and
the solution within was controlled with the FP50-MV sending
water of the desired temperature through the double wall. The
redox potentials of the couple [MO(dt)2]2− ↔ [MO(dt)2]− were
determined by cyclic voltammetry (scan rate = 100 mV s−1) in
the cases of 1ab, 2a, 3ab and by differential pulse voltammetry
(step potential = 2.55 mV; modulation amplitude = 25.02 mV;
modulation time = 0.05 s) in the cases of 2b and 4ab. The samples
were stirred prior to the measurements for at least 10 min at the
desired temperature.
Fig. 3 (a) non-innocence of a dithiolene ligand; (b) redox interaction
between metal center and dithiolene ligand.
It is remarkable that the preparation methods obviously
depend strongly on the ligands used resulting in completely
different syntheses for the four pairs of molybdenum and
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0
7 1 5

View Article Online
Fig. 4 Synthesis of the compounds [PPh4]2[MoIVO(fdt)2]·EtOH 4a and [PPh4]2[WIVO(fdt)2]·EtOH 4b following a method developed by Joule and
Garner.37
tungsten compounds that seem not to be exchangeable. The
syntheses of the six known molybdenum (1a, 2a, 3a) and
tungsten (1b, 2b, 3b) complexes were exactly like those described
in the respective literature.22,28–32 The syntheses of the two new
complexes (4a, 4b) (see Fig. 4) followed largely a procedure
developed by Garner and Joule.37
Figs. 5 and 6. Only the DPV oxidations are shown here but the
reductive DPV shows in both cases the same two peaks similar
in height.
The flavanone was chosen as starting material because it
is commercially available and promised the synthesis of a
ligand system that not only includes the dithiolene group but
also the pyrane feature of the enzymatic molybdopterin and
therefore a more accurate modelling of the natural ligand system.
The syntheses of the intermediates were carried out according
to the literature procedures of similar compounds with the
exception of the ring closing step with sulfuric acid. Here
the isolation of the product turned out to be rather difficult.
Recrystallization from ethanol was not possible because with
this solvent the aminodithiolium hydrogensulfate and probably
some side products formed a honeylike residue that even in
vacuo could not be dried to give a solid product. Instead the
residue was stirred several times vigorously in acetone in which
some side products of the reaction were soluble, leaving the
desired product as a microcrystalline precipitate. The synthesis
of the molybdenum and tungsten complexes followed again the
literature procedures with a rather poor yield for the tungsten
compound. Crystals of these complexes could not be obtained
possibly due to the fact that these complexes form cis- and trans-
(ligand orientation) and R- and S- (position of the phenyl group
at the pyrane ring) isomers i.e. a mixture of diastereomers, which
usually results in a poor crystallization behaviour. However,
the identity of the new complexes was clearly established by
elemental analyses and NMR spectroscopy as well as by IR
Fig. 5 Differential pulse voltammograms of 4a (solid) and 4b (dotted)
at 25 ◦C. Data is referenced against Fc/Fc+ (= 0 V in this diagram).
+
spectroscopy, showing dominating bands for the PPh4 cation
=
and characteristic though less intense bands for the M O
and M–S valences.
Electrochemistry
The compounds 1a, 1b, 2a, 3a and 3b were examined by cyclic
voltammetry (CV), the compounds 2b, 4a and 4b by differential
pulse voltammetry. The obtained CV spectra of the latter were
of inferior quality (for 4a and 4b due to the poor solubility
of these compounds). Therefore the peak determination was
problematic. The differential pulse voltammograms of 4a and 4b
and for comparison the cyclic voltammogram of 4a are shown in
Fig. 6 Cyclic voltammogram of 4a referenced vs. Fc/Fc+ (= 0 V in this
diagram).
7 1 6
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0

View Article Online
Table 1 E1/2 (MIV↔MV) vs. Fc/Fc+ and vs. SCE of this work, the differences between corresponding molybdenum and tungsten compounds and
comparison with literature data
Complex
E1/2 vs. Fc/Fc+/Va
E1/2 vs. SCE/Vb
DE1/2/V
0.18
E1/2 vs. SCE/V Literature
DE1/2/V Literature
0.26/0.17/0.17f
0.24/0.28/0.29
0.29
[MoO(mnt)2]2− 1a
[WO(mnt)2]2− 1b
+ 0.09 (0.005)
−0.09 (0.015)
+0.46
+0.28
+0.48c/+0.395d
+0.2238e
[MoO(bdt)2]2− 2a
[WO(bdt)2]2− 2b
−0.78 (0.006)
−1.30 (0.016)
−0.41
−0.93
−0.39c/−0.35g/−0.34h
−0.63h
0.52
[MoO(S2C2Me2)2]2− 3a
[WO(S2C2Me2)2]2− 3b
−0.74 (0.037)
−1.33 (0.005)
−0.37
−0.96
−0.62i
0.59
−0.91j
[MoO(fdt)2]2− 4a
[WO(fdt)2]2− 4b
−1.33 (0.034)
−0.96
−1.36 (0.042)
−0.99
0.03
a Average over all data at 25 ◦C for each compound (standard deviation in parentheses). b Calculated by adding 0.37 V43 to obtained data vs. Fc/Fc+.
c Ref. 38. d Calculated from data vs. Ag/AgCl, ref. 39. e Calculated from data vs. NHE, ref. 40. f Value of difference from ref. 35. g Ref. 29. h Ref. 30;
i Ref. 31. j Ref. 32.
The DPVs of 4a and 4b are very similar with the respec-
tive peaks of the oxidation process MIV→MV as well as for
the MV→MVI oxidation close together. As expected both redox
potentials for the molybdenum compounds are (at 25 ◦C) more
positive than the potentials for the tungsten compound but the
difference is extraordinarily small.
or vs. Ag/AgCl was recalculated to obtain the potentials vs.
SCE by subtracting 0.2412 V or 0.045 V, respectively, as
these are the differences to the potential of the SCE.44 The
difference between the literature data and the new data for
2b may also be based on the fact that different solvents were
used for the measurements (acetonitrile vs. dimethylformamide);
other details of the experiment are not mentioned and further
comparison of the experimental set-up not possible.
The temperature dependent measurements were carried out
as follows. Of each compound about 100 data points have been
◦
◦
measured at different temperatures from 5 C to 65 C. Mea-
As expected for all ligands at 25 ◦C the redox potentials
for the molybdenum compounds are more positive than those
for the tungsten compounds, but the differences vary. Garner
et. al. stated that the difference of redox potentials between
corresponding molybdenum and tungsten dithiolene complexes
is usually in the range of 225 mV, but the compounds they
are referring to are not very diverse with respect to size and
mesomeric potential.37,45,46 On the other hand it was said by
Holm et. al. that the difference of the redox potentials for
the molybdenum and the tungsten complexes depend upon
the strength of the non-innocent character of the respective
ligands. That means: the more pronounced the non-innocence
the smaller the difference of the redox potentials.35,36 Within
the pairs of molybdenum and tungsten compounds examined
here the mnt2− ligand (complexes 1a and 1b) with the two
strongly electron withdrawing CN− groups, that facilitate the
breaking of the dithiolene C–C–p-bond, was expected to be the
most non-innocent ligand of all. But the pair with the smallest
difference between the redox potentials for the molybdenum
and the tungsten complex, extraordinary 30 mV, is that with the
fdt2− ligand (complexes 4a and 4b) which is not only the most
voluminous ligand with the highest mesomeric potential but
also the ligand system that resembles the natural molybdopterin
ligand closest by incorporating the pyrane subunit. Therefore
it is likely that this feature plays an important role within the
natural molybdopterin with respect to the tuning of the redox
properties of the enzymes’ active sites.
surements at temperatures up to 75 ◦C (about 7 below the
◦
boiling point of acetonitrile) resulted in a substantial change
of the potentials due to the loss of solvent, therefore only a
temperature range of 60 ◦C could be examined with reliable
results. This temperature region was measured for every sample
in circles by changing the temperature at each step by 10 ◦C
and by 5 ◦C at the turning points. For instance starting with
◦
◦
◦
◦
◦
25 C the routine was as follows: 25 C, 35 C, 45 C, 55 C,
65 ◦C, 60 ◦C, 50 ◦C, 40 ◦C, 30 ◦C, 20 ◦C, 10 ◦C, 5 ◦C,
15 ◦C, 25 ◦C. The starting points of the routine were also
varied for every new sample. This was carried out to achieve
neutralization of any phenomenon that might occur due to
deterioration or other time dependent change of the set-up.
Every sample was finally measured again at 25 ◦C after ferrocene
had been added to the solution for referenciation. Then the
ferrocene/ferrocenium couple was measured at 25 ◦C. (A typical
course of the measurements of the MIV↔MV redox potentials
with time and with temperature can be found in the ESI.†)
The compound 3a was deteriorating during the measurements
possibly due to an electrode mediated disproportionation which
could be observed by the fading of the CV signal. Of this
compound no full circle could be measured. Of each sample
of 3a only three to eight CVs could be obtained. Therefore of
this compound only 80 instead of about 100 data-points were
collected and the quality of the data is substantially lower than
that for the other compounds.
The compounds average redox potentials at 25 ◦C determined
in this work as well as a comparison with literature data is
shown in table 1. Partly there are some discrepancies between
literature and the new data (differences up to 300 mV for
[WO(bdt)2]2− 2b). This may occur due to the fact that there is no
truly reliable reference electrode for non-aqueous solutions.41
In accordance with an IUPAC recommendation42 potentials
should be referenced against the ferrocene/ferrocenium couple
for better comparability of different investigations, which was
performed here and the data obtained by this method are
listed in the first column of table 1. To be able to compare
the data of this work and the literature data, the potentials were
recalculated against SCE (standard calomel electrode) assuming
that the potential of Fc/Fc+ against SCE in acetonitrile solution
is 0.37 V.43 The literature data that was referenced vs. NHE
The differences of redox potentials for 2a/2b and 3a/3b are
similar with a slightly smaller difference for the former (the
bdt2−-pair) as was expected. The differences for 2a/2b and
3a/3b observed in this study are substantially larger than those
reported in the literature. This may be based on the fact that the
reported potentials were not measured consecutively under
the exact same conditions as was performed here and also on
the above mentioned problem with the reference electrodes.
The results of all temperature dependent measurements of the
eight compounds are shown in Fig. 7. The redox potentials E1/2
were plotted against the temperature, linear fits employed and
these fits extrapolated as far as necessary.
The gradients of the eight graphs and the virtual cross-points
ofthe particular twographs ofeachpair are assembledin Table 2.
Based on eqn. (2)47 (with dE/dT = gradient; DS = entropy
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0
7 1 7

View Article Online
Fig. 7 Plots of the redox potentials E1/2 (MIV↔MV; vs. Fc/Fc+) against temperature and their linear fits. Note that the scales of the diagrams vary.
change for the reduction process; z = number of transferred
electrons = 1; F = Faraday constant = 9.64853 × 10−4 C mol−1)
the DS values for the reduction process were calculated for each
compound and inserted into Table 2 as well.
with rising temperature, as the influence of DS increases, the
redox potentials of 1a and 1b decrease. Oxidation becomes easier
while the gain of Gibbs free energy for the reduction process
declines. For the other compounds (2a/2b, 3a/3b and 4a/4b)
DS for the reduction is positive and therefore its contribution
to DG negative. Thus the entropic part of the Gibbs equation
favours the reduction over the oxidation.
dE/dT = DS/zF
(2)
The most striking characteristic is the behaviour of com-
pounds 1a and 1b that show a reverse response to rising
temperature compared with the six other compounds. Obviously
for 1a and 1b the gradient dE/dT and thus the entropy change
for the reduction is negative. Therefore the term −TDS is
positive as well as its contribution to DG. This means that
from the entropic point of view the oxidative form MV is
thermodynamically favoured over the reduced form MIV. Thus
Comparison of the average redox potentials at 25 ◦C leads
to the following conclusion. The redox potentials of 1a and
1b are the most positive ones within this series and hence the
reduction process is easier than for all the other compounds.
While DS for the reduction is negative (favouring the oxidation),
obviously DH has to be substantially more negative (favouring
the reduction) than for the other compounds, and hence the
Table 2 Gradients, cross-points of the linear fits for the temperature dependence of the redox potentials E1/2 (MIV↔MV) and from the gradients
calculated values of the entropy change DS for the respective reduction processes
Complex
Gradient/mV K−1a
Cross-point temperature/Kb
Non existent (−31(30))
4179 (1000)
DS/J mol−1 K−1c
[MoO(mnt)2]2− 1a
[WO(mnt)2]2− 1b
−1.14 (0.043)
−1.70 (0.102)
−110.0 (4.15)
−164.0 (9.84)
[MoO(bdt)2]2− 2a
[WO(bdt)2]2− 2b
0.17 (0.033)
0.31 (0.094)
16.4 (3.18)
29.9 (9.07)
[MoO(S2C2Me2)2]2− 3a
[WO(S2C2Me2)2]2− 3b
0.71 (0.344)
0.80 (0.072)
68.5 (33.19)
77.2 (6.95)
6722 (1500)
[MoO(fdt)2]2− 4a
[WO(fdt)2]2− 4b
0.65 (0.131)
1.10 (0.149)
62.7 (12.64)
106.1 (14.38)
343 (26)
a Standard deviation in parentheses as obtained from the linear fit. b Errors in parentheses estimated by the calculation of extreme values. c Errors in
parentheses calculated from the standard deviation of the gradients.
7 1 8
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0

View Article Online
resulting DG is more negative as well. The difference of the
DH values for instance between 1a and 3a (the molybdenum
compound with a redox potential most closely to that of 1a) is
about 133 kJ mol−1 as can be calculated from the Gibbs equation
with the knowledge of the DS values and the relation between
DG and the redox potential. The reason for this particular
behaviour has to be the electron withdrawing effect of the
CN groups at the dithiolene, which reduces electron density
at the dithiolene group and eventually at the metal centre, thus
facilitating reduction. In contrast to the other complexes for
compounds 1a and 1b the reduced is the more stable form (redox
potentials referenced vs. NHE are positive) and an unrestricted
distribution of electrons in a system solely driven by the enthalpy
would lead to a complete reversion of the oxidized to the reduced
form. This would mean a minimum of disorder to which DS
for the reduction process is likely to be opposed and therefore
negative, as can be observed here. For the other compounds the
non-innocence is manifested in an electron distribution to the
metal (see Fig. 3) and therefore in a facilitated oxidation. This
matches a redox potential referenced vs. NHE that is negative.
Such a system solely driven by the enthalpy would lead to a
complete reversion of the reduced to the oxidized form, resulting
in a minimum of disorder to which the entropy is opposed.
Therefore DS for the reduction process, which increases the
disorder, is positive.
The virtual cross-points of the graphs for 2a and 2b and 3a
and 3b are in temperature regions that have no significance for
any form of life; that for 1a and 1b (at calculated −31 K) does not
even exist. But the cross-point for the graphs of 4a and 4b, the
models that resemble the natural enzymes active sites closest, is
at 343 K (70 ◦C) and hence in a region in which usually tungsten
is used within the active sites of the molybdopterin containing
enzymes. For these model compounds the usual characteristic
that the redox potential for the molybdenum compound is more
positive than that for the tungsten compound is reversed at
temperatures above 70 ◦C. This could be an indication for a
redox property related change of metal within molybdopterin
containing enzymes. It should be noted though that the redox
potentials for 4a and 4b are only separated by a minimum of
30 mV even at 25 ◦C, and thus the redox potential may not be
the main reason for the incorporation of either of the two metals.
In all cases the gradient of the tungsten compound is of greater
value than that for the corresponding molybdenum compound,
meaning that the tungsten compounds’ response to temperature
change is always more drastic than that of the molybdenum
compound. The stronger dependence on the temperature of
the redox potentials of the tungsten compounds seems to be a
disadvantage with respect to the function of the enzymes unless
of course the response of the substrates redox potentials is about
equal to that of the respective tungsten enzyme.
therefore in a temperature region in which organisms usually
utilize tungsten instead of molybdenum. This may hint at a
change of metal within molybdopterin containing enzymes
caused by a reversal of the usual order (E1/2(Mo) > E1/2(W)) for
the redox potentials of the metal sites at a certain temperature.
But because even at 25 ◦C the redox potentials of 4a and 4b are
exceptionally close together it can not yet be concluded that this
reversal would be the main reason for the metal exchange within
theses enzymes.
The cross-points of the other three pairs are in temperature
regions that are of no importance for any kind of life or
even do not exist (<0 K for 1a and 1b) indicating that the
pyrane subunit of the natural molybdopterin ligand plays an
important role with respect to the tuning of the enzymes redox
potentials. Whether this is generally valid will be explored
further by the synthesis and electrochemical examination of
more pyranedithiolene ligands and their respective molybdenum
and tungsten complexes.
The gradients of the linear fits of the tungsten compounds
redox potentials are in every case of greater value (positive
or negative) than those of the molybdenum compounds. Their
redox potentials are more susceptible to temperature change
than those of the molybdenum complexes. Should this also be
the case within the molybdopterin enzymes, this characteristic
would probably be unfavourable unless the substrates potentials
behave in similar ways. This will be the subject of further
investigations.
Acknowledgements
Generous financial support by the DFG (Deutsche Forschungs-
gemeinschaft) and the Institut fu¨r Anorganische Chemie, Georg-
August-Universita¨t Go¨ttingen is gratefully acknowledged as
well as the opportunity that Prof. Felix Tuczek, Institut
fu¨r Anorganische Chemie, Christain-Albrechts-Universita¨t Kiel
gave me to start this work in his group.
References
1 D. W. Hough and M. J. Danson, Curr. Opin. Chem. Biol., 1999, 3(1),
39–46 and literature cited therein.
2 K. V. Rajagopalan, Adv. Enzymol. Relat. Areas Mol. Biol., 1991, 64,
215–290.
3 J. H. Enemark, J. J. J. A. Cooney, J.-J. Wang and R. H. Holm, Chem.
Rev., 2004, 104(2), 1175–1200 and literature cited therein.
4 C. Lorber, J. P. Donahue, C. A. Goddard, E. Nordlander and R. H.
Holm, J. Am. Chem. Soc., 1998, 120(32), 8102–8112.
5 E. I. Stiefel, J. Chem. Soc., Dalton Trans., 1997, (21), 3915–3923.
6 S. Gardlik and K. V. Rajagopalan, J. Biol. Chem., 1991, 266(8), 4889–
4895.
7 F. E. Inscore, R. McNaughton, B. L. Westcott, M. E. Helton, R.
Jones, I. K. Dhavan, J. H. Enemark and M. L. Kirk, Inorg. Chem.,
1999, 38(7), 1401–1410.
8 M. E. Helton, N. E. Gruhn, R. L. McNaughton and M. L. Kirk,
Inorg. Chem., 2000, 39(11), 2273–2278.
Conclusion
9 R. Hille, Chem. Rev., 1996, 96(7), 2757–2816.
Only the compounds [MoO(mnt)2]2− 1a and [WO(mnt)2]2− 1b
show gradients dE/dT that are negative. For these compounds
the redox potentials and the gain of Gibbs free energy decreases
with rising temperature. They are also the only complexes within
this study that show positive redox potentials vs. NHE. There-
fore DS as well as DH for the reduction process are negative. This
in comparison with the other compounds reverse behaviour is
based on the fact that the mnt2−-ligand contains two strongly
electron withdrawing CN groups which from the enthalpic point
of view support the reduced forms of the complexes. This
indicates that 1a and 1b are not as good for modelling the
molybdopterin containing enzymes, in which no such groups
occur, than the other dithiolene complexes.
10 M. K. Johnson, D. C. Rees and M. W. W. Adams, Chem. Rev., 1996,
96(7), 2817–2839.
11 F. A. M. de Bok, P.-L. Hagedoorn, P. J. Silva, W. R. Hagen, E. Schiltz,
K. Fritsche and A. J. M. Stams, Eur. J. Biochem., 2003, 270(11), 2476–
2485.
12 J. C. Boyington, V. N. Gladyshev, S. V. Khangulov, T. C. Stadtman
and P. D. Sun, Science, 1997, 275, 1305–1308.
13 I. Yamamoto, T. Saiki, S.-M. Liu and L. G. Ljungdahl, J. Biol. Chem.,
1983, 258(3), 1826–1832.
14 J. C. Deaton, E. I. Solomon, G. D. Watt, P. J. Wetherbee and C. N.
Dufor, Biochem. Biophys. Res. Commun., 1987, 149(2), 424–430.
15 R. A. Schmitz, S. P. Albracht and R. K. Thauer, Eur. J. Biochem.,
1992, 209(3), 1013–1018.
16 P. A. Bertram, R. A. Schmitz, D. Linder and R. K. Thauer, Arch.
Microbiol., 1994, 161(3), 220–228.
17 P. A. Bertram, M. Karrasch, R. A. Schmitz, R. Bocher, S. P. Albracht
and R. K. Thauer, Eur. J. Biochem., 1994, 220(2), 477–484.
18 A. Hochheimer, R. Hedderich and R. K. Thauer, Arch. Microbiol.,
1998, 170(5), 389–393.
The redox potentials of 4a and 4b (the compounds that
include the pyrane subunit of the molybdopterin l◦igand) are
exceptionally close together with a difference at 25 C of only
30 mV. The cross-point of their linear fits is at 343 K (70 ◦C) and
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0
7 1 9

View Article Online
19 A. Hochheimer, R. A. Schmitz, R. K. Thauer and R. Hedderich,
Eur. J. Biochem., 1995, 234(3), 910–920.
20 A. Hochheimer, D. Linder, R. K. Thauer and R. Hedderich,
Eur. J. Biochem., 1996, 242(1), 156–162.
35 G. C. Tucci, J. P. Donahue and R. H. Holm, Inorg. Chem., 1998,
37(7), 1602–1608.
36 D. V. Fomitchev, B. S. Lim and R. H. Holm, Inorg. Chem., 2001,
40(4), 645–654.
21 M. K. Johnson, D. C. Rees and M. W. W. Adams, Chem. Rev., 1996,
96(7), 2817–2840.
22 S. K. Das, D. Biswas, R. Maiti and S. Sarkar, J. Am. Chem. Soc.,
1996, 118(6), 1387–1397.
23 J. Bernholc and E. I. Stiefel, Inorg. Chem., 1985, 24(9), 1323–1330.
24 H. Oku, N. Ueyama and A. Nakamura, Chem. Lett., 1996, 1131–
1132.
37 E. S. Davies, R. L. Beddoes, D. Collison, A. Dinsmore, A. Docrat,
J. A. Joule, C. R. Wilson and C. D. Garner, J. Chem. Soc., Dalton
Trans., 1997, (21), 3985–3996.
38 J. P. Donahue, C. R. Goldsmith, U. Nadiminti and R. H. Holm,
J. Am. Chem. Soc., 1998, 120(49), 12869–12881.
39 R. Maiti, K. Nagarajan and S. Sarkar, J. Mol. Struct., 2003, 656(1–3),
169–176.
25 N. Takeno, T. Fukushima and S. Takeda, Bull. Chem. Soc. Jpn., 1985,
40 S. Sarkar and S. K. Das, Proc. Indian Acad. Sci., Chem. Sci., 1992,
104(4), 533–534.
58(5), 1599–1600.
26 A. Roodt, S. S. Basson and J. G. Leipoldt, Polyhedron, 1994, 13(4),
599–607.
41 K. Izutsu, Electrochemistry in Nonaqueous Solutions, Wiley-VCH,
Weinheim, 2002, pp. 171–174.
27 J. P. Smit, W. Purcell, A. Roodt and J. G. Leipoldt, Polyhedron, 1993,
42 G. Gritzer and J. Kuta, Pure Appl. Chem., 1984, 56(4), 461–466.
43 D. K. Gosser, Jr., Cylic Voltammetry—Simulation and Analysis
of Reaction Mechanisms, Wiley-VCH, New York, USA, 1993,
p. 34.
12(18), 2271–2277.
28 S. K. Das, P. K. Chaudhury, D. Biswas and S. Sarkar, J. Am. Chem.
Soc., 1994, 116(20), 9061–9070.
29 S. Boyde, S. R. Ellis, C. D. Garner and W. Clegg, J. Chem. Soc., Chem.
Commun., 1986, (19), 1541–1543.
44 A. J. Bard, L. R. Faulkner, Electrochemical Methods—Fundamentals
and Applications, Second Edition, Wiley & sons, New York, USA,
2001, inner back cover.
45 C. D. Garner and L. J. Stewart, in Metal Ions in Biological Systems—
Molybdenum and Tungsten: their Roles in Biological Processes, ed.
A. Sigel and H. Sigel, Marcel Dekker, New York, USA, 2002,
vol. 39, pp. 699–726.
46 E. S. Davies, G. M. Aston, R. L. Beddoes, D. Collison, A. Dinsmore,
A. Docrat, J. A. Joule, C. R. Wilson and C. D. Garner, J. Chem. Soc.,
Dalton Trans., 1998, (21), 3647–3656.
30 N. Ueyama, H. Oku and A. Nakamura, J. Am. Chem. Soc., 1992,
114(18), 7310–7311.
31 B. S. Lim, J. P. Donahue and R. H. Holm, Inorg. Chem., 2000, 39(2),
263–273.
32 C. A. Goddard and R. H. Holm, Inorg. Chem., 1999, 38(23), 5389–
5398.
33 A. Thapper, C. Lorber, J. Fryxelius, A. Behrens and E. Nordlander,
J. Inorg. Biochem., 2000, 79(1–4), 67–74.
34 Y. Hu, S. Faham, R. Roy, M. W. W. Adams and D. C. Rees, J. Mol.
Biol., 1999, 286(3), 899–914.
47 D. K. Gosser, Jr., Cylic Voltammetry—Simulation and Analysis of
Reaction Mechanisms, Wiley-VCH, New York, USA, 1993, p. 11.
7 2 0
D a l t o n T r a n s . , 2 0 0 5 , 7 1 3 – 7 2 0