The first structurally authenticated organomercury(1+) thioether complexes - Mercury-carbon bond activation related to the mechanism of the bacterial enzyme organomercurial lyase

Article abstract of DOI:10.1002/ejic.200300907

The new compounds [MeHg([9]aneS₃)](BF₄) (1), [MeHg([12]aneS₃)](BF₄) (2), and [(MeHg) ₂([14]aneS₄)](BF₄)₂ (3) have been prepared and their crystal structures determined. In 1, the thioether acts as a tridentate ligand [Hg-S 2.611(2)-2.768(2) A] and thus the metal atom is tetrahedrally coordinated, which is rare in organomercury chemistry. Temperature-dependent ¹H and ¹³C NMR spectra showed that this coordination is retained in acetonitrile solution. In crystalline 2 and 3, linear-coordinated Hg^II occurs with Hg-S bond lengths of 2.441(4) and 2.425(2) A. [MeHg([9]aneS₃)]⁺ was found to be stable towards ligand substitution by CF₃SO₃^- in dimethyl sulfoxide, whereas the thioether was partly displaced by CF ₃CO₂^- and completely by CH₃CO ₂^-. Protonolysis by the very strong Bronsted acid CF ₃SO₃H in [D₃]nitromethane transformed the methanido ligand into methane. The degree of Hg-C bond cleavage was ca. 25% for the four-coordinate [MeHg([9]aneS₃)]⁺ after 1 h, whereas no reaction was observed for two-coordinate [MeHg(SEt₂)]⁺ or MeHgCl under similar conditions even after 24 h. The product Hg²⁺ was trapped as [Hg([9]aneS₃)₂]²⁺, which is a six-coordinate complex, as shown by a crystal structure analysis of [Hg([9]aneS₃)₂](BF₄)₂·2 CH₃CN. Quantum chemical calculations [MP2/SDD + SDD ECP Hg, 6-31+G(d,p) other elements] confirmed that hydrogen transfer activation barriers are significantly lower for high -coordinate (CN > 2) complexes. Hg-C bond activation by the enzyme organomercurial lyase is possibly also based on multiple (cysteinyl) sulfur ligation. We propose a hypothetical reaction mechanism that involves an OH-containing amino acid side chain or a water molecule simultaneously serving as a proton acceptor (from Cys-SH) and donor (to R^- of RHg⁺). This mechanism is supported by quantum chemical calculations on a model system. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejic.200300907

FULL PAPER
The First Structurally Authenticated Organomercury(1؉) Thioether Complexes
؊ Mercury؊Carbon Bond Activation Related to the Mechanism of the
Bacterial Enzyme Organomercurial Lyase
Michaela Wilhelm,[a][‡] Stefan Deeken,[a] Eva Berssen,[a] Wolfgang Saak,[a] Arne Lützen,[a]
Rainer Koch,[a] and Henry Strasdeit*[b]
Keywords: Bioinorganic chemistry / Enzyme models / Ligand effects / Mercury / S ligands
The
new
compounds
[MeHg([9]aneS3)](BF4)
(1),
or MeHgCl under similar conditions even after 24 h. The
product Hg2+ was trapped as [Hg([9]aneS3)2]2+, which is a
six-coordinate complex, as shown by a crystal structure ana-
lysis of [Hg([9]aneS3)2](BF4)2·2 CH3CN. Quantum chemical
calculations [MP2/SDD + SDD ECP Hg, 6-31+G(d,p) other
[MeHg([12]aneS3)](BF4) (2), and [(MeHg)2([14]aneS4)](BF4)2
(3) have been prepared and their crystal structures deter-
mined. In 1, the thioether acts as a tridentate ligand [Hg−S
˚
2.611(2)−2.768(2) A] and thus the metal atom is tetrahedrally
coordinated, which is rare in organomercury chemistry. Tem- elements] confirmed that hydrogen transfer activation bar-
perature-dependent 1H and 13C NMR spectra showed that
riers are significantly lower for ‘‘high’’-coordinate (CN Ͼ 2)
this coordination is retained in acetonitrile solution. In crys- complexes. Hg−C bond activation by the enzyme organomer-
talline 2 and 3, linear-coordinated HgII occurs with Hg−S
bond lengths of 2.441(4) and 2.425(2) A. [MeHg([9]aneS3)]
curial lyase is possibly also based on multiple (cysteinyl) sul-
fur ligation. We propose a hypothetical reaction mechanism
that involves an OH-containing amino acid side chain or a
water molecule simultaneously serving as a proton acceptor
(from Cys−SH) and donor (to R− of RHg+). This mechanism is
supported by quantum chemical calculations on a model sys-
tem.
+
˚
was found to be stable towards ligand substitution by
−
CF3SO3 in dimethyl sulfoxide, whereas the thioether was
−
−
partly displaced by CF3CO2 and completely by CH3CO2
.
Protonolysis by the very strong Brønsted acid CF3SO3H in
[D3]nitromethane transformed the methanido ligand into
methane. The degree of Hg−C bond cleavage was ca. 25%
for the four-coordinate [MeHg([9]aneS3)]+ after 1 h, whereas
no reaction was observed for two-coordinate [MeHg(SEt2)]+
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
mercury, methylmercury is partly incorporated by plankton
species and accumulates in the food chain, especially in
predatory fish.[7,8] Methylmercury is known to be an ex-
treme neurotoxin.[9Ϫ16] A recent study estimates that in the
USA alone, over 60,000 children are born each year at risk
of neurological problems caused by in utero exposure to
methylmercury.[9,17]
The high toxicity of various mercury compounds is a
constant issue of environmental and public health concern.
Consequently, the global cycle of mercury, including the
considerable effects of anthropogenic activities, has been in-
tensely studied.[1Ϫ6] It was found that in the marine environ-
ment, the reduction of inorganic mercury (Hg2ϩ) to the el-
emental form (Hg0), and the biomethylation of Hg2ϩ to
methylmercury (MeHgϩ) and dimethylmercury (HgMe2)
are major reactions. While Hg0 is released into the atmos-
phere where it represents the dominant transport form of
One of the main factors responsible for the facile bio-
magnification of methylmercury in the food chain is the
relatively high stability of the HgϪCH3 bond under physio-
logical conditions.[18,19] It is therefore interesting that the
bacterial enzyme organomercurial lyase or MerB (gene:
merB) is capable of accelerating the protonolytic cleavage
of methylmercury and other RHgϩ cations by a factor of
106Ϫ107.[20] The products of the enzymatic reaction are the
hydrocarbon RH and Hg2ϩ which is eventually complexed
by exogenous thiolates. Organomercurial lyase is part of a
broad-spectrum mercury resistance system in which a se-
cond enzyme, termed mercuric reductase or MerA (gene:
merA), catalyzes the reduction of Hg2ϩ to Hg0 with the
latter finally diffusing out of the bacterial cell.[21Ϫ26] The
[a]
Institut für Reine und Angewandte Chemie der Universität
Oldenburg
Carl-von-Ossietzky-Str. 9Ϫ11, 26129 Oldenburg, Germany
Fax: (internat.) ϩ 49-(0)441-798-3329
Institut für Chemie der Universität Hohenheim
[b]
Garbenstr. 30, 70593 Stuttgart, Germany
Fax: (internat.) ϩ 49-(0)711-459-3950
E-mail: h-strasd@uni-hohenheim.de
[‡]
Present address: Universität Bremen FB 4/IW3, Keramische
Werkstoffe und Bauteile
Am Biologischen Garten 2, 28359 Bremen, Germany
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
DOI: 10.1002/ejic.200300907
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2301

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
three-dimensional structure and the enzymatic mechanism
In tetrahydrofuran or acetonitrile/diethyl ether, the thio-
of organomercurial lyase are still unknown. However, a ethers [9]aneS3, [12]aneS3, and [14]aneS4 are capable of dis-
general mechanistic hypothesis proposed by Walsh and co- placing the aqua ligand from [MeHg(OH2)]ϩ {Equation (1);
workers is available.[20,27] One of its key features is the coor- L ϭ [9]aneS3 or [12]aneS3: n ϭ 1; L ϭ [14]aneS4: n ϭ 2}.
dination of the substrate RHgϩ to one or two cysteinyl-
SϪ groups, enhancing the polarity of the HgϪC bond. A
n [MeHg(OH2)](BF4) ϩ L Ǟ [(MeHg)nL](BF4)n ϩ n H2O
(1)
Brønsted-acidic group of the enzyme, possibly a cysteinyl-
SH, is thought to be the proton donor in the subsequent
protonolysis step (for further details see Results and Dis-
cussion). In model studies, Barbaro et al. have demon-
strated that a high ‘‘primary’’ coordination number of mer-
cury can contribute to the activation of HgϪC bonds. They
found that the tetrahedral phosphane complexes
[RHg{N(CH2CH2PPh2)3}]ϩ (R ϭ Me, Ph) were unusually
susceptible to protonolytic HgϪC bond cleavage.[28,29] Re-
sults of a theoretical study, comparing HgϪC bond cleav-
age in [MeHg(PH3)]ϩ and [MeHg(PH3)3]ϩ, agreed with the
experimental observations.[30] Metal ion coordination by
phosphanes, however, is ‘‘non-physiological’’. Synthetic
analogues that more closely mimic the proposed substrate
binding by organomercurial lyase are lacking.
In this paper we describe methylmercury(1ϩ) thioether
complexes with properties relevant to the suggested mecha-
nism of organomercurial lyase. Our experimental and
theoretical study addresses, among other things, the ques-
tion of HgϪC bond activation by multiple sulfur ligation.
Perhaps surprisingly, MeHgϩϪthioether complexes have
low thermodynamic stabilities with typical log K values for
the formation of 1:1 complexes falling in the range Ϫ1 to
2.[31,32] In marked contrast, the stability constants of thi-
oether complexes of Hg2ϩ are by about ten orders of mag-
By this route, we prepared the new coordination com-
pounds [MeHg([9]aneS3)](BF4) (1), [MeHg([12]aneS3)]-
(BF4) (2), and [(MeHg)2([14]aneS4)](BF4)2 (3). They were
isolated as analytically pure, crystalline solids in yields of
85% (1), 68% (2), and 92% (3), respectively. We did not
succeed in preparing complexes with different numbers of
MeHgϩ cations per ligand molecule by variation of the
starting material ratios. NMR experiments showed, how-
ever, that [12]aneS3 can probably bind more than one
MeHgϩ group (see below). Compounds 1Ϫ3 are air-stable
and soluble in strongly polar organic solvents such as aceto-
nitrile, nitromethane, and dimethyl sulfoxide.
Crystal Structures of Compounds 1؊3
Crystals of [MeHg([9]aneS3)](BF4) (1) consist of discrete
Ϫ
nitude larger (log K ϭ 8Ϫ12).[33] This difference may be [MeHg([9]aneS3)]ϩ cations and BF4 anions. The com-
attributed to ‘‘anti-symbiosis’’ which means that the soft pound crystallizes with two formula units in the asymmetric
methanido ligand reduces the affinity of mercury() for an- unit. The two symmetry-independent [MeHg([9]aneS3)]ϩ
other soft ligand in the trans position.[34] So far, only a few complexes have essentially the same structure (Figure 1).
organomercury(1ϩ) thioether complexes have been iso- Each of the Hg atoms is in a distorted tetrahedral coordi-
lated[35] and, to the best of our knowledge, no crystal struc- nation environment formed by a methanido and a triden-
tures have been determined.[36] In the following, we report tate thioether ligand.[29] One HgϪS bond is shorter than
the syntheses and X-ray crystal structures of three new the other two. Nevertheless, all three are ‘‘primary’’ bonds.
MeHgϩ complexes with meso- and macrocyclic thioether The shortest HgϪS bond is involved in the largest
ligands. Part of this work has been communicated pre- SϪHgϪC angle [complex 1: 142.7(4)°; complex 2:
viously.[37]
136.4(4)°], demonstrating a slight distortion towards the lin-
ear coordination usually preferred in RHgϩ complexes. As
expected, the mean HgϪS bond length of the four-coordi-
˚
Results and Discussion
nate complexes in 1 (2.71 A) is considerably larger than the
HgϪS bond lengths of the two-coordinate complexes in 2
[2.441(4) A] and 3 [2.425(2) A]. The HgϪC bond lengths
Syntheses of the Methylmercury Complexes
˚
˚
A suitable methylmercury starting compound was ob- are, however, not significantly different [1: 2.080(8), 2.075(8)
˚
˚
˚
tained by treating methylmercury hydroxide with a slight A; 2: 2.11(2) A; 3: 2.068(7) A] and correspond well with the
˚
excess of tetrafluoroboric acid in aqueous solution. After average value of 2.07 A that was derived from a large num-
removal of the solvent, a colorless oil remained. It is known ber of MeHgϩ compounds.[29b] As in the free form,[40] the
that in strongly acidic aqueous solutions [MeHg(OH2)]ϩ is ligand [9]aneS3 in 1 adopts the [333] conformation.[41,42]
by far the main MeHgϩ species.[38] It therefore seemed Coordination to the MeHgϩ group resulted in a small in-
safe to assume that the obtained product was crease of the mean transannular S···S distance from
[MeHg(OH2)](BF4)·xH2O. The complex [MeHg(OH2)]ϩ 3.451(2) to 3.489 A and a decrease of the mean SϪCϪCϪS
˚
should have a similar structure to [MeHg(DMSO)]ϩ, which torsion angle from 58.5 to 56.0°. The maximum change of
contains O-coordinated dimethyl sulfoxide in solution as an individual torsion angle was Ϫ4.4° at a CϪC bond and
well as in the solid state.[39]
ϩ4.6° at an SϪC bond.
2302
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312

The First Structurally Authenticated Organomercury(1ϩ) Thioether Complexes
FULL PAPER
tacts in 2 i.e. two Hg···S and one Hg···F, are below the lower
ends of these ranges. In 3, two Hg···F contacts are at least
˚
0.10 A shorter than the minimum van der Waals distance.
The respective SϪHg···X and CϪHg···X angles (X ϭ S, F)
fall within a range reasonable for Hg···X bonding interac-
tions [76.6(4)Ϫ101.4(6)°]. Small distortions from linearity
of the SϪHgϪC angles [2: 168.1(6)°; 3: 174.7(2)°] may be
attributed to the influence of the ‘‘secondary’’ bonds. Com-
pared with the free ligand molecules,[45,46] the basic confor-
mations of [12]aneS3 and [14]aneS4 ([3333] and [3434],
respectively)[41] are retained in the MeHgϩ complexes. The
maximum change of a torsion angle between the free and
the coordinated forms amounts to Ϯ6Ϫ7° both for
[12]aneS3 (at SϪC) and [14]aneS4 (at SϪC and CϪC).[47]
[14]aneS4 in 3, as in the free form, possesses a crystallo-
graphically imposed center of symmetry.
Figure 1. Ellipsoid plot (50% probability level) of the two sym-
metry-independent [MeHg([9]aneS3)]ϩ complexes in 1 showing
their mutual orientation in the crystal (hydrogen atoms are omit-
˚
ted); bond lengths [A] and angles [°]: Hg1ϪS1 2.611(2), Hg1ϪS2
2.768(2), Hg1ϪS3 2.760(2), Hg1ϪC1 2.080(8), Hg2ϪS4 2.720(2),
Hg2ϪS5 2.732(2), Hg2ϪS6 2.677(2), Hg2ϪC8 2.075(8);
S1ϪHg1ϪS2 81.1(1), S1ϪHg1ϪS3 79.6(1), S2ϪHg1ϪS3 79.0(1),
S1ϪHg1ϪC1 142.7(4), S2ϪHg1ϪC1 124.9(3), S3ϪHg1ϪC1
127.3(4), S4ϪHg2ϪS5 79.9(1), S4ϪHg2ϪS6 80.4(1), S5ϪHg2ϪS6
80.4(1),
S4ϪHg2ϪC8
132.2(3),
S5ϪHg2ϪC8
126.6(4),
S6ϪHg2ϪC8 136.4(4)
In contrast to the rare tetrahedral coordination observed
in 1, the much more common diagonal coordination of the
Hg atom[29] occurs in [MeHg([12]aneS3)](BF4) (2) and
[(MeHg)2([14]aneS4)](BF4)2 (3). In addition to the HgϪS
and HgϪC ‘‘primary’’ bonds, however, intermolecular ‘‘sec-
ondary’’ bonds to the S and F atoms exist in crystals of 2
and 3 (Figures 2 and 3). The sum of the van der Waals radii
(RvdW) of the participating atoms usually serves as reference
point for establishing the presence of such weak
Figure 2. Ellipsoid plot (50% probability level) of the
[43]
metalϪligand bonds. The RvdW values of HgII (1.75 A),
˚
[MeHg([12]aneS3)]ϩ complex in crystals of 2 (hydrogen atoms are
˚
[44]
˚
˚
omitted); bond lengths [A] and angles [°]: HgϪS1 2.441(4), HgϪC1
S (1.60Ϫ2.03 A), and F (1.30Ϫ1.38 A)
give Hg···S and
2.11(2), Hg···S2I 3.176(4), Hg···S3II 3.256(4), Hg···F1 2.97(1);
S1ϪHgϪC1 168.1(6); symmetry transformations: I: 0.5 ϩ x, 0.5 Ϫ
˚
Hg···F van der Waals distances of 3.35Ϫ3.78 A and
3.05Ϫ3.13 A, respectively. The three intermolecular con- y, 0.5 ϩ z; II: 0.5 ϩ x, 0.5 ϩ y, z
˚
Figure 3. Ellipsoid plot (50% probability level) of the [(MeHg)2([14]aneS4)]2ϩ complex in crystals of 3 (hydrogen atoms are omitted);
bond lengths [A] and angles [°]: HgϪS1 2.425(2), HgϪC1 2.068(7), Hg···F1II 2.947(5), Hg···F4III 2.946(5); S1ϪHgϪC1 174.7(2); symmetry
˚
I
transformations: : Ϫ x, 2 Ϫ y, 2 Ϫ z; II: Ϫ0.5 ϩ x, 1.5 Ϫ y, Ϫ0.5 ϩ z; III: 0.5 Ϫ x, 0.5 ϩ y, 1.5 Ϫ z
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2303

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
compares well with coupling constants found for other two-
coordinate thioether complexes, e.g. [MeHg(SMe2)]ϩ
(220.7 Hz),[35b] [MeHg(methionineH)]2ϩ (223 Hz),[32] and
[MeHg(SEt2)]ϩ (218 Hz; see Exp. Sect.). Surprisingly,
[(MeHg)2([14]aneS4)]2ϩ, which exhibits primary two-coor-
dination in crystals of 3 (see above), has a relatively large
coupling constant (231.1 Hz) which is strongly temperature-
dependent. Raising the temperature from Ϫ30 to ϩ60 °C
resulted in an increase by 9.7 Hz (1: 1.9 Hz; 2: 2.6 Hz). One
possible explanation is that 3 is more strongly dissociated
than, for example, 2. In this case, the contribution of the
dissociation product [MeHg(NCCD3)]ϩ would increase the
observed (average) 2J value. Another explanation is sup-
ported by the observation that the HgϪC bonds are acti-
Properties in Solution and Behavior Towards Competing
Ligands
At room temperature, 1Ϫ3 showed rapid ligand exchange
resulting in averaged NMR signals for coordinated and free
thioether
ligands.
In
suitable
solvents,
the
[MeHg([9]aneS3)]ϩ complex of 1 is rather stable against sol-
volysis. This was demonstrated by an NMR experiment in
which increasing amounts of [MeHg(OH2)](BF4)·xH2O
were added to a 0.06  solution of [9]aneS3 in CD3CN.
The 2J(1H,199Hg) value was found to be independent of the
MeHgϩ/[9]aneS3 molar ratio in the range of 0.5Ϫ0.95:1 and
identical to the coupling constant measured for solutions
of 1 in CD3CN. At ratios Ͼ 1:1, the carbon atoms of the
thioether showed a constant chemical shift identical to that
vated
towards
protonolytic
cleavage
in
2
observed for 1. The two NMR quantities J(1H,199Hg) and
[(MeHg)2([14]aneS4)]2ϩ but not in [MeHg(SEt2)]ϩ (see be-
low). It is conceivable that in solution, three-coordinate iso-
mers of the former complex exist where the [14]aneS4 ligand
forms five-membered HgS2C2 chelate rings. This binding
δC(CH2) differ clearly between 1 (237.4 Hz; 30.4 ppm) and
the ‘‘free’’ components ([MeHg(OH2)]ϩ: ca. 260 Hz;
[9]aneS3: 35.4 ppm). A marked dissociation equilibrium
would be sensitive to an excess of either of the two starting
materials, resulting in non-constant values of the NMR
parameters. The experimental results thus enabled us to
conclude that [MeHg([9]aneS3)]ϩ does not strongly dis-
sociate in CD3CN.
situation
has
a
precedent
in
the
complex
[(HgCl2)2([14]aneS4)], in which the thioether is simul-
taneously coordinated to two HgCl2 moieties.[51] Higher
temperatures may favor the accompanying ligand confor-
mations which are probably energetically less favored than
the one observed in crystalline 3.
An analogous experiment with [12]aneS3 showed that the
[MeHg([12]aneS3)]ϩ complex in 2 is markedly dissociated
in CD3CN. At a MeHgϩ/[12]aneS3 molar ratio of 0.5:1 the
2J(1H,199Hg) coupling constant was 221.9 Hz (2: 224.2 Hz),
indicating a more complete complexation of methylmercury
due to the excess of thioether. When the ratio was increased
from 1:1 to 4.5:1, a continuous increase in the 2J(1H,199Hg)
value resulted. However, this increase was significantly
smaller than expected from the contribution of
[MeHg(OH2)]ϩ or [MeHg(NCCD3)]ϩ. This discrepancy
can be explained by the binding of more than one methyl-
mercury group per thioether molecule. From the δC(MeHg)
values {8.9 ppm at MeHgϩ/[12]aneS3 ϭ 1:1; 4.5 ppm at
4.5:1; [MeHg(OH2)]ϩ: ca. 0.0 ppm} one can easily estimate
that at the highest MeHgϩ/[12]aneS3 ratio (4.5:1) each thio-
ether ligand binds, on average, slightly more than two
MeHgϩ cations. Consistent with this picture, the δC(SCH2)
value of the ligand continuously increased from 30.3 ppm
(at 1:1) to 32.6 ppm (at 4.5:1). The value for free [12]aneS3
is 28.9 ppm.
1
The H NMR spectrum of 1 at Ϫ30 °C shows that the
tetrahedral structure observed in the solid state is retained
in solution. The signal of the CH2 groups forms a charac-
teristic AAЈBBЈ (or AAЈXXЈ) pattern[52] in accordance with
the presence of three five-membered chelate rings which
rapidly alternate between the λ and δ conformations
(Scheme 1).
Scheme 1
Upon raising the temperature, this pattern collapses into
a broad and finally, at 60 °C, a sharp singlet, obviously
because of fast exchange of the thioether ligand. Between
Some years ago, Rabenstein suggested that the Ϫ30 °C and room temperature, 2J(1H,199Hg) (see above)
2J(1H,199Hg) coupling constant of the methylmercury group and the coordination-induced chemical shifts of 1 are prac-
may be utilized as a general indicator of the presence of tically temperature-independent. For example, the relative
chelating ligands.[48] Indeed, it has been reported that com- shift of the CH2 groups ∆δC ϭ δC(coordinated) Ϫ δC(free
plexes of potentially chelating ligands of the 2,2Ј-bipyridyl ligand) has a constant value of Ϫ5.0 ppm in CD3CN. It can
and 1,10-phenanthroline type have higher values therefore be concluded that the tetrahedral structure is also
1
(235.1Ϫ239.8 Hz) than those of monodentate pyridine de- present at room temperature. In the room temperature H
rivatives
(225.2Ϫ229.6 Hz).[49]
A
comparison
of and 13C NMR spectra of 2 and 3, the number of observed
[MeHg{N(CH2CH2PPh2)3}]ϩ (174 Hz, CN 4)[28] with resonances is less than would be expected from the solid-
[MeHg(PR3)]ϩ complexes (167Ϫ173 Hz, CN 2)[50] shows state structures. This can be explained either by fast thio-
the same trend, albeit less pronounced. Accordingly, we ob- ether ligand exchange as in the case of 1, fast intramolecu-
served a higher 2J value for the tetrahedral complex lar migration of the MeHgϩ group(s) at the macrocycle via
[MeHg([9]aneS3)]ϩ (237.4 Hz) than for the two-coordinate intermediate chelates, or both. At Ϫ30 °C, the number and
[MeHg([12]aneS3)]ϩ species (224.2 Hz). The latter value multiplicity of the signals remain unchanged. Thus, at this
2304
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312

The First Structurally Authenticated Organomercury(1ϩ) Thioether Complexes
FULL PAPER
temperature, at least one of these processes cannot be ‘‘froz- anion of cysteine or glutathione (RЈSH), thereby binding
en’’ as opposed to the ligand exchange in 1.
the RHgϩ cation at the active site. There the key step fol-
The relative stability of [MeHg([9]aneS3)]ϩ towards a lows, namely the protonolytic HgϪC bond cleavage, in
number of neutral or monoanionic potential ligands, L(Ϫ)
which a second SH group of the enzyme is involved (step
,
was studied by NMR spectroscopy. In this study, fast ligand 2). In the last two steps, exogenous thiolates coordinate to
exchange at room temperature proved beneficial because it the product Hg2ϩ and finally remove it from the enzyme.
resulted in only one 13C NMR signal for mixtures of coor- Pitts and Summers proposed that Cys159[54] could be the
dinated and released [9]aneS3. The position of this signal primary ligand for the substrate RHgϩ (step 1), while Cys96
between the two extremes, δC(coordinated) ϭ 29.5 ppm and donates the proton in step 2.[53] They also pointed out that
δC(free ligand) ϭ 33.9 ppm, gave a qualitative indication of further amino acid residues, for example Tyr93 and Cys117,
the position of equilibrium [Equation (2)].
are possible candidates as ligands and/or proton donors.
The latter is also one of three highly conserved cysteines,
besides Cys96 and Cys159. However, there are indications
that Cys117 may have no catalytic function.
[MeHg([9]aneS3)]ϩ ϩ L(Ϫ) Ǟǟ [MeHgL](ϩ) ϩ [9]aneS3
(2)
From the viewpoint of coordination chemistry, step 2 is
the most intriguing one. It raises the question of how the
HgϪC bond is activated for protonolysis. From model stud-
ies on phosphane[28] and thioether complexes (this work), it
seems less likely that sufficient activation of the HgϪC
bond is possible in the binding situation depicted in
Scheme 3 (A). Here, Cys96 can only act as a weak thiol
donor and has a spatial orientation typical of secondary
ligands. Scheme 3 (B) shows the original proposal[20,27] ex-
tended by recent biochemical results.[53] The mercury atom
is three-coordinate, and therefore an enhanced polarity and
a significant activation of the HgϪC bond may be expected.
In Scheme 3 (C) we suggest a model that carries features of
both previous ones and takes into account the occurrence
of several conserved OH-containing amino acid residues in
the enzymes. In this model, activation and protonation de-
pend on the participation of YϪOH, which represents an
alcoholic or a phenolic amino acid residue. Alternatively
YϪOH could be a water molecule, which may be involved
in hydrogen bonding to an OH-containing side chain. The
oxygen atom of YϪOH simultaneously serves as proton ac-
ceptor (from Cys96ϪSH) and donor (to RϪ). Proton with-
drawal from Cys96ϪSH imposes thiolate character on the
sulfur atom. This, together with the overall spatial arrange-
ment, would allow the formation of a second primary
HgϪS bond. Candidates for YϪOH include the conserved
residues Thr77, Thr81, Tyr93, Ser115, and Thr120. In ad-
dition, the ligand AAϪXϪ (shown in Scheme 3, B) could
also be present here. The model in Scheme 3 (C) is purely
hypothetical but it is supported by results of quantum
chemical calculations, which are presented below. It might
be tested experimentally on suitable site-directed mutants.
The δC values are given for solutions in [D6]DMSO,
which was used for solubility reasons. The solutions were
initially 50 m in each of the two components 1 and L(Ϫ)
.
LϪ anions were introduced as their Et4Nϩ salts. It appeared
that the position of the equilibrium according to Equation
(2) was far to the left side when LϪ was an oxo anion with
Ϫ
a
largely delocalized charge such as CF3SO3 , p-
CH3C6H4SO3Ϫ or NO3Ϫ. With L(Ϫ) ϭ CH3CO2Ϫ, FϪ, ClϪ,
BrϪ, IϪ or PPh3, however, complete substitution of the
thioether ligand was observed. An intermediate situation
occurred with L(Ϫ) ϭ CF3CO2Ϫ or NEt3 where the equilib-
rium concentrations of [MeHg([9]aneS3)]ϩ and [MeHgL](ϩ)
were of the same order of magnitude. As expected, the
monodentate ligand [12]aneS3 cannot displace the triden-
tate [9]aneS3 (experiment conducted in CD3CN). In sum-
mary, it can be concluded that [9]aneS3 is an unusually
strong ligand for MeHgϩ when compared with other thio-
ethers, but still a rather weak ligand on an absolute scale,
comparable for example with trifluoroacetate.
Hypothetical Reaction Mechanism of Organomercurial
Lyase
The bacterial enzyme organomercurial lyase catalyzes the
simple reaction shown in Equation (3) (R ϭ alkyl, aryl).
A model for the reaction mechanism of this enzyme was
proposed some years ago[20,27] (see Introduction) and has
very recently been refined.[53] It consists of at least four
steps (Scheme 2).
RHgϩ ϩ Hϩ Ǟ Hg2ϩ ϩ RH
(3)
Protonolysis of Hg؊C Bonds in Methylmercury Thioether
Complexes
The four-coordinate complex [MeHg([9]aneS3)]ϩ offered
the opportunity of studying, for the first time, the effect of
multiple sulfur ligation on the reactivity of the HgϪC bond.
The very strong Brønsted acid CF3SO3H in [D3]nitrome-
thane was used as proton donor.[55] An additional equiva-
lent of [9]aneS3 was employed in order to trap the reaction
product Hg2ϩ as [Hg([9]aneS3)2]2ϩ. After the protonolysis
Scheme 2. Reaction steps in the mechanism of organomercurial
lyase (Enz ϭ enzyme)
In the first step, a cysteinyl-SH group of the enzyme pro- step, the thioether is thus functionally similar to the exog-
tonates and displaces an exogenous thiolate, such as the enous thiols in the mechanism of organomercurial lyase
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2305

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
Scheme 3. Proposals for the key step of the reaction mechanism
of organomercurial lyase; A: with only Cys96 and Cys159 being
involved; B: with a further primary ligand AAϪXϪ, for example
Tyr93 (X ϭ O) or Cys117 (X ϭ S); C: with participation of an
additional amino acid (or water) YϪOH as a simultaneous proton
donor and acceptor
(steps 3 and 4). The overall reaction can be expressed by
Equation (4).
[MeHg([9]aneS3)]ϩ ϩ CF3SO3H ϩ [9]aneS3 Ǟ
(4)
[Hg([9]aneS3)2]2ϩ ϩ CF3SO3Ϫ ϩ CH4
Figure 4. Ellipsoid plot (50% probability level) of the
[Hg([9]aneS3)2]2ϩ complex in crystals of 4 (hydrogen atoms are
˚
omitted); bond lengths [A] and angles [°]: HgϪS1 2.677(1), HgϪS2
2.660(1), HgϪS3 2.711(1); S1ϪHgϪS2 83.69(3), S1ϪHgϪS3
82.49(3), S1ϪHgϪS1I 180.0, S1ϪHgϪS2I 96.31(3), S1ϪHgϪS3I
97.51(3), S2ϪHgϪS3 83.14(3), S2ϪHgϪS2I 180.0, S2ϪHgϪS3I
Methane was identified as one of the reaction products
by using deuterated acid. In the CF3SO3D-containing sys-
tem, a mixture of CH3D and CH4 formed due to the pres-
ence of traces of H2O (see below). For CH3D we measured
I
96.86(3), S3ϪHgϪS3I 180.0; symmetry transformation: : Ϫ x, 1
Ϫ y, Ϫz
2J(H,D)
ϭ 1.9 Hz and a deuterium isotope effect
activities of different MeHgϩ complexes. We found for ex-
ample, that under the conditions described in the Exp. Sect.,
the degree of protonolysis reached ca. 25% for both 1 and
3 after 1 h. In contrast, [MeHg(SEt2)]ϩ and MeHgCl
showed no detectable reaction even after 24 h. MeHgCl is
known to be resistant towards strong acids; concentrated
hydrochloric acid, for example, causes less than 1% pro-
tonolysis after 100 min.[58] This is despite the fact that chlo-
ride forms a stronger 1:1 MeHgϩ complex than [9]aneS3,
as our competition experiments have shown (see above). It
therefore seems very likely that the higher coordination
number in [MeHg([9]aneS3)]ϩ compared with MeHgCl is
the most crucial factor in HgϪC bond activation. The large
δH(CH3D) Ϫ δH(CH4) ϭ Ϫ15.3 ppb in [D3]nitromethane.
These results are in good agreement with very accurately
measured values in different solvents, for example in [D6]-
acetone: 1.9361 Ϯ 0.0008 Hz and Ϫ15.555 Ϯ 0.002 ppb.[56]
The reaction product [Hg([9]aneS3)2]2ϩ could be isolated as
crystalline [Hg([9]aneS3)2](BF4)2·2 CH3CN (4) when the re-
action was carried out in acetonitrile with HBF4 instead of
CF3SO3H [Equation (5)].
(5)
difference in reactivity between [(MeHg)2([14]aneS4)]2ϩ
,
Compound 4 was obtained in 72% yield. Its identity was
confirmed by comparison with a sample that had been pre-
pared independently from HgPh2, HBF4, and [9]aneS3.
Compound 4 was characterized, inter alia, by an X-ray
structural analysis. In the crystals, the [Hg([9]aneS3)2]2ϩ cat-
ion is located on a crystallographic center of inversion. The
six sulfur atoms form a distorted octahedral coordination
environment around the metal atom (see Figure 4). The mo-
lecular structure will not be further discussed here, since it
is similar to those of other [Hg([9]aneS3)2]2ϩ salts whose
structures have already been described in the literature.[57]
which is two-coordinate in crystals of 3, and [MeHg(SEt2)]ϩ
appears to contradict this conjecture. However, the
2J(1H,199Hg) value indicates that in solutions of 3, the coor-
dination number of mercury may exceed 2, as already dis-
cussed. In order to better understand the role of the coordi-
nation number, quantum chemical calculations were car-
ried out.
Results of Quantum Chemical Calculations
In the following, structural properties[59] such as metrical
The progress of the reaction according to Equation (4) data as well as NMR chemical shifts and mechanistic path-
was monitored by NMR spectroscopy, but attempts to de- ways for HgϪC bond activation[60] have been examined. In
termine kinetic data were unsuccessful because the meas- order to obtain reliable data, for example on coordination-
urements were impaired by unavoidable traces of water (ca. induced chemical shifts, it is of the utmost importance to
0.3 equiv.). The effect of water could be demonstrated by find a method/basis set combination that allows the accu-
intentionally adding 1 equiv., which almost entirely sup- rate description of the molecules studied herein. In a first
pressed the protonolysis reaction. However, as the water attempt, we therefore employed several ab initio and density
content was the same in all experiments, it was possible to functional theoretical approaches to reproduce the struc-
obtain semi-quantitative values to compare the relative re- ture of the [MeHg([9]aneS3)]ϩ cation of 1, the main focus
2306
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312

The First Structurally Authenticated Organomercury(1ϩ) Thioether Complexes
FULL PAPER
ϩ
˚
Table 1. MetalϪligand bond lengths [A] in the complex [MeHg([9]aneS3)] , determined at different levels of theory, compared with the
experimental values
Method
HgϪC
HgϪS
X-ray diffraction
AM1
PM3
2.075(8), 2.080(8)
2.057
2.076
2.611(2)Ϫ2.768(2) (mean: 2.711)
2.738Ϫ2.752
2.590
B3LYP/LANL2DZ
2.237
2.143
2.138
2.134
2.137
2.136
2.110
2.111
2.986Ϫ2.997
2.886Ϫ2.912
2.856Ϫ2.864
2.839Ϫ2.844
2.858Ϫ2.870
2.856Ϫ2.866
2.771Ϫ2.776
2.768Ϫ2.774
B3LYP/SDD ϩ SDD ECP Hg
B3LYP/SDD ϩ SDD ECP Hg, 6-31G(d) C,H,S
BP86/SDD ϩ SDD ECP Hg, 6-31G(d) C,H,S
B3LYP/SDD ϩ SDD ECP Hg, SDD ϩ 6-31G(d) C,H,S
B3LYP/SDD ϩ SDD ECP Hg, SDD ϩ 6-31ϩG(d,p) C,H,S
MP2/SDD ϩ SDD ECP Hg, 6-31G(d) C,H,S
[a]
MP2/SDD ϩ SDD ECP Hg, 6-31ϩG(d,p) C,H,S
[a]
In the discussion this level of theory is denoted MP2/I.
being on the HgϪC and HgϪS bond lengths. Table 1 sum- has been found experimentally at 0.7 ppm, was also satis-
marizes the results. In each case, the basic structural feature factorily reproduced by theory (at δ ϭ Ϫ3.1 ppm).
of the complex, i.e. the tetrahedral coordination of mercury,
The protonolytic cleavage of HgϪCH3 bonds by tri-
is reproduced. The necessity of including a large basis set fluoromethanesulfonic acid has been investigated at several
as well as an effective core potential can be clearly seen levels of theory (for pertinent experimental results see
from the DFT calculations.[61] However, the HgϪS bonds above). This includes the transition state for the proton
were still much too long. Only the treatment at the corre- transfer from CF3SO3H to methylmercury complexes of the
or
0
lated level[62] gave good agreement for these distances. type [MeHgL]ϩ
Therefore the last method from Table 1 was used for the
subsequent NMR calculation. Interestingly, the bond
lengths were also well reproduced by the AM1 method.
However, despite the good semiempirical description of the
experimental structure, all evaluated ab initio calculations
of NMR chemical shifts based on both AM1 and PM3 geo-
metries failed to give satisfactory results.
leading to methane and, initially,
As mentioned before, a reliable structure is indispensable
for the evaluation of NMR chemical shifts in general,[63]
and for organomercury compounds in particular.[64] The
approach used herein applied the MP2-GIAO method[65]
together with the SDD basis set and an ECP on mercury,[66]
and the 6-31ϩG(d,p) basis set[67] on the remaining atoms
based
on
the
MP2-optimized
geometries
of
[MeHg([9]aneS3)]ϩ and [9]aneS3. Experimentally, a coordi-
nation-induced shift of Ϫ5.0 ppm (from 35.4 to 30.4 ppm)
was observed for the CH2 carbon atoms of the thioether
ligand (see above). Such a high-field shift is perhaps coun-
ter-intuitive. However, it was reproduced quite accurately
by our theoretical model: the averaged 13C NMR shift
changed by Ϫ7.4 ppm (from 38.9 to 31.5 ppm) when
[9]aneS3 underwent coordination. The 13C NMR signal po-
sition of the methanido ligand of [MeHg([9]aneS3)]ϩ, which
Scheme 4. Pathway of the reaction between [MeHgL]ϩ or 0 and tri-
Figure 5. MP2/I-calculated transition states for the protonolytic
fluoromethanesulfonic acid (L ϭ ClϪ, SEt2, MeSCH2CH2SMe, or HgϪC bond cleavage in MeHg(SMe) by methanethiol with (TS2),
[9]aneS3)
and without (TS1) assistance by methanol
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2307

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
[(CF3SO3)HgL]ϩ or 0 (Scheme 4). The latter can further re- plays a key role in the mechanism of organomercurial lyase.
act with another L to yield [HgL2]2ϩ or 0
.
Assistance by an OH group, for example from an amino
The results of the MP2/I calculations (see Table 1 for acid side chain, could further facilitate the protonolytic
basis set explanation) show that the hydrogen transfer acti- HgϪC bond cleavage in the enzyme’s active site.
vation barriers for the two complexes with digonal mercury
coordination are similar, namely 23 kcal·molϪ1 for MeHgCl
and 20 kcal·molϪ1 for [MeHg(SEt2)]ϩ. Significantly lower
Experimental Section
activation barriers were found for the complexes with
higher coordination numbers of mercury: 14 kcal·molϪ1 for
Safety Precautions: Soluble mercury() compounds are highly
[MeHg(MeSCH2CH2SMe)]ϩ (CN 3) and 12 kcal·molϪ1 for
toxic. In addition to the precautions routinely taken when toxic
[MeHg([9]aneS3)]ϩ (CN 4). This confirms the above de-
metal species are handled, further protection measures are neces-
sary when working with methylmercury (MeHgϩ).[10,11,68] In par-
scribed experimental finding that multiple sulfur ligation
activates HgϪCH3 bonds towards protonolytic cleavage.
The possible relevance for the enzymatic reaction mecha-
nism is obvious.
ticular, laminate gloves that are specially designed for chemical re-
sistance (Silver Shield) should be worn under neoprene or nitrile
gloves. Disposable latex gloves are not sufficient! Medical surveil-
lance of the mercury concentration in the blood should be con-
Finally, we sought an answer for the possible involvement
sidered for those repeatedly working with methylmercury com-
of OH-containing amino acid residues in the enzymatic re-
action mechanism. The model investigated consists of
methylmercury methanethiolate i.e. MeHg(SMe), to which
a proton is transferred from methanethiol, either directly or
by assistance from methanol. Our model corresponds to the
situations depicted in Scheme 3 (A and C). Optimization of
the two proton transfer transition states at the MP2/I level
of theory (Figure 5) demonstrates a clear preference of
14 kcal·molϪ1 for the six-membered transition structure
TS2 which involves the additional hydroxy group. The
activation energies for the two pathways are 39 and 25
kcal·molϪ1, respectively. Although these values are still high
for an enzymatic reaction this model shows, that the in-
volvement of OH groups can significantly lower the acti-
vation barrier for protonolytic HgϪC bond cleavage, and
should therefore be considered when discussing possible re-
action mechanisms of organomercurial lyase.
pounds. Methylmercury-containing wastes can be treated with
strong oxidants, for example aqua regia, to give inorganic mercury
(Hg2ϩ). Using the ‘‘supertoxic’’ compound dimethylmercury as an
Hg NMR standard is not to be recommended (see below).
Instrumentation: 1H and 13C{1H} NMR spectra were recorded at
300 K with a Bruker DPX 300 or a DRX 500 instrument. Chemical
shifts are given relative to TMS (δ ϭ 0 ppm). 199Hg{1H} NMR
spectra were recorded with the DRX 500 spectrometer at 290 K. A
0.10  solution of Hg(ClO4)2 in 0.10  perchloric acid in D2O
served as an external standard.[69,70] The standard was prepared as
follows: in a 1-mL volumetric flask, D2O was added to yellow HgO
(21.7 mg, 0.100 mmol) and DClO4 (26.4 µL of a 68% solution in
D2O, ρ ϭ 1.694 g·cmϪ3, 0.300 mmol); after ca. 30 min, the HgO
had completely dissolved, and the flask was filled to the mark.
199Hg chemical shifts were referenced to neat HgMe2 (δHg
ϭ
0 ppm) with this solution (δHg ϭ Ϫ2250 ppm on the HgMe2 scale).
IR spectra (KBr pellets) were obtained with a Bio-Rad FTS 7PC
spectrometer. Mass spectra were measured with a Finnigan MAT
212 instrument. Melting points were determined in unsealed glass
capillaries. Elemental analyses were performed by the Mikroanaly-
tisches Labor Pascher, Remagen (Germany).
Conclusions
Starting Materials: The ligand 1,5,9-trithiacyclododecane
([12]aneS3) was prepared according to a literature method starting
from bis(3-hydroxypropyl) sulfide.[45] The other starting materials,
including methylmercury hydroxide (1  aqueous solution, Alfa
Aesar), were purchased from commercial sources and used as re-
ceived. The solvents were of reagent grade.
The following are the principal results and conclusions
of this work: (i) Complexes between the methylmercury(1ϩ)
cation and the cyclic thioethers [9]aneS3, [12]aneS3, and
[14]aneS4 are isolable as crystalline tetrafluoroborates
[MeHg([9]aneS3)](BF4) (1), [MeHg([12]aneS3)](BF4) (2),
and [(MeHg)2([14]aneS4)](BF4)2 (3), respectively, in good
yields. In 2 and 3, HgII has an unexceptional linear coordi-
nation. In 1, however, the thioether acts as a tridentate li-
gand. The resultant tetrahedral metal coordination is rarely
found in organomercury compounds. (ii) The unusual coor-
Preparation of [MeHg([9]aneS3)](BF4) (1): HBF4 (150 µL of a 50%
aqueous solution, ρ ϭ 1.38 g·cmϪ3, 1.2 mmol) was added to
MeHgOH (1.00 mL of a 1  aqueous solution, 1.0 mmol). After
12 h, the solution was concentrated in vacuo to give a colorless
oil, which was redissolved in tetrahydrofuran (2 mL). A solution of
dination has a distinct influence on the reactivity of [9]aneS3 (180 mg, 1.00 mmol) in tetrahydrofuran (5 mL) was added
[MeHg([9]aneS3)]ϩ. This complex is significantly more
dropwise whilst stirring. A white, microcrystalline solid formed im-
mediately. After stirring for 12 h, the reaction was complete, and
stable to substitution of its sulfur ligand than other
MeHgϩϪthioether complexes. Besides, its HgϪC bond is
much more susceptible to cleavage by a strong Brønsted
the solid was collected on a glass filter, washed with tetrahydrofu-
ran and dried under vacuum. Yield: 412 mg (85%). M.p. ca. 190
1
°C (dec.). H NMR (300 MHz, CD3CN, 40 m): δ ϭ 1.09 [s with
acid than the HgϪC bonds in linear coordinated
satellites, 2J(1H,199Hg) ϭ 237.4 Hz, 3 H; HgCH3], 3.09 (br. s, 12 H,
[MeHg(SEt2)]ϩ or MeHgCl. Quantum chemical calcu-
lations confirm that the coordination number is an import-
ant factor in HgϪC bond activation. (iii) Our experimental
and theoretical findings strengthen the hypothesis that mul-
CH2) ppm. 13C NMR (75.5 MHz, CD3CN, 40 m): δ ϭ 0.7
(HgCH3, superposed by solvent signal[71]), 30.4 (CH2) ppm. 199Hg
NMR (89.6 MHz, CD3CN, 40 m): δ ϭ Ϫ194 ppm. IR: ν˜ ϭ 2920
(w), 1449 (m), 1410 (m), 1287 (m), 1161 (w), 1144 (w), 1047 (s, br),
tiple sulfur coordination, in this case by cysteinyl S atoms, 1036 (s, br), 926 (m), 889 (m), 816 (m), 783 (m), 521 cmϪ1 (m). MS
2308
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312

The First Structurally Authenticated Organomercury(1ϩ) Thioether Complexes
FULL PAPER
(CI, isobutane): m/z (%) ϭ 397 (2) [MeHg([9]aneS3)]ϩ, 181 (100)
[[9]aneS3 ϩ H]ϩ. C7H15BF4HgS3 (482.8): calcd. C 17.41, H 3.13,
Hg 41.55, S 19.93; found C 17.35, H 3.03, Hg 41.2, S 19.8.
24 H, SCH2) ppm. 13C NMR (75.5 MHz, CD3CN): δ ϭ 28.3
(SCH2) ppm. 199Hg NMR (89.6 MHz, CD3CN, 40 m): δ ϭ
Ϫ305 ppm. IR: ν˜ ϭ 2911 (w), 2899 (w), 2249 (w), 1443 (m), 1408
(m), 1300 (sh), 1285 (m), 1059 (s, br), 1036 (s, br), 920 (m), 883
(m), 822 (m), 534 (m), 521 (m), 424 cmϪ1 (m). MS (CI, isobutane):
m/z (%) ϭ 181 (100) [[9]aneS3 ϩ H]ϩ, 121 (97) [C4H8S2 (dithiane)
ϩ H]ϩ. C16H30B2F8HgN2S6 (817.0): calcd. C 23.52, H 3.70, Hg
24.55, N 3.43, S 23.55; found C 23.58, H 3.57, Hg 25.0, N 3.15, S
23.9. Method B: HBF4 (136 µL of a 54% solution in diethyl ether,
ρ ϭ 1.19 g·cmϪ3, 1.0 mmol) was added whilst stirring to a solution
of 1 (96 mg, 0.20 mmol) and [9]aneS3 (36 mg, 0.20 mmol) in aceto-
nitrile (5 mL). After stirring for ca. 30 min, the reaction mixture
was stored at ca. 0 °C. After 24 h, colorless crystals of 4 had
formed, which were collected on a glass filter, washed with a small
amount of an ice-cold acetonitrile/diethyl ether mixture (1:1, v/v)
Preparation of [MeHg([12]aneS3)](BF4) (2): HBF4 (75 µL of a 50%
aqueous solution, ρ ϭ 1.38 g·cmϪ3, 0.6 mmol) was added to
MeHgOH (0.50 mL of a 1  aqueous solution, 0.5 mmol). After
12 h, the solution was concentrated in vacuo to give a colorless oil,
to which [12]aneS3 (111 mg, 0.50 mmol), dissolved in acetonitrile
(5 mL), was added. The reaction flask was connected to another
flask that contained diethyl ether. During the next 7 d the diethyl
ether was allowed to diffuse into the reaction mixture through the
gas phase. A colorless, crystalline solid formed, which was collected
on a glass filter, washed with tetrahydrofuran and dried under vac-
uum. Yield: 179 mg (68%). M.p. ca. 166 °C (dec.). 1H NMR
(300 MHz, CD3CN): δ ϭ 1.11 [s with satellites, 2J(1H,199Hg) ϭ
1
and briefly dried under vacuum. Yield: 117 mg (72%). The H and
3
224.2 Hz, 3 H, HgCH3], 1.99 (quint, JH,H ϭ 6.6 Hz, 6 H,
13C NMR spectra and the IR spectrum were identical to those of
the material prepared by Method A.
3
CH2CH2CH2), 2.93 (t, JH,H ϭ 6.6 Hz, 12 H, SCH2) ppm. 13C
NMR (75.5 MHz, CD3CN):
δ
ϭ
8.9 (HgCH3), 27.0
(CH2CH2CH2), 30.3 (SCH2) ppm. 199Hg NMR (89.6 MHz,
CD3CN, 40 m): δ ϭ Ϫ886 ppm. IR: ν˜ ϭ 2926 (w), 1449 (m), 1431
(m), 1416 (m), 1300 (w), 1260 (m), 1250 (m), 1177 (w), 1049 (s, br),
1030 (s, br), 793 (m), 756 (m), 519 cmϪ1 (m). MS (CI, isobutane):
m/z (%) ϭ 439 (4) [MeHg([12]aneS3)]ϩ, 223 (100) [[12]aneS3 ϩ H]ϩ.
C10H21BF4HgS3 (524.9): calcd. C 22.88, H 4.03, Hg 38.22, S 18.33;
found C 22.88, H 3.90, Hg 38.4, S 18.3.
Estimation of the Rates of Protonolysis: Solutions for the NMR
measurements were prepared in a stream of dry nitrogen as a pro-
tection against atmospheric moisture. A typical experiment was
conducted as follows. 1 (48.3 mg, 100 µmol) and [9]aneS3 (18.0 mg,
100 µmol) were weighed out in a 1-mL volumetric flask, which had
been flushed with dry nitrogen. Hexamethylbenzene (300 µL of a
50 m stock solution in [D3]nitromethane, 15.0 µmol) was then ad-
ded, which served as a reference standard for determining the con-
centration change of methylmercury. After addition of [D3]nitro-
methane (ca. 0.3 mL), the flask was shaken until a clear solution
was obtained. Trifluoromethanesulfonic acid (98 µL of a 1.02 
stock solution in [D3]nitromethane, 100 µmol) was then added.
After briefly shaking, the flask was filled to the mark with [D3]ni-
tromethane. After shaking again, most of the solution was immedi-
ately transferred to an NMR tube, which had been flushed with
dry nitrogen. The tube was sealed with a standard cap and para-
Preparation of [(MeHg)2([14]aneS4)](BF4)2 (3): HBF4 (150 µL of a
50% aqueous solution, ρ ϭ 1.38 g·cmϪ3, 1.2 mmol) was added to
MeHgOH (1.00 mL of a 1  aqueous solution, 1.0 mmol). After
12 h, the solution was concentrated in vacuo to give a colorless oil,
to which [14]aneS4 (134 mg, 0.50 mmol), dissolved in acetonitrile
(5 mL), was added. The reaction flask was connected to another
flask that contained diethyl ether. During the next 4 d the diethyl
ether was allowed to diffuse into the reaction mixture through the
gas phase. Small, plate-like crystals formed, which were collected
on a glass filter, washed with a small amount of tetrahydrofuran
and dried under vacuum. Yield: 403 mg (92%). M.p. ca. 129 °C
1
film. The initial H NMR spectrum was measured typically 8 min
after addition of the acid. Subsequent spectra were recorded every
3 min during the first 30 min, and later at increasingly longer time
intervals. The temperature was kept at 20 °C. NMR spectroscopic
(dec.). 1H NMR (500 MHz, CD3CN): δ ϭ 1.16 [s with satellites,
3
2J(1H,199Hg) ϭ 231.1 Hz, 6 H, HgCH3], 2.10 (quint, JH,H
ϭ
1
data were collected over a period of ca. 6 h. The H NMR spectra
3
7.3 Hz, 4 H, CH2CH2CH2), 3.05 (t, JH,H ϭ 7.3 Hz, 8 H,
were analyzed by measuring the change of the integral intensity
of the MeHgϩ signal (at δ ϭ 1.10 ppm) relative to the signal of
hexamethylbenzene (at δ ϭ 2.19 ppm). For 3, MeHgCl, and
[MeHg(SEt2)]ϩ the sample preparations and measuring conditions
were similar to those described above. However, the concentrations
were adjusted less rigorously, and much longer time intervals be-
tween successive measurements were chosen because the main pur-
pose was to establish whether the reaction had occurred at all. Be-
tween measurements, the NMR samples were stored at 20 °C. In
the case of 3 (0.04  solution), two moles of trifluoromethanesul-
fonic acid per mole of complex were applied and no hexameth-
CH2CH2CH2), 3.26 (s,
8
H, SCH2CH2S) ppm. 13C NMR
(125.8 MHz, CD3CN): δ ϭ 7.6 (HgCH3), 30.2 (CH2CH2CH2), 32.7
(CH2CH2CH2), 33.3 (SCH2CH2S) ppm. IR: ν˜ ϭ 2942 (w, br), 1445
(m), 1287 (w), 1215 (w), 1182 (m), 1084 (s, br), 1051 (s, br), 1036
(s, br), 789 (m), 521 cmϪ1 (m). MS (CI, isobutane): m/z (%) ϭ 485
(2) [MeHg([14]aneS4)]ϩ, 351 (6) [MeHg([7]aneS2)]ϩ,[72] 107 (100)
[C3H6S2 (dithiolane) ϩ H]ϩ. C12H26B2F8Hg2S4 (873.4): calcd. C
16.50, H 3.00, Hg 45.93, S 14.69; found C 16.75, H 3.00, Hg 45.7,
S 14.7.
Preparation of [Hg([9]aneS3)2](BF4)2·2 CH3CN (4). Method A: Di-
phenylmercury (177 mg, 0.50 mmol) and [9]aneS3 (180 mg, ylbenzene was added. Instead, the intensity of the CH2CH2CH2
1.00 mmol) were dissolved in tetrahydrofuran (10 mL). Addition of ligand signal served as a reference. It occurred at δ ഠ 2.45 ppm
HBF4 (152 µL of a 54% solution in diethyl ether, ρ ϭ 1.19 g·cmϪ3
,
and, later, split to give an additional signal at δ ϭ 2.36 ppm, which
was assigned to the product complex(es) that formed as a result of
the protonolysis. The MeHgCl solution was 0.07 , and contained
1.1 mmol) resulted in the immediate precipitation of a white, mi-
crocrystalline solid. After 6 h, the solid was collected on a glass
filter, washed with tetrahydrofuran and redissolved in acetonitrile
(20 mL) in a flask, which then was connected to another flask con-
taining diethyl ether. During the next 5 d the diethyl ether was
allowed to diffuse into the solution through the gas phase. Color-
a
known concentration (0.02 ) of hexamethylbenzene.
[MeHg(SEt2)]ϩ was prepared in situ. To this end, aqueous solutions
of MeHgOH (0.30 mmol) and HBF4 (0.36 mol) were mixed. After
1 h, the reaction solvents were evaporated in vacuo. The residue
less, elongated tabular crystals formed, which were collected on a was dissolved in nitromethane (ca. 0.7 mL), and the resultant solu-
glass filter, washed with tetrahydrofuran and briefly dried under
tion was concentrated again. This procedure was repeated four
vacuum. Yield: 358 mg (88%). M.p. ca. 189 °C (dec.). 1H NMR times in order to remove water and excess HBF4. The residue was
(300 MHz, CD3CN): δ ϭ 1.96 (s, ca. 6 H, CH3CN), 2.9Ϫ3.3 (m, then dissolved in [D3]nitromethane (1.0 mL). Addition of 1 equiv.
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2309

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
of diethyl sulfide resulted in a sharp decrease in the J(1H,199Hg)
of the starting materials [MeHg(OH2)](BF4)·xH2O and [14]aneS4.
2
coupling constant from 251 to 218 Hz, indicating the formation of 4: Suitable single crystals were obtained by slow diffusion of
[MeHg(SEt2)]ϩ {1H NMR (300 MHz, [D3]nitromethane): δ ϭ 1.17
diethyl ether through the gas phase into a solution of
[PhHg([9]aneS3)](BF4) in acetonitrile. Under these conditions,
[s with satellites, 2J(1H,199Hg) ϭ 218 Hz, 3 H, HgCH3], 1.55 (t,
3
3JH,H ϭ 7.3 Hz, 6 H, CH2CH3), 3.49 (q, JH,H ϭ 7.3 Hz, 4 H; symmetrization of the PhHgϩ compound to
4 and HgPh2
occurred.[37a] The BF4 anion is disordered; its F atoms were re-
fined on two positions with an occupancy ratio of 2:1. CCDC-
207378 (1), -207379 (2), -207380 (3), and -207381 (4) contain the
supplementary crystallographic data for this paper. These data can
be obtained free of charge at www.ccdc.cam.ac.uk/conts/
retrieving.html [or from the Cambridge Crystallographic Data
Center, 12 Union Road, Cambridge CB2 1EZ, UK; Fax: (internat.)
ϩ 44-1223/336-033; Email: deposit@ccdc.cam.ac.uk].
Ϫ
SCH2)}. Finally, 1 equiv. of trifluoromethanesulfonic acid was ad-
ded. The intensity of the SCH2 signal of the thioether served as
a reference.
X-ray Crystallography: Crystal data and details of the data collec-
tions and structural refinements are summarized in Table 2. The
crystals were embedded in perfluoropolyalkyl ether (viscosity 1600
cSt., m.p. Ϫ20 °C; ABCR, Karlsruhe, Germany) and placed into
glass capillaries. The capillaries were sealed and transferred into
the cold gas stream of a STOE IPDS area detector diffractometer.
Computational Methods: The semiempirical calculations (AM1[75]
and PM3[76]) employed the program package MOPAC 2002,[77]
while the ab initio and DFT calculations were performed with
˚
Mo-Kα radiation (λ ϭ 0.71073 A) was used for intensity data col-
lection. The measuring temperature was 193 K. Numerical absorp-
tion corrections were applied to the data. The structures were
solved by direct methods and refined on F2 values. Anisotropic
displacement parameters were refined for the non-hydrogen atoms,
except for the C atom of the MeHgϩ group of compound 2, which
gave unreasonable values on anisotropic refinement. All hydrogen
atoms were included in idealized positions by use of riding models.
Programs used were those of the SHELX-97 software package[73]
and DIAMOND for the graphical representations.[74] 1: Single
crystals were grown by diffusion of diethyl ether through the gas
phase into a solution of 1 in acetonitrile. The four largest peaks in
Gaussian 98.[78] With the exception of
a single B3LYP/
LANL2DZ[79] optimization, all calculations at the correlated MP2
or the DFT BP86,[80] and the hybrid functional B3LYP[81][82] levels
of theory used the Stuttgart/Dresden SDD basis set and the effec-
tive core potential for Hg.[66] The remaining atoms were described
with the 6-31G(d) or 6-31ϩG(d,p) basis sets.[67] In two cases the
SDD basis sets were added to these Pople bases. Local minima and
transition states were verified by determination of the number of
imaginary frequencies. Details of the NMR chemical shift calcu-
lations are given in the Results and Discussion section.
˚
the final difference map are located at distances Ͻ 0.9 A from the
Acknowledgments
Hg atoms. The remaining difference peaks are smaller than 0.7
e·AϪ3. 2: A method similar to that described above for this com-
We thank Prof. Anne O. Summers and Dr. Irene Wagner-Döbler
for providing information about structural and mechanistic details
of organomercurial lyase. We also thank Prof. Martin Kaupp for
helpful discussions on the computation of chemical shifts in Hg-
˚
pound yielded suitable single-crystals. The structure was success-
fully refined in the non-centrosymmetric space group Cc. The abso-
lute structure parameter was Ϫ0.05(2) for the final model. 3: Suit-
able single crystals were obtained using a procedure similar to that containing systems. Financial support from the Deutsche For-
described above for this compound, however with use of a 1:1 ratio schungsgemeinschaft and generous allocation of computer time at
Table 2. Crystallographic data and details of the data collection and structure refinement for compounds 1Ϫ4
1
2
3
4
Empirical formula
Mr
C7H15BF4HgS3
482.8
monoclinic
P21/c
7.7802(4)
42.319(2)
8.6996(5)
109.098(6)
2706.7(2)
8
C10H21BF4HgS3
524.9
monoclinic
Cc
10.306(1)
12.449(1)
12.590(1)
92.067(10)
1614.2(2)
4
C12H26B2F8Hg2S4
873.4
monoclinic
P21/n
8.6904(6)
13.951(1)
9.8959(9)
102.86(1)
1169.7(2)
2
C16H30B2F8HgN2S6
817.0
orthorhombic
Pbca
8.8396(3)
21.463(1)
15.137(1)
90
2871.9(3)
4
Crystal system
Space group
˚
a [A]
˚
b [A]
˚
c [A]
β [°]
3
˚
V [A ]
Z
ρcalcd. [g cmϪ3
]
2.369
11.852
2.160
9.946
2.480
13.526
1.890
5.858
µ(Mo-Kα) [mmϪ1
]
Crystal size [mm]
Exposures
0.40 ϫ 0.33 ϫ 0.21
259
0.70 ϫ 0.16 ϫ 0.06
139
0.36 ϫ 0.17 ϫ 0.12
136
0.44 ϫ 0.27 ϫ 0.18
200
Oscillation angle for each exposure [°]
0.7
52.06
95.0
19449
1.4
51.56
94.7
5868
1.4
51.54
93.5
8270
1.0
51.96
94.5
21248
2Θmax. [°]
Completeness to 2Θmax. [%]
Reflections collected
Independent reflections (Rint
Data/restraints/parameters
R1 [I Ͼ 2σ(I)]
)
5087 (0.1899)
5087/0/289
0.0448
0.0907
0.983
2872 (0.0891)
2872/2/167
0.0499
0.1323
1.024
2095 (0.0651)
2095/0/128
0.0258
0.0597
0.940
2662 (0.1139)
2662/0/197
0.0232
0.0553
0.922
wR2 (all data)
Goodness-of-fit on F2
Ϫ3
˚
Largest difference peak/hole [e·A
]
ϩ2.06/Ϫ2.35
ϩ1.16/Ϫ1.53
ϩ1.05/Ϫ0.84
ϩ0.31/Ϫ0.53
2310
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312

The First Structurally Authenticated Organomercury(1ϩ) Thioether Complexes
FULL PAPER
[27]
T. P. Begley, A. E. Walts, C. T. Walsh, Biochemistry 1986, 25,
7192Ϫ7200.
HLRN, Hannover (Germany), and APAC, Canberra (Australia),
are gratefully acknowledged.
[28]
[29]
P. Barbaro, F. Cecconi, C. A. Ghilardi, S. Midollini, A. Orland-
ini, A. Vacca, Inorg. Chem. 1994, 33, 6163Ϫ6170.
Tetrahedral coordination is very rare in organomercury chem-
istry. Instead, RHgϩ cations strongly prefer linear coordination
of the Hg atom. Additional coordinate bonds occur frequently,
but usually they are weak, ‘‘secondary’’ bonds: [29a] J. S. Casas,
M. S. Garcıa-Tasende, J. Sordo, Coord. Chem. Rev. 1999,
193Ϫ195, 283Ϫ359.
ganomet. Chem. 1995, 495, 1Ϫ31.
V. Barone, A. Bencini, F. Totti, M. G. Uytterhoeven, J. Phys.
Chem. 1995, 99, 12743Ϫ12750.
R. D. Bach, H. B. Vardhan, J. Org. Chem. 1986, 51,
1609Ϫ1610.
M. T. Fairhurst, D. L. Rabenstein, Inorg. Chem. 1975, 14,
1413Ϫ1415.
T. E. Jones, L. S. W. L. Sokol, D. B. Rorabacher, M. D. Glick,
J. Chem. Soc., Chem. Commun. 1979, 140Ϫ141.
R. G. Pearson, Inorg. Chem. 1973, 12, 712Ϫ713.
See for example:
Gavrilova, O. A. Reutov, Dokl. Akad. Nauk SSSR 1982, 262,
108Ϫ113; English translation: Dokl. Chem. 1982, 262, 4Ϫ7.
[35b] P. L. Goggin, R. J. Goodfellow, S. R. Haddock, J. G. Eary,
J. Chem. Soc., Dalton Trans. 1972, 647Ϫ653.
[1]
C. H. Lamborg, W. F. Fitzgerald, J. O’Donnell, T. Torgersen,
Geochim. Cosmochim. Acta 2002, 66, 1105Ϫ1118.
United States Environmental Protection Agency, Mercury
[2]
Study Report to Congress, vol. 3 (‘‘Fate and Transport of Mer-
cury in the Environment’’) to be found under http://www.epa.
gov/oar/mercury.html, 1997.
´
[29b]
´
C. E. Holloway, M. Melnık, J. Or-
[3]
W. F. Fitzgerald, R. P. Mason, in Metal Ions in Biological Sys-
[30]
[31]
[32]
[33]
tems, vol. 34 (Eds.: A. Sigel, H. Sigel), Dekker, New York, 1997,
pp. 53Ϫ111.
[4]
R. P. Mason, W. F. Fitzgerald, F. M. M. Morel, Geochim. Cos-
mochim. Acta 1994, 58, 3191Ϫ3198.
[5]
J. O. Nriagu, Nature 1989, 338, 47Ϫ49.
[6]
J. O. Nriagu, J. M. Pacyna, Nature 1988, 333, 134Ϫ139.
R. A. Bodaly, V. L. St. Louis, M. J. Paterson, R. J. P. Fudge,
[7]
B. D. Hall, D. M. Rosenberg, J. W. M. Rudd, in Metal Ions in
Biological Systems, vol. 34 (Eds.: A. Sigel, H. Sigel), Dekker,
New York, 1997, pp. 259Ϫ287.
[34]
[35]
[35a]
V. A. Nikanorov, V. I. Rozenberg, G. V.
[8]
A. Boudou, F. Ribeyre, in Metal Ions in Biological Systems,
vol. 34 (Eds.: A. Sigel, H. Sigel), Dekker, New York, 1997,
pp. 289Ϫ319.
[9]
Committee on the Toxicological Effects of Methylmercury,
National Research Council, Toxicological Effects of Methyl-
mercury, National Academy Press, Washington, DC, 2000; also
to be found under http://www.nap.edu/books/0309071402/
html/.
[36] [36a]
Cambridge Structural Database (CSD), version 5.24 (No-
vember 2002, 272066 entries) and updates (FebruaryϪJuly
2003, 24361 entries). [36b]F. H. Allen, Acta Crystallogr., Sect. B
2002, 58, 380Ϫ388.
M. Wilhelm, Doctoral Dissertation, University of Olden-
burg, Germany, 2001; also to be found under http://docserver.-
bis.uni-oldenburg.de/publikationen/dissertation/2001/wilneu01/
[10]
D. W. Nierenberg, R. E. Nordgren, M. B. Chang, R. W. Siegler,
[37] [37a]
M. B. Blayney, F. Hochberg, T. Y. Toribara, E. Cernichiari, T.
Clarkson, New Engl. J. Med. 1998, 338, 1672Ϫ1676.
[11]
H. Strasdeit, Nachr. Chem. Tech. Lab. 1998, 46, 846Ϫ848.
[37b]
wilneu01.html.
H. Strasdeit, S. Deeken, W. Saak, M.
[12]
United States Environmental Protection Agency, Mercury
Wilhelm, GDCh-Jahrestagung Chemie 2001, Würzburg, Ger-
many, September 23Ϫ29, 2001, Abstract WÖHLER-044.
H. Strasdeit, M. Wilhelm, S. Deeken, E. Berssen, R. Koch,
35th Int. Conf. Coord. Chem. Heidelberg, Germany, July
21Ϫ26, 2002, Abstract P1.128.
Study Report to Congress, vol. 5 (‘‘Health Effects of Mercury
and Mercury Compounds’’), to be found under http://
www.epa.gov/oar/mercury.html, 1997.
[37c]
[13]
N. K. Mottet, M. E. Vahter, J. S. Charleston, L. T. Friberg, in
Metal Ions in Biological Systems, vol. 34 (Eds.: A. Sigel, H.
Sigel), Dekker, New York, 1997, pp. 371Ϫ403.
R. Hamada, K. Arimura, M. Osame, in Metal Ions in Biologi-
[38] [38a]
`
A. De Robertis, C. Foti, G. Patane, S. Sammartano, J.
Chem. Eng. Data 1998, 43, 957Ϫ960. [38b] D. L. Rabenstein, C.
[14]
A. Evans, M. C. Tourangeau, M. T. Fairhurst, Anal. Chem.
1975, 47, 338Ϫ341.
Helv. Chim. Acta 1965, 48, 28Ϫ46.
cal Systems, vol. 34 (Eds.: A. Sigel, H. Sigel), Dekker, New
York, 1997, pp. 405Ϫ420.
R. Hamada, M. Osame, in Toxicology of Metals (Ed.: L. W.
[38c]
G. Schwarzenbach, M. Schellenberg,
[15]
[39] [39a]
C. A. Barradell, H. G. M. Edwards, D. N. Smith, J. Or-
Chang), CRC, Boca Raton, FL, 1996, pp. 337Ϫ351.
The Toxicity of Methyl Mercury (Eds.: C. U. Eccles, Z. Annau),
[39b]
ganomet. Chem. 1991, 421, 137Ϫ146.
H. Schmidbaur, H.-
[16]
J. Öller, S. Gamper, G. Müller, J. Organomet. Chem. 1990,
Johns Hopkins University Press, Baltimore, 1987.
394, 757Ϫ764.
[17]
M. Schrope, Nature 2001, 409, 124.
[40]
[41]
[42]
[18]
R. S. Glass, G. S. Wilson, W. N. Setzer, J. Am. Chem. Soc.
1980, 102, 5068Ϫ5069.
For a definition of the conformational nomenclature see: J.
Dale, Acta Chem. Scand. 1973, 27, 1115Ϫ1129.
A detailed conformational study of [9]aneS3 in metal com-
plexes is available: J. Beech, P. J. Cragg, M. G. B. Drew, J.
Chem. Soc., Dalton Trans. 1994, 719Ϫ729.
H. Strasdeit, A. von Döllen, W. Saak, M. Wilhelm, Angew.
Chem. 2000, 112, 803Ϫ805; Angew. Chem. Int. Ed. 2000, 39,
784Ϫ786.
[19]
L. Magos, in Metal Ions in Biological Systems, vol. 34 (Eds.:
A. Sigel, H. Sigel), Dekker, New York, 1997, p. 327.
M. J. Moore, M. D. Distefano, L. D. Zydowsky, R. T. Cum-
[20]
mings, C. T. Walsh, Acc. Chem. Res. 1990, 23, 301Ϫ308.
[43] [43a]
[21]
P. Pyykkö, M. Straka, Phys. Chem. Chem. Phys. 2000, 2,
2489Ϫ2493.
In the absence of organomercurial lyase, mercury resistance is
[43b]
restricted to Hg2ϩ (‘‘narrow-spectrum’’ resistance).
S. M. Miller, in Essays in Biochemistry, vol. 34 (Ed.: D. P. Bal-
A. J. Canty, G. B. Deacon, Inorg. Chim. Acta
[43c]
[22]
1980, 45, L225ϪL227.
The possibility of a larger RvdW
value of HgII (around 2.0 A) has been discussed, see for ex-
ample: K. R. Flower, V. J. Howard, S. Naguthney, R. G. Pritch-
ard, J. E. Warren, A. T. McGown, Inorg. Chem. 2002, 41,
1907Ϫ1912; S. S. Batsanov, J. Chem. Soc., Dalton Trans.
1998, 1541Ϫ1545.
˚
lou), Portland Press, London, 1999, pp. 17Ϫ30.
A. M. Osborn, K. D. Bruce, P. Strike, D. A. Ritchie, FEMS
[23]
Microbiol. Rev. 1997, 19, 239Ϫ262.
J. L. Hobman, N. L. Brown, in Metal Ions in Biological Sys-
[24]
tems, vol. 34 (Eds.: A. Sigel, H. Sigel), Dekker, New York,
1997, pp. 527Ϫ568.
[44]
Because van der Waals radii of sulfur and fluorine have been
found to be anisotropic, ranges are used here instead of single
values: S. C. Nyburg, C. H. Faerman, Acta Crystallogr., Sect.
B 1985, 41, 274Ϫ279.
[25]
A. O. Summers, Annu. Rev. Microbiol. 1986, 40, 607Ϫ634.
[26]
Transgenic plants engineered to express merB seem promising
for the degradation of methylmercury at polluted sites; see for
[45]
[46]
[26a]
S. C. Rawle, G. A. Admans, S. R. Cooper, J. Chem. Soc., Dal-
ton Trans. 1988, 93Ϫ96.
R. E. DeSimone, M. D. Glick, J. Am. Chem. Soc. 1976, 98,
762Ϫ767.
example:
S. P. Bizily, C. L. Rugh, R. B. Meagher, Nat.
Biotechnol. 2000, 18, 213Ϫ217. [26b] S. P. Bizily, C. L. Rugh, A.
O. Summers, R. B. Meagher, Proc. Natl. Acad. Sci. USA 1999,
96, 6808Ϫ6813.
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312
www.eurjic.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2311

M. Wilhelm, S. Deeken, E. Berssen, W. Saak, A. Lützen, R. Koch, H. Strasdeit
FULL PAPER
[47]
[67b]
Only the non-disordered α and β1 molecules of free [14]aneS4
M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M.
S. Gordon, D. J. DeFrees, J. A. Pople, J. Chem. Phys. 1982,
77, 3654Ϫ3665.
M. B. Blayney, J. S. Winn, D. W. Nierenberg, Chem. Eng. News
1997, 75, 7.
were used for comparison, see ref.[46]
[48]
[49]
D. L. Rabenstein, Acc. Chem. Res. 1978, 11, 100Ϫ107.
A. J. Canty, A. Marker, Inorg. Chem. 1976, 15, 425Ϫ430.
[68]
[69]
[70]
[50] [50a]
P. L. Goggin, R. J. Goodfellow, D. M. McEwan, A. J.
Griffiths, K. Kessler, J. Chem. Res. (S) 1979, 194Ϫ195; J.
B. Wrackmeyer, R. Contreras, Annu. Rep. NMR Spectrosc.
1992, 24, 267Ϫ329.
[50b]
Chem. Res. (M) 1979, 2315Ϫ2343.
J. Schnittker, R. Smith, M. C. Baird, Inorg. Chim. Acta 1978,
K. Stanley, J. Martin,
T. V. O’Halloran, C. P. Singer, 199Hg NMR Standards,
to be found under http://www.chem.nwu.edu/ϳohallo/
HgNMRStandards/index.html, 1998.
[50c]
27, L111ϪL113.
H. Schmidbaur, K.-H. Räthlein, Chem.
Ber. 1973, 106, 2491Ϫ2507.
The signal position was inferred from the proton-coupled 13C
NMR spectrum in which the CH3 quadruplet is only partly
[71]
[51]
[52]
N. W. Alcock, N. Herron, P. Moore, J. Chem. Soc., Dalton
Trans. 1978, 394Ϫ399.
See for example: H. Friebolin, Basic One- and Two-Dimensional
NMR Spectroscopy, 3rd ed., Wiley-VCH, Weinheim, 1998, p.
124f.
1
superposed. The J(13C,1H) values are 138.5 Hz (HgCH3) and
143.9 Hz (CH2).
[72]
[73]
[74]
The isotopic pattern unambiguously showed that this signal
was not due to [(MeHg)2([14]aneS4)]2ϩ
.
[53]
[54]
[55]
[56]
K. E. Pitts, A. O. Summers, Biochemistry 2002, 41,
G. M. Sheldrick, SHELX-97, Universität Göttingen,
Göttingen, Germany, 1997.
10287Ϫ10296.
The numbering of the amino acid residues refers to the organo-
mercurial lyase that is encoded by plasmid R831b, see ref.[53]
T. Fujinaga, I. Sakamoto, J. Electroanal. Chem. 1977, 85,
185Ϫ201.
DIAMONDϪϪVisual Crystal Structure Information System,
version 2.1, Crystal Impact, Bonn, Germany, 1999.
[75] [75a] M. J. S. Dewar, E. G. Zoebisch, E. F. Healy, J. J. P. Stewart,
[75b]
J. Am. Chem. Soc. 1985, 107, 3902Ϫ3909.
M. J. S. Dewar,
F. A. L. Anet, D. J. O’Leary, Tetrahedron Lett. 1989, 30,
[75c]
Y. C. Yuan, Inorg. Chem. 1990, 29, 3881Ϫ3890.
Dewar, C. Jie, Organometallics 1989, 8, 1547Ϫ1549.
J. J. P. Stewart, J. Comput. Chem. 1989, 10, 209Ϫ220.
J. J. P. Stewart, J. Comput. Chem. 1991, 12, 320Ϫ341.
J. J. P. Stewart, MOPAC 2002, Fujitsu Limited, Tokyo, Japan,
2002.
M. J. S.
2755Ϫ2758.
[57] [57a]
J. Pickardt, J. Shen, Z. Naturforsch., Teil B 1993, 48,
[76] [76a]
[76b]
969Ϫ972. [57b] W. N. Setzer, Q. Guo, G. J. Grant, J. L. Hubbard,
R. S. Glass, D. G. VanDerveer, Heteroat. Chem. 1990, 1,
[77]
[78]
[57c]
317Ϫ325.
A. J. Blake, A. J. Holder, T. I. Hyde, G. Reid,
M. Schröder, Polyhedron 1989, 8, 2041Ϫ2045.
C. Walsh, T. Begley, A. Walts, Pure Appl. Chem. 1987, 59,
295Ϫ298.
M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M.
A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgom-
ery, R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C.
Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe,
P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, J. L.
Andres, M. Head-Gordon, E. S. Replogle, J. A. Pople, GAUS-
SIAN, Gaussian, Inc., Pittsburgh, PA, USA, 1998.
[58]
[59]
For theoretical studies on mercury-second-row element com-
plexes see: [59a] Ref.[30][59b]T. Rambusch, K. Hollmann-Gloe, K.
Gloe, J. Prakt. Chem. 1999, 341, 202Ϫ217.
For theoretical studies on HgϪC bond cleavage see: [60a] V. B a -
[60]
rone, A. Bencini, F. Totti, M. G. Uytterhoeven, Organometall-
[60b]
ics 1996, 15, 1465Ϫ1469.
V. Barone, A. Bencini, F. Totti,
M. G. Uytterhoeven, Int. J. Quantum Chem. 1997, 61,
[60c]
361Ϫ367.
J. A. Tossell, J. Phys. Chem. A 1998, 102,
3587Ϫ3591.
[61] [61a]
J. A. Altmann, N. C. Handy, V. E. Ingamells, Int. J. Quan-
[61b]
tum Chem. 1996, 57, 533Ϫ542.
J. A. Altmann, N. C.
Handy, Phys. Chem. Chem. Phys. 1999, 1, 5529Ϫ5536.
M. Kaupp, Chem. Ber. 1996, 129, 535Ϫ544.
G. Schreckenbach, T. Ziegler, Theor. Chem. Acc. 1998, 99,
71Ϫ82.
[79] [79a] T. H. Dunning, Jr., P. J. Hay in Modern Theoretical Chem-
istry, vol. 3 (‘‘Methods of Electronic Structure Theory’’) (Ed.:
H. F. Schaefer, III), Plenum, New York, 1977, p. 1Ϫ27. [79b]P.
J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270Ϫ283.
[62]
[63]
[64] [64a]
[80] [80a]
[80b]
M. Kaupp, O. L. Malkina, J. Chem. Phys. 1998, 108,
A. D. Becke, Phys. Rev. A 1988, 38, 3098Ϫ3100.
J.
[64b]
Perdew, Phys. Rev. B 1986, 33, 8822Ϫ8824. [80c] J. Perdew, Phys.
Rev. B 1986, 34, 7406.
3648Ϫ3659.
J. Vaara, O. L. Malkina, H. Stoll, V. G. Mal-
kin, M. Kaupp, J. Chem. Phys. 2001, 114, 61Ϫ71.
[65]
[66]
[81]
[82]
Review: J. Gauss, J. F. Stanton, Adv. Chem. Phys. 2002, 123,
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785Ϫ789.
A. D. Becke, J. Chem. Phys. 1993, 98, 5648Ϫ5652.
Received December 5, 2003
355Ϫ422, and references cited therein.
W. Kuechle, M. Dolg, H. Stoll, H. Preuss, Mol. Phys. 1991,
74, 1245Ϫ1263.
Early View Article
Published Online April 7, 2004
[67] [67a] P. C. Hariharan, J. A. Pople, Mol. Phys. 1974, 27, 209Ϫ214.
2312
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
Eur. J. Inorg. Chem. 2004, 2301Ϫ2312