Investigation of the mercury(II) coordination chemistry of tris[(1-methylimidazol-2-yl)methyl]amine by X-ray crystallography and NMR

Article abstract of DOI:10.1039/b300001j

The coordination chemistry of Hg(II) with tris[(1-methylimidazol-2-yl)methyl]amine (TMIMA) was investigated. The structures of [Hg(TMIMA)₂](ClO₄)₂ (1), [Hg(TMIMA)(NCCH₃)](ClO₄)₂ (2) and [Hg(TMIMA)Cl]₂(HgCl₄) (3) were characterized by X-ray crystallography. Complex 1 has six strong Hg-N_imidazoyt bonds ranging from 2.257(5) to 2.631(6) A. Ligand geometry suggests the Hg-N_(NR3) distances of 2.959(6) A in 1 reflects weak bonding interactions. This complex has a ¹⁹⁹Hg chemical shift of -1496 ppm, significantly upfield from nitrogen coordination complexes with lower coordination numbers. The five-coordinate complex 2 has Hg-N_(NR3), Hg-N_imidazoyl and Hg-N_acetonitrile bond lengths of 2.642(8), 2.198(5) and 2.264(11) A, respectively. Complex 3 is also five coordinate, with Hg-N_(NR3), Hg-Cl and average Hg-N _imidazoyl distances in the cations of 2.758(7), 2.424(2) and 2.29(4) A, respectively. Conditions for slow exchange on the J(HgH) coupling time-scale were found for both 1: 1 metal-to-ligand complexes in acetonitrile-d₃. Observed heteronuclear coupling constants were similar to those associated with Hg(II) substituted proteins with histidine-metal bonds. Solution and solid-state comparisons to the Hg(II) coordination chemistry of tetradenate pyridyl ligands are made. Relevance to development of ¹⁹⁹Hg NMR as a metallobioprobe is discussed. The Royal Society of Chemistry 2003.

Full text of DOI:10.1039/b300001j

View Article Online / Journal Homepage / Table of Contents for this issue
Investigation of the mercury(II) coordination chemistry of
tris[(1-methylimidazol-2-yl)methyl]amine by X-ray crystallography
and NMR†
Deborah C. Bebout,*a Melissa M. Garland,a Geoffrey S. Murphy,a Edith V. Bowers,a
Christopher J. Abelta and Raymond J. Butcherb
a Chemistry Department, The College of William & Mary, Williamsburg, VA 23187, USA.
E-mail: dcbebo@wm.edu; Fax: 757 221 2715; Tel: 757 221 2558
b Chemistry Department, Howard University, Washington, DC 20059, USA.
E-mail: rbutcher@howard.edu; Fax: 202 806 5442; Tel: 202 806 6886
Received 3rd January 2003, Accepted 15th April 2003
First published as an Advance Article on the web 12th May 2003
The coordination chemistry of Hg() with tris[(1-methylimidazol-2-yl)methyl]amine (TMIMA) was investigated.
The structures of [Hg(TMIMA)2](ClO4)2 (1), [Hg(TMIMA)(NCCH3)](ClO4)2 (2) and [Hg(TMIMA)Cl]2(HgCl4) (3)
were characterized by X-ray crystallography. Complex 1 has six strong Hg–Nimidazoyl bonds ranging from 2.257(5) to
2.631(6) Å. Ligand geometry suggests the Hg–N(NR ) distances of 2.959(6) Å in 1 reflects weak bonding interactions.
3
This complex has a 199Hg chemical shift of Ϫ1496 ppm, significantly upfield from nitrogen coordination complexes
with lower coordination numbers. The five-coordinate complex 2 has Hg–N(NR ), Hg–Nimidazoyl and Hg–Nacetonitrile
3
bond lengths of 2.642(8), 2.198(5) and 2.264(11) Å, respectively. Complex 3 is also five coordinate, with Hg–N(NR )
,
3
Hg–Cl and average Hg–Nimidazoyl distances in the cations of 2.758(7), 2.424(2) and 2.29(4) Å, respectively. Conditions
for slow exchange on the J(HgH) coupling time-scale were found for both 1 : 1 metal-to-ligand complexes in
acetonitrile-d3. Observed heteronuclear coupling constants were similar to those associated with Hg() substituted
proteins with histidine–metal bonds. Solution and solid-state comparisons to the Hg() coordination chemistry of
tetradenate pyridyl ligands are made. Relevance to development of 199Hg NMR as a metallobioprobe is discussed.
Development of 199Hg NMR would complement 113Cd NMR
Introduction
methods since the metals have different donor atom prefer-
A wide variety of multidentate ligand systems have been used to
ences. Furthermore, 199Hg NMR has important advantages
simulate protein metal coordination environments. Many aro-
over 113Cd NMR including shorter relaxation times,2 larger and
matic amines have been used in these ligands to emulate metal
longer range heteronuclear coupling constants3 and greater
coordination by the 4-methylimidazoyl side chain of histidine,
including various derivatives of pyridine, imidazole and
pyrazole. Typically the coordination chemistry of the physio-
logically essential metal ions with each ligand is methodically
chemical shift dispersion. Hg() is slightly larger than Cd()
and significantly larger than the physiologically essential
metal ions. However, Hg() has been substituted for native
Fe(),4 Cu()2,5–7 and Zn()5,8 ions without significant
investigated for both biomimetic behavior and similarity to the
changes in overall protein structure or identity of the metal
spectroscopic signatures of metals in protein coordination
ligands.
environments. These studies have provided considerable insight
Although rapid exchange is commonly associated with Hg()
into the structure and function of numerous metalloproteins.
coordination chemistry, we have found that tri- and tetra-
Recently we have been investigating the coordination chem-
dentate amine ligands exhibit slow exchange behavior on the
istry of Hg() in biologically relevant ligand systems to investi-
chemical shift time-scale under appropriate conditions. Our
gate the potential of 199Hg NMR as a metallobioprobe of
previous studies have focused on tripodal and dipodal
proteins. The ideal metallobioprobe would have similar co-
amines with 2-methyl- and 2,6-dimethylpyridyl substituents.9–12
ordination properties to the native metal ion yet offer a sensitive
The Mn(),13 Fe(),14 and Cu()15 coordination chemistry
alternative means of protein characterization. NMR is one of
of tris[2-(1-methylimidazol-2-yl)methyl]amine (TMIMA) has
the most powerful spectroscopic techniques because it can pro-
been investigated recently. Since the 1,2-dimethylimidazoyl
vide specific bonding information. Since none of the physio-
substituents of these ligands are more similar to histidine than
logically essential metal ions has an isotope with favorable
the pyridyl derivatives associated with ligands previously exam-
NMR properties, detailed NMR information about protein
ined, we decided to compare the coordination properties of
metal binding sites requires metal substitution. For example, a
these ligands. In this paper we report the structures of [Hg-
wide variety of Zn() proteins have been substituted with Cd()
(TMIMA)2](ClO4)2 (1), [Hg(TMIMA)(NCCH3)](ClO4)2 (2) and
[Hg(TMIMA)Cl]2(HgCl4) (3), the first structurally character-
ized complexes of Hg() with polyimidazoyl ligands. In addi-
tion, the acetonitrile-d3 solution-state NMR of TMIMA in the
and investigated by 113Cd NMR spectroscopy to gain valuable
insight.1 The wealth of information potentially available from
NMR, including the identity of specific metal ligating amino
acids, has motivated Cd() substitution of less similar metal
presence of Hg(ClO4)2 and HgCl2 is reported. These results are
ions as well. These studies have taken advantage of the limited
compared with previous studies of multidentate pyridyl ligands.
coordination geometry preferences of d10 metal ions. Although
also a d10 metal ion with a favorable isotope for NMR studies,
Hg() has not been used extensively as a metallobioprobe.
Experimental
Methods and materials
† Electronic supplementary information (ESI) available: Cell packing
diagrams for 1–3 and NMR figures showing trends in chemical shift as
a function of [Hg()]/1 for both Hg(ClO4)2 and HgCl2. See http://
www.rsc.org/suppdata/dt/b3/b300001j/
Starting materials were of commercially available reagent qual-
ity. Elemental analyses were carried out by Atlantic Microlab,
Inc., Norcross, Georgia.
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
2578

View Article Online
Table 1 Selected crystallographic data for 1–3
Complex
[Hg(TMIMA)2](ClO4)2 (1)
[Hg(TMIMA)(NCCH3)](ClO4)2 (2)
[Hg(TMIMA)Cl]2(HgCl4) (3)
Empirical formula
Formula weight
Crystal system, space group
a/Å
b/Å
c/Å
α/Њ
C30H42Cl2HgN14O8
998.27
Monoclinic, P21/c
12.3821(13)
13.6576(15)
12.5346(19)
90
C17H24Cl2HgN8O8
739.93
Rhombohedral, R3c
12.3849(8)
12.3849(8)
57.739(4)
90
C30H42Cl6Hg3N14
1413.25
Orthorhombic, Pbna
9.093(3)
15.381(4)
30.119(9)
90
¯
β/Њ
110.358(9)
90
1987.3(4)
2
1.668
4.072
90
120
7669.4(9)
12
1.922
6.287
293(2)
90
90
4213(2)
4
2.228
11.329
168(2)
γ/Њ
V/Å3
Z
Dc/Mg mϪ3
Absorption coefficient/mmϪ1
Temperature/K
296(2)
Independent reflections (Rint
R1,a wR2b [I > 2σ(I )]
R1,a wR2b (all data)
)
4453 (0.0351)
0.0425, 0.0967
0.0784, 0.1147
1972 (0.0352)
0.0391, 0.0924
0.0824, 0.1108
4771 (0.0487)
0.0319, 0.0648
0.0635, 0.0923
2
a R1 = Σ||Fo| Ϫ |Fc||/Σ|Fo|. b wR2 = [Σ[w(Fo2 Ϫ Fc2)2]/Σ[w(Fo )2]1/2
.
diethyl ether (4 × 30 mL), the pH of the aqueous phase was
adjusted above 9.0 with sodium carbonate and extracted with
chloroform (4 × 30 mL). The combined organics were dried
with MgSO4, filtered, and concentrated in vacuo to a yellow
solid. The solid was suspended in acetonitrile (2 mL), vacuum
filtered, then dried to afford 339 mg (18%) of a white solid; mp
202.5–203.5 ЊC (literature value unavailable). δH (400 MHz;
solvent CD3CN; standard SiMe4) 6.89 (3 H, d, J(HH) 1.3 Hz,
Hb), 6.81 (3 H, d, J(HH) 1.4 Hz, Ha), 3.66 (6 H, s, Hd), 3.05
(9 H, s, Hc).
All of the perchlorate salts of mercury() complexes included
in this work were stable for routine synthesis and purification
procedures. However, caution should be exercised because
perchlorate salts of metal complexes with organic ligands are
potentially explosive.16
NMR measurements
Solutions for NMR analysis were prepared by adding stock
solutions of mercuric salts in acetonitrile-d3 to solutions of
TMIMA or isolated complexes in acetonitrile-d3 using cali-
brated autopipets. NMR spectra were recorded in 5-mm o.d.
NMR tubes on a Varian Mercury 400VX operating in the pulse
Fourier transform mode. The sample temperature was main-
tained by blowing chilled nitrogen over the NMR tube in the
probe. Proton chemical shifts were measured relative to internal
solvent but are reported relative to tetramethylsilane. Chemical
shifts of 199Hg were measured relative to an external reference
of 0.1 M phenylmercuric acetate in acetonitrile-d3 (Ϫ1439.5
ppm)19 but are reported relative to dimethylmercury (0 ppm).
Synthesis of 1-methyl-2-imidazolecarbaldehyde oxime (MICAO)
MICAO was prepared by variation of the procedure of Ober-
hausen et al.17 Hydroxylamine hydrochloride (19.0 g, 0.273
mol) and sodium carbonate (14.5 g, 0.137 mol) were dissolved
in cold water. The solution was placed in an ice-bath and
1-methyl-2-imidazolecarbaldehyde (28.0 g, 0.255 mole) dissol-
ved in ethanol (100 mL) was added dropwise. The solution was
refluxed for 3 h then stored at 4 ЊC to precipitate the product.
The white solid (29.5 g, 87%) isolated by vacuum filtration,
rinsing with cold ethanol (30 mL) and drying in a vacuum
desiccator was adequately pure for further manipulations; mp
168–170 ЊC (lit.,17 176 ЊC). δH (400 MHz; solvent DMSO-d6;
standard SiMe4) 8.07 (1 H, s, CHoxime), 7.27 (1 H, s, CHimidazoyl),
7.00 (1 H, s, CHimidazoyl), 3.82 (3 H, s, CH3).
X-Ray crystallography
Selected crystallographic data are given in Table 1 and complete
data are given in the CIF file. Data for 1 and 2 were collected at
room temperature on a Siemens P4 four-circle diffractometer
using a graphite-monochromated Mo-Kα X-radiation (λ =
0.71073 Å). During data collection three standard reflections
were measured after every 97 reflections. Both crystals turned
light gray in the beam, but the decay in the intensity of the
standards was less than random error so no decay correction
was performed on the data. Intensity data for 3 were collected
on a Bruker SMART CCD diffractometer with graphite-mono-
chromated Mo-Kα radiation (λ = 0.71073 Å) at Ϫ105 ЊC with
absorption corrections performed using the SADABS pro-
gram.20 Although 3 turned light gray in the beam, re-collection
of the initial 50 frames and analysis of these frames showed
that no decay correction was needed. The structures were
solved by direct methods21 and refined on F 2 by full-matrix
least-squares using the SHELXTL97 program package.22 All
non-hydrogen atoms were refined as anisotropic and, except as
noted for the CH3CN hydrogens of 2, the hydrogen atomic
positions were constrained in an idealized geometry relative to
the bonded carbons and the isotropic thermal parameters were
fixed.
Synthesis of tris[2-(1-methylimidazol-2-yl)methyl]amine
(TMIMA)
TMIMA was prepared from three equivalents of 1-methyl-2-
imidazolecarbaldehyde by first converting two equivalents to
the oxime MICOA17 and preparing bis[2-(1-methylimidazol-
2-yl)methyl]amine (BMIMA) by catalytic hydrogention.18
BMIMA (1.0 g, 4.88 mmol) was added to a magnetically stirred
solution of glacial acetic acid (1.25 mL, 21.8 mmol) and meth-
anol (80 mL). A solution of 1-methyl-2-imidazolecarbaldehyde
(0.73 g, 6.62 mmol) in methanol (40 mL) was added dropwise.
The reaction mixture was cooled to 0 ЊC and sodium
cyanoborohydride (1.38 g, 21.8 mmol) was added. After stir-
ring for 18 h at 0 ЊC, the solution was acidified by addition of
concentrated HCl. The reaction mixture was filtered to
remove the white precipitate. The filtrate was concentrated in
vacuo and dissolved in 30 mL water. Following extraction with
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2579

View Article Online
Table 2 Selected bond distances (Å) and angles (Њ) in 1–3
[Hg(TMIMA)2](ClO4)2 (1)a
[Hg(TMIMA)(NCCH3)](ClO4)2 (2)b
[Hg(TMIMA)Cl]2(HgCl4) (3)c
Hg–N
2.957(6)
2.631(6)
2.257(5)
2.537(5)
Hg–N
Hg–N(1)
Hg–N(11)
2.642(8)
2.198(5)
2.264(11)
Hg–N
2.758(7)
2.273(7)
2.250(6)
2.340(6)
2.424(2)
Hg–N(1A)
Hg–N(1B)
Hg–N(1C)
Hg–N(1A)
Hg–N(1B)
Hg–N(1C)
Hg–Cl
N–Hg–N(1A)
N–Hg–N(1B)
N–Hg–N(1C)
N–Hg–N(1AA)
N–Hg–N(1BA)
N–Hg–N(1CA)
N(1A)–Hg–N(1B)
N(1A)–Hg–N(1C)
N(1B)–Hg–N(1C)
N(1A)–Hg–N(1AA)1
N(1B)–Hg–N(1BA)1
N(1C)–Hg–N(1CA)1
N(1A)–Hg–N(1B)1
N(1A)–Hg–N(1C)1
N(1B)–Hg–N(1C)1
59.8(2)
63.5(2)
60.5(2)
N–Hg–N(1)
N–Hg–N(11)
N(1)–Hg–N(1)
N(1)–Hg–N(11)
Hg–N(11)–C(11)
71.83(11)
180.00(2)
110.75(11)
108.16(11)
180.00(2)
N–Hg–N(1A)
N–Hg–N(1B)
N–Hg–N(1C)
N(1A)–Hg–N(1B)
N(1A)–Hg–N(1C)
N(1B)–Hg–N(1C)
Cl–Hg–N
Cl–Hg–N(1A)
Cl–Hg–N(1B)
Cl–Hg–N(1C)
68.2(2)
68.3(2)
67.9(2)
119.4(2)
101.5(2)
98.8(2)
170.5(2)
102.46(17)
117.25(17)
117.04(17)
120.2(2)
116.5(2)
119.5(2)
100.37(19)
99.59(17)
96.31(18)
180.0(2)
180.0(4)
180.0(3)
79.63(19)
80.41(17)
83.69(18)
a Symmetry transformation used to generate equivalent atoms: Ϫx, Ϫy ϩ 1, Ϫz ϩ1. b Symmetry transformations used to generate equivalent atoms:
Ϫx ϩ y ϩ 1, Ϫx ϩ 1, z; Ϫy ϩ 1, x Ϫ y, z; x Ϫ y ϩ 1/3, Ϫy ϩ 2/3, Ϫz ϩ 7/6; Ϫx ϩ 4/3, Ϫy ϩ 2/3, Ϫz ϩ 2/3; x Ϫ y ϩ 1/3, x Ϫ 1/3, Ϫz ϩ 2/3; y ϩ 1/3, Ϫx
ϩ y ϩ 2/3, Ϫz ϩ 2/3. c Symmetry transformation used to generate equivalent atoms: x, Ϫy ϩ 1/2, Ϫz ϩ 1.
CCDC reference numbers 200560–200562.
See http://www.rsc.org/suppdata/dt/b3/b300001j/ for crystal-
lographic data in CIF or other electronic format.
Preparation and X-ray diffraction of [Hg(TMIMA)Cl]2-
(HgCl4) (3). HgCl2 (158 mg, 0.334 mmol) was dissolved in 8 mL
acetonitrile and TMIMA (100 mg, 0.334 mmol) was added.
Toluene (2.5 mL) was added slowly with stirring. The solution
was separated into 2 mL aliquots and set aside for slow evapor-
ation. Upon standing for one week, some vials contained color-
less plates (mp 210–212 ЊC (decomp.)) and others contained
colorless needles suitable for X-ray crystallography (mp 215–
217 ЊC (decomp.)). Crystals were collected by drawing off the
mother-liquor and air drying. Yield 88 mg (28%). The two crys-
tal types were indistinguishable by NMR (Found C, 26.18;
H, 3.13; N, 13.90. C30H42Cl6Hg3N14 requires C, 25.50; H, 2.99;
N, 13.88%); δH (400 MHz; solvent CD3CN; standard
SiMe4; 2 mM; 20 ЊC) 7.13 (6 H, d, J(HH) 1.5 Hz, Hb), 6.99 (6 H,
d, J(HH) 1.5 Hz, Ha), 4.02 (12 H, s, Hd), 3.58 (18 H, s, Hc);
(Ϫ40 ЊC) 7.15 (6 H, d, J(HH) 1 Hz, J(HgH) 10 Hz, Hb), 6.95
(6 H, d, J(HH) 1 Hz, J(HgH) = 9 Hz, Ha), 4.01 (12 H, s, J(HgH)
= 11 Hz, Hd), 3.55 (18 H, s, Hc).
Preparation and X-ray diffraction of [Hg(TMIMA)2]-
(ClO4)2 (1). Hg(ClO4)2ؒ3H2O (75.7 mg, 0.167 mmol) was dissol-
ved in 8 mL acetonitrile and TMIMA (100 mg, 0.334 mmol)
was added. The solution was stirred and slowly diluted with
12 mL toluene and set aside for slow evaporation. Colorless
X-ray quality crystals formed upon standing for five days.
Crystals were collected by drawing off the mother liquor
and air drying. Yield: 120 mg (28%); mp 229-230 ЊC
(decomp.) (Found: C 36.06; H 4.18; N 19.48. C30H42Cl2-
HgN14O8 requires C 36.09; H 4.24; N 19.63%); δH (400 MHz;
solvent CD3CN; standard SiMe4; 2 mM; 20 ЊC) 6.98 (6 H, s,
Hb), 6.50 (6 H, s, Ha), 4.02 (12 H, s, Hd), 3.60 (18 H, s, Hc); δHg
(71.588 MHz; solvent CD3CN; standard 0.5 M phenylmercuric
acetate in DMSO-d6;19 50 mM with excess TMIMA; 20 ЊC)
Ϫ1496.
A crystal measuring 0.08 × 0.12 × 0.30 mm was glued on the
end of a glass fiber. The data were collected using the θ–2θ
technique over a θ range of 1.35–28.33Њ. The final data to
parameter ratio was 20 : 1.
A crystal measuring 0.25 × 0.75 × 0.55 mm was glued on the
end of a glass fiber. The data were collected using the θ–2θ
technique over a θ range of 2.98–27.5Њ. The final data to
parameter ratio was 15 : 1.
Results
Preparation and X-ray diffraction of [Hg(TMIMA)(NCCH3)]-
(ClO4)2 (2). Hg(ClO4)2ؒ3H2O (152 mg, 0.335 mmol) was dissol-
ved in 8 mL acetonitrile and TMIMA (100 mg, 0.334 mmol)
was added. Toluene (2.5 mL) was added slowly with stirring.
The solution was set aside for slow evaporation. Colorless
X-ray quality crystals formed upon standing for 4 days. Crys-
tals were collected by drawing off the mother-liquor and air
drying. Yield: 79 mg (32%); mp 205–207 ЊC (decomp.) (Found:
C, 27.64; H, 3.18; N, 15.37. C17H24Cl2HgN8O8 requires C, 27.59;
H, 3.27; N, 15.14%); δH (400 MHz; solvent CD3CN; standard
SiMe4; 2 mM; 20 ЊC) 7.18 (3 H, d, J(HH) 1.4 Hz, Hb), 7.08 (3 H,
d, J(HH) = 1.4 Hz, Ha), 4.11 (6 H, s, Hd), 3.55 (9 H, s, Hc);
(Ϫ40 ЊC) 7.18 (3 H, d, J(HH) 1.4 Hz, J(HgH) = 20 Hz, Hb), 7.05
(3 H, d, J(HH) = 1.4 Hz, J(HgH) 17 Hz, Ha), 4.03 (6 H, s,
J(HgH) 7 Hz, Hd), 3.55 (9 H, s, Hc).
The structures of [Hg(TMIMA)2](ClO4)2 (1), [Hg(TMIMA)-
(NCCH3)](ClO4)2 (2) and [Hg(TMIMA)Cl]2HgCl4 (3) are
reported herein. Selected bond distances and bond angles are
given in Table 2.
Crystal structure and solution-state NMR of [Hg(TMI-
MA)2](ClO4)2 (1). The Hg() ion is located at an inversion cen-
ter within a bicapped trigonal antiprism of nitrogen atoms (D3d
core symmetry, D3 ion symmetry) (Fig. 1). The overall conform-
ation of the ligand indicates that the lone pairs of all eight
nitrogens are directed towards the metal center. The Hg–N(NR )
distances of 2.957(6) Å are fairly long, but shorter than the sum
3
23
of the van der Waals radii (rvdw(N) = 1.60 Å,
r
vdw
(Hg) = 1.73–
2.00 Å).24 The Hg–Nimidazoyl bonds form a trigonal antiprism
and are significantly shorter, consistent with differences in
donor capacity, the steric demands of ligand architecture, co-
operative optimization of ligand–metal interactions and
adjustments for interactions between opposing ligands. The
A crystal measuring 0.40 × 0.63 × 0.59 mm was glued on the
end of a glass fiber. The data were collected using the θ–2θ
technique over a θ range of 2.03–27.50Њ. The final data to
parameter ratio was 14 : 1.
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2580

View Article Online
solutions of 1 in CD3CN, possibly due to the complexity
of solution equilibria at [Hg()]/[TMIMA] = 0.5 (vide supra;
Fig. S2). However, in the presence of excess ligand a 199Hg
resonance was detected at Ϫ1496 ppm. Previous studies have
suggested that ligand exchange for complexes of type HgL2
involves an associative exchange mechanism.9,12 The ligand
proton chemical shifts were linearly dependent on the metal-to-
ligand ratio for [Hg()]/[TMIMA] < 0.375 (Fig. S2). This indi-
cates that the lower metal-to-ligand ratio complex(es) involved
in ligand exchange with [Hg(TMIMA)2]2ϩ are kinetically
accessible but thermodynamically unstable. The position of the
199Hg peak was independent of the amount of excess ligand
present confirming the transient existence of lower metal-to-
ligand ratio complexes and permitting assignment of the 199Hg
peak observed to [Hg(TMIMA)2]2ϩ
.
Crystal structure and solution-state NMR of [Hg(TMIMA)-
(NCCH3)](ClO4)2 (2). The coordination sphere of Hg()
has three-fold symmetry in 2 (Fig. 2). The TMIMA ligand is
tetradendate and the acetonitrile is trans to the amine nitrogen,
thus forming a distorted trigonal bipyramid (C3v core sym-
metry, C3 ion symmetry). As expected for a lower coordination
number complex, the Hg–Nimidazoyl and Hg–N distances are
Fig. 1 Thermal ellipsoid plot of the [Hg(TMIMA)2]2ϩ cation of 1.
Ellipsoids are at 50% probability. Only one ligand has been completely
labeled.
shorter than those found in 1. However, in 2 the Hg–N(NR )
3
bond lengths are still significantly longer than the Hg–Nimidazoyl
bond lengths (0.444 Å) consistent with the reduced donor
capacity of aliphatic nitrogens in comparison with imidazoyl
nitrogens as well as the steric demands of ligand architecture.
Cooperative optimization of ligand–metal interactions with the
large Hg2ϩ ion also likely contributes to the elongation of the
Hg–Nimidazoyl bond distances range from 2.259(6) to 2.628(7) Å,
bracketing the Hg–N distance for the one structurally charac-
terized complex of Hg() with a simple imidazole derivative
(2.45 Å)25 and the range of Hg–N distances to histidine imid-
azoyl groups in Hg() substituted proteins (2.3–2.5 Å).7 Since
the previously reported complexes of Hg() with imidazoyl
derivatives are all four coordinate, longer Hg–Nimidazoyl bonds
are expected in 1. A reason for the short Hg–Nimidazoyl bond
length in 1 was not readily apparent.
metal-N(NR ) bonds in 2.
3
This is the first 1 : 2 metal-to-ligand complex of TMIMA to
be structurally characterized. In addition, there are very few
structurally characterized high coordination number complexes
of Hg() in which all the donor nitrogens are part of chelating
organic ligands. Previous investigation of Hg() coordination
chemistry with a potentially tetradentate tripodal nitrogen
ligand provided the complex [Hg(TMPA)2](ClO4)2 (4, TMPA =
tris[(2-pyridyl)methyl]amine).12 While the symmetries and
formal coordination numbers of 4 and 1 are similar, the
Hg–N(NR ) distance in 4 is 2.560(3) Å, slightly shorter than the
3
average Hg–Npyridyl distance of 2.583(8) Å. Hence, in Grdenic
terminology,26 1 is the first example of characteristic 6 co-
ordination and effective
8 coordination for a nitrogen
coordination compound of Hg().
The solution-state behavior of a dilute CD3CN solution of
complex 1 was also investigated for comparison with 4. In solu-
tion, 1 exhibits a single set of resonances for the four proton
types indicating effective three-fold symmetry through rapid
solution exchange processes which eliminate the differences
between the Hg–Nimidazoyl groups observed in the solid state.
Proton chemical shift properties were comparable for 1 and 4,12
including fluxionally exchanged methylene protons, shielding
of the protons closest to the binding aromatic nitrogens by the
ring current of the opposing ligand, and deshielding of remain-
ing ligand resonances by σ donation to the metal cation. How-
ever, ligand proton chemical shifts were inversely dependent on
the metal-to-ligand ratio below [Hg(ClO4)2]/[TMIMA] = 0.5
and then increased at higher metal-to-ligand ratio, reflecting
the precedented solution speciation of tripodal tetradentate
amines with Hg(ClO4)2 (ESI,†Fig. S2).9,12 Rapid intermolecular
exchange processes with free ligand and higher metal-to-ligand
ratio complexes precluded detection of J(HgH) for 1. In con-
trast, extensive heteronuclear coupling between ligand protons
and Hg was detected for 4 in the presence of excess ligand.
Additional investigations of 1 were conducted by solution-
state 199Hg NMR. No resonances were detected for 20–50 mM
Fig. 2 Thermal ellipsoid plot of the [Hg(TMIMA)(NCCH3)]2ϩ cation
of (2). Ellipsoids are at 50% probability. Only one (1-methylimidazol-2-
yl)methyl group has been completely labeled.
This structure can be compared to the one other reported
five-coordinate complex of TMIMA and other five-coordinate
complexes of Hg() involving potentially tetradentate amines,
as well as Hg() substituted proteins. The Cu–N(NR ) and aver-
3
age Cu–Nimidazoyl bond lengths in [Cu(TMIMA)Cl]PF6ؒ
CH3CN15 are 2.167 and 2.03 Å, respectively, a difference of
approximately 0.14 Å. In this copper complex, the axial Cl is
2.234 Å from the metal ion,15 comparable to the Hg–Nacetonitrile
bond distance in 2. The differences between Hg–N(NR ) and
3
Hg–Npyridyl distances of five-coordinate complexes involving
tripodal amines with pyridyl donors have been small as well. In
[Hg(TMPA)Cl]2(HgCl4)12 and [Hg(TLA)Cl]2(Hg2Cl6)10 (TLA =
tris[(2-(6-methylpyridyl))methyl]amine), the average Hg–N(NR )
3
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2581

View Article Online
bond length was smaller than the average Hg–Npyridyl bond
length by only 0.040 and 0.132 Å, respectively. The average Hg–
Cl distance is 2.355 Å and 2.366 Å, respectively, in these cations,
slightly longer than the Hg–Nacetonitrile bond distance in 2. The
2.198(8) Å Hg–Nimidazoyl bond length in 2 is below the range
of Hg–N distances to histidine imidazoyl groups in four-
coordinate Hg() substituted proteins (2.3–2.5 Å).7
peaks in the proton NMR spectrum precluding meaningful
characterization.
Crystal structure and solution-state NMR of [Hg(TMIMA)-
Cl]2HgCl4 (3). There are two identical five-coordinate [Hg-
(TMIMA)Cl]ϩ ions and a nearly tetrahedral [HgCl4]2Ϫ anion in
the asymmetric unit of 3 (Fig. 4). Although the Hg–N(NR ) dis-
tance of 2.758(7) Å is fairly long, it is significantly shorter3 than
the sum of the van der Waals radii23,24 and the overall conform-
ation of the ligand indicates that the lone pair on the aliphatic
nitrogen is directed towards the metal center for a bonding
interaction. The Hg() coordination environment is distorted
trigonal bipyramidal with the chloride trans to the aliphatic
nitrogen. The Hg–Cl distance of 2.424(2) Å in the cation
is slightly longer than the average Hg–Cl distances found in
the cations of related complexes [Hg(TMPA)Cl]2(Hg2Cl6)
(2.354(12) Å)12 and [Hg(TLA)Cl]2(Hg2Cl6) [2.366(8) Å].10
Although the overall geometry of 3 is very similar to 2, the
metal–ligand bond lengths are uniformly larger. The Hg–
Nimidzoyl bond lengths are similar (average 2.29(4) Å), bracketing
the lower end of the range of Hg–N distances to histidine imid-
azoyl groups in Hg() substituted proteins (2.3–2.5 Å).7
In CD3CN solution, 3 gave rise to a single set of (1-methyl-
imidazol-2-yl)methyl proton resonances consistent with
expected three-fold symmetry and exchange averaged methyl-
ene protons. The resonances were shifted downfield 0.17–0.53
ppm with respect to free ligand with generally the same trend as
observed for 2. The ligand chemical shifts were linearly depend-
ent on the metal-to-ligand ratio below [HgCl2]/[TMIMA] = 1.0
(ESI,†Fig. S3), indicating that a 1 : 1 metal-to-ligand complex
was the most thermodynamically stable TMIMA complex of
HgCl2. Although the metal-to-ligand ratio in 3 is 1.5 and the
neutral complex [Hg(TMIMA)Cl2] is potentially consistent
with the NMR data, isolation of 3 from a 1 : 1 metal-to-ligand
ratio solution indicates [Hg(TMIMA)Cl]ϩ is present under
these conditions. The later is consistent with the NMR assum-
ing both that ion pairing is negligible and that the cation is
formed by a dissociative process which is independent of the
presence of excess HgCl2 as opposed to an equilibrium of type
(1).
Complex 2 exhibited a single set of proton resonances for
each type of ligand proton in 2 mM CD3CN solutions indi-
cating preservation of three-fold symmetry and fluxionally
equivalent methylene protons. Ligand protons were deshielded
0.24–0.50 ppm through σ donation to Hg(). At room temper-
ature, dilute solutions of complex 2 exhibited the fast exchange
behavior on the coupling constant time-scale usually associated
with Hg() complexes due to the equilibrium formation of
complexes with lower metal-to-ligand ratios (ESI,†Fig. S2).
However, slow exchange conditions on the coupling constant
time-scale could be achieved by cooling the sample to Ϫ40 ЊC
(Fig. 3) or by addition of a 5% excess of Hg(ClO4)2 at 20 ЊC.
Coupling of 199Hg to both of the imidazoyl protons and
the methylene protons was apparent. Coupling satellites
had the same multiplicity as the parent resonance and appeared
to be approximately one-fifth the size of the main reson-
ance consistent with the 16.8% natural abundance of 199Hg.
Interestingly, the coupling expected to be generated predomin-
antly through a three-bond pathway to Ha was 17 Hz while the
coupling to Hb, for which only four- and five-bond coupling
pathways are available, was 20 Hz. Since J(HgH) are prece-
dented over four and five bonds for nitrogen coordination
compounds of Hg(),9 the larger longer range coupling may
reflect additive contributions from both potential pathways.
The methylene protons of 2 were also coupled to Hg with
J(HgH) 7 Hz. These heteronuclear coupling constants are
somewhat smaller that those recently reported for Hg() com-
plexes of other potentially tetradentate amines,9,10,12 but com-
parable in magnitude to those reported for Hg() substituted
proteins.4–6
2Ϫ
2[HgLCl2] ϩ HgCl2
2[HgLCl]ϩ ϩ HgCl4
(1)
The above equilibrium was evident with the more sterically
demanding ligand TLA. The neutral complex [Hg(TLA)Cl2], in
which one of the lutidyl arms is pendant, was prevalent at [Hg]/
[TLA] = 1 and [Hg(TLA)Cl]ϩ was prevalent at [Hg]/[TLA] >
1.5.10 Interestingly, with TMPA chemical shift discontinuities at
[HgCl2]/[TMPA] = 0.5 and 1.0 indicated that [Hg(TMPA)Cl]ϩ
was less thermodynamically stable than a 1 : 2 complex.12
At Ϫ40 ЊC, slow exchange conditions on the coupling con-
stant time-scale were found for 2 mM 3 in CD3CN (Fig. 3). As
in 2, comparable heteronuclear coupling constants were
observed between 199Hg and both imidazoyl protons. Hetero-
nuclear couplings of 10 and 11 Hz were detected to Ha and Hb,
respectively. Interestingly, 3 had a longer Hg–N(NR ) bond than
3
2, yet was associated with a stronger heteronuclear coupling
constant to the methylene protons (J(HgH) 11 Hz). The limited
solubility of 3 in CD3CN thwarted attempts to investigate it by
199Hg NMR.
Fig. 3 Proton NMR spectra recorded in CD3CN for nominally 2 mM
2 at Ϫ40 ЊC (upper) and nominally 0.8 mM 3 at Ϫ40 ЊC with 15% excess
HgCl2 added (lower). Singlet resonances without 199Hg coupling
satellites for the methyl protons at 3.55 ppm for both complexes are not
shown. Magnitudes of J(HgH) are indicated.
Discussion
Tetradentate binding modes for TMIMA were found for three
complexes of Hg() in the solid state. Although related tripodal
amines with pyridyl donors were observed to have comparable
Hg–N distances for all donors, the TMIMA complexes gener-
ally had significantly longer distances to the aliphatic nitrogen
than the imidazoyl nitrogens. The same trend has been observed
with the five previously reported complexes of TMIMA,
Characterization of 2 by 199Hg NMR was attempted with 20–
68 mM CD3CN solutions. These samples did not exhibit slow
exchange on the J(HgH) time-scale at Ϫ40 ЊC. Furthermore,
addition of excess Hg(ClO4)2 to these solutions resulted in for-
mation of a white precipitate and appearance of additional
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2582

View Article Online
Fig. 4 Thermal ellipsoid plot of [Hg(TMIMA)Cl]2(HgCl4) (3). Ellipsoids are at 50% probability. Only one [Hg(TMIMA)Cl]ϩ has been completely
labeled.
although the differences have been less substantial,13–15
reflecting differences in ion size and hardness. The shorter Hg–
Nimidazoyl bonds are consistent with the stronger σ and π donor
ability of imidazoyl nitrogen compared to pyridyl and aliphatic
nitrogen. Furthermore, the metal chelate ring C–N double
bond is slightly shorter with imidazoyl donors than pyridyl
donors leading to additional architectural constraints on chel-
ation. Stronger binding to the aromatic amines enhances the
biorelevance of TMIMA in comparison to related pyridyl lig-
ands since histidine is much more common in protein metal
binding sites than lysine.
Although correlations between solid-state structures and
solution-state NMR properties must always be made care-
fully,27 the symmetry of tripodal ligands facilitates inter-
pretation of spectra within the context of crystallographic
structures. In CD3CN solution, each complex gave rise to
a single set of four proton resonances indicating three-
fold symmetry on the chemical shift time-scale. Although 1
and 3 are not three-fold symmetric in the solid-state, packing
effects that disrupt complex symmetry would be absent in
solution.
characterization by 199Hg NMR spectroscopy due to solubility
related issues.
Although slow exchange conditions on the J(HgH) time-
scale were inaccessible for 1 due to rapid ligand exchange, the
complex was characterizable by solution-state 199Hg NMR.
Solid-state 199Hg NMR is preferable for characterization of
Hg() coordination compounds to avoid exchange compli-
cations.5 However, the concentration independent Ϫ1496 ppm
199Hg chemical shift of 1 in the presence of excess ligand sup-
ported the structural assignment. The 199Hg chemical shift of 1
is approximately 450 and 250 ppm upfield of the 199Hg chemical
shifts for [Hg(N(SiMe3))2] and [Hg(ethylenediamine)2]2ϩ
,
respectively, the only nitrogen coordination compounds of
Hg() for which 199Hg chemical shifts have apparently been
reported.28,5 Interestingly, the 199Hg chemical shift of 4 is Ϫ1996
ppm in CD3CN at 20 ЊC.29 Both 1 and 4 are formally eight
coordinate but in 1 the axial Hg–Namine bonds are significantly
longer than the Hg–Nimidazoyl bonds while in 4 all Hg–N are
comparable. Increases in coordination number are associated
with upfield shifts in 199Hg resonance positions with other
ligand types.28
The trends in proton chemical shift as a function of the
metal-to-ligand ratio complemented structure assignments. The
observed trends for the perchlorate system (ESI,†Fig. S2) were
consistent with the series of linked equilibria established for
related tetradentate amines,9,12 in the context of the lower steric
demands of TMIMA. For the chloride system, the proton
NMR behavior was consistent with the prevalence of either
[Hg(TMIMA)Cl2] or [Hg(TMIMA)Cl]ϩ in solution, assuming
ion pairing was negligible for the later.
Slow exchange conditions on the J(HgH) time-scale were
found for the two lower coordination number complexes of
TMIMA. In general, the heteronuclear coupling constants
observed for TMIMA complexes were smaller than those
observed with related pyridyl ligand systems and more compar-
able to those observed for Hg() substituted proteins. There are
many contributing factors to the magnitudes of indirect nuclear
spin–spin coupling constants including electronegativity of
substituents, hybridization, bond angles and bond lengths.
Coupling between 199Hg and nuclei up to five bonds away has
been observed in Hg() coordination compounds with nitrogen
ligands9 (seven bonds away in alkylmercurials28). Although
there is no standard theory to explain long-range couplings, it is
generally accepted that coupling constants are additive, which
can lead to significant enhancements if contributions are of the
same sign. Additivity over as many as five bonds is proposed to
account for the comparable magnitudes of 199Hg coupling con-
stants to Ha and Hb. Neither 2 nor 3 proved to be suitable for
Conclusions
Generally ligands with pyridyl donors have been exploited
much more extensively than ligands with imidazoyl donors in
bioinorganic model studies, in part due to differences in ease of
synthesis and stability. However, there are clearly significant
differences in electronics and sterics of these aromatic amines.
The three structures reported in this work highlight the differ-
ences in donor capacity of imidazoyl and aliphatic amine
ligands as well as the demanding sterics of five-membered
chelate rings involving imidazoyl donors. Additional studies
of TMIMA and related ligands with physiologically essential
metal ions would complement the existing literature.
Heteronuclear coupling is not routinely observed between
199Hg and protons of small organic ligands in solution, even
with multidentate ligands. However, each of the tripodal ligand
systems with pyridyl and imidazoyl donors we have investigated
has permitted detection of J(HgH) for two or more specific
complexes. The magnitudes of J(HgH) observed with TMIMA
were more comparable to the magnitudes of J(HgH) observed
in Hg() substituted proteins.4–6 These results encourage
investigation of the Hg() coordination chemistry of additional
tripodal ligand systems involving imidazoyl donors, as well as
ligand systems with more diverse donors. In addition, tripodal
ligand systems may advance solution-state NMR studies of
other high natural abundance spin I = ½ nuclei, such as
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2583

View Article Online
109Ag, for which heteronuclear coupling with protons of organic
ligands are rarely observed.
9 D. C. Bebout, J. F. Bush II, K. K. Crahan, E. V. Bowers and
R. J. Butcher, Inorg. Chem., 2002, 41, 2529–2536.
10 D. C. Bebout, J. F. Bush II, K. K. Crahan, M. E. Kastner and
D. A. Parrish, Inorg. Chem., 1998, 37, 4641–4646.
11 D. C. Bebout, A. E. DeLanoy, D. E. Ehmann, M. E. Kastner,
D. A. Parrish and R. J. Butcher, Inorg. Chem., 1998, 37, 2952–
2959.
12 D. C. Bebout, D. E. Ehmann, J. C. Trinidad, K. K. Crahan,
M. E. Kastner and D. A. Parrish, Inorg. Chem., 1997, 36, 4257–4264.
13 K. J. Oberhausen, R. J. O’Brein, J. F. Richardson, R. M. Buchanan,
R. Costa, J.-M. Latour, H.-L. Tsai and D. N. Hendrickson,
Inorg. Chem., 1993, 32, 4561–4565.
14 J. Wang, M. S. Mashuta, Z. Sun, J. F. Richardson, D. N.
Hendrickson and R. M. Buchanan, Inorg. Chem., 1996, 35, 6642–
6643; R. M. Buchanan, S. Chen, J. F. Richardson, M. Bressan,
L. Forti, A. Morvillo and R. H. Fish, Inorg. Chem., 1994, 33, 3208–
3209; R. H. Fish, M. S. Konings, K. J. Oberhausen, R. H. Fong,
W. M. Yu, G. Christou, J. B. Vincent, D. K. Coggin and
R. M. Buchanan, Inorg. Chem., 1991, 30, 3002–3006.
15 K. J. Oberhausen, R. J. O’Brien, J. F. Richardson and R. M.
Buchanan, Inorg. Chim. Acta, 1990, 173, 145–154.
16 W. C. Wosley, J. Chem. Educ., 1973, 50, A335; K. N. Raymond,
Chem. Eng. News, 1983, 61(49), 4.
Acknowledgements
We thank Monte Helm for helpful discussions regarding 199Hg
NMR and Robert Hinkle for assisting with the NOE experi-
ments to confirm proton assignments. This research was
supported by a Bristol–Meyers Squibb Company award of
Research Corporation and NIH AREA award number
1R15GM59043–01. E. V. B. was supported in part by a Howard
Hughes Medical Institute grant through the Undergraduate
Biological Sciences Education Program to the College of
William and Mary and a Beckman Foundation Research
Scholarship. M. M. G. was supported in part by a Batten
Scholarship for Pre-Honors Research Summer Fellowship from
the College of William and Mary. The NSF chemical instru-
mentation program partially funded purchase of the NMR
instrumentation. R. J. B. acknowledges the DoD-ONR instru-
mentation program for funds to upgrade the Siemens P4
diffractometer and the NIH-MBRS program for funds to
maintain this diffractometer. We are thankful to Dr Travis
Holman for permitting us to use a Bruker SMART 1K CCD for
diffraction analysis of 3.
17 K. J. Oberhausen, J. F. Richardson, R. M. Buchanan and W. Pierce,
Polyhedron, 1989, 8, 659–668.
18 M. Rodriguez, I. Morgenstern-Badarau, M. Cesario, J. Guilhem,
B. Keita and L. Nadjo, Inorg. Chem., 1996, 35, 7804–7810.
19 M. Borzo and G. E. Maciel, J. Magn. Reson., 1975, 19, 279–
282.
20 G. M. Sheldrick, SADABS, Program for Empirical Absorption
Correction of Area Detector Data; University of Göttingen,
1996.
21 G. M. Sheldrick, SHELXS 97, Program for Crystal Structure
Solution; University of Göttingen, 1997.
References
1 M. F. Summers, Coord. Chem. Rev., 1988, 86, 43–134; P. D. Ellis,
Science, 1983, 221, 1141–1146.
2 L. M. Utschig, J. G. Wright, G. Dieckmann, V. Pecoraro and
T. V. O’Halloran, Inorg. Chem., 1995, 34, 2497–2498.
3 D. C. Bebout, S. W. Stokes and R. J. Butcher, Inorg. Chem., 1999, 38,
1126–1133.
4 P. R. Blake, B. Lee, M. F. Summers, J.-B. Park, Z. H. Zhou and
M. W. W. Adams, New J. Chem., 1994, 18, 387–395.
5 L. M. Utschig, J. W. Bryson and T. V. O’Halloran, Science, 1995,
268, 380–385.
22 G. M. Sheldrick, SHELXL 97, Program for Crystal Structure
Refinement; University of Göttingen, 1997.
23 S. C. Nyburg and C. H. Faerman, Acta Crystallogr., Sect. B, 1985,
41, 274–279.
24 A. J. Canty and G. B. Deacon, Inorg. Chim. Acta, 1980, 45, L225–
L227.
25 F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard,
C. F. Macrae, E. M. Mitchell, G. F. Mitchell, J. M. Smith and
D. G. Watson, J. Chem. Inf. Comput. Sci., 1991, 31, 187–204;
Z. Popovic, D. Matkovic-Calogovic, Z. Soldin, G. Pavlovic,
N. Davidovic and D. Vikic-Topic, Inorg. Chim. Acta, 1999, 294,
35–46.
26 D. Grdenic, Quart. Rev., 1965, 19, 303–328.
27 J. A. Davies and S. Dutremez, Coord. Chem. Rev., 1992, 114, 201–
247.
6 L. M. Utschig, T. Baynard, C. Strong and T. V. O’Halloran, Inorg.
Chem., 1997, 36, 2926–2927.
7 F. X. Gomis-Rueth, F. Grams, I. Yiallouros, H. Nar, U. Kuesthardt,
R. Zwilling, W. Bode and W. Stoecker, J. Biol. Chem., 1994,
269, 17111–17117; W. B. Church, J. M. Guss, J. J. Potter and
H. C. Freeman, J. Biol. Chem., 1986, 261, 234–237.
8 D. C. Rees, J. B. Howard, P. Chakrabarti, T. Yeats, B. T. Hsu,
K. D. Hardman and W. N. Lipscomb, in Zinc Enzymes,
ed. I. Bertini, C. Luchinat, W. Maret and M. Zeppezauer, Birkhäuser
Verlag, Basel, 1986, p. 155.
28 B. Wrackmeyer and R. Contreras, Annu. Rep. NMR Spectrosc.,
1992, 24, 267–329.
29 D. C. Bebout, unpublished result.
D a l t o n T r a n s . , 2 0 0 3 , 2 5 7 8 – 2 5 8 4
2584