2952
J. Phys. Chem. B 2009, 113, 2952–2960
Factors Governing Metal-Ligand Distances and Coordination Geometries of Metal Complexes Gopi Kuppuraj,†,§ Minko Dudev,† and Carmay Lim*,†,‡ Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan, and Department of Chemistry and College of Life Sciences, National Tsing-Hua UniVersity, Hsinchu 300, Taiwan ReceiVed: September 8, 2008; ReVised Manuscript ReceiVed: NoVember 7, 2008
The metal-ligand (M-L) distances play a central role in determining the structure and reactivity of the metal complex and in metal discrimination in proteins. They reflect properties of the metal ion and the ligand during their coordination as well as the environment. However, the variation of the M-L distances as a function of the properties of the metal cation and its donor atoms had not been systematically analyzed, and no public website listing the M-L distances from metal complexes was available. Herein, the distances around 63 different types of metal ions in ∼200 000 high-resolution crystal structures in the Cambridge Structural Database as well as their preferred coordination geometries have been determined. The dependence of these distances on (i) the donor atom’s charge and charge-donating ability, (ii) the donor atom’s size, (iii) the metal ion’s oxidation state and charge-accepting ability, and (iv) the metal ion’s size has also been determined. The concept of ligand coordination number (CN) was introduced to take into account the effect of the number of constituent ligand atoms on the M-L distances in addition to the number of metal-bound ligands. We propose grouping the M-L distances according to both the metal CN and the ligand CN. The mean M-L distance corresponding to a given metal CN and ligand CN was found to be linearly correlated with the metal’s ionic radius in going down a main group or in going across the first three rows of the periodic table. The M-L distances as a function of the metal and ligand coordination numbers are available via http://bioit.ibms.sinica.edu.tw/CSD.htm Introduction Metal ions offer a multitude of functions in biological macromolecules.1 They are essential cofactors in proteins, providing structural rigidity and/or stability for performing crucial biological processes.2-5 The metal-ligand (M-L) distances reflect properties of the metal complex, namely, the electronic distributions of the metal and the ligands, their polarizing abilities and interactions, as well as the metal coordination number (CN) and coordination geometry,6 which are key factors governing the structure and reactivity of a metal complex.7,8 On the basis of statistical analyses of metal complex structures in the Cambridge Structural Database (CSD),9 the metal CN was found to be dictated by the size of the metal ion and the charge-transfer tendency of the ligand.10 A subsequent theoretical study on group IA and IIA metal hydrates revealed the physical basis for the trends in the metal hydration number with increasing size of the metal ions.11 Herein, the statistical analyses of metal complex structures are extended to the metal coordination geometries and the M-L distances, which are analyzed in terms of the properties of the metal ion and its ligand donor atoms. The M-L distances play a central role in metal discrimination in proteins.12,13 Native metal ions fulfill the best “fit” conditions in metalloproteins, and when substituted with an alien ion, their structures adopt different arrangements.14-16 Studies have shown that heavy-metal ions such as Cd2+ can replace the native Zn2+ * To whom correspondence should be addressed. E-mail: carmay@ gate.sinica.edu.tw. † Academia Sinica. ‡ Department of Chemistry, National Tsing Hua University. § College of Life Sciences, National Tsing Hua University.
cofactor from Zn-finger motifs, but they cannot maintain the requisite conformation for the DNA-binding function of Znfinger proteins.17 For example, Cd-bound proteins, though tetrahedrally coordinated like Zn2+, cannot maintain the native conformation because of the increase in the Cd-N (2.3-2.5 Å) and Cd-S (2.6 Å) distances, compared to the Zn-N (2.1 Å) and Zn-S (2.3 Å) distances,18 respectively. In contrast to cadmium, Hg2+ and Pb2+ adopt totally different nontetrahedral geometries when bound to proteins.19,20 The characteristic M-L distances and metal coordination geometry have not only revealed plausible mechanisms of heavy-metal poisoning in cells, but they have been used to distinguish between a catalytic versus a structural role for Zn2+ in proteins.21 This is especially useful in cases where the Zn ligands are identical, but Zn2+ plays a catalytic role in one protein (catalytic Zn site) but a structural role in another (structural Zn site). For example, Zn2+ plays a catalytic role in human carbonic anhydrase II but a solely structural role in interferon-β, even though both proteins contain a common [Zn(His)3H2O]2+ core. This implies that the protein matrix rather than the Zn ligands dictates the specific role of the Zn2+. The effect of the protein matrix is reflected in the Zn-L distances in the Protein Data Bank (PDB)22 structures and can be assessed by comparing the latter with those in the CSD structures. The mean Zn-L distances corresponding to a given metal CN in catalytic Zn sites are generally longer than those in structural Zn sites and CSD structures; this creates space for binding another ligand during the enzymatic reaction.21 For example, the mean Zn-N distance in human carbonic anhydrase II (Zn-N ) 2.11 Å, PDB entry 2cba) is longer than that in
10.1021/jp807972e CCC: $40.75 2009 American Chemical Society Published on Web 02/06/2009
Metal-Ligand Distances and Coordination Geometries interferon-β (Zn-N ) 1.88 Å, PDB entry 1au1). The Zn-L distances have been used to identify Zn’s role by computing bond valence sum values,23,24 as described in our previous work.21 It is thus important to elucidate the trends in the M-L distances and the key factors governing these distances and the metal coordination geometries. This can be achieved by surveying the CSD,9 which currently incorporates roughly 400 000 high-resolution crystal structures containing around 100 different types of metal ions in the periodic table. Indeed, previous workers have carried out such statistical analyses, but they were limited to certain metal ions and ligand types.25-27 To aid interpretation of electron density maps in newly resolved metalloprotein structures, Harding28-31 has analyzed various M-L distances in the CSD and compared them with the respective distances in e1.25 Å PDB structures to provide “target distances” for biogenic metal ions and their ligands. Rulı´sˇek and Vondrasek18 have determined the preferred coordination geometries of six dications (Co2+, Ni2+, Cu2+, Zn2+, Cd2+, and Hg2+) from CSD structures in order to compare them with those derived from the respective PDB structures. Zheng et al.32 have computed the mean M-L distances in medium(between 2.0 and 2.5 Å resolution) and high-resolution (
