W0031

Identification of Metal Atoms in Protein Structures using the Bond-Valence Method. Peter Müller, Sinje Köpke and George M. Sheldrick, UCLA-DOE-Laboratory, 201 MBI, Box 95157, Los Angeles, CA, 90095-1570, peterm@mbi.ucla.edu.

Inspired by the ‘bond number concept’ of L. Pauling [1], N. E. Brese and M. O’Keeffe developed a formula to relate bond lengths and bond valences [2, 3]. Thereby the valence, v_ij of a bond between two atoms i and j is given as a function of the bond length d_ij:

v_ij = exp[(d₀ - d_ij)/b]

Here d₀ is the so called ‘bond-valence parameter’, describing the expected bond length of a perfect single bond between the atoms i and j. The factor b is a more or less ‘universal constant’ equal to 0.37 Å [4]. The bond valences of all bonds from an atom i sum up to the valence V_i of that atom. For ions V_i is equal to the oxidation state:

V_i = (_jv_ij

This bond-valence concept is useful in a number of ways, e.g. in predicting bond lengths from a given bond valence or in calculating the oxidation state of an atom. Another beneficial use is the identification of metal ions in protein structures: By calculating and summing up over the bond valences of an unidentified atom, which is coordinated e.g. to some O-atoms one derives the hypothetical valence of that atom. Performing this calculation with several d₀-values leads to a list of atom valences for several different metals.

For most of the metals frequently present in protein structures only one oxidation state is reasonable (e.g. 1 for potassium or sodium, and 2 for calcium). A comparison of the calculated atom valences with the expected values of different possible ions normally leads to a clear answer to the question ‘do I have a metal ion on that position and, if yes, what type of?

1. Pauling, L., J. Am. Chem. Soc. 69 (1947) 542-553.

2. Brese, N. E. & O’Keeffe, M., Acta Cryst. B47 (1991) 192-197.

3. O’Keeffe, M. & Brese, N. E., J. Am Chem. Soc. 113 (1991) 3226-3229.

4. Brown, I. D. & Altermatt, D., Acta Cryst. B41 (1985) 244-247.