Two-Angstrom Crystal Structure of Oxidized Chromatium High Potential Iron Protein

Abstract

Abstract The 85-residue structure of Chromatium high potential iron protein (HiPIP) has been determined by x-ray diffraction methods at 2.0 A resolution, and is currently undergoing crystallographic refinement. The partially refined HiPIP structure reported here has an R factor of 0.24. Moreover, bond distances and angles have been constrained to their expected values, so that a model may be constructed from standard brass parts. Thus, it is an unusually well determined protein structure. The Fe4S*4 cluster (Carter, C. W., Jr., Freer, S. T., Xuong, Ng. H., Alden, R. A., and Kraut, J. (1971) Cold Spring Harbor Symp. Quant. Biol. 36, 381–385) is covalently attached to the protein by Fe—S bonds at cysteine residues 43, 46, 63, and 77. Residues 1–42 fold up upon this cluster binding segment. Most of the interface between these NH2- and COOH-terminal segments consists of packed nonpolar side chains. The chain segment containing Cys 77 is trapped inside this interface where it forms a stretch of antiparallel β sheet with residues 17–20 in the NH2-terminal half of the chain. This unusual architecture probably accounts for the relatively high stability of HiPIP (Dus, K., DeKlerk, H., Sletten, K., and Bartsch, R. G. (1967) Biochim. Biophys. Acta 140, 291–311). The Tyr 19 side chain abuts the Fe4S*4 cluster in a manner similar to that observed for Tyr 2 and Tyr 28 in Peptococcus aerogenes ferredoxin (Adman, E. T., Sieker, L. C., and Jensen, L. H. (1973) J. Biol. Chem. 248, 3987–3996). The polypeptide chain conformation may be described, almost in its entirety, as a sequence of α helical or extended conformations and hairpin turns. Moreover, most of these secondary structures are surprisingly close in their detailed geometry to those predicted to be of lowest energy. More than 75% of the main chain hydrogen bonding sites are bonded either within secondary structures (47%) or to water (30%). Thus these bonds may be formed prior to complete assembly of the entire structure. Hairpin turns are conspicuously associated with the protein-water interface: they contain a large majority of the polar side chains and their main chain carbonyl oxygen and amido nitrogen atoms bind a disproportionately large fraction of the fixed water molecules. Sequence-structure correlations of the type proposed by Robson and Pain (Robson, B., and Pain, R. H. (1971) J. Mol. Biol. 58, 237–259) and Lewis et al. (Lewis, P. N., Momany, F. A., and Scheraga, H. A. (1971) Proc. Nat. Acad. Sci. U. S. A. 68, 2293–2297) turn out to be quite accurate, and the most serious discrepancy between predicted and observed structures can be rationalized convincingly in terms of Fe4S*4 cluster binding requirements. These observations appear to weigh heavily against theories of protein folding which involve early, nonspecific micelle formation (Robson, B., and Pain, R. H. (1971) J. Mol. Biol. 58, 237–259).