The crystal structure of a host defense system chemotactic factor, interleukin 8, has been solved by molecular replacement using as a model the solution structure derived from nuclear magnetic resonance experiments. The structure was refined with 2 A x-ray data to an R factor of 0.187 (0.217 at 1.6 A). A comparison indicates some potential differences between the structure in solution and in the crystalline state. Our analysis also predicts that residues 4 through 9 on the amino terminus and the (-bend, which includes His-33, may be important for receptor binding.Nuclear magnetic resonance (NMR) has been established in the last 5 years as the only viable alternative to x-ray crystallography for determining detailed three-dimensional structures of macromolecules up to 20 kDa molecular mass (1, 2). Although test calculations using crambin (3), tendamistat (4), and ubiquitin (5) have established the feasibility of utilizing NMR-derived models as starting points for solving x-ray structures by molecular replacement, no unknown crystal structures have been previously solved by that route. We present here the results of such an investigation. The highresolution x-ray structure ofinterleukin 8 (IL-8) was solved by a combination of NMR (6, 7) and crystallography, and the models obtained by these two techniques were compared.tt IL-8 is a small, chemotactic factor released by monocytes and a number of other cell types in response to inflammatory stimuli (8). The factor promotes the directed migration of neutrophils, basophils, and T cells (9) along a concentration gradient to the inflammatory focus. IL-8 has also been reported to initiate the oxidative burst response in neutrophils. The factor has been known as monocyte-derived neutrophil chemotactic factor (10) or neutrophil-activating protein (11,12) among others, but it is now generally designated IL-8 (9). IL-8 is a member of a family of similar immune system proteins including the monocyte chemoattractant factor (13), human melanoma growth-stimulating activity (14), murine macrophage inflammatory protein 2 (15), and bovine platelet factor 4 (BPF4) (16). Each of these protein sequences shows a conserved region containing four cysteines that form two disulfide bridges. Under the experimental conditions of the structural investigations IL-8 is a dimer (6, 7) and BPF4 is a tetramer (17). Materials and MethodsRecombinant human IL-8 that was used in the crystallographic studies was purified from Escherichia coli (18). All of the x-ray data used for this structure determination were collected on crystals that were grown using protein recycled from the NMR experiments of Clore and Gronenborn and coworkers (6, 7). Crystals belonging to the space group P3121, a = b = 40.3 A, c = 90.1 A, were grownfrom40%o ammonium sulfate (pH = 8.5) as described (19). Identical crystals have also been grown using chemically synthesized protein. Recently, workers at Sandoz Pharmaceutical have reported that crystals apparently identical to the ones described here could be grown from poly...

The glycolytic enzyme glucokinase plays an important role in the regulation of insuln secretion and recent studies have shown that mutations in the human glucokinas gene are a common cause of an autosomal domint form of non-inssulin-dependent (type 2) diabetes meilitus (NIIDDM) that has an onset often during childhood. The majority of the mutations that have been identified are missense mutations that result in the syntheis of a glucokinas molecule with an altered amino acid sequence. To characterize the effect of these mutations on the catalytic properties of human 3-cell glucokinase, we have expressed native and mutant forms of this protein in Eschenchia coli. AU of the missense mutations show chnges in enzyme activity including a decrease in V.. and/or increase in K. for glucose. Using a model for the threedimensional structure of human glucokinas based on the crystal structure of the related enzyme yeast hexokinae B, the mutations map primarily to two regions of the protein. One group of mutations is lcated in the active site cleft separating the two domains of the enzyme as wefl as in surface loops leading into this cleft. These mutations usually result in large reductions in enzyme activity. The second group of mutations is located far from the active site in a region that is predicted to undergo a substrate-induced conformational change that results in closure of the active site cleft. These mutations show a smail -2-fold reduction in V., and a 5-to 10-fold increase in K. for glucose. The characterization of mutations in glucokinase that are asted with a distinct and readily recognizable form of NIDDM has led to the identification of key amino acids involved in glucokinas catalysis and locaized functionaUly important regions of the glucokinas molecule.

A model of heparin bound to bovine platelet factor 4 (BPF4) was completed using a graphically designed heparin molecule and the crystallographic coordinates of the native bovine platelet factor 4 tetramer. The oligosaccharides had a chain length of at least eight disaccharide units with the major repeating disaccharide unit consisting of (1----4)-O-(alpha-L-idopyranosyluronic acid 2-sulfate)-(1----4)-(2-deoxy-2-sulfamino-2-D-glucopyranosyl 6-sulfate). Each disaccharide unit carried a -4.0 charge. The structure of BPF4 was solved to 2.6 A resolution with R = 0.237. Each monomer of BPF4 contains an alpha-helix lying across 3 strands of antiparallel beta-sheet. Each helix has four lysines, which have been implicated in heparin binding. These lysine residues are predominantly on one side of the helix and are solvent accessible. Electrostatic calculations performed on the BPF4 tetramer show a ring of strong, positive charge which runs perpendicularly across the helices. Included in this ring of density is His-38, which has been shown by NMR to have a large pKa shift when heparin binds to BPF4. Our model of heparin bound to PF4 has the anionic polysaccharide perpendicular to the alpha-helices, wrapped about the tetramer along the ring of positive charge, and salt linked to all four lysines on the helix of each monomer.

No abstract