Studies on the antigenic determinants in the self-association of IgG rheumatoid factor.

The number, location, and other characteristics of the antigenic determinants for self-association of IgG-rheumatoid factors (IgG-RF) were examined using the IgG-RF isolated from the plasma of one patient as a model system. Affinity chromatography was employed for isolation of the IgG-RF. Sedimentation equilibrium ultracentrifugation was used to study the various interactions. The antigenic valence of IgG-RF Fc, normal human Fc, and rabbit Fc fragments was two for the interaction with Fab fragments from IgG-RF, as might be expected from the molecular symmetry of IgG. The antigenic valence of intact normal IgG, however, was only one, indicating that when one of the available antigenic determinants interacted with the Fab fragment of IgG-RF, the other determinant becomes sterically inaccessible. Reduction and alkylation, known to increase the flexibility of the hinge region, did not alter the antigenic valence of IgG for Fab fragments of IgG-RF. The antigenic valence of IgG-RF in self-association could not be experimentally determined but must be two to permit the observed concentration-dependent further polymer formation of IgG-RF dimers. Unique antigenic determinants on the Fc fragments of IgG-RF were sought and not found, thus reaffirming the formation of two antigen-antibody bonds as the basis for dimerization of IgG-RF molecules. The pFc' and Fc' fragments, representing Cgamma3 domains of IgG, failed to show significant interaction with Fab fragments of IgG-RF, indicating that the antigenic determinants were not expressed by the Cgamma3 regions but are located either on Cgamma2 region or require intact Cgamma2 and Cgamma3 regions for expression. These conclusions were corroborated by the antigenic valence of one for the Fc(i) fragment, a new papain-generated intermediate fragment of Fc, composed of two intact Cgamma3 domains and one intact Cgamma2 domain. Normal IgG, because of its valence of one for interaction with IgG-RF, would effectively terminate further polymerization of IgG-RF dimers. This may well in part explain the finding of smaller IgG-RF complexes in the serum than in synovial fluid of patients with rheumatoid arthritis.

IgG rheumatoid factors and staphylococcal protein A bind to a common molecular site on IgG.

The antigenic determinant on the Fc region of human IgG for two IgG rheumatoid factors (IgG-RF) from patients with rheumatoid arthritis were investigated in detail. The RF did not interact with IgG fragments that contained the C gamma 2 or C gamma 3 region alone, but required the presence of both regions for binding. The RF binding to solid-phase IgG were poorly inhibited by the IgG3 subclass and strongly inhibited by staphylococcal protein A (SPA) (42 kD), and fragment D of SPA (7 kD), indicating that the binding site is most likely the same as the Ga antigenic determinant described for IgM-RF, and is in the same location as the site on IgG that binds SPA. pH titration studies of the RF binding to IgG indicated the involvement of histidine and lysine or tyrosine side chains. Chemical modification studies showed the histidines were involved on the Fc side of the interactions, and tyrosines were involved on both the antigenic and antibody sides of the interactions. Lysines were not involved. The above information, and the knowledge of the number and position in space of the amino acid residues involved in the C gamma 2-C gamma 3 interface region of IgG, the binding site for SPA, and the amino acid substitutions in IgG3 that account for its inability to bind protein A, allowed the identification of the site on IgG that bind IgG-RF. This binding site involves some of the same amino acid side chains, His 435, Tyr 436, and one or both His 433 and 310, and is in the same location as the site that binds SPA. The same site is likely to be a common antigenic determinant for other RF. Furthermore, the described molecular mimicry suggests a biological relationship between bacterial Fc-binding proteins and the production of RF in rheumatoid arthritis.

The 2.3 angstrom X-ray structure of nitrite reductase from Achromobacter cycloclastes.

The three-dimensional crystal structure of the copper-containing nitrite reductase (NIR) from Achromobacter cycloclastes has been determined to 2.3 angstrom (A) resolution by isomorphous replacement. The monomer has two Greek key beta-barrel domains similar to that of plastocyanin and contains two copper sites. The enzyme is a trimer both in the crystal and in solution. The two copper atoms in the monomer comprise one type I copper site (Cu-I; two His, one Cys, and one Met ligands) and one putative type II copper site (Cu-II; three His and one solvent ligands). Although ligated by adjacent amino acids Cu-I and Cu-II are approximately 12.5 A apart. Cu-II is bound with nearly perfect tetrahedral geometry by residues not within a single monomer, but from each of two monomers of the trimer. The Cu-II site is at the bottom of a 12 A deep solvent channel and is the site to which the substrate (NO2-) binds, as evidenced by difference density maps of substrate-soaked and native crystals.

Crystal structure of rhodopsin: A G protein-coupled receptor.

Heterotrimeric guanine nucleotide-binding protein (G protein)-coupled receptors (GPCRs) respond to a variety of different external stimuli and activate G proteins. GPCRs share many structural features, including a bundle of seven transmembrane alpha helices connected by six loops of varying lengths. We determined the structure of rhodopsin from diffraction data extending to 2.8 angstroms resolution. The highly organized structure in the extracellular region, including a conserved disulfide bridge, forms a basis for the arrangement of the seven-helix transmembrane motif. The ground-state chromophore, 11-cis-retinal, holds the transmembrane region of the protein in the inactive conformation. Interactions of the chromophore with a cluster of key residues determine the wavelength of the maximum absorption. Changes in these interactions among rhodopsins facilitate color discrimination. Identification of a set of residues that mediate interactions between the transmembrane helices and the cytoplasmic surface, where G-protein activation occurs, also suggests a possible structural change upon photoactivation.

Crystal structure of a 30 kDa C-terminal fragment from the gamma chain of human fibrinogen.

BACKGROUND: Blood coagulation occurs by a cascade of zymogen activation resulting from minor proteolysis. The final stage of coagulation involves thrombin generation and limited proteolysis of fibrinogen to give spontaneously polymerizing fibrin. The resulting fibrin network is covalently crosslinked by factor XIIIa to yield a stable blood clot. Fibrinogen is a 340 kDa glycoprotein composed of six polypeptide chains, (alphabetagamma)2, held together by 29 disulfide bonds. The globular C terminus of the gamma chain contains a fibrin-polymerization surface, the principal factor XIIIa crosslinking site, the platelet receptor recognition site, and a calcium-binding site. Structural information on this domain should thus prove helpful in understanding clot formation. RESULTS: The X-ray crystallographic structure of the 30 kDa globular C terminus of the gamma chain of human fibrinogen has been determined in one crystal form using multiple isomorphous replacement methods. The refined coordinates were used to solve the structure in two more crystal forms by molecular replacement; the crystal structures have been refined against diffraction data to either 2.5 A or 2.1 A resolution. Three domains were identified in the structure, including a C-terminal fibrin-polymerization domain (P), which contains a single calcium-binding site and a deep binding pocket that provides the polymerization surface. The overall structure has a pronounced dipole moment, and the C-terminal residues appear highly flexible. CONCLUSIONS: The polymerization domain in the gamma chain is the most variable among a family of fibrinogen-related proteins and contains many acidic residues. These residues contribute to the molecular dipole moment in the structure, which may allow electrostatic steering to guide the alignment of fibrin monomers during the polymerization process. The flexibility of the C-terminal residues, which contain one of the factor XIIIa crosslinking sites and the platelet receptor recognition site, may be important in the function of this domain.

Direct determination of the stoichiometry of charge transfer complexes.

The stoichiometry of the charge transfer complex between N-acetyl-L-tryptophan and 1-methylnicotinamide chloride has been determined to be precisely 1:1 by direct measurement of the molecular weight of the complex. The result is of interest both in terms of a general method for determining the stoichiometry of charge transfer complexes, and in terms of the probable stoichiometry of specific charge transfer complexes between 1-methylnicotinamide chloride and the exposed tryptophyl side chains of certain proteins. In the latter case, the result provides experimental proof for the assignment of extinction coefficients of specified magnitudes to the homomorphic model complexes which serve as the basis for the interpretation of results with proteins.

Crystal structure of rhodopsin: a template for cone visual pigments and other G protein-coupled receptors.

The crystal structure of rhodopsin has provided the first three-dimensional molecular model for a G-protein-coupled receptor (GPCR). Alignment of the molecular model from the crystallographic structure with the helical axes seen in cryo-electron microscopic (cryo-EM) studies provides an opportunity to investigate the properties of the molecule as a function of orientation and location within the membrane. In addition, the structure provides a starting point for modeling and rational experimental approaches of the cone pigments, the GPCRs in cone cells responsible for color vision. Homology models of the cone pigments provide a means of understanding the roles of amino acid sequence differences that shift the absorption maximum of the retinal chromophore in the environments of different opsins.

The corn inhibitor of blood coagulation factor XIIa. Crystallization and preliminary crystallographic analysis.

A 13.6 kDa protein from corn seeds is known to be a highly selective inhibitor of human blood coagulation Factor XIIa (or activated Hageman factor). We have crystallized this inhibitor at 23 degrees C and pH 7.5 from a solution of 30% polyethylene glycol 400, 0.2 M MgCl2, and 0.1 M Hepes. The crystals diffract to at least 2.1 A resolution. The space group is P4(2)2(1)2 with a = b = 57.15 A and c = 80.5 A. The crystals contain 51% solvent. Two heavy atom derivatives have been identified.

Domain structure, stability and domain-domain interactions in recombinant factor XIII.

The process of heat denaturation of recombinant factor XIII (rFXIII), as well as its C-terminal 24 kDA and 12 kDa elastase-produced fragments starting at Ser514 and Thr628, respectively, was investigated in a wide range of conditions by fluorescence, CD and differential scanning calorimetry (DSC). It was found that the intact protein melts in two distinct temperature regions reflecting unfolding of different parts of the molecule with different stability. The less stable structures unfold in a low temperature transition with a tm of 69 degrees C or lower depending on conditions. Unfolding of the more stable structures was observed at extremely high temperatures, tm > 110 degrees C at acidic pH < 3.5 and tm = 90 degrees C at pH 8.6 with 2 M GdmCL. Thermodynamic analysis of the low and high temperature DSC-obtained heat absorption peaks indicated unambiguously that the first represents melting of three thermolabile independently folded domains while two thermostable domains melt in the second one giving a total of five domains in each a subunit of rFXIII. Both 24 kDa and 12 kDa fragments exhibited a sigmoidal spectral transition at comparatively high temperature where the thermolabile structures are already denatured, indicating that two thermostable domains are formed by the C-terminal portion of rFXIII and correspond to the two beta-barrels revealed by crystallography. The remaining 56 kDa portion forms three thermolabile domains, one of which corresponds to the N-terminal beta-sandwich and the other two to the catalytic core. Fast accessible surface calculations of the X-ray model of rFXIII confirmed the presence of two structural subdomains in the core region with the boundary at residue 332. The thermolabile domains appear to interact with each other intra- and/or intermolecularly resulting in dimerization the a subunits. At acidic pH, where all domains became destabilized but still remained folded, interdomainial interactions seemed to be abolished, resulting in the reversible dissociation of the dimer as revealed by ultracentrifugation analysis.

Fc intermediate (Fci), a papain-generated fragment of human IgG, intermediate in charge, molecular weight and cleavage between the Fc and Fc' fragments of IgG.

Fc intermediate (Fci) is a papain-generated fragment of human IgG which is intermediate in charge, mol. wt and state of cleavage between the Fc and Fc' fragments of IgG. It is composed of two polypeptide chains of unequal mol. wt held together by non-covalent bonds between the C gamma 3 regions. The larger polypeptide chain has both a C gamma 2 and C gamma 3 domain and its N-terminus is at Leu 235 (60%) and Leu 234 (40%) (IgGl Eu numbering). The smaller polypeptide chain is composed of a C gamma 3 domain with its N-terminus at Gly 341. The carboxy-termini obtained by carboxypeptidase digestion and by a computer program which determined the most probable sequences by fitting the amino acid compositions to the sequence of IgG Eu Fc were heterogeneous involving residues 440-446 for the large polypeptide chain and 429-436 for the small one. The calculated mol. wt of the large polypeptide chain was 26,183, assuming the N-terminus at Leu 234 and C-terminus at 446 and including the carbohydrate moiety. The calculated mol. wt for the small polypeptide chain was 10,682, with the N-terminus at 341 and assuming the C-terminus at 434, for a combined mol. wt of 36,865 for the Fci fragment. Sedimentation equilibrium ultracentrifugation of Fci under non-dissociating conditions showed an Mn of 36,200 +/- 1200, an Mw of 36,400 +/- 600 and an Mz of 37,000 +/- 300 g/mole. The best yields of Fc were obtained with a 6-hr digestion and the best yields of Fcl and Fc' were obtained with digestion for 18 hr in phosphate buffer. Digestion in Tris buffer for 18 hr gave results similar to the 18-hr digestion in phosphate buffer except the yields of Fc' were less. This fragment may be useful for exploring biological functions of human IgG Fc. In addition, some Fc fragments obtained by papain digestion of human IgG either in phosphate or Tris buffer are not covalently bonded and are probably cleaved on the carboxy-terminal side of the interchain disulfide bonds at Leu 234 or Leu 235, the N-termini for the large polypeptide chain of Fci. This indicates that, if disulfide bonded Fc fragments are needed, gel filtration under dissociating conditions will be necessary to remove non-covalently bonded Fc.

Transglutaminase factor XIII uses proteinase-like catalytic triad to crosslink macromolecules.

The X-ray crystal structure of human transglutaminase factor XIII has revealed a cysteine proteinase-like active site involved in a crosslinking reaction and not proteolysis. This is among the first observations of similar active sites in 2 different enzyme families catalyzing a similar reaction in opposite directions. Although the size and overall protein fold of factor XIII and the cysteine proteinases are quite different, the active site and the surrounding protein structure share structural features suggesting a common evolutionary lineage. Here we present a description of the residues in the active site and the structural evidence that the catalytic mechanism of the transglutaminases is similar to the reverse mechanism of the cysteine proteinases.

A reexamination of the reaction between colchicine and tubulin.

The rate of reversible dissociation of the colchicine-tubulin complex has been determined for purified porcine brain tubulin and for sea urchin tubulin from 100,000 times g supernatant. The rate constant, which is essentially the same for both species, is extremely slow; it corresponds to a half-life of approximately 36 hours for the reaction. This rate was the same whether or not sucrose was present, and was not influenced by low concentrations of vinblastine or by millimolar concentrations of colchicine. The association rate constant was determined for porcine brain tubulin with and without CaCl2. The sample without CaCl2 had been maximally polymerized prior to the addition of colchicine. A 4.7-fold difference was observed between the association rate constants under these two conditions. Equilibrium constants calculated from the measured rate constants are 5 to 20 times greater than directly measured values; this suggests that the interaction of microtubules and colchicine is more complex than was previously thought.

Structural evidence that the activation peptide is not released upon thrombin cleavage of factor XIII.

The three-dimensional structure of the recombinant human factor XIII a2 dimer after cleavage by thrombin has been determined by X-ray crystallography. Factor XIII zymogen was treated with bovine alpha-thrombin in the presence of 3 mM CaCl2, and the cleaved protein was crystallized from Tris buffered at pH 6.5 using ethanol as the precipitating agent. Refinement of the molecular model of thrombin-cleaved factor XIII against diffraction data from 10.0 to 2.5 A resolution has been carried out to give a crystallographic R factor of 18.2%. The structure of thrombin-cleaved factor XIII is remarkably similar to that of the zymogen: there are no large conformational changes in the protein and the 37 residue amino terminus activation peptide remains associated with the rest of the molecule. This work shows that the activation peptide, upon thrombin cleavage, has the same conformation and occupies the same position with respect to the rest of the molecule as it does in the zymogen structure.

Self-association of lyophilized horse liver alcohol dehydrogenase.

The molecular weights of lyophilized and non-lyophilized horse liver alcohol dehydrogenase have been compared by quasi-elastic light scattering, and ultracentrifugation. Whereas the non-lyophilized enzyme has the expected molecular weight of 78 000, the lyophilized enz)me has an initial molecular weight of about 10(6) which increases with time by an endothermic process. This result shows that any physical measurement using lyophilized liver alcohol dehydrogenase to investigate the enzyme mechanism, which relies upon the molecular size, will be invalid.

Magnetic resonance imaging of the facial nerve during Bell's palsy.

Twenty-five patients with Bell's palsy were evaluated to assess the efficacy of gadolinium (Gd+)-enhanced MRI in determining: (1) the site of facial nerve enhancement, (2) the relationship between EMG findings and Gd+ MRIs, and (3) the usefulness of Gd+ MRI in predicting recovery of facial function. Eighteen of twenty-five patients had enhancement of the facial nerve during Gd+ MRI whereas seven did not. The most common areas of facial nerve enhancement were the labyrinthine, geniculate ganglion, and proximal tympanic segments of the facial nerve. EMGs were performed on ten patients who lost nerve excitability. The segments of facial nerve enhanced during Gd+ MRI varied in location and intensity in patients who maintained nerve excitability and in patients who lost nerve excitability. There was no correlation between EMG findings and location of facial nerve enhancement in patients who lost nerve excitability. The location of facial nerve enhancement during Gd+ MRI was not useful in predicting recovery of facial paralysis.