Three dimensional structure of human C-reactive protein.

The structure of the classical acute phase reactant human C-reactive protein provides evidence that phosphocholine binding is mediated through calcium and a hydrophobic pocket centred on Phe 66. The residue Glu 81 is suitably positioned to interact with the choline group. A cleft on the pentameric face opposite to that containing the calcium site may have an important functional role. The structure provides insights into the molecular mechanisms by which this highly conserved plasma protein, for which no polymorphism or deficiency state is known, may exert its biological role.

The human myosin light chain kinase (MLCK) from hippocampus: cloning, sequencing, expression, and localization to 3qcen-q21.

Myosin light chain kinase (MLCK), a key enzyme in muscle contraction, has been shown by immunohistology to be present in neurons and glia. We describe here the cloning of the cDNA for human MLCK from hippocampus, encoding a protein sequence 95% similar to smooth muscle MLCKs but less than 60% similar to skeletal muscle MLCKs. The cDNA clone detected two RNA transcripts in human frontal and entorhinal cortex, in hippocampus, and in jejunum, one corresponding to MLCK and the other probably to telokin, the carboxy-terminal 154 codons of MLCK expressed as an independent protein in smooth muscle. Levels of expression were lower in brain compared to smooth muscle. We show that within the protein sequence, a motif of 28 or 24 residues is repeated five times, the second repeat ending with the putative methionine start codon. These repeats overlap with a second previously reported module of 12 residues repeated five times in the human sequence. In addition, the acidic C-terminus of all MLCKs from both brain and smooth muscle resembles the C-terminus of tubulins. The chromosomal localization of the gene for human MLCK is shown to be at 3qcen-q21, as determined by PCR and Southern blotting using two somatic cell hybrid panels.

Conformational studies on beta-amyloid protein carboxy-terminal region (residues 34-42): strategic use of amide backbone protection as a structural probe.

Analogues of beta-amyloid (32-42) peptide, containing N-(2-hydroxy-4-methoxybenzyl) (Hmb) amide backbone substitutions at various positions have been prepared using fluoren-9-ylmethoxycarbonyl (Fmoc)-polyamide based solid phase peptide synthesis. On-line N alpha-Fmoc deprotection monitoring during assembly exhibited hindered release in the native and beta A(34-42, (Hmb)Gly38) analogue syntheses. No such hindrance was observed during the synthesis of beta A(34-42, (Hmb)Gly37) nor beta A(34-42, (Hmb)Val36). However, the latter contained an exceptionally slow coupling reaction. Cleaved peptides were analysed for solubility in a variety of solvents and insoluble pellets tested for congophilic staining. X-ray analysis of Fmoc (and H-) beta A(34-42) and the corresponding (Hmb)Gly38 analogues as dimethylformamide swollen gels gave very similar structures. Secondary structure prediction and model-building of ordered arrays, compatible with our results, suggest that beta A(34-42) forms a beta-hairpin structure, with the reverse turn at Val36-Gly37-Gly38-Val39 both in solution and on the resin during synthesis.

Binding of the dye congo red to the amyloid protein pig insulin reveals a novel homology amongst amyloid-forming peptide sequences.

The three-dimensional structure has been determined of a complex of the dye Congo Red, a specific stain for amyloid deposits, bound to the amyloid protein insulin. One dye molecule intercalates between two globular insulin molecules at an interface formed by a pair of anti-parallel beta-strands. This result, together with analysis of the primary sequences of other amyloidogenic proteins and peptides suggests that this mode of dye-binding to amyloid could be general. Moreover, the structure of this dye-binding interface between protein molecules provides an insight into the polymerization of amyloidogenic proteins into amyloid fibres. Thus the detailed characterization, at a resolution of 2.5 A, of the dye binding site in insulin could form a basis for the design of agents targeted against a variety of amyloid deposits.

Crystal structure of ovalbumin as a model for the reactive centre of serpins.

The serpins are a widely distributed family of proteins with diverse functions; they include the key serine protease inhibitors of human plasma as well as noninhibitory homologues such as hormone-binding globulins, angiotensinogen and egg-white ovalbumin. Sequence alignment based on the crystal structure. On the cleaved form of the archetypal serpin, alpha 1-antitrypsin, indicates that the serpins share a common highly ordered structure. On cleavage of the reactive centre peptide bond, they characteristically undergo a remarkable conformational change, the newly generated C terminus moving some 70 A to the opposite pole of the molecule. The structure of this post-cleavage form is known, but the conformation of the intact serpins and in particular that of their reactive centre is not. Wright et al.'s structure of plakalbumin (ovalbumin cleaved by subtilisin) has provided evidence for the conformational change that results from cleavage. We have now determined the structure of native ovalbumin to 1.95 A resolution and have found that the intact peptide loop forming the analogue to the reactive centre of the inhibitory serpins takes the unexpected form of a protruding, isolated helix. This model of the intact structures of the serpins suggests how they may interact with their target proteases.

The use of pseudosymmetry in the rotation function of gamma IVa-crystallin.

Bovine lens gamma IVa-crystallin crystallizes in space group C222(1) with cell dimensions a = 35.1, b = 46.2, c = 186.2 A, and contains one molecule in the asymmetric unit. The structure was determined at 3.0 A resolution using cross-rotation functions and R-factor searches with the bovine lens protein gamma II-crystallin as the model structure. The rotation function appears to be very sensitive to the resolution range and type of coefficient employed; the use of normalized structure-factor amplitudes gave the best results. The potential problem of a pseudo solution due to an internal pseudo-twofold axis was put to advantage by aligning this axis parallel to z. The results of the R-factor search were well defined. The molecular replacement solution was improved by rigid-body least-squares refinement, initially of the whole molecule, then for the two domains. The R factor at this stage was 39.4% at 2.3-10.0 A. The gamma IVa structure has an even higher internal symmetry than gamma II, since the two domains are related by a rotation around the pseudo-twofold axis of 178.7 degrees as compared with 176.2 degrees for gamma II.

Acute-phase high-density lipoprotein in the rat does not contain serum amyloid A protein.

Serum amyloid A protein (SAA) is an acute-phase apolipoprotein of high-density lipoprotein (HDL). Its N-terminal sequence is identical with that of amyloid A protein (AA), the subunit of AA amyloid fibrils. However, rats do not develop AA amyloidosis, and we report here that neither normal nor acute-phase rat HDL contains a protein corresponding to SAA of other species. mRNA coding for a sequence homologous with the C-terminal but not with the N-terminal part of human SAA is synthesized in greatly increased amounts in acute-phase rat liver. These observations indicate that the failure of rats to develop AA amyloid results from the absence of most of the AA-like part of their SAA-like protein, and that the N-terminal portion of SAA probably contains the lipid-binding sequences.

Is the serum amyloid A protein in acute phase plasma high density lipoprotein the precursor of AA amyloid fibrils?

Serum amyloid A protein (SAA), an apolipoprotein of high density lipoprotein (HDL), is generally considered to be the precursor of AA protein, which forms the fibrils in reactive systemic amyloidosis in man and animals. This view is based on amino acid sequence identity between AA and the amino-terminal portion of SAA. However, in extensive and well-controlled studies of experimentally induced murine AA amyloidosis, we were unable to demonstrate a direct precursor-product relationship between SAA, in SAA-rich HDL preparations from acute phase or amyloidotic mouse or human serum, and AA protein in the amyloid deposits. This raises the possibility that SAA in its usual form, as an apolipoprotein of HDL synthesized during the acute phase response, may not be the major precursor of AA fibrils. The amyloidogenic forms of circulating SAA molecules may not be isolated during the preparation of HDL. Alternatively, particularly in the light of recent evidence that SAA mRNA is expressed in many different tissues throughout the body of appropriately stimulated animals, amyloidogenic SAA may be derived from sources other than the liver cells in which SAA-rich HDL is synthesized.