Correspondence of homologies in amino acid sequence and tertiary structure of protein molecules

According to the method developed previously (Kubota, Y., Takahashi, S., Nishikawa, K. and Ooi, T. (1981) J. Theor, Biol. 91, 347-361), homology among proteins may be estimated quantitatively. We extended the method to investigate the relationship of an amino acid sequence to its teritary structure and identify homologous segments which have homologous native conformations in proteins. First, we selected proper indices for the computation of correlation coefficients from 32 properties inherent to amino acids, such as hydrophobicity. The arithmetic average of correlation coefficients using six indices gave rise to a good correlation for the CD- and EF-hand regions (Ca2+ binding sites) in carp parvalbumin, but poor ones for other segments. We then applied the method to homologous proteins, the three-dimensional structures of which are known: horse hemoglobin alpha-chain and beta-chain; cytochrome c and c2; serine proteases, chymotrypsinogen and elastase; alpha-lytic protease and protease A from prokaryotic organisms. The results show that the sequence homology estimated by the present method has a good correspondence to the homology in three-dimensional structures and therefore the method is promising for the identification of important sites in sequences which have similar native conformations. For an example of the application of the method, two sequences of human interferon, one from fibroblast and the other from leukocyte, are compared, suggesting functional sites in the molecule.     

Production and characterization of functional domains of human fibronectin expressed in Escherichia coli.

An efficient expression system was constructed in Escherichia coli that produced a 33-kDa fragment, C-274, of human fibronectin with a strong cell-adhesive activity. The entire sequence of the heparin-binding domain with 271 amino acids, H-271, was also expressed. Deletion analysis of the type III repeats showed that the heparin-binding site was at type III-13. The cell-adhesive activity of a fusion protein, CH-271, containing the cell- and the heparin-binding domains was twice that of C-274 when BHK but not B16-F10 melanoma cells were tested; H-271 alone was inactive. Recombinant proteins containing the CS1 sequence of the IIICS region were more active than C-274 and CH-271 with B16-F10. However, H-296, which contained both H-271 and CS1, was almost inactive with BHK. CH-296, which contained CS1 at the C-terminus of CH-271, was more active with B16-F10 than H-296 and C-CS1, which was produced by the deletion of H-271 from CH-296. Thus, the cell-binding domain was active with both kinds of cells. The heparin-binding domain promoted the adhesion of both kinds of cells only when linked to the cell-binding domain or CS1. CS1 was specific for the adhesion of B16-F10 but was not essential.     

Rat cathepsin H-catalyzed transacylation: comparisons of the mechanism and the specificity with papain-superfamily proteases.

We found that rat cathepsin H showed strong transacylation activity under physiological conditions. It is a feature of cathepsin H to utilize amino acid amides not only as acyl-acceptors but also as acyl-donors in the reaction. The pH-dependence of the transacylation activity was distinct from those of other papain-superfamily proteases. The alkaline limb (pKapp = 7.5) could be regarded as the pKa of the alpha-amino group of the acyl-donor, which was also involved in the original amino-peptidase activity. The acidic limb (pKapp = 5.8) was suggested to be involved in the deacylation step, where amino acid amide attacked the acyl-intermediate as a nucleophile in place of water in the hydrolysis. Although the N alpha-deprotonated acyl-acceptor, which is supposed to govern the nucleophilic attack, has a small population in the acidic pH range (above pH5), the transacylation was detectable even at the acidic pH-range because of the high S1'-site binding ability and suitable nucleophilicity of the acyl-acceptor. In the transacylation between various amino acid amides, the S1 and S1' site appeared to prefer hydrophobic residues without and regardless of a branch at beta-carbon, respectively. From these results and the sequence homology in the papain superfamily, we concluded that the reaction was governed by the acyl-donor having a protonated amino group, the acyl-acceptor having a deprotonated amino group and the remarkable hydrophobic character (especially favoring tryptophan amide) of the S1' site, presumably reflecting the good conservation of Trp177 in papain-superfamily proteases.     

Endogenous basic fibroblast growth factor-dependent induction of collagenase and interleukin-6 in tumor necrosis factor-treated human microvascular endothelial cells.

Tumor necrosis factor-alpha (TNF-alpha) has been shown to enhance the synthesis of interleukin-6 (IL-6) and collagenase in human omental microvascular endothelial (HOME) cell (Mawatari, M., Kohno, K., Mizoguchi, H., Matsuda, T., Asoh, K., Van Damme, J. V., Welgus, H. G., and Kuwano, M. (1989) J. Immunol. 143, 1619-1627). In the present study, we have examined whether the TNF-alpha-induced synthesis of IL-6 or collagenase in HOME cells is mediated by an inducible growth factor. Among several growth factors examined, we found that the expression of basic fibroblast growth factor (bFGF) mRNA was the one most prominently enhanced when HOME cells were treated with TNF-alpha. The increase of bFGF mRNA by TNF-alpha in HOME cells was observed in both a dose- and time-dependent manner when assayed by Northern blot analysis. The induction of bFGF mRNA was observed by 3 h after incubation with TNF-alpha, and the maximal increase of 5-fold was obtained after 12 h of incubation with 100 units/ml TNF-alpha. Western blot analysis confirmed the enhanced synthesis of bFGF by TNF-alpha. Metabolic labeling and immunoprecipitation assays of bFGF showed that exposure to TNF-alpha enhanced secretion of bFGF into culture medium and also that TNF-alpha stimulated the production of bFGF molecules with molecular masses of 18, 21, and 22.5 kDa in HOME cells. TNF-alpha induced the expression of collagenase mRNA and IL-6 mRNA in HOME cells as well, and the coadministration of neutralizing anti-bFGF antibody almost completely blocked the effects of TNF-alpha. The treatment of HOME cells with exogenous bFGF significantly stimulated the expression of collagenase mRNA and IL-6 mRNA in HOME cells. Therefore, the biological effects of TNF-alpha on HOME cells may be mediated, at least in part, by TNF-alpha-induced bFGF.     

Two genes, atpC1 and atpC2, for the gamma subunit of Arabidopsis thaliana chloroplast ATP synthase.

Arabidopsis thaliana has two genes (atpC1, atpC2) coding for gamma subunits of chloroplast ATP synthase. The atpC1 and atpC2 were cloned and sequenced. They had no introns within the reading frames and coded for proteins of 373 and 386 amino acid residues, respectively, including putative transit sequences (50 and 60 amino acid residues, respectively). In contrast, the spinach gamma subunit gene had two introns within the reading frame. The mature sequences coded by the two genes of A. thaliana (atpC1, 323 residues; atpC2, 326 residues) were homologous with that of spinach (J. Miki, M. Maeda, Y. Mukohata, and M. Futai (1988) FEBS Lett. 232, 221-226): the homologies of gamma subunits coded by atpC1 and atpC2 were 72%, those of the subunits coded by atpC1 and spinach cDNA were 84%, and those of the proteins coded by atpC2 and spinach cDNA were 71%. Like the spinach subunit, the gamma subunits coded by the two genes had unique regulatory domains not found in mitochondrial or bacterial subunits. Poly(A)+ mRNAs corresponding to atpC1 (1.5 kilobases) and atpC2 (2.5 kilobases) were detected in illuminated plants, the amount of the former being at least 140 times that of the latter. The atpC1 mRNA was not found in dark-adapted plants. Nuclear protein(s) specifically bound to the upstream region of atpC1 was detected by gel shift assay and its binding was shown to be inhibited by the GT-1 element of the gene encoding the ribulose-1,5-bisphosphate carboxylase small subunit, which is expressed under illumination (P. J. Green, S. A. Kay, and N. H. Chau (1987) EMBO J. 6, 2543-2549). Consistent with these findings, an increased amount of the gamma subunit was detected immunochemically in illuminated plants.     

Immunochemical approach to characterize advanced glycation end products of the Maillard reaction. Evidence for the presence of a common structure

Reaction of protein amino groups with glucose (the Maillard reaction) leads from early stage products such as Schiff base and Amadori products to advanced glycation end products (AGE), structures implicated in diabetic complications and the aging process. We have prepared the polyclonal anti-AGE antibody and the monoclonal anti-AGE antibody against AGE-bovine serum albumin and made an immunochemical approach to characterize AGE structures. Both polyclonal and monoclonal antibodies reacted with AGE-proteins such as AGE-bovine serum albumin, AGE-human serum albumin, and AGE-hemoglobin but not with unmodified counterparts. Treatments of these AGE-proteins with borohydride had no effect on the immunoreactivity. Moreover, fructosyl-epsilon-caproic acid, a synthetic Amadori compound, did not serve as an antigen, indicating that these antibodies were specific for AGE products but not for early stage products of the Maillard reaction. In addition, these antibodies were also able to recognize AGE products prepared either from alpha-tosyl-1-lysine, alpha-tosyl-1-lysine methyl ester, monoaminocarboxylic acid such as epsilon-aminocaproic acid, gamma-amino-n-butyric acid, and beta-alanine. Thus, these results strongly suggest the presence of a common structure in AGE preparations, regardless of whether AGE products are generated from proteins, amino acids, or monoaminocarboxylic acids.     

Direct effects of ionizing radiation on integral membrane proteins. Noncovalent energy transfer requires specific interpeptide interactions

The 12 transmembrane alpha helices (TMHs) of human erythrocyte glucose transporter were individually cut by pepsin digestion as membrane-bound 2.5-3.5-kDa peptide fragments. Radiation-induced chemical degradation of these fragments showed an average target size of 34 kDa. This is 10-12 x larger than the average size of an individual TMH, demonstrating that a significant energy transfer occurs among these TMHs in the absence of covalent linkage. Heating this TMH preparation at 100 degrees C for 15 min reduced the target size to 5 kDa or less, suggesting that the noncovalent energy transfer requires specific helix-helix interactions. Purified phospholamban, a small (6-kDa) integral membrane protein containing a single TMH, formed a pentameric assembly in sodium dodecyl sulfate. The chemical degradation target size of this phospholamban pentamer was 5-6 kDa, illustrating that not all integral membrane protein assemblies permit intersubunit energy transfer. These findings together with other published observations suggest strongly that significant noncovalent energy transfer can occur within the tertiary and quaternary structure of membrane proteins and that as yet undefined proper molecular interactions are required for such covalent energy transfer. Our results with pepsin-digested glucose transporter also illustrate the importance of the interhelical interaction as a predominating force in maintaining the tertiary structure of a transmembrane protein.     

Copolymerization of pNcollagen III and collagen I. pNcollagen III decreases the rate of incorporation of collagen I into fibrils, the amount of collagen I incorporated, and the diameter of the fibrils formed

Previous observations suggested that pNcollagen III, the partially processed form of type III procollagen, coats fibrils of collagen I and thereby helps regulate the diameter of fibrils formed by collagen I. The previous observations, however, did not exclude the possibility that pNcollagen III was deposited on preformed collagen I fibrils after the fibrils were assembled. Here, mixtures of pNcollagen III and collagen I were generated simultaneously by enzymatic cleavage of precursor forms of the proteins. The results demonstrated that pNcollagen III forms true copolymers with collagen I. The presence of pNcollagen III both inhibited the rate at which collagen I assembled into fibrils and decreased the amount of collagen I incorporated into fibrils at steady-state equilibrium. In addition, the results demonstrated that copolymerization of pNcollagen III with collagen I generated fibrils that were thinner than fibrils generated under the same conditions from collagen I alone. Increasing the initial molar ratio of pNcollagen III to collagen I in the solution-phase increased the amount of pNcollagen III copolymerizing with collagen I and progressively decreased the diameter of the fibrils. Therefore, the copolymers were heterogeneous in that the stoichiometry of the two monomers in the fibrils varied. The results are consistent with a model in which pNcollagen III can regulate the diameter of collagen I fibrils by coating the surface of the fibrils and thereby allow tip growth but not lateral growth of the fibrils.     

Fusion of sphingomyelin vesicles induced by proteins from Taiwan cobra (Naja naja atra) venom. Interactions of zwitterionic phospholipids with cardiotoxin analogues

Egg sphingomyelin vesicles were used to assay aggregation/fusion activities of proteins from Taiwan (Naja naja atra) venom to avoid the problem of phospholipase A2 contamination during protein purification. It led to the identification of a new cardiotoxin (CTX) analogue protein (CTX V) with major aggregation/fusion, but few hemolysis, activities. On the contrary, cardiotoxin (CTX III) induced significant hemolysis of human red blood cells but exhibited few aggregation/fusion activities. To study the structure/activity relationship of these CTX-induced processes, the amino acid sequence of CTX V was determined and its aggregation/fusion activity was compared with that of CTX III by transmission electron microscopy, quasielastic laser light scattering, differential scanning calorimetry, and fluorescence spectroscopy. The results show that the CTX-induced fusion process at temperatures slightly above that of the gel to liquid-crystalline phase transition of sphingomyelin vesicles can ultimately convert small sonicated vesicles into large fused vesicles with sizes of 1-2 microns. The abilities of CTX V to induce the leakage of sphingomyelin vesicles content and to cause the fusion of vesicles are approximately 10-fold higher than those of CTX III. Based on the CTX structures determined in the present and other studies, it is suggested that the amino acid residue X within the well conserved sequence of -Cys-Pro-X-Gly-Lys-Gln-Leu-Cys- plays a role in the interaction of CTX with lipid molecules. The lipid phase transition could further enhance the protein-lipid interaction in the process leading to the fusion of vesicles.     

Alteration of N-terminal residues of mature human lysozyme affects its secretion in yeast and translocation into canine microsomal vesicles

Signal sequences play a central role in the initial membrane translocation of secretory proteins. Their functions depend on factors such as hydrophobicity and conformation of the signal sequences themselves. However, some characteristics of mature proteins, especially those of the N-terminal region, might also affect the function of the signal sequences. To examine this possibility, several mutants of human lysozyme modified in the N-terminal region of the mature protein were constructed, and their secretion in yeast as well as in vitro translocation into canine pancreatic microsomes were analyzed using an idealized signal sequence L8 (MR(L)8PLAALG). Our results show the following. (1) Change in the charge at the N-terminal residue of the mature protein does not affect secretion drastically. (2) Substitution of a proline residue at the N terminus prevents cleavage of the signal sequence, although translocation itself is not impaired. (3) Excessive positive charges in the N-terminal region delay translocation of the precursor protein across the membrane. (4) Polar and negatively charged residues introduced into the N-terminal region affect the secretion of the mature protein by preventing its correct folding.