Fibrillation of human insulin A and B chains

Human insulin, which consists of disulfide cross-linked A and B polypeptide chains, readily forms amyloid fibrils under slightly destabilizing conditions. We examined whether the isolated A and B chain peptides of human insulin would form fibrils at neutral and acidic pH. Although insulin exhibits a pH-dependent lag phase in fibrillation, the A chain formed fibrils without a lag at both pHs. In contrast, the B chain exhibited complex concentration-dependent fibrillation behavior at acidic pH. At higher concentrations, e.g., >0.2 mg/mL, the B chains preferentially and rapidly formed stable protofilaments rather than mature fibrils upon incubation at 37 degrees C. Surprisingly, these protofilaments did not convert into mature fibrils. At lower B chain concentrations, however, mature fibrils were formed. The explanation for the concentration dependence of B chain fibrillation is as follows. The B chains exist as soluble oligomers at acidic pH, have a beta-sheet rich conformation as determined by CD, and bind ANS strongly, and these oligomers rapidly form dead-end protofilaments. However, under conditions in which the B chain monomer is present, such as low B chain concentration (<0.2 mg/mL) or in the presence of low concentrations of GuHCl, which dissociates the soluble oligomers, mature fibrils were formed. Thus, both A and B chain peptides can form amyloid fibrils, and both are likely to be involved in the interactions leading to the fibrillation of intact insulin.     

Interaction and combination of separate A and B chains of insulin effects of D-and L-tryptophan in position A1

The S-thiomethyl derivatives of insulin A chain with A1-Gly replaced by D- or L-Trp have been prepared and their respective interaction and combination with the S-thiomethyl B chain studied. The UV difference spectra of the mixed against the separated [Trp1]A chains with the B chain at pH 10.8 are similar to those obtained for the unmodified chains except that the 295-nm-negative peak for ionized Tyr residue appears to be less marked. Fluorescence studies show very little environmental changes at the A1-Trp residues when mixed with the B chain. The intact hormone with A1-Gly replaced by D-Trp is known to be considerably more active than the analog with L-Trp replacement. However, for both derivatives the resynthesis of the whole molecules correctly joined by disulfide bridges starting from the separated reduced chains, gives similar low yields as shown by HPLC analysis and by receptor-binding assay. The replacement of A1-Gly by D-Trp appears to affect the separated A chain more than the intact hormone and replacements at A1 by both D- and L-Trp probably lead to significant conformational changes of the A chain so as to prevent its correct pairing with the B chain.     

Identification of A chain cleavage sites in intact insulin produced by insulin protease and isolated hepatocytes

The degradation of insulin by the enzyme insulin protease and by isolated hepatocytes results in proteolytic cleavages in both the A and B chains of intact insulin. Previous studies have shown that one of the A chain cleavages is between A13 leucine and A14 tyrosine and that a second cleavage occurs carboxyl to the A14 residue. In the present study we have used insulin specifically iodinated on the A19 tyrosine and examined the A chain cleavages by the enzyme and by hepatocytes. Insulin degradation products were purified by HPLC and sequenced by automated Edman degradation. Only two A chain cleavage sites were identified, one the previously reported A13-A14 and the other between A14 tyrosine and A15 glutamine. These data thus identify the second A chain cleavage site and further support the role of insulin protease in hepatic metabolism of insulin.     

Resynthesis of insulin from its A and B chains in the presence of denaturants

Reoxidation of the reduced insulin A and B chains in 8 M urea leads to a recovery of native insulin of 2-7% whereas in 6 M guanidine or 0.5 mM dodecylsulfate very little, if any, resynthesis of insulin could be detected. Considering all the possibilities of different oligomeric forms containing one or both of the chains, the yield in 8 M urea is well over the yield as calculated from random joining of the chains and the yield in guanidine or dodecylsulfate is to be expected. It is concluded that some interaction and pairing of the chains occur even in the presence of 8 M urea.     

The N terminus of laminin A chain is homologous to the B chains

A major proteolytic fragment (E1/E1-4) of the basement membrane protein laminin, comprising the three short arms with some terminal globules missing, was isolated by elastase digestion, and partial protein sequence data were determined for several tryptic peptides. Sequences which corresponded to A-chain structures were used to synthesize oligonucleotides for the construction and screening of a primer-extended cDNA library from mouse PYS-2 cells. A clone of 1.1 kb was obtained and shown by sequencing to correspond to the 5' end of the 10-kb mRNA of the A chain of laminin. The clone contains 77 nucleotides of 5' untranslated sequence and a region coding for 334 amino acids, including a presumptive signal peptide of 24 amino acids. The sequence is 30% homologous to the corresponding N-terminal part of the B1 chain of laminin, suggesting the same structure for both domains. The data present further evidence for a recent structural model which postulates that each of the three laminin polypeptide chains forms a distinct short arm.     

Influence of imprinting with A and B chains of insulin on binding and functional changes in tetrahymena

Insulin and its A and B chain increased the quantity of intracellular PAS-positive material (glycogen) in tetrahymena, whereas the combined A+B chains decreased it. Imprinting—previous interaction—with insulin, its A and B chains in themselves and with the A+B chain increased the hormone binding capacity of tetrahymena, but the functional effect of imprinting (storage or breakdown of glycogen) showed a different tendency with insulin and A+B chain on the one hand, and A chain and B chain on the other. Since the imprinting potential of a molecule promotes the induction of receptor formation, the fact remains that both component chains of insulin were able to act as potential imprinters, although the A chain was superior to the B chain in this respect throughout, and combined treatment with the A+B chain ultimately induced the formation of a similar binding site as insulin itself.     

The insulin A and B chains contain sufficient structural information to form the native molecule

Refolding studies show that native insulin can be reformed from A and B chains only. This suggests that the A and B chains contain the necessary structural information and that the C peptide is not required for this process.     

Cloning and expression of the B chain of volkensin, type 2 ribosome inactivating protein from Adenia volkensii harms: co-folding with the A chain for heterodimer reconstitution

Type 2 ribosome inactivating proteins (RIPs) include some potent plant toxins, among which ricin from Ricinus communis and abrin from Abrus precatorius seeds, have been known for more than a century. Two other type 2 RIPs belong to this class of proteins, both isolated from plants of the same family (Passifloraceae), modeccin and volkensin, from Adenia digitata and Adenia volkensii roots, respectively. Volkensin is probably the most potent plant toxin known, with an LD50 for rats of 50-60 ng/kg. Here we report the cloning, expression and renaturation of recombinant volkensin B chain. Furthermore, starting from separately expressed A and B chains, a co-association procedure was set-up, leading to in vitro heterodimeric volkensin reconstitution. The recombinant heterodimer was characterized by N-terminal sequence analysis and its hemagglutinating activity assessed. In parallel, we have explored the carbohydrate-binding properties of native volkensin with the aim to correlate toxin-specific properties (i.e., axonal transport along neurons) to lectin's sugar-binding preferences.     

Limited proteolysis patterns of the B chain of insulin.

The oxidized B chain of insulin was used as a simple model for further consideration of limited proteolysis with low substrate:enzyme ratios. With low B chain:trypsin ratios, the ordinarily slower cleavage rate of the -Lys₂₉-Ala₃₀ bond essentially equaled the cleavage saturation rate of the -Arg₂₂-Gly₂₃ bond. This led to the disappearance of octapeptide which ordinarily forms most rapidly. Heptapeptide and alanine, formed mainly by cleavage of the octapeptide, decreased somewhat at high enzyme relative levels. Trypsin added to B chain formed a single chromatographic peak.     

The pairing of the separated A and B chains of insulin and its derivatives, FTIR studies.

From the amide I bands of their deconvolved FTIR spectra, the S-thiomethyl derivatives of the insulin A, B, despentapeptide(26-30) B and desoctapeptide(23-30) B chains all contain significant amounts of ordered secondary structure. The intact B chain is considerably more ordered than either the A or the truncated B chains. Comparison of the spectra of the separated and mixed intact chains of insulin suggests further folding upon mixing of the chains leading to significant increases in ordered secondary structures, presumably because of stabilization by interaction of the chains. The interactions of the A chain with the DPI B chain appear to be weaker as compared to that with the intact B chain. The above results suggest that only the intact A and B chains contain sufficient structural information to recognize each other and interact to form a native-like structure which make the correct formation of the disulfide linkages possible.     

Action of human liver cathepsin B on the oxidized insulin B chain.

The lysosomal cysteine proteinase cathepsin B (from human liver) was tested for its peptide-bond specificity against the oxidized B-chain of insulin. Sixteen peptide degradation products were separated by high-pressure liquid chromatography and thin-layer chromatography and were analysed for their amino acid content and N-terminal amino acid residue. Five major and six minor cleavage sites were identified; the major cleavage sites were Gln(4)-His(5), Ser(9)-His(10), Glu(13)-Ala(14), Tyr(16)-Leu(17) and Gly(23)-Phe(24). The findings indicate that human cathepsin B has a broad specificity, with no clearly defined requirement for any particular amino acid residues in the vicinity of the cleavage sites. The enzyme did not display peptidyldipeptidase activity with this substrate, and showed a specificity different from those reported for two other cysteine proteinases, papain and rat cathepsin L.     

Interaction of zinc protoporphyrin with intact oxyhemoglobin

In erythropoietic protoporphyria and lead poisoning, free protoporphyrin (PPIX) and zinc protoporphyrin (ZPP), respectively, accumulate in erythrocytes. That PPIX and ZPP bind to human hemoglobin A (Hb4) is established, but the site of binding is still a matter of controversy. We investigated the interaction of ZPP with intact, tetrameric oxy Hb4, using batch microcalorimetry, front-face fluorometry, absorption difference spectroscopy, oxygen equilibrium studies, and isoelectric focusing (IEF). In the presence of oxy Hb4 (pH 7.35, 0.05 M phosphate), the fluorescence emission maximum (excitation at 420 nm) of ZPP immediately shifts from 587 nm (ZPP alone) to 594 nm, as expected when binding to protein. The fluorescence intensity increases with time and is correlated with the ZPP:Hb4 mole ratio. A slow, time-dependent reaction is also observed with microcalorimetry: the rate of heat of reaction exhibits both a fast and a slow component. The heats of reaction range from -2.1 to -14.8 mcal depending upon the ZPP:Hb4 ratio of 4:1 (0.4 mM:0.1 mM) to 38:1 (3.8 mM:0.1 mM), respectively, and are typical of weak, noncovalent protein-ligand interactions. The optical difference spectra are a function of the ZPP:Hb4 molar ratio and also exhibit a slow increase in intensity over time. No time-dependent optical difference spectra are observed with ZPP or with Hb4 alone. The oxygen affinity of Hb4 in the presence of ZPP decreases with increasing mole ratio. During IEF, all ZPP separates from Hb4, consistent with a weak, noncovalent interaction at a non-heme pocket site. We conclude that ZPP binds to intact, tetrameric hemoglobin at non-heme pocket sites in a nonspecific, weak, noncovalent interaction.     

Synergy between immunotoxins prepared with native ricin A chains and chemically-modified ricin B chains

Ricin B chains treated with chloramine-T in the presence or absence of NaI show a 100-fold to 200-fold reduction in their ability to bind to the galactose-containing protein asialofetuin. Such treated B chains do not form covalently associated homodimers with treated B chains or heterodimers with native ricin A chains. Furthermore, they cannot enhance the toxicity of a ricin A chain-containing rabbit anti-human immunoglobulin (RAHIg-A) for Daudi cells. However, when such B chains are coupled to goat anti-rabbit Ig (GARIg), they potentiate the killing of RAHIg-A-treated Daudi cells only slightly less effectively than GARIg coupled to native B chains. Furthermore, if GARIg-B chain conjugates are treated with chloramine-T after coupling, they fail to bind to asialofetuin but enhance the killing of Daudi cells treated with RAHIg-A. These results demonstrate that the ability of ricin B chains to bind to galactose and to enhance the toxicity of ricin A chains (in the form of an antibody-A chain) can be operationally separated. Thus, the two functions of the B chain may reside on separate domains of the molecule.     

The insulin A and B chains contain structural information for the formation of the native molecule. Studies with protein disulphide-isomerase.

It has been shown previously [Tang, Wang & Tsou (1988) Biochem. J. 255, 451-455] that, under appropriate conditions, native insulin can be obtained from scrambled insulin or the S-sulphonates of the chains with a yield of 25-30%, together with reaction products containing the separated A and B chains. The native hormone is by far the predominant product among the isomers containing both chains. It is now shown that the presence of added C peptide has no appreciable effect on the yield of native insulin. At higher temperatures the content of the native hormone decreases whereas those of the separated chains increase, and in no case was scrambled insulin containing both chains the predominant product in the absence of denaturants. Both the scrambling and the unscrambling reactions give similar h.p.l.c. profiles for the products. Under similar conditions cross-linked insulin with native disulphide linkages can be obtained from the scrambled molecule or from the S-sulphonate derivative with yields of 50% and 75% respectively at 4 degrees C, and with a dilute solution of the hexa-S-sulphonate yields better than 90% can be obtained. The regenerated product is shown to have the native disulphide bridges by treatment with CNBr to give insulin and by the identity of the h.p.l.c. profile of its peptic hydrolysate with that for cross-linked insulin. It appears that the insulin A and B chains contain sufficient information for the formation of the native molecule and that the role of the connecting C peptide is to bring and to keep the two chains together.     

Primary structure of the Drosophila laminin B2 chain and comparison with human, mouse, and Drosophila laminin B1 and B2 chains

Laminin, a major component of basement membranes, is a large glycoprotein consisting of three disulfide-bonded subunits, A, B1, and B2. We have isolated and sequenced a Drosophila laminin B2 chain cDNA clone that spans 5737 nucleotides. The deduced amino acid sequence predicts that the mature and nonglycosylated polypeptide has a chain length of 1606 residues (Mr = 178,665). This B2 chain contains 100 half-cystine residues, most of which are located in two cysteine-rich domains, and 11 N-X-S or N-X-T sequences which are potential sites of N-linked glycosylation. The predicted secondary structure reveals the presence of six structurally distinct domains, of which two are mainly alpha-helical, two are cysteine-rich with homologous repeats, and two are globular regions. The Drosophila B2 chain is 40.3 and 41.1% identical to the human and mouse B2 chains, respectively, and 29.6, 30.0, and 29.4% identical to the Drosophila, human, and mouse B1 chains, respectively.     

Synthesis of an insulin-like compound consisting of the A chain of insulin and a B chain corresponding to the B domain of human insulin-like growth factor I

An insulin-like hybrid molecule consisting of the A chain of insulin and a B chain corresponding to the B domain of human insulin-like growth factor I (growth factor I sequence 1-30) has been synthesized essentially by the procedures developed in this laboratory for the synthesis of insulin and analogues. The hybrid competed with 125I-insulin for insulin receptors in rat liver plasma membranes and was a full agonist in stimulating incorporation of [3(-3)H]glucose into lipids in rat adipocytes. In both assays, the compound displayed ca. 2% of the potency of insulin. The compound was recognized by anti-insulin antibodies but was only ca. 0.25% as potent as insulin in this activity. The hybrid exhibited growth-promoting activity in fibroblasts, displaying 3-8% of the activity of insulin. In contrast, the compound was recognized by insulin-like growth factor carrier proteins, a property not associated with insulin. Two points of nonhomology between the B chain of insulin and the B domain of insulin-like growth factor I are considered in connection with these observations.