Analytical biotechnology of recombinant peptides and proteins: II. A confirmation of the primary structure of fusion protein containing human proinsulin and optimization of its proteolysis by trypsin

The kinetics of trypsin proteolysis of the fusion protein (FP) containing human proinsulin was studied by a set of analytical micromethods. These were the microcolumn reversed-phase HPLC and the qualitative identification by MALDI-TOF mass spectrometry and amino acid sequencing. The first stage of the proteolysis was shown to be the cleavage of FP into the leader fragment and proinsulin. The subsequent splitting off ofC-peptide from proinsulin results in the formation of Arg^B31-Arg^B32-insulin. The effect of temperature on the formation of de-Thr^B30-insulin, a by-product, was also studied. The structure of FP was confirmed by the peptide mapping technique, and the leader fragment was shown to contain noN-terminal Met residue. For communication I, see [1].     

Photoaffinity labelling of a nitrobenzylthioinosine-binding polypeptide from cultured Novikoff hepatoma cells.

Incubation of pig desoctapeptide-(B23-30)-insulin with trypsin in solvent systems consisting of dimethyl sulphoxide, butane-1,4-diol and Tris buffer resulted in the formation of an extra peptide bond between Arg-B22 and Gly-A1 in the DOPI molecule. This DOPI derivative can also be regarded as pig des-(23-63)-proinsulin. The structure of the new, previously unreported, proinsulin analogue was determined on the basis of amino acid analysis, dansylation and digestion with Staphylococcus aureus V8 proteinase. Receptor-binding ability of des-(23-63)-proinsulin was 20% of that of pig desoctapeptide-(B23-30)-insulin and 0.02% of that of pig insulin.     

Cloning a replication function from the streptomycete bacteriophage FP43.

Streptomyces griseofuscus cells carrying a 4.4-kb SphI DNA fragment from bacteriophage FP43 inhibited plaque formation (Pin) by FP43, and the Pin function was localized to a 0.96-kb SacII fragment. The same 4.4-kb SphI fragment was able to replicate freely in several streptomycetes, including S. griseofuscus, and the replication (Rep) function was localized to a 1.2-kb SphI-FspI fragment. Plasmids with FP43 Rep function are unstable and are present at about 20-50 copies per chromosome. Plasmids with FP43 Rep function are compatible with SCP2* plasmids.     

Acidic pH generated by H+-ATPase pumps triggers the activity of a fusogenic protein associated with rat liver endoplasmic reticulum.

Fusogenic protein (FP) is a glycoprotein ( approximately 50 kDa), previously purified by us from rat liver endoplasmic reticulum, which explicates fusogenic activity at acidic pH in vitro. To suggest a possible role of FP in membrane fusion, the topology of the protein in the membrane and the conditions in which FP is operating in microsomes have been investigated. Anti-FP polyclonal antibodies inhibited pure FP activity, but not the protein activity in microsomes, suggesting interaction of antibodies with a part of FP concealed in intact membranes. FP activity in microsomes was lost after treatment with Pronase. Western blot analysis of Pronase-treated microsomes showed that the proteolysis removed a fragment ( approximately 5 kDa). This fragment is exposed on the outer surface of microsomes and involved in fusogenic activity, whereas the largest part of FP is embedded in microsomal vesicles. Therefore, FP can be affected by modifications on the cytosolic and luminal sides of microsomal membranes. Indeed, when microsomal lumen was acidified by H+-ATPase activity, binding and fusion of fluorescent labelled liposomes to microsomes occurred. Direct involvement of FP in the fusogenic event was observed by reconstituting pure FP in liposomes with a preformed H+ gradient. FP triggered a fusion process in response to the acidic interior of liposomes, despite an exterior 7.4 pH unable to promote fusogenic protein activity. As intracellular membrane fusion occurs at neutral pH involving the cytosolic sides of membranes, FP may participate in this event by exploiting the acidic pH formed in the lumen of endoplasmic reticulum through H+-translocating ATPase activity.     

Increased production of human proinsulin in the periplasmic space of Escherichia coli by fusion to DsbA

The production of human proinsulin in its disulfide-intact, native form in Escherichia coli requires disulfide bond formation and the periplasmic space is the favourable compartment for oxidative folding. However, the secretory expression of proinsulin is limited by its high susceptibility to proteolysis and by disulfide bond formation, which is rate-limiting for proinsulin folding. In this report we describe a method for the production of high amounts of soluble, native human proinsulin in E. coli. We fused proinsulin to the C-terminus of the periplasmic disulfide oxidoreductase DsbA via a trypsin cleavage site. As DsbA is the main catalyst of disulfide bond formation in E. coli, we expected increased yields of proinsulin by intra- or intermolecular catalysis of disulfide bond formation. In the context of the fusion protein, proinsulin was found to be stabilised, probably due to an increased solubility and faster disulfide bond formation. To increase the yield of DsbA-proinsulin in the periplasm, several parameters were optimised, including host strains and cultivation conditions, and in particular growth medium composition and supplement of low molecular weight additives. We obtained a further, about three-fold increase in the amount of native DsbA-proinsulin by addition of L-arginine or ethanol to the culture medium. The maximum yield of native human proinsulin obtained from the soluble periplasmic fraction after specific cleavage of the fusion protein with trypsin was 9.2 mg g(-1), corresponding to 1.8% of the total cell protein.     

Radioautographic and quantitative study of insulin binding sites in the rat brain

In the present study, we describe the specificity and the autoradiographic distribution of insulin binding sites in the rat central nervous system (CNS) after in vitro incubation of brain sections with [125I]-14A insulin. Increasing concentrations of unlabeled insulin produced a dose-dependent inhibition of [125I]-insulin binding which represented 92 +/- 2% displacement with 3 X 10(-5) M, whatever the brain sections tested. Half-maximum inhibition with native insulin was obtained with 2.2 X 10(-9) M, with 10(-7) M proinsulin whereas glucagon had no effect. Under our experimental conditions, no degradation of [125I]-insulin was observed. Autoradiograms obtained by apposition of LKB 3H-Ultrofilm showed a widespread distribution of [125I]-insulin in rat CNS. However, quantitative analysis of the autoradiograms with 10(-10) M of labeled insulin, showed a high number of [125I]-insulin binding sites in the choroid plexus, olfactory areas, in both cerebral and cerebellar cortices, the amygdaloid complex and in the septum. In the hippocampal formation, the dorsal dentate gyrus and various subfields of CA1, CA2 and CA3 were labeled. Moreover, arcuate, dorso- and ventromedial nuclei of the hypothalamus contained high concentrations of [125I]-insulin whereas a low density was observed in the mesencephalon. The metabolic role of insulin in the CNS is supported by the large distribution of insulin binding sites in the rat brain. However, the presence of high affinity binding sites in selective areas involved in perception and integrative processes as well as in the regulation of both feeding behavior and neuroendocrine functions, suggests a neuromodulatory role of insulin in the brain.     

Biosynthesis of betagranin in pancreatic beta-cells. Identification of a chromogranin A-like precursor and its parallel processing with proinsulin.

The biosynthesis of insulin and betagranin, a 20-21 kDa co-secreted chromogranin A-related protein, were investigated in isolated insulinoma cells and islets. The insulinoma tissue processed proinsulin to insulin with kinetics similar to those reported in islet tissue. Unlike islets, however, the insulinoma released almost one-quarter of the newly synthesized proinsulin into the medium 10-40 min after its formation. Betagranin was initially immunoprecipitated as a 100 kDa precursor form, which was indistinguishable from chromogranin A in size and immunoreactivity and by peptide mapping. After an initial lag of 10-20 min, the precursor was converted progressively into betagranin, which appeared to be a stable end product. Formation of betagranin and insulin from their respective precursors followed a parallel course and could likewise be inhibited by NH4+, chloroquine and monensin, added either before labelling or at any point of time up to 15 min after labelling. As with proinsulin, approximately one-quarter of the betagranin precursor was released 10-40 min after synthesis. It is concluded that betagranin is produced by limited proteolysis from a chromogranin A precursor in pancreatic beta-cells by a cellular pathway indistinguishable from that of insulin from proinsulin. Chromogranin A is highly conserved in the N-terminal region represented by betagranin, further suggesting that the biological activity of chromogranin A may reside in a derived peptide rather than in the parent molecule.     

Structural and Functional Study of the GlnB22-Insulin Mutant Responsible for Maturity-Onset Diabetes of the Young

The insulin gene mutation c.137G>A (R46Q), which changes an arginine at the B22 position of the mature hormone to glutamine, causes the monogenic diabetes variant maturity-onset diabetes of the young (MODY). In MODY patients, this mutation is heterozygous, and both mutant and wild-type (WT) human insulin are produced simultaneously. However, the patients often depend on administration of exogenous insulin. In this study, we chemically synthesized the MODY mutant [GlnB22]-insulin and characterized its biological and structural properties. The chemical synthesis of this insulin analogue revealed that its folding ability is severely impaired. In vitro and in vivo tests showed that its binding affinity and biological activity are reduced (both approximately 20% that of human insulin). Comparison of the solution structure of [GlnB22]-insulin with the solution structure of native human insulin revealed that the most significant structural effect of the mutation is distortion of the B20-B23 β-turn, leading to liberation of the B chain C-terminus from the protein core. The distortion of the B20-B23 β-turn is caused by the extended conformational freedom of the GlnB22 side chain, which is no longer anchored in a hydrogen bonding network like the native ArgB22. The partially disordered [GlnB22]-insulin structure appears to be one reason for the reduced binding potency of this mutant and may also be responsible for its low folding efficiency in vivo. The altered orientation and flexibility of the B20-B23 β-turn may interfere with the formation of disulfide bonds in proinsulin bearing the R46Q (GlnB22) mutation. This may also have a negative effect on the WT proinsulin simultaneously biosynthesized in β-cells and therefore play a major role in the development of MODY in patients producing [GlnB22]-insulin.     

Insulin receptors on rat pancreatic islets: specificity of 125 I-insulin binding

Binding sites of isolated rat pancreatic islets have been shown to interact with insulin. Employing various species-insulins, insulin analogues and substances not being structurally related to insulin, structure-specificity as well as pH- and temperature-dependence of insulin binding to rat pancreatic islets have been studied. Rat insulin displaced 125 I-insulin from its binding sites in the same concentration-dependent manner as pork insulin did, whereas the insulin analogue des-(phe-val-asp)B1-3-p-glu B4-insulin was less effective. Pork C-peptide hardly competed for binding and pork proinsulin did not compete at all. Both the species' insulins inhibited glucose (16.7 mM)-induced insulin secretion. The inhibitory effect was less when des-(phe-val-asp)B1-3-p-glu B4-insulin was employed and no inhibition of insulin secretion was observed by the use of pork C-peptide or proinsulin. Glucagon and somatostatin did not affect insulin binding. pH optimum of insulin binding appears to be in the range between 7.0 and 8.0. Binding was augmented with increasing temperature up to 37 degrees C. It is concluded that rat pancreatic islets possess insulin because binding and biological potency of substances related to insulin were in harmony. Moreover pH- and temperature-optimum of insulin binding are in a physiological range.     

Photoconversion of the Chromophore of a Fluorescent Protein from Dendronephthya sp.

A green fluorescent protein from the coral Dendronephthya sp. (Dend FP) is characterized by an irreversible light-dependent conversion to a red-emitting form. The molecular basis of this phenomenon was studied in the present work. Upon UV-irradiation at 366 nm, the absorption maximum of the protein shifted from 494 nm (the green form) to 557 nm (the red form). Concurrently, in the fluorescence spectra the emission maximum shifted from 508 to 575 nm. The green form of native Dend FP was shown to be a dimer, and the oligomerization state of the protein did not change during its conversion to the red form. By contrast, UV-irradiation caused significant intramolecular changes. Unlike the green form, which migrates in SDS-polyacrylamide gels as a single band corresponding to a full-length 28-kD protein, the red form of Dend FP migrated as two fragments of 18- and 10-kD. To determine the chemical basis of these events, the denatured red form of Dend FP was subjected to proteolysis with trypsin. From the resulting hydrolyzate, a chromophore-containing peptide was isolated by HPLC. The structure of the chromophore from the Dend FP red form was established by methods of ESI, tandem mass spectrometry (ESI/MS/MS), and NMR-spectroscopy. The findings suggest that the light-dependent conversion of Dend FP is caused by generation of an additional double bond in the side chain of His65 and a resulting extension of the conjugated system of the green form chromophore. Thus, classified by the chromophore structure, Dend FP should be referred to the Kaede subfamily of GFP-like proteins.     

Processing of synthetic pro-islet amyloid polypeptide (proIAPP) 'amylin' by recombinant prohormone convertase enzymes, PC2 and PC3, in vitro.

Islet amyloid polypeptide (IAPP), amylin, is the constituent peptide of pancreatic islet amyloid deposits which form in islets of Type 2 diabetic subjects. Human IAPP is synthesized as a 67-residue propeptide in islet beta-cells and colocalized with insulin in beta-cell granules. The mature 37-amino acid peptide is produced by proteolysis at pairs of basic residues at the C- and N-termini of the mature peptide. To determine the enzymes responsible for proteolysis and their activity at the potential cleavage sites, synthetic human proIAPP was incubated (0.5-16 h) with recombinant prohormone convertases, PC2 or PC3 at appropriate conditions of calcium and pH. The products were analysed by MS and HPLC. Proinsulin was used as a control and was cleaved by both recombinant enzymes resulting in intermediates. PC3 was active initially at the N-terminal-IAPP junction and later at the C-terminus, whereas initial PC2 activity was at the IAPP-C-terminal junction. Processing at the basic residues within the C-terminal flanking peptide rarely occurred. There was no evidence for substantial competition for the processing enzymes when the combined substrates proinsulin and proIAPP were incubated with both PC2 and PC3. As proinsulin cleavage is sequential in vivo (PC3 active at the B-chain-C-peptide junction, followed by PC2 at A chain-C-peptide junction), these data suggest that proteolysis of proIAPP and proinsulin is coincident in secretory granules and increased proinsulin secretion in diabetes could be accompanied by increased production of proIAPP.     

Synthesis of a human insulin gene. VII. Synthesis of preproinsulin-like human DNA, its cloning and expression in M13 bacteriophage

A 74-bp DNA sequence coding for the pre sequence of human preproinsulin and containing EcoRI termini was synthesized by the chemical enzymatic method, joined with previously synthesized proinsulin DNA, and cloned in the M 13mp8 vector. A clone pNB82 -121 was identified by DNA sequence which confirmed the correct orientation of the pre sequence to the proinsulin DNA. The EcoRI site at the junction of pre- and proinsulin DNA was eliminated by removing a triplet ATT using a synthetic 19-mer primer. To simplify preproinsulin isolation and to study its expression in the M 13 system, a 25-bp affinity leader sequence coding for (glu)7 was inserted at the remaining EcoRI site; this put the preproinsulin DNA in a correct reading frame with the AUG initiation codon of beta-galactosidase. Preproinsulin was expressed under lac promoter control as analyzed by a radioimmunoassay (RIA) against C-peptide.     

Characterization of insulin binding to slices of slow and fast twitch skeletal muscles in the rabbit

Insulin binding was studied in rabbit semimembranosus proprius and psoas major muscles composed of slow-twitch oxidative (SO) and fast-twitch glycolytic (FG) fibers, respectively. For this purpose, we developed a technique using cryostat microtome muscle slices. Degradation of 125(I)-insulin during the incubation period was prevented by the addition of 1 mM bacitracin in the buffer. Specific binding to muscle slices plateaued by the 24 hrs. of incubation at 4 degrees C. It increased as a function of the amount of muscle, with a maximum binding occurring at about 5 mg of muscle slices. Triton X-100 has been shown to increase specific binding from a critical concentration of 10(-4) M with a maximum effect occurring at 3.3 10(-4) M. Under this condition, the binding was specific since displacement studies showed no inhibition of 125(I)-insulin binding by GH, HCG, ACTH and glucagon, whereas half maximal inhibition was achieved using 5 10(-10) M insulin, 3 10(-9) M IGF1 and 2 10(-8) M proinsulin. The analysis of the binding data yielded curvilinear Scatchard plots. The number of high affinity insulin receptors was higher in the SO muscle than in the FG muscle (4.3 +/- 0.7 vs 0.7 +/- 0.2 fmol/mg fresh muscle; P less than 0.001) with similar high affinity dissociation constants (Kd = 1.5 10(-10) M). Analogous results were obtained using muscle microsomal fractions. The differences in insulin binding might be related to the more intense metabolism of SO fibres which contract more often than FG fibres in vivo.     

Processing of proinsulin by transfected hepatoma (FAO) cells.

Rat hepatoma (FAO) cells were stably transfected with the gene encoding either rat proinsulin II (using the DOL retroviral vector) or human proinsulin (using the RSV retroviral vector). Using the DOL vector, production of insulin immunoreactive material was stimulated up to 30-fold by dexamethasone (5 x 10(-7) M). For both proinsulins, fractional release of immunoreactive material relative to cellular content was high, in keeping with the absence of any storage compartment for secretory proteins in these cells. Pulse-chase experiments showed kinetics of release of newly synthesized products in keeping with release via the constitutive pathway. High performance liquid chromatography analysis showed immunoreactivity in the medium distributed between three peaks. For rat proinsulin II, the first coeluted with intact proinsulin; the second coeluted with des-64,65 split proinsulin (the product of endoproteolytic attack between the insulin A-chain and C-peptide followed by trimming of C-terminal basic residues by carboxypeptidase); the third (and minor peak) coeluted with native (fully processed) insulin. For human proinsulin, by contrast, the second peak coeluted with des-31,32 split proinsulin (split and trimmed at the B-chain/C-peptide junction). Analysis of cellular extracts showed intact proinsulin as the major product. The generation of the putative conversion intermediates and insulin was not due to proteolysis of proinsulin after its release but rather to an intracellular event. The data suggest that proinsulin, normally processed in secretory granules and released via the regulated pathway, may also be processed, albeit less efficiently, by the constitutive pathway conversion machinery. The comparison of the sites preferentially cleaved in rat II or human proinsulin suggests cleavage by endoprotease(s) with a preference for R/KXR/KR as substrate.     

Study of the structure of the thylakoid membrane-bound chloroplast coupling factor CF1.

The structure of thylakoid membrane-bound chloroplast coupling factor CF1 was studied by limited proteolysis followed by sodium dodecylsulfate polyacrylamide gel electrophoresis and N-terminal sequence analysis. The N-terminal fragment of the alpha-subunit was shown to have an exposed area including the peptide bond R21-E22. The cleavage of this peptide bond caused the alphaK24-V25 bond to be exposed to the outside. In the N-terminal fragment of the beta-subunit, the L14-E15 bond was identified and found to be subject to trypsinolysis. Also, the alphaR140-S141, alphaG160-R161, and betaG102-G103 bonds were accessible to the proteolytic attack. In general, the beta-subunit of membrane-bound CF1 is more sensitive to proteolysis than that of solubilized CF1. The products of proteolysis of the alpha-subunit did not contain the polypeptides typical of the reaction of cleavage of the alphaE17-G18 and alphaE22-V23 bonds in isolated CF1. These results suggest a significant structural difference between soluble and membrane-bound CF1. A number of peptide bonds, alphaG160-R161 in particular, were shown to be shielded from proteolytic attack by papain in illuminated thylakoid membranes, probably as a result of membrane energization. In contrast, the light-induced reduction of the gamma-subunit caused an increase in the accessibility of some peptide bonds to this protease, including the alphaG160-R161 bond.     

Renaturation of human proinsulin--a study on refolding and conversion to insulin

The production of human proinsulin in Escherichia coli usually leads to the formation of inclusion bodies. As a consequence, the recombinant protein must be isolated, refolded under suitable redox conditions, and enzymatically converted to the biologically active insulin. In this study we describe a detailed in vitro renaturation protocol for human proinsulin that includes native structure formation and the enzymatic conversion to mature insulin. We used a His(8)-Arg-proinsulin that was renatured from the completely reduced and denatured state in the presence of a cysteine/cystine redox couple. The refolding process was completed after 10-30 min and was shown to be strongly dependent on the redox potential and the pH value, but not on the temperature. Refolding yields of 60-70% could be obtained even at high concentrations of denaturant (3M guanidinium-HCl or 4M urea) and protein concentrations of 0.5mg/ml. By stepwise renaturation a concentration of about 6 mg/ml of native proinsulin was achieved. The refolded proinsulin was correctly disulfide-bonded and native and monomeric as shown by RP-HPLC, ELISA, circular dichroism, and analytical gel filtration. Treatment of the renatured proinsulin with trypsin and carboxypeptidase B yielded mature insulin.     

Gc (vitamin D-binding protein) binds the 33.5 K tryptic fragment of actin

Limited proteolysis of G-actin was performed with trypsin and chymotrypsin to compare the binding sites for Gc and DNase. DNase I bound to the N-terminal area corresponding to the major cleavage site on G-actin (residues 62-68) and inhibited proteolysis, but did not bind the 33.5K C-terminal fragment (G-actin33.5) generated. In contrast, Gc did not exert any inhibitory effect upon proteolysis of the intact native G-actin42.0 molecule, although its presence protected G-actin33.5 from further proteolysis. This was shown by gel filtration to be due to the formation of complexes between Gc and G-actin33.5.     

Disulphide bridge formation of proinsulin fusion proteins during secretion in Streptomyces

To study disulphide bridge formation by Streptomyces lividans TK 24 in secreted single chain precursors of insulin a fusion protein (PTF 1) was investigated consisting of monkey proinsulin and the aminoterminal sequence Asp1 to Gly43 of the alpha-amylase inhibitor tendamistat from Streptomyces tendae. The purified soluble protein PTF 1 has a molecular mass of 14.4 kDa. The primary structure was elucidated after digestion with lysyl endopeptidase and fragment analysis. In this system, disulphide bond formation occurs in a way that the first cysteine in proinsulin is linked to the next following cysteine in the amino-acid chain resulting in a non-natural folding of the insulin part of the fusion protein. Re-folding of PTF 1 by reduction and re-oxidation followed by proteolytic digestions led to insulins which are identical to authentic material. The ease of correct disulphide bond formation in solution and incorrect processing during secretion suggests involvement of yet unknown factors leading to an unfavourable folding of proinsulin.     

Single turnovers of the EcoRI restriction endonuclease.

Limited proteolysis of the arom enzyme complex of Neurospora crassa by trypsin or subtilisin yielded a stable fragment of Mr 68000. This fragment, which was purified by two-dimensional polyacrylamide-gel electrophoresis, was shown by activity staining to contain the shikimate dehydrogenase active site, and by substrate labelling with 3-dehydroquinate and NaB3H4 to contain the 3-dehydroquinase active site. The fragment thus constitutes a bifunctional domain containing the two enzymic activities that are known, from genetic evidence, to be located adjacently at the C-terminal end of the pentafunctional arom polypeptide.