Validation of a method for monitoring the manufacture of human insulin

A method for monitoring the manufacture of human insulin by HPLC was developed. The method was validated by the estimation of its linearity, accuracy, precision, specificity, and robustness; the limits of detection and quantitative assessment were also determined. It was proven that HPLC analysis enables reliable and reproducible results to be obtained and can be used for monitoring insulin manufacture.     

Recombinant human insulin V Optimization of the reversed-phase high-performance liquid chromatographic separation

The applicability of reversed-phase high-performance liquid chromatography (HPLC) to the analysis of the products of recombinant insulin was studied. The influence of several mobile phases in reversed-phase and ion-pair HPLC on selectivity, resolution and sensitivity was investigated. Optimum conditions for the separation of insulin-related proteins on commercial and laboratory-made supports were established by means of three-dimensional optimizations of selectivity and resolution as a function of pH and ionic strength (μ). A mechanism for the separation of proteins with a mobile phase containing a high salt concentration and a pH near the isoelectric point of proteins is proposed. The questions of scaling up are considered. The proposed techniques allow the analysis of the main impurities and ensures a high quality of active insulin production.     

Additives to biological substances. IV--Lyophilization conditions in the preparation of international standards: an analysis by high-performance liquid chromatography of the effects of secondary desiccation

Many International Standards, Reference Preparations and Reference Reagents are routinely prepared by lyophilization in the presence of 'inert' carriers followed by an extensive period of secondary desiccation. In this study we have used high-performance liquid chromatography (HPLC) to analyse the effects of lyophilization and secondary desiccation on initial degradation and subsequent stability of a model protein, insulin. Secondary desiccation was found to promote a reaction of the insulin with a carrier consisting of non-volatile buffer salts and a sugar. Secondary desiccation did not improve the stability of the insulin as determined by accelerated thermal degradation and analysis using the Arrhenius equation. We conclude that careful consideration needs to be given, on a case-by-case basis, to the selection of the procedures for the preparation of International Standards, particularly those ampouled in the absence of carrier proteins and intended for physicochemical analysis such as HPLC.     

Insulin-stimulating peptide from tryptic digest of bovine serum albumin

Insulin-stimulating peptide was isolated from a tryptic digest of bovine serum albumin by gel permeation, SP Sephadex column chromatography, reversed phase HPLC and cation-exchange HPLC. This peptide, with a molecular weight of about 8,400, had no insulin-like activity by itself, but enhanced fatty acid synthesis from glucose in rat adipose tissue explants in the presence of suboptimal concentrations of insulin. It also stimulated the effect of insulin on CO2 production from glucose in rat adipocytes, without affecting insulin binding. These stimulations were dose-dependent and were observed at concentrations of more than 2 X 10(-7) M peptide only in the presence of a suboptimal concentration of insulin.     

Purification and functional characterization of pancreatic insulin from camel (Camelus dromedarius)

Large-scale production of insulin still represents the key step in helping diabetic patients throughout the world. Many species and approaches have been used for the production of insulin. In this study, we purified and characterized for the first time pancreatic insulin from the Arabian camel (Camelus dromedarius) using a modified acid-alcohol extraction method. After extraction insulin was purified using a one-step gel filtration on a Sephadex G-50 column leading to a purification yield of 80 mg/kg (20%) of camel pancreas. The purity of camel insulin was assessed by SDS–PAGE and HPLC using insulin from human, bovine and porcine as standards. Molecular weight was determined for purified camel insulin as 5800 Daltons and its amino acid composition is similar to that known for other species. The functional characterization of purified crude camel insulin was demonstrated in vitro by positive competition by radioimmunoassay and in vivo showing camel insulin inducing acute hypoglycaemia in mice. Together, our study reports for the first time the successful purification of functional insulin from camel pancreas with similar properties compared to other insulin species. This is of great interest given that the camel represents considerable economic worth in many countries.     

Analytical biotechnology of recombinant peptides and proteins. I. Analysis of the purity, composition and structure of human, porcine and bovine insulins

A method for analysis of the type, purity, and possible structural modifications of insulins of bovine, porcine, and human origin was proposed. It is based on a combination of narrow-bore reversed-phase HPLC and mass spectrometry. The hydrolysis of insulins with highly specific Glu-protease V8 from Staphylococcus aureus followed by peptide mapping of the hydrolysis products and mass spectrometry of the isolated fragments helps rapidly and reliably localize and identify substitutions of amino acid residues in insulin structure by using insulin samples of less than 1 nmol.     

Distinct alpha-subunit structures of human insulin receptor A and B variants determine differences in tyrosine kinase activities.

Human insulin receptor isoforms (HIR-A and -B) differ in their alpha-subunit structures which result from alternatively spliced precursor mRNAs. This structural difference causes distinct binding affinities for insulin. To determine the impact of the structural difference on receptor signaling, we characterized the tyrosine kinase activity of HIR-A and HIR-B in vitro and determined the insulin stimulated beta-subunit phosphorylation and tyrosine kinase activation in the intact cell. When 32P incorporation in beta-subunits of equal amounts of isolated HIR-A and HIR-B was measured, an increased 32P incorporation in tyrosine residues of the beta-subunit of HIR-B (2.5-fold) compared to that of HIR-A was found after in vitro insulin stimulation. This was paralleled by an increased rate of phosphorylation (2.0-fold) or poly(GluNa,Tyr 4:1). In vitro analysis of Km values for ATP were similar for HIR-A (Km = 14.3 microM +/- 3.8) and HIR-B (Km = 20.2 microM +/- 8.6), whereas the Vmax of HIR-B was significantly increased (HIR-A Vmax = 5.5 mumol/60 min micrograms-1 +/- 1.4, HIR-B Vmax = 42.5 mumol/60 min micrograms-1 +/- 19.2). HPLC analysis of tryptic beta-subunit phosphopeptides revealed identical patterns, suggesting that the difference in kinase activities is not due to an alteration of the phosphorylation-activation cascade within the beta-subunit. However, when cleavage of the alpha-subunit by short-time trypsinization was used to activate the tyrosine kinase, the differences in 32P incorporation between HIR-A and HIR-B were abolished.(ABSTRACT TRUNCATED AT 250 WORDS)     

Synthesis of biologically active porcine secretin and [ITyr10] porcine secretin.

Porcine secretin, [Tyr10] secretin, and [Tyr13] secretin were synthesized by solid phase methodology and purified by stepwise gradient elution from a short reversed-phase column with ethanol and acetic acid as organic modifiers. [Tyr10] secretin and [Tyr13] secretin were iodinated by the chloramine-T method and nonmono-, and di-iodinated products separated and isolated by reversed-phase HPLC. Batch incubation analysis is isolated mouse pancreatic islets revealed that secretin and the [Tyr10] analogue were indistinguishable in their effect on the glucose-induced insulin release and cAMP accumulation. [Tyr13] secretin in contrast was significantly less potent in its effect on the glucose-induced insulin release.     

Methods for protecting silica sorbents used in high-performance liquid chromatography from strongly adsorbed impurities during purification of human recombinant insulin

One of the main stages of human recombinant insulin (HRI) production is the hormone purification by means of reversed phase high-performance liquid chromatography (RP HPLC). The optimization of this stage determines the increase of the total manufacturing yield. Therefore, the cost of the sorbent used in HPLC influences the cost of the manufacturing product, i.e. HRI substance. However, resolution between HRI and its admixtures decreases with time. The reason for this is the sorbent contamination with strongly adsorbed impurities (SAI) which are accumulated during elution. In the following research several methods for sorbent protection are studied. The opinion that SAI are mainly high-molecular weight compounds was examined using gel filtration. Different sorbent types were tested for the use in guard column. The results obtained were applied and improved at preparative level.     

Implementation of a thermal biosensor in a process environment: on-line monitoring of penicillin V in production-scale fermentations.

The production of penicillin V was monitored in 0.5 m3 and 160 m3 bioreactors. The thermal biosensor was an enzyme thermistor modified for split-flow analysis. The heat signal generated in the enzyme column was corrected for any nonspecific heat with the use of an identical but inactive reference column. The on-line monitoring was performed in the fermentation pilot plant and in a fermentation plant of Novo Nordisk A/S. Immobilized beta-lactamase was used to monitor three consecutive 0.5 m3 penicillin fermentations. Broth samples were continuously filtered through a tangential flow filtration unit in a sterile external loop. The on-line penicillin V values were 10% higher than those obtained by off-line HPLC analysis. Alternatively a polypropylene filtration probe was inserted into a 160 m3 bioreactor and samples were withdrawn at 0.5 ml/min. The same experiments were repeated with purified and immobilized penicillin V acylase. The on-line penicillin V values obtained with this enzyme correlated very well with those from HPLC analysis. The on-line monitoring was controlled and analysed by a software program written in Labtech Notebook.     

Effects of cysteine to serine substitutions in the two inter-chain disulfide bonds of insulin

Using site-directed mutagenesis we deleted the two inter-chain disulfide bonds of insulin, separately or both, by substitution of the cysteine residues with serine. Deletion of A20-B19 or both of the two inter-chain disulfide bonds resulted in the complete loss of secretion of the mutant single-chain porcine insulin precursor (PIP) from Saccharomyces cerevisiae cells. Removal of the A7-B7 disulfide bond resulted in a large reduction of secretion, but we could obtain the mutant for analysis of its biological and some physico-chemical properties. The A7-B7 disulfide bond deleted insulin mutant retained only 0.1% receptor-binding activity compared with porcine insulin, and its in vivo biological potency measured by mouse convulsion assay was also very low. We also studied some physico-chemical properties of the mutant using circular dichroism, native polyacrylamide gel electrophoresis and reversed-phase HPLC, which revealed some structural changes of the mutant peptides compared to native insulin. The present study shows that the two inter-chain disulfide bonds are important for efficient in vivo folding/secretion of PIP from yeast, especially the A20-B19 disulfide bond, and that the A7-B7 disulfide bond is crucial for maintaining the native conformation and biological activity of insulin.     

Biogenic amine production by wild lactococcal and leuconostoc strains

Two qualitative and one quantitative HPLC methods were evaluated for the detection of biogenic amine producers among wild dairy lactococcal and leuconostoc strains. High tyramine producers ranging from 370 to 807 mg l⁻¹ were detected by qualitative methods and confirmed by HPLC analysis. Tyramine levels detected throughout the incubation time depended on the concentration of the amino acid precursor available and no tyramine production was observed when strains were grown in milk. However, increasing amounts of tyramine were detected in cultures grown in milk supplemented with different concentrations of tyrosine. Qualitative methods failed to detect weak producers so that tryptamine production (<7 mg l⁻¹) could only be determined by HPLC. None of the tested strains was able to produce histamine. Simultaneous production of different amines was observed by HPLC although no colour change was observed in the specific decarboxylase media. Thus, it was concluded that the amine forming ability should be taken into account when selecting starters for milk fermentations. Qualitative methods could be used as a first screening step to eliminate the highest amine producers while the quantitative methods would detect any producing strain.     

Detection of a new hormone contact site within the insulin receptor ectodomain by the use of a novel photoreactive insulin.

We have used a preparation of soluble human insulin receptor ectodomain and a novel photoreactive, biotinylated derivative of insulin (4-azidosalicyloyl(B1-biocytinyl-B2-lysine)-insulin) to identify a new hormone contact site within the extracellular domain of the insulin receptor. The ectodomain was photoaffinity-labeled and digested to completion with trypsin, and the resulting tryptic fragment was purified by either HPLC or by streptavidin-affinity chromatography. The amino terminus of the fragment was identified as Gly390 within the alpha-subunit. These results suggest that residues that are carboxyl-terminal to the cysteine-rich domain, in addition to previously identified regions within the amino terminus of the alpha-subunit, contribute to the insulin binding site. The implications of these results for the de novo folding of the insulin receptor to constitute the hormone binding site are discussed.     

Solid phase isotope exchange of deuterium and tritium for hydrogen in human recombinant insulin

Reaction of a high-temperature solid-phase catalytic isotope exchange in peptides and proteins under the action of the catalytically activated spillover hydrogen was studied. The reaction of human recombinant insulin with deuterium and tritium at 120–140°C resulted in an incorporation of 2–6 isotope hydrogen atoms per one insulin molecule. The distribution of the isotopic label by amino acid residues of the tritium-labeled insulin was determined by the oxidation of the protein S-S-bonds by performic acid, separation of polypeptide chains, their subsequent acidic hydrolysis, amino acid analysis, and liquid scintillation counts of tritium in the amino acids. The isotopic label was shown to be incorporated in all the amino acid residues of the protein, but the higher inclusion was observed for the FVNQHLCGSHLVE peptide fragment (B_1–13) of the insulin B-chain, and the His5 and His10 residues of this fragment contained approximately 45% of the whole isotopic label of the protein. Reduction of the S-S-bonds by 2-mercaptoethanol, enzymatic hydrolysis by glutamyl endopeptidase from Bacillus intermedius, and HPLC fractionation of the obtained peptides were also used for the analysis of the distribution of the isotopic label in the peptide fragments of the labeled insulin. Peptide fragments which were formed after the hydrolysis of the Glu-Xaa bond of the B-chain were identified by mass spectrometry. The mass spectrometric analysis of the isotopomeric composition of the deuterium-labeled insulin demonstrated that all the protein molecules participated equally in the reaction of the solid-phase hydrogen isotope exchange. The tritium-labeled insulin preserved the complete physiological activity.     

Production and separation of peptides from proteins stained with Coomassie brilliant blue R-250 after separation by sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

Proteins stained with Coomassie brilliant blue on polyacrylamide gels were digested with lysylendopeptidase in the presence of sodium dodecyl sulfate. Peptide production was similar to that under ordinary conditions of digestion. Peptides were recovered easily and efficiently from the gel pieces and separated by HPLC. The present method for preparation of peptides from proteins separated by sodium dodecyl sulfate gel electrophoresis is quite simple and can be used for sequence analysis of proteins in general at the subnanomolar level.     

Comparison of the T cell receptors on insulin-specific hybridomas from insulin transgenic and nontransgenic mice. Loss of a subpopulation of self-reactive clones.

Transgenic mice expressing the human insulin gene do not produce insulin-specific antibody after injection of human insulin. Nevertheless, they have some peripheral T cells that proliferate to human insulin in vitro. To investigate the nature of these T cells, human insulin-specific T cell hybridomas were produced from transgenic and nontransgenic mice. Transgenic hybridomas required more insulin to achieve maximum responses and they produced lower levels of lymphokines than nontransgenic hybridomas. The majority of nontransgenic hybridomas recognized only human and pork insulin whereas transgenic hybridomas recognized beef, sheep, and/or horse insulin in addition to human and pork insulin. The TCR expressed by transgenic and nontransgenic hybridomas were determined by Northern analysis. Both types of hybridomas used several different V alpha and V beta gene families and no favored association between V alpha and V beta gene usage was detected in either type. V beta 1 was used by 7 of 16 nontransgenic hybridomas but only by 1 of 16 transgenic hybridomas. V beta 6 receptors were predominantly expressed by the transgenic hybridomas and all V beta 6-bearing hybridomas recognized beef as well as human insulin. The differences in Ag reactivity and TCR gene usage suggest that V beta 1-bearing human insulin-reactive T cells were clonally deleted or inactivated in the transgenic animal. Other clones, representing a minor subpopulation in nontransgenic mice, were recovered from transgenic mice.     

Synthesis and characterization of photoreactive insulin-like growth factor 1 derivatives for receptor analysis

Recombinant human insulin-like growth factor 1 (hIGF-1) was reacted with 4-azidosalicylic acid (Asa) to synthesize photoreactive IGF-1 derivatives. Depending on the time of reaction and the Asa/IGF-1 ratio, a mixture of mono-, bis-, and trisacylated derivatives and one tetraacylated derivative was produced, which was separated by reversed-phase HPLC. HPLC, PAGE, and MALDI mass spectrometry were used to determine the degree of the acylation and location of the photolabel insertion. After iodination of the three monoacylated photoreactive IGF-1 derivatives, the specific labeling of the receptor could be proved. Together with a comparative investigation in which B29-Asa-insulin was used, the results suggest corresponding contact areas for IGF-1 and insulin with the insulin receptor ectodomain.     

Enzymatic semisynthesis of porcine despentapeptide (B26-30) insulin using unprotected desoctapeptide (B23-30) insulin as a substrate. Model studies

Unprotected porcine desoctapeptide(B23-30) insulin (DOPI) and the synthetic Gly-Phe-Phe were used as substrates for the trypsin-catalyzed synthesis of despentapeptide(B26-30) insulin (DPPI). The DPPI synthesis was accompanied by a moderate oligomerization and by the formation of a side produce which was identified as a DOPI derivative having an extra peptide bond between the Gly(A1) and Arg(B22) and which was named des(23-63) proinsulin (1). Despite side reactions, the conditions were found where the overall DPPI yields were comparable to those obtained via di-Boc DOPI, and these procedures were faster and simpler since the Boc protection and deprotection steps were omitted. The reaction progress was directly monitored by HPLC.