The Insulin Approach in
Protein Folding Research
Report by Holger Blasum
Composed on Nov. 28, 1991, for the Seminar "Enzymes and
Enzyme-Related Problems" by Professor Arnold Schwartz
at Bonn University in Winter Term 1991/92.
I. Introduction: Insulin Research
II. Hormonal Mechanisms of Insulin
III. Pancreatic Biosynthesis
IV. Molecular Structure
V. Investigating Protein Chain Interaction
I. Introduction: Insulin Research
As the insulopenia diabetes mellitus is a very widespread disease with grave consequences, it is not surprising that the interest in developing methods to cure it has lead to intensive world-wide research.
The Greek term diabetes (meaning "running through") was coined about 2000 years ago in a desciption by Aretaios. However, the necessary foundations for a scientific understanding of the disease were not laid until the second half of the 19th century. In 1869 while studying the pancreas the pathologist Langerhans discovered small islets of cells that were phenomenologically different from the surrounding tissue. Their function remained unclear. It was only after further evolution of endocrinology that the investigation of their secretory functions began, leading to the suggestion that bovine insulin could be used to cure human diabetes. Banting and Best finally managed to isolate applicable insulin in 1922; this work was promptly honoured in 1923 with the awarding of the Nobel Prize to MacLeod (leader of the institute) and Banting. The large application of insulin stimulated research so that in 1953 its amino acid sequence was one of the first to be cleared up by Sanger (Nobel Prize winner in 1956). In the mid-sixties the total synthesis of insulin could be achieved simultaneously by research groups in the USA, Germany and China, but this process has remained until today far more expensive than extraction from bovine or porcine pancreas. The animal insulin gene was among the first to be isolated in 1977, by genetic engineering human insulin can be produced by Escherichia coli since 1978.
In addition to such spectacular achievements, these vast endeavours also provided new insights into the biosynthesis of proteins. A factor leading to this was the discovery of proinsulin by Steiner in 1967 (Tager, Steiner and Patzeld 1981:73) which had important impacts on disulphide bond research. After a brief presentation of the presently known basic principles of insulin mechanisms and biosynthesis in the subsequent two sections we will thus resume the discussion about insulin disulfide formation...
II. Hormonal Mechanism of Insulin
Insulin is a hormone which has become well known for its function as a regulator of glucose levels in blood. High glucose level in blood and thereby enlarged glucose concentration in the pancreatic á-cells induce the release of the hormone insulin. Insulin molecules are transported by the blood circulation to membranes of liver or muscle cells. Interacting with receptors, they enable glucose and cations (sodium, potassium and calcium) to cross the membranes. Thus, by change of membrane potential and phosphorylation (Roth 1990:170) they also activate a cascade of second messengers (including G proteins, diacylglycerol and inositol triphosphate; Hollenberg 1990:187), which e.g. in complex insulin-dependant concurrence regulate lipogenesis (Zhang et al. 1990:60), gluconeogenesis and the transport of amino acids.
The most obvious effect caused by the absence of insulin is represented by the adjective in medical term "diabetes mellitus": The postprandial glucose levels in blood cannot be transfomed into highly energetic anabolic products as glycogen or fats, so the glucose is in a large extent secreted by diuresis. This diurnal secretion may reach several hundred grammes causing a vast dehydration. Consequently, energy has to be provided by the extreme oxidation of available material so that many keto groups are formed causing serious damage by their high chemical activity.
III. Pancreatic Biosynthesis
The only site of insulin gene transcription is in the pancreas á-cells. These secrete several products such as chromogranin A or an islet amyloid peptide (IAP; Steiner 1990:67), but insulin is by far the dominant secretion. The insulin gene transcription and translation result in the formation of a peptide of 110 amino acids (human) called "preproinsulin". However, a lipophilic sequence consisting of 24 amino acids is simply used to pass through the membrane of the endoplasmic reticulum and then cleaved so that an 86 amino acid "proinsulin" remains. In the endoplasmic reticulum protein folding seems already to be accomplished (Freedman 1984:438); soon the proinsulin passes via the cytoplasma to the Golgi vesicles, where it receives a clathrin coating so that can be stored in the cell. Within these granules the proinsulin is cleaved into the insulin (51 amino acids) and a "C-peptide" (35 amino acids). This insulin is stored as a hexameric insulin grouped around two zinc cations (Brader and Dunn 1991:341). If required, these granules will be released into the bloodstream.
IV. Molecular Structure
The insulin excreted by the á-cells is a polypeptide of 51 amino acids (ca. 6 kD) that consists of two chains called A and B (21 and 30 amino acids respectively) which are crosslinked by two cysteine-cysteine disulfide bonds. "Crosslinking" is defined as a "durable combination of (usually large) distinct elements at specific places to create a new entity that has different properties as a result of the union" (Friedman 1977:v). On the A chain, there two further cysteine residues that intracatenarily link to each other. Mammalian insulins vary slightly, usually not in the formation of the disulfide bonds, but rather in variations of different amino acids in the primary sequence: Bovine and porcine insulin differ from human insulin in three and one amino acids resp., but generally both can be used for human medication. Of course, there are some rare cases of allergic reactions to one of the forms. In this case a "dealienated" porcine insulin with the different amino acid replaced by enzymatic treatment will be used.
Other mammalian insulins differ more than these, but their industrial relevance can be neglected. Also human structural variants usually resulting in insulinopathies are found and identified (Steiner 1990:82), throwing new light on the functional necessity of certain amino acid sequences. Another approach is the in vitro creation of despeptide insulin by partial enzymatic hydrolysis, so that the declining or remaining activity can be observed (Lu and Yu 1980:1592). Nevertheless, be it that the primary structure of insulin has been known for over 30 years, its secondary and tertiary conformations are still in discussion. We will thus have a look at the mechanisms of protein folding in the insulin molecule...
V. Investigating Protein Chain Interaction
After the early decoding of the insulin amino acid sequence, speculations about the folding of this (and other) proteins began as early as in the fifies. These were dominated by the idea "that the polypeptide chain tests, by trial and error, many conformations until it reaches ist thermodynamically most stable state" (Goldberg 1985:388). In this way, mathematical models were established predicting that by random crosslinking of insulin chains the yield of correctly linked insulin would be comparably low (Kauzmann 1959:97). But at the same time experiments with messenger-ribonuclease showed that this stochastical approach was not true for all proteins, as the yield was far higher than statistically expected (Sela and Lifson 1959:477). Essential for the folding of the proteins is the formation of disulphide bonds. So experiments were performed by reducing the disulphide bonds (often by mercaptoethanol or dithiothreitol), causing the chains to break apart. Then the reduced chains were crosslinked again under oxidizing conditions (sometimes air oxidation was sufficient). Alternatively, already "scrambled" protein (i.e. a protein where this has already been once performed and which is thus having a reduced activity) can be used as basis for these experiments.
By developing techniques in amino acid sequence formation, synthetic insulin could be composed in the sixties. However, the yields were only ineconomic. When the C-peptide was discovered in 1967, the precedent concepts assuming that the insulin chains were biosynthesized seperatedly were suddenly outdated. In this way the experiments using this peptide began. In 1970 Varandani and Nafz first unscrambled scrambled proinsulin increasing its immunological reactivity by about 12.4% in the presence of the enzyme glutathione insulin-transferase (GIT; this will later be called protein disulphide-isomerase, "PDI") at 37øC. Insulin reactivity could not be increased by the same treatment. It was only in the presence of the C-peptide that the insulin activity increased (Varandani and Nafz 1970:535). So they concluded that "information present in the C-peptide is required for the folding of insulin" (op.cit., p. 536).
Recent experiments however demonstrate that insulin can be gained from scrambled insulin in the absence of the C-peptide at lower temperatures. Tang, Wang and Zou also used PDI, but they lowered the temperature down to 4øC. The obtained insulin had a yield of 25-30% from scrambled insulin originally only containing 4% of the native hormone (Tang, Wang and Zou 1988:454). By the lately developed technique of high performance liquid chromatography (HPLC) anaysis using a column with very small particles through which the eluent is pressed at 10 to 250 atm (M zor 1983:vol.15,263) very exact elution curves can be achieved. For example, the column they used contained activated carbon with a particle size of 10 æm (Tang, Wang and Zou 1988:451). Yet scrambled insulin artificially crosslinked at Gly-A1 and Lys-B29 by carbonyl-bis-methionine-bis-p-nitrophenylester ("CBM-insulin") provides even higher yields ranging from 50% to 90% (Tang and Zou 1990:432).
Wang and Zou point out that "during renaturation process, the yield of any refolded protein decreases with increasing temperature" (Wang and Zou 1991:281) which may be related with energetic relations: The loss in entropy caused by the folding of the protein is proportional to the temperature (Zaccai and Eisenberg 1990:333).
Their experiment thus proves that chemical synthesis of proteins can be optimized by temperature variation, so that in future times insulin total synthesis may become more attractive again. If this will be the case, the role of insulin research might reveal further fundamental informations about techniques in protein synthesis.
Furthermore, the possibility of reforming native disulfide bonds by relatively simple conditions (presence of one enzyme only) in the absence or replacement of the C-peptide by other substances throws new light on a debate about the ways proteins are folded. In the 1980s a model has been developed describing the protein formation by an evolution of small local structures that grow in a processus resembling the solution of a jigsaw puzzle (Harrsion and Durbin 1985:4028). This allows constant editing; many path-ways are possible to form finally the (required) most stable configuration. So by circular dichroism the formation of secondary structures in single reduced insulin chains has been observed in insulin, too (Hua, Qian and Zou 1985:854).
Yet a contrary theory has also been elaborated: As some molecules can only be folded via intermediates of an energetic level higher than themselves it has been suggrested that protein formation usually proceeds pursuing unique sequential pathways (Baldwin 1989:293). However, the possibility to produce correctly folded proteins by enzymatic reshuffling at ideal temperatures seems to give evidence of the validity of the first theory. So further practical research on appropriate conditions for protein formation may provide stimulating ideas on protein folding in general.
As world-wide research on insulin has been extensive it has provided many important ideas on methods and theories in protein research in general. However, the synthesis of insulin has not yet been achieved at reasonable yields. This is partially due to the lack of knowledge how the mechanisms of chain combination and disulfide bonds exactly work. It has been assumed that a polypeptide occuring in natural biosynthesis, the C-peptide, was necessary to provide the correct formation of the disulfide bonds. Recently, experiments have shown that at unphysiological temperatures the chain combination of insulin may be combined in the absence of the C-peptide. The high yields thus gained throw new light on technical opportunities of insulin production and on the theory of protein chain folding in general.
Robert L. BALDWIN
1989 "How Does Protein Folding get Started?", in: Trends in Biochemical Sciences (TIBS), Vol. 14 No. 7, pp. 291 - 294.
Mark L. BRADER and Michael F. DUNN
1991 "Insulin Hexamers: New Conformations and Applications", in: TIBS, Vol. 16 No. 9, pp. 341 - 345.
1990 "Insulin Chemistry", in: Pedro CUATRECASAS and Steven JACOBS [Edts.], Handbook of Experimental Pharmacology, Vol. 92, Berlin, Heidelberg, New York, pp. 3 - 22.
U. DEREWENDA et al.
1990 "Insulin Structure", in CUATRECASAS and JACOBS, op. cit., pp. 23 - 39.
Robert B. FREEDMAN
1984 "Native Disulphide Band Formation in Protein Biosynthesis: Evidence for the Role of Protein Disulphide Isomerase", in: TIBS, Vol. 9 No. 10, pp. 438 - 441.
1977 Protein Crosslinking, New York.
Michel E. GOLDBERG
1985 "The Second Translation of the Genetic Message: Protein Folding and Assembly", in: TIBS, Vol. 10 No. 10, pp. 388 - 391.
Stephen C. HARRISON and Richard DURBIN
1985 "Is There a Single Pathway for the Folding of a Polypeptide Chain?", in: Proceedings of the National Academic Society of the USA, Vol. 82, pp. 4028 - 4030.
M. D. HOLLENBERG
1990 "Insulin Receptor-Mediated Transmembrane Signalling", in: CUATRECASAS and JACOBS, op. cit., pp. 183 - 208.
HUA Qingxin, QIAN Yanqiu and ZOU Chenglu
1985 "A Circular Dichroic Study of the Interaction of Insulin", in: Scientia Sinica, Vol. 28 No. 8, pp. 854 - 862.
1959 "Relative Probabilities of Isomers in Cystine-Containing Randomly Coiled Polypeptides". in: Reinhold BENESCH et al. [Edts.], Sulfur in Proteins, New York and London, pp. 93 - 108.
LU Zixian and YU Ronghua
1980 "Preparation and Crystallization of des(B-chain c-terminal) Heptapeptide Insulin", in: Scientia Sinica, Vol. 23 No. 12, pp. 1592 - 1596.
L szl¢ MAZOR
1983 Methods in Organic Analysis, Budapest.
R. A. ROTH
1990 "Insulin Receptor Structure", in: CUATRECASAS and JACOBS, op. cit., pp. 169 - 180.
Michael SELA and Shneior LIFSON
1959 "On the Reformation of Disulfide Bridges in Proteins", in: Biochimica et Biophysica Acta, Vo,l. 36, pp. 471 - 478.
Donald F. STEINER
1990 "The Biosynthesis of Insulin", in: CUATRECASAS and JACOBS, op. cit., pp. 67 - 92.
Howard S. TAGER, Donald F. STEINER and Christoph PATZELT
1981 "Biosynthesis of Insulin and Glucagon", in: Methods in Cell Biology, Vol. 23, pp. 73 - 88.
TANG Jianguo, WANG Zhizhen and ZOU Zhenlu
1988 "Formation of Native Insulin from the Scrambled Molecule by Protein Disulphide-Isomerase", in: Biochemical Journal, Vol. 255, pp. 451 - 455.
TANG Jianguo and ZOU Zhenlu
1990 "The Insulin A and B Chains Contain Structural Information for the Formation of the Native Molecule", in: Biochemical Journal, Vol. 268, pp. 429 - 435.
P. T. VARANDANI and Mary A. NAFZ
1970 "Enzymatic Destruction of Immunoreactivity in Proinsulin and Insulin and Activation of Their Scrambled Forms", in: Archives of Biochemistry and Biophysics, Vol. 141, pp. 533 - 537.
WANG Zhizhen and ZOU Zhenlu
1991 "The Insulin A and B Chains Contain Sufficient Structural Information to Form the Native Molecule", in: TIBS, Vo. 16 No. 8, pp. 279 - 281.
Giuseppe ZACCAI and Henryk EISENBERG
1990 "Halophilic proteins and the influence of solvent on protein stabilization", in: TIBS, Vol. 15 No. 9, pp. 333 - 337.
ZHANG Shirong et al.
1990 "Lipogenesis Stimulatory and Inhibitory Activities of the Insulin Mediators", in: Scientia Sinica, Vol. 33 No. 1, pp. 60 - 66.