George Walden spent much of September trying to solve the issue. After carefully controlling all parameters of the process, he discovered that the stability of insulin was dependent upon the hydrogen ion concentration, the pH, of the solution in which it was contained. By keeping the pH of the solution within a well controlled range, he was able to prevent the degradation of the insulin.
Once he was aware that pH had an important effect on insulin in solution, Walden made an even more startling discovery. They had been keeping the wrong fraction! Recall that biochemical purification proceeds by separating the solution into two fractions. In Toronto, Collip had found that by mixing the pancreatic extract with 80% alcohol and 20% water, the insulin was contained in the solution and not the precipitate.
Walden discovered that the solubility of insulin not only depends upon the content of alcohol, but also upon the pH of the solution. Insulin, and all proteins, can be either positively or negatively charged in solution. The charge they carry in solution is a function of the solution’s pH. At low pH, they will be positively charged. At high pH, they will be negatively charged. The crossover point, where the protein has zero charge, is called the isoelectric point for the protein. It is an inherent characteristic of the protein. The isoelectric point of bovine insulin is 5.7, while the isoelectric point of porcine insulin is 6.0. The isoelectric point is also where the protein has minimum solubility. Since they have no net charge, proteins in solution tend to clump together and fall out of solution.
When purifying insulin using Collip’s alcohol method, the pH was uncontrolled. The problems that beset production occurred because the pH was coincidentally near 5.7 in their 80/20 alcohol solution. Collip, and Walden, were expecting the insulin to remain in the 80/20 solution, but because the pH was near the isoelectric point, much of the insulin was precipitating out of solution and ended up in the wrong fraction. This effect led to an apparent overall loss of insulin activity. [ http://www.clinchem.org/content/48/12/2270.full]
Walden’s discovery of isoelectric precipitation was the final step in solving the insulin problem. It allowed Eli Lilly to scale up production in early 1923. By April of 1923, Eli Lilly was producing 180,000 units of insulin per month over 100 times as much as they had been producing ten months earlier. This production level provided enough insulin for any patient who needed it.