James Collio works out a biochemical method of purifying insulin.

James Collip was a young, hardworking scientist.  Born in 1892, he had just turned 29.  He had obtained his PhD from the University of Toronto, then moved to the University of Alberta.  After spending seven years in Edmonton at the University of Alberta, he took a year of sabbatical back at the University of Toronto working with MacLeod.  He began his sabbatical work on an unrelated project in April 1921, but in the fall MacLeod suggested he tackle the problem of purifying insulin from a pancreas.

The problem is straightforward.  A pancreas is ground up into small bits; the process is called homogenization.  Once the tissue is homogenized, it contains a uniform mixture of cells, proteins, fat, and other molecules.  The separation process is to separate the mixture into two fractions and determine which fraction contains the insulin.  The fraction containing insulin is retained, while the other is discarded.  To increase purity, the process can then be repeated by somehow splitting the insulin fraction into two fractions, again determining which fraction contains the insulin, and once again retaining the insulin fraction while discarding the other.

Standard ways to separate a complex solution into two fractions exist.  One is to mix the homogenized tissue with an alcohol/water mixture.  Some molecules will fall out of the solution, or precipitate, while others will stay in solution.  The fraction of alcohol in the mixture can be varied, to pinpoint exactly at what concentration the insulin will fall out of solution.

While the separation of a complex solution into two fractions was standard laboratory technique at the time, determining which fraction held the insulin was not.  Collip leveraged off of Banting and Best’s work to develop an assay that measured insulin activity.  Starting with insulin isolated using Banting’s method, he found that insulin administered to a rabbit would drop its blood glucose concentration.  Then, he used the rabbit assay to test each fraction for the presence of insulin.

Within a few short months, Collip discovered a method to purify insulin based on its solubility in alcohol.  He found that insulin remained in solution at an alcohol concentration of 80%, but fell out of solution at a concentration of 90%.