Oliver Smithies:

[00:00:00] This is book G.  Capital G.  Starts in 1961 and is a continuation of the work that was going on in book — the preceding book, F with splitting tests resumed.  The first page sets out the hypothesis as to what I was considering was happening at that time.  Looking at it now, it wasn’t very far wrong.  Basically you could take the haptoglobins and reduce them to a single species or virtually a single species of molecule with mercaptoethanol at a high pH.  But if you had urea then you began to split them into the two chains, [00:01:00] the constant chain and the variable chain, which we later called Hp, but which by then we were already calling little hp1 and little hp2.  So the hypothesis was fairly good and tests were continued to try to show that it was correct in the following pages for quite a while.  Tests of different pHs and so on.  So by page five it’s talking about the results suggest a big effect of Versene or [tryp?] on electrophoresis, etc.  And I’m still thinking about a resynthesis as I call it.  The gels on page seven, opposite page seven, are interpreted [00:02:00] I think as showing that there is some resynthesis.  But I think it’s probably all failure to cleave properly.  On page nine I’m talking about the results show that pH 8.2 (inaudible) mercaptoethanol shows incomplete split, etc.  Going — talking about how to get splitting into the still native monomers.  Higher pH test on page 11 July 26th looks from hindsight very promising, and so on continuing this type of work.

[00:03:00] For example page 17 shows all of the gels there now which are with Versene are showing splitting into the native component and so on and so forth.  Higher pH has been tested.  Higher mercaptoethanol and longer time.  The time (inaudible) continuing in the same vein.  Trying the effect of starch concentration on Wednesday August 2nd.  Fairly good up to 160% is practical, although difficult.  [00:04:00] Still no re-solution.  A rather obscure comment on page 29.  New buffer systems tested on page 31, lysine and phosphate.  And talking about resynthesis in both phosphates but with a caution.  Lower pH is tested and continued.  And then a little intermission.  On Saturday August 5th I took purified brilliant (inaudible) blue, lower pHs in two-dimensional gels tested in August 6th.  [00:05:00] With some thought on page 43 that resynthesis has occurred on the gel that had been run on the previous page, 41.

Constantly worrying about this resynthesis.  Wednesday August 9th.  Reversible synthesis.  Some idea that it might be done by reversing the direction of migration of the protein.  But the result on page 47.  Evidence of incomplete splitting yet reed temperature acetone synthesis and so on.  [00:06:00] Keeping struggling with it.  Repeated resynthesis sample page 55.  Many repeats of the same sort of thing.  Page 61 Saturday August 12th repeat runs on everything.  Repeated again.  Not at all trying the repeating thing.  It’s not always getting anything out of it.

For example on page 63 said if promising repeat with 1-1 and 2-2 on Tuesday.  [00:07:00] Another very promising statement.  Quite common.  Page 65.  Very promising with mercaptoethanol.  Therefore repeat the whole procedure, etc., etc.  Doesn’t look very promising from where I’m looking now.  August 15th some acetone freeze-dry repeats, which have been looked at on page 73 with a comment these dilute solution resyntheses are much superior.  From what I know was the answer I’m quite doubtful about that interpretation.  [00:08:00] The gels look as if nothing had ever happened, not that there was a resynthesis.  But now here is a big change because between page 73 and 75, 73 was Saturday August 19th, and page 75 is Sunday August — excuse me.  Sunday October 8th.  And what happened in between was that George Connell, Gordon Dixon, and I went to a meeting in Rome of the Human Genetics Society.  And we had by this time reached the conclusion that haptoglobin 2 was bigger than haptoglobin — that Hp2 was bigger [00:09:00] than Hp1.  I don’t remember where that conclusion was reached but by then we had reached that conclusion.  And Gordon Dixon and George Connell were in Toronto and I was in Wisconsin.  And we agreed that a critical test would be to — using purified Hp1, little hp1, little hp2, we would test the molecular size of the Hp2 by running an ultracentrifuge experiment, which George would do.  He was in Toronto and he could do that experiment nicely.  And then we would meet the evening before our talk in Rome and discuss [00:10:00] how we were going to present the material.  The general idea being that George would talk first about the purification and quantification of haptoglobin and the separation into different chains, the constant chain and the variable Hp1 Hp — little hp1, little hp2 subunits.  And Gordon would talk about the fingerprint tests which led us to a conclusion that Hp2 was bigger.  And we would present them in that order and then I would follow by talking about our hypothesis which by then had been formed.  [00:11:00] And so I’m going to pause here and go back a little while because I haven’t given the history here of how we got to the hypothesis.

This happened as a result of my getting ready for a talk I was going to give in Japan.  And looking very carefully at Gordon Dixon’s fingerprints, I could see there was some rather strange relationships in there that there were two peptides which were present in Hp1 fingerprints and in Hp2 fingerprints that appeared to be less common in less amount of them [00:12:00] relative to the total amount of protein, less amount of these two peptides in Hp2 than there was in Hp1.  And there was an additional peptide in Hp2 which was not present in Hp1.  And that peptide had amino acids that looked to be similar in composition to the two peptides which were in reduced amount.  And so it looked as if the extra peptide in Hp2 was a junction between the two peptides that were in reduced amount.  And we later called it the junction peptide.  And then it looked — also we could see that the characteristic peptides that [00:13:00] defined Hp1F versus Hp1S, that those two peptides were both present in the Hp2 sample.  So it looked as if Hp2 was a combined Hp1F and Hp1S.  And this was very difficult to understand.  And I visited Toronto sometime in that period and I remember extremely well what happened because in talking about these results I suddenly said, “Well, let’s see what happens if we believe our results,” because we couldn’t believe this.  And then suddenly it became clear that Hp2 was a junction between Hp1F and Hp1S and the junction wasn’t quite perfect.  So the junction peptide was not the [00:14:00] N-terminal and C-terminal exactly, the sum of those two.  And etc., etc.  So we had a good hypothesis already by the time we were getting ready for the meeting in Rome.

Well, we got together that evening in Rome and George Connell had a very very long face and he said, “I’ve done the ultracentrifuge test and the sedimentation coefficient of Hp1 — little hp1 versus little hp2, the sedimentation coefficients are the same.  And so our hypothesis is wrong.”  I said, “Well, George, I don’t agree with you.  I think what we should do is we should do exactly what we said.  We’ll get up there and you can present your material and [00:15:00] Gordon will present his and I will present our hypothesis.  And then I’ll say, ‘And we’ve tested this hypothesis by doing an ultracentrifuge run.  The ultracentrifuge says that the hypothesis is wrong.’  But I’ll say that no, I think the ultracentrifuge is wrong and the hypothesis is correct and I’ll go back to Wisconsin, invent a new method of determining molecular size.”

And I knew by then that starch concentration affected large molecules more than it affected small molecules.  So on page 75 of this notebook there is a Sunday experiment, October 8th.  Hp — little hp1F, little hp1S and little hp2 mobilities versus starch concentration.  And I’m running a mixture of these three [00:16:00] preparations in starch varying from 100% starch up in steps to 145% starch.  And the prediction is I hope to see that 1F and 1S don’t — the separation of 1F from 1S is not affected by starch concentration whereas the large Hp2 should be slowed down relative to.  And there is the prediction on Sunday October 8th on page 75 and on page 77 and 76 there’s a graph.  And the statement is made.  The results are as predicted, exclamation mark.  So this was a first real definite proof [00:17:00] that Hp2 was bigger than Hp1F and 1S.  It was a very beautiful and very exciting result and was subject to a lot of cheering.

Then continuing in a more mundane manner thereafter some tests to get other data on the molecular weight determination.  For example October 11th page 85 repeat the haptoglobin in quotation mark molecular weight determinations.  And then [00:18:00] beginning to (inaudible) realize that this was a general method for determining molecular size and might be useful for other things than just for haptoglobin.  And so on page 89 there is the beginning of testing the effect of starch concentration on molecules of different sizes, lysozyme, insulin, alpha-chymotrypsinogen or alpha-chymo, I don’t remember if it was trypsinogen or not.  Yes.  Alpha-chymotrypsinogen and beta-lactoglobulin.  And this led eventually to a paper which I published rather unwisely in a symposium for — in remembrance of — or in recognition.  Yeah, I think he was still alive.  In recognition of [00:19:00] Tiselius’s work with electrophoresis.  And I’ll stop there for a moment to find that reference.  Yes, this was beginning on page 91 of experiments attempting to use this idea of starch concentration to determine molecular size which led to a publication in which we shall no doubt come to the gels that were in the publication.  So this was what was going on in that period.  For example page 97 wide range test of 70% to 140% starch gels, etc. [00:20:00] with alpha-chymotrypsinogen, insulin, lysozyme, and so on being tested.  Eight molar urea, formic acid gel which of course is a good gel.  Different concentrations and so on.

Some data on page 107 trying to get there to something worthwhile.  Continuing getting this method to work for the next several pages.  Glucagon and ribonuclease survey on page 115 with a letter from William Bromer sending us some pork glucagon.  [00:21:00] People were always very willing to share.  This was from the Eli Lilly and Company in Indianapolis.  William Bromer sending us glucagon for the molecular weight studies.  Data getting a little better.  Page 118.  Page 125 quite a nice set of gels where the data is — where the concentration is varied [00:22:00] with three proteins.  Insulin, glucagon, and trypsin, as plotted on page 126.

Going back to the subtyping of the haptoglobins on Friday November 3rd page 129.  Just some further thoughts on this procedure.  Quite reasonably happy with the results, but some complications. [00:23:00] Page 137 small molecule molecular weight additives.  Thinking about adding glucose or glycerol instead of just varying the starch concentration.  And glycerol migrations were plotted on page 139.  Never led to anything.  [00:24:00] And continuing in the work.  And on page 149 Tuesday November 21st new longer eight-gel apparatus was used to make eight gels as usual.  And there is an insulin B-chain and Hp1F alpha and Hp1S alpha and Hp2 alpha and Hp beta gel.  Really very pretty gel on that page.  So the method was really working well.  [00:25:00] And we end this book with that as the — as it were the final gel for this particular book.  It’s a beautiful set of gels on page 148 and they even now make my mouth water.  [00:25:25]