Oliver Smithies:

Is it ready now? Yeah, OK.  This is a continuation of Book XII, starting with May 22nd, when I’m still struggling with purifying haptoglobin using a Dowex column, and I’m trying to wash the column with acid and water, after removing previous material with concentrated sodium chloride.  I’m evidently having trouble, but it looks as if there’s going to be a solution later on, because it says, “But see May 27th.”

Then the next couple of pages are some more [farmer’s?] studies, or starch [00:01:00] studies.  I think they’re actually starch studies.

Then on Thursday, May 23rd, I have two interesting things.  I have a circuit diagram that comes from my friend, Les [Pinterich?], which is, in that we’re having a speaker make a sound with the conductivity measurements, I think that’s basically what it’s for, although there isn’t much comment on it, the only comment being that it’s by my friend, Les Pinterich, who was an absolute wizard with radio and tube radios, and tube electronics in those days. [00:02:00] Of course, the transistor had not bee invented yet, so I’m still trying to get samples prepared as I enter on Thursday, May 23rd.

Continuing the next page or two, no improvement.  The serum level is OK.  Friday, May 24th, a different Dowex, or at least it made Dowex-2 hydrochloride as positively-charged resin, and set up with lot one of Dowex, and lot two, etc.  I think, it is the Dowex, the same Dowex, just, [00:03:00] some different lots of Dowex.  Though I think, obviously, one is better than the other.

Continuing, same, the first [tri?]-column application, and a fairly complicated set of entries here, all aimed at the same problem.  With the comment on page — the results are still not good enough, but one factor has not yet been watched carefully enough, that is the minimum amount of resin to be used.  On one experiment it was 10mLs; in another one, the best of all, it was 45mLs.  And yet, another was good, and this is 10mLs, so I’m not understanding it. [00:04:00] So, with the comment, however the Dowex-2, hydrochloride lot two, with 10mLs is fully satisfactory, and study it carefully.  Here’s one notice, that those lots of haptoglobin obtained on May 16th, May 21st entries, haptoglobin 3 pluses, plus or minus, so lots of haptoglobin.  Whether the acetate is as good as the hydrochloride, and looking at 5[mL?] resins instead of the larger amounts, continuing [00:05:00] with this general problem.

We turn to a different sort of problem with inquiring about the changes of outdated plasma or serum.  Got some large, two bottles, 500mLs of outdated whole blood that I could get and try for making haptoglobin, basically, is what I’m trying to do.  Also, asking whether hemoglobin-haptoglobin can be removed from serum by the resin, either as haptoglobin, or haptoglobin/hemoglobin complex.  [00:06:00]

May 29th, I got some material from Dr. Charles to sample, but abandoned, too much hemoglobin.

More entries, more samples, from different persons.  Nothing very remarkable.

An attempt to understand the effect of pHs, Thursday, May 30th, a pH of 4.3, put in quotation marks, and 4.5, meaning that probably, that’s nominal, not measured pH, and 5, and 5.5, [00:07:00] and 4.5, with a note that it’s almost pure haptoglobin, so it’s fairly clear that the pH matters.  And the conclusion is that rough examination suggests that the haptoglobin level is constant, but the [first?] impurities decrease markedly as you decrease pH.  And there aren’t any impurities at pH 4.5, so repeat [performing hazard?], etc., etc..

Noting on Friday May 31st, and it’s quite common.  I used to be very happy with this idea that formic acid as a pKa of 3.75, near enough for, a [CG hazard?] that’s 5, and so on, and using different acids around their pKas to get good control of pH.  There are some [00:08:00] calculations of the pH, and using formic acid on Friday, May 31st.

And we were still working on the weekends, actually, Saturday June 1st, some [well-washed?] samples.  Nothing very remarkable in here.  Several pages more of this general approach, though it’s fairly clear here that I’m narrowing down to using a fixed pH, and using that, I can get pretty pure haptoglobin with the Dowex resins, and nothing to do with [finishing?] chromatography, just ordinary ionic absorption and elution.

Telegram from Charlie Hickman, that [00:09:00] about the animal samples.  That was just prior to June 4th, the date is probably given there.  And I can’t read the date on the telegram.  I dialyzed some serum to get it to ph 4.6.  But having some troubles, note in this all the impurities are missing. [00:10:00] pH 4.7.  May have to work in the cold. That proved to be completely unnecessary.

Have some very nice plasma, or, on June 4th.  Still worrying about pHs in this period.  Checking whether there was [00:11:00] chromatography along the resin on June 6th, and they’re still continuing, continuing.  With the resin, [extra?] resin, abandon this method, comment, not at all uncommon to find that abandoned, or abandon that, pH 4.3 experiment is confirmed exactly, and as usual tube #1 is almost pure haptoglobin.

Comments on dialysis, still worrying about whether there’s chromatography on June 7th.  Further sodium chloride tests.  I must settle the chromatographic question, [00:12:00] where there’s chromatography occurring or not, or just absorption and elution.  Tried an inverted column for some reason or another.  See I’m fairly happy with the method.  Not much is changing, it just looks like micro fine-tuning.

Other strange result [00:13:00] on Tuesday, June 11th, at pH 4.3, trying to extract with thousands molar sodium chloride, then nothing came off, nothing, that’s completely negative.  And, with a comment shortly thereafter which is outlined in red crayon, “Sodium chloride must be greater than 2000M.”

Another bottle, and another — blood — another, keep going, and experiment now using 0.5M sodium chloride.  [Let’s hope for?] zero haptoglobin. [00:14:00] And this is 0.02M and 0.05M.  I think this is related to washing the column, not eluting, but we’ll see.  Three bottles obtained June 17th, that would be my sister’s birthday.

TGH is noted here, and that would be Toronto General Hospital, it’s referring to a family studied [00:15:00] June 24, June 25th.  On no, these are cattle, cattle, actually.  These are almost certainly cattle as well.  A23, 22, A9 — 18, 17, a large number of sample.

That ends this book, with also a little family observation.  Friday June 28th is the end of the book.

This is XIII, physical 13 book starting in July 1957, and running through September, the first entry being July 4th, 1957, with [00:16:00] some notes on a family there, and the same in the next two, three pages.  Several family studies are recorded.  Then the usual, confirmatory gel being run, on a few days later.  Evidently, at this time I began to receive samples from Aborigines, and Torres Islanders, the Aborigines, Australian Aborigine, and the Torres Islanders, I should know where that is, but I don’t know where it is anymore, but I can find out.  I’m receiving samples from Aborigines which are Australian, [00:17:00] and Torres Islanders, and they’re from the islands between Australia and New Guinea, and have been isolated for a long time.  Both the Aborigines, the Australian Aborigines, and the Torres Islanders, have long ago been in these areas before their British, or similar invaders came.  So, there’s a whole set of observations on these, still concerned mainly with haptoglobin on the right-hand side, but on the left-hand side of this page, I’m talking about clear, split-beta, very clear [00:18:00] split-beta, fairly clear split-beta, and this is the beginning of understanding all typing variations in transferrin.

Family study in here a couple of pages, and there’s still more Aborigine samples being looked at.  Thus the Aborigines’ data are self-consistent, in various checks.  Looking at the frequency of the haptoglobins, this is what this refers to, the observed distribution, where there’s two 1s, ten 2a’s, and 11 2b’s, and that assuming that they were randomly distributed, that gives a frequency of the two genes, [00:19:00] can be calculated.  It’s a rather naïve calculation, I would now say.  But split betas, which are the transferring, recorded very clearly, and compared with the Douglas [Nigro?] split beta from a long time earlier.

And some fat stains are studied, July 15th.  I was always continuing those sorts of things on the side, different stains, while doing other things, that Oil Red O, and Oil Blue and are still the best.  IR-red always comes out well.

And now, the testing in the Aborigine samples with two-dimensional electrophoresis, which is the [00:20:00] only reliable way of establishing the transferrin differences.  So, Wednesday — Tuesday, Wednesday, July 16th, 17th, there are two very nice two-dimensional gels, just some comments on another type of filter paper, but it’s very easy to see that the split-beta, as I was calling it.

More lipid stains being talked about today.  Adopt Oil Red O, and Oil Blue, and a mixture for staining for lipids.  In trying to stain [ferrins?], Gordon [Romans?] who was a colleague in the [00:21:00] production side of the [Canote?] labs, has a green die which appears stable, so he gave me this to try.  Trying a new lot of starch at the same time.

More studies of insulin again, still we’re not completely giving up all insulin work on July 18th, various tests with insulin, to see whether I could establish something of importance.

Back to [00:22:00] two-dimensionals again, Thursday, Friday, July 19th, two very nice two-dimensional gels.  Both of these samples, as the comment says, have beta-1 and beta-2, as I’m beginning to call them, and photograph them, and good photographic images.  And nice to see the records in this way.  I remember how enjoyable it was to get good photographs.  And some just, still trying to improve the two-dimensional separations, by doing various things to keep the first dimension on filter paper in good shape. [00:23:00]

Complicated insulin staining tests.  More dye tests, nothing very striking.  More studies, insulin staining again, Thiazine Red R, there was a formula for it, and Bromophenol Blue, the formula for it, and Amido Black 10B, which of course was, became our — and remained our standard for many years.

Some sketches of the insulin results, July 22nd, and again, more two-dimensional, checking of Aborigines again shows, A21 has split beta, missed it in checking [00:24:00] originally, so due to two-dimensions on it, and also another one, so, a good result obtained with A21 confirms split beta.

Sketches of insulin separation, that never led to anything.  Full review of two-dimensional thing — studies, looking at the haptoglobin typing, and with one little comment, “This one is a classic [Nigro?] 2A, but also [00:25:00] has beta-1 and beta-2,” so that’s a haptoglobin 2A modified, I think we later called it.

And on the following page, some good photographs of two-dimensional gels, pointing out where there’s hemoglobin, where there’s no hemoglobin, etc., etc., and the two beta-globins.

More two-dimensional gels, and one that is not — beta-2 gels are [00:26:00] run showing no beta-2, so it’s showing some sample that are on two-dimensional gels, just show a single, as we later find out, transferrin.

At this point, I began to use a simplified technique, instead of running a two-dimensional gel on the whole of — taking a large square gel, and taking the whole of a filter paper strip, I realized that in order just to type the beta-globins, it was easier to use a small gel, and cut the filter paper sample to include only the region where beta-globin runs, and in this way, one could run them much easier, and run their samples much more easily, because you could do [00:27:00] essentially two-dimensional gels, using small, one-dimensional, the one-dimensional apparatus, because I was only wanting to look at that region of the total pattern.  So from Thursday, Friday, July 26th, it was actually quite an important beginning of what later on enabled Otto Hiller to study several hundred samples by two-dimensional gels, which would have been extremely laborious using the big gels, but were relatively easy using the smaller gel.  It was still hard work, but not impossible.

Tested another dye on July 29th.  Now here is the beginning, again, now realizing this, [00:28:00] pursuing this, what I call “mass two-dimensional tests,” Monday, Tuesday, July 30th, where many samples were run at the same time and only the beta-globin part was picked out, and so there are three, four gels there, not really perfect, but I think one can see that three of them are — or have an ordinary transferrin, or ordinary beta, as I was calling it still, and one had the double one.  Not actually very good result yet, but it’s beginning of the (inaudible).

Getting better, [00:29:00] Tuesday, Wednesday July 31st, with a comment in red crayon that all the two-dimensional multiple gels this date were photographed, or have been photographed, and there are some gels from [name redacted] and [name redacted], which are just beta globin C as I called it only, beta globin C.

More tests again on Thursday, Friday August 2nd, with two-dimensional, and [00:30:00] what I call them, multiple two-dimensional tests.  All four samples on that day had only a single beta globin.  Always testing starch, August 2nd.  A little family.  Tried some Tris buffer, which these suggest that these suggest (inaudible) Tris was not very useful.

These were samples of insulin. [00:31:00] Nice set of results, Tuesday, Wednesday August 6th, where there, it’s clear that the entries were — that these are more being done by Otto, I think, although I was probably doing the photography, I’m not sure exactly.  But all of these were — none of them were — had the double-beta.

Here’s an interesting observation, Wednesday August 7th, haptoglobin purification with a 55mL resin, [00:32:00] and 22mLs of serum, and it was dialyzed serum that had been left nearly two weeks in the cold, and it froze, and thawed, and rather poorly handled.  But nonetheless, using the usual method of putting onto a washed Dowex column, and extracting with 0.1M sodium chloride, collected five tubes, and ran them, and it says enough haptoglobin in #s 1 and 2 to use, with a comment, “surprisingly stable protein.”  So it was a fairly, now a fairly well-established method of making haptoglobin.  Absorption to Dowex resin, and eluting with sodium chloride, hemoglobin [00:33:00] was not involved at all.

Still trying different filter papers, for the first dimension of the filter paper electrophoresis on Wednesday August 7th and 8th, and comments on which filter papers were better, whether they streaked, etc., etc.  Whether there was anything left at the origin, whether they streaked, etc., etc.

Trying to make a synthetic medium instead for the electrophoresis, in applying the sample. [00:34:00] I tried Perspex spheres, vinyl chloride, vinyl acetate, and starch grains, etc.  [Try a centered block?], etc., etc.  Not the answer.  Treated filter paper again, (inaudible) narrowing the choice of filter paper down, 598YD, and 3MM papers treated, was being [00:35:00] investigating.  And I tried plaster, CSH plaster.  CSH is Charlie Haynes.  This is for use in applying the sample to the gel.

Modified filter paper, etc., on August 12th.  Human results are disappointing, August 13th, although the bovines are very good.  This was actually a test on the August 15th, or 13th, was just testing a new Danish starch, so the human results weren’t [00:36:00] as nice as I wanted, but the results with the cattle blood were very good.  More test with zinc insulin.

Friday August 16th, talking about in view of the rather disappointing results on aqueous Danish starch, which gives a highly glazed final gel after staining, and because unhydrolyzed Idaho gave the best results on bovine beta-globin, I decided to try adding unhydrolyzed starch to the Danish, and [00:37:00] see whether it helped.  I still don’t — I think I had not realized, or maybe I never did completely realize that a lot of the problems with the beta globin was on the amount of iron that was present in the sample, or in the starch that was used.

Freshly prepared, [Dr. Hill?] bat serum was testes Thursday, Friday August 23rd, and fairly good, as far as beta-globin is concerned, two-dimensional gel, with and without hemoglobin, was also, didn’t affect the beta-globin, the photographs there.  [00:38:00]

And then on Saturday, August 25, I’m trying to control the hydrolysis of this starch, the aim being to produce a chemical test of correctly hydrolyzed starch.  So testing the total glucose of the supernatant during the hydrolysis, but it looked unpromising.  It didn’t work very well, though I wanted to find a better controlling thing.

Working quite small differences, as indicated by Monday, August 26, but that’s starch with 21 and a half, 22 and a half, and 25 and a half hour hydrolysis [00:39:00] so it’s fairly critical to get the best hydrolysis.  Trying to do some dye stability tests with Amido Black, Thiazine Red, and Nigrosin, but it was not very well done, although I do have the comment, “Abandon Amido Black,” and I know that wasn’t the final result, but must have thought that the Thiazine Red was much better in the stability of ability to stand washing with methanol and acetic acid, 45% methanol, or 30% methanol, and with or without tannic acid in this case. [00:40:00] Trying to add the dye with a comment, “Not a good idea.”  And I thought that in general, it’s quite interesting, because it’s quite contrary to what eventually happened, a general comment, “No use testing Amido Black anymore,” or maybe it said that I always get a certain result, although I obviously have some worry about it because I have a circled comment that Amido Black is no good on these tests either.  50/50/10, was what we eventually adopted, which was 50 volumes of water, 50 volumes of methanol, and 10 volumes of acetic acid, glacial acetic acid, [00:41:00] 50/50/10 was our final washing solution, and we did use Amido Black.

August 28, still worrying about it, set up a new series of tests with 50/50/10 versus 35/65/10, increasing tannic acid with three different dyes.

Thursday, Friday, August 29, 30th, some samples, in two dimensions of Aborigine samples. [00:42:00] The [name redacted] family and the [name redacted] family.

Wondering about fractionating the starch, Tuesday September 3rd.  Charlie Haynes was an expert on starch fractionation, so I talked to him quite a lot about that type of topic. [00:43:00] No indication that this was helping.

Pathological sample on September 4th, the Torres Strait Islanders are being looked at.  They were all 2b.  The two-dimensional gel was not very good, but they don’t [00:44:00] have any variation in the beta-globin in this particular group.  Confirms that they were 2b, a rather poor result I would say, by [current?] standards.

Trying mixtures of amylose and amylopectin, which is in fact the controlling factor in the variation in starches from different potatoes, the ratio of the linear polysaccharide amylose versus the branched polysaccharide amylopectin, controls the cross-linking of the gel, and therefore its gelling properties, so if it’s pure amylose, my [00:45:00] recollection is that it will not gel, although that’s not what I say here in this entry, September 4.  Increasing amylose leads to gel formation, is amylopectin and amylose.  And probably the hydrolysis forms more linear changes and less branching network.  My current interpretation is confounded a little bit by knowing that you can change polyacrylamide properties very much by varying the amount of bifunctional polyacrylamide versus linear — monoacrylamide [00:46:00] to change the cross-linking.  But here I’m struggling with this, a gel being made with 100% amylopectin, 80/20, 60/40, 40/60, 20/80, down to 100% amylose, and the viscous solution was obtained with amylopectin, and the gel was obtained with amylose, so this is the reverse of what I would have expected from polyacrylamide, but I did understand that the ratio mattered.  Six percent amylose gave an excellent gel, 4% quite good, not very strong, and 2% [00:47:00] didn’t gel at all, so the effect of the ratio was there.  The samples were Rees samples, presumably given to me by Rees.

It is a person, W.R. Rees, who had given me these samples, amylose and amylopectin. [00:48:00] And towards the end of the book, some more mini two-dimensional gels, or of various aborigine samples.  Bovine serum, and family, possible mother.  [Abandoned?] starch papers, September 5th, which had been tried in addition to 3MM paper, and still interested in the amylose business.  September 5, repeated the amylose preparation. [00:49:00] I’m trying to put out a dilute agar around the 3MM paper in the sample [slot?], but didn’t work very well.

Variously, tests of that same type, 3MM paper [00:50:00] plus bovine [four?], starch and distilled water, 3MM, plus bovine four, etc., and with a comment on, “The left-hand gel was not kept; the right-hand gel was kept.”  That’s the only technical problems for slow alpha-2 now, but beta was uncertain; I’m not clear what that meant, but still not quite happy with separation.

And this book ends with an experiment using polyethylene filter pads.  And the 0.5mm sheet is indistinguishable from the best filter paper results, but shows that the slow alpha-2 has mechanical streaking, meaning that the filters pads have places where [00:51:00] they touch and don’t touch.  The betas are excellent, but that is the end of this book, Wednesday September 11th.