Oliver Smithies:
[00:00:00] Start on the Book XX, April 1959 through August of the same year — Roman XX. It’s beginning by continuing the experiments that are leading to the depolymerization of the haptoglobin polymers using 8‑[molar?] urea and thioglycolic acid. So the first entry, Fri‑‑ April 17, shows a gel of that type, only run for a short time. And it’s clear that there are the polymers there and that the — that the 1‑1 haptoglobin, and appears to give one band and the 2‑2 appears to give two bands, with the faster one a little bit fainter than the slower one. And the [00:01:00] heterozygote looks like some intermediate. But the gel is run not a sufficient length of time. In fact, these gels are all part of — or became part of a publication that I wrote with George Connell, in the Ciba Foundation Symposium, on Biochemistry of Human Genetics, 1959, with editors Wolstenholme and O’Connor. And we have a long — a paper there, which, the images are now the ones that are coming in this book.The first [00:02:00] image in the book, in the publication is facing page 179, is a gel run on Tuesday, April 28th, looking at the haptoglobin types in a simple gel. And though it’s the reverse image or laterally switched, it is the same one — and includes a haptoglobin type that I have not yet talked about, what was called the Johnson phenotype — proved later on to be a very interesting sample. But we didn’t yet understand what was special about it. It was discovered by Elo Giblett. [00:03:00] And so the — facing page 179 article is Tuesday, April 28th, gel.
So continuing back to the Book XX, we see some more tests of this general type. Some difficulties with the thioglycolic acid being oxidized. But the gel run on Thursday, April 23rd through 24th, is one of the gels that is in the publication. I’ll just find the page in the publication. There is a urea, thioglycolic acid, plus borate gel — [00:04:00] is Figure 5. And Figure 5 is almost certainly the gel that was run on Thursday, April 23rd to 24th, the lower gel, which is thioglycolic acid and 8‑molar urea, showing that the haptoglobin polymers have gone, but the rather messy pattern. Still not able to make very much sense out of the pattern, because it was so smeary. But that’s the gel, on April 23rd, 24.
Continue with the problem, trying to get them to go back together again, a recombination, full-scale test, [00:05:00] making the solution and then adding oxidized glutathione, to try to reverse the polymerization. But not successful. I did later find how to get the reversibility. But not yet able to do that.
We’re now to the Tuesday, April 28th, gel, which I’ve already said was part of the CIBA symposium publication. Some purity tests. And so now the haptoglobin preparatio‑‑ the very nice haptoglobin that George Connell was making at this time. So nearly all of these purified haptoglobins are from George Connell. He was a real expert at making [00:06:00] pure haptoglobin. (pause) An 8‑molar urea gel, the various haptoglobin types, including the new one from Elo Giblett, on April 29 — which may be in the Ciba… No, they’re not in the Ciba publication. (pause) [00:07:00] But continuing to be sure what’s happening. Dissolving solid haptoglobins directly in urea and urea/thioglycolic acid/borate shortly before use. And ran the samples. Eight-molar urea alone, as usual, didn’t dissociate the polymers, depolymerize them. But urea and thioglycolic acid did. Quite a pretty gel.
Did some auto centrifuge tests on the material. And they were quoted in the Ciba symposium. No particular [00:08:00] record of them in this notebook, just a comment that they were run. (pause) No evidence that time made any difference. Some older solutions, now about 48 hours, April 30th, May 1st.
An unusual sample on April 30th, May 1st, from [name redacted]. Special haptoglobin test from [name redacted]. And the pattern is quite strange. It says, “It looks like a [00:09:00] 2‑1/2‑2 chimera.” So, “Test mixes of 2‑1 and 2‑2 and compare them with the serum from [name redacted].” And here are the — on May 1st — no — yes — May 1st through 2nd, some tests with [name redacted], saying that “[name redacted] is most like an 80/20 mixture of” the two haptoglobins. Being 20% [00:10:00] haptoglobin 2‑2 and 80% haptoglobin 2‑1 is the closest to matching her sample. And they really are very much alike. The results show the Hp 2‑2 and Hp 2‑1 bands line up with [name redacted] but the proportions are wrong. “Therefore a nondisjoined chromosome,” and 2‑1 type. I wasn’t quite happy with the mixture, compared to [name redacted]. But looking at it at this point, really are remarkably close. [00:11:00] Interesting. I don’t remember it ever being published. Maybe it was.
Auto centrifuge tests commented on, on April 1st. No diagram. Except there is a diagram in the Ciba symposium, of the auto centrifuge results with haptoglobin. (pause) Purified type 2‑2 haptoglobin was — the auto centrifuge — was published. [00:12:00] Continuing with this thioglycolic acid — the thioglycolic acid experiments. Still the key to solving the problem is not yet apparent. But I think, before the end of this book, unless I’m mistaken, the key is discovered. June 9th, Tuesday, “Interested in trying to find out if the split haptoglobins have any hemoglobin binding power.” (pause) There is a benzidine stain and the protein stain. And [00:13:00] this is the order of adding. So I set up a 0.5% solution of haptoglobin 2‑1 with the thioglycolic acid and added hemoglobin to it and tested it. And the result says that, “Rather overloaded with hemoglobin. But all the usual haptoglobin, hemoglobins are formed and overrun.” So that this is just thioglycolic acid, not thioglycolic acid and urea. So thioglycolic acid does not prevent the binding. But, of course, thioglycolic acid alone doesn’t dissociate the polymer.
So the following page [00:14:00] is the experiment with the urea-thioglycolic acid-dissociated — or ‑depolymerized haptoglobins — but for a rough test. There is no comment on what the result is, in this particular page. The hemoglobin, urea and thio and urea was dialyzed against thioglycolic acid — dilute thioglycosic and then set up with or without hemoglobin. [00:15:00] (laughs) Says it’s a “rough test and an odd result.” Repeated again on Wednesday, June 10th. “The hemoglobin patterns of 2‑1 and 2‑2 are indistinguishable.” But beginning to think about getting mixed disulfides, protein SSCH2COOH, which later on became critical in understanding what was happening. But these experiments were rather confusing, urea, thioglycolic acid and dialyze and the urea, thioglycolic acid, plus hemoglobin and dialy– But [00:16:00] they were still able to bind hemoglobin, from the look of it. Or it might have been just the hemoglobin itself. I suspect it’s just the hemoglobin itself. That there is no binding would be the interpretation I would make now. But let’s see how it worked out. It says, “Repeat the 2‑2 experiment above with a hemoglobin control.” Because it’s likely that the only thing that was really giving a benzidine reaction was the hemoglobin. And attempting this, on the following page. A hemoglobin titration with [00:17:00] 0.25% hemoglobin, on Thursday, June 11th, with the interpretation, “Hemoglobin is split. Must repeat with new hemoglobin. Also try increasing the ionic strength,” etc., etc. Obviously, not very happy about it. (laughs) Odd little inclusion in the — in the book of “The danger is now an MEND report” — and the military medicine. Something about general warfare, etc., Greek culture. Why this is in there, I don’t know. But it had been sent to me in the Genetics Department, University of Wisconsin, Madison. [00:18:00] (phone rings) Still trying to work out the problems.
Interviewer:
Smithies, [Montoyas?].
OS:
And various dialysis experiments —
I:
I’m sorry.
OS:
— with hemoglobin —
I:
I didn’t get your name, sir.
OS:
— Monday, June 15.
Nobuyo Maeda:
Who…?
I:
[Police?]?OS:
And so…
NM:
Oh, tha– [T. Bontham?]?
OS:
Oh, yeah. Yes, hi, T.
(pause; break in audio)
OS:
Still struggling with the possibility that [00:19:00] hemoglobin binds to the split haptoglobin. Very unlikely but still trying. June 16th, to determine the effect of increasing the thioglycolic acid on the SS cleavage. And so going from 0.003-molar up to 0.05-molar. And there is a gel on that topic, not very revealing. Continuing these types of experiment. Decreasing SH. Going down to 1 millimolar. And [00:20:00] some progressive changes. But, “No evidence of hemoglobin binding if the thiol is greater than 0.003-molar. So use 0.01-molar –” 10 millimolar — ”again and dialyze, since this worked before. And still trying to get them to bind to the dissociated polymer. Here we are, June 18th, the same type of experiment. It says, “Initially, very clean hemoglobin migration.” This is into a gel containing 0.002 thioglycolic acid. But, “No sign of binding.” And, “All colors are brown after running but not initially,” [00:21:00] indicating that the hemoglobin is being converted to [mat?], though I don’t say that in the — in the book. More experiments of the dialyzing, etc., still not getting very far.
Family sent to Dr. Smithers, at Connaught Laboratory, from Johns Hopkins, from Carmen Merriman, and sending some samples that they wanted to have me [00:22:00] test. (pause) And this looked like a family where the haptoglobin was low-concentration. And they had type, it says, 0‑0, a couple of them. But I typed them all as 2‑2. So probably their benzidine wasn’t working properly.
Transferrin iron tests. I did 5 gamma of Fe3+-per-milliliter to serum, [00:23:00] approximately equivalent to 1 gamma of Fe per ml. And dilute [a seru?]‑‑ with serum. “Tested with pH 4.4,” etc. The original image of the result. I was testing with alpha-alpha-dipyridine and nitroso-R salt. And I have a diagram comparing the structure of this diazo dye with Amido Black. “Clear binding on all the proteins, of NRS.” [00:24:00] Trying to see if I could detect iron with the diazo compound, evidently. [And?] a stain for transferrin, Friday, July 3rd, continuing this. And [for?] the gel. But there is no image there — although it says, “The photo suggests that all of the solutions were equally good,” with that. (pause) More tests of that same type, trying to get [00:25:00] something to stain the iron.
Monday, Tuesday, July 7th, samples from J.P. Garlick, the Department of Anthropology, in Boston University College. No, I’m sorry. That’s [Euston?]. (laughs) I thought it wouldn’t make sen– University College, London, Gower Street, Department of Anthropology. He had some variants from Nigeria and wanted to see how they compared with the samples that we had been running.
It’s clear at this point that I’ve begun to understand that it’s better [00:26:00] to add iron to the sample, in order to get clearer bands. Because when the iron binds to transferrin, it’s stable, whereas, if there is incomplete saturation, the migration is likely to be poorer. So that, Monday, Tuesday, July 7th, I’m talking about “Good sample with fresh iron” and comparing Helen McGuire, which was the interesting person, with various other samples. Quite a nice gel. And typed her as — comparing, with Helen McGuire, which was used [00:27:00] for the Fe‑59 experiment, with this family, the Alberta family. And the typing is written down.
Some antisera tests. Never came to anything. (pause) [00:28:00] (pause)
Trying to get ceruloplasmin better, again, Monday, Tuesday, July 14th, say ceruloplasma — “Cerulo. is good” — and a comment, “None better than the standard,” which is 500 NRS for 15 seconds and then [50:50:10?]. [00:29:00] And Tuesday, Wednesday — Tuesday, Wednesday, July 15th, shows some of the staining patterns, obtained with this method, there. Let’s just see what it is. So this is using 500 NRS, the structure of which was previously compared with Amido Black, back several pages. Let me get that. [00:30:00] Back on June 30th or thereabout, comparing the structure of 500 NRS, which was thou‑‑ which was, in fact, able to stain for iron, and a rather sensitive stain. And then here it is, on July 15th, tests of some standard haptoglobin — sorry — some standard transferrin types, Campbell, [00:31:00] Ashley, etc, various transferrins, with the stain. And there is a gel stained with protein. And then there is the staining with the 500 NRS. And the transferrins are really quite well stained. So one could detect the transferrins without radioactivity, quite specifically. And testing different persons. And one of the individuals was… “Check on Elo Giblett’s B1B1 homozygote.” And it checks with what she had said. [00:32:00] Quite nice images, without radioactivity. There are some loose photographs, at this point, in the book, that I will attach, to prevent them being lost. This isn’t critical, since it’s just a rough…
And tested on some small gels, on the following day, to try different amounts of iron — and what the result was of adding different amounts of iron. And it’s saying that the “0.25μg of [00:33:00] iron per milliliter looks the best in sensitivity.”
Back to dialyzing haptoglobin again and testing for binding. More transferrin tests with the NRS 500, on Tuesday, Wednesday, July 22nd. Not very good but usable. Still dialyz‑‑ haptoglobin, after depolymerization, July 22nd, with a [00:34:00] comment that “The hemoglobin entry is very good but the haptoglobin entry is poor. Stick to undialyzed for haptoglobin comparisons.” Quite a few comments on what was happening.
And on Thursday, July 23rd, is the beginning of the solution to the haptoglobin depolymerization tests. It’s an overnight gel run with 4‑molar urea, showing the… [00:35:00] And the haptoglobins were treated with urea and thioethane — “EtSH overnight.” And they’re fully dissociated, into the more or less usual pattern, in a 4‑molar urea gel of pH 8.5, giving the sort of semi-smeary patterns that I had by then associated with alkaline-reduced haptoglobin in alkaline gel. But there is a pH of 3.62 gel, made with 4‑molar urea, 0.05-molar formic aci‑‑ 20% neutralized, with sodium hydroxide. [00:36:00] Comparing the haptoglobins that have been treated with urea and thioethane overnight. And the result is very interesting, that now the patterns are not smeary anymore. So the comment on the left is, “Very promising for U/EtSH. Two bands, possibly four, in — Faster band in the single band, in 1‑1, and a single slower one in 2‑2.” It looked as if those — much better migration. But I can see material that was at the origin. So it was the beginning of a [00:37:00] solution. With and without hemoglobin, these are samples. There’s a sample with hemoglobin, sample without. And a comment, “Free hemoglobin migration is the same. No binding.” But the beginning of the use of acidic gels. So that’s a very important day, although I don’t think I had fully realized how important it was, yet. (pause)
But [00:38:00] it’s clear that I’m beginning to understand. Because I’ve got the test of low pH and urea on haptoglobin. So I made up 0.5mls of 1% aqueous haptoglobin and added the pH‑2.9 formic acid — formate — and ran the gels with the haptoglobin that had been just treated with water or hemoglobin — or had been treated with 8‑molar urea. I can’t… And then with a comment that I canceled this experiment, “in view of the result over the page.” And so over the page is pH 3.7 and pH 3.15 — in big red crayon, “A 4‑molar urea, formic acid, sodium hydroxide –” [00:39:00] and just acid gel alone. And although I don’t yet understand the patterns, I’m beginning to see that there’s a markedly different haptoglobin 1‑1 as a fast component and 2‑2 as a diffuse component, etc., etc. But the patterns are beginning to look clearer, pa– In particular, pH 3.15, the haptoglobin 1‑1 has a fast band and haptoglobin 2‑2 has a slower, rather diffuse band. And the heterozygote has both of them. So at last I’m seeing something that looks like a [00:40:00] component that is a simple component, that’s different with the 1 gene and the 2 gene, and something that stays back at the origin and doesn’t migrate. This is beginning of the solution to the haptoglobin and realizing that there are two chains, and the alpha chain and the beta chain, and that the variation is in the alpha chain. So this is actually a very important part of understanding what’s going on.
Friday, July 24th, making up acidic gels, with acetic acid, pH 5.9 and pH 5.75, 4‑molar urea. And just saying that the [00:41:00] gel without urea is not very promising, “Much insoluble material,” but the 4‑molar urea gel is very promising, “Very little, if any, insoluble material, although the run –” or was run the wrong way. “Test properly, the other way round and run longer.” Because, of course, the charges are now the opposite way, at the — of this pH. But they are migrating to the negative pole. “Color migrating to the positive pole in urea. Sticks positive origin in aqueous.” [00:42:00] Penetrable remark.
But now beginning to work on it some more. Still not making sense quite yet. Monday, July 27. And then, the following day, the images from a formic acid gel and an oxalic acid gel. And there’s a c– Actually, oxalic acid gel looks nicer. There isn’t much to choose between them. But it’s probably more in the photography than in anything else. [00:43:00]
Tuesday, Wednesday, July 28th, 29. Four gel, on a multiple vertical set‑‑ was set up, with 8‑molar urea, oxalic acid, 4‑molar urea, oxalic acid, pH 3.25 for the first, 2.65, 2.3, 1.7, different amounts of urea, so going from no urea to 2‑molar, 4‑molar, and 8‑molar urea, in the presence of oxalic acid. And stained these and beginning to see… All of these were… Doesn’t say which haptoglobin was being used. All [00:44:00] three haptoglobins, perhaps one… No, there are four gels. And there was a comment, “The 2‑molar oxalic acid is approxima‑‑” the gel — “is approximately equivalent to the 4‑molar urea, formic acid gel.” So both a decrease in pH and an increase in urea encourage the entry of the proteins into the gel. And they’re clearly all done on one haptoglobin type. But the record doesn’t show there which it is.
Except, next page, “Made formic acid gels.” And it was new [00:45:00] 1% haptoglobin 2‑1 in 8‑molar urea. “Solution complete. Dissolved directly in 8‑molar urea,” the solid, and then added EtSH, 20 millimolar, just in urea, without any comment on what the pH of the solution is of the 8‑molar urea. But the migrations are now beginning to be clear, that… But… For the 0‑molar urea, the migration is — “Result confirmed previous — Variable part of molecule splits into two.” And there the constant part is [00:46:00] back at the origin. Etc. So beginning to get understandable results.
July 30th, which is formic acid gels with the different amounts of urea. Continuing in the same way, with rather more concentrated samples. “8‑molar urea, 0.05-molar formic acid, 0.01-molar sodium hydroxide, pH 4.09,” Thursday, Friday, July 30, 31st. (laughs) With some insulin in the same [00:47:00] gel. One-point-five-molar urea the next day, with formic acid. Rather a confusing gel. Time test, on Tuesday, August 4th. Over-the-weekend gel, etc. August 4th, 5th, purifying haptoglobin. One-point-five-molar urea, formic acid gel, showing very much variation between the different types. Now it’s very clear that, when… In this gel, one can see that there is a constant region, that migrates [00:48:00] fairly well into the gel, and that there is a faster-migrating component, which is varying according to the haptoglobin genotype. But not completely understandable, at this point. But getting closer. And so, on Wednesday, August 5th, trying purified 1‑1, etc., various gels. A formic acid gel — or a formic acid, urea gel, 1.5-molar, in this case. So struggling but making progress in understanding what urea concentration and what pH to use.
And the final [00:49:00] entry of this book is, “Test stopping the reaction with iodoacetamide.” This again, turned out to be very significant, to stabilize the product, after the urea and thiol had depolymerized the hemoglobin. Stabilizing the product by blocking free SH groups with iodoacetamide proved to be very useful. Although I don’t know that this is yet clear, at this stage of the game. But anyway, here is the beginning of stopping the reaction with iodoacetamide.
The results indicate that iodoacetamide [00:50:00] stopped the reaction and, even more so, does a decrease in pH. So you could stop the reaction with a decrease in p– “The reaction is incomplete at 30 minutes, on this test. Both stored haptoglobins are ve‑‑ are very different.” But getting there. pH 3.4, and with iodoacetamide. And the end of Book XX. [00:50:31]