Oliver Smithies:
[00:00:00] This is book Roman numeral XI, starts January 1957 and runs through to April of 1958, begins with some bovine samples, and by Otto Henderson family and other families, run by Otto. Some copying of exposure times, as usual, with, again, Otto’s writing, January 17th, a larger — [00:01:00] up to January 22nd, all of the entries by Otto. January 27, Tuesday, and January 23rd, I’m still fiddling with photography. There was a constant concern to trying to get better images of the photo paper. And more families following, with Otto’s entries. And January 29th, “Filter paper electrophoresis photography using Microfile, various exposures, with and without filter. The red filter has little effect on the final print. F4 paper may be too contrasty. F2 is better but maybe F1 –” etc., etc. Anyway. “Take a set of negatives…” Starch tests going on, January. [00:02:00] More [Herb?] families on February 27th. And families with Otto, Otto’s entries. A lot of Otto entries here. February 5th, more photography. And rather beat up record on Thursday, February 7, and nothing very remarkable. Wednesday, February 13, a fairly extensive set of samples, from Charlie Hickman. “Received 38 samples from C.J. Hickman.” And ran them on two [jowels?], [00:03:00] 1 to 11 and 13 through 39. And all are typed, with the numbering system that we’d used for cattle serum groups, which, as I mentioned earli‑‑ we showed, shortly afterwards, transferring. More checks, February 14th, bovine checks on them, as usual, checking preliminary typing. Two-dimensionals again, February 14th, but no images.A summary of cattle types and numbers and pedigrees, where, on the right-hand-side page, there are all of the types, [00:04:00] written over with… Clearly, some have been overwritten. As we changed the nomenclature, I think, is the probable reason. And the Holstein pedigrees are shown on the left page there. This is an entry, around about February 15th, where the types are all given, and checkmarks meaning that they had been checked against each other and the sire and the dame, to try to work out the inheritance. But looking down there, as I look down, the sire and the dame, a [5-by-5?] who give a 4… And a 2-by-5 could give a 5, what was called 5. And a 5-by-5 gave a 5. And a [00:05:00] 5-by-5 gave a 5. So we were beginning to understand the inheritan‑‑ A-4-by-a-3 could give you a 3 but it could also give you — a-4-by-a-3 could give you a 3. So we began to understand who was heterozygous and who was homozygous. But there were multiple bands in the homozygotes. So what appear to be about five proteins would appear in a homozygotes. And eventually, inheritance was surprising, because they segregated as single alleles but there are five proteins were involved in the group. Presumably, there were five copies of the [00:06:00] gene. Never did try to do it at the DNA level. But it would be interesting to go back and see that.
Crosschecks of family sample, more families. Otto is running a lot of these. (laughs) I don’t know what I was doing while these were going on. Tuesday, February 26, made an apparatus for doing two-dimensional electrophoresi‑‑ so that the first dimension was in starch grain electrophores‑‑ starch powder, the idea being to remove the spaces, to change the del to electrophorese. And two or three pages of trying to [00:07:00] do the… Thursday, 28th, some vaccine tests, Newcastle virus vaccine test, on certain important… Again, Thursday, 28th, 51 bovine samples this time. So he really did send me plenty of sample. And there were, altogether, four gels run to give these. Well, more than four gels. There were five, six. Six gels were run to type these sample — with a note, however, gel #4, [00:08:00] that samples 26 and 27 — completely typical and 3 and 4 types but the pedigrees suggest that these two switched. And later word from CGH that 26 is really 2303, not 23‑‑ I’m sorry — is 5303, not 5304. So the gel would detect labeling errors, as well, of the t‑‑ a main indication that our hypothesis was almost certainly correct. Six gels on that. Some starch tests or — [00:09:00] on Friday, March 1st. And then a big summary, the following day, of cattle numbers and types and pedigrees, again, and various checks — a rather large number of checkmarks here, one set of checks and double-checks and thing. Because the new… No, I withdraw that. A large compilation of data.
Pedigrees, continued. More data. So —
Interviewer:
More — and…
OS:
— maybe 40 or 50 samples of cattle serum. And the bovine checks, following on March 1st, again, several gel. [00:10:00] And sample #40 was a little bit atypical, and a comment that, “In view of the importance of #40, must do a triple test with #40 versus the least typical samples,” etc., etc. A diabetic family, February 28. A comment on Wednesday, the 13th, “Bovine checks repeated,” with… This is #15 sample. “This animal is of about the worst 4 versus 5 typing ever seen. Doubtful value as a 5. But occurs in an unimportant pedigree.” More samples [00:11:00] from Charlie Hickman, on March 7th, and three samples that were needing confirmation. Continuing pages, Sunday, March 3rd, Monday, March 4, more work on the bovine sample. A human family, in there. And from Wednesday, March 6, a rough test on a new type of two-dimensional, without spaces, rather complicated, from the look of it. I know that it was never followed up, on the… It says, “The gel was extremely fragile.” And th‑‑ and there’s a sketch of [00:12:00] the results and they’re pretty awful. So I don’t wonder that I didn’t pursue it. More samples — Otto’s notation. Then checks plus a hemoglobin, on a family, again, Tuesday.
So here’s a beginning of thinking about quantitating two-dimensional electrophoresis, the aim being to use a fluorescent dye and then radiate a cleared stain gel with UV, and the emission, then, being visible proportional to the number of molecules of dye and independent of their distribution — in other words [00:13:00] integrated fluorescence. An‑‑ (laughs) quite a good thought, and, of course, is much used, anyway, these days. But it was a nice thought, never brought to fruition. But that started, with various dyes tested, acriflavine and National Erie Flavine S and National Phosphene GN. Tested these for fluorescence versus dye quantity, on filter paper, first of all. Made accurate solutions of them and tested these three dyes for their staining powers on some gels, and in the next page or two. And with a comment that the — [00:14:00] “Stained the gel parts,” etc., and then, “After 60 to 90 minutes of watering, only EF,” which is [E.?] Flavine, “looks promising, by visual dye intensity. The other two show generalized stain,” in other words, too heavy a background to be stained. So EF, which is Erie Flavine S, National Erie Flavine S was the only one that looked promising.
Interestingly enough, there were poor results had been obtained with Idaho 80-minute starch, with bovine serum. I wonder if that might have been due to the amount of iodine, the starch. But I don’t know. Becau‑‑ An interesting thought. Decided to [00:15:00] test new lots of the Idaho starch, of 50-minute hydrolysis and 45-minute hydrolysis, with bovine serum. And is a comment that “They’re mechanically indistinguishab‑‑ but still incorrect in separation,” incorrect meaning not like I was getting before. Again, it might have been due to the iodine content of the starch. But I didn’t know that, at the time. Some more tests of starch, Otto, March 14th.
And some new dyes to test, following, Uvitex — Uvitex WGS and WS and National Uranine, more fluorescent dye. [00:16:00] “All are visible at 0.002%, possibly at even” three zeroes-two percent, “But Erie Flavine S is still the best.” “Try with protein gels.” More samples with Otto, more serum samples. And going on with starch tests, “In view of the success of soaking 45 minute, Lot 3322, try soaking 3311 and see if it makes any difference.” “Very doubtful observations. Probably the order of making the gels is — im‑‑ is more important. No evidence of changing the grain size in the final gel.”
Continue into fluorescence, on March 18th, [00:17:00] Monday. “Fluorocell and Uvitex were used.” “Fluorocell extremely good. Can abandon — the — all the others temporarily.” And started inquiries for fat dyes. “But before discarding any dye,” I have a comment, “check with another UV source other than the mercury arc.” This — starch tests. “[You do?] [00:18:00] Idaho,” etc. “Starch tests continued. The results show that the betas are quite strongly dependent on the total borate.” Trying borate, slightly different concentration, slightly different pHs, 60% neutralized, pH 8.66, up to 70% neutrali‑‑ Oh, starting at 50%, pH 8.45, going up to 70% neutralized, pH 8.81, [00:19:00] with the comment there for “Use the rest of 3323, 3322, 45-minute, at 0.025-molar, 50% neutralized.” And it beat the 50% neutralized, which we’d been using a lot of the time. Dye tests beginning on Wednesday, March 20th. Received from Ciba Uvitex ER and [ER-Conc.?] and Uvitex U and National aniline. Gets into auramine and rhodamine. Or I got hold of them from somewhere. “Some very promising,” etc. The fat and dye tests continued, [00:20:00] Thursday, March 21st and 22nd. “Uvitex ER‑Conc. Seems the most favorable.” And, “Clearly positive result,” with a rough test, with two‑ to five-minute staining and rapid rinsing. Another family, on March 21st. Starch tests continuing, March twenty‑sec‑‑ Always struggling with starch — which is, of course, why people [00:21:00] eventually decided to use polyacrylamide instead, because it’s easier to handle and not so much variability. Fat, dye tests, March 25th. “Probably ER‑Conc. can be made good.” Starch tests. Starch tests continue. Checks on 2As and 2Bs, a big test of crosschecks. More starch tests. The usual sort of work. Starch tests continued, March 28th. More family. [00:22:00] Concentrated polio vaccine tested, on April the 3rd, for doubtful reasons. Oil Red O again and Thiazine Red, for lipid staining, April 4th. And somewhere around about April 5th or so, fat staining continued. I couldn’t get the gels to stain very well, except at the slot entry. Probably the lipoproteins didn’t get into the gel. [00:23:00] Comment on Friday, April 5th, “Oil Red O is as good as any.” And made a mixture of Thiazine Red and Amino Black. “Amino Black mixture no improvement. Not kept” — the gel was not kept — Monday, April 8th.
At that time, I saved every gel that we made, in tubes. So these gels were kept for a long period of time. And when I left Toronto, there was a huge stack of these saved gels was discarded. But I still have — [00:24:00] now, in 2015, I have three or four gels that are stained and still in the original tubes and visible. For example, there is one with Otto Heller, clearly, O.H. — sample, that’s still visible, in a tube, in a drawer in my lab. So they could be kept a long time.
On April 9th, it’s clear that I’m now beginning to realize that the protein that’s binding hemoglobin is haptoglobin. Because it’s… The heading Tuesday, April 9th, says, “Hp –” haptoglobin — [00:25:00] “purification test.” The idea is to adsorb hemoglobin irreversibly to charcoal, for example, and wash it and then treat with serum and wash it and treat with acid, to displace the haptoglobin — or maybe with excess hemoglobin, to elute the haptoglobin by competition. So I was doing affinity chromatography, trying to invent affinity chromatography. And I tried [Nor‑IT?] acid, which is acid-washed charcoal and diluted aqueous hemoglobin. And “It appears to remove some hemoglobin, washed with 0.9% sodium chloride and treated with serum,” etc., etc., concentrated and [00:26:00] eluted. And it says the result is “Negative for haptoglobin.” The hemoglobin came out fine. And there is a beautiful band. “Looks like a pure component. Find out what it is.” The next experiment must use a serum control alongside. Because there was a sketch of hemoglobin gel with a band that was faster than that. Might have been haptoglobin, actually.
And so, April 10th, “Overnight Norit A plus hemoglobin,” etc., etc., etc. And it s‑‑ “Negative. No sharp band came out.” I didn’t get anything from that experiment, trying to do [00:27:00] affinity chromatography. It did, in the end, lead to a method of purification. And that’s beginning to become apparent. Wednesday, April 10th, haptoglobin preparation continued. Now I’m using Dowex 1 or Dowex 2, which have an — amino groups, positively charged. pH is less than 10 or 11. And so th‑‑ so they bind a negatively charged protein well. And so activated alumina is also a good protein adsorbent. And Dowex 2 is available in dilute acetic acid. So alumina, etc., etc. And I tried these. And “Both are much reddened by treatment.” But alumina was [00:28:00] awkward to re‑suspend and “appears to wash colorless.” And the “Dowex stays colored.” So it looks to be binding the hemoglobin already. And the second washing of the [subanic?] discolor‑less. So I am getting the hemoglobin to bind to the haptoglobin — I’m sorry — to the Dowex, prior to trying to — the haptoglobin with it. And a sketch here showing that it — “to the alumina was” plus or minus for “hemoglobin, haptoglobin but the Dowex was” three pluses, “haptoglobi‑‑ hemoglobin and some –” query — [00:29:00] “prealbumin.” So it looks as if it’s beginning to work.
And then here is the reason that eventually I did not become the person to describe affinity chromatography, when I make this comment here that I “sub-dry the Dowex plus hemoglobin” and washed it, centrifuge, “Can’t see any binding.” So it looks as if the Dowex didn’t bind the hemoglobin but bound the haptoglobin. “Can’t see any binding.” But “Continue on re‑‑ on repeat of previous wo‑‑” etc., etc.
[00:30:00] The sodium hydroxide-treated resin, water-washed, gave a pH of 8.8. So following page, I set up for an overnight experiment of hemoglobin binding, with five drops of hemoglobin. So water-washed Dowex. “Water-washed Dowex made alkaline with the dilute borate buffer.” And also with alkaline-washed Dowex. Because I’ve got a comment here that “The Wednesday results could be due to general serum binding, followed by specific elution of haptoglobin rather than specific binding.” Specific elution, in other words that the resin was just binding haptoglobin and other protein, rather than [00:31:00] doing it via hemoglobin. “Therefore, include controls of no hemoglobin.” This is more likely, as the — as the serum could bring the pH up to the binding level. So rather quickly, I found out that the Dowex could be used to purify haptoglobin. But it had nothing to do with being used as an affinity gel. So on Thursday, the supernatants, etc., were set up, decanted and — “added water,” etc. And the — (laughs) [it’s?] washed — there — and the gels are set up but doesn’t say what the answer is — on the Thursday. The results are on the following page. [00:32:00] “Alkaline control, pH 8 through 9, almost zero haptoglobin but some impurities. Borate-washed gel, almost zero haptoglobin but some impurities. Water-washed gel,” the control, “5.‑‑ pH 5.1 to 5.2, very good haptoglobin.” So acidic the acidic pH, 5.1 to 5.2, the haptoglobin binds. But a pH 8 or 9, it does not, since… So my theory of — or a hypothesis would be a better word, is that this looked as if, at pH 5, the binding of serum proteins is strong but not so binding as a [00:33:00] hemoglobin-haptoglobin link. Therefore, the haptoglobin can be removed. I’m still thinking that there is hemoglobin on the column. “Try pH 6 and one of pH 4, with benzoic acid or formic acid.” pHs are washing [too?] (inaudible). So that was my hypothesis. An interlude of checking a family.And then April 15th, haptoglobin purification continued. “Resins, Dowex — left over the weekend — Dowex 2 left over the weekend in formic, benzoic, and boric acids,” and, “Sodium hydroxide-treated resins used.” So the pHs [00:34:00] of the formic one were 3.65, the benzoic 5.1, acetic 5.1, 5.2, and the boric acid 7.35. And the resins were suspended in water, etc. And then gels were run, with the gels being called “Acetic resin,” “Acetic borate resin,” “Hemoglobin [control?],” “Formic resin,” “Benzoic resin,” “Boric resin.” And the formic is the best, all ha‑‑ all haptoglobin. So by using a pH of around 3.65, I could get haptoglobin. Acetic acid with boric gave nothing.
[00:35:00] The last two pages of this book say, “In view of the formic acid result, I made a resin and made the hydrochloride of it and washed it to pH 3.8.” So this is a positively-charged resin that’s been made into the acidic form with HCl, and treated with 20% serum and washed four times. The pH came down to 4.6, 4.5. Treated it with the hemoglobin, pH 4.4, hemoglobin, 4.8, etc. “Can probably use a higher p‑‑” I haven’t quite got the solution but I’m getting very clo‑‑ “Can probably use a higher pH, [00:36:00] at a higher ionic strength, to weaken the haptoglobin resin binding.” “Hemoglobin acts according to mass action.” So I’m still thinking I have to elute with hemoglobin. I haven’t got to the point of realizing that hemoglobin has nothing to do with this success of that resin. End of Book XI. [00:36:27]