Oliver Smithies:

So this is book Roman XVI, Roman 16, physical, January 1958 through April.  It starts off in a straightforward way, with a large, one-dimensional gel, with many samples, typing haptoglobin, nothing unusual about the beta globin, and then two-dimensional of the same samples on the following page.  Similarly, on Wednesday, Thursday, January 23rd, just large gels with multiple samples in open slots.  And followed by two-dimensionals of the same [one?].  Thursday, Friday, January 24th, [00:01:00] again, similar experiments with flat gels, horizontal gel, open slots.  And experiment on January 23rd.  These are just one-dimensional gels, photographed, the small gels that have been sliced in the ordinary way, looking for a haptoglobin zero sample.  And one (inaudible), it says, “Suspected negro.  On examination, extremely probable that he has the negro ancestors, judging from the [passage of the 2B?].”  [00:02:00] I don’t know why “suspected” is used in writing about it.  I might cut that out in the transcript eventually.  I don’t particularly like that comment, but that’s what it was written up as.

Friday, January 25, still horizontal gels, quite nice horizontal gel.  And then on Sunday, Monday, January 27th, beginning to realize that it might be a good idea to have a vertical gel, this being the beginning of thinking that you could get over a lot of troubles that way.  [00:03:00] So I made a Saturday, Sunday, Monday five votes for centimeter of vertical gel, with half inch by 30, [one 37-inch?] parallel slot, but there was trouble due to gel shrinkage, and bridge solutions siphoning down the gel, but it was beginning of a vertical gel electrophoresis, certainly.  At last I’d begun to realize that that was the way to get over the problem.  There were some difficulties in making contact between the filter paper bridges to the electrode vessels, and that would have to make it wider at the top, so the gel was not actually kept, although that was the first vertical gel.  I tried some [00:04:00] B samples, royal gel, just for curiosity, I imagine.  More still horizontal gels.  I haven’t yet made an apparatus suitable for vertical (inaudible), though I’d begun to think about it.  Still horizontal, big horizontal gels going on, January 28th, with two-dimensional to confirm things.

Now, here’s a vertical gel with hemoglobin, Tuesday, Wednesday, January 29th.  I’m a little bit puzzled by the result, because [00:05:00] evidently these gels were being run so that you could go backwards as well as forwards.  In other words, the gamma globulin was there.  The slow alpha 2 is very good.  Some loss of contact and siphoning is the comment, but generally very promising, and the horizontal distribution is good.  The resolution is quite nice.  Some more gel tests, and still running horizontal gels here in the routine way.  Thursday, Friday, January 30th.  And then Thursday, Friday, January 30th, a vertical test with a new gel is designed, [00:06:00] which got rid of some of the siphoning problems by having a higher tray for the bridge solution, and having rather thick gel at the end so that you wouldn’t have trouble with contact between the bridging filter papers and the gel itself.  So this is really one of the first good vertical gels, vertical just where new gels is designed, in an attempt to get a more perfect slow alpha 2.  The results are quite nice, not much different from the horizontal, but definitely good in slow alpha 2, and the haptoglobins, the 2B haptoglobins is getting so that I can see [00:07:00] many bands.  I can probably see one, two, three, four, five, six, seven, eight — something like eight or nine bands in the 2B samples.  So it really is a good improvement.  And, in fact, it became very common to use, vertical gels, and I published it as a new method.  We’ll probably come across that.

Friday, Saturday, February 1st, 0.027 [more?] the vertical, half the usual votes for centimeter, 2.5 votes, to try to see what happened to slow alpha 2.  And it just says that slow alpha 2 [00:08:00] is poorer than the fast method.  Looking at it, I can’t see why I thought it was poorer, but at that time they did a very faint pregnancy band in one of the one female type one haptoglobin.

[00:09:00] Some attempts to measure electro-osmosis, George [Cardinal?] serum, with glucose.  The results are glucose is migrating backwards.  It wasn’t very clear.  It isn’t any longer very clear in the image, but it says that the glucose migrating due to borate minus complexion, both detection systems OK, but the silver was best to try to find where the glucose was, and glucose bound to borate.  Glucose can be used, [00:10:00] it says, but the results aren’t terribly convincing.  Electro-osmosis tests on Thursday, February 6th.  Two deoxyribose tests also made, the results comparable to those with formaldehyde.  I was trying formaldehyde.  The formaldehyde one, it was very clear electro-osmosis with [femo?].  So [00:11:00] looking at this experiment, it’s quite an interesting experiment.  February 6th and Thursday, one test was with 0.2% phenol, which doesn’t move, and I have become suspicious.  Two percent glucose, unable to detect any glucose.  The formaldehyde was very clear, move backwards about 15 millimeters, and I presume I’m testing with silver to find these compounds, though there isn’t any note of that particular…  The formaldehyde is quite interesting.  [00:12:00]

Looking at some new starch.  More samples.  Quite nice resolution in the post-[arguments?], February 8th to 9th.  [00:13:00] Some images on February 9th, 10th.  It looks as if I’d used a benzidine stain, although it doesn’t say so, looking for hemoglobin, and the hemoglobin haptoglobin, compared with the staining with the Amido black on the left-hand side of the page.  Fairly clearly, it is very high hemoglobin, rather low haptoglobin in some of them, but I can see clearly type 1, type 2A, and type 2B differences on the benzidine stain.  [00:14:00]

Here and there, [some more family?].  Then, for example, February 13th, Wednesday, Thursday, where the sample — the 659 and 661 are BC, faster betaglobins, quite clearly faster.  And checked with two-dimensional is the comment, and the two-dimensional is on the following page.  [00:15:00] So, in reading these results of this long time ago, I’m currently less clear on the notation, but not in the interpretation that there were two, at least two types of betaglobin, faster and slower.  And vertical gel on Thursday, Friday, February 13th, 14th, with a comment, a note of some contamination with hemoglobin.  But clearly, the two betas are — two of the samples are [00:16:00] the faster beta, BC.  This was a very hot gel.  Evidently, this gel — because there’s a comment — the heating was very marked, and the slow alpha 2 is essentially wrecked, but the resolution is very good, because of looking at the [optical?] in 2B.  One can see one, two, three, four, five, six, seven, eight, something like eight, more than eight bands.  But it was just too hot for the gel.  Eight votes per centimeter.  Vertical, water-cooled gel was attempted on February 20th, and the diagram of how I set it up is shown.  No difficulty with keeping the temperature down, and [00:17:00] volts per centimeter, about 16.  It was nothing entirely satisfactory, but at least it was an attempt.

Thursday, Friday, February 21st, water cooled, run, with flowing water, because — about one liter an hour, six volts per centimeter, etc.  But I don’t see an image of the results at this point.  [00:18:00] And one sample with odd results like 2AN, which was what are called the 2A modified N; N was standing for “negro,” which was not necessarily specific at all, but it was a variant of reduced expression of a 2 gene is what it turned out to be.

Some two-dimensional tests with Tris buffer on Wednesday, Thursday.  [00:19:00] The comment that Tris/[oxo?] is as good as claimed, many, many bands in the (inaudible) acetate paper.  This was [oxide?] cellulose acetate paper, and giving many bands.  I can see one, two, three, four, five, six, seven, or eight bands that were there.  And then some two-dimensional gels — the Tris, two-dimensional Tris filter paper in two-dimensional gel versus XXM buffer, the usual buffer that we use [00:20:00] for two-dimensional gels, with Tris versus XXM.  And they are different, but the XXM is looking at it here, looks better than the Tris, in fact.  Tris, again, compared on — with a 598YD paper on Thursday, Friday, February 28th, with and without hemoglobin.  Not a very good result, looking at it now.  And two-dimensional gels, all of them that — I doubt if I pursued it, but they don’t look — no particular comment, but they don’t look [00:21:00] very good.

Monday, Tuesday, March 2, trying to improve the two-dimensional — Monday and Tuesday, March 4th, decided to test [bridge and barbiturate?] at the same [ion?] extreme.  That’s borate versus barbiturate.  There’s been some preference in the literature for barbiturate buffers.  And [when you?] see the result on Monday, Tuesday, March 24th, a gel, neither particularly good, the two-dimensional results.

[00:22:00] A whole bunch of samples on Tuesday, March 6th.  Six volts per centimeter, with a dog sample for growth hormone tests.  (inaudible) some wider slots, the dog serum, with an excellent result comment.  Vertical gel on Thursday, Friday, March 7th, “OK, at last” is the comment.  This is quite a nice result.  The image is not terribly good, but the resolutions are really quite nice, vertical gels, [00:23:00] which I did publish.  Some two-dimensional on Thursday, Friday, March 7th, with a comment, “not kept,” and just a sketch of the result.  And a slower double beta on Monday, March 10th.  I think I…  Yeah.  I think I withdraw that comment; [00:24:00] I don’t know, really…  They were labeled BC, 2A, BC, 2A.  They don’t look the same on the gels, so I’m not sure quite what that one was about.  Oh, I see what it was.  I misread it.  It’s just with and without hemoglobin on a straightforward gel, one-dimensional.

[00:25:00] Water-cooled vertical gel on Wednesday, March 12th.  Very good setting up with old serum, etc., etc.  Excellent result.  Even the high molecular rate haptoglobins are showing.  It’s a bit curved.  It was just a single — not a very big gel, and the diagram shows how it was set up, but running them vertical allows the large proteins to be resolved very easily, and when I published it I used purified haptoglobin 2B to demonstrate the resolution of the gel, and probably we’ll come across that image in a little while.  Wednesday, Thursday, March 13th, more samples.  [00:26:00] Water-cooled and the hemoglobin color appeared to reticulate, and a little diagram of what that meant, when viewed from about 10 degrees above horizontal, trying to understand this convection that is occurring of some sort in the slots.  So may have to increase the volts per centimeter initially, though it’s obvious that I crossed out decrease, and I think probably decrease is the right answer.  This is the cause of the appearance of the February 1st [00:27:00] vertical gel.  So just some thoughts on varying the voltage initially versus later.

Tuesday, March 18th, tried some rice starch, and some periodic acid.  I don’t really know why I was trying to treat it with periodic acid, but I was.  I probably had talked to Charlie [Haynes?] about something.  [00:28:00] Try antibodies.  I’m beginning to think about antibodies, injected with hemoglobin, and albumin, and serum proteins.  Thursday, March 20th, hemoglobin-coated rice.  That’s what I was trying to do, evidently, to put some groups on that would allow it to bind to the starch for glutination.  I think I’m looking for a way of having an antibody assay.  [00:29:00] That’s the bovine sample on water-cooled vertical gel.  Results very satisfactory.  The image is not particularly good.  Trying to bind things to the rice again.  Oxidation, and coupling with [nitrous O?], it looks like a hydrazine of some sort.  Not nitrous O, nitrophenylhydrazine.  [00:30:00] Going to test with an anti-hemoglobin antibody, made earlier, and with these particles, evidently interested in getting an antibody assay.

[00:31:00] Some rather obscure experiments with starch and periodic acid.  And then Monday, Tuesday, March 25th…  Some rather pretty images of March 27th, 28th, Friday, a whole bunch of unattached photographs and ordinary comparison gel, but there are maybe 10 — well, [00:32:00] perhaps half a dozen, at least — different images of the gel, gel number one, July 4th, it says, and this is March 27th, so they’re probably in the wrong place in terms of where they are in my collection here.  I suspect that these are the ones were used in publishing the vertical gel electrophoresis, and I’ll pause a minute to check on that.  (break in audio)

In this book, in book 16, around March 27, 28, there were a whole bunch of images, photographs, taken of gels one and gel two, and labeled number [00:33:00] 21, July 4, and in checking I find that this collection of images is in the wrong place, so in the scanning, we may have to correct the scans for that date.  The samples scanned in book 16, March 27th, 28th, there were some loose photographs that may have been scanned at that point, gel one, gel two, and gel three.  They belong, really, in book XVII, XVII, 17, June 30th, July 1st.  And I will reinsert them in that position at the present time.  [00:34:00]

So we were continuing with book 16, March 27, 28 there, nothing particularly striking there.  Anti-hemoglobin tests on the next page, results disappointing.  And following day:  “Badly conceived test.  Can’t expect to find antihuman hemoglobin with human cell.”  As a trace of hemolysis will give inhibition.  Also, too little tannic acid used by mistake.  So not very happy about those experiments.  They never went anywhere, anyway.

Reverse photograph here of some samples on Monday, Tuesday, April 1st.  [00:35:00] More tests of the starch granules.  They’re attempting to bind things, all negative, April 1st.  More tannic acid tests on the same day.  And a quite pretty gel to cheer one up on Wednesday, Thursday, April 3rd, just a large gel with multiple open slots, quite good results showing both haptoglobin types, 1, 2A, 2B, and a sample where there was a BC, [00:36:00] beta BC in that one.  And the book ends with some straightforward, large gel on Tuesday, April 8th.