Oliver Smithies:

[00:00:00] This is book Z beginning October 9th, 1968 and running to April 1969.  It’s important to put these experiments into context.  And I’ll do this at a later time.  My immediate recollection is that all of these experiments were related to some theoretical papers that I published in that time suggesting how antibody variability might arise and that I was trying to do experiments that test this hypothesis.  But just looking at the notebooks I’m not really able to resynthesize that mental position as it were, the gestalt if you like of the experiments.  And I’ll have to do that at some later stage when I can integrate the [00:01:00] published work with the notebooks.


[00:00:00] I hadn’t started the book yet, just that preliminary thing, right?  So here I am with book Z.  N terminals on the original 19-24 peptide of heterozygote.  So I’m trying to find out something about Inv at this point.  Talks about Dave.  And I have to go and try to think which Dave that would be at this point.  Not Dave Poulik.  I’ll pause for a moment.  Yes, the Dave is Dave Gibson.  And in fact we’re getting close to [00:01:00] a paper which we later published.  I’m having a little difficult in the timing because the paper I’m looking at was published in December of 19 —




Oh, 1970.  And we’re in October of ’68.  It’s possible that that was what it was.  Because a paper published in Science volume 176 I think.  “Deletions in Immunoglobulin Polypeptides as Evidence for Breakage and Repair in DNA,” which I’ll talk about at some later stage.  But this is Dave Gibson.  The [00:02:00] context of this work is perhaps best explained by reading some parts from the papers.  Which — one of the papers at least — that Dave Gibson and I and Michael Levanon published later in 1971.  This sets the context for quite a lot of the work that had been going on in the past several books.  The statement in the paper is we have started an investigation into the structure of light chains of immunoglobulins from normal human individuals in an attempt to answer several questions which have arisen from structural studies on Bence-Jones proteins and light chains of myeloma proteins.  [00:03:00] In particular it’s important in understanding the origin of antibody variability to know how much if any of the sequence variation seen in Bence-Jones proteins is due to genetic differences between the various myeloma donors and whether the Bence-Jones proteins are in any way nonrepresentative of normal antibody light chains.  We have consequently developed and used methods which allow us to characterize portions of both the variable and the constant regions of normal individual light chains by — and that means by cutting the light chains so that the constant part of the light chains is separated from the variable part.  And as the abstract says, immunoglobulin light chains from normal humans were prepared as their [00:04:00] mixed disulfides with aminoethanethiol, cysteamine, from the IgG fractions of 31 individual units of plasma.  And then goes on to say about the tryptic digests of them were made, etc., etc.  And goes on to talk about the Inv polymorphism and talking about the light chains of kappa and lambda.  Most of the peptides from the constant region of kappa and lambda Bence-Jones proteins were identified in the light chain digests, etc.  Peptides corresponding to the lambda chains were identified in all of the light chains examined, etc.  And it goes on to talk about what this means.  So that’s the context of this work.  It was trying to understand [00:05:00] from normal people what was happening in antibody variability and making sure that the Bence-Jones protein differences were not unusual in some strange way.  So book Z starts with — continues with this task.  N terminals on the original 19-24 peptide of a heterozygote, an Inv heterozygote.  And talks about Dave as I mentioned.  That’s David Gibson.

And then more extensive dansyl tests are shown on page five.  Just checking still the [00:06:00] amino acids.  Various mixtures of amino acids.  So as to identify amino acids in a digest presumably.  Repeat of chromatograms October 11th and 14th.  Thin layer chromatography of 19-24 peptide.  Results are quite clear.  Silica Eastman is better than Gelman chloroform.  Most promising yet on page 17.  Peptide [cleose?] thin layer chromatography page 19 working [00:07:00] with David to get a useful result.  And Michael Levanon I should add was a very enjoyable technician from Israel and later on — and her husband also.  And later on I met them in Israel at a visit much later in my career.  I don’t remember the exact date.  But she was an enjoyable technician, very bright.

I think at one stage in here, I don’t know whether I’ll come across it, but Ellen Fanning was also my technician and they were two of them, young women, and Ellen, I remember at one stage [00:08:00] commenting on her work and telling her that she was a smart lady and this work wasn’t up to her standard.  And I was being quite cross with her.  And she was repentant and worked hard.  And later on was very distinguished in her career.  She went to — she married a German and went to Germany and became a full professor in Germany.  So very conversant with the language and smart enough to become a full professor in Germany, at that time in science was quite a distinction.  Again a grand lady.  Good people.

Large-scale preparations of [Clark?], [Ralph?] Clark, a little comment, Negro.  [00:09:00] And then making peptides from that individual.  Pages and pages of similar sort of work (inaudible) formic acid series on page 39.  Large-scale preparations, 41, 42, 43, etc.  Not very promising on page 45.  [Double front double?].  More peptides and so on.  Peptides from the [basic area?] of 100 milligrams of Ralph Clark light chains.  Iodoacetamide derivatives on October 25th.  It always was [00:10:00] one of my favorite reagents for modifying proteins, iodoacetamide, very reliable.  And still use it.  With lying around in the lab here and there still bottles of recrystallized iodoacetamide.  Always a little bit of iodine formed from time to time as it [decompose?] and [so?] one would recrystallize it to get rid of the impurities.  It was easy to crystallize.

Cyanogen bromide treatment.  This later on was used in one instance and resulted in a paper which I think we may come across in due course.  Where I ran an amino acid sequence for Harold Deutsch and came across a protein [00:11:00] that was eventually realized to be the light chain of the histocompatibility factors.  And I thought I’d found evidence for reinitiation of protein synthesis after a methionine.  But really was a different protein.  We’ll probably come across that later.  Anyway here is cyanogen bromide treatment on page 59.

Long list of possible procedure for [00:12:00] other recombinant from N terminal useful for Edman degradation.  Technical thing.  Iodoacetamide on page 63.  Separating kappa light chains from lambda light chains.  Actually quite important in this general work that I needed to know whether I was dealing with kappa or lambda.  [Frank’s?] procedure 1967.  Procedure was used or was commented on.  And trying to [00:13:00] set that procedure up.  Iodoacetamide derivatives again page 71.  More cyanogen bromide time series on page 73, 75.  Going on next few pages the same type of thing being repeated.  Cyanogen bromide again.  Kappa from lambda on Wednesday, November 13th using copper sulfate.  Cyanogen bromide page 81.

Here we go.  Repeating these various experiments on page — Friday, November 15th, no better.  Best is clearly hydrochloric acid.  [00:14:00] Thioglycolic acid-treated cyanogen bromide being considered and a sketch of the peptides.  Cyanogen bromide cleavage again being attempted.  My recollection being that — and I’m sure others know this well enough — that cyanogen bromide cleaves in front of methionine.

Yes.  I just checked and of course I should have remembered but I didn’t that after cleavage with cyanogen bromide the [00:15:00] methionine is — there is a cleavage between the methionine and the next amino acid.  So one recovers a cleavage product where the amino terminus of the new — of one fragment is the next amino acid to the methionine on the downward side.  And on the upward side the amino acid has been — the polypeptide has been cleaved.  And [so in result?] gets to fragments if there is one methionine.  [00:16:00] Test I think on page 93 is a leucine aminopeptidase I think, LAP, with [Bates’s?] protein.  Trying this scheme on page 95 and the preceding page.  [00:17:00] Saying that the conclusion is not likely to work with precipitated protein.  Try again with a new LAP and perhaps soluble Bence-Jones protein.  More cyanogen bromide tests the following day and thereafter.  Iodoacetamide derivatives page 105.

New N terminal method being tested on page 105, which is an acetylation.  [00:18:00] And comment that Michael tested the result on page 105.  Michael is not spelt correctly.  And continuing to break up these proteins into their fragments and identify the peptides.  Nitrous acid, HONO, again on page 113.  Don’t remember the earlier use of it, but anyway it is again.  [00:19:00] Continuing on the following page.  Cleaning up hemoglobin for some not obvious reason on Saturday, February 22nd.  Lysate made with [Watts?] cells, etc. Deionized pH and [run?] down the column.  Continuing on page 119.  Something odd with this DEAE cellulose method.  [00:20:00] (inaudible) top of the column briefly holds up the hemoglobin and then the character changes and elution occurs.  Seems likely a pH effect, etc.

Now there’s a bit of light on the nitrous acid experiment, HONO, page 121.  General position is that nitrous acid clearly inactivates about 90% of N-terminal amino groups but can also inactivate the next amino acid during hydrolysis, etc.  Some comments of the new idea on the following page.  [00:21:00] Nitrous acid and acetic acid on page 125.  Following pages.  It was straightforward peptide chemistry, such as it was at the time.  Now beginning to use a sequenator because here we have on Monday, March 3rd X149/Z90 Bence-Jones proteins for the sequenator.  The sequenator became an encompassing big stir at about that time as far as my work was concerned.

This was based on a paper published by Edman that I’ll refer to in a moment.  There isn’t any particular introduction here to the Edman-Begg sequenator.  But there is a history related to it.  [00:22:00] Edman-Begg sequenator was published in 1967 for the automated degradation of proteins starting from the amino terminus and cutting off one amino acid after another in a sequential manner.  And it was achieved by having protein in a rotating cup designed in such a way that solvents or solutions could be added to the cup and because they were centrifuging you could scoop them out again with a little device at the top of the spinning cup.  It was a tricky apparatus but very successful eventually.  And I read about the procedure [00:23:00] and I heard that one of these machines was going to be built in Chicago by the Illinois Tool Company of all companies as a result of their being approached by — I guess this is the first mention in the book so far of the sequenator.  There is a very interesting series of experiments that I did for my personal part of it using the sequenator.  That was a procedure for automatically carrying out an Edman degradation which degraded proteins from the amino terminus cutting off one amino acid at a time.  [00:24:00] And Edman and Begg had automated this.  And Emanuel Margoliash of Northwestern University who had a long career investigating the sequences of cytochrome c and their relation to evolution with Walter Fitch in Wisconsin had contacted of all places the Illinois Tool Works to make an updated Edman-Begg sequenator.  And probably through Walter Fitch I heard about this and said I would join with Margoliash in having one of these machines made.  I remember when it came.  [00:25:00] Was very heavy and about a meter square in cross section and maybe one and a half meters high.  Very heavy aluminum body.  And when you opened the door of the sequenator you could see that on the left-hand side was an impossible — at that time an impossible array of electronic cards.  These were in those days integrated circuits but they could have very little on each circuit board.  For example one circuit board might just be there to control one valve in the sequenator on command from a previous part of the program.  [00:26:00] And I remember thinking this thing will never work.  It’s just too complicated.  In fact it turned out to be extraordinarily reliable and very good and the person from Illinois Tool Works who had designed it knew it so intimately that I remember calling him one Saturday afternoon using the machine when it had quit working in a certain way, and I was able to tell him how it had stopped and he said, “OK, look at card number four or five from the left on row six of these,” or whatever it might have been, “on row six of these multiple cards and pull out that card, and you’ll find that the transistor in the bottom right-hand corner has probably [00:27:00] come loose or something.”  And sure enough, when I got there the contact of that transistor was bad and just applying a soldering iron to it and repairing it and back in business.  But that’s how well he knew the system that from a symptom he could tell you just which transistor was wrong and what to do about it.  Quite incredible.  I don’t remember his name.  But it may come back to me in due course.

But that’s the beginning of the first mention here of sequenator.  We did publish quite a nice paper on use of it and we’ll come to that no doubt later.  Paper has some 500 references.  It had some rather good computer programs in it also.

So going on with the chemistry, [00:28:00] we’re doing this precipitating for the sequenator on Tuesday, March 4th, 25 milligrams of Bence-Jones protein, etc.  Dissolved in formic acid easily.  And applied to the sequenator presumably.  Or at least it was crystallized.  Said this gave large crystals by the morning, so it was a pure enough protein that it would crystallize.

With problems oddly enough of in a sense too pure on the left-hand page.  It talks about Friday, March 7th, that’s three days later.  [00:29:00] Applied to the sequenator to avoid a [gel just so driest?] powder obtained.  But doesn’t hold up too well.  May have to alkylate or use DTT, etc., etc. and so on.  Dry hydrolysis being attempted.  We did eventually set up a method so that the — to hydrolyze the derivatives, the phenylhydrazine derivatives, that came from the — I believe they were phenylhydrazine — that came from the Edman-Begg sequenator.  They had to be hydrolyzed to be run on an amino acid analyzer, that was a complicated business too.  [00:30:00] We used to hydrolyze them with hydroiodic acid under vacuum in large desiccators.  So it was a large desiccator sealed with silicone.  Not silicone glue but silicone gel.  And a vacuum applied.  So here’s hydroiodic acid, which would not be colored.  And we’d know in the morning.  And we put those in the autoclave or the oven, I don’t remember which, overnight.  And if things were good in the morning it would be colorless.  If you’d had a leak and air had got in, then it would be brown with iodine.  So you’d get a result in the morning, which gave you some idea of what was going on.  Now I’m amazed at the complexity of the system that we were using at that time.

[00:31:00] And maybe some diagrams will appear in the book to show it.  Performic acid notes on [Dylan 63?] treated in order to get an SO3- derivative.  Formic acid precipitation.  Performic acid precipitation.  On page 113 with little notes.

The theory of better acid precipitation on page 145.  With a comment bad, therefore use IEP, isoelectric focusing, or acidic form, or dissolve in Quadrol buffer.  The results were quite unsatisfactory.  [00:32:00] Derivatives of [Dylan?] iodoacetic and iodoacetamide being looked at on page 147.  And performic acid series on April 1st, formic acid again, etc.  Trying to understand the structure of these proteins.  And that’s the end of book Z.