Oliver Smithies:

[00:00:00] So here we begin on perhaps the best notebook that I have in my collection of notebooks, April 7th, 1982, beginning, and ending on August 20th of the same year.  So here we start with continuing the immunobiotin work and summarizing to date the conclusions, on page one, Wednesday, April 7th.  And various comments on the different pages of previous work.

And continuing, on page three, to begin check of a new vector, “DH beads, bio, and pH 4 elution,” still all in the [00:01:00] same vein.  And the supernatants are extremely hot, as shown on page five.  But the eluates are very, very low in amount.

Conversation with Dave Ward.  Because I’m getting so frustrated.  Talking to him again about things.  And so that usual yellow page of — two yellow pages recording the conversation and his telephone number — and confirming that only very high pH works with immunobiotin, due to the pK being about 12.5, and confirms “low substitution level leads to lower recovery” but warns that “high substitution leads also to non-elution with biotin,” [00:02:00] so that there’s some sort of a maximum.  “Recommends using biodUTP and antibody plus protein A Sepharose.”

And here I am, April 9th, still at it, with heavy labels.

And then, on page 11, is the last test with biotin and attempt to label, and on page 13, which is April 9th to April 22nd.  There’s a gap of about two weeks — when we come to page 13, in Book γ, Thursday, April 22nd, my [00:03:00] change to working no longer with immunobiotin and origins of revocation but to gene targeting.  The history of that transition is really quite simple, that I was very frustrated with the immunobiotin work and not enjoying it.  And then I was teaching at the time and came across a paper published April 1st, 1982, in Nature, which I used in my course, which is called 711.  And this is Paper 37 in that course.  And the paper was by Mitchell Goldfarb, Kenji Shimizu, [00:04:00] Manuel Parucho, and Michael Wigler — Mitchell Goldfarb in Michael Wigler’s lab.  And in this paper, they isolated and characterized, in a preliminary way, a human transforming gene, which they isolated from T24 bladder carcinoma cell.  And their procedure was quite complicated.  And it took me some time to understand it well enough to be able to teach it.  But it was right in line with what was going on in my teaching of the Molecular Genetics course in Wisconsin, at that time, which…  And I always liked to use original papers, in this course.  And so this was one of those papers.  And I had to work hard to understand [00:05:00] it.

And they had used DNA from a cell line called T74, a bladder carcinoma cell.  And they knew from previous work that this can induce transformation of the cells from contact inhibition, using NIH 3T3 cells…  They normally stop growing when the dish is full.  But if they’ve been transformed, then they produce colonies which are rather difficult to score, just a clump of cells which are rather higher than normal and not easy to score.  But nonetheless, this transformation is possible.  And it was known from previous work that this was [00:06:00] due to DNA from a transformed cell.  Therefore, there was a transforming gene in the DNA.  It was also known that, that DNA, the transforming principle or gene, if you like — and that’s what they called it, a transforming gene — was not cut by some restriction enzymes and was cut by others.  And we will see what restriction enzymes they used.  But the principle of the method was very straightforward.  From transformed cells, they isolated the DNA, cut it with a restriction enzyme that did not cut within the transforming gene [00:07:00] but, of course, reduced the sizes of fragments very markedly, and then ligated to this digested material a small piece of DNA, the SupF gene, which is a suppressor of amber mutations.  And so they have now ligated to the transforming gene a fragment of DNA that will enable a bacteriophage with an amber mutation to grow.  So their idea was take the — I think it was HindIII-digested material — yeah, HindIII-digested material was ligated to the SupF gene, used for transformation again.  And then isolate [00:08:00] the DNA, once again, from the transformed cell.  But then cloning that DNA into a lambda bacteriophage, that could accept the HindIII fragment.  And only bacteriophages that picked up the SupF gene would propagate.  And so this would enrich for fragments to which SupF was bound.  And then take the phage DNA, prepare the phage DNA, and use it to transform again, so that this round would now enrich for the transforming principle.  And by doing several successive transformations [00:09:00] and lambda bacteriophage cloning, using a bacteriophage that needed suppressor F in order to grow, they were able to isolate the transforming principle.

My way of describing that in lecturing, these days, is that I talk about the SupF being a blue piece of DNA and the target gene, in their case the transforming principle, being in the red DNA.

Nobuyo Maeda:



They had a method of —


You can stop some‑‑


— finding red —


— you can stop somewhere.


— yeah — finding red next to blue.  And then I realized that that same principle could be used as an assay [00:10:00] for gene targeting.  And page 13 of this notebook, which is only two weeks or thereabouts after their paper came out, is an assay for gene placement.  And the “Aim:  To place corrective DNA in the right place.  And need an assay for developing the technique.”  And then a diagram of the proposal.  And I’ll discuss that a little bit more later.  So —  [00:10:33]


[00:00:00] OK.


So the next page is still a leftover from the biotin work.  Stanley Kline sent me some biotin dUTP.  But, in fact, I never did use it.  Because as we see on page 17, Tuesday, May 11th, I’m now beginning to think about how to make a construct of the type illustrated on the plan on page 13.  In that [00:01:00] plan, I did not say exactly what the construct was that I intended to make.  But it was going to be a large piece of DNA, extending from partway into the A gamma gene and going all the way to downstream of the pseudo beta gene, followed by the blue piece of DNA, the tk SupF gene, tk being thymidine kinase, needed to make sure that the cells that I was transforming had taken up the incoming DNA — and the SupF being the blue piece of DNA that I was going to use to help me recognize that I reached the target [00:02:00] gene, the target gene being beta, and which I call the red piece of DNA.  So the whole aim was to find the blue DNA, which was on the incoming DNA, next to the red DNA, which was the target gene, by finding a piece of DNA — illustrated at the bottom — which I call the recombinant fragment, SupF gene, next to the beta gene, so the blue next to the red.

So the first need, of course ‑‑ was to make that plasmid.  This was a going to be a rather big plasmid.  And I knew some things about what was available.  And those things are listed on yellow sheet, on page 16, [00:03:00] talking about what I might use, with a map, which was a very important map for me, of cosmids, very large plasmids, that were available, particularly HG3, a 41‑kilobase of DNA, extending from — and covering almost all of the beta globin genes, starting with the epsilon embryonic gene and going as far as the delta gene.  So that these were DNAs that I hoped to be able to use.  And particularly, on page 17, Tuesday, May 11, the comment that Clone 3 — [00:04:00] Clone 3, which is a HG3 — is a source of a Cla1 fragment, extending from within the epsilon gene to downstream of the pseudo gene.  So that was a Cla piece, which is quite easy to see in the restriction map there.  So I was going to use that.  And with a comment that “It probably doesn’t need to be purified.”  And now where am I going to get the suppressor gene from?  Well, there was a plasmid, p3pi vx as a source of the suppressor gene.  So I had to ask for somebody to help me by supplying that gene.  And no doubt we’ll come across who that was.

So still, however, [00:05:00] a little bit of reminiscence, one might almost say, of the bipolar Alu repeat variation.  An assay for that talked about, on page 19.

And here, on page 21, three DNA samples are being talked about, regarding p3pi vx supF fragment, that were available from Elio Vanin and Natalie Borenstein.  And so here my aim was to make a construct, which is outlined at the bottom of that page, Tuesday, May 25th, to [00:06:00] isolate the suppressor gene on a ColEI plasmid, with a polylinker upstream, into which I could put whatever I wish.  And so here, the next few pages are devoted to finding that piece of DNA.  And the typical diagram, on page 25, of a single cutout comes, of ColEI, cut with various things, and what the expected products might be.  Carrying out these experiments — or these steps, on May 26, page 27.  We go on with the straightforward matter of making something.  [00:07:00] With the conclusion this was a time series that, “25‑minute digests should be close to ideal,” for a single cut.

Large-scale preparation of the plasmid, on Wednesday, May 26, page 29.  So going quite quickly.  And continuing, on the following pages, with a preparative gel, on May 27th, page 33, obtaining a supernatant in which there was a ClaRI fragment, that was what I was wanting to make.  Testing at least [00:08:00] this fragment, on page 35, May 28.  And straightforward work in purifying and characterizing.

Makes me smile — on page 39, which was a gel, on Wednesday, June the 9th — “Still crazy dye.”  I don’t know what that means.  But the dye must have been a little bit odd.  And certainly the photograph of the image of the gel shows what looks like three dye bands.  There must have been too much dye.  But on the left‑‑ the comment that “This is fine.  No curvature, no heating.  Could go to 40 V/cm.”  Always enjoying gel.

[00:09:00] Then, on page 41, June 21st, Monday, I’m thinking of a test scheme for general recombination in mammalian cell.  Talking to Steve Shapiro, the idea was developed of checking for correct-site recombination, using two different site mutants, etc.  So here is another possible way of doing things.  But my recollection is I did not follow that up.

Again going back to biotin experiments for Dixie Mager, who was doing some, still, biotin antibody tests, on Tuesday, June 22nd, a [00:10:00] day before my birthday.

Now, on Thursday, June 24, looking for pSV2 Neo mutants — and in thinking of…  I’ll pause here.  (break in audio)  Yes.  This, on page 47, Thursday, June 24th, I am following up.  I was mistaken.  At least I did follow up, to some extent, the outline of scheme on page 41, and which was [00:11:00] going to use the neomycin resistance gene for testing purposes, in these experiments.  So I’m looking at that sort of DNA, pSV2 Neo, on June 24.  Following on the following pages.

Again, the usual introduction of Jake for concentrations of DNA that would be the Stockmayer relationship, on page 51.  “The Jake concentration for 5.7 KB,” so and so, “40 mg/ml.”

Looking for a probe [00:12:00] for the neomycin resistance gene, on June 25th.  And, “See also NM/L‑139.”  That is presumably Nobuyo Maeda, Book L, 139.  So thinking about transformation with Neo and having a Neo probe, on this page.  And indeed, it was with Nobuyo, because, on page 55, it says, “Nobuyo has fresh but untested C600 cell.  And Debbie –” Debbie Reynold — “Debbie…”  [00:13:00] Yeah.  (break in audio)  So I’m still on page 55.  Debbie Endean has a fragment, etc., etc.

So Saturday, June 26, still continuing to screen for the pieces of the Neo gene, the left-hand side and the right-hand side, etc., etc.

Now here, on page 59, be talking about getting pSV2 Neo Gpt, from [00:14:00] M. Dieckman, in Paul Berg’s lab.  There was very easy and good collaboration between everybody in the field would send whatever DNA they had, to help other people.  So here a message from Steve Shapiro — “Dear Oliver.  Having closed,” etc. — Clone 3, which is the big piece of DNA that I wanted, with the Cla1 fragments.  So this is really a critical piece of DNA for the large targeting construct — not really for the pSV2 Neo experiment.  This…  So let me repeat that.  The letter on page [00:15:00] 58 is from Steve Shapiro, telling me that he has sent me the Clone 3 piece of DNA that I wanted — as he said, “which contains most of the epsilon 1 gene and about 15 KB 5’ of the –”  Can’t really see.  But it — like it was of the delta gene.

And the opposing page is not in the same vein.  That’s page 59, June 28.  It’s [00:16:00] continuing to talk about pSV2 Neo now.  Here we go on, on page 61, Tuesday, June 29, looking at the piece of DNA that is critical for my construct of an R1, BamHI fragment, containing the suppressor gene, and next to delta part of the beta globin gene, but not the whole of the beta globin gene.  So here is a review, on page 60, of the SupF vector for the tests that I was wanting to do and a [00:17:00] map of what I wanted the final construct to look like.  This is the vector containing the neomycin gene.  Continuing in the same vein, on the following pages.  Typically, page 65, Wednesday, June 30th, with a map of the construct continued, EcoRI fragment, at the end with delta beta in it.

[00:18:00] On with the process.  “Complete R1 digest to get a 5.2‑KB fragment.”  So the 5.2‑kilobase is OK, on page 71.  I have to check again what the 5.2‑kilobase fragment is.  (break in audio)  Returning, for a moment, back to page 60, reviewing the fragments that were needed.  The second of my outline [00:19:00] diagrams that are attached to page 60 shows that the ClaRI SupF fragment, with an Eco‑‑ correction — a Cla EcoRI fragment containing a SupF gene is made.  And it’s necessary to make a longer piece — next to which is going to be the 5.2‑kilobase fragment that’s referred to later.  And so some further experimental tests begin, and on that page.  [00:20:00] (break in audio)  So as we were continuing, after page 69, “Complete R1 digest to get” the “5.2‑KB fragment,” a necessary part of the construct.

Pointing out that, although the 5.2‑kilobase fragment is OK for one purpose, it contains too much pBR322 sequence for use as a probe, on page 71 — that sort of comment.

Trying to make linkers.  Because one couldn’t buy linkers, at that time.  So a small linker prep gel was made, to get a [00:21:00] fragment that was a Sau Bam fragment and a Cla Sau fragment, etc.  So at the end of the page, there’s a 280 linker Sau Bam, gamma 73, and a Cla Sau linker, of 625.  Trying to get over what would now be so easy to synthesize.

And then, Saturday, July 3rd, through Monday, July 5th, page 77, partially digesting this 5.2‑kilobase fragment to get out the pieces that I was wanting to get.  But not a very satisfactory outcome, with the comment, on [00:22:00] page 76, that “The 5.2‑KB partial is complex.  May have been too impure.  Keep it as a probe but start again, using single-Bam cut route.”  So that is what is done on the next page, Monday, July 5th.  BamHI single cut.  And on page 81, July 5th — Thursday — or Tuesday, July 6.

Some comment, “Some material that Fred Blattner had of pK‑7 DNA” — and a sketch of it — of it.  Derivation is there on a letter to Fred Blattner, from “Tom.”  I do‑‑ I don’t whether…  That’s probably Tom Maniati‑‑  But I don’t know for sure.

Continuing to make these little pieces needed for assembly of the targeting construct, on page 85, with a diagram of the R1 fragment, containing SupF and delta beta, again.  And continuing, [00:24:00] and pages and pages of the same quest.

So tests of the various pieces made, on the stage, on page 91, Thursday, July 8th.  The conclusion that “The 4.2‑KB fragment looks clean but did not digest,” etc.  Usual technical problem.  Retesting the fragments, on the following page.

Thinking about ligating things together now but — on Thursday, July 8th.  And then thinking about doing it, again, on Saturday, July 17.  And finally, evidently, doing it on Tuesday, July 20, [00:25:00] and first ligation obtaining an EcoRI Bam Sau linker fragment.  (pause)

And various Neo fragments, delta left and delta right of Neo.  And, “Alert Natalie had to clone up…”  I think this is probably the beginning of a digression, which did take place at that time, making fragments of [00:26:00] the Neo gene, here called delta left and delta right, which together could make a complete gene but, by themselves, neither was effective.  So one could use delta left and delta right recombination to find evidence for plasmid-by-plasmid, as we called it, recombination.  I think this is probably in that vein.  But it’s not specifically stated there.  Yes, I think that’s fairly correct.  Because on page 101, the digestion of a pSV2 Neo Gpt, [00:27:00] “for delta left and delta right tests.”  The cutting looks OK.  With, Saturday, July 10th, page 103, a conclusion, “Not convincing that the fragments are as hoped for but probably OK,” etc.  So delta left and delta right.  (pause)  [00:28:00] So there’s continuing to make cells that transform delta-left Neo and delta-right Neo, on Sunday, July 11th — transformations.

And so there’s a — somewhat jumping around a little bit between making what’s needed for the major test and this, perhaps, partial digression of making plasmids that could be used for testing plasmid-by-plasmid recombination.  I remember at that time we became very much involved in thinking about what was possible.  So a plasmid-by-plasmid recombination in a eukaryotic cell, in a [00:29:00] mammalian cell was considered one of the things that was necessary.  And then introducing one of the fragments into mammalian cells, so one would have a plasmid‑by‑‑ incorporated plasmid, chromosomal plasmid.  And then eventually getting up to the real experiment, which would be chromosomal DNA-by-chromosomal DNA recombination.  So, in fact, we did write some papers of that type, which we’ll probably come across in a little while — and certainly I can refer to.

So Monday, July 12th, replating [00:30:00] the Sunday transforms, getting delta-left and delta-right cells.

And with intermissions, again, to get the things needed for the chromosomal targeting experiments, a large supply of the 4.4‑kilobase fragment, on July 14th, Wednesday, page 119.  Reviewing, again, the slow progression to make the targeting construct.  Friday, July [00:31:00] 16th, page 127, a review of the EcoRI Bam Bam suppressor F-containing fragment, joined to part of the beta globin gene, delta beta.

So here I am again trying to put a linker on this 4.4‑kilobase fragment, on page 129, July 20th.  Continuing to test, as the fragment got longer and longer, by now having [00:32:00] two linkers on this fragment, to get the SupF delta beta fragment now with two linkers, that let the end become a Cla end.  So an EcoRI Cla1 fragment laboriously constructed.  A prep gel for that fragment now, on page 137, Monday, July 26th.  (break in audio)

This stage, my notes are somewhat confusing.  Because two things are going on at the same time.  And that is the preparation of the necessary targeting construct [00:33:00] to test the idea of targeting by using a large piece of DNA as the incoming DNA, carrying with it the suppressor gene, and targeting the beta globin gene, as described on my favorite page, γ‑‑whatever it is — I — yes, γ‑13, page 13 in γ, that being one set of experiments, the other set of experiments being related to making plasmids, in which we could test what we called [00:34:00] plasmid-by-plasmid recombination.

And the easiest way for anyone that’s interested in sorting that out is to go and look at a paper that I published with — myself, Mike Korelewski, and (break in audio) and, this paper, with Mike Korelewski, who was my technician, Kui-Yung Song, who was Raju’s student, and Raju, where we talked about “Homologous Recombination with DNA Introduced into Mammalian Cell‑‑” which was published…  I think it was published in — in Cold [00:35:00] Spring Harbor Symposium.  But we can check that reference later.  But that paper describes the various things that we were thinking about.  Plasmid-by-plasmid recombination was the pSV2 Neo, the Neo delta-left, delta-right.  That could be done with plasmid-by-plasmid recombination.  And so this was making those tools.  And then the next level was what we called plasmid-by-chromosomal plasmid.  In other words, a plasmid has been introduced into some region in the genome, where it could enter, and then supplying perhaps a right arm, and the left arm could be brought in.  And one could demonstrate plasmid-by-chromosomal [00:36:00] plasmid.  But the problem with that being that, of course, it was not a general feature of gene targeting.  Because perhaps the chromosomal plasmid had only entered the genome at, I’d say, a susceptible site.  So that was the next level but was not really what we were trying to get.  And then there were plasmid-by-native gene recombination, which is the one that we most wanted to be able to demonstrate.

And then in that paper we show a test of — that we could really score that type of construct.  So we made an artificial test, where [00:37:00] one of the plasmids was the same general structure as we hoped to use in our chromosomal work.  We called it a mock target — contained the suppressor gene and the beta globin gene.  And then used a rather simplified form of the incoming DNA — and demonstrated that we could find the recombinant fragment that would become part of our real assay, in other words, where the red and the blue DNA were together.  Now this was, in a way, a rather dangerous experiment, as we later realized, because it meant that we had already made [00:38:00] a plasmid that was to be our recombinant fragment and might later be accidentally introduced as an impurity — following some other experiences that I’ve talked about.  So Figure 6, and that paper, which shows that you can indeed find the blue-by-red fragment, was also a dangerous experiment.  And the paper in Cold Spring Harbor stops at the point of the real experiment, because we’d not yet made the real experiment work, which was incoming plasmid-by-chromosomal — a native chromosomal target.  That’s the difference, a native chromosomal target rather than artificial [00:39:00] one.  But this Cold Spring Harbor paper can help sort out what was happening, if anyone wishes to pursue that.

So page 137 is just going on making the various fragments needed for these tests.  (pause)  So here is a fragment being made, on page 139, Tuesday, July 27, trying to get a 5.6‑kilobase fragment, which would have the suppressor gene in it and the beta globin [00:40:00] sequences.  And with the 5.6‑kilobase ligation being continued, on page 145, Friday, July 30th.  (pause)

So in these various experiments, we’re at least producing a product.  Because, on page 149, we see that Debbie continued the screen and two very strong positives were obtained — which were being looked at as possible maps of what might be the fragments.

And continuing, with the end of this book being Sunday, August 22nd, more tests of many preparations, with the conclusion that [00:42:00] “Candidates #27, #359, and #373 probably have the same insert that cuts out with Cla,” etc.  So we end Book γ, probably the best book in my whole collection.