Oliver Smithies:

[00:00:00] This is the beginning of book sigma, which begins October 19th, 1989 and goes through November of the same year, a fairly busy time, making various constructs that would — could be used for targeting  in ES cells.  So it starts off at the beginning just by continuing work within the previous book, digesting various plasmids, nothing particularly remarkable. It’s the continuing of the work to get the necessary tools. [00:01:00]

And so, continues over the beginning part of this book. [00:02:00] Just mainly technical details of making the plasmids, ligations, for example, of LF-c-kit into p284, et cetera, and beta, h-beta EC Monday, October 30th, page 31, just trying to insert these various elements into a suitable vector.  In this case, this was trying to express c-kit with the idea that, previously developed, that this might help get germline-competent [00:03:00] ES cell.

An interesting paper summarized, or appended on page 32, Monday, October 30th of an important paper, really, from Marilyn Monk, Alan Handyside, Kate Hardy, and Dave Wittingham, that pre-implantation diagnosis of deficiency of HPRT deficiency in a mouse model, meaning that it’s possible to diagnose pre-implantation, and this sort of work later became important, distinguishing embryos, or selecting — [00:04:00] choosing embryos that were suitable for helping families where there was a deficiency, or a mutant gene.

More minis, Wednesday November 1st, page 37, picture 24, minis of ligations of various 2 types, nearly all of the 284-kit look good, pic 21, et cetera.  And continuing that type of work.  Thinking again of getting HPRT- ES cell, Tuesday, December 19th, [00:05:00] page 43, that’s a current status of getting HPRT- from 98-12.  98-12 was, as we’ve seen earlier, an ES cell line that Bev Koller had found gave good results in, after reintroduction into blastocysts, it was a good ES cell line, 98-12.  And trying to find the necessary sequences.

Page 45, December 19th, 1989, is beginning to worry about [00:06:00] methylation possibilities on HPRT, and beta-2M genes. These were the two primary targets of our work at that time, the HPRT locus and the beta-2 microglobulin locus.  HPRT, Tom Doetschman, particularly, and beta-2 microglobulin, was Bev Koller.  And some comments on 9-8-12 with a PAT-153, which was a cloning vector that had been developed by various people, and it was available commercially.  It’s a deletion derivative of pBR322, and gives [00:07:00] good plasmid copy number.  So there, on this page is a summary of what I was thinking about over time, in regard to methylation.

And, here is an insert on page 47, talking about varied amplification across beta-2 microglobulin primers, et cetera. [00:08:00] And a diagram showing on the opposite page, where in the beta-2 microglobulin gene, we proposed to insert a neo sequence so that it would inactivate the locus and give us at the same time a selectable marker for isolating the clones.

So it continues on the following page, 49, Thursday, December 28th, with digesting, 98-12 DNA with a view to inserting the neo gene.  And, this continues over the next [00:09:00] several pages.

Usual technical problems encountered, for example, Tuesday, January 2nd, page 61.  “The results show strong concentration dependence of digestion of the plasmid with MboII, HpaII, and MspI.”

PCRing 98-12, following page.  But worrying about contamination, as [00:10:00] detailed on page 62, “Beginning to show contamination or getting better at detecting enzyme problem, or both.  Try to be more careful still,” worrying about the efficiency of PCR, and that one can be misled by contaminants.

And that is exemplified on page 67, Wednesday, January 3rd, “98-12 PCR, super-careful.  New distilled water, no sleeve, washed the containers,” et cetera.  “28 cycles.  Results still either water and enzyme contamination,” [00:11:00] so worrying about contamination.

And here is some comments on Hooper’s HPRT- cell line.  This, in the end, proved to be, of course, the one that was most valuable for us in getting correction of the HPRT locus and germline transmission that Martin Hooper’s cell line which has a deletion in it, and here is a series of maps on page 68, presumably from Hooper, although it doesn’t say so, [00:12:00] giving a map of normal mouse HPRT, and the map of Hooper’s HPRT- gene, which had a deletion.  This meant that it couldn’t revert spontaneously, since it was a deletion.  You have to supply the missing part, and a critical element in our success was going from HPRT-, a deletion, to HPRT+, where the background was — essentially was zero.  Not “essentially zero,” it was zero.  But still trying to make our own HPRT- on page 69. [00:13:00]

With amplification or emphasis of this problem of contamination on page 74, an extract from nature, January 4th, 1990, “Shedding light on PCR contamination: the most pernicious problem plaguing the widely-used technique of PCR is contamination of reagents with previously amplified material, so pointing out that, again, the importance of never having [00:14:00] the product of a PCR reaction in the same area as setting up an experiment in order to detect that product, because it would lead to contamination.  Routine precautions on page 75.

And going on with this, Monday, January 15th, et cetera, “Suspect one of the reagents blotting of HPRT- samples,” page 79.  [00:15:00] That’s Vicky’s probe, Vicky Valencius made a very important contribution as a graduate student later by showing that one could correct various size deletions, that it was quite possible to correct big deletions in a plasmid, in her case, HPRT- plasmid, with different deletions, and that one could repair them.

The following pages are with Vicki’s probe on the HPRT- samples obtained from 98-12, four or five pages of blots, with a conclusion on Monday, January 15th, [00:16:00] page 87.  The best conclusion is that, “None of the clones are as expected.  All appear to have lost the plasmid, and must have reverted to the originally 14,” et cetera, usual failures.  And so, repeating.

Page 91, Wednesday, February 7th, trying to insert the wild type HPRT+ gene into E14. [00:17:00] So, the E14 from Martin Hooper is, the gene has, ends at the — after the exon 3, and it’s missing the beginning part or so.  This is talking about inserting the wild type gene to get a crossing over, which would, generally at HPRT+, [00:18:00] where the crossing over is 5’ to the gene in a region 5’ to the gene, and crossing over has to occur close to exon 3, which is in E14, with wild type DNA.  This is the critical design, actually, that was eventually successful.  And I’m not sure at that point we realized quite how important it was to do it this way.

And Laura Reid had a plasmid called pMP5, which is across the junction between the — across the deletion junction of E14. [00:19:00]

So here on the following page, page 93, Friday, February 16th, we’re thinking about, beginning to think about correcting the cystic fibrosis gene.  And, so that this is some thoughts on how to get a knockout of the cystic fibrosis gene. [00:20:00]

With comments the following day, page 95, Sunday February 18th, from Beverly Koller pointing out various problems.  But, definitely this whole business of getting homologous recombination in ES cells, or in fertilized eggs was being considered generally, so page 97, Friday, June 22nd, talked about homologous recombination in fertilized eggs, the aim being to transfer the ES cell type technology to fertilized eggs, beginning [00:21:00] to think, was that possible?  And also, in bone marrow stem cells, when we were suggesting it would be useful to use a white locus mutation to help assaying the whole animal, or in a Dexter culture, because we had now with us Sallie Boggs, who was a visiting professor who came to work with us, and she was a great expert at Dexter cultures, as they were called, bone marrow cells.

Further thoughts on increasing the frequency of homologous recombination on page 99. [00:22:00] I’m thinking it might help to have a partial region that was single-stranded.  So, on page 101, Tuesday, August 7th, there’s a type of scheme being considered to have an extrusion of a hairpin that looks a bit wild at this point, but no doubt at the time, it seemed quite logical. [00:23:00]

Thinking about single-stranded page, PMP8, Wednesday, August 8th, page 107, thinking about getting PMP8 to grow as single-stranded — [00:24:00]

Various gifts from people, from William Crain, a pair of beta-actin primers, page 112, and the following page on 114 [00:25:00] material from Allan at Baylor.  That was Allan Bradley, I’m not sure.  But it was pBR322, plasmid containing part of the HPRT gene.

Reminiscence a little bit on page 117, Tuesday, November 6th, on ways to increase homologous recombination.  Looking at a paper that had come out September 7th on meiotic gene conversion and crossing over, and their relationship to each other and to chromosome [00:26:00] synapsis, et cetera, paper from — by JoAnne Engebrecht, Gene Hirsch, and Shirleen Roeder, published in Cell, and in the discussion, they talked about the relationship between gene conversion and synapsis, and that synapsis is a direct consequence of homology search mediated through standing base with a comment that, “This was our idea,” and it reminds me of Smithies and Power paper, and I was thinking about it, of strand invasion being important for synapsis [00:27:00] and leading also to gene conversion.  So I got interested again in single-strand invasion, and began to think about doing experiments of that type.

It’s talked about on page 121, Tuesday, November 6th, “feelers #1, the simplest experiment is to test the nicking, and then go on to add synthesis,” et cetera, testing the idea that feelers might increase the frequency.  So this is, idea is taking up experimentally [00:28:00] on Wednesday, November 7th, page 125, nicking, typical nick translation et cetera, uses, and so on, talking about beginning by nicking the DNA. And on the following pages, and gel checks, for example, Wednesday, November 7th, page 129, the conclusion is, the new range is very good, but could use old, although losses were severe note, the presence of DNA-sensitive sites in the DNA, et cetera, trying to get this idea of feelers into practice. [00:29:00]  And just continuing here with developing techniques to do it.  Saturday, November 10th, page 135, conclusion is, “The changing buffer changed the nicking, borderline enough, and no cross-linking,” et cetera.  Just experiment after experiment.

And the following page 136, conclusion from the nicking experiment is, “Proceed, but remember that nicked DNA becomes shear-sensitive.”  I managed to obtain some E. coli single-strand binding protein. [00:30:00] Let’s see who it was from.  Jack Chase.  And so, he wrote a feeler Stage 2, further discussion with Jack Griffith, and his friend, Jack Chase.  Best test was to use sequenase plus a single-stranded binding protein as a test.  And so we continue with Tuesday, November 13th, page 141, trying to make [00:31:00] these feelers.  Adding feelers on Tuesday November 13th, page 143.  “Conclusion: no sign of action.  Repeat with a new lot of enzyme.”

Feelers again on Wednesday, November 14th.  Beginning to think it might be working.  “Conclusion: the single-stranded binding protein certainly has an effect, [00:32:00] and it looks good.”

And so, then taking the material we got to Jack Griffith for looking in the electron microscope, Friday November 16th, page 153, EM data, and the results are clear.  Polymerase plus single-stranded binding protein definitely makes single-stranded DNA branches.  Looks like about five to six for — per full length plasmid.  So, a rather pleasing result with the single-stranded binding protein being — [00:33:00] appearing in the electron micrographs as a blob, a rather dark blob attached to the DNA, which of course appears as a single line.

And so the book ends on Friday, November 16th, with tests of plus or minus single-stranded binding protein after extraction with phenol.  The results are essentially as predicted, the pattern returns to pre-phenol, so the clean-looking DNA is still branched, was the conclusion.  And that ends book sigma.