Oliver Smithies:

[00:00:00] So this is the beginning of book tau starting November 15th, 1990 and ending April of the following year, 1991.  A very active time.  Continuing the thought of looking for feelers, feeler length experiment, etc., on page one.  Feelers for the EM on page three.  Nicking tests again on page five.  Working on Thanksgiving Day, nicking continued.  More seriously nicked DNA on Monday, November 27th [00:01:00] page 11.  But the conclusion, it was too gentle.  Repeat with twice the enzyme, etc.  So repeated nicking on the next page and EM again on page 14 and 15 with single-stranded binding protein added before the electron microscopy at different levels.  And so at zero level, and 1x and 2x and 4x.  Let’s see what those are.  It’s not very clear from just looking at them.  Tuesday, November 27th then,  page 15.  The results from tau 2 results.  Simple molecule, etc.  [00:02:00] The different products obtained.  The different stages of the reaction.  First scale-up of the feelers on page 19, Thursday, 29th of November, with many — with the EM showing extensive regions of single-stranded feelers on the DNA.  Conclusion, looks very good on the gel.  Guess about one nick every three kilobases.  But Jack and Carl tested and found the result very good.  [00:03:00] So the gel was suggesting things were good and the EM showed they were indeed good.

Whether it helped still has not been done.  Just making them.  So the plan was made Monday, December 3rd and executed Tuesday, December 4th, page 21.  Bulk test by Denise and a gel and EM work and this is the gel part of it.  [00:04:00] Concluding that the branches are complete.  The feelers are good with about 5% of the molecule being full-length, deproteination works as expected, etc.  So here we see the results of the — so that page 21 was the gel results and page 23 was the EM result.  And single-stranded binding protein added to all of them before fixation.  And sample 67.  Branches visible but burnt.  The protein is common at the forks.  Not just at the ends.  Not quite as expected.  So the [00:05:00] protein might be polymerase.  Made me think of other ways the experiment might have proceeded.

So here is the first test of the ability of this DNA to transform with and without the feeler.  Page 25, Tuesday, December 4th, the first look on December 12th.  Eight days.  Single-stranded binding protein experiment [00:06:00] 6.7 SB 6.7 C various preparations.  Clear colonies but no difference.  With a conclusion that the result is pretty clear.  Branching double-stranded DNA and single-stranded DNA plus single-stranded binding protein are poorer than simple nicking.  Jack suggested that 5’ feelers are not what we want which is quite possibly correct.  That you need a feeler that ends in 3’ end so that it can be extended by scissors of polymerase.

[00:07:00] So on page 27, continuing these thoughts, Wednesday, December 12th.  The data begin to suggest that another product other than double-stranded DNA is causing recombination.  How else can explain the results from tau 12?  And plan for Friday, December 14th executed then and look at the result.  Various preparations P, Q, R, S, T, U.  And with the excellent series.  [00:08:00] And various causes.  And all of these are reminiscent of my tendency has always been to test hypotheses experimentally and sometimes come back again to look at something that had been an idea earlier.  So this is as I said already several times a testing of the idea that strand invasion by a single-stranded DNA could increase — could cause synapsis and increase the frequency of recombination.

Now working on nicking for 3’ feelers Thursday, December 20th.  [00:09:00] Better type of feeler.  Nicking about correct.  The Bam and Sam and Sal was not yet enough.  So 3’ feelers again Thursday, December 20th with a comment, don’t do this.  So on page 41 [00:10:00] Friday, December 21st looking at the gel being repeated.  And single-stranded binding was evidently OK.  Looks like a modest amount of feelers.

Some complicated ideas being pursued on Friday, December 21st.  [00:11:00] Plan page 45.  And looking at colonies again.  First conclusion was no sign of increase.  But it’s still remarkable how extensive the degradation can be and still things work.  With a new conclusion, still same effect but less marked.  Denaturing does not help.  Beginning to think of other 5’ exonuclease using that different ways of making these feelers.  [00:12:00] Going now to use lambda exonuclease on page 57 Friday, January 14th.  No sign of activity under these conditions.  Larger test on page 59.  Needs EM to interpret fully.  But proceed with the tissue culture anyway.

Lambda exo at a lower temperature after trying higher Sunday, January 6th.  Temperature control is [00:13:00] effective but try 25 degrees.  So testing exonuclease at 25 on page 65 Monday, January 7th.  Progression is very rapid.  So large-scale test trying to make 3’ ends with a little diagram of what they would look like.  They would be double-stranded DNA on both ends.  There would be single-stranded DNA with 3’ terminus.  So could be used to invade and then extend.  [00:14:00] But the results are not very pleasing because although the gel suggests that the products are just about perfect, when the colonies are looked at, the conclusion is that putting on a modest 3’ end has only a slight effect, and going over to an extensive 3’ single-stranded region reduces targeting.  So not too promising.  Trying to be more specific on page 71, Tuesday, January 8th.  With the simple 5’ exo [00:15:00] that would degrade the 5’ strand and leave feelers.  Single-stranded 3’ ends.  Or one could go complete and get single-stranded DNA.  Compared A with B.  Partial versus complete single-stranded.  And as we go back to look at the EM on page 75, January 9th, Wednesday, the conclusion that ssDNA does not work.  Looking at the number of colonies.  [00:16:00] The structures reform but don’t help much even though the EM shows single-stranded branches.  So there are fairly good single-stranded branches, but no help in transformation.

3’ exo of MP1.  One of the colonies.  One of the plasmids showed that no double-stranded DNA left.  But the single-stranded DNA is fairly homogeneous.  So large-scale tests, etc. on the following page.

[00:17:00] Continuing with the work.  So on page 83, Monday, January 14th, looking at various possibilities there.  Some gel that was complicated.  In this case it did look like the feelers had an effect.  Looks like a real increase, but the homology is now larger.  Tantalizing.  So yes possibly, no possibly.

So [00:18:00] continuing these different ideas on the following page.  5’ exo test again.  Thursday, February 7th.  With a conclusion that unfortunately 3’ ends never give any increase.  But the decrease is modest.  Retest 5’ end.  Quite possibly with histones on the next page.  Getting pretty desperate.

[00:19:00] These experiments are going on Thursday, February 19th with the histone.  And very little note on the bottom right-hand corner of page 95.  Friday, February 27th, 22/2/91 [00:20:00] NM and OS married.  That’s a story I might have told earlier that we got married at the local police station one Friday afternoon.  And people asked, “Where have you been?”  Oh, we just got married.  But anyway there it is on Friday, February 22nd.  Very happy occasion.  So that’s in 1991.  That’s 24 years ago.  Pretty good 24.  May it continue.

So we go on.  Testing nucleosomes.  [00:21:00] So this was the motivation of the experiments with the histones is a little bit unclear.  See if we can find out what was the thought.  [00:22:00] I think the idea must have been that maybe by making the DNA into nucleosomes it might be possible to increase recombination.  But it’s not clearly stated here.

Nucleosome assembly.  Starting with previously made MP8H digested with Bam and Sal.  So there it is alone and trying to assemble histones onto this DNA.  But [00:23:00] tested with standard procedure.  This is Monday, February 25th, page 103.  But no evidence of nucleosomes.  Need controls for the nucleosomes February 27th.  Trying to make some controls.  Some of the thinking is listed on page [00:24:00] 109 Friday, March 1st.  Some thoughts on what’s going on.  And attached to it a paper from Ramdas, Mythili, and Muniyappa in PNAS.  “Nucleosomes on linear duplex DNA allow homologous pairing but prevent strand exchange promoted by RecA.”  etc.  This was part of the thinking.  The preliminary look at the results previously tau 93 and 95 show that histones don’t help.  However, I don’t know that at this ratio of histone to DNA that the nucleosomes were made.  [00:25:00] I don’t know if the electrophoresis worked with this form of DNA, etc.  So many things that I didn’t know.

The comment that after writing these thoughts, interestingly, the paper that I referred to on the previous page turned out lately.  Testing whether nucleosomes can be obtained with DNA.  No problem in seeing that.  Micrococcal nuclease.  [00:26:00] Everything is now working.  Wednesday, March 6th, fresh attempt at nucleosome assembly.  Not very happy with the state of things page 119, Wednesday, March 6th.  Opposite the conclusion is trace of H4, histone four, experiment.  Trace at H4 but not enough for a good test.  [00:27:00] So wasn’t really getting nucleosome from the DNA that I had.

Looking at the electron microscope.  Page 125.  Not very nice-looking electron micrograph.  So continuing in the same vein.  [00:28:00] Page 129, Tuesday, March 12th.  Rather large number of thoughts about what might be done.  But something a little more practical on Monday, April 8th at the bottom of that is [00:29:00] CFTR.  That’s cystic fibrosis.  I forget what it’s called.  Cystic fibrosis TR.  CFTR.  The bottom of the page Monday, April 8th, page 129 talking about targeting the cystic fibrosis gene.  The cystic fibrosis transmembrane conductance regulator or whatever it’s called.  CFTR.  Which work was primarily undertaken by Bev Koller.  And trying to think of ways of inactivating the gene in order to replicate, to get a mouse model for cystic fibrosis.  Maybe this is a good time just to mention that that work [00:30:00] proved to be very competitive.  There were a lot of people trying to target the gene.  And yet none of us was really very confident about what was the best way to target the gene.  For example whether one should use what we called an O-type recombination which was a nick, a circular DNA cut once.  Or an omega-type recombination which would insert something rather.  A different sort of orientation of the plasmids.  So what we called O-type and omega-type recombination.  And Bev chose to use simple targeting and began to do these experiments, and just kept doing one after another.  Didn’t succeed, didn’t succeed.  But kept doing it step by step.  [00:31:00] Just repeating the experiment over and over again.  Whereas various other people who were working at the method kept changing their approach and sometimes trying this and sometimes trying that.  And it turned out in the end that Bev’s persistence worked very well.  And in fact her paper was the first to obtain gene targeting and an animal model of cystic fibrosis.  But the competitive atmosphere was very unpleasant.  And neither Bev nor I really enjoyed that area.  And in fact I deliberately decided I didn’t want to work anymore in a field where there was such cutthroat competition.  One incident at the meeting where [00:32:00] I went to talk about Bev’s work, and quite by accident a note was passed to me to give to somebody else.  Or I don’t know how I got the note.  I didn’t search for it or anything.  But it said in the note, “Shall we buy Smithies?”  Meaning give us a grant and buy us to work on this.  This was somebody in the Cystic Fibrosis Foundation.  I felt very unhappy about this whole business and was glad to be rid of it later, although we did publish a paper and the gene-targeted animal was used quite extensively later on by other people.  So here we are beginning that type of construction on page 131, Tuesday, April 9th.  [00:33:00] Cystic fibrosis polyNeo construction is being talked about.  And Bev has two new DNA samples, etc. and what can be done with them.  And I’m trying to help.  So the polymer is not very clear.  [00:34:00] Perhaps I can find out what that means.  And Tuesday, April 9th, page 131.  Looking into the construction of the plasmids which were going to be used by Bev.  There were basically two plasmids eventually made.  One called CF polyNeo because it was more than one neo gene copy in one of the plasmids.  [00:35:00] So two plasmids were eventually made with DNA that was primarily homologous to the region around exon 10 of the CFTR gene and interrupted by insertion into an Xba site within exon 10.  So this would cause a frameshift mutation and at the same time would introduce either one or two copies of the neo gene.  One pMC1neo, Mario Capecchi’s neo gene or pGKneo driven by the pGK promoter.  So it was one or two copies of the neo gene disrupting exon 10 and making a frameshift mutant.  [00:36:00] That was the construct that was being made.  And these are beginning experiments to help make the plasmids.  For example just making the pGKneo fragment is on page 133, Friday, April 12th.  To make the fragment used in making the construct.  And continuing to make these constructs over the next several pages.  [00:37:00] And preparative gel to make the fragments.  Ligating the pGKneo fragment into the plasmid on Wednesday, April 17th, page 147.  Something was wrong on that page.  Very odd.  The precipitation didn’t work.  Dilute more before precipitating maybe or the gel was bad.  With a comment that the marker looked odd.  [00:38:00] So ligating pGKneo into the plasmid.  And repeating again on the following pages.  Coming towards the end of the book.  Yes.  E. coli cells were purchased.  E. coli that were easily transformed.  They’re called SURE cells.  And the results obtained.  [00:39:00] Not very satisfactory at that point because on Monday, April 27th, page 157, the gel is run on colonies from the two experiments.  With a conclusion not very promising.  Most look like the vector.  Except possibly number 10 which is unstable.  Try again.  [00:40:00] So we go on and the book ends on page 163, Monday, April 29th with minis from transforming the SURE colonies with either DH5 alpha recipient cells or SURE cells.  With a comment that it’s pushing it a bit.  Some are pretty faint.  But continuing to work towards [00:41:00] getting the necessary plasmids for the experiment.  And that ends book tau.  [00:41:12]