Oliver Smithies:

[00:00:00] So we are now at April 9th, 1992, the beginning of book chi, and ending at August of the same year.  Starting up by reviewing the data on the silencer, there’s multiple copies.  And, checked on the candidates, and sequencing the inserts on Friday, April 10th, page 5, just working on the idea.  Some linkers needed, Monday April 13th, page 9, and thinking of enrichment for recombinogenic sequences. [00:01:00]

Some tests of LAR effects on April 14th, Tuesday, page 13, getting ready by cutting the sequences. [00:02:00]

Continuing the same work with LAR and [K-lock?], and so on.  Nothing very significant, just making these various vectors, or targeting constructs. [00:03:00]

Looking at page 33, Friday, April 24th, getting some sequencers, DNA sequencers these days, at last, to help decide whether anything is any good, looking at various things from book phi, phi123, 23, phi123.3, phi127.4, and 137.4 is exactly as predicted, but no insert.  123.3 is as (inaudible) as predicted, but the insert in junk, and 123.23 [00:04:00] is exactly as predicted, one copy of the inverted silencer, so 14.12, or another of this group can be used.  It tells us which to go on with the (inaudible), but box inverted.

And now on page 37, the review of the data obtained with delta hn, plus or minus locus activating region, and the ratio of non-homologous to homologous with the various constructs. [00:05:00] And the conclusion being that “chi13 is a better result.  Disregard the low (inaudible) KH20,” et cetera.  But, very little difference except that the LAR may enhance Neo a little.  No indication, no clear [00:06:00] indication of differential protection between on the [R?], so not very happy conclusion after all that work.  The number of colonies there, the number of [G48?] colonies, and the ratio of non-homologous to homologous, varying from 96 to 845, so roughly a 10-fold variation.  [Sonny’s?] tests of the same thing, not agreeing with my tests, so not very appealing. [00:07:00]

Retesting LAR for other ideas.  Some more ideas on LAR sequences, page 41, Thursday, May 14th.  Some thoughts on trying to see if synchronizing the cells at different stages of the cell cycle might [00:08:00] improve recombination on page 43.

Been to Japan between Monday, May 18th, page 51, and Sunday June 7th, page 53, back from Japan, having had a happy visit.  More sequencing beginning to be talked about. [00:09:00]

Cloning the silencer sequences into KH21 on page 73.


So Wednesday, June 17th, page 77, putting silencer into KH21, one of the many targeting vectors.  Continuing these same experiments, trying to dimerize the silencing sequence.

Page 87, Friday June [00:11:00] 19th, third, at least, attempt at dimerization.  Continuing with this type of experiment.  Still more minutes.

Tuesday June 23rd, OS, RWS 67, page 67.  Bud Rodger, of course, is (inaudible).  Diagnostics on KNH21 S1, [00:12:00] and continuing with this interpretation of the results on page 97.

Still attempting to double the silencer on page 99, Thursday, June 25.  “New attempts on doubling silencer.  Doubtful value.”  Looks as if I thought on Monday, June 29th, that I might have S2 going to S3.  Let’s see. [00:13:00] S3 to S4 on 105.  It’s the idea again of cell replicating slowly might recombine more easily, page 109 [00:14:00] and 108 to test the possibility that cells replicating more slowly or arrested might recombine homologously, more non-homologously, et cetera, but to no significant change with either treatment low serum and then normal, et cetera.  However, one experiment, the 12/12, looks better than usual, so do freshly-fed cells do better, because the ratios are definitely better than previously.  In fact, this was the beginning of, well not the beginning, but the continuation of a constant struggle to try [00:15:00] to find ways of increasing the frequency of homologous recombination, which in general never resulted in anything very striking that wasn’t already known.

And the idea with the next series of tests was cDNA, page 111, Thursday, July 9th.  “Put in poly-CG, or poly-TG, and test,” but this idea was canceled because of the next page.  “But if you’re going to do it, better have the DNA next to the homology but not within it.” [00:16:00] Page 113, so a modified plan.

First silencer tests, Friday, the 10th.  Try sequencing it with and without digestion.  Recheck of the sequences, page 117.  “The products are good, concentrations are close,” et cetera.  And there on the opposite page is the ratio of homologous to non-homologous, or non-homologous to homologous, 190, 151, and [00:17:00] 147.  Conclusion: “No significant effect of the silencer sequence.”  Hence, the actual sequences being now confirmed on page 118.  Current sequences compiled, and new sequencing primer designed, but S1/S2/S4/S4 were obtained, so there was S4, S3, S2, S1, and there wasn’t any doubt that we had four copies of the silencer, and we had one copy of the silencer, and that they were correct, but they didn’t work. [00:18:00]

And page 121, Monday July 13th, reviewing the possibilities with the thought on delta-HN that in all the other work in book τ, the most interesting and most intriguing was simple nicking, when there was a decrease in targeting, but a much slower rate than decreasing in visible construct bound, et cetera, et cetera, so various, and of (inaudible) thoughts. [00:19:00] With a comment on page 120 that it’s remarkable how long homologous recombination persists, but no significant improvement in the ratio, probably gets worse in looking at the treatment in different ways in this particular experiment, not very cheerful.  Many people must have spent many hours trying to improve the ratio, and many unpublished experiments, of which these are some. [00:20:00]

Repeat on page 125, Wednesday, July 15th, some old experiments with the tails having the 3’ tails, decided to try that again for the ratios, and the results are given, and there’s no significant difference between having the tails and not.  Conclusion: “The results are quite remarkable.  Zero effect of quite extensive degradation on either the non-homologous recombination or homologous, plus or minus single-stranded [00:21:00] binding protein has no effect,” in pencil.  Long series of experiments leading to no positive result.

3’ ends, plus single-stranded binding protein being tested on Wednesday, July 22nd, page 137.  The conclusion being that, [00:22:00] “Confirms chi125 results that extensive single-stranded regions with 3’ ends does not affect homologous recombination appreciably.  There may be a decrease in non-homologous recombination, plus or minus single-stranded binding protein has no effect.  Too many times that result.”

Test of 5’ tails again, not quite giving up with a 3’-5’ [00:23:00] [exo?] on page 147, Tuesday, August 4th, and the conclusions are, as usual, 5’ single-stranded region is very poor, because the number of homologous recombinants decrease very extensively from no digestion around about GH homologous over the ratio of non-homologous to homologous was around about 100 without digestion, and after digestion, it got up to 1,000, or even infinite, because there were no homologous recombinants.  [00:24:00] 5’ tail generated by 3’-5’ [exo?] was definitely not good.

More thoughts of 3’, still can’t give that end up because of the feeler ideas, page 149.

An old idea revisited on page 151, Wednesday August 5th, cut and exo, cut and exo.  And have the palindrome, with the idea that it will partly make of an old rolling circle, not really, but anyway, trying to have the idea that you could get a rolling circle. [00:25:00]

cDNA tests on page 155, Monday, August 10th, with a comment that since they’re very, very satisfactory, recombination frequency, where the ratio of homologous to non-homologous was now [00:26:00] around 4 instead of around 40, and with a comment, “Perhaps this is one of those days when the ratio is ten times better than usual.  Didn’t believe that it was real.”

Repeated this experiment again.  Not that experiment yet. [00:27:00]

The book ends with, on Wednesday, August 12th, with thinking about steps for mismatching.  That ends book chi.