Oliver Smithies:

[00:00:00] Beginning of book little s which starts March 22nd, 1978 and goes through July of the same year.  Beginning with looking again at Bam/HindIII digests of this plasmid gamma 2b parenthesis 11, control and Bam and Bam and HindIII digest.  Then thinking about the dideoxy method with a reference to December PNAS.  So this was Bam digest [00:01:00] again and Bam and HindIII digest of this clone.  The Bam/HindIII piece has poly(A), poly(T) tails.  Can use poly(A) or poly(T) feature for quick strand separation.  On an oligo(dT) gel or whatever.  But Bam/HindIII or even Bam alone looks good.  The gel is quite clear.

Trying an 8 molar urea TBE gel 5% acrylamide.  So a rather poor gel.  And it says the conclusion, [00:02:00] despite the poor electrophoresis there looks to be excellent strand separation.  So trying to separate the two strands by a gel containing 8 molar urea.  So instead of using those gels I began to try agarose denatured gels.  Using 0.7% agarose TBE and 1.5% agarose TBE gels with the denatured DNA, concluding that it’s uncertain.  Denaturation is uncertain.  [00:03:00] And so retest with the standard gel.  Trying to get an easier gel to use.  So going back to acrylamide gels again.  Not very satisfactory.  So repeat the digest, etc., etc.  Always gels that were important in the work that I did.

So Wednesday, March 29th, conclusion is that the loading problem is now OK but there is no denaturation.  So go to formamide, etc.

[00:04:00] Continues trying that separation on March 30th.  pH greater than 12 gel on March 30th, Thursday, with sodium hydroxide.  Migration is very odd from this gel on page 16, March 3rd.  The [front?] about two-thirds down, etc., etc.  So the conclusion is try formamide, exclamation mark, with a checkmark.  Cannot use high pH because of deamidation.  Saturday, April 1st, so doing some alkaline gels.  [00:05:00] Possible.  Try increased ionic strength, etc.  Doubtful if the pH of these gels was as high as expected.

Gel pH nearer 10.  So more high pH gels.  Analyzed on page 22.  pH 12 at this salt concentration does not denature the DNA.  So how did the old denaturing gels work?  Possibly higher EDTA in the sample.  [00:06:00] And [Sarah?] commented the Kpn/Bam/Bam does become clean.  The first gel may have denatured.  See later, etc.  So struggling with denaturation.

Continuing on Wednesday, April 5th.  Still no or incomplete denaturation with 0.1 molar EDTA at pH greater than 12.  Lower ionic strength.  Page 29.  With a comment that the clear denaturation now, etc.  Marker is gone.  [00:07:00] Lower ionic strength.  Lower pH gels.  April 7th.  Page 31.  Continuing to try to find a way to get denaturation and strand separation.  More and more tests of the same type.  Trying phosphate trivalently negative and [00:08:00] ascorbic acid, two negative charges, at pHs 11.5 to 12.8.  Phosphate buffer and [EEKA?].  And the conclusion is that nothing works below pH 12.2.

So some banks data.  Southern blots.  Tuesday, April 11th.  [00:09:00] [Beta?] with human.  Various comments.  With normal and Lepore hemoglobin.  What the restriction map should look like.  EcoRI and [batpis?] and Bam, etc.  All of these with normal hemoglobin and with Lepore hemoglobin.  The different restriction maps that had been obtained by banks.  [00:10:00] So still trying to separate strands on Wednesday, April 12th.  Deionized agar on the same day.  pHs 11.6 to 12.0.  Temporarily abandoned this deionized DNA.  On and on with the same general business that has always been characteristic of the work.  Repeat and repeat and repeat until you find conditions that work.  Some conclusion about transferring DNA on page 52.  [00:11:00] And films of gel again on the following pages.  [Summary one?] comment that some of the probes have E. coli DNA contamination which gives a positive on lysed E. coli.  The solution.  Add cold E. coli to make sure that you don’t have any just general hybridization.  So [Anne Bleckel’s?] [00:12:00] [miniplates share an Hb4 reconstruction complex?].  Set of results on Tuesday, April 25th.  With films and outlines of gels.  [00:13:00] Presumably some solution to it.  The film is shown on the left.  Where the film was from is shown on the right.  Shaded area was in the developer that — it would have been [a?] complete development of the film.  The film stuck so that’s what the outline is about.  [00:14:00] [00:15:00] Tests on Monday, May 1st, 27, EcoRI test on the minis 25 tubes.  I’ll pause there a moment.  We already know what it is that is — [00:15:44]


[00:00:00] I’m going to pick up book little s and go back a little ways.  I want to pick up at page 63, Wednesday, April 26th.  This was preparing E. coli DNA so that we had plenty of cold E. coli DNA to use to block hybridization to E. coli in the various hybridizations and so this is eventually the material is called little s62 E. coli DNA 1.2 micrograms per microliter.  Doesn’t say how much volume there is there.  But it’s quite clear why this was made.  And then we go on the following page, Monday, May 1st, [Howard Goodman’s?] clones.  [00:01:00] A human chorionic [go his?] somatotropin, etc.  And human growth hormone and rat growth hormone.  The two human ones are very homologous.  And the rat one is different and would only cross-react weakly.  So he says, “Don’t use the rat growth hormone for human growth hormone work.”  And so he’s talking about that.

Going on then with the test of E. coli on the minis to make sure that it will suppress hybridization by adding cold DNA.  I think that that’s what it is.  Yeah.  So this is restriction enzymes.  Each have 25 tubes.  And added mini lysates from s61.  [00:02:00] These are mini lysates made from various clones.  So page 67.  They transferred as usual and hybridized to Jerry’s probes.  Hemoglobin probe.  HCS probe.  PRO probe.  RI.  Rat insulin.  Rat growth hormone.  PRO, I’ve forgotten what that is.  But not very critical.

[00:03:00] Five gels.  Not clear what that means.  So the following page, May 3rd, Wednesday, has a list of the probes used on Wednesday in someone else’s handwriting, 5/3/78, and in my handwriting on the opposite side larger-scale minis on [30-5A?].  [00:04:00] And on pages 70 and 71 there are the gels and the hybridization autoradiographs.  But on the HCS [only positive is a little?] on the control hemoglobin.  But the control plasmid did not hybridize until longer exposure.  But overnight exposure the control is strong.  [00:05:00] And [30-5A?] is clear with the RI piece about 9.2% of lambda length and still looks the best for the hemoglobin.  So [30.5?] is looking pretty good for hemoglobin clone.

Continuing with the work on page 72 and 73.  RI digest fairly strong positive on [28/2ABC?].  With PRO being completely negative.  [00:06:00] Page 75 the hemoglobin control exceptionally strong and so on.  Continuing to try to find the probes.

On page 76 HCS virtually negative and PRO essentially indistinguishable from Hb except Hb control is weak.  Perhaps need to check what PRO is.  It’s not very important.  We’ll find out what it is later on.  Page 78, 79.  [00:07:00] [RG?] completely negative.  Looking at [hepa one?] digest of eight minis on page 81.  Transferred as usual but didn’t [rack and bake?], etc., etc.  Looking for positive clones.  Going back to [30-5A?].  Test on [30-5A?].  The size of the RI insert is 4.5 kilobases.  It doesn’t have a [hepa one?] site in it.  But [00:08:00] [later fill late fill late as?] mouse is 7 kilobase pairs.  Has two [hepa?] sites.  Further confirmation of the difference there.  So further tests on [30-5A?] on the next page.  Set of digests with HindIII, [hepa?], EcoRI, [bagel two?], Xho, and rather nice and simple autoradiographs on the opposite page.

[00:09:00] Then translation of the probes on page 87.  Elution of the probes.  s87 [Hb4?] 2.2 times 10 to the 6 or 22 times 10 to the 6 total counts.  And s87 [Hb117?] total of 21 times 10 to the 6 counts.

[00:10:00] On Wednesday, May 10th, page 89, test of alpha versus beta.  Using EcoRI and HindIII and RI alone.  Rather nice gel.  Conclusion.  Very good hybridization for the lambda-specific and clone-specific sequences.  [00:11:00] No E. coli background.  Looks to be an alpha clone as [Hb117?] is much greater than [Hb4?].  Try a second hybridization with a lambda DNA competitor.  Making a large amount of [30-5A?] on page 91.  And competing out the lambda and E. coli sequences from the nick translational minis on page 93 Saturday.  However, the background was very high.  So therefore cut out a gel band and nick translation from the cut band.  [00:12:00] So on page 95 Monday, May 15th, and on the preceding page, 94, evidently rather pleased with this because it’s nicely mounted.  There’s a nick translated [50-5A?] probe with 117 delta [lack four?] alpha and beta and 117 delta [lack four?] alpha and beta.  [00:13:00] So this is a test with s91 [3-5A?] DNA.  And lambda [DGR one?] EcoRI digest DNA.  With a remark on the preceding page that Jerry Slightom continued this experiment and nick translated a fifth of the recovered RI piece, about 2 times 10 to the 6 count per mini.  [00:14:00] Hybridized it to [Hb4?] and [Hb117?] and it gave the autoradiographs that are shown on the left page there.  Those are evidently Jerry Slightom’s experiments.

Continuing in the same vein Tuesday, May 16th, page 97.  Human hemoglobin candidates.  Mini-lysates [00:15:00] using materials which are labeled for a slide.  The conclusion is that [24 exposure?] is too little on this experiment but [nothing?] equivalent to the control.  Doubtful if anything.  Still trying to be sure what’s going on.  More minis on page 99.  [00:16:00] With 36, 23, 36, 24, 26, 27, 28, 29, 30, up to 36.  No likely candidates but repeat with more DNA and shorter run time.  Try human cDNA on the repeat.  Looking for more colonies.

[00:17:00] A little bit worried about hybridizing to the tail.  May 31st, Wednesday, page 101.  Suppose that the odd clones are rich in A, T.  Then they would all hybridize to poly(T) or poly(A) tailed cDNA.  All the probes would cross-react and would hybridize to [Hb4?].  Only [Hb30-5A?] would be clean.  So test with a poly(A) poly(T) competitor.  Or increase the stringency.

And understanding the problem and getting more stringent condition.  [00:18:00] Showing what I expected with the hybridization on page 103.  That [Hb117?] with the poly(A) poly(T) tail should give strong hybridization under nonstringent conditions.  But stringent conditions with nick translated plasmid as the probe, etc.  Various results to be expected.  Continue with this general experiment on page 105 and 104, June 3rd.

[00:19:00] And the results of these various experiments on page 107.  And the comment OS and LKS divorced.  Fourteen-hour autoradiograph shows hemoglobin cDNA nonstringent, etc., and so on.  Provisional conclusion hybridization is surprisingly dependent on the concentration of salt.  But nonspecific hybridization is quite easy to detect under nonstringent conditions and so on.  So understanding.

[00:20:00] [Did one?] 20-hour autoradiograph shows no sign of improved specificity so not quite as good as I had hoped.  Continuing on June 5th more hybridization tests.  And poly(dA) and poly(dT) were available.  [00:21:00] More hybridization thoughts.  June 6th.  And mini-lysates of human [MP36?] now plaque-purified on page 113.  So this is about 20 more clones, [36-1-2 aka 12A, 14A?], etc., etc.  All have been plated four times.  And the last [pick?] was from a positive hybridizing isolated plaque.  [00:22:00] But the conclusion is the results show that little hybridization was obtained.  With interpretation of the results on Friday, June 9th.  That none of the nonspecifics show.  Only the hemoglobins are still working.  Trying some experiments with S1 nuclease [00:23:00] hybridization on Tuesday, June 13th.

And then reprint on page 121.  A reprint from Jeff Ross.  There’s a precursor of globin mRNA paper.  Nice paper.  Jeff Ross’s.  [00:24:00] With some protocols on the left.  Continuing the saga.  S1 results.  Page 123, June 14th.  [The 30-1C?] has something in it.  On June 16th repeat of June 8th experiment and of the June 5th experiment.  [00:25:00] More and more tests of the same general type.  With page 129 June 25th review of the hybridization data.  Looking at all of the salt concentration, temperature, etc. effects suggests that 12 times [S of C?] going to 3 times [S of C?] is not very effective with discriminating specific or nonspecific clones.  But hybridization is faster or better [00:26:00] in 6 times [S of C?] so testing the condition for hybridization.  With a conclusion that both an increase in temperature, a slight effect, or a decrease in salt, a strong effect, do increase specificity somewhat.  But at the serious expense of count with salt.  Not very promising.  Effect of poly(rA) with the conclusion on page 130 that it’s clear result with poly increases the specificity by a very good factor.

So when using probe that has poly(A) poly(dT) tails [00:27:00] it helps greatly to add the cold competitor.  S1 nuclease tests again on Wednesday, June 28th with a large array of tests set up on page 135.  Titrating poly(rA) on June 29th, 30 micrograms per ml looks near optimum.  Greatly decreases the background.  [00:28:00] Two films, one very smeary, the other quite clean, that are from two tests with and without poly(rA).  Though the details are not given.  Going on with the human candidates.  Latest human globin candidates July 3rd, 139 page.  [MP35, MP35 2A1, 1, 5A1, 1, 6A1, 1 8A3?].  Conclusion.  The gel suggests that all are the same size as [30-5A?].  [00:29:00] [2R?] film shows minuscule but recognizable hybrid [205?] with and without poly(A) again very similar.  So it looks like we got four [phages?] again that had the same length of DNA as [30-5A?].

And that is the conclusion of a gel on page 143 run on July 5th with RI Bam and RI HincII or HindIII.  And RI [bagel two?] and RI alone.  [00:30:00] And these are [30-5A?] compared with [2A1, 5A1, 6A1, and 18A3?].  With a conclusion that the stain shows all three are the same and the autoradiographs with exclamation marks contamination underlined.  Perhaps we were reisolating the same clone is what is being hinted at there.

But there seems to be little doubt about [30-5A?] and continuing work with HindIII digest of it on page 145.  And with a map now of the [piece?].  [00:31:00] Monday, July 10th, using EcoRI fragments as probes, [30-5A?] EcoRI fragment cut out of the gel.  Nick translated.  Did quite well for checking alpha versus beta, etc.  [The Sharon 3A?] hemoglobin 4 and 117.  The lambda background was too severe for use as a cloning probe.  So some ideas to remove unwanted material and prepare the digest today.  On the following page is the result showing that RI on pBR322 [was not quite complete?] RI, etc.  [00:32:00] So this was some problems in digesting DNA.  Not very relevant to the [rate of progress?].  Ending with the last page of this notebook, page 151.  Ligation of pBR322 and the RI fragment of [30-5A?].  So instead of trying to purify it I decided to clone it into pBR322.  The aim to ligate at a high DNA concentration and RI template.  [00:33:00] I don’t know what it is quite.  Check the product for [extent of size?], etc.  It’s just a way of getting the pBR322 clone of the EcoRI fragment from [30-5A?].  Trying to get the best Stockmayer concentrations so that the fragment will integrate rather than just circularize which we used to fuss about in those days.  That’s the end of book s.  [00:33:43]