Oliver Smithies:

[00:00:00] This is now Book ε, which starts on October 25th, 1982, and runs through April of 1983.  I think I’ll begin this book by jumping a number of pages and describing what we were trying-‑what we were trying to make.  It gets a little bit lost in the details of making different fragments.  So starting with Thursday, December 12th, page 87 — sorry — page 57, this is talking about the clone Cosos.  That’s a cosmid os, Cosmid [00:01:00] #17.  And gives a map of that cosmid.  This was the enormous cosmid that I wanted to make to test the ability to recombine using a big piece of DNA.  And the map is shown there.  And it starts, basically, at the epsilon gene — or part of the epsilon gene, which is on a Clafragment.  And the Clafragment is a large fragment, going all the way through about 20 kilobases further along the chromosome.  So the epsilon gene, the G gamma gene, the A gamma gene, and the pseudo-beta gene, all on this [00:02:00] Clafragment, which is shown quite clearly in this construct.  So the Cla‑Cla fragment is coming from the cosmid which we received from…  (break in audio) Yes, this large piece of DNA, the Cla-Cla fragment, or, hardly fragment – cosmid, was obtained from Frank Grosveld, in Dick Flavell’s lab — was the basis for this huge construct that we were making.  So on page 57 is the map, that helps — or let one understand the various fragments that were being made, which may be [00:03:00] confusing in looking at individual pages.  So this was now the cosmid, using the Neo gene as the selectable marker.  So although, in the original idea, in the δ book…  The idea was to use tk.  This selection is using the Neo gene, driven by an SV40 promoter, as can be seen there.  So confusion is helped by looking at page [00:04:00] 57.

So going back to the beginning of the book now, we are looking for various fragments and hybridizations, and nothing particularly remarkable.

Packaging tests — or packaging being continued, on page 7.  “The ligations are much improved.  Package and try again.”  So we’re learning the — or doing what later became a very long-term set of experiments, of taking DNA, packaging it into bacteriophages, and testing whether the phages contained [00:05:00] the recombinant fragment, with the red and the blue DNA in them, supF — well, the other way around.  The blue DNA is supF, and the red DNA being now the whole beta globin gene and not the delta-beta of the construct.

And on page 15, Wednesday, November 3rd…  And page 14 is perhaps one of the prettiest pages in my notebooks, trying to sort out the orientation and sites, by Sal and Kpn digests, of these two clones, 17 and 12.  And it’s [00:06:00] quite a difficult set of diagrams to decipher.  But if the reader is concerned, could try to reconcile this with the map that we talked of on page 57.  But looking at these maps, it said, “The required products — are — for 17 are 3, 6, and 38 KB long and the ones for 12 are also 3, 6, and 38…”  But by using Sal and Kpn, one can distinguish the orientation.  [00:07:00] (pause) Let’s stop it.  (break in audio) It’s an intellectual puzzle for the reader, to figure out which is which, (laughter) because I can’t anymore — but to spend the effort to do it.  But I’ve often used a picture of these two pages, 14 and 15 in Book ε as an example of the lengths to which we had to go in order to get the big piece of DNA called Cosos 17, for the macro tests, as it [00:08:00] were, that ‑‑ Eventually, as we’ll see, in the real world we used a much smaller plasmid — in the end.  But the experiments, nonetheless, began with Cosos 17.  And here we are, continuing to make it.

(laughs) A typical comment, on page 21, Tuesday, November 9th, “Repeated Kpn‑Sal, with Xba replacing Sal” and, at the bottom, “That was a mistake.  Xba cuts too many other places.  What have I got wrong?” etc., etc., but “Horrendous.  Anyway, likely the map is wrong.”

[00:09:00] And trying Sal again, on Wednesday, November 10th, with the conclusion that “Sal1 must be working.  The problem is rather that none of the 3, 6, 38 category are represented.  #19 is probably 6, 11, 30,” meaning different set of cuts.  I don’t recognize the other…  “May be too short.  Screen another 96.”

Minis again.  And then here, on page 25, Wednesday, November 10 — one, two, three, four — about 40 more minis [00:10:00] being looked at, and 40 or so, on page 29, where there are four or five gels all being imaged.  And reviewing the data, on Sunday, November 14th, pointing out what the desired products are, with a difference between Clone 12, and one orientation, and Clone 17, in a different orientation.  Screening of minis with a beta probe, to distinguish delta-beta [00:11:00] plus G gamma and A gamma from G gamma and A gamma without the delta-beta, in order to make sure I had that little piece of DNA that was important.

So on page 35 a test grid is set up.  And on Tuesday, November 16th, Kpn and Sal digests on some of the selected minis, and with the conclusion that — “Looks like I got it.  Seven months to make,” with, “Cla 17 has the right fragments,” to include this huge piece of DNA.  [00:12:00] (pause) Looking again at 12, on page 39.  Plenty of colonies.  Looking at Cla-Cla 12.  On Thursday, November 18th, again.

Friday, December 3rd, page 45, getting back to Cla-Cla 17 [00:13:00] or what I began to call Cosos 17 — in a square box with arrows.  “Testing #124, ε‑25 Cla-Cla 17 candidate.”  Natalie grew up the plasmid.  And had an upper and a lower band, preparing the DNA.  (pause) With a conclusion that “Both upper and lower bands are equivalent but the more E. coli in the upper than [00:14:00] in the lower.”  So presumably continuing with the lower.

So back to the plasmid-by-plasmid recombination experiments and the like, sent to Raju Kucherlapati, Friday, December 3rd, page 47.  A whole bunch of plasmids were sent to Raju, pSV2, delta left Neo, delta right Neo, delta left Neo-Gpt, delta right Neo‑Gpt, etc., delta-beta 1‑2, delta-beta 1‑7, and Cosos 17.  We sent him Cosos 17.  Ten micrograms per ml.  So in other words now, the material that we’d made, including the [00:15:00] important material, at this point, which is Sample H, 2 milliliters, the whole of the sample for the massive construct, which we were hoping to use.  So to be specific, “Cosos 17, ε‑45, #124, upper,” etc.  And this is the Cla-Cla 17.  Something in return came from Raju.  “Including the two DNA samples that you wanted.  One is pSV2 delta left Neo and the other is a modified version, pSV2 Neo delta right,” etc.  But that’s a different topic.  The main thing is that we sent him the material which was going to be used in testing for recombination.  So the scheme was that Raju would use the Cosos 17 and transform bladder carcinoma cells, the same type of cell that had been used by Michel Goldfarb, and with a DNA transformation by calcium phosphate precipitation, select for the Neo resistance to G‑folinate resistance to Neo, [00:17:00] and then isolate the DNA and send it back to me, to look for the recombinant fragment with the phage assay.  So this is now in Raju’s hands.  And I don’t remember what I was doing immediately thereafter.  But we’ll find out.

Here is a construction of a tk plasmid, being continued, tk epsilon through pseudo-beta plasmid — using the Cla digest of tk [00:18:00] pBR322, from Book v, page 69, going back a long way.  So a ligation for tk epsilon through pseudo-beta or tk Cla 33, Monday, December 6, page 51.

And now we come again to that page 57 and further checks on Cosos 17.  Preliminary map, etc.  “Is about as good for checking as anything,” HpaI.  [00:19:00] And with a “Conclusion:  The first probe suggests that this is really Cosos 12.  But –” etc.  However, in the end it was Cosos 17.

A change in topic here, on Friday, December 10th, page 61.  “Expression vectors for PPP.”  That’s parotid proline-rich proteins, which Nobuyo had been working on and Ed Azen — together had worked on these.  Two rat cDNA clones were available.  And etc.  And beginning [00:20:00] to get expression vectors for them.  So a different topic.

But back to Cosos 17 again, on page 65, Sunday, December 12th, taking the rest of ε‑57 Cosos 17, Hpa digest, and trying to get the right material.  (break in audio) And so here we are, on Sunday, December 12th, back working with Cosos 17, and with, this time, a conclusion, “It looks as predicted for Cosos.”  And so on.  Clone 3 DNA.  [00:21:00] Has another sample of a Clone 3 DNA.  It gives a Cla pattern without the extra Cla band, etc.  And so a little sidetracking.  (pause) More Cosos checks, Wednesday, December 15th, with the now familiar map of the Cosos 17.  And what fragments should be there.  And they are.  “As expected,” the fragments are correct.  [00:22:00] (pause) Cosos being continued, page 79, Thursday, December 16th.  (pause) Making a fragment for nick translation.

Minis for orientation, on Saturday, December 17th, page 81.  [00:23:00] Nothing very remarkable.  A whole bunch of minis.  (pause) Continues, page after page, of similar types.  Wednesday, December 22nd, page 89, more mini tests.  [P‑33?] fragment, from ε‑85 gel.  So following, apparently, in the same orientation, etc., [00:24:00] etc.

(break in audio)

So these tests are going on, two sets of tests going on at the same time, and further tests of Cosos and the making of the expression vectors for the parotid proline-rich protein.  [00:25:00] So 8 and 33 are two candidates for the PPP material.  And this is what’s being considered on page 81, Saturday, December 17, looking at the orientation of Plasmid 8, upper and lower, and Plasmid 33 — or — yes, Plasmid 33.  And Plasmid 8 is…

Continuing, on page 83, as mentioned already, I believe, mini tests for orientation of these proline-rich protein fragments.  On page 85, probably [00:26:00] 30 or 40 plasmids being looked at.  On page 89, Wednesday, December 22nd, continuing to look for the orientation of the minis and their maps.  So F33 is secure, according to the plan on Thursday, December 23rd, page 95.  And continuing with [00:27:00] F8L‑24.  Further mini tests of F8 — of 91 mini.  Better control of temperature, on page 103, December 24, having made the remark on the preceding page that the wrong temperature had been used.

[00:28:00] So reverting back to Cosos work, Friday, December 24th, New Year’s — I mean, Christmas Eve, page 107, nick translation for the epsilon probe for Cosos, labeling and cutting out a fragment, which becomes epsilon IVS — intervening sequence — nick translated, ε‑107.  Quite a nice count.  “1.4×105 cpm/ml.”  And 1.5ml of material.  And used it as a [00:29:00] probe.  And everything checks out OK.  That “The 19 KB Hpa fragment hybridizes to beta, also hybridizes to epsilon Sau, which shorts it slightly,” etc., “This is correct.  Everything checks out OK.”  So, happy with the Cosos.

Continuing again, on page 109, Sunday, December 26, “Test for Cosos 12 vs. Cosos 17.”  A still remaining problem is the orientation of one part of it.  [00:30:00] Work of the same type, in the next few pages.  For example, again, page 115, Tuesday, December 28, “Recheck of Cla 12 and Cla 17.”  Never seem to give up.  And on pages 117 and 116, good maps being drawn of the fragments and attached add-ons, with the maps of 12 and the map of 17, and the Neo gene and Coso‑‑ etc., and [00:31:00] with a circled comment, “All is well.”

And again, the map that is so familiar, of the Cla region of the whole of the beta globin sequences, from Dick Flavell and Frank Grosveld — map that is reproduced many times in these notebooks.

Now, then, beginning of progress, in a way.  Because — simplified materials, [00:32:00] starting with Cosos 17, deciding to make a simpler construct, that could be used instead of the whole sequence.  These are Delta-Beta 17 and Delta-Beta 12, which are now taken from Cosos 17.  But only part of the sequence is used, a very much reduced size.  (pause)

Let’s see what was the thought at that time.  But this is all in relation to [00:33:00] page 117, Wednesday, December 29.  “Final checks on Cosos 17.”  The conclusion is that “Delta 93 maps of Cla 2 and 17 are correct.  The linker disagrees with the orientation of Debbie’s Delta 45 map.  And therefore, the map of Cosos 17 is as follows.”  And that’s the insert, where the “All is well” is commented on.  But there isn’t an immediate realization that Delta-Beta 17 and Delta-Beta 12 are going to be used in place of Cosos.  This is just a description of [00:34:00] the parent, as it were, of Cosos 17, which has a huge Cla1-Cla1 fragment in it.  Underneath that, it’s still a sequence that might be useful.  Because it has the suppressor gene next to a fragment of the globin gene.  But we shall see what happens.

(break in audio)

So delta-beta plasmids still contained a cos site.  So they could be packaged [00:35:00] into bacteriophages or could be used later on for that sort of purpose.  And we’ll go on and see how this was exploited.

More ampicillin ideas, on Saturday, January 1st, and Sunday, January 2nd.  But t‑ and beta-lactam.  But not pursued, I think, in further work.  Although, as usual, a couple of experiments.

And here is beginning of what I commented on with regard to the Cold Spring Harbor plasmid.  [00:36:00] And that Cold Spring Harbor paper — a recombination test plasmid — and dangerous, actually.  And this is Wednesday, January 5th, page 123.  Raju Kucherla‑‑ Raju Kucherlapati suggested we make an SV40 RA vector, and the whole of the beta globin, to test recombination at the plasmid level in cos cells.  So the beginning of making a test plasmid, which would have a large piece of DNA in it that we could test whether it would recombine.  Plasmid experiments.  So began to make this type of plasmid.  [00:37:00] (pause) Continuing, on page 127, ligation.

And transformation, the following page, 129, with on the left a little box saying, “Diamond altitude.”  That was after January 11th.  Because it was February 27th, 1983.  Fourteen thousand mean sea level, the low point in a glider, and 30,600 mean sea level, high point, with a gain of 16,600 feet.  It was a very exciting [00:38:00] week, near Denver, waiting for the wave, as it’s called, which develops over high mountains or when a high mountain or high plateau gives way to a lower one.  So, near Denver, one can find that sort of condition.  And I went with a friend to a glider port near Denver and waited for a week for getting the right strong winds and was able to achieve the distinction of getting higher than 30,000 feet in a glider, of course, [00:39:00] with oxygen — but also cold.  I froze my heels in the experiment.  But is a good experiment.

Back to work again.  Still haven’t yet gone, I guess.  The “Diamond altitude” is for February but I’m still working on Thursday, January 27th.  Candidates for proline-rich protein.  Tests by Jim L. and Ed Azen.

[00:40:00] Further tests of the beta Xba fragment, on page 135, Thursday, February 3rd, with a big circle, “pSV beta.”  “Conclusion:  Worked fine this time.  Streak out #6, a single colony for growth,” of this type.  (pause)

[00:41:00] Trying to get different plasmids for use in these experiments.  And received from Jürgen Brosius, who had been part of a paper with Talmadge and Gilbert, back in 1980, BroTS fragments, of which there were a number.  So cDNA expression vectors, a map by Brosius, unpublished, on page 138, with my calculation and measurement that it was 4.758 [00:42:00] base pairs long.  (long pause)

A check of minis of the — of the GF gamma series.  Twenty different minis there.  “Results show that all are OK at this stage.”  So cloning from the minis, on [00:43:00] page 143.  They were rather degraded but well worth pursuing.

So F33 cloning is taken up again, on Tuesday, March 29th, page 145.  Competent cells being looked for.  DH20.  Without competent cells, one can’t do transformations.  A new lot of DH cells being tested, Thursday, March 31st, page 149.  [00:44:00] Still looking at orientation tests, on Sunday, April 3rd, page 151.  And ending the book on the following two pages, so with more orientation tests and more mini tests, which are in reverse orientation.  Ending Book ε, on page 155, Tuesday, April 5th.  [00:44:38]