Oliver Smithies:

[00:00:00] So this is the beginning of book phi.  Starts September 26th, 1991 and goes through April 7th of 1992.  The beginning of this book is continuing the work on trying to clone HPTO HPRTNeo tel the telomeric sequence as diagrammed on page three, Thursday, September 26th, the beginning of trying to make the fragment with the [00:01:00] telomeric sequence.  Sal Kpn fragment.  And trying to ligate it into the correct form on Friday, September 27th.  Page five.  With a figure there showing the two telomeric sequences pointing in opposite directions with a Sal insert of an HPRTNeo gene.  So it would be tel HPRTNeo tel.  And the desired SalI single is plenty represented.  And S tel K is good and [00:02:00] with a comment go, go.  And here we are transforming it on Friday, September 27th, page seven, tel HPRTNeo tel.  All are either correct or vector.  So test to see which is the answer.  And Monday, September 30th, page nine the conclusion.  Looking very good.

All checks are confirmed except check four.  There were four tests.  Diagnostic tests listed there.  [00:03:00] And some more checks on page 11, Friday, October 4th.  Somewhat concerned that it’s not as long as I expected.  Conclusion is the suspicions were correct.  Go back to the start.  Do not pass go.  Do not collect $200.  Usual business.  Start again.  Steps to repeat on the following page, Friday, October 4th, page 13, reclone, etc. with several steps needed to get to the product which is clearly diagrammed at the bottom of the page.  So the different steps.  Step one on page 15 is OK.  [00:04:00] Prep gel again for the elements needed.  Results on page 57.  Quite some ways away.  Turning to that later.

Some naming problems and steps again needed to produce the result.  [00:05:00] Transformants being sought on page 25, Monday, October 14th.  Looking for two particular plasmids.  On the same type of experiment.  Rescue HPRTNeo tel on Wednesday, October 16th  page 31.  [00:06:00] Two of the plasmids being looked at more carefully.  Both look fine, page 33.  So the different steps again.  So back to preparative gel once more.  Friday, October 18th, page 37.  Sal tel Bam Kpn piece being made.  Sal tel Kpn.  Repeated the following day on following page, 39, Saturday, October 19th.  Transformation on page 41 with the usual sort of conclusion.  Either all the large ones are too small or they are [00:07:00] very much overloaded.  Former more likely, etc.  Test again.  Check on HNT versus HNT squared.  But as I feared both number 22 and number 25 are HNT, not HNT squared.  So religating page 45.  Couple of them are looking good on page 46.  Checking on the candidates on page 49, Wednesday, October 21st.  Looks good this time.  Size is [00:08:00] increased over HNT.  But check that the two tels are on either side.  So check on it and that is going on on page 51.

Page 53.  Thursday October 24th.  Picking various minis, of which one was number 33.  And the conclusion:  Number 33 is OK.  With a parenthesis there.  Dudwell Lane – because I lived as a teenager — well, from 11 until I went to college — at 33 Dudwell Lane in Halifax.  So I like the number 33.  [00:09:00] So number 33 you see.  And then there is a very significant change occurs between page 53 and page 55.  Complete change in type of work, which turned out to be a very wise choice.  The page is Tuesday, October 29th.  And it talks about ACE, angiotensin-converting enzyme.  I knew something about it because in order to get my pilot’s license I had to control my blood pressure and one of the agents I took was an ACE inhibitor, angiotensin-converting enzyme inhibitor.  So I knew a little bit about ACE but not much.  [00:10:00] I had a visit from a prospective post MD as I put it in — I thought his name was John Klege but in fact it was John Krege, K-R-E-G-E.  And he came to talk to me about the possibility of working in my lab as a visiting scholar for a little while.  And I, always a bit skeptical about that, sort of teased him about well, I know about you guys come spend half time the summer learning molecular biology.  And I always say, “Well, I’ll spend half time next summer and you can teach me surgery.”  And he got the idea.  He said, “No.  I’d like to work for quite a long time.”  And so I said, “Oh.  Well, what are you interested in?”  [00:11:00] He said, “Well, I’ve been interested in angiotensin-converting enzyme.”  Which I didn’t really know much about as I said.  But a little.  No, that’s incorrect.  He didn’t say that at first.  He said, “I’ve been interested in working on preeclampsia.”  And I didn’t even know what preeclampsia was.  I had to ask him, “What is it?”  Which is really quite significant as now at the present time it’s a very important part of my current research in 2015.  Preeclampsia.  But at the time I didn’t know what it was.  It’s hypertension and proteinuria that occurs in association with pregnancy.  And anyway then I said, “Well, I’ve just been reading a paper about angiotensin-converting enzyme.  [00:12:00] There’s a paper been published that people have been trying to understand the control of blood pressure and mapping genes and had mapped a gene, angiotensin-converting enzyme in rats as a possible gene that affects blood pressure.”  So I said, “Let’s have a look at it and see what we can see.”  And there were a couple papers in Nature and in Cell October 1991 describing this possible linkage of ACE gene — it was rats.  I should look that up.  [00:13:00] And it turned out that I had just read a couple of papers or a couple of papers had just come out and I had read them I suppose.  One by Jacob in Eric Lander’s lab.  And the other by Hilbert in George Lathrop’s lab in which they had genetic mapping of a gene causing hypertension in the stroke-prone spontaneously hypertensive rat, the first paper, and the other one was chromosomal mapping of two genetic loci associated with blood pressure regulation in hereditary hypertensive rats.  So one in Cell and one in Nature.  And they had mapped the gene ACE.  A mapping [00:14:00] which in the end turned out wrong.  But it didn’t matter at the time.  This was stimulating.  And so I said to John Krege, “Well, why don’t we think about knocking out the enzyme and see what will happen?”  And I read enough about it to understand that the gene had two domains and two catalytic sites, one of which was particularly active in the testis and one in somatic tissues or something of that general type.  And so I thought about attempting to knock out the upstream domain, both domains, and only the testis exon.  [00:15:00] And that seemed like an interesting experiment to do.  And later try altering the gene to mimic the rat that is dominantly hypertensive as a result of a hypothesized dominant mutation in the ACE gene, etc.  And so had put down some thoughts about targeting the exon 12 and 12T which is the testis-specific exon  and mapped.  And this is the beginning of working on the genes controlling hypertension.  Very important change in my work because that later became the focus of a lot of work.  All stimulated by John Krege.  But those experiments are not yet in progress.  Here we are on page 55 [00:16:00] and it says wait until page 117 before those experiments are considered again.  But not yet.  So this is continuing on page 57, Wednesday, October 30th.  Continuing to try to get delta tel HPRTNeo delta tel.,  etc.  Delta being a little indication of slight variation in the telomere sequence.  But nonetheless on page 58 here is a possible scheme [00:17:00] for the ACE gene experiments.  Schematic of the gene using just letters of the alphabet.  Target one, etc.  Cut and remove.  So on.  Very detailed idea with no real sequences.  Just imaginary A, B, C, D, E, F, G sequences.  For example take the exon, the sequenced exon 12, exon 13 region and design whichever is more convenient to inactivate the gene.  [00:18:00] Using a pPNT as a positive control and as a vector pGKneo poly(A) addition sequence pGK-TK gene, poly(A) pPNT as then pGK driving neo and then the thymidine kinase gene which was going to be used in some experiment.  And with a comment that [00:19:00] getting the genomic Kpn fragment that has exon 12T would allow all but the long ends to be made of the — in order to inactivate both ACE functions.  So this genotype.  [00:20:00] Some new plasmid vectors thought about.  But here we go.  Wednesday, November 13th.  HNT number 33 phi 63, phi 52 tel HPRTNeo tel.  Not as good as the previous delta tel material.  And the results are talked about on page 75.  Looking to get that HPRTNeo with a tel sequence.  Now also evidently been reading more about the ACE gene.  In fact I think I made myself a diagram of the whole system [00:21:00] that was later very helpful in thinking about things.  But I don’t have a note of it here.  I remember making that diagram that proved very helpful to understand the whole system.  So I began to understand more about the control of blood pressure.  And here on Wednesday, November 13th, page 65,  I’m talking about the angiotensin II receptor.  About three papers describing the identification of type 1 angiotensin II receptor in rats and cattle, etc.  And references.  And also thinking about the atrial natriuretic [00:22:00] peptide.  Published sequence of the human and mouse atrial natriuretic factor genes in Science 226, year presumably the year that we’re in which is 1991.  So now thinking about it enough on page 67.  ANF is coded by a single sequenced gene with three exons, etc.  Under details of the gene and beginning to think about it.

Maybe can [00:23:00] isolate the gene.  And this was work that would be done by my postdoc.  Would be done by Simon John.  I think he was the person involved in this.  We’ll see.  So the receptor also being described here.  Receptor A by Pandey and Singh.  And I later did a lot of collaboration with Kailash Pandey.  And as he once [00:24:00] said to me in describing his work with ANF receptor, he said, “ANF receptor A is my life.”  He dedicated all his scientific work to that receptor.  But anyway I’m now aware of this and trying to think about cloning them.  And four oligos have been ordered to make some probe here.  So this probe would be exon 1 probe.  And one pair of primers.  And the other would be an exon 3 3’ untranslated region sequence.  So these are pages 65 and 69 are both thinking about what sort of primers to use in order to isolate the genes.

[00:25:00] Little diversion on Monday, November 18th.  Having read about the use of electrofusion to make tetraploid embryos.  You could buy a rather complicated apparatus.  And I made an extremely simple one with some platinum wire probe.  And could with my PCR — not PCR —  with my electroporation machinery have a pulse of 100 microseconds at 1,000 volts per centimeter, etc.  Tested it with electroporator and varied the [00:26:00] voltages.  Tuesday, November 19th, page 71.  The voltages were varied from around about 300 to about 500 on dead embryos.  On live embryos 400, 425, 450 are all close.  Semiexplosive arcing does not necessarily kill but it’s messy.  So Sarah Bronson will continue this work.  How to make fused embryos that had various advantages in terms of use in embryonic stem cell work.

Back to trying to clone the ACE gene Wednesday, November 20th  [00:27:00] page 73.  Kim reports that the ACE probe didn’t give any strong candidate but appears to be cross-reacting.  So try again with a different exon.  And the best I could find was in exon 8, and made some primers that were ACE 8 left and ACE 8 right that could be used to look for that.  Still working on tel HPRTneo tel from phi 63.  This is phi 75.  Where comment was made that the results were on this page.  [00:28:00] So looking at the results.  The result being that this says once again and even more so an increase in non-homologous recombination by using the telomeres and a decrease in homologous.  So it looks as if this idea is dying a natural death.  With the telomeres the ratio of non-homologous to homologous was about [00:29:00] 59.4 to 1 in favor of non-homologous.  And without the telomere the ratio was 11.3.  So thus with the telomeres increases the ratio about fivefold.  And it says compare this with the delta telomere sequences when the ratio increased about 1.9 times.  So it looks as if this idea with telomeres is dying a natural death.

The following page, 76 and 77, have nothing on them.  But between them is a collection of rather long letters to [00:30:00] my collaborator Christine.  I don’t remember her surname, but Christine.  So I’m going to go over that letter because it’s talking about the telomeric experiment.  So “Dear Christine, the long silence on my part is not due to inactivity but to the long lag time of the experiment.  And I now say, ‘I have definite results to report.’”  Referring to a diagram.  And then there’s a large sheet there at the back with a diagram.  Looking for recombination between the deleted [00:31:00] chromosome that’s present in E14 remembering that that has the promoter exons 1 and 2 of the HPRT gene deleted but has exon 3 to the end of the gene OK and has upstream sequences OK.  So there’s an unknown deletion there.  And that we corrected the gene with a plasmid of MP8 is my recollection of its name.  Not important.  But the plasmid which in this letter I call neo HPRT.  This has the neo gene in it for selection and it has promoter exon 1 and exon 2 which are missing in the HPRT- cell and exon 3 as a crossing over sequence and then upstream sequence.  And [00:32:00] leading to a corrected gene by homologous recombination.  So that’s the diagram.  Straightforward diagram.  Then construct which is the same as neo HPRT except that it has a telomeric sequence as well which I’m going to call tel neo HPRT 5 in this diagram.  And then other sequences.  [00:33:00] So going back to the letter.  It talked about homologous recombination to correct the gene.  Any HPRT cells are necessarily also G418-resistant.  But you can also get G418 resistance without homologous recombination.  So the ratio of non-homologous recombination to homologous recombination is obtained by counting the number of HAT-resistant colonies and the number of G418-resistant colonies.  And the ratio is number of G418 [00:34:00] resistance minus number of HAT resistance divided by the number of HAT resistance and so forth.  And talking about making this correction in my construct.  Two experiments were done each in triplicate comparing Bam and Sal digests of delta tel neo HPRT delta tel and tel neo HPRT tel.  And the resulting colony numbers are given.  And the conclusion was that the ratio of non-homologous to homologous [00:35:00] with delta tel was 21.4 and the ratio without delta tel was 11.5.  In other words, adding the tel made things worse.

So homologous recombination rate was not significantly affected by delta tel but the non-homologous rate increased nearly twofold.  And I interpreted this to mean that the delta tel may have protected the neo gene from degradation but delta tel did not improve the ratio of homologous.  So we’re talking about the death of the tel idea.

So we have a pretty definite although unpublishable result.  The addition of tel does not decrease non-homologous recombination.  [00:36:00] Perhaps it would have been publishable but we didn’t try.  If you can see anything wrong with all of this please let me know.  If any of the constructs are useful, etc.

Ending up with all in all I feel the experiment was well designed and well executed and I’m satisfied.  I’m a little disappointed but not greatly so.  After all, when one tries way out experiments, as I often do, one gets used to having negative results.  Give my greetings to Peter and show him the results.  All the best.  Oliver and Denise.  And I will find out who Christine and Peter are.

Yes.  I don’t at the present time remember who the people were that I was talking to.  But it was Christine and Peter and Howard  [00:37:00] and I will in due course find out who they were and where they were from.

Now going back to the progression of the idea of working with hypertension.  A very important change again in my direction occurred on Wednesday, December 4th, page 79, when I was writing up a grant for the hypertension work.  In a little digression, that grant was turned down and rescued by an administrator in NIH.  I think I’ve told that story before but we’ll see.

Anyway this is [00:38:00] Wednesday, December 4th, page 79.  During the writing of a grant for hypertension the possibility that lack of ACE might not completely block the angiotensin II production arises.  Also a preference to be able to change the molecular nature of angiotensin II itself would be very nice, etc., etc.  So will also try thinking about preproangiotensinogen.  And there was the molecular cloning of the mouse angiotensinogen gene described in 1988 in Genomics by Clouston, Evans, Haralambidis, and Richards.  So beginning to think about angiotensinogen.  Not yet quite to the [00:39:00] idea that was later to be important in the work.  But here Thursday, January 30th, page 81.  The grant application was submitted to do the work and the long term objective was to help unravel the genetic complexities of essential hypertension.  And there were five specific aims.  One was to modify the angiotensinogen gene so that it makes no product.  Or that the portion coding for angiotensin II is altered.  The second was to [00:40:00] inactivate the gene coding for the type 1 receptor of angiotensin II.  The third was to modify the gene coding for pro-atrial natriuretic factor to make no product or to alter the ANF portion.  Fourth to study the phenotype generated by the mutations described above singly and in combination.  And the fifth was to study the responses of the mutant animals to manipulations that disturb homeostasis.  And eventually that grant got funded.

[00:41:00] Some correspondence February 31st, 1992 regarding apolipoproteins, and saying that my wife, Dr. Nobuyo Maeda, has a construct that efficiently targets the apolipoprotein A-I gene in ES cells.  We could use that, etc., etc.  [00:42:00] Still getting excited about different genes related to hypertension here.  Two new candidates being considered on page 85 Friday, February 7th.  The cyclooxygenase genes.  Thinking about inactivating that gene.  For example, knowing about COX1 and COX2.  Considered Tuesday, February 11th, page 87.  Primers with a view to modifying the structure of the enzyme.  Perhaps changing [00:43:00] serine 530 which is normal target of aspirin acetylation to alanine so that it can no longer be inhibited by aspirin.

So more thoughts on improving the strategy for homologous recombination by decreasing the nonhomologous portion.  After talking with Nobuyo decide to work on a new way of decreasing non-homologous recombination by having a silencer sequence [00:44:00] upstream of the pMC1neo gene so that only after recombination was the silencing sequence removed.  So after homologous recombination the silencer sequences would be locked, but non-homologous recombination the neo gene would be silenced based on an idea that had been published by Arizumi that negative transcriptional regulatory elements that function in embryonal carcinoma cells and negative silencer.  [00:45:00] Negative factor.  More thoughts on the silencer the next page.

And beginning to modify MP8 neo to insert these sorts of sequences.  New uses on page 99 for a sequence that I’m going to call K delta HN.  [00:46:00] Rough test of the idea.  Need to have some measure of non-homologous recombination versus homologous recombination with this short vector which was I going to call delta HN.  Thinking about the LAR sequence again.  Putting it into different thing not continued, [00:47:00] abandoned this.

Going back to page 91.  Thinking about using a very much cut down version of pMP8neo, the construct used for correcting the HPRT gene deletion so it would have now only 200 base pairs of homology [00:48:00] upstream and only 1 kilobase downstream of the neo gene and the promoter.  So that it’s going to be called — I don’t say yet what it’s going to be called.  But it’s a cut down version of the homologous recombination sequence with only 200 base pairs upstream and 1 kilobase downstream of homology.  Surrounding the neo gene, the promoter exons 1 and 2 missing sequences needed to correct the HPRT deletion.  [00:49:00] How to go about getting this is talked about on page 93.  Trying to get this very small targeting construct with the diagram on page 95, Wednesday, February 26th at the bottom.  Needed to make some modifications to change the sequence so that the sites were Kpn and HindIII [00:50:00] rather than HindIII HindIII.  Just a step needed.

Now then on page 99 I’m looking now at the sequence which I’m going to call K delta HN.  Meaning K for Kpn, delta H for a short version of HPRT, and neo.  So it’s the small construct with a very limited homology.  For example I could put into it the LAR sequence and so on.  Thinking about testing for homologous versus non-homologous recombination.  [00:51:00] Could wait for K delta HN but meanwhile testing linearized version delta HN.  Putting I think LAR sequences again.  Putting it into changing the linkers.  Modifying the construct so that I could control where the LAR or other sequence was put in in the following pages.  Not very important except as necessary steps towards the end.  [00:52:00] For example here is a sequence.  On Sunday, March 1st, page 115, thinking about the hypersensitive site two on the 1.9-kilobase fragment.  Putting that into one of these vectors.  Thinking about the orientation also of the sequence.  [00:53:00] And here we have the further ideas of the silencer fragment on page 121 Tuesday, March 3rd, cloning polymers of the silencer sequencing.

And tests of these on Sunday, March 8th.  But a mistake – should have used PvuI instead of Pvu and HindIII.  Redigesting the minis.  So it continues.  [00:54:00] Thinking some more about silencer sequences and about the locus-activating region strategy, messing around.

Just trying to find ways of improving the frequency of homologous recombination.  Test of silencers again.  [00:55:00] Expansion.  Page 145 Saturday, March 14th.  Complex result, etc.  Going back to the mini delta HPRT construct.  Page 151, Friday, April 3rd.  Construction of LAR delta HN.  So putting in the LAR sequence, locus-activating region to make K LARK Kpn.  [00:56:00] There’s a special image on page 150 recombinant fragment PCR assay on tail DNA from agouti mouse.  CFTR- heterozygote.  This is showing Bev Koller’s results that she had got heterozygotes for the knockout of the HPRT gene.  Showing it was a happy time to get that result.  [00:57:00] Even though we didn’t continue Bev’s PCR results on the first agouti offspring of CFTR targeted ES cells.  March 17th, 1992.  So a knockout of the cystic fibrosis gene, establishing that they were heterozygotes.

Faint hand but it can be seen.  That’s an addendum to the page which was continued the work on K LAR K LAR delta HN [00:58:00] and so on it goes.  Review of the experiment on page 157, April 6th, Monday.  Various problems.  End of the silencer experiment.  Consideration of that too.  So they go on.  Silencer experiment.  Expansion continued.  Page 161.  And finally the book ends with K LARK delta HN one or two page 163 [00:59:00] Wednesday, April 8th.  The number six, a monomer, which has a reverse orientation, produces LAR tail delta HN, etc.  Grow up number six.  And so ends book phi.  [00:59:29]