Oliver Smithies:

[00:00:00] Starting a book, E’, December 26th, 1995, and ending October 17th, the following year, so quite a long time in one book here.  Boxing Day, December 26, preparative gels, following upon the work that was going on.

PCR screen Tuesday January 2nd.  And, Saturday January 6, page 13, [00:01:00] a note that there was a big snow for North Carolina on that day.  But, came to work.  Nothing particularly remarkable, just slogging away at making various constructs.  But, again, a nice trip to Japan there, between January 16th, and the next entry is February 5th, and the note, “Back from six talks, two in the US and four in Japan.  And vacation with Nobuyo in Sendai, [00:02:00] her hometown, with her sisters.”

KH21 neo-dup on this page, duplicating the neo gene, the loxP site between the two copies. [00:03:00]  Different use of a different strain of bacteria, K802 recA-minus for E’30 mini, #8.  “Something is very wrong,” conclusion on page 36.

Restarting the ATR1a duplication, Wednesday, February 21st.  [00:04:00] With various plans, usual site changes, Bgl2 to Ask1.  Removing ATG in dup-del, the following thought on  [00:05:00] the previous page.

Going back to page 39, Thursday, February 15th, on K8O2-recA-minus transformation with E’, 31 ligation with a comment that the conclusion, “Still no good.” [00:06:00] The only stable insert is in #14, et cetera.  “Abandon this project.  That’s enough work on it for now.  Stop thinking about it.”  Changing Bgl2 to Ask1 in dup-del on page 43.  And unfortunately, can’t prevent getting an ATG start signal at the end of dup-del, because part of Ask1 when it’s next to the loxP site, so decided on Monday, February 26th, [00:07:00] to remake dup-del without any ATGs before continuing.  I didn’t want to have starts of translation somewhere that I didn’t understand.

And so then, we get to page 49, Wednesday, February 28th, “Removal of ATG in dup-del.”  So the two lox sequences, both have ATG in the anti-neo and anti-lox directions.  Can’t remove these.  The scheme seems to be to treat dup-del as a cloning vector, and excise RSR1 so finding ways around this [00:08:00] sort of problem, modifying dup-del on page 51, Thursday, February 29th.  Did a lot of work on dup-del, had various forms of it.  It was a very useful vector, and was modified many times.

Continuing to go on.  Here we are back to PLE7 again, on Tuesday, March 5th, page 59. [00:09:00]

Sunday March 10th, page 63, redesigning to avoid polA.  Decided that the experiment is too complex to risk AATAAA vs. AATAAA, poly-A addition site.  So that wasn’t a correct poly-A addition site.  So, a computer simulation, et cetera, replaced this and that and see what it looks like.  Take dup-del, that was to see if it would work, just doing the digests and imagining the pieces. [00:10:00]

So, as on page 65, dup-del with RSR, Bam cuts.  I remember that, doing those experiments, vaguely, with RSR1.  It was useful because it was the only site, only single site in dup-del, and a rather unusual site too. [00:11:00] That’s RSR2.

So more thoughts on the AGTR-1a duplication, with some added comments on page 70, Masaki’s clone, B’145, and et cetera, how to, various thoughts of what to make. [00:12:00] Having some sort of – for having a diagram, been reading about Feynman’s work, and Feynman diagrams, I wanted to make a Smithies diagram, something like the following will be needed, Sunday March 17th, page 73, to sort of think about how the genes were changing their expression, as I’ve tried earlier. [00:13:00] On page 72, there was a diagram of that sort that I was thinking of, starting with the copies of AGT, renin, AT1, ACE, angiotensin-2 receptor, angiotensin-2 receptor, all the steps, angiotensinogen being acted on by renin to give AT1 being acted on by ACE to give AT2, which combines with the receptor ATR1a bound receptor, and changing the blood pressure level.  And with a little note there that is hidden away, ACE, the bradykinin level is altered.  This is really quite an important note that we had come to the conclusion from thinking about Km and work of other people, that [00:14:00] changing the level of ACE would not change the level of angiotensin-2.  The logic was fairly straightforward, but we hadn’t realized it for quite a long time, and that is that if you have a steady state system with short half-life, as you have going from angiotensinogen cut with renin to give angiotensin-1, which is rapidly converted to angiotensin-2, and the short half-lives of angiotensin-1 and -2 are short, and the half-life, and therefore the concentrations are steady state concentrations, and I began to realize that if you increase the amount of — if, for example, you decrease the amount of ACE, [00:15:00] that what will happen is that angiotensin-1 level will rise, and if you’re not saturated in the Km, when angiotensin-1 rises, the conversion to angiotensin-2 is returned to normal, with the following sort of thought that one times ACE level leads to a certain amount of ACE-2, and a certain amount of angiotensin-2, and a certain amount of angiotensin-1.  If you have the ACE, then the angiotensin-1 level will rise, and not angiotensin-2.  It backs up, and when it gets to twice the level, so twice the level of angiotensin-1 times half the level of ACE-converting enzyme, is the same as 1×1, so ACE does not affect the concentration [00:16:00] of angiotensin-2, but it does affect the concentration of bradykinin, because bradykinin is a substrate of ACE, and angiotensin-2 is a product, so the product of ACE doesn’t change when the level of ACE changes, but the concentration of the substrate changes, so bradykinin will increase when ACE decreases, because bradykinin is broken down by ACE just the same way as angiotensin-1 is broken down to amino acids, clipped off by ACE, so here is the logic behind that little statement on page 72, that the bradykinin level will be altered by ACE. [00:17:00]

And then, beginning to try to think of ways of showing this sort of progressive system, as I called them, renin and angiotensin system path diagrams, one showing what I thought would happen with ACE-1-0, and what would happen with the angiotensin-1-0, on page 75, Sunday, March 24th.  So that when ACE went down to zero, or to one copy, the level of angiotensin-2 would not fall, but the level of angiotensin-1 would rise [00:18:00] and if it was angiotensinogen, everything would stay low.  Beginning of trying to get a logical way of looking at the whole system.  So, this rather poor description I made of the effect of ACE levels is written down more clearly on page 77, Sunday April 7th.  ACE 1 copy will show an increase in angiotensin-1 due to less usage by half ACE compared with decrease in angiotensin — angiotensinogen, in the angiotensinogen 1-0 copies [00:19:00] due to greater usage by increased renin, so that the level of angiotensinogen in the 1 copy, angiotensinogen animals, is less than 50%.  It’s 35%, because renin has gone up and more is being used by renin.  So, we need to use the concentration in terms of fraction of Kms, and think of binding constants to get a true pattern on page 77, Sunday, April 7th.

Thinking about renin release more on Thursday, April 25th, the next page 79, looking at [00:20:00] renin mRNA in ENOS minus/minus mice, Ed Shesely’s minus/minus mice, ENOS.

On June 3rd, there is a block there because — not a big one, April 25th.  Next entry is Monday, June 3rd with a little comment that hypertension, competitive renewal approximately complete.  So back to the lab again.  Some involvement of Randy Thresher in the lab, talking about making and testing [00:21:00] methylated DNA, et cetera, for increasing homologous recombination, so still thinking about increasing.  The test being delta-HPRT locus, usual correcting vector, and try to see if methylation would make any difference.  Not immediately clear what was happening, but several pages of thinking, page 83 and 82. [00:22:00]

Trying to modify the end of the construct to alter recombination, on page 87, [00:23:00] Wednesday, June 5th, the possibility of adding a base there in 6-hexylamine, and to try cross-linking.  But, getting into difficulties, and on page 89, the big letters, “ABANDON.  Hyung-Suk Kim has a better scheme.”  With a spacer. [00:24:00]

Still trying to put a tail on the end to add something to the 5’ end, or to the end of the construct, hexylamino, being the idea to put on that end, or malido side chain, again in the same position.  Go back and see what the rationale was.

Thursday, June 6th, page 91, more thoughts of adding hexylamino side chain to cut talgine construct.  The [00:25:00] logic of this is, I don’t any longer remember, and probably not worth trying to decipher, because these experiments were, in any case, all negative, but they were a continuation of the long dream of increasing homologous recombination, one way or another.  And so, we go on with that.  And we’ll leave the topic.

A rather more interesting thought on Thursday, June 13th, page 97, nuclear localizer peptide. “Decided to test whether a nuclear localizing peptide put in the right place might increase the frequency of homologous recombination by [00:26:00] directing the DNA to the nucleus,” a rather interesting thought.  I think it was negative again, but we’ll see.

Monday, June 17th, Nancy’s birthday, page 99.  Nuclear localization DNA into the EcoRI site is being attempted.  Another thought on page 101, Friday, June 21st, is Randy Thresher’s idea, can use the reverse reactions, thiol coupling, if EcoRI, SH is not accessible, to couple a restriction enzyme to the DNA, [00:27:00] is I think the motivation, although at the bottom of the page, Tuesday, July 30th, it says, “Decided to stop this work, at least temporarily, and return to the duplication of coding regions.”  And that is the thought on the next page, 103, Wednesday, July 31st, “Coding duplications.”  Possibility that seems worth exploring is to use a more sophisticated construct for allowing two sequences to both be translated, mRNA, start-AUG, stop translation, restart translation, and poly-A; in other words, an internal code that’s sequenced, and internal [00:28:00] ribosome entry site of the type described in a couple of references.  So internal ribosomal entry site was the thought now between two copies of gene.

So, here it is, that scheme drawn out more fully on Wednesday, August 7th, page 105.  AGTR1a duplication with the possibility of having an internal ribosomal entry site to allow better expression of the gene.  And this being pursued to make this construct on page 107, dup-del [00:29:00] trying to put in P-H-A-M-A-I-R-E-S, Internal Ribosome Entry Site.

Here is what became eventually a quite useful idea, a dup-card sequence, on Friday, August 9th, the idea being to be able to test whether, have two coding sequences to increase expression.  And looking at the IRES insert on page 111. Inserting [00:30:00] linkers, et cetera, removing sites, as we go through the pages.

Then, beginning of a new thought here which turned out to be very important.  So let’s just review the situation. Thursday, August 22nd, page 125, my page of the notebook is talking about a linker insertion, into the neo-IRES plasmid.  But attached to that page on the opposite page, 124, is an important letter, that was involved in my beginning to make models, [00:31:00] computer models of the system, which were very revealing.  So on August 22nd, I wrote a letter to Marshall Edgell, with my various ideas up to that point, that it starts off, “The following two got diagrammed somewhere in my view of the renin-angiotensin system, which was generally acknowledged to be one of the most important, perhaps the most important means of controlling blood pressure.”  And it talks about angiotensinogen, and the difficulties of modeling this system.   And, my path diagram idea [00:32:00] near the end of this letter to him, complex diagram of a sort that I talked about before, where we start with the level of angiotensinogen, and then the renin responds, and then what’s the angiotensin-1 level, what’s the ACE level, what’s the angiotensin-2 level, the receptor, ligated receptor to the angiotensin-2, and the effect of response, et cetera, and a diagram for angiotensinogen one copy, and saying that this is a flow or path diagram for a heterozygous plus/minus AGT gene disruption, and [00:33:00] what I’m asking Marshall for is his help in making, the letter doesn’t appear, oddly enough, to be complete.  Let me just see, yes.  It starts to say, “What’s happening?” but there is no conclusion to this letter.  But I know what the conclusion was.  I asked him for his help.  Yes, I don’t know why it’s been lost, the page has been lost here, unfortunately, and having the transcript —


It’s not on the back, is it?


No, it’s not on the back, Jenny.  Oh, yes it is.  On the back, the letter has, it’s double-sided, and the back page is there, [00:34:00] so it goes on about the asking really whether he can help.  Is there any way of modeling this sort of system?  I must have talked to him about it earlier.  The letter is not really quite explicit enough, but the idea is very straightforward.  I thought that I — Marshall could help me because of his computer experience, and modeling, he could help me develop a computer simulation of this system.  And, so that letter was written on August 22nd.  But I’m still experimenting on the following days. [00:35:00] As we see, several pages going on in this general way, August 28th, for example, page 133, removing an SK1-site from dup-del.

And rescue the SK-deletion on page 137.  And more diagrams of the OS-dup card on page 139, Sunday, September 22nd.  And continuing to experiment. [00:36:00] With the idea that on page 145, that a gene might be, what I’ve been calling a “virtual gene” to an actual gene.  For example, data from Curt Sigmund’s lab shows that renin-angiotensinogen-plus animals, et cetera, [00:37:00] renA-plus animals.  The logic of this page is a little bit difficult, but the comment is, it suggests that renin compensation in RA-plus is already maxed out.  So there’s angiotensin-2 effects, angiotensinogen effects are easier to see. [00:38:00] ?Therefore, would it be worthwhile to generate a mouse in which all the renin regulation is absent, but renin is present?  So that actually is the beginning of an idea that turned out to be very productive.  In other words, to make an animal where renin is controlled by promoters, et cetera, which are not responsive to regulations, so it would be a renin mini-pump gene, as it were, gene which would produce renin at a constant rate?  And that idea was later developed with Kathleen Caron, to make a renin gene driven by the albumin promoter, but that’s sometime in [00:39:00] the future; it’s not as far back as this, but there is the seed of the idea in my head.  I expect to get constant expression in the liver free from renin feedback, any type-2 gene, a type-2 gene being a gene which has an effect on blood pressure, but is regulated.  So, making animals in which the unimportant component of the system was constant, and could not be altered by physiological feedback, which we did eventually use. [00:40:00]

More on this general business of renin homeostasis, Thursday, October 10th, page 149, with some comments on what Kim had found, that the adrenal and submandibular renin mRNA has greatly increased in male F1s, because of the ren2 gene, because mice have a ren1 and a ren2 gene, which caused some difficulty at various times.

The experiments [00:41:00] went on, Thursday, October 10th, and Friday, October 11th, step one and step two, making the next part of card work, preparative gel, October 11th, and more experiments of the same type on October 11th, and October 17th, with the IRA — with the internal ribosomal entry site, IRES on page 159, Thursday, October 17th, just on with making these things.

And the book ends on Monday, October 21st [00:42:00] with a prep gel.  Good run, and a good cut.  Quite a nice way to end the Book E’.