Oliver Smithies:

[00:00:00] So this is book K’ beginning June 14th, 2000, and ending in March of 2001.  Beginning with the continuation of the Alu-rich element preparation, PCR for ARE continued from J’-157.  And, the conclusion, it looks like the VM purification leads to polymers, and then a null conclusion, “Proceed.  The polymer looks as if salines were OK,” etc., etc., trying to polymerize the ARE, I want to more pure, [00:01:00] get more than one copy of the ARE entities, into pSKB1.  So there is b-actin EpoR checking on colony here, 28, and colony 31, from B101.

Assembly of the ARE, Friday, June 15th, page 5.  Conclusion: “Doubtful, but possible.”  Conclusion two, “The cuts look good.”

Going on with the assembly on the following page 7, June 18th. [00:02:00] And further work on the colonies earlier obtained.

So sequencing ARE and pSKB1-AVR2, Tuesday, June 20th, page 13.  “The sequences confirm that Blue 10 is correct, though some slight mix-up of labeling, very, very unlikely to be wrong.”  Assembling it into pSKB1-AVR2 the b-actin-EpoR insert, on page 15, June 21st. [00:03:00]

Friday, June 23rd, my brother and my birthday, 75, and I flew both my aircraft, November Eight Four Kilo Sierra, the Grob motorglider, and November Seven Five Six XP, the Cessna-182.  So it was a good day to celebrate by flying my airplanes, the same time doing experiments.

Following page, 19, Saturday, Sunday, July 7th, 2nd and 3rd, some thoughts on Ariel Gomez’s cell recruitment ideas, related to the renin gene, [00:04:00] or the renin-producing cells, formed during the development of an afferent arteriole.  Wondering if cells change.  Fairly complicated argument.

Going back to constructs, pSKB1 with the enhanced green fluorescent protein in it to backup the T-EpoR work, I’ll make a vector [00:05:00] for EGFP and cell sorting.  Just, so that I had a positive cell with EGFP inserted into the HPRT locus.

So, various steps needed to do this.  Inserting a HindIII site in the GFP plasmid, July 11th, Tuesday, page 23. And with, after being July 11th, and then the next entry being Saturday, July 22nd, with “Back from Japan today.”  So I must have [00:06:00] had a visit to Japan, with Nobuyo probably, to meet her parents, or to see her parents go to a meeting.  There was an insertion of an HindIII site to EGFP, a simple modification.

Some more switching thoughts on Sunday, July 23rd, page 27.  “Because renin synthesis is greatly suppressed in the kidneys of the renin transgene mice, Katherine’s, albumin-type of promoter in the liver, they can be used [00:07:00] to see intrarenal increases.  So, those already have almost no expression of renin, so you can look to see if you can increase the expression.  So, some thoughts on this type.

Some thoughts on page 29, Sunday July 21st, various Japan trip thoughts, so various things that I’d thought of during that trip, I just listed, presumably from bits of paper that were lying around to make sure I didn’t forget what the ideas were.  [00:08:00] Thinking about modeling 13, the possibility of two different – of 13 different genes, each having a lower or higher expression, and what would happen to blood pressure if it was controlled by multiple – by small changes of many genes, so it may be worthwhile to simulate a 13-card locus, multigenic blood pressure system with two alleles, plus being red and minus being blue — black, versus blood pressure, so the blood pressure would be low, if all 13 genes were in the form, and blood pressure would be very high if all genes were in the red form, but it all would be a sort of Gaussian distribution [00:09:00] between the two.  “And the frequencies of each pressure would be a binomial and would be normal, an interaction between loci is possible,” etc.  So thinking about modeling his — I don’t know that I ever did, it was fairly obvious that what the result would be, it wasn’t really necessary to replicate it in practice.

Back to vector preparation B-actin — beta-acting EpoR vector, Monday, July 24th, page 31, going on page 33, the same thing, 35 also, and 37.  And preparation of the parts needed, still going on, page 39, and final assembly of the [00:10:00] pSKB1, with the enhanced green fluorescent protein lacking the poly-A sequence on page 41.  And I decided that #12 is good.  So, turned everything over to Seigo.  So, a little comment that, “I change the EpoR to EGFP in about three weeks,” rather how happy about how fast it had gone.

And, on [00:11:00] page 43, Friday, August 4th, I’m talking about a sorting antigen, and Stuart Bentley had identified on my request an antigen for which magnetic beads are available, that could be used for cell sorting, and thought it might be useful to use this.  And he found a trunk ADH to K^K antigen, expressed normally in only a few mouse strains.  “Needs beta-2M for expression,” etc.

Now on page 45, Wednesday, August 9th, thinking about a fluorescent indicator for the atrial natriuretic [00:12:00] peptide, ANF factor, atrial natriuretic factor, in order to better study cell recruitment on the volume overload, et cetera, of ANP in the heart, we need in situ with a good mRNA probe.  In the interim, can use the fluorescent surrogate for ANP, and the thought being, how to go from BNP-ANP to BNP-fluorescent marker, and then ANP to put in a fluorescent marker between the BNP gene and the ANP gene.  To execute this I’ll need more sequenced data upstream of the ANP gene.  [00:13:00] And so, on page 44, I show the two genes, beta-ANP — BNP and ANP, and the construct that I’m trying to make.

Making an enhanced yellow fluorescent protein cassette, Friday, August 11th, page 47, a possible assembly route being talked about starting from Clontech plasmid.  These are wanting to get different plasmids, with the yellow fluorescent protein, [00:14:00] and the cyan fluorescent protein, and cyan being expressed in the nucleus.  So here are a bunch of plasmids being made that would have different colors of fluorescent protein for use in tracking what’s going on in the heart, or in other places, which we did quite a lot of work later on.  And this is making the plasmids with different fluorescent proteins.  All going into pSKB1.  At least, I presume they’re going to pSKB1.  We’ll see.  Anyway, page 51, Sunday August 20th, [00:15:00] pSKB1, Epo without poly is being looked at again, with an error there.  Pointed out on page 50, “Too big a hurry.  Costs 10/3 days and very obvious.  Oligo should be this,” it didn’t really cost that much time, but it cost some time.  It was the wrong way around the sequence.  And some of [00:16:00] the plasmids were, I was having difficulty because some were kanamycin-resistant, and some were ampicillin-resistant, and I think I got that mixed up a bit.  As talked about on page 53.

An IRES into pSKB1 EGFP, Tuesday, August 27th, page 55.  Still working on fluorescent protein vectors.  August 28th, a rather strange entry, page 59, Monday, “Recent failures.  The simplest hypothesis is that the current batch of ligases is bad.  Others [00:17:00] are having a problem.  So what to do?”  Found an obvious error in the end in the IRES linker.  But the other linkers were OK, so it’s only part of the answer.

pSKB1-EGFP reviewed on Tuesday, August 29th, page 61.  Trying to sort out the problems.  Continuing the same sort of thing on page 63 IRES and poly for pSKB1. [00:18:00] That continues with Sunday, September 3rd, page 71, IRES into pSKB1 EGFP.  Similar experiments continuing, Tuesday September 5th, page 77.  IRES insertion in EGFP, still haven’t got it in yet, and still working with poly on the following page.  So, assembly of pSkb-EGFP poly [00:19:00] and at that point, my — why am I making this?  I think it’s the beginning of the idea that you could test poly-A sequences.  Not really quite sure why we’re doing it.  I know what we eventually did, that we made some pSKB1 variants that would allow us to check what a poly-A sequence, how strong it was in expressing, or checking promoters.  So it’s a promoter assay, and a poly-A assay [00:20:00] could be made with pSKB1.  Maybe this is the beginning of that work.  It probably is the beginning of that work, to get vectors which could be used to tell you how good a certain poly-A type sequence was, or how good a promoter was, by inserting them all into the same locus in single copies.

Trying to get the sequence into pSKB1, GF1, EGFP, [00:21:00] poly-A on page 89, Thursday, September 14th, where the sequence is very clearly shown, and what I’m trying to get the poly-A sequence in there.  And here are, on page 91, is reversed, or page 90 is reverse-poly, so we could have either direction.

Testing pSKB1-EGFP poly, October 24th, page 92.  They give strong fluorescent signal with Hat resistant colonies and normal homologous recombination frequency. [00:22:00] And note that S340-poly-A sequence works in either direction.  Didn’t matter which way around it was.  So pSKB1-EGFP was a good plasmid, and gave good colonies.

Thinking about blood pressure data, protein, and expression, Saturday October 7th, and Sunday October 8th, page 95 and analysis of Kim’s RT-PCR data, with — there were two most striking [00:23:00] results are that AGT-mRNA effect on renin mRNA and aldosterone synthesis mRNA, et cetera.  And thinking about what this all means, and diagrams of the interpretations, as we published this type of sequence later on, or this type of analysis of what happened in expression with various modifications.  I’ll try that again. [00:24:00]

On Saturday, October 7th, and Sunday October 8th, page 95, an analysis of Kim’s reverse-transcript PCR, real-time PCR data on expression of AGT mRNA, and on renin mRNA, and on aldosterone synthase mRNA, and their interaction, various ideas on how these may be related to each other, rather nice three-dimensional type diagram on page 94, summarizing the data, [00:25:00] which is continuing to be thought of on the following page, 97, Monday, October 9th.  Past experience is that one copy x2 is equivalent to a two-copy — I’ll backtrack on that.  This is 1-0 vs. 1-1, and 2-0 vs. 2-2, because 1-0 is a half, usually.  1-1 is normal, and 2-0 might be like a 1-1, versus 2-2.  So, here is a thought  that the past experience is that 2×1 copy, which is wild type, is equal to a two-copy [00:26:00] when the genes and background are identical.  And though I was seeing that the interaction of these genes is very strong, because the aldosterone synthase levels increase almost linearly with the increase in AGT copy number, not really understanding that at this point.

Thinking more about Kim’s blood pressure data on page 99, Thursday, October 12th.  Kim measured the blood pressure of all the AGT animals in the 1-1, 2-1, and 2-2 groups, and various initial observations are that the [00:27:00] blood pressure is markedly affected by AGT in both male and female.  And though, that the blood pressure in the higher AGT animals is not correlated with renin, and gender has no effect, etc.  So these various things about what, how the blood pressure data fitted to the gene copy number, which eventually led to a very nice publication with Kim.  So this is Kim measuring the blood pressure on all the AGT animals which he has, and various results of that type being talked about, to see if one can make, understand the interactions, and more data [00:28:00] analysis on the following pages.

Tests of the effect of AGT by genotype, renin by genotype, and AS by genotype, et cetera, all sorts of correlations being looked at, on the analysis of the data for feedback.  Tests of MW’s log transformation. [00:29:00]

Interviewer:

It says Mike West here.

OS:

Mike West over here?  So this is Mike West data, that he suggested log, that this is — I’ll backtrack a little bit.  This is analyzing Kim’s blood pressure data.  There is some suggestion from Mike West that sometimes log values can be used to bring variances into better shape.  We’ve used several times in the future, in future work.

So sometimes using log transformation is helpful, and on page 101, Monday, October 17th, [00:30:00] Tuesday, October 17th, some of these log transformations are being tested to get the data to fit more consistently.  Looking for feedback from AS to AGT-negative feedback, still borderline.  Renin results are very clear.  Increase or decrease of renin with changes in blood pressure is progressive in either direction, and has no detectable curvature.  In other words, if AGT goes up, renin goes down, and if AGT goes down, renin goes up, and it was very strong homeostasis with the renin gene, as everybody now would agree. [00:31:00] But we were seeing it in the mice.  More analyses of the same type on page 103.

Even trying Bazian treatment, always a difficult task for me, on October 24th, page 105, Tuesday, looking at these various correlations.  Data analysis continued with some fairly good results on Tuesday, October 31st, page 107 and 106; for example a very nice [00:32:00] effective log, blood pressure by log AGT, is a reasonable ellipse, log by log, but still a good slope and a good r-square.  Correlation: 0.75, or 0.76.  That’s not r-squared; that’s just r, but still, quite good.

And looking at males and females, showing different effects on the bottom on page 106.  So the questions being, on page 107, “Are the blood pressure data distributed normally?  As blood pressure, or log blood pressure, and is there a gender effect?” [00:33:00] And the answer is, “The log blood pressure’s slightly better than blood pressure, but there’s a neglible gender effect.”  Question two, “Using combined data and log blood pressure, which variable is most correlated?”  Very clear answer, “Log AGT 57% of the variance due to AGT.”  But this was log-renin, predicted by log AGT, etc., trying to do this.

And then, on page 109, the following page, Friday November 6th-Friday November 10th, “Beginning to learn how to use my glass micro-arrays,” trying to work with micro-arrays for the [00:34:00] first time.

It looks like there’d been a fire at some time, at the bottom of these two pages. (laughter) Distinctly, a burned look about them on page 111, Saturday, November 11th, Armistice Day.  More now microarray analysis on page 113, Monday, November 17th.  Another grant sent off on page 115, February 28th, Wednesday, and with a unifying tested hypothesis, [00:35:00] the cell type switching hypothesis, to try to get to stable states, and the transition I made a stellar program that would do that type of switching, though it had various logical flaws in it, but I did make a program that would switch when stress increased would certainly switch to a different state, and then when you reverse the stress, it would have to go quite a long way back before it came down again, so that this was a switch between two states.  And grant, all that, that wasn’t the only thing in the grant, but the grant was funded.  Still having burns on this [00:36:00] book on the following page, and now a lot of construct with different colored EGFP genes, and this testing of using BSKb1 as what was called, “code-test, pol-A and no pol-A.”  These are the constructs that will allow one to try what happens if a coding sequence is altered, what happens to expression, or what happens if the pol-A sequence is changed.  Quite severe burns in these pages here, quite a lot is missing, making these [00:37:00] many constructs.

I remember at that time fairly well, and they were rather pleasing to make, and quite effective in a way.  Here, for example, on page 129, Friday, March 9th, is a cyan fluorescent protein, to get the cyan fluorescent protein in the nucleus, that one could do that.  So that was called K’129, the page number, CN. [00:38:00]

And then, a different sequence HRGN, page 131, different fluorescent plasmids in, and one with IRES, page 133.  Continuing many pages thereafter.  Page 143, Thursday, March 15th, red fluorescent protein is being reconsidered.  And cold tests plasmid reconsidered.  And Seigo, for example, saying that Seigo now has accumulated substantial evidence [00:39:00] that the beta-actin promoter, co-directional with the HPRT promoter, is four times less effective than the reverse orientations, therefore continue to make fragments, pending the availability of pSKB1 test code.  These are the pSKB1 constructs that allow one to test different parts of the gene, varying at different parts of the gene.  And this is coming to the end of this book, still making these fluorescent tags into pSKB1 code test pol-A, for example, putting in a red protein, [00:40:00] red fluorescent protein on page 157, Tuesday, March 27th.

And the book ends on page 162 with a summary of five weeks’ work, April 9th.  That’s now I have seven different color plasmids available.  Green fluorescent protein, Clontech strategy, an HRGN, Stratagene, HRGU, Clontech Y, Clontech Y nucleus, Clontech C nucleus, Clontech C, Clontech R, Clontech XRM.  I think it’s mitochondrial, but I’m not sure at this point, but different constructs.  And that ends this book with seven available constructs, Book K’. [00:41:00]