Oliver Smithies:

[00:00:00] This is now the beginning of the normal alphabet again, capital A, but now A’.  Starts on June 22nd, 1993, which would have been a day before my birthday.  1993, let’s see, that would be, that’s —


Twenty-two years ago.  So you would have been —


Twenty, I must have been about 70.




I’d be about 70.  Still happily working.  I made a note on June 23rd.  “Roger’s birthday and mine, page 3.” Anyway, this book begins on [00:01:00] June 22nd, 1993, and goes on to the end of December 31st, just continuing the various constructs that were in progress, PCR tests on the reverse PCR control on page 1, Tuesday, June 22nd.

And then, on Roger and my birthday, Wednesday, June 23rd, looking at the review of the various probes that were going to be used to detect the fragments that we wanted, and the recombinant. [00:02:00] To test the fragments that we wanted is a better description of that.  And the gene after-targeting is what I mean by the recombinants.  Purifying the various fragments.

A different type of photography being considered on page 9, for recording the results, a camera called [00:03:00] a Dage camera, can equal Polaroid as judged by the video scan.  Just always trying to have good methods of recording the results, because at this time, they were still using Polaroid all the time.

And nothing particularly notable in the next few pages, just slogging away. [00:04:00]

Revisiting the PCR reactions, Friday July 2nd, page 33, various tests.  Conclusion: not very good on 30 cycles, but the reverse control is doing better.  Concentrate on it anyway, et cetera, trying to make the PCR of the necessary pieces to work, realizing that we have a positive control and a negative control.  This means that we [00:05:00] have means of telling that the PCR is working correctly in both directions, so that the results of the candidates can be believed.

Still thinking about these controls, Monday, July 5th, PCR control DNA.  And with a comment, “Error, this was, the PCR was the blue PCR control, the forward control, now trying to get the reverse,” et cetera.

More PCRs, [00:06:00] page 43.  More check on primers, page 45.  A different type of PCR being tried on Thursday, July 8th, which at least later on was quite often useful, a hot start PCR, meaning that the fragments only, the primers, et cetera, and the reagents were all assembled after the temperature was high to avoid background during the warm-up.  So I used this a great deal later on, what we call a hot PCR start. [00:07:00]

Longer primers being made on page 49, Friday, July 9th.  Some thoughts on the homologous recombination on Wednesday July 21st, page 51.  Since BK-4, the ES cell line BK-4, which is HPRT-minus, is several times better than the standard E14TG2a, it would be wise to test all clones of [00:08:00] ES cells, for poor or good frequency of homologous recombination, et cetera,  with the aim to find an inducing agent that increases homologous recombination, as it’s non-homologous, et cetera.  Don’t need to worry about cell quality.  It’s only to identify a factor, et cetera, so still trying to achieve the dream of improving H homologous recombination.

Trying this camera, [00:09:00] instead of using Polaroid.  The Polaroid was, of course, a direct visualization of the fluorescent DNA bands photographically, whereas the camera would take the image and then print the image.  It was not a film image; it was the image generated by the video camera, which later on, we used entirely.  We stopped using Polaroid film.  I remember telling Polaroid about this that they would lose their business once these cameras became suitable, the electronic camera with a printout from the camera. [00:10:00]

Now then, here, on page 55, looking at something that I don’t remember seeing yet, but it’s important.  PCR and candidates for Ace-dup duplication.  In other words, let’s duplicate the gene to increase its expression.  Up to this point, not much has been said in the notebooks that I remember about that, but this became a very important part of the work that instead of knocking out genes, we wanted to do what we referred to as gene titration.  So here on page 55, Friday July 23rd, that PCR on candidates for [00:11:00] the duplication gene program is 35 cycles, with various comments on what the colonies look like.

A change in focus, or beginning of a new type of work, Dennis Lubahn was interested in the estrogen receptor, and the possibility of knocking this gene out.  On page 57, Friday July 23rd, begins to talk about the endogenous estrogen receptor locus, and targeting construct, aimed at introducing [00:12:00] the neo gene in the middle of exon 2, which would inactivate the gene.

More screening for the Ace duplication on page 61, Sunday, July 24th.  At the bottom of that page 61 with a comment that, “270 colonies, Ace-duplication colonies, have been examined today, all from one electroporation [00:13:00] and next electroporation is ready for screening.”

And we see that continuing on page 67, Wednesday July 28th, zap number two with ACE duplication to PCR on the pools, with the conclusion that “PCR is excellent, and #2 looks very good.  #s 4-7 could be, but the clustering suggests contamination.” One shouldn’t find a whole run, 4, 5, [00:14:00] 6, 7 are positive. You can see this in the gel, on page 66, that two is a strong positive and 5, 6, 7 are also positive.  So a bit skeptical about the results.

PCR screen on the A’ 67 pools, of the Ace duplication are continued.  Not happy about the results.  Getting a little bit more pleased by page 75, Monday, August [00:15:00] 2nd, PCR ACE duplication 2 individual colonies, and the amplification was good.  #6D is a candidate, proceed, and to repeat a PCR for some of the other pools.

On page 77, Tuesday, August 3rd, the repeats are done, with the conclusion, “It’s a very good result.  All the past positives are confirmed.  And 2B and 5A are added, the past ones being 6D, 7C, and 15A,” so a whole bunch of good-looking [00:16:00] colonies.  Should say candidates.

Now diagnostic Southerns, need to confirm the PCR results with the diagnostic Southern.  So here is a good summary of what one is expecting, on Wednesday, August 4th, page 79.  The wild type gene is shown there, and the duplicate gene, or the duplicate chromosome, is shown below, and what one might expect.  For example, a simple test of diagnostics [00:17:00] should be with SstI, SstI + HindIII, and HindIII alone, blotted with the 3′ untranslated gene region, or with the short untranslated segment, et cetera, how to – but need to develop a better map, and so that’s considered on page 83.

But there is some interesting thoughts coming on page 81, Thursday, August 5th, on the gene titration, now specifically called, John Krege suggests stressing the genes, with dosages to be measured; in other words, stressing the animal so that it stresses the gene.  For example, estrogen [00:18:00] or glucocorticoids, or nephrectomy, might push the angiotensinogen gene to its limit and allow discrimination of a phenotype of one copy, normal two copy, and three copy ACE genes to be readily distinguished.  Even if the resting level feedback tends to equalize the three constructs, the three phenotype [drug?].

And here is an attempt at getting better maps, more maps on page 83, Wednesday, August 11th.  All the way now back, doing the first Southerns on page [00:19:00] 85, Thursday, August 12th.  There is problems, “Forgot to filter the gel before pouring, but proceed anyway.”  First film shows a very promising band in 7C, in that particular cell line.  So we have a candidate for the duplication. [00:20:00]

More candidate Southerns following a few days, Wednesday, August 18th through Saturday, August 21st.  And now, we’re doing PCR on Ace dup 3 on page 89, Saturday, August 21st.  But, sadly, despite testing some 40-odd candidates, sadly only the 7 — 5 control worked.

Now, on page 91, [00:21:00] Wednesday, September 1st, I’m starting to look at the angiotensin receptor, type 1B, AT1B, which Masaki Ito, visiting professor from Japan, had been making a targeting construct.  Masaki Ito had finished the targeting construct, but had not made a positive control.  So, I’m looking at that situation and various maps on the opposite page, presumably, prepared by Masaki. [00:22:00]

Some considerations on Thursday, September 2nd, page 95, of improved diagnostics for duplications.  And so here is some theoretical gene, with 5′ and 3′, with looking for gap repair following Vicky Valancius’s ideas, so that one could use a targeting construct that was a reasonable size, even to duplicate a large segment, because only the 5′ and 3′ parts of the homology need be present.  Gap repair would fill in the whole of the missing piece, even if it was 20 or 30 kilobases long, and crossing over could then lead to a duplication. [00:23:00] This was particularly being considered for 12T.

Technical steps on the following pages.  Cleaving the vector on Friday, September 3rd, page 99, straightforward progression, and the preparation of the vectors. [00:24:00] Now, primers for the PCR-positive control of ATR1b.  This is at a time when we really felt very much the need for a positive control, and for that matter, a negative control, so one could detect contamination, or failure, to detect a positive, or a false negative because the PCR didn’t work.  So this is looking at the ATR1b positive control on page 105, it continues, Tuesday, September 4. [00:25:00]

Again, testing the computer printout, on page 107, and with a conclusion that the biophotonics system is easy to use, but is only about 50% the sensitivity of the Polaroid.  So Polaroid is still ahead.

PCR-positive control again, Friday, September 14th, #3 A’105, candidate #3, A’ 105 is correct.  Zapping for ATR1b, page 111, September 20th, and Southern blots [00:26:00] on page 113.  And a probe being isolated for the Southern blot probe on this page, with the conclusion, “The fragment is fine, somewhat larger than expected, and to use 3μL for the probe.”

Still trying to get improved assays for the ACE duplication, page 115, October 6th, Wednesday, “Desirable to have a single probe test for the duplication, using fragments of similar size.”

Thinking about combining knockout and duplication [00:27:00] constructs to make life easier, page 117, Tuesday, October 12th.  If you use one primer, the construct would show you one pair of primers, you could detect, if you had a duplication, and with another pair of primers, you could test if you had a knockout.  And you can use the same building blocks for both.  But, which sites must be unique, and which fragments would be diagnostic.

On page 119, Tuesday, October 26th, there’s a gap from Tuesday, October 12th, page 117 [00:28:00] to Tuesday October 26th, page 119, with a Gairdner 2, “Celebrations are over.”  I’d been awarded the Gairdner prize from Canada for the homologous recombination, the first time being for the gel electrophoresis.  So this was the second Gairdner Award.

And I’m now looking at a beta-globin gene cluster deletions, et cetera, or thinking about them on that page.

So looking for PCR of ATR1b, the fourth electroporation [00:29:00] for that construct, and #16 is positive, and the positive controls work OK.  So, going to do a Southern blot on them.

So, here on page 125, Friday, November 5th, Guy Fawkes Day, looking at the AGT duplication, preliminary round of tests with Hyung-Suk Kim on animals that we thought are 2-1, and animals that are [00:30:00] 1-1, using radio-immuno assay for angiotensin-1, after a renal extraction, et cetera, et cetera, look to see whether the plasma levels of angiotensinogen were different and, it looks very promising.  “Collect more plasmid and repeat,” including some heterozygous knockout animals.  But the results show what was being seen.  1-1 points, and the 2-1 points that all the 1, 2, 3, 4 of the 2-1 points are higher than all of the rest of the [00:31:00] 1-1 points, which there must have been some sort of test against microliters of the plasma, with more or less straight line, with the 0, 1, 2, for the 1-1, and then testing 2μL of plasma for 2-1 and the points are higher.  So it’s looking very promising.  Kim’s work now, to establish that the angiotensinogen duplication worked, and that we have really got a good start at the new type of work.  And indeed, that is projected on the following page 127, Friday, November 5th, thinking about duplicating [00:32:00] the atrial natriuretic peptide receptor A, A and PRA, and atrial natriuretic peptide receptor A had been cloned and characterized in the rat.  And, Paula Oliver and Simon John in the lab have isolated and characterized several phage clones, have isolated — made various subclones, and that some of the pieces could be used for knockout and for duplication, using the idea already talked about on page A’117.  So, duplication and [00:33:00] knockout being considered.  The PCR, the controls on page 126.

So beginning of a very successful usage, again, or extended, not the beginning, extending the usage of the idea of gene titrations to understand what’s going on.  This was used successfully for several genes.

Going back to this idea of finding a universal plasmid for inactivation and duplication on page 133, Saturday, November 6th. [00:34:00]

Beginning of some idea here of what I was thinking of as pressure coefficients, Wednesday, November 10th, page 139.  Slope of the angiotensinogen concentration, curve this as blood pressure.  At normal blood pressures, will be defined as the pressure coefficient for AGT.  This idea never came to anything, but it was a beginning thought anyway.  It is an effective partial derivative of the blood pressure pi with respect to concentration.  And so, [00:35:00] a little equation there that delta pi = A*delta AGT + R*delta renin + C*delta ACE, et cetera.  In other words, the blood pressure changes, with little changes in a lot of genes, beginning of this multi-gene thought, that blood pressure can change as a result of just many genes, a little change in angiotensin, a little change in renin, a little change in ACE, et cetera.  And these are pressure coefficients; in other words, the constant in front of the delta, AGT delta being, et cetera, could be [00:36:00] positive or negative, and assume, initially, that they are added in.

And some dreamy thoughts on page 138.  Despite the fact that this was, these pressure coefficients never went anywhere, the idea was very real, and very documented, what we called, “many little things.”  In other words, the normal blood pressure depends on many little things, not on single big changes.  Nobuyo and I wrote an article on that topic, which we called, I think, “Many Little Things,” I have to look up that reference. [00:37:00]

So this general idea that the blood pressure was a result of changes in many little things was actually an essay that I wrote for March of Dimes, a Perspective, after being awarded a research prize in developmental biology.


We had one together in PNAS saying that gene targeting approaches to complex genetic disease.


[00:38:00] OK, let’s start again, right.  So this was continuation of the idea of this, of many things, “Many Little Things.”  We first published this idea in a review, by myself and Nobuyo, on gene targeting approaches to complex genetic diseases, atherosclerosis and essential hypertension, and we talked about the value of knockout in studying atherosclerosis, and the effect of varying the copy number of angiotensinogen and exploring the effect of the concentration of angiotensinogen on blood pressure, and then the actual [00:39:00] blood pressure is measured on the conscious and unrestrained animals with one copy, two copies, three copies, or four copies of the angiotensinogen gene, and their blood pressures arranged in almost a straight line from 122mm of mercury up to 147mm, so it was a very clear statement of the fact that varying gene level, varying expression of genes could be important in controlling of blood pressure, not with the idea that humans have a variable copy number, that in humans there may be other changes which cause differences in levels, and this can be mimicked by, [00:40:00] or modeled by the gene titration method.

So, the end of the abstract of that review by myself and Nobuyo is that gene targeting experiments, when properly designed, can test the effects of a precise genetic change, completely free from the effects of differences of any other genes, and allowing proof of causation, and that one could use it for changes in level, as well as absence of a gene.

It would probably be a good idea to scan into the record, a much later review that [00:41:00] I wrote in 2005 in response to being awarded a research prize in developmental biology by the March of Dimes, because this was a sort of review of all of the philosophical and actual thoughts and experiments in this area. So it starts with the meeting with John Krege, which I talked about.  It talks about thinking about the reading angiotensin system, and all of the genes that were in that system that can be modified, and how the gene duplication idea was introduced by [00:42:00] a Germ Matrix paper, then the work with Hyung-Suk Kim in which a gene was duplicated.  And, a summary of the results of angiotensinogen on amount of product, and on the blood pressure, and of the ACE gene copy number, and its failure to make any difference in blood pressure.  So I’m suggesting that we scan that into the record to help interpret what’s going on in the notebooks, where my 2015 memory has not been as good as [00:43:00] it might have been.

So going back, I got another prize on page 141, North Carolina Prize Day, very nice to have an award from the state, North Carolina Prize Day, November 12th.

So here is a summary of the angiotensinogen data on blood pressure, the 1st measurements, on page 145, Monday, November 15th, with one copy of AGT.  The AGT level is 53%, and the [00:44:00] two copies by definition, 100%, and the three-copy AGT has 119%, so the increase in level of angiotensinogen is modest.  And then, on the right-hand side of the page are the level of AGT on the left-hand page, and on the right-hand page, the blood pressure.  So the mean of all of these were male, the mean of the wild type was 117.6mm of mercury, the mean of 1-0 was 118.7, and 2-1 was 119.0, so it looks as if it didn’t have any effect on blood pressure in the initial tests. [00:45:00] With the comment that, “Although a level of angiotensinogen change, this initial work, it looked as if the blood pressure did not change.  The BP data look as if there is compensation!”  That later on turned out to be not correct, with more extensive work, and more careful measurements.  But we did find the change in blood pressure.

So, this, November 15th, attempt, was negative, as far as blood pressure control was concerned. [00:46:00] But when we really got down to it and completed the work, which was published by Hyung-Suk Kim and many others in 1995, there wasn’t any doubt that the level of the gene product changed. But really, if one measured the tail-cuff pressure data showed an increase from the knockout through one copy, two copy, three copy, four copy, was decreased, but if the mean arterial pressure was measured, then there was a straightforward increase in blood pressure with genotype, one copy, two [00:47:00] copy, three copy, four copy.  Don’t know why one was — it was the method of measuring the blood pressure, the tail-cuff measure, which were, in this case, not quite so good as catheterization by the carotid artery, and measuring directly.  So, as a method of measuring the blood pressure, it was important.  But we didn’t know that, of course, on page 145, November 15th.

Had a trip to Japan between November 15th, and December [00:48:00] 1st back from Japan, visiting Nobuyo’s family, maybe gave seminars, I don’t know what happened exactly at that time, but it was a happy visit.

Experiments starting on A’143, transformation, changing a site, a simple experiment, but with the usual sort of conclusion, “Start again.”  So repeated on the following page 149.  Looking at the orientation of the positive PCR control with Simon John [00:49:00] for the atrial natriuretic peptide receptor on page 150.

On page 159, or on the opposite page, a large set of data entered from table of the animal types, the sex, [00:50:00] the angiotensinogen genotype, and the various assay extracts to measure the level of angiotensinogen.  And with the provisional conclusion, “December 31st, 1993, that the data from the males on 12/17 look very good.  Females on 12/21 are off-scale on the radio-immunoassay, so repeat the assay.”

And on the following page, the last page in this book, Friday, December 31st, the new angiotensinogen data are plotted, and the number of gene copies there are listed as 1, 2, or 3, and the percent of normal angiotensinogen level of 35% [00:51:00] in the one copy, by definition, 100% in the two-copy, and 112 to 24% in the three-copy. So covering the range that was observed by Jeunemaitre and was sent to PNAS, January 6th, 1994, and that’s the paper that I referred to from Hyung-Suk Kim, which was published eventually — we must have the citation here.  Anyway, it was sent to PNAS January 6th, 1994, but [00:52:00] maybe it was — I don’t know what happened.  Maybe it was rejected or something, because the actual contribution to the paper is listed as November 30th, 1994, so there must have been rather a long period of getting the data, from the way that it went back to the reviewer.  I don’t have the answer.

And so that ends Book A’.