Oliver Smithies:[00:00:00] So this is Book o, starts December 16th, 17th, 1976, and goes through April, to the beginning of April 1977. So, again, a fairly short period of time. And this begins with repeating the experiment that had been done earlier with new DNA. Test Charon 3A, plus lac, for ligation and nicks. Fresh sample of R1, from [Harris Sugden?] obtained at the beginning of the book, page 1. [00:01:00]
And then, various combinations of R1, HindIII and ligase, etc., etc., trying to find out safe conditions. So, on page 5, it shows a DNA with RI and DNA with RI plus HindIII, etc., with and without the ligase. The gels don’t look terribly good. The conclusion is, “Not bad for a first try. Protein gets in the way of the electrophoresis. Control looks to be in trouble, with no enzyme. So there may be nuclease in the cesium chloride [00:02:00] which can survive 85 degrees centigrade ?!not very likely.” I didn’t think it was when I wrote it.
So further 3A ideas, Charon 3A ideas, reminding me “The 3a lac has an RI site, and then two HindIII sites downstream. So therefore, EcoRI to completion, plus any type of HindIII that gives preference of the left site is a big help.” That’s probably not 3A minus lac; it’s probably “3A lac,” between with lac. Not sure which; it’ll become clear. [00:03:00]
Further thoughts on page 9, on the cuts, HindIII cuts, and ligating into the bacteriophage DNA. With HindIII ends, the problem is background from self-closure, so better to use something with tailed G or C to avoid self-closure.
As, talking more about the same sort of ideas on page 11, “Therefore test embryonic DNA, increase [00:04:00] concentration plus lambda visible, and then lambda at trace levels, for nick translated, translation 32P, transfection,” etc., etc. So trying hybridization on December 20th, with a gel, and nitrocellulose. Looks like a Southern gel.
So, on page 15, starts with sodium chloride/sodium hydroxide, solution 1, for about 30 to 45 [00:05:00] minutes. Something about 60, hour, not very clear. So, 3M sodium chloride, solution 2. Solution 1 is sodium chloride/sodium hydroxide; solution 2 is sodium chloride with Tris buffer, and then nitrocellulose; it’s a protocol for a Southern blot, showing how to do it, with hybridization [solution?].
And, on page 19, some results. [00:06:00] Genomic DNA with RI, and a little bit of lambda phage in it, not enough to see on the ethidium bromide stains.
And remaking a nick translated probe, on Tuesday December 21st, and a characterization of the lambda probe on page 23. “[Bam?] buffer, Charon phage, probe, 20,000 counts per minute, etc., Bam digest, OK to use, but clearly 15 degrees would be better.” That’s not very important. [00:07:00] Just trying to get hybridization to work. Just a general smear with, marked where there should have been bands.
There’s [Sugden’s?] tests, tips on page 25, and Friday, Christmas Eve, an experiment set up. The first use that I remember, Denhardt’s solution, which we used all the time, on [00:08:00] page 25, transferred to 150μL of Denhardt’s 6x SSC Ficoll buffer. It contains BSA, and polyvinylprrolidone, etc. “Forgot to denature the probe, mistake. Should have denatured the probe first,” on page 25. So here are some tests of 3A lac, and HP4, HP9, HP11 clones, with the conclusion that [00:09:00] exposing them to DNA and ligase, and R1.
With a comment that, the conclusion on page 29, “The ligation is good in the 3-hemoglobin cloned, and adequate in the 3-lac. There are no RI sites in the hemoglobin probe.”
Transfer tests on Christmas Day. [00:10:00] These are the Southern blots. Nice, heavy signals on page 30. “The results show even on first try, reasonably good transfer, and sensitivity. No background. So very prolonged washing is not needed,” etc., and so what was the washing? However, it’s clear that the probes need to be at least one order of magnitude [hotter?] or more DNA, can just see one copy on 10^-4 dilution, presumably. [00:11:00] Next step is to used precipitated RI DNA at a higher concentration, but there are quite heavy signals, and, more difficulty in seeing the signals when there is enough label.
Eukaryotic DNA testing, December 27th. Yes, it now talked about it being a Southern, “In view of lambda hybridization Southern test success,” so it was a Southern blot test, “decided to try mouse DNA with the cDNA probe.” So I isolate from mouse liver, (inaudible)-inate it, pronase it, and try Stafford’s conditions. That was [very?] North Carolina, Darrel Stafford [00:12:00] who had very good conditions for making good DNA from eukaryotes.
And yet, there on the next page is an outline of this procedure, page 35, “Stafford’s procedure.” And, we’re using it. “By the morning, the proteinase K looked good. Pronase didn’t go, and wouldn’t get viscous even with sarkosyl, so abandon.” So the proteinase K looked good.[DNA?] continued, etc. “No viscosity, try again. Too low sarkosyl was suspected, I added 0.5% of 30% solution,” and so added [00:13:00] later more sarkosyl.
Page 39, evidently getting happier, because they’re very much better. “Strings up to four feet, 3x phenol, pouring off the DNA, extremely viscous. Centrifuged and left with the last phenol [in the cold?].” So here was very good DNA being made on trial 3.
Continuing on the next page, DNA continued, with chloroform, isoform, isobutanol extractions, and phenol extractions, and dialysis. So, here we get to the material now with this little Book o, “O41 mouse DNA dialyzed.” The length of the string [00:14:00] has dropped substantially, since the RNA is now about two inches, having been initially about four feet. But I think the four feet is actually, probably means it wasn’t really free from protein.
New nick translated probe being made on page 43. Charon 3. [00:15:00] So on page 42, 3,285 counts per microliter of 32 labeled phage O43, phage labeled material.
EcoRI digests on the marked DNA on the next page, and then following page, some filter papers that are not described with blue plaques presumably in prints of [00:16:00], after lac digestion of a bacterial plate, after lac development of galactosidase.
So, very heavily labeled material, more hybridized, on page 47, conclusion that, “The digestion goes OK, but there are quite a lot of RNA fragments in this material. So, the starting DNA looks good.”
So DNA [00:17:00] transfer tests. This must be Southern blotting. Bulk of (inaudible) at one, plus 1/100 of R1 lambda phage made a thick gel, Helling’s x4, and transfer and hybridize to both probes. [00:18:00] And, the result of hybridization.
New Year’s Eve, working New Year’s Eve, not R1 transfer. And, more experiments on the transfer, on January 1st, Saturday, so working Saturday, January 1st, 1977. And I think there are several other times, similar sort of timing, with [00:19:00] January and Saturday.
Gel overloaded for transfer and before and after transfer images to show that the DNA is being transferred nicely in the Southern procedure, with more experiments of the same type showing that the results show excellent transfer on successful hybridization, so the Southern blots are working.
Real tests now on Tuesday, January 4. The rest of globin phage 53 strip, of [00:20:00] — not clear what the phage 53 must be. I see, bacteriaphage on 53. Yeah, this is a DNA strip that has been transferred with mouse DNA, containing trace of bacteriophage in it. That’s what I call a phage 53 lambda strip, and so that phage 53, or globin strip was transferred to Denhardt’s, etc., and hybridized to cDNA single-stranded from Harvey Faber, so I had some cDNA [00:21:00] 32-labeled globin probe.
This is the first mention of that, but in fact, the experiments were done — the successful experiment isolating the globin genes were done using probes which had been made either from human or mouse mRNA that had been, from which complementary DNA had been prepared by reverse transcriptase, which were the probes, and that’s presumably what this is being referred to, on page 57, 1.8mLs of 32p-labeled cDNA, single-stranded globin from Harvey, Harvey Faber, 2.2×10^6 counts; thousands more [00:22:00] EDTA. And, hybridized, so that was what the probes were. So here I’m trying to look for a globin fragment in the DNA.
Must have been a period of grand-writing in there, because Tuesday January 4, and then next is February 15th, with a comment, “Grants all done, meeting attended. So probe for direct hemoglobin clone detection,” [00:23:00] so it’s talking about the clones. If the clone hemoglobin is about 1% of the bacteriophage lambda, then it would mean about 100x more cold lambda to decrease the background to equality, so the idea being to have a large excess of cold phage DNA and using a bacteriophage that contained a cDNA, and relying on the cold DNA to prevent hybridization to the clone. So therefore, try S1 formamide on the hemoglobin clone, [00:24:00] etc. Add “stuffer DNA,” and S1 buffer, etc., etc. S1 nuclease is single-stranded nuclease, is being [00:25:00] also considered. But we’ll just stop there for a moment again.
So this is, on page 59, Book o, talking about the probe for direct hemoglobin probe detection. And if a clone DNA is 1% of lambda, then it would need 100x cold lambda to decrease the background, so the idea being to have a large amount of clone that doesn’t contain the cDNA, and the clone with the cDNA in it, and label the clone with the cDNA, and use it as a probe by getting rid of the common part by having unlabeled DNA [00:26:00] to compete. And, talks about since there are no RI sites left, and HHA leaves still about 5% of lambda, need to try S1 nuclease on the dA/dT tails. Not immediately clear why I wanted to do that, but it might become clearer later on.
So the experiment is done partly on page 61, with a possible faint band on [84?], [00:27:00] which is called, [Fi?]-61. So the Southern blot, repeat on S1 tests without the stuffer DNA. So, nuclease being tested at concentrations from 3-1, 1/3 and 1/10 of the S1 nuclease. And looking at what happens without the stuffer DNA. [00:28:00] With a conclusion that the same piece in hemoglobin 4, it’s about 3.4% of lambda phage, trying to find out what was there, and not immediately clear.
So, on page 71, preparative run, DNA probe from O-69, which was made by precipitation. Let’s see what [00:29:00] it was. Let’s go back, on Saturday February 19th, preparation of the probe. And so here is the actual experiment, which was dATP, dCTP and dTTP, and dGTP, all radioactive. 52 microcuries total. And nick translation buffer, and then hybridized to — I mean used to label 3AHb4 DNA, 20μL of that material. And that is called, [00:30:00] the final product is called, O-69 [post?] S1 probe. Relying on S1 not to digest the cDNA that is now — not the cDNA — yes it is, cDNA still hybridized to its template.
So a [00:31:00] comparative run on Sunday, February 20th, page 71, where a band is cut out at the expected size, as the probe. The pattern is dark enough, (inaudible) it says, but the bands don’t show; therefore stained, and cut out. But anyway, the probe being made for the first time.
Being tested in page 73, and contest the five potential probes, [00:32:00] which I have, HHA digests of Charon 3A, 3A-delta-lac, Charon 3AHb3, Charon 3aHb4, transfer and hybridize. And the gel are shown before and after transfer. [00:33:00] And transferred as usual, February 22nd.
February 23rd, page 77, probe extraction. “We’ll try boiling bath to melt Helling’s 4x containing gel.” And also trying freezing to get the material out of the gel. And the gel can be melted in boiling water OK. And it resets, [00:34:00] and then, sodium hydroxide was added. And, I got a semi-solid product of the probe.
And probe tests of the material [method?] isolated on page 72, or 72 probe tests, sodium hydroxide, etc. “Used half for the test,” etc., “and froze the rest.” A few of the notes here are a little bit confusing, but it looks like [feta?], but it’s really [00:35:00] all, and with a slash through it, so it isn’t confused with a zero. I don’t know why I was being so funny about it.
So probe tests continued on Thursday, February 24th, with a conclusion that “The agarose method is very impressive, but the probe is doubtful as specific; therefore repeat with S1 nuclease first, and try all DEAE-HCl method to clean up the agar; the agar might not have been clean.” I remember being able to clean agar with DEAE. [00:36:00]
And Benton and Davis protocol for screening lambda GT recombinant clones by in situ hybridization to single plaque, which eventually became the sort of technique we would use, hybridizing to plaques, taking [00:37:00] nitrocellulose replicas, etc. That was used in later stages when it got to the time of trying to do recombinant DNA work. Cleaning up the agarose to get rid of anything that was negatively-charged.
Testing two gels, [00:38:00] one not treated, 0.7% Helling’s x4, and 0.7% Helling’s x4, treated with DEAE to see what difference this made. Let’s see what the conclusion is. Before the gels, and got some supernatant from, and tried with HindIII, with a DEAE-tested sample, and cut from the gel, and [00:39:00] the usual agarose cut from the gel, this is not being treated, with a comment on the bottom, “No digests in any of these samples,” which were being digested with RI, EcoRI.
An odd result on page 90, from nick translation from the gel, etc., just struggling to make a good probe.
Odder still on the next page, 92, nick translation again, just repeats of recent gels on page 95, [00:40:00] and cut out a band for lac.
Further nick translation ideas on page 97, still working at the same problem, trying to get a good probe. Hydroxyapatite being used on page 97, and [00:41:00] on HTP on page 99. Continuing with [work?].
Page 101, Friday March 4th, trying to see whether the material could be improved by [00:42:00] using hydroxyapatite, HAP, or the Tiselius type of hydroxyapatite, which is called “HTP.” [00:43:00]
On page 97, Thursday, March the 3rd, further translation ideas, talking about trying to improve the method of getting the gel material, material taking from the gel to be well nick-translated. And, comment here that, “The idea is gel electrophoresis freeze, and nick-translate, and hybridize. This is OK if the probe is already hot.” (inaudible) comment on, “Trying Kelly’s procedure, which is described on page 74 of this book, and that was banding potassium iodide,” [00:44:00] etc., etc., “and using hydroxyapatite to bind the DNA and then elute it.” So it’s a hydroxyapatite procedure, and so the idea being then, which is Kelly’s procedure, is to dissolve the gel in 6M potassium iodide, 10mM phosphate at pH 7, and absorb to a small hydroxylapatite column, and wash with the potassium iodide to remove the agarose, and then with low phosphate, to get rid of the potassium iodide, and then elute wit high phosphate, still would need to dialyze, or Sephadex, a fairly complicated [00:45:00] procedure. But before giving up my own idea, try adding solid potassium iodide to the frozen gel. But as expected, it clogs, the difficulty being to get the DNA out of the gel. That’s what this is all directed towards. So page 99 goes on with using HTP which is a different form of hydroxyapatite from Bio-Rad, treated with a procedure, that had been devised by [Tizaleas?], so HTP is hydroxyapatite with [Tizalias’s?] modification. And, [comment?] comes out to say that, the gel is treated in this way. “And so it works,” is the conclusion. [00:46:00] “Need to do B-32 to see what the yield is.” It just shows a gel before and after cutting out the sample, and Southern blotting.
So page 101, continuing with using this type of test. Saturday, March 5th, not liking it too much. It is clear that the S1 method on AT is not very good. It’s also clear that nick translation on an S1 piece is not going to be easy, and therefore alternate generations, etc., etc. [00:47:00] So it’s talking now about beta-chain hemoglobin clones. “Take out [putative?] beta chain clones, cut with HaeIII, which has sites between 26, 27 glutamine, glutamic acid, alanine, and a Bam site at valine/aspartic 98/99, and another HaeIII site at 134/135,” etc., and talking about thinking about the actual sites in the DNA. So, [v?] clone, etc., etc.
And the present maps are, here is a map of hemoglobin 11, Hb11, [00:48:00] is a clone obtained — no, I don’t really know. But it’s chromosome 11 is the hemoglobin, the beta chain locus is on chromosome 11. Hemoglobin 11 is a map of that region which is, perhaps, evidently had already been published, showing band sites at various places. And HaeIII, etc. And, isolating fragments and purifying them. [00:49:00] So a scheme to do (inaudible).
Nick translation continuing on page 107, nick translating of hemoglobin-4.
And so on page 109, have Leslie — that’s Leslie [Furlong?]. She was later on the first person to get any DNA sequence that we ever had. It will come up in due course, I’m sure. [00:50:00] This was, nick translating, almost certainly, on page 109, Harvey Faber’s cDNA, that’ll be made from messenger RNA, Harvey Faber 9-133. Nick translational buffer — no, it’s not. That’s not correct. This is just the buffer. And anticipating a result which we haven’t got to. [00:51:00] So the hemoglobin phage 3aHb4 DNA, 1.5μg is then labeled, and is available as O-109, 32P, CH Hb4, and has about 6,000 counts per minute, per microliter, and Leslie Furlong has Charon 3A at 18,000 counts per minute, per microliter.
And so, [Hae-Hae?] digests on page 111, [00:52:00] with some tests of the [Hae-Hae?] digests on page 111, with at least labels suggesting that those are hemoglobin-4, and the delta-lac, and [Hae-Hae?], mainly delta-lac and hemoglobin-4 [00:53:00] giving two different bands, and not very clear, although the comment is, “The product is adequate.”
I’m considering using other vectors, instead of phage vectors, considering using pBR322, every reason to pick it, because EK2 and — [00:54:00] chi1776, and can be used semi — EK2 being the degree of safety, to pass the conditions allowable for cloning DNA, human DNA or older DNA, EK2 and chi1776, very weak bacteria made by Roy Curtis, so considering using pBR322 instead of bacteriophage with the conclusion that, “Probably will find it best to stay with lambda for all genomic clones, and use the plasmid only for probe preparation once cut-out-able pieces [00:55:00] are in use.” Quite a wise decision.
The reason for using lambda phage at all was Fred Blattner’s argument that you get a large amount of DNA, and you didn’t have to have the bacterial cells able to replicate, so that clones that might prevent growth of a bacteria would not likely be poisonous enough to prevent lysis — lytic growth of bacteriophage, which is done in less time than one round of phage growth — of bacterial growth, that’s to say. So the phage growth could be done without the bacteria having to divide, whereas the clones of plasmids required amplifying bacteria, and much more difficult [00:56:00] in terms of [poison?] sequences. In fact, pBR322 had a sequence which made it difficult to grow under some circumstances, so we made a vector later which was deleted, pBR322.
Remake of the [Hae-Hae?] piece, 32P of HaeIII, on page 115, part of the scheme. And following two pages, “Abandon, abandon.”
No good on page 119. Page 121, a comment on why it didn’t work, “So I poisoned the reaction with isotope yesterday.” [00:57:00]
Another scheme still being considered on page 123, all getting a good probe. And this sort of problem, being that on page 125, “Looks as if HaeIII cuts the clone anyway, and Bam makes a small piece, too small to stick, or find. On and on with the same type of problem, 127.
Autoradiograph on page 129 again, of the [Hae-Hae?], of labeled delta-lac [00:58:00] and hemoglobin Hb4, and the cloned Hb4.
Page 131, an important technical comment. Not — doesn’t look very important. Although, a good result, saved this probe, page 131. cDNA hemoglobin (inaudible) test on [Gal?] (inaudible) a bit more. [00:59:00]
(inaudible) gel, Fred suggested electrophoresing it out of the acrylamide. But, tried it and, said no [01:00:00] detectable transfer by hybridization 48 hours auto-[rad?] of the label.
More serious gel transfer experiment on page 135. The results show that transfer occurred on both attempts, 15 minutes and 45 minutes, so if the gel (inaudible) improved, the method will work. This is electrophoresis, which some people had used, to electrophorese the sample out of the gel, and getting transfer other than Southern-type transfer. [01:01:00]
And continuing with the same general problem, trying to get a good (inaudible). [01:02:00] Grunstein-Hogness protocol on page 148, a replica plating onto our Millipore filter, and growth colonies, etc. And, so they’re hybridizing directly from a column as — [01:03:00]
An experiment on Monday, April 4th, using papers to draw colonies on them. Interesting, because I always — a collection of Whatman filter papers, and here is a comment that, on page 151, Monday, April 4, “70mm filter papers are available from stores, Whatman 1, 2, 40, 42, and 50,” and then the following page, [01:04:00] a photograph of the same colonies. And it counts for something else that we’ve noticed earlier on. That filter papers earlier encountered really should be on this page. The photograph is there, the actual images themselves, not quite the same, as the photograph, but these filter papers probably belong here anyway, and the two filter papers [01:05:00] with colonies stained blue. So the colonies are allowed to go on the filter paper, and then stained, and then various things can be done to transfer them to, or hybridize to the colonies. So I’m going to put these unknown filter papers in this page, it’s right at the end, on page 152 of this book. And so we end Book o.