Developing GFP as a Biological Marker
Transcript of Part 1: Developing GFP as a Biological Marker
00:00:17.15 Hello, I am Martin Chalfie. I am in the department of 00:00:19.14 Biological Sciences at Columbia University. 00:00:22.08 Green fluorescent protein and other fluorescent proteins 00:00:24.27 are now used quite commonly in biology, 00:00:29.11 particularly cell biology. What I would like to do 00:00:31.29 in the next few minutes is tell you a little bit about 00:00:34.27 my lab's involvement in the origin of using GFP as a biological marker. 00:00:40.11 I did not start out being interested in developing a new biological marker. 00:00:45.28 Actually, my lab has been interested for many years in mechanosensation, 00:00:50.13 particularly trying to understand the genes 00:00:52.16 that are needed for touch sensitivity. 00:00:54.28 But in the late 1980s, my lab had cloned several genes 00:01:01.10 that were needed for touch sensitivity in the nematode, Caenorhabditis elegans 00:01:05.13 and we wanted to know where those genes were being expressed. 00:01:09.17 And particularly, we wanted to know 00:01:11.13 which of our cells were, in the end, touch sensing cells. 00:01:16.05 At this time, there were several methods that could 00:01:18.25 and did help us to answer these questions. 00:01:21.14 These included antibody staining, 00:01:25.17 markers such as beta-galactosidase, 00:01:28.10 in situ hybridization. All of these could answer the question: 00:01:33.22 what cells activated what particular gene? 00:01:37.01 But there were problems with these methods. 00:01:39.27 First, they actually required a fair amount of work. We had to prepare 00:01:44.17 the tissues, and that meant killing them, fixing them, permeabilizing them so that 00:01:49.13 we could get the antibody or the substrate for beta-galactosidase, 00:01:53.22 or the probe for in situ hybridization into the tissue. 00:01:56.24 But the main problem I think was the fact that although we could tell 00:02:02.02 at the moment we prepared the sample that genes were being expressed there, 00:02:07.02 we couldn't look at more than just this static picture. 00:02:11.15 If we wanted to look over time, we'd have to do preparations 00:02:14.24 for each time point. Then on April 25th, 1989, 00:02:19.07 I heard a seminar that really changed my life. 00:02:21.14 It was a talk being given by Paul Brehm, who at the time, 00:02:25.06 was a neurobiologist at Tufts University. 00:02:28.06 And in the introduction of his talk, he started talking 00:02:32.02 about the work of this man, Osamu Shimomura, 00:02:33.16 and the work he had done in isolating proteins 00:02:37.29 from the jellyfish, Aequorea victoria. 00:02:40.14 In the early 1960s, Shimomura had discovered 00:02:46.00 the protein aequorin, which was a bioluminescent protein. 00:02:49.06 It produced light. 00:02:50.23 What he found was that when aequorin and calcium were together, 00:02:56.03 they would produce light, and this was the problem he was trying to address. 00:02:59.27 But he had one slight difficulty, and that was that the light 00:03:05.04 that was produced by this reaction 00:03:06.21 was blue. The jellyfish, however, gave off a light that was green. 00:03:13.05 And he knew that there had to be something else. 00:03:15.18 So he went back to his protein fractions to look at 00:03:16.23 what was there, and he found that there was another protein, 00:03:21.11 that when added to aequorin and calcium, now gave green light. 00:03:26.26 And this protein, he called the green protein. 00:03:29.10 We now call it GFP or green fluorescent protein. 00:03:31.24 And because I was working on a transparent animal, 00:03:37.00 Caenorhabditis elegans, I was looking at gene expression, 00:03:41.20 I suddenly realized that this protein, GFP, 00:03:45.10 would make a terrific marker for our experiments. 00:03:49.00 And I spent the whole rest of the seminar fantasizing about 00:03:52.11 this work. I actually don't remember what the rest of the seminar was about. 00:03:56.14 The next day, I got in touch with this man, Douglas Prasher, 00:04:00.08 and found out that he was in the process 00:04:03.17 of cloning the cDNA for GFP from the jellyfish. 00:04:07.20 We had a wonderful conversation, culminating in a decision 00:04:11.26 to collaborate to see whether GFP could be used in other organisms. 00:04:16.03 A few years later, I was lucky enough to get a very talented graduate student 00:04:21.03 into my lab, Ghia Euskirchen, 00:04:24.18 who came to the lab having just finished a Master's degree in chemical engineering 00:04:29.04 working on fluorescence. She came into the lab, 00:04:31.21 to do a rotation, and this was the project that she was given. 00:04:34.23 We got back in touch with Douglas who had cloned the gene by this time; 00:04:39.07 he sent it to us, and Ghia proceeded to see whether this would work in, at first, E. coli. 00:04:46.13 When she did these experiments, there was hanging over us a problem. 00:04:51.11 And that problem was that we weren't really sure that it was going to work 00:04:55.22 because of what was known about the GFP molecule. 00:04:58.12 By this time investigators had found that GFP had a rather unusual 00:05:03.19 post-translational modification. The peptide backbone in GFP 00:05:10.04 cyclized, and the making of this five-membered ring was real mystery. 00:05:16.03 People speculated that it might take one, two, or maybe even more 00:05:19.29 converting enzymes to make the mature protein from the translated product. 00:05:26.00 So it wasn't a sure thing that this was going to work. 00:05:29.07 Nonetheless, Ghia did try the experiment, 00:05:32.14 and to our great joy and excitement, it worked very nicely. 00:05:36.10 This is a page from her lab notebook that day, 00:05:39.19 approximately one month after she entered graduate school 00:05:43.05 where she found strongly fluorescent E. coli. 00:05:46.10 and took this picture. She was also lucky in another sense, and that is how 00:05:50.22 we did the particular experiment. 00:05:51.25 Douglas's clone had a cDNA from the jellyfish that was really the coding sequence 00:06:03.01 which I have diagrammed here in green, 00:06:04.24 and non-coding sequence, which is in red. 00:06:08.12 He had gotten this as an EcoRI restriction fragment. 00:06:13.13 Now we could have simply taken that restriction fragment. 00:06:17.18 and used that in our experiments, but I decided 00:06:20.09 that I didn't want to get extra stuff, 00:06:22.12 and that it would be better to use PCR to amplify 00:06:25.15 just the coding region. I wasn't sure what the rest was going to be, 00:06:31.01 but this has turned out to be a very lucky choice, 00:06:33.25 because it turns out that at least three other labs that I know of 00:06:37.10 tried the experiment, but used the EcoRI fragment. 00:06:41.01 And when they did that, they never got any fluorescence, 00:06:44.08 so there's something in these extra sequences 00:06:47.00 that seemed to interfere with the production of a fluorescent protein. 00:06:49.28 We didn't do that. We didn't have that problem. 00:06:52.29 Within one month we knew that this was going to work. 00:06:55.09 We were very excited, and it now meant that we had to do... 00:06:58.18 we were going to go off and do many other things. 00:07:01.13 We put this into worms, and we decided that this was time now 00:07:05.06 to publish the material, and we had a little bit of difficulty with publishing the paper. 00:07:09.25 So we sent the paper in to be reviewed 00:07:13.18 in Science, and the first thing we found was that the editors 00:07:17.11 were not going to send it out to reviewers 00:07:20.20 for consideration because they did not like the title. 00:07:24.14 The original title was "Green Fluorescent Protein: A New Marker for Gene Expression" 00:07:28.13 and the editors told me that all the papers in the journal were new 00:07:32.12 and novel and so we couldn't use new in the title and would I change the title. 00:07:36.25 I was a little bit miffed at this, so in retaliation, 00:07:42.05 the title I gave them was considerably longer, 00:07:45.02 "The Aequorea victoria Green Fluorescent Protein 00:07:47.23 Needs No Exogenously-Added Component to Produce 00:07:50.15 A Fluorescent Product in Prokaryotic and Eukaryotic Cells". 00:07:54.08 This is essentially the entire paper. 00:07:57.11 The paper got reviewed; the reviewers liked it. 00:08:00.15 And then the copy editor got in touch with me and said, 00:08:02.25 "You know your title is a little long, 00:08:04.21 could you possibly shorten it?" 00:08:07.08 And I said, "I think I can do that," 00:08:08.20 and I changed it so the final title is 00:08:11.03 "Green Fluorescent Protein as a Marker for Gene Expression". 00:08:14.10 This was not the end of our troubles though, 00:08:16.16 we had sent in this picture that you see here 00:08:18.18 that eventually made it to the cover of Science 00:08:20.17 and I was very proud of this picture because what it shows 00:08:25.10 is a growing nerve cell in a living larva of the animal. 00:08:29.28 And I wanted to point out the fact that GFP could be used in living tissue. 00:08:35.11 The art editor, the cover editor called me up and said, 00:08:39.14 that they really liked the picture, that they wanted to use it on the cover 00:08:42.15 but there was one problem. And that problem was that they 00:08:46.03 never like to use the color green on the cover, 00:08:48.17 and would I consider changing the color 00:08:51.10 of the picture. I said, "no, I really wanted it to be green," 00:08:55.00 and fortunately they kept it that way. 00:08:57.16 The final problem that we had in publication 00:08:59.18 was by this time we had already given out samples 00:09:02.26 of the plasmid to people to try for themselves. 00:09:04.24 and we were getting wonderful reports back from people saying that 00:09:07.18 it had worked. So I wanted to put that as personal communication into our paper, 00:09:13.27 and the people that we asked were all uniformly very generous, 00:09:19.22 and one person, however, asked for some additional 00:09:24.06 considerations, specifically, and you probably can't read it in this letter, 00:09:30.27 but the letter asked that I prepare coffee every Saturday morning 00:09:36.03 for two months, prepare a special French dinner, 00:09:39.21 and take out the garbage nightly for the next month, 00:09:42.16 these were requirements set out by my wife, Tulle Hazelrigg. 00:09:45.01 But I really wanted to use her work, and although we still debate about whether 00:09:49.23 I have actually paid up on this. 00:09:51.23 The work she did was really wonderful, 00:09:54.17 and it was published a few months later. 00:09:56.19 But she's the person that made the first protein fusion with GFP, 00:10:01.22 so she was able to show that GFP could be linked to other proteins, 00:10:05.25 and that one could follow those other proteins, check their localization, 00:10:08.24 but also see their movements within cells and tissues. 00:10:13.21 A beautiful example of the use of GFP protein fusions can be seen 00:10:18.06 in this movie by Rosalind Silverman-Gavrila 00:10:21.04 of the nuclear divisions in an embryo 00:10:24.16 of Drosophila. I have taken this from the cell image library 00:10:27.28 of the American Society for Cell Biology. 00:10:30.05 It's on their website. It's a really beautiful picture 00:10:34.06 showing, in this time-lapse film, cell division in which you can see the spindle labeled 00:10:39.21 and you can see the spindle forming and then dissolving and forming once again. 00:10:44.26 GFP, other fluorescent proteins, and their derivatives, 00:10:49.18 have been used for thousands upon thousands of experiments 00:10:53.06 throughout biology. Their small size and inheritability 00:10:57.03 provide a dynamic and essentially a non-invasive 00:11:00.12 means of following biological processes in living cells. 00:11:05.17 We've already learned much, and we will continue 00:11:07.16 to learn more by using these molecules in the future. 00:11:10.29 I want to close with some more general 00:11:13.11 lessons that I take from the story of the development of GFP. 00:11:17.02 First, many discoveries are accidental. 00:11:22.29 Certainly the discovery of GFP is one of those cases. 00:11:26.29 Shimomura was not looking for a fluorescent protein; 00:11:31.04 he was looking for a bioluminescent protein. 00:11:32.28 But his experiments led him to this rather wonderful molecule. 00:11:36.25 Second, scientific progress is cumulative. 00:11:42.00 It is not the product of just Shimomura or myself or Douglas Prasher, 00:11:48.05 or Tulle Hazelrigg, or many other people- 00:11:51.05 Roger Tsien, who developed the first molecules with different emission colors, 00:11:58.17 who also developed the first FRET based 00:12:00.26 molecular monitors using the fluorescent proteins, 00:12:04.29 or the Lukyanovs or Mikhail Matz who discovered the first red fluorescent protein 00:12:10.03 by looking at corals. And really the thousands of people 00:12:14.20 who have added to the usefulness of GFP. 00:12:17.09 Third, I think this is a good example of saying that all life should be discovered 00:12:23.01 We should not simply be working with model organisms. 00:12:26.05 I think we can now all be very grateful that Shimomura 00:12:29.21 was interested in a fundamental biological question, 00:12:32.17 that had nothing to do with human biology or human health, 00:12:35.27 but how it is that some organisms can produce light. 00:12:40.04 And it was that curiosity that led him to work on the jellyfish, 00:12:44.05 and then l ed him to discover GFP. 00:12:47.19 And finally, the GFP story is a good example of how essential basic research is. 00:12:57.05 When Shimomura discovered GFP, 00:13:01.02 I don't think anyone would have imagined the usefulness 00:13:05.26 that it would have for other discoveries and basic research 00:13:08.05 throughout the biological sciences, or for investigations into human health. 00:13:13.24 It's been used to study HIV, AIDS, inherited diseases. 00:13:17.25 Or that it would be useful in biotechnology 00:13:21.03 The usefulness comes from other people and what they've done, 00:13:24.09 but it all starts with that basic research that was so necessary. 00:13:28.14 So what I want to do is close with one of my favorite quotes about basic research. 00:13:33.09 It comes from Robert R. Wilson who is the physicist 00:13:37.05 that was the first head of the Fermilab, 00:13:39.18 the particle accelerator in Batavia, Illinois. 00:13:41.22 He was asked in 1969 to go before a congressional committee to 00:13:47.17 justify why they should pay for this rather enormous science experiment. 00:13:53.03 He was asked many times to try to describe the security benefits, 00:14:01.15 the benefits for national defense that would come from this accelerator, 00:14:07.11 from what was going to be learned there. 00:14:08.27 And every time he was asked he said no, and then he finally 00:14:11.28 said that the accelerator had "only to do 00:14:15.15 with the respect with which we regard one another. 00:14:18.10 The dignity of men, our love of culture. 00:14:20.21 It has to do with whether we are good painters, 00:14:23.14 good sculptors, great poets. 00:14:25.22 I mean all the things we really venerate in our country 00:14:28.24 and are patriotic about. It has nothing to do directly with defending 00:14:33.03 the country except to make it worth defending." Thank you.