Today’s scientists can watch in real time as glowing cancer cells migrate through tissue. They can examine a colorful network of neurons that trace the neural pathways across a mouse’s brain and even identify the individual sperm involved in the fertilization of fruit flies.
Each of these breakthroughs can be attributed to a single protein hidden for millions of years along the coasts of the Pacific Ocean that became the go ahead for new biological research.
Since long before scientists harnessed it as a research tool, the protein has illuminated the bioluminescent jellyfish Aequorea victoria. As it drifts through the otherwise dark ocean waters, A. victoria emits a gentle glow that brightens its immediate surroundings. Nobody knows for certain why the jellyfish produce the unusual fluorescent protein, although some scientists suggest that the light is meant to help in defense against predators.
Enter Osamu Shimomura, a marine biologist at Princeton. While he was studying the bioluminescent properties of the jellyfish, Shimomura discovered a strange protein. At first, the revolutionary molecule—later called green fluorescent protein, or GFP for short—was only mentioned in a footnote in Shimomura’s 1962 paper on the jellyfish’s bioluminescence.
The protein was largely ignored by the scientific community for 30 years.
However, when Martin Chalfie, a Columbia biology professor, first heard about the protein in the introduction to a 1989 seminar on bioluminescent organisms at Columbia, he was immediately struck by the possibilities. Chalfie was investigating touch sensation in a small, clear roundworm called C. elegans and was particularly interested in locating where in the worm a specific set of genes were expressed.
Chalfie realized that if he could make the C. elegans express GFP alongside the genes he wanted to target, the fluorescence could act as a visual marker shining through the transparent organism, indicating where the genes were expressed.
“It wasn’t a great leap of genius to come up with the idea that if I put this protein into the worms that I was studying, I would be able to have that gene not make what it’s supposed to make but make GFP instead,” Chalfie said. “Wherever GFP was, I could shine blue light on it and see it because I would see green light coming back. And so I got very excited about this possibility.”
However, the research applications of GFP were limited without the sequence of its gene, so Chalfie made no further progress in using GFP to map gene expression in C. elegans.
After the seminar, Chalfie connected with molecular biologist Douglas Prasher at the Woods Hole Oceanographic Institution, who was trying to identify the sequence. With no leads, the two lost contact with each other for the next three years.
In 1992, Prasher published a groundbreaking paper in which he detailed the sequence of the GFP gene and demonstrated his success in cloning the gene. Chalfie and Prasher immediately reconnected.
With the GFP gene provided by Prasher, Chalfie, along with several other scientists, were able to move forward with developing techniques to create GFP-expressing cells.
Chalfie was able to insert the gene into E. coli bacteria, creating bacterial cells that expressed GFP and fluoresced under ultraviolet light.
For their work on GFP, Shimomura, Chalfie, and biochemist Roger Tsien, who engineered variations of GFP with different colors, shared the Nobel Prize in Chemistry in 2008.
Chalfie received the prize for being the first to successfully genetically modify cells so that they express GFP, paving the way for tagging cells with the fluorescent protein.
Getting to GFP
All living organisms—from Chalfie’s E. coli bacteria to Chalfie himself—rely on DNA as a blueprint for existence.
Each gene in DNA tells the cell how to make a protein, which are essential in the function of the cell. Whenever a cell needs to make a specific protein, it locates the required gene in the genetic library of its DNA and makes a copy of that gene. DNA lookalikes, known as RNA, then direct the formation of proteins.
GFP is one such protein encoded by a gene in A. victoria DNA. Proteins like GFP are made of strings of 22 different organic molecules called amino acids, each of which has different properties, and the order of these amino acids give the protein its shape and function—in GFP, that function is the fluorescent glow seen in A. victoria.
To use GFP in an organism, Chalfie had to hack into the bacteria’s existing genome and insert the gene for GFP, a process called transformation.
He wasn’t alone—after Prasher published the gene sequence that encoded GFP, many scientists conducted experiments attempting to introduce the gene into new cells. However, the complexity of the protein made adding a functional gene sequence extremely difficult.
Many scientists assumed that the complicated structure of GFP would require additional enzymes to fold the amino acid chain into the functional form of the protein. Because of this, researchers attributed failed bacterial transformation to a missing enzyme rather than to an error in the process of gene insertion.
“They said, ‘Well, it’s gonna require an enzyme, maybe two, maybe ten, who knows? We don’t know how it’s made, but it looks complicated to us, it’s gonna need something else,’” Chalfie said. “So if you do the experiment and you don’t get any result … you could conclude, ‘I failed—I don’t know what I’m doing,’ or you could conclude, ‘Aah—it needs something else, I’d better go find that something else.’”
Expectations for an additional protein-folding component led other scientists astray when the first experiments failed to create bacteria that produced GFP.
Before scientists insert a gene into bacteria, they must make many copies of it—enough to expose all the microscopic cells to the molecule and allow for the poor efficiency at which cells uptake the DNA molecules.
At the time, there were two methods of replicating the DNA that differed in accuracy. The more accurate technique of copying a sequence of DNA was to insert a large segment of DNA into the genome of bacteria and allow the bacteria to divide, thereby replicating the DNA sequence which could be extracted later.
Despite the accuracy of this method, the extraction process was far less precise, and scientists ended up with a sequences of DNA that contained the GFP gene along with extra segments.
“It turns out … that those extra bits prevent the thing from being made,” Chalfie said. “So those people that wanted to be careful, they would have put, they in fact did put it, this in their systems and said, ‘We don’t have anything, there’s no fluorescence. Therefore, there must have been a converting enzyme.’”
Instead, Chalfie used a less accurate method of DNA replication called the polymerase chain reaction, which replicates sequences of DNA inside a test tube. Without the proofreading machinery of the cell, far more errors are made, but Chalfie was able to specify exactly what sequence he wanted to copy.
He decided that 100 percent accuracy didn’t matter because he didn’t need every E. coli cell to express GFP—he just needed enough to show that the gene works.
“I didn’t care whether some of [the DNA] was bad—most of it was bad—all but just a little bit was bad,” Chalfie said. “We put it into bacteria, and did all the bacteria become fluorescent? Of course not! But enough of them did to make it really obvious that this worked.”
Chalfie’s successful experiment showed that no extra enzymes were required to fold GFP into its active state.
Four months after Chalfie established GFP as a practical cell marker, Columbia biology professor Tulle Hazelrigg developed techniques to use GFP to tag proteins. By creating hybrid functional proteins with an extra GFP component, Hazelrigg could track the movement of proteins throughout organisms in the same way that Chalfie could identify cells.
“She did the second, really important, experiment,” Chalfie said about Hazelrigg, who is now his wife. “The protein has GFP as a lantern on the back of it, and so wherever the protein goes, it drags along this light with it so that you can see where it is.”
A new way of seeing
The ability to trace the movements of molecules in living tissue became a green light for new biological research, allowing scientists to investigate pathways that couldn’t be seen previously.
Before GFP was developed, organisms had to be killed and treated so that dyes could enter the cells and bind to the relevant proteins.
Because the dyes had to be applied to dead tissue, scientists had only a static view of life, a snapshot of the cells and proteins present at the death of the organism.
In stark contrast, GFP’s glow traces cells and proteins in still living organisms, so that scientists can watch life’s processes as they happen.The thousands of scientists who use GFP in these experiments have a debt of gratitude to a lab on the 10th floor of Fairchild Hall.
Chalfie’s experiment demonstrated for the first time that GFP was not toxic to cells, and that cells expressing GFP can continue to carry out life processes while producing the protein. He also demonstrated that the gene alone was sufficient to induce a cell to produce functional GFP. Thus, no other molecules needed to be added to an experiment in order for GFP to be used.
“[The experiment] said this is going to work without any other component being added. This is a one component system,” Chalfie said. “This protein can be activated, that is produce green light, simply by shining blue light on it. You don’t have to kill the organism … all we have to do is shine the light.”
Impact and innovation
After these foundational studies were conducted—many at Columbia—GFP became an essential molecule in biological research. Tsien’s further engineering of the molecule has allowed scientists to tag different cells in the same organism with different colored fluorescent proteins so that their activities can be compared.
According to Chalfie, it is estimated that by 2014 there were 160,000 scientific papers that used GFP as part of their research.
Scientists who use GFP consistently cite its ability to visualize cells and molecules within a living organism as a defining feature.
Peter Canoll, a professor of pathology and cell biology, studies the formation of brain tumors and uses GFP identify cancerous cells and map their spread through the brain.
“We use the GFP to track where the tumor cells go and how to distinguish them from the other cells in the environment,” Canoll said. “We can use it to actually image the cells in living brain tissue and watch the cells move and extend their processes and crawl through the living brain tissue.”
Chalfie continues to use GFP in his research on the C. elegans roundworm, investigating the development of touch-sensing neurons as well as the mechanistic process behind sensing touch.
Felix Qiaochu Jin, CC ’16, has worked in Chalfie’s lab since his sophomore year, studying the factors that lead to the differentiation of six specific types of neurons that detect gentle touch. Jin, who was recently chosen as valedictorian, uses GFP to help identify his six target neurons among all the other cells in the worm.
“Say you want to know where the touch neurons are in the worm—to do that you would simply take one of the touch genes, which are only found in those six neurons, and then take its promoter, which is the piece of DNA that drives the gene’s expression, and put that in front of a GFP protein sequence,” Jin said. “Now you have that same promoter driving the GFP expression, and so all six of these neurons will make a lot of GFP, but none of the other cells will.”
Some applications of GFP have been especially innovative. One experiment aimed to developing bacteria which could detect land mines leaking TNT by expressing GFP in the presence of the explosive.
“If you could make a bacteria that could detect TNT, you could spray the ground and come back at night with an ultraviolet lamp and see if you could detect the fluorescence which would have meant there was leaking TNT around,” Chalfie said. “Since land mines hurt a lot of innocent people, the idea of being able to have a nice detector of the land mines would be wonderful.”
Unfortunately, the experiment was not successful because bacteria could not be created that had a 100 percent success rate at detecting TNT.
“This is the one case you want it to work every single time. You don’t care if you get a false positive, but you absolutely do not want to get a false negative,” Chalfie explained. “But I like the example because it is an example of people being creative of thinking about useful tools and doing something I would never have thought of.”
Though Chalfie says he couldn’t imagine every area in which GFP would be applied, he did understand the monumental nature of the protein when he first conceived the idea of tagging cells in that 1989 seminar.
“We were going to be able to see things that we normally would not be able to see,” Chalfie said. “In fact, I fantasized so much at that seminar when I first heard about it, I didn’t listen to the seminar. I have no idea what they said. But I certainly was excited about the possibilities … I knew it was going to be very important.”