The jellyfish revolution
This essay was written by Donald Bell and was first published in the 2013 Mill Hill Essays.
About 50 years ago a group of researchers from Princeton University took a field trip to Friday Harbor, a small bay town in Washington state, to find samples from a local jellyfish population. This expedition was the beginning of a journey that would revolutionise and influence many threads of biological research. It led to a step forward in technology and tools that have facilitated new discoveries.
The researchers, led by Frank Johnson, were interested in the biology of luminescence. This is a complex chemical reaction in which energy is released in the form of light. Their target was to collect specimens of Aequorea victoria, a slightly dowdy crystal jellyfish (figure 1). Aequorea is about 10cm in diameter and has a ring of cells around its umbrella that glow green when agitated. The researchers set out to identify and isolate the substance responsible for producing this light in what turned out to be an epic piece of basic research that had a far-reaching influence way beyond just understanding how the jellyfish glowed.
The isolation procedure was a grand and rather messy job involving fishing with nets for the abundant jellyfish, then taking thousands of them back to the lab to extract the light-producing component. A production line was created to process the thousands of jellyfish, cutting and sieving extracts. They yielded only a tiny amount of glowing extract but this was enough for Osamu Shimomura to be able to establish that the glowing material was a protein with two components. The first they called aequorin, a protein that yields blue light due to a chemical reaction (luminescence). The second was a protein that glows bright green when exposed to higher energy light. In the jellyfish it is the aequorin that donates its dim blue light energy to the other protein thereby causing it to fluoresce green. They unimaginatively, but accurately, named this second protein green fluorescent protein (GFP).
The key to GFP’s potential in research is its ability to convert blue light into green light on its own without the need for any extra factors. Martin Chalfie saw its potential and was next to take up the baton to establish whether GFP could work in organisms other than the jellyfish it was isolated from. Working initially in bacteria and then in a tiny roundworm, Caenorhabditis elegans, he used genetic cloning and transfection techniques to make pieces of DNA describing how a cell might make GFP and then introduced this DNA into roundworms. By doing this he showed that, given the correct instructions, specific types of worm cells (touch receptor neurons) could indeed make their own GFP, that glowed green, without any apparent effect on the normal function of the cells. Martin Chalfies’ wife, Tulle Hazelrigg, then demonstrated that GFP could be genetically joined onto specific proteins in the cell. Joining GFP to naturally occurring proteins did not seem to affect the way they behaved and so could be used as a beacon reporting (by emitting a green glow) precisely where in the cell they resided. This set of experiments demonstrated the potential of GFP for directly visualising proteins and cells in biological systems, paving the way for an explosion of diverse applications.
The next step was to determine the shape of GFP in order to gain clues as to how it works. The molecular structure was determined and shown to be a small barrel-like form supporting a tiny light-producing complex, or chromophore, at its centre (figure 2). This knowledge helped to explain how GFP functions. Researchers from Roger Tsien’s lab in San Diego used protein engineering techniques as they attempted to understand and modify the light-absorbing and light-emitting properties of the native form of GFP. Tiny alterations were found that could dramatically increase brightness, produce light of different colours, and make the protein more soluble. Solubility is a useful property for research in a cellular environment as the cell is mostly water. Soon it was possible to produce a variety of proteins glowing with different colours across the spectrum from blue to yellow and so researchers had the potential to visualise a range of colours in a single organism or cell.
Despite great efforts it proved difficult to extend the colour palette into the red end of the spectrum, by further mutation of GFP. A breakthrough into the red finally came after finding GFP-like proteins in non-luminescent corals and anemones. The expedition to find these, headed by Sergey Lukyanov, was rather less laborious than that to find GFP. The raw material for the research was harvested, or rather purchased, from aquarium shops in Moscow. The first red fluorescent protein, DsRed, was rapidly isolated from the coral Discosoma species but it proved not to be ideal for biological research. Roger Tsien then stepped in again with his lab systematically isolating mutants of DsRed, refining this red form into one far more suitable as a biological reporter by reducing its size and toxicity and increasing its brightness. Although much improved, further slight structural changes in this new version resulted in a range of red coloured fluorescent proteins, with names like tomato, cherry and plum reflecting their shade of fluorescence. Along with GFP variants biologists now had a palette of beacons covering the visible spectrum.
Collectively Shimomura, Chalfie and Tsien received the Nobel Prize in Chemistry in 2008 for the discovery and development of green fluorescent protein. The story is far from over, however: research continues to find novel fluorescent proteins. To date more than 150 fluorescent variants have been described and 45,000 scientific papers refer to this little protein as the key to their content.
As the palette of fluorescent proteins has evolved there have been efforts made to develop the techniques and equipment used to visualise them. Applications exploiting these techniques have also blossomed, providing tools for the biologist to explore the function and mechanics of living things.
Light microscopy has long been employed to detect and measure the fluorescence emitted from a range of chemical dyes. Whilst the light given off from these synthetic compounds tends to be much brighter than that of the fluorescent proteins, they are generally difficult to introduce into living cells and this is where the advantage of the fluorescent proteins lies. Many researchers use fluorescent proteins as an easily visualised, inert marker of living proteins or cells in culture, tissues or even in whole animals. They have become invaluable detection tools, especially in connection with the wide variety of genetic manipulation methods available. Researchers continue to develop ever more ingenious methods using the fluorescent proteins, pushing the boundaries of biology, physics and chemistry in their efforts to eke out further information from their system of interest.
What follows is a small summary of some of the more exciting and potentially influential techniques made possible by the discovery and development of the fluorescent proteins. Key to many of these applications is the ability to microscopically examine cell behaviour and function by a method that does not interfere with that behaviour and function.
One of the earliest observations of the characteristics of GFP was made when it was first purified and crystallised. A crystal is a highly ordered structure which, in this case, means that all the barrel-like GFP molecules pack next to one another in the same orientation and form long needle-like crystals. Fluorescence from this crystal is six times greater when the light waves used to excite it are polarised parallel to the long axes of this needle (as opposed to perpendicular). This is called anisotropic or direction-dependent behaviour. Hence the difference in light produced depends on the orientation and can be used as a measure of changes in the shape of proteins in response to stimuli.
The outer surface of the cell is a membrane made predominantly from fat molecules but it is studded with proteins that perform a variety of important functions. Fusing GFP to these membrane proteins means that the little barrel is held at a constant angle with respect to the membrane, just like in the crystal. If the membrane protein joined to GFP changes its shape in response to a stimulus then the angle of the attached GFP will also change and this tiny difference can be detected by polarised light microscopy. In this way GFP can be used to measure the response of a cell to its environment. Using the same anisotropic activity and detection techniques GFP can also be used as a miniature thermometer within a cell. Monitoring temperature change is a useful way to follow the activity of numerous intricate cellular processes.
Growth and development of all organisms happens in a highly ordered manner. At the heart of this process is the cell division cycle which describes the typical steps that a cell follows in order to reproduce. Cell division occurs at different times and speeds, and coordination of cell division is critical for the correct formation or maintenance of organs and tissues. A failure to precisely regulate cell division can lead to a variety of diseases such as cancer. It is therefore of great interest to biologists to be able to observe this process in living tissue and to understand how growth is influenced by the dynamics of cell division. Until recently this has been difficult to achieve but thanks to the technical wizardry of a group of Japanese scientists, who developed a couple of fluorescent proteins isolated from coral, the task has become far easier. They developed a pair of beacons causing cells to fluoresce in different colours depending on where the cell is in its division cycle. During the first phase a red fluorescent protein is produced which then gives way to a second, green form, in the next phase. As cells cycle through these steps they change from red to green via an amber colour defining a period when both proteins are active in the cell (figure 4). This progression of colours provides a readily detectable (and rather beautiful) signal that can be used to visualise, for the first time, the intricacies of cell cycle dynamics in living tissue. This tool is likely to significantly influence research into cancer, development and ageing.
The first application of GFP in an animal was to visualise a small group of nerve cells in C. elegans. This was an impressive achievement but biologists soon acknowledged that if they labelled all cells of a particular type they could simply confuse the story in a complex system such as the brain. In 2007, in answer to this problem, researchers created a multicoloured genetic reporter in mice, and coined the name ‘Brainbow’ (figure 5). In its simplest form this system can be used to label a set of brain cells with one of four different fluorescent proteins. With a slight technical modification, however, cells can each individually express multiple fluorescent proteins. In this way individual cells may be discriminated from each other by appearing to be one of a hundred different hues depending upon the combination of the fluorescent proteins they express. This is analogous to the rainbow of colours achieved by mixing different hues of paint. With this a nod was given to the work of the Italian scientist Camillo Golgi who, more than a hundred years previously, had developed a silver stain that could label a limited number of brain cells at random. The ability to pick out a single cell in a field of many others allowed individual neuron types to be catalogued in exquisite detail and for this Golgi, along with Spanish neuroscientist Santiago Ramon y Cajal, received the Nobel Prize for Physiology or Medicine in 1906. Brainbow may also be used to catalogue the connections and pathways that neurons take but it has the advantage that the type of cell labelled can be controlled more easily and that the fluorescence arises in a living tissue. Brainbow has since been developed further and similar systems have been introduced for research in fruit flies and fish.
Another useful technique allows researchers to visualise functional neural networks in a living brain, but on a single cell level. Here the response to a stimulus of an individual nerve cell can be measured electrically. Other cells pass information about the stimulus to the cell being measured and they can be identified by a clever application of a modified rabies virus. This virus is only passed to cells actively communicating with the measured one which means that only these, amongst the millions of others, make GFP and so can be imaged to produce a map of connections in the context of the whole brain. It is hoped that studying fluorescent cells in mice as well as other simpler animals will pave the way to a greater understanding of the human brain and perhaps to a map of all its trillions of connections.
The multicolour fluorescence methods described above provide powerful techniques with which to investigate a plethora of biological questions but they fundamentally rely upon the introduction and visualisation of distinct, single coloured, proteins. There is a separate class of fluorescent proteins, however, that can be individually coaxed to change their optical properties upon exposure to specific wavelengths of light by way of a very slight structural alteration. These are the photo-activatable or photo-switchable fluorescent proteins. They may change from a dark state to a bright one or from one colour to another and can be used in ways that offer exquisite precision to seeing proteins or organelles within cells.
Fluorescently tagging one protein type in a cell yields a good overall picture of where those proteins are but little information about their movement or turnover. Using a photo-switchable version allows the overall structure to be seen in one colour in live cells. Applying a precisely targeted laser to a small region of these labels can change their fluorescence colour and allows the behavior of small subsets of proteins or even single molecules to be observed. Perhaps one of the most exciting developments involving this class of fluorescent proteins is one which is beginning to break one of the longest held barriers in optical physics.
A hundred and forty years ago the German physicist Ernst Abbe was the first to recognise that the resolution of optical instruments is fundamentally limited by the diffraction of light. Abbe described this diffraction limit as roughly equating to half the wavelength of the light used to visualise it. This means that the light emitted by a single GFP molecule is spread out and is detected as a fuzzy blob appearing to be about a hundred times bigger than it actually is. When a set of proteins in a cell are tagged with GFP and imaged all at once the resulting picture is an amalgamation of many thousands of these fuzzy blobs. This is analogous to pointillist pictures, painted by late nineteenth century artists, where detailed scenes are depicted by painting many thousands of different coloured spots on the same canvas. For many biologists these fuzzy blob pictures are not a problem. The diameter of an average cell is about forty times larger than the diffraction limit and even collections of these fuzzy blobs, merging with one another, can give an excellent impression of many subcellular structures. However, to discern the finest detail improvements known as super resolution techniques are needed.
Photo-activatable or photo-switchable fluorescent proteins are used to produce images with a resolution far exceeding that possible with standard light microscopy. To attain this resolution it is necessary to acquire many thousands of images of a single cell. A different set of only a few fluorescent proteins is switched on each time, so each picture shows a limited number of sparsely separated fuzzy blobs. Merging all these pictures together at this stage would produce a standard, diffraction-limited image of the cell. For super resolution this collection of images is first analysed and the centre of each blob is mathematically estimated and then plotted as a smaller spot, representing the source of the fluorescence, on a single new image. As these smaller spots accumulate so a super resolution image of that cell is created. Currently, this is an extremely active field of research. Advances require innovation in computing power and analysis, improvements in speed and sensitivity of cameras and of course a bigger range of brighter fluorescent proteins.
It has been a long journey since the first isolation of GFP from Aequorea victoria but its availability as a tool has revolutionised many areas of biological research. This modest little protein has had a profound effect upon basic research into neuroscience, development, immunity and cancer. Further expansion of the palette of colours alongside other useful properties will no doubt continue, with the next great step being clinical applications. Meanwhile GFP and Aequorea victoria still hold some secrets. The biggest of these is why this jellyfish, and other marine creatures, have the ability to fluoresce at all. It was long since assumed that it was just a way of signaling but it is becoming apparent that this might not be its primary role. Clues are given by the fact that some of the brightest fluorescent protein variants have never been discovered in nature and also that some animals seem to require very similar proteins that do not fluoresce at all. To this end we are grateful to the researchers such as Osamu Shimomura for pursuing basic scientific questions. After his retirement he moved his laboratory to his home and to this day the 84 year old still investigates his jellyfish. In a telephone interview after he received the Nobel Prize he explained his motivation:
… I don’t do my research for application or any benefit. I just do my research to understand why jellyfish luminesce.