NOT all the fruit flies in Gerry Rubin鈥檚 lab see as clearly as they should. 鈥淛ust to look at the fly from the outside it looks normal,鈥� says Rubin, 鈥渂ut it鈥檚 colour-blind.鈥� With other flies, however, the problem is more obvious. 鈥淧eople use the word 鈥榬ough鈥�,鈥� Rubin says, describing flies with eyes covered in misshapen lumps rather than the normal, even array of facets.
Rubin鈥檚 group at the University of California at Berkeley has found that both the colour-blind and rough-eyed flies suffer from the same deep-seated biological flaws 鈥� genetic defects that disrupt the way biochemical information flows along a 鈥渃ommand chain鈥� of molecules within some of the flies鈥� cells. The chain starts at the surfaces of cells, finishes at their genetic hub, the nucleus, and functions as one of the many biochemical lifelines that allow cells to respond to the world outside their membranes. In Rubin鈥檚 case, the research is about far more than how well flies can see. The same command chain has turned up in organisms ranging from nematode worms to mice, and when it goes wrong in human cells it leaves them vulnerable to cancer. Nor is Rubin鈥檚 team the only one intent on cracking the biochemical secrets of the chain. With so much at stake, research groups worldwide have spent much of the past decade working on the problem, eavesdropping on the proteins and messenger molecules that make the chain work. The result is a milestone in the history of cell biology. For the first time, textbooks can publish a map showing how biochemical messages penetrate cell membranes and are passed, from protein to protein, to genes in the nucleus. The big question now is whether this map can inspire new approaches to treating cancer.
The answer could depend on one particular component of the chain, a protein called Ras that has become something of a superstar in research into chemical communication within cells. Genes that make Ras proteins first came to light in the 1960s, in viruses that cause cancers in rats. We now know the reason for those cancers: the Ras proteins produced by such viruses enter the command chains inside cells and put out false information. How they are able to do so became clear in about 1980, when biologists found that humans and other animals carry their own ras genes and produce their own versions of the Ras protein. The viruses were simply using corrupted versions of these genes to wreak havoc in the cells they infected.
Advertisement
It soon became clear that the Ras protein produced by humans is not entirely innocent. In the early 1980s researchers discovered that mutations in human ras genes contribute to up to one third of all cases of cancer. When ras genes were later found in everything from plants to single-celled organisms such as yeast, the race to discover their function became even more urgent.
The first clue emerged in 1984 when several groups decoded the DNA sequences of ras genes and discovered that the protein they code for is similar to a group called the G proteins. 鈥淭hat was the first biochemical landmark,鈥� says Frank McCormick, a cancer researcher and vice-president of Onyx Pharmaceuticals, a biotechnology company in Richmond, California. 鈥淕 proteins were already known and it showed the sort of thing that Ras might do.鈥�
G proteins, so called because they link up to a molecule called guanosine, are part of a cell鈥檚 internal signalling system. They float in the cell membrane where they ferry messages from receptors to other molecules in command chains. The trick is that G proteins only do so when the guanosine they are attached to is in an active chemical form. The guanosine functions like a switch. By the mid-1980s, researchers had discovered that Ras proteins also float in the cell membrane and that they too seemed to use the same switch mechanism. 鈥淩as is a very simple protein; it doesn鈥檛 do anything other than be active or inactive,鈥� says Alan Hall of University College London, who studies proteins similar to Ras.
But although switches are simple, the circuits they are wired into can be very complex. At first Ras seemed to be an on/off switch dedicated to controlling cell division. Rapidly dividing cancer cells often carry mutated Ras proteins that are abnormally active. But in 1985 the picture started to look more complicated. A pair of researchers in New York and a team in Japan showed that Ras has a wider role in controlling cell behaviour.
Both groups had been working with cultured rat cells designated PC12. Under the right conditions, these ball-shaped cells keep on dividing indefinitely. But if they are exposed to nerve growth factor, a hormone, they stop dividing and change shape to look like nerve cells. The two groups supplied dividing PC12 cells with extra copies of active Ras protein. Had Ras been dedicated to promoting cell division, this extra dose of the protein would have speeded up cell division. Instead, and to the researchers鈥� surprise, cells with extra Ras stopped dividing and began to change into nerve cells. The message was clear: flipping Ras to the 鈥渙n鈥� position didn鈥檛 just make cells divide. In some circumstances it could stop cells dividing and make them more specialised. In other words, it seemed that Ras might play a pivotal part in controlling cell differentiation.
Confirmation of this came at the beginning of the 1990s. In rapid succession, groups of biologists studying how animals such as fruit flies and nematode worms develop from embryos showed that Ras influences cell differentiation in whole animals as well as in cultured cells. In fruit flies, for example, Rubin鈥檚 group has shown how Ras helps to control the development of the creature鈥檚 multifaceted compound eyes. Beneath each facet lies a cluster of eight light-sensitive cells, one of which cannot develop without Ras. This particular cell only develops correctly if the right steps are performed in a complex dance of messenger molecules. In the first step, the cell is stimulated by a signalling protein produced by a neighbour cell. As a result, a signal is sent cascading down a command chain inside the cell. The signal is finally delivered to the cell nucleus where it switches on genes that change the unspecialised cell into a colour-sensitive rhabdomere. The involvement of Ras in the command chain was clear: when researchers disabled the protein, the cell developed into a light-focusing crystalline cone, leaving the fly colour-blind. Conversely, when they artificially stimulated Ras in cells that normally develop into crystalline cones, the cells grew into rhabdomeres, leaving the fly with rough eyes.
Since then, researchers like Rubin have been trying to figure out just how a protein that looks like a simple on/off switch manages to be so influential. Which command chains are wired into the Ras switch? How is the switch turned on and off? And so on. Fortunately, since biologists from several different fields now have a stake in understanding Ras, results have piled in with dizzying speed.
The first, crude picture of how Ras works dates back to 1986 when Dennis Stacey and his colleagues at the Roche Research Center in Nutley, New Jersey, investigated the effects of jamming Ras in the 鈥渙ff鈥� position. They injected cancer cells with antibodies against Ras, so blocking it. On a microscope slide the cells looked bloated, like balls or sausages. But once Ras had been switched off, most cells become temporarily normal 鈥� flattened rather than rounded.
As Stacey鈥檚 team also discovered, jamming Ras in the 鈥渙ff鈥� position has knock-on effects on other proteins involved in transmitting commands in cells, especially ones attached to the membrane. It is as if Ras forms a turnstile through which information from the membrane proteins must pass before reaching proteins in the cell cytoplasm. Lock Ras shut and the command can鈥檛 get through.
Information highway
So far so good. But attempts to map the flow of commands farther away from Ras were frustratingly unsuccessful. It was the early 1990s before researchers began to find the key to the problem. Proteins carry information in their shape and chemical structure: change either of these and you change the information carried by the protein 鈥� and by extension the signal it will transmit to its partners in a command chain. Nature鈥檚 favourite way of changing the shape or chemical structure of a protein is to stick a phosphate group onto its surface or to remove one 鈥� tasks for which it has evolved a fleet of specialised enzymes, or kinases. For proteins involved in command chains, a phosphate tag often acts as an on/off switch. And the chains themselves often consist of a series of kinases, each acting in sequence to stick a phosphate group onto the next member in the series. It turned out that flipping the Ras switch to the on position kick-starts just such a kinase chain.
At one end of this chain lies a kinase that springs into action, switching on genes in the nucleus, when cells react to substances that trigger division. First identified in the mid 1980s, this enzyme is now known to be a key accomplice of Ras. 鈥淚t suddenly all started to come together about a couple of years ago,鈥� says Julian Downward who studies the Ras signalling network at the Imperial Cancer Research Fund laboratories in London.
The kinase delivers the all important message to the nucleus, but it doesn鈥檛 take its orders from Ras directly. There are intermediaries. Fortunately, however, like workmen digging from both ends of a tunnel, biologists studying the kinase and those working on Ras met in the middle. It soon became apparent that the kinase receives its orders from another kinase, a go-between. And it became apparent that Ras delivers its orders to a protein called Raf, which swims in the cell cytoplasm.
From there the last links in the chain fell quickly into place. Researchers found that Raf 鈥渢alks鈥� to the go-between kinase. 鈥淭hen the main thing was how does Ras activate Raf,鈥� says McCormick. 鈥淚n 1993 four groups showed that Ras binds directly to Raf.鈥� One theory is that all Ras needs to do to switch on Raf is to bring it from the cytoplasm to the cell membrane.
Fitting Ras and the kinase into the same picture meant that researchers had mapped an entire chain of command from the cell鈥檚 outer membrane all the way to genes in the nucleus. Only one puzzle remained. Ras floats on the inside of the cell membrane. So how does information from outside the cell cross the membrane to switch Ras on or off?
Back in 1987, McCormick 鈥� then at the Department of Molecular Biology of Cetus Corporation in Emeryville, California 鈥� and Meg Trahey showed that normal cells like to turn off Ras unless stimulated to do otherwise. We now know that there are several ways to flip Ras back on, each beginning with receptors sticking out from the cell membrane. The shortest route is probably via the receptor for the hormone epidermal growth factor. When this hormone docks into the receptor, it stimulates it to add phosphate groups to its own tail 鈥� the section of the receptor poking into the cell. This makes the tail a target for proteins that act as a kind of double-sided sticky tape, attaching themselves to the tail on one side and pulling proteins out of the cytoplasm on the other. The proteins fished out in this way happen to be the very ones that can switch on Ras.
The overall picture is perhaps less like a chain and more like a funnel. A wide range of receptors on the surfaces of cells channel commands to Ras, but thereafter all commands flow down the same pipeline to the nucleus (see Diagram). This begs the question why signals from different receptors produce different effects on cells if all the commands end up in the same place.FIG-mg4301.jpg
One clue comes from looking more closely at the role of Ras in PC12 cells. Nerve growth factor makes the cells grow into nerve-like cells, and epidermal growth factor makes them divide rapidly. Yet in both cases, Ras is involved in transmitting commands to the cell nuclei. How so? Some point to the length of time Ras is switched on. Alan Hall explains: 鈥淣GF gives sustained Ras activation and EGF gives transient activation. It turns out that if you over-express EGF to give sustained activation you get differentiation rather than growth.鈥�
Another possibility is that Ras triggers a range of effects depending on the type of cell and its stage of development. For example, in fruit flies, Ras is not just involved in eye development. It also marks out which parts of the fly embryo become the head and tail. The commands sent from Ras to the nucleus produce different results because different sets of genes are primed for action in the two cell types concerned.
Now that the first cell command network has been mapped out, research into Ras and its fellow messenger molecules seems set for an explosive diversification. Close relatives of these molecules are cropping up in odd places, suggesting an ancient evolutionary origin. In brewer鈥檚 yeast, for example, command chains leading to different versions of the kinase control the building of cell walls, each cell鈥檚 mating type (in effect its sex) and how much dissolved salts in its cytoplasm. In cells from mice, a similar command chain involving Ras and the kinase appears to handle signals produced by stress.
But the biggest front of all will be against cancer. Mutations that jam Ras in the 鈥渙n鈥� position contribute to the uncontrolled division of cancer cells. Could medical researchers intervene? One approach may be to use gene therapy to switch off ras genes in tumour cells. For example, Jack Roth and his colleagues at the University of Texas MD Anderson Cancer Center, in Houston, are attempting to treat lung cancer by injecting tumours with a virus that carries an 鈥渁ntisense鈥� ras gene. In theory, this back-to-front version of ras prevents the normal gene from making Ras protein. Roth has yet to publish his results, but in earlier tests on mice the approach stopped the growth of tumours.
Hard nut to crack
Other groups are following a more conventional route and searching for chemicals that will inactivate Ras. 鈥淎ll the drug companies are working on this,鈥� says Michael Gould, a cancer researcher at the University of Wisconsin in Madison. What they are finding is that understanding exactly how potential drugs kill cancer cells is a hard nut to crack. For example, Gould and his colleagues showed in the early 1980s that a chemical from orange peel inhibits several cancers in rats. Recently, the team found that the chemical, limonene, stops cells from installing Ras in their outer membranes. This breaks the flow of commands through Ras, giving what appears to be a clear-cut explanation for why limonene may inhibit tumour growth. But the true picture is more complex, says Gould. Limonene acts by preventing the cell from attaching a molecular 鈥渁ddress鈥� to newly created Ras proteins, which normally sends them to the cell membrane. But this same address tag is used to direct many other proteins from their place of synthesis to the cell membrane. And which of these proteins is limonene鈥檚 most important lever against cancer is unclear.
Fortunately, this does not mean that they have to give up the approach. Limonene is currently undergoing preliminary trials in Britain by Charles Coombes, professor of medical oncolology at Charing Cross Hospital, London, on patients with breast cancer. And in the US, researchers hope to begin trials of a more potent form of the chemical.
It is hardly surprising that researchers cannot trace every effect of a molecule like limonen. Mapping all the command chains through which cells and their genes 鈥渟ense鈥� the outside world will be a daunting task. Biologists have already found about 50 proteins that together form a 鈥淩as superfamily鈥�. This host of Ras lookalikes helps to control processes ranging from the structure of a cell鈥檚 internal skeleton to the identity of any chemicals it might secrete. Ras appears to have evolved so long ago that it has become integrated into almost every aspect of a cell鈥檚 behaviour.
