FOR most people, Candida albicans literally evokes irritation. Often called “thrush”, this microbe is to blame for 85 per cent of vaginal yeast infections and many throat infections, sending a good many people scurrying to the doctor’s for a prescription.
But in the world of science, this loathsome creature is now inspiring awe. That’s because about 270 million years ago, its ancestor did what most biologists have considered nearly impossible. It changed its genetic code. Like jamming a Windows disk into a Macintosh machine, this abrupt shift in the operating system that deciphers genes should have spelled disaster. “You wouldn’t expect a cell to fix itself after that,” says Manuel Santos of the University of Aveiro in Portugal. “This was a textbook example of something you’d expect to be eliminated by natural selection.”
It might be time to send those biology textbooks back for a rewrite. Santos has found that C. albicans is not alone – it and other code-changers are alive and thriving out there. They are a slap in the face of one of the most basic tenets of biology: the unshakable stability and ubiquity of the universal genetic code.
Advertisement
Dogma has it that the genetic code is the lingua franca of all living things, and has been so for nearly four billion years. Put another way, it means that every cell is reading from the same genetic recipe book – the same piece of DNA code won’t mean “cat” to one cell and “dog” to another. The code is the perfect lexicon for life’s instruction manual and to change it in all but the most trivial ways is to spell disaster.
But the mystery of C. albicans and similar organisms turning up in a variety of places is prompting some experts to suggest that these long-held beliefs about the code are myths. Instead, they say, the code is neither fixed nor ubiquitous. It may have undergone major overhauls in the past, could still be changing, and is more diverse than previously suspected.
If these heresies hold up, they have broad implications. The most fundamental is the way a flexible code rewrites the story of life. “The creative era of the code is supposed to be history,” says Dieter Söll of Yale University, who was part of the team that first cracked the genetic code back in the 1950s and 60s. That lab’s leader, H. Gobind Khorana, shared a Nobel prize for the work in 1968. “But there is no reason to think its evolution is done,” says Söll. “If we have this discussion in a million years, we might be talking about a very different code.”
There are practical implications of a changing code, too. As genome projects amass data on far-flung species, the lack of a universal code could make the interpretations harder. On the other hand, a more malleable code suggests that genetic engineers might be able to do far more with cells than was imagined only a few years ago.
When the structure of DNA was discovered 50 years ago, it was found to comprise a backbone studded with a long, linear string of four chemical bases or “letters”, adenine (A), cytosine (C), guanine (G) and thymine (T). The obvious next question was what those letters meant. The answer, worked out over the next two decades, was that the letters formed sequences of triplets, or codons, that together provided the instructions for making proteins. This “language” was dubbed the genetic code.
On paper, the code is simple. Its entire vocabulary is the 64 possible codons. Sixty-one of these codons encode the 20 amino acids from which all proteins are made (some amino acids have several codons), while three “stop” codons mark the termination of protein production.
Here’s how the cellular clockwork of the code ticks. When a DNA gene is turned on, enzymes produce an RNA version called a messenger RNA. RNA has a similar structure to DNA although the base uracil (U) is used instead of T. The mRNA’s linear sequence of codons is read by molecules called transfer RNAs, or tRNAs. Each of the different tRNAs binds to a specific codon and carries a specific amino acid, which it then glues into place on the growing protein chain.
This process is so fundamental to every cell’s biology that intuition suggests that changing the code should be disastrous. To understand why, consider a fairly common genetic mistake called a point mutation, which changes only one single codon in a gene. As a result, the encoded protein contains just one incorrect amino acid, and yet that minor change may weaken or completely inactivate the protein, so harming or killing the cell.
In contrast, a sudden change in the code would alter the meaning of a codon across the board. Every gene containing that codon would end up making a protein with at least one incorrect amino acid. Thousands of proteins would be corrupted and the chance of a total meltdown of the cell would skyrocket.
So it came as no real surprise when biologists discovered that organisms as diverse as bacteria, tobacco plants and humans used exactly the same code. By the late 1960s the conclusion was that all living things were straitjacketed into using the one true code, and no deviation was possible.
Minor exceptions did subsequently emerge, but biologists were quick to write these outliers off, or cite them as exceptions that proved the rule. For example, mitochondria, the energy factories of our cells, often have different genetic codes. But since each one only contains a few dozen genes, and there are hundreds of mitochondria in each cell, corrupting a codon in one wouldn’t necessarily cause a meltdown. They simply aren’t under the same evolutionary pressure as complex genomes with thousands of genes.
Some microbes were also found to have redeployed one of the three “stop” codons to encode one of the 20 amino acids. But this isn’t as disruptive as altering the meaning of an amino acid’s codon, and these organisms didn’t father a new group of species with an alternative code. So these renegades weren’t seen so much as great trendsetters, but rather as misfits that had turned off down dead ends.
In fact, the near universality of the code implied it had remained stable since ancient times. That’s because biologists believe that all modern life traces back to what is known as the “last universal common ancestor”, or LUCA, primordial cells at the very base of the tree of life. If the code has been evolving since the time of LUCA, you would expect to find some organisms with more primitive versions of the code – ones with only 19 amino acids, say.
Yet these “missing link” organisms have never been found. The code appears to have been fixed by the time of LUCA. And if it has remained unchanged for four billion years, there is no reason to think it could ever change.
But in the past few years, cracks in the theory have started to appear. One leading iconoclast is Michael Syvanen at the University of California, Davis. He questions whether the standard set of 20 amino acids really were in place in LUCA, or whether some have been added since. His scepticism comes from his recent genetic analysis of biochemical pathways for synthesising two amino acids, arginine and tryptophan, which suggest these pathways evolved no more than two billion years ago (Trends in Genetics, vol 18, p 245).
The fact that this is about half the age of the code itself has been noted before – and generally written off as a statistical error. But more and more data has supported this date and Syvanen argues it should be taken seriously. “This means the code underwent major change billions of years after cells were crawling around the planet,” he says. But if these amino acids were latecomers, how could they have got into the code of every living organism?
Syvanen suspects that both amino acids gave the organisms that acquired them huge selective advantages. Arginine has a striking biochemical property – the amino acid can tightly bind RNA without breaking it. He thinks tryptophan, too, might have unique properties, although what they are isn’t clear yet. Syvanen reasons that both amino acids were such important innovations that they spread from species to species in what is called “horizontal gene transfer”, in the same way a killer application is quickly adopted by web users.
And those that missed the code upgrade simply couldn’t compete, so they disappeared from the face of the Earth. If that’s the case, the current code might not be life’s original software, but version 2.0, or perhaps 3.0.
Supposing Syvanen is right and the code experimented a little in its youth, is it still capable of acquiring new tricks today? It seems unlikely: at some point, the argument goes, any advantage gained by using an extra amino acid would be balanced by the problems a shifting code would cause for its host cell.
And until recently, most experts argued there wasn’t any special advantage a new amino acid could bring. Life already populates the chilly Antarctic, the crushing pressure and boiling temperatures of deep-sea vents, and almost every available niche in between. The standard repertoire of 20 amino acids seemed to be enough to build any kind of protein you could wish for.
Ripe for an upgrade
But this theory was undermined in 1986 when researchers discovered a 21st natural amino acid. Called selenocysteine, it proved to be genetically encoded in a vast number of organisms – including humans. The codon for selenocysteine is in fact the “stop” codon UGA in the universal code, marked with a nearby looping structure in the RNA. August Böck at the University of Munich, Germany and his colleagues discovered that cells have elaborate machinery to bring the amino acid’s tRNA to the mRNA and insert it at the special UGAs. All this fuss seems worthwhile, says Böck, because selenocysteine gives enzymes a new property, making them up to 400 times as powerful as ones that contain the chemically similar amino acid cysteine.
Even so, selenocysteine is generally regarded as having secondary status in the pantheon of amino acids. Böck says he understands why biologists are reluctant to expand the list of amino acids to 21. “For two years, I didn’t believe this result myself,” he says. “Finding a new amino acid in the code wasn’t supposed to happen.”
But just last year, it happened again. This time Joseph Krzycki at Ohio State University in Columbus and his colleagues discovered an amino acid he dubbed pyrrolysine, in a protein belonging to the microbe Methanosarcina barkeri, which dwells in stagnant water (Science, vol 296, p 1462). Like selenocysteine, this amino acid appears to use a specially marked stop codon – this time UAG. And it also appears to bestow a special chemical power on proteins: when the amino acid’s synthetic pathway is damaged the microbe can no longer digest methylamines, a common source of energy in the environments it occupies.
So are selenocysteine and pyrrolysine the next big innovations in the code? Probably not. Selenocysteine is common in microbes and animals, but not plants, and only appears in a handful of proteins per species. Böck says that makes him think that it may not be a new addition to the code, but an old one – the original code may have had 21 amino acids, and over time selenocysteine has been pushed to the periphery. Likewise, pyrrolysine has been found in only a few methylamine-sucking species.
All the same, Krzycki points out that however rare these amino acids are, they challenge the notion of a single code. If they are present in a select group of organisms, then nature seems willing to invent regional dialects of the code to fit particular niches. And the fact that pyrrolysine escaped detection for so long suggests that other amino acids may also have been overlooked.
How many amino acids there really are in the code clearly depends on how hard it is for cells to evolve new pathways to insert them. Just a few years ago, conventional thinking suggested it should be extremely difficult. Then researchers such as chemist Peter Schultz of The Scripps Research Institute in La Jolla, California, suggested they might be able to engineer new genetically encoded amino acids into living cells. They weren’t testing any theory of code evolution, but were simply trying to make proteins with new chemical properties. Schultz had had great success at getting synthetic amino acids into proteins in the test-tube, but many doubted he could do the same thing in living organisms (żěè¶ĚĘÓƵ, 30 September 2000, p 32).
To carry off this trick the researchers would need to construct a new tRNA and a new enzyme to fuse the amino acid to it without interfering with all the similar reactions taking place in the same cell. “It was possible that cells just didn’t have the capacity to juggle more amino acids,” admits Schultz. “Having 20 balls in the air might have been the limit.” But two years ago they announced their first success at throwing a new ball into the mix. They inserted with 99 per cent accuracy a synthetic amino acid called O-methyl-L-tyrosine into a UAG stop codon in the bacterium Escherichia coli (Science vol 292, p 498).
As impressive as that is, it was only the beginning. Schultz told żěè¶ĚĘÓƵ that his team has now introduced 14 other synthetic amino acids one at a time into different E. coli strains and has been able to engineer new enzymes in as little as two weeks. His team has also used the same technique to insert new amino acids into mammalian cells and will soon tackle their first multicellular organism, the microscopic worm Caenorhabditis elegans.
Even though Schultz’s lab is renowned for its technical wizardry, he says they cannot take all the credit. “This turned out to be far easier than we thought,” he says. “We’re not bad, but we’re not competitive with God. The real question is why nature doesn’t invent new and improved amino acids all the time.”
Of course, pyrrolysine might be a case where cells did just that. But if every change in the code required the evolutionary advantage of a brilliant new innovation in protein function, and was limited to redeploying stop codons, that alone would seem to limit how much it could change.
This is why the C. albicans code change is so intriguing. It didn’t involve slotting a new amino acid into a redundant stop codon, but something more profound. The yeast has reassigned the CUG codon, the universal code for leucine, to encode another of the 20 standard amino acids, serine. When this odd change was first reported by Santos and others in the early 1990s, people suggested that some special feature of C. albicans must have allowed it to make this dramatic shift in its code. Perhaps, for instance, the yeast’s forebear simply didn’t use this codon for leucine (there are five other codons for leucine), so it was easily reassigned as serine.
But Santos has found the code is altered not only in C. albicans but in other disparate Candida species, suggesting the change happened independently more than once. And his team’s recent work reveals that it was no easy feat. By comparing the genome sequence of C. albicans with two distantly related yeasts, they concluded that accommodating the code change involved a major upheaval, involving at least half of its 7000 genes.
The analysis also shows that early on in the transformation, the CUG codon stood for both amino acids. Conventional thinking suggests this ambiguity must have been highly detrimental. Yet Santos’s analysis shows that the ambiguity was tolerated for 100 million years until the leucine-coding CUG tRNA was lost (Genome Research, vol 13, p 544). That suggests cells have a much greater capacity to deal with changes in the code than has been suspected.
More challenging to explain is what selective advantage could have driven the change. “We just don’t know what it is,” Santos says.
So could a similar change happen again? Paul Schimmel of the Scripps Research Institute, who studies the evolution of tRNA, says: “The Candida change shows the genetic code is plastic. What it’s capable of doing in another billion years is not something you or I can contemplate.”
However, evolutionary biologist Carl Woese of the University of Illinois at Urbana-Champaign says it would be a mistake to think the universal code’s days are numbered simply because we are discovering exceptions. “Nature is a curio shop full of interesting things with no general significance,” he says. “Finding exceptions isn’t a major sign that the code is changing its form.”
Patrick Keeling of the University of British Columbia in Vancouver, who has studied organisms with alternative codes, agrees. “The code still appears to be one of the most highly conserved characteristics of life,” he says. “Understanding these changes is only going to help us understand the elements in the cell that generally hold it so steady.”
Keeling admits that the real distribution of the universal code is still unclear. It may simply be luck that the organisms first studied had the same genetic code. “Our knowledge of the majority of microbes is pathetic,” he says. “We haven’t even scratched the surface.”
But there is a certain irony if some people are unwilling to accept the idea of a changing genetic code, says Syvanen. He points out that multicellular organisms didn’t exist when the arginine/tryptophan additions went global. And because of animals’ slow reproductive cycles and our specialised reproductive organs, animals are far less able to adopt a new code. “So the next time a code change sweeps the biosphere, we’re going to be left out,” he says.
What that would mean for our cohabitation with these upgraded bugs Syvanen isn’t certain. But whatever the effect, we’ve had fair notice. Be warned, humans, your operating system may soon be obsolete.
