The genetics of a marine protozoan may overturn one of the long-held tenets of protein synthesis. According to conventional wisdom, the genetic code is unambiguous: each DNA triplet, or codon, corresponds to a single amino acid. But a study in this week's Science reports that in the wee waterborn creature Euplotes crassus, a single codon can code for two different amino acids, even within the same gene.

This two-pronged coding could be a universal feature of all domains of life, said Vadim Gladyshev, a biochemist at the University of Nebraska in Lincoln, who led the study. "Euplotes is just one example, but the implications are that this situation [one codon coding for two amino acids] might exist in other organisms," he told The Scientist. "If Euplotes can use a codon for targeted and specific insertion of two amino acids into different positions in a protein, why not other organisms, too?"

The universal genetic code generally encodes the same 20 amino acids and three stop signals in all organisms, although a small number of genes in many organisms, including humans, contain two additional amino acids -- selenocysteine and pyrrolysine. What's more, some species and organelles use alternative genetic codes, including many protists that rewire certain triplets that normally code for stop codons to code for specific amino acids instead. Euplotes crassus is a tiny, marine ciliate -- about the size of the width of a human hair -- that does both of these things simultaneously, the new study reports. In E. crassus, the UGA triplet, which is a stop codon in most eukaryotes, codes for both cysteine and selenocysteine.


Differential interference contrast micrograph of E. crassus.
Gladyshev and his colleagues analyzed the E. crassus genome sequence and found eight genes for proteins that incorporate selenocysteine, four of which contained multiple UGA codons. By expressing these genes in human embryonic kidney cells and using mass spectrometry on the E. crassus proteins, Gladyshev's team showed that UGA differentially coded for either cysteine or selenocysteine. They also found that which amino acid was incorporated into the protein depended on the proximity of the UGA codon to a specific structure in the 3' untranslated region called the "Sec insertion sequence," or SECIS.

John Atkins, a molecular biologist at the University of Utah in Salt Lake City and University College Cork in Ireland, who was not involved in the study, said that dual coding might be a "specialist situation," rather than a general phenomenon. Still, the observation that the SECIS element was differentially affecting the ribosome at various UGA triplets has major mechanistic implications about the fine-tuned control of ribosome-based protein synthesis, he said.

"It shows that there must be very specialized folding of RNA in the 3' untranslated region to talk to the ribosome," he told The Scientist. "The redefinition [caused by SECIS] is not of the ribosome as a whole; rather, the signal is determining a localized definition."

But the paper's findings might not be all that novel, cautioned Yale University microbiologist Dieter Söll. He noted that in some Candida species, the CUG codon is translated as both leucine and serine, even in the same gene, albeit by a single ambiguous tRNA rather than two separate tRNAs as Gladyshev's team found in E. crassus. "This already showed that you can have the same codon in one gene [encode] two amino acids," Söll told The Scientist. "Really, it simply shows that nature has more than one way of doing the same thing."

Gladyshev disagrees. "It's really a completely different situation," he said. In Candida, the tRNA doesn't discriminate between amino acids, he noted. Rather, it randomly inserts either serine or leucine, "whereas in our case there is a specific insertion of one amino acid or the other, depending on the presence and availability of the RNA element."

Söll also noticed a technical error in the paper. In Figure 2, the putative cysteine tRNA that Gladyshev identified from the genome sequence had a cytosine residue at a terminal "charging site." But all other known cysteine tRNAs have uracil instead, Söll noted. Gladyshev said the figure contained a mistake: the true nucleotide was, indeed, a uracil. "I would like to thank... the expert who noticed this," he wrote in an email. The sequence he submitted to Genbank was correct, he noted, and he has contacted Science's editors to issue a correction.

Gladyshev is now searching for other organisms that display dual coding. He is also investigating mammalian selenocysteine-containing genes to investigate whether the position of the SECIS element affects the incorporation of selenocysteine at different codon sites.

Image courtesy of Lawrence Klobutcher


Related stories:

Nirenberg's genetic code chart, 1961-66
[June 2007]

Extending the genetic code
[15th August 2003]

And then there were 22
[7th June 2002]