You are on page 1of 1

research highlights


Hidden code in the protein code
Apparently redundant codons may not be Researchers led by Eran Segal at the
redundant after all. Weizmann Institute of Science took a
The instructions for building proteins are computational approach to find evidence
given in triplets of the four nucleotides. This for additional codes in the protein code
yields 64 codons, each dictating either the (Itzkovitz et al., 2010). They defined all pos-
addition of a particular amino acid or the sible 6-base-pair or 7-base-pair sequences
end of protein synthesis. With only 20 amino in real genomes. They then tallied up the
acids to choose from, several different codons frequencies of occurrence of the defined
code for the same protein residue, but recent short sequences in randomly generated cod-
research shows that so-called synonymous ing sequences and in actual genomes cod-
codons may provide subtly different infor- ing for the same proteins. They looked for
mation to a cell’s gene-reading machinery. enrichment or depletions across phyla and
Researchers led by Anna Kashina at the found pronounced differences between real
University of Pennsylvania found that dif- and randomized coding regions, particu-
ferent codons for the same amino acid may larly for bacteria. The technique identified
© 2010 Nature America, Inc. All rights reserved.

affect how quickly mRNA transcripts are previously known patterns, such as enrich-
translated, and that this pace can influence ment for microRNA target sites in eukary-
post-translational modifications (Zhang et otic genomes as well as for codons contrib-
al., 2010). Despite being highly homologous, uting to the three-dimensional shape of
the mammalian cytoskeletal proteins beta- mRNA molecules, and also suggested that
and gamma-actin contain notably different whereas some ‘overlapping codes’ have been
post-translational modifications: though evolutionarily conserved, others have been
both proteins are actually post-translation- enriched in different phyla.
ally arginylated, only arginylated beta-actin Segal and colleagues have made their ran-
persists in the cell. This difference is essential domization software available and believe
for each protein’s function. that it could be useful to other scientists who
To investigate whether synonymous codons also want to search for additional informa-
might have a role in how arginylated forms tion: perhaps, for example, identifying pat-
persist, Kashina and colleagues swapped the terns in specific families of genes and assess-
synonymous codons between the genes for ing whether these patterns may represent
beta- and gamma-actin and found that the binding sites of yet-uncharacterized tran-
patterns of post-translational modification scription factors or RNA-binding proteins.
switched as well. Next, they examined trans- A more practical application, Segal says,
lation rates for the wild-type forms of each is in synthetic biology: “Designing the ‘opti-
protein and found that gamma-actin accu- mal’ synthetic gene could become a fasci-
mulated more slowly. Computational analysis nating combinatorial problem that could
suggested that differences between the folded integrate the constraints that evolution had
mRNA structures might cause differences converged upon and that we believe our
in translation speed. When the researchers study reveals.” Kashina also believes that
added an antibiotic that slowed down trans- scientists will eventually be able to pick
lation rates, accumulation of arginylated the right codons to engineer desired post-
actin slowed dramatically. Subsequent work translational modifications. “The idea that
indicated that N-arginylated proteins may, if nucleotides encode more than the protein
translated slowly, be subjected to ubiquitina- primary structure is one of those very sim-
tion, a post-translational modification that ple and very fundamental mechanisms that
targets proteins for destruction. have been out there on the surface all the
Thus, these apparently synonymous time,” says Kashina. The code within a code
codons can help explain why some arginy- surely has more secrets to reveal.
lated proteins but not others accumulate in Monya Baker
cells. “One of the bigger implications of our RESEARCH PAPERS
work is that post-translational modifica- Itzkovitz, S. et al. Overlapping codes within protein-
tions are actually encoded in the mRNA,” coding sequences. Genome Res. advance online
publication (14 September 2010).
says Kashina. “Coding sequence can define a
Zhang, F. et al. Differential arginylation of actin
protein’s translation rate, metabolic fate and isoforms is regulated by coding sequence-dependent
post-translational regulation.” degradation. Science 329, 1534–1537 (2010).

874 | VOL.7 NO.11 | NOVEMBER 2010 | nature methods