INSIGHTS | PERSPECTIVES
By Boris Zinshteyn and Rachel Green
In the age of computational biology, it is easy to envision the genetic code as a set of immutable instructions that the cell follows without exception. The recent discovery of ciliate (1, 2) and trypanoso- matid (3) species in which all three stop
codons, which normally act to terminate
protein translation by the ribosome, encode
amino acids instead is a reminder that decoding the information in messenger RNA
(mRNA) depends on molecular factors that
we do not entirely understand. In these organisms, stop codons specify amino acids
by default, and termination of mRNA translation only occurs in close proximity to the
polyadenylate [poly(A)] tail. How do these
species differentiate “true” stop codons from
identical ones that encode amino acids? The
answers to this puzzle may provide insights
into translation termination and gene regulation in all eukaryotes.
Since the discovery of the “standard”
genetic code, small deviations have been
discovered in various organisms and organ-
elles (4). Ciliated protists have a particular
propensity for reassigning one or two of the
three standard stop codons (UAA, UAG, and
UGA) to encode amino acids. Recent tran-
scriptome sequencing (all the mRNA mole-
cules expressed) revealed that in two ciliates,
Condylostoma magnum and the unclassified
Parduzcia sp., all three stop codons encode
amino acids as well as signal translation
termination (1, 2). This is an altogether pre-
viously unknown paradigm containing a po-
tentially disastrous ambiguity. The genomes
of these species encode cognate transfer
RNAs (tRNAs), which are complementary to
the stop codons, and the balance between re-
cruiting a tRNA to incorporate an amino acid
or a release factor to terminate translation is
resolved by the surrounding context of each
particular stop codon. The sequence struc-
ture of ciliate transcripts provides a clue as to
how this balance is accomplished. The unusu-
ally short 3′ untranslated regions (UTRs) be-
tween the “true” stop codon and the poly(A)
tail (1, 2) of these ciliates suggest a model in
which cognate tRNAs decode stop codons by
default, but in close proximity to a poly(A)
tail, release factors outcompete tRNAs (see
the figure). This hypothesis is supported by
in vitro data showing that poly(A)-binding
protein (PABP) (5) stimulates translation ter-
mination by recruiting release factors.
The idea of context-dependent termina-
tion sheds light on another ciliate coding
oddity. More than 11% of genes in Euplotes
octocarinatus contain an in-frame stop co-
don followed by a frameshift in the coding
sequence (6, 7). These stop codons occur in a
“slippery sequence” [AAA UA(A/G), stop co-
don underlined], which allows the tRNA in
the ribosomal P site (which is decoding the
AAA codon) to pair with the codon in the +1
frame (AAU). At these positions, ribosomal
frameshifting efficiently outcompetes termi-
nation (8), much like cognate tRNA recogni-
tion outcompetes termination in C. magnum
and Parducia sp. This suggests that position-
dependent stop codon recognition exists in
all of these species and may be a general fea-
ture of ciliate biology.
The context-dependence of stop codon
recognition in ciliates bears a striking resemblance to the nonsense-mediated mRNA
decay (NMD) pathway in humans and other
eukaryotes. NMD degrades mRNAs with premature stop codons in early exons or long
3′ UTRs, but the mechanism by which premature stop codons are differentiated from
normal ones is unclear. Several studies have
argued that proximity to the poly(A) tail is
one measure by which the translational machinery identifies a normal stop codon (9,
10), and that decay is the default outcome of
stop codon recognition in the absence of suppressing factors (11), much as amino acid incorporation appears to be the default for stop
codons in C. magnum.
The distinct stop-codon recognition mechanism in these ciliates raises many questions
about termination and NMD throughout the
eukaryotic lineage. We still do not understand how proximity to the poly(A) tail is
determined, particularly when RNA secondary structure and protein binding can affect
the conformation of a mRNA in three-dimensional space. Are release factors only found
in proximity to the poly(A) tail, or are they
somehow activated by its presence? Is the system tightly tuned to prevent the production
of truncated and extended proteins, or are
ciliates constantly cleaning up the mistakes
of a sloppy translational machinery? What is
the role of conserved NMD factors in ciliates
with ambiguous stop codon usage? Answering these questions will require both detailed
biochemical studies and genome-wide analyses and will likely yield fundamental insights
about gene expression in all eukaryotes. j
1. E. C. Swart, V. Serra, G. Petroni, M. Nowacki, Cell 166, 691
2. S. M. Heaphy et al., Mol. Biol. Evol. 33, 2885 (2016).
3. K.Záhonová et al., Curr. Biol.26,2364(2016).
4. S. Sengupta, P. G. Higgs, J. Mol. Evol. 80, 229 (2015).
5. A. Ivanov et al ., Nucl. Acids Res.44, 7766 (2016).
6. R. Wang et al., Sci. Rep. 6, 21139 (2016).
7. L. A. Klobutcher, P. J. Farabaugh, Cell 111, 763 (2002).
8. A.V.Lobanov et al., Nat. Struct. Mol. Biol. 10.1038/
9. N. Amrani etal .,Nature 432, 112 (2004).
10. A. B. Eberle et al., PLOS Biol. 6, e92 (2008).
11. Z.Ge,B.L.Quek, K.L.Beemon,J.R.Hogg, eLife 5,e11155
When stop makes sense
RNA sequence context matters in the termination
of protein translation
Department of Molecular Biology and Genetics, Johns
Hopkins University School of Medicine, Baltimore, MD, USA.
Inefcient termination Efcient termination
Amino acid incorporation Frameshifting
factor Polypeptide chain
In some ciliate species, termination of mRNA translation is only efficient in close proximity to the poly-A tail. C.
magnum has cognate tRNAs for the standard stop codons, so they are translated like normal sense codons. In E.
octarinatus, which lacks cognate tRNAs, premature stops stall the ribosome, which is resolved by +1 frameshifting.