STOP! Every protein-coding RNA needs a termination codon or translation will run right on through the poly-adenine tail, producing a run of superfluous lysines (encoded by AAA). Such defective proteins could build up in the cell, potentially leading to neurodegeneration, according to Claudio Joazeiro of The Scripps Research Institute in La Jolla, California. Fortunately, there is a system in place that identifies nonstop proteins on their way out of the ribosome and tags them for destruction. Researchers have a good handle on how bacteria deal with nonstop proteins, but to this day do not understand how eukaryotes do it. In the September 12 Nature online, Joazeiro and first author Mario Bengtson report that a homolog of the mouse ubiquitin ligase listerin forms the heart of this policing system in yeast, suggesting that the ligase might be the ticket for run-on proteins in mice and humans as well. Curiously, listerin-mutant mice have symptoms resembling motor neuron disease, and these results might just explain why.
In an ARF interview, Joazeiro suggested the protein might also be involved in human neurodegenerative disease. Other mRNA-wrangling proteins, including TDP-43 and FUS, are relevant to neurodegeneration (e.g., see ARF related news story on Kwiatkowski et al., 2009 and Vance et al., 2009), though it is not clear if any of them interact with listerin. Listerin brings up no hits in AlzGene or other neurodegenerative gene association databases.
The listerin story began with a large-scale screen of mutant mice run by the Novartis Institute for Biomedical Research. Laboratory technicians noticed that some of the animals developed paralysis as they aged. Steve Kay and others at Scripps, including Joazeiro, analyzed these animals, which they dubbed “lister” for their listing gait (Chu et al., 2009). (Now, dear reader, you can finally stop thinking “mouthwash?”) The animals showed signs of neurodegeneration that reminded the researchers of animal models for amyotrophic lateral sclerosis, with soluble, hyperphosphorylated tau in the brain and vacuolated mitochondria in the spinal cord. The mutation mapped to a gene for a ubiquitin ligase, christened listerin.
Since then, Joazeiro has studied listerin function. Because the gene is conserved throughout eukaryotes, he and Bengtson decided to study its yeast homolog, Ltn1. Previous work in yeast suggested that Ltn1 was involved in nonstop protein metabolism (Wilson et al., 2007).
The Scripps scientists expressed two gene constructs in yeast: one a green fluorescent protein (GFP) labeled with both Flag and HIS motifs, and the other the same sequence sans the stop codon. The nonstop GFP chimera co-immunoprecipitated with Ltn1 at a higher level than the normal, stop-containing reporter, suggesting Ltn1 and the runaway protein physically interact. The scientists also found that if they knocked out Ltn1, the nonstop protein was expressed at higher levels.
Since Ltn1 shares sequences with other genes that encode ubiquitin ligases, the researchers hypothesized that Ltn1 ubiquitinated nonstop proteins. When they examined their GFP reporters, they found that the nonstop version, but not the normal one, was ubiquitinated. This was true in wild-type cells; in Ltn1-deficient cells, ubiquitination of the nonstop protein was reduced. Therefore, the researchers concluded that Ltn1 is essential for full ubiquitination and degradation of defective, unending proteins.
But how does Ltn1 identify run-on proteins? Previous research suggested that the extra lysines, coded by the mRNA’s normally untranslated poly(A) tail, give those offenders away (Ito-Harashima et al., 2007). To test this, Bengtson created a GFP-Flag-HIS3 gene with 12 consecutive lysine codons just before the stop codon. If the lysines were the Ltn1 signal, then this protein should be degraded like a nonstop protein. Sure enough, the cell ubiquitinated and broke down the extra-lysine version, but only if Ltn1 was present.
Other scientists have found that electrostatic interactions between the nascent protein and the ribosome can arrest translation (Lu and Deutsch, 2008). Lysines are positively charged; hence, Joazeiro and Bengtson hypothesized that nonstop proteins with a string of lysines might get stuck in the ribosome.
To look at that, the researchers fractionated cells and used a sucrose gradient to separate components by density. Normally, the few nonstop proteins that escape degradation float in the top, lightest layer. But in Ltn1-deficient cells, the nonstop proteins came down with the ribosomes. Ltn1 co-immunoprecipitated with the ribosomal protein Rpl3, suggesting that Ltn1 spends at least some of its time attached to ribosomes as well. These results, the authors write, suggest that Ltn1 associates with the ribosome and ubiquitinates nonstop proteins as they are synthesized. Without Ltn1, these defective proteins never manage to exit the ribosome. How Ltn1 finds the nonstop proteins is unclear; the authors hypothesize that lysine stretches might alter the ribosomal structure, opening up an Ltn1 binding site.
To understand how Ltn1 affect yeast’s ability to survive, Bengtson and Joazeiro treated cells with an antibiotic that speeds up how fast ribosomes read through stop codons. Ltn1 knockouts struggled to survive—a struggle, Joazeiro suggested, that might also occur in aging neurons.
“Defects in protein quality control are a hallmark of neurodegeneration,” Joazeiro noted. He suspects listerin may be involved in human neurodegenerative diseases. Which one is uncertain. The lister mice’s motor symptoms suggest ALS, though their generalized gliosis and hyperphosphorylated tau could imply Alzheimer’s.
Ambro van Hoof, who also studies nonstop protein degradation at the University of Texas Health Sciences Center in Houston, called the paper “exciting,” but noted that the connection between listerin and human disease remains unproven. In an e-mail to ARF, Joazeiro wrote: “Time and more research on listerin will clarify its involvement in human neurodegenerative disease.”—Amber Dance