18 February 2010. Prions have a bad reputation. They are responsible for a variety of neurodegenerative diseases including Creutzfeldt-Jakob disease in humans and a plethora of spongiform encephalopathies in animals. But there may be more to neuronal prions than neurotoxicity. In the February 5 Cell, researchers led by Kausik Si at the Stowers Institute for Medical Research, Kansas City, Missouri, report that in the marine mollusk Aplysia Californica, prion-like activity of an RNA binding protein controls long-term facilitation (LTF), a strengthening of synaptic transmission that has been linked to memory formation. The protein oligomerizes into amyloid fibers when Aplysia sensory neurons are stimulated with serotonin and, in turn, the amyloid seems to strengthen LTF. This is the first example of eukaryotic prion activity having a physiological function, according to the authors. They also suggest that it “might be representative of a new class of proteins that utilize a self-perpetuating multimeric state to create self-sustaining altered activity state in the cell in response to specific stimuli.”
Whether other amyloids, such as Aβ, might have similar properties, or have evolved from proteins that did, remains to be seen. “This is a fascinating example of what appears to be another "functional amyloid," wrote Charlie Glabe, University of California, Irvine, in a comment to ARF. He noted that the properties of prions may be relevant to learning. “They are stable and ‘self sustaining,’ at least as long as their assembly from monomer balances their disassembly by chaperones or their degradation,” he wrote (see full comment below).
The study was a collaboration with Howard Hughes Investigator Eric Kandel and colleagues at Columbia University, New York. It continues work carried out when Si was in Kandel’s lab. Then, working with Susan Lindquist at MIT, the researchers found that Aplysia cytoplasmic polyadenylation element binding protein (ApCPEB) has prion-like activity in yeast, and that it is required for long-term facilitation in mollusk sensory neurons (see ARF related news story). But whether the prion activity was necessary for the facilitation was unclear. Now, using a variety of biochemical and cellular approaches, Si and colleagues show that the two seem to go hand-in-hand.
First, Si and colleagues wanted to know if ApCPEB has prion activity in Aplysia. When they overexpressed an enhanced green fluorescent protein (EGFP)-tagged ApCPEB in mollusk sensory neurons, they found these neurons to be riddled with distinctive green puncta, suggesting that the protein formed some kind of multimer. Tagged proteins lacking ApCPEB’s prion-like N-terminal domain did not form puncta, supporting the idea the aggregates were formed in a prion-like manner. Though the authors could not rule out lower expression as a reason for the absence of puncta, mouse CPEB, which normally did not form aggregates, did when fused to the N-terminal of ApCPEB. In addition, to ensure that the multimeric ApCPEB was not an artifact of overexpression, the researchers took advantage of an antibody, Ab464, that they raised to aggregated ApCPEB. This recognized the puncta and also immunoprecipitated a large molecular weight (over 180 kDa) entity from normal Aplysia CNS. The complex was recognized by antibodies to monomeric ApCPEB, and it was also resistant to boiling in 10 percent sodium dodecyl sulfate, consistent with it being a stable protein aggregate and not an RNA/protein mixture. Since ApCPEB binds to RNA, the researchers wanted to eliminate the possibility that the puncta were simply RNA protein complexes.
To confirm that RNA was not part of the picture, Si and colleagues expressed protein lacking the RNA-binding domain in sensory neurons and still saw aggregate formation. They also used a split fluorescent protein, which only fluoresces when the two halves come together, to test if ApCPEB can multimerize. With two halves of the split fluorescent protein attached to the C-terminal on different ApCPEB constructs, the researchers were able to reconstitute the fluorescence, and it appeared as puncta, again suggesting ApCPEB forms large aggregates. Finally, to show that the protein can form amyloid, the researchers turned to the electron microscope (EM) and thioflavins, dyes that have a characteristic fluorescence once they bind amyloid β-sheets. In vitro, the ApCPEB folded into complexes that stained with thioflavin T and had the appearance of fibers under the EM. The researchers also found puncta in ApCPEB-overexpressing cells that fluoresced upon binding thioflavin S. They did not find any thioflavin S puncta in normal sensory neurons, but they attribute this to a limited number and smaller size of multimers in vivo.
The biochemical and cellular analyses are consistent with ApCPEB being a bona fide prion in Aplysia, the authors write. The protein also passed a further test, albeit a neuronal version. Prions are self-sustaining and can be passed down from one generation to the next. In non-dividing neurons, the equivalent would be to show that a prion multimer can recruit monomers, note the authors. Si and colleagues demonstrated this by taking advantage of KiKGR, a green fluorescent protein that switches to red fluorescence when briefly excited with blue light. In sensory neurons they expressed an ApCPEB chimera tagged with the switch hitter and then zapped it with blue light, turning it into a red fluorescent reporter. They then sat back and waited to see what would happen to newly synthesized, green fluorescing chimeras. After about four hours, they found green protein incorporated into the red multimers, consistent with the idea that the multimers are self-sustaining.
And what of the physiological role? In Aplysia, long-term facilitation can be induced by pulses of serotonin (5-HT), which is released during learning. The scientists found that in neurons expressing the ApCPEB-EGFP chimera, the number of green puncta jumped by 15 percent after five pulses of 5-HT. To test if endogenous ApCPEB also forms multimers in response to 5-HT, Si and colleagues used a filter trap assay commonly used to detect rare oligomers such as those of amyloid-β and α-synuclein. After 5-HT pulses, the filter trapped fivefold more protein as judged by Ab464 immunoreactivity. Together, the different experiments suggest to the authors that 5-HT induces multimerization of ApCPEB.
If ApCPEB forms multimers in response to 5-HT, then do those multimers contribute to neuronal activity? To test this, the researchers injected sensory neurons with the Ab464 antibody. This had no effect on basal neurotransmission, nor on long-term facilitation measured 24 hours after 5-HT pulses. But it did block facilitation 48 hours after induction, when excitatory post-synaptic potentials were down by 64 percent compared to cells free of the antibody. “These results indicate that the multimeric form of Aplysia CPEB is involved in long-term stabilization of activity-dependent change in synaptic efficacy,” write the authors.
Exactly how the multimers contribute to long-term facilitation is unclear, but since ApCPEB is an RNA-binding protein and protein synthesis is part and parcel of the process of memory formation, the multimers may lead to a period of sustained translation at activated synapses, suggest the authors.
What does this work say about prions in general, which heretofore have turned out to be pathological when active? the authors ask. Si and colleagues think that ApCPEB is distinct in that multimers retain activity and are not detrimental to the organism. “It is possible that the ability of a prion-like protein to assume a self-propagating conformational state is part of an evolutionarily conserved molecular mechanism that can create two distinct activity states from a single polypeptide,” they write. Whether this holds true in higher organisms is unclear. Si and colleagues found that the mouse homolog of ApCPEB does not form multimers, but the fruit fly homolog Orb2RA forms puncta in Aplysia neurons that are stimulated with 5-HT. This suggests that there is some machinery in mollusk neurons that induces multimerization of exogenous CPEB. Other researchers agreed that the findings are certainly interesting but were more cautious, noting that this work has not yet been replicated by other labs and that the phenomenon seems to be limited to Aplysia. There was also some concern that much of the findings rely on overexpression of an enhanced GFP-tagged protein, since EGFP is known to multimerize. Though the authors found that ApCPEB fused to hemagglutinin or KiKGR also formed puncta in Aplysia neurons, they acknowledge that some of the prion properties may be due to overexpression.
“I think it is likely that most amyloids have prion-like properties, such as replication and transmission, but the difference may be that CPEB has been selected by evolution to do this in a highly regulated fashion, which is why it may be functional, while the same properties in unregulated amyloids, such as Aβ, may be pathological,” wrote Glabe.—Tom Fagan.
Si K, Choi Y-B, White-Grindley E, Majumdar A, Kandel ER. Aplysia CPEB can form prion-like multimers in sensory neurons that contribute to long-term facilitation. Cell. February 5 2010; 140:421-435. Abstract