28 April 2011. In the movie Casablanca, protagonists Rick and Ilsa would always have Paris. Parkinson’s disease researchers always will as well, if a different one. Parkin interacting substrate (PARIS) was not born of the movie screen, but of a screen for parkin substrates. Its debut was a highlight at Neurodegenerative Diseases: The Molecular and Cellular Basis of Neurodegeneration, a Keystone Symposium held 21-26 February 2011 in Taos, New Mexico. PARIS may be a lead actor in mitochondrial dynamics, an underlying theme at the meeting. In particular, it seems that researchers are finally getting a grip on the long-suspected link between mitochondria and not only PD, but other neurodegenerative diseases as well. Many molecular players take part in this link, and quite a few appear to have a penchant for regulating fission and/or fusion of these organelles. This story features new data on a large cast of characters involved. It opens with PARIS and parkin, then moves on to PGC1α and Pink1, mitochondrial anchored protein ligase (MAPL) and Drp1, then further on to mitofusin, PARL, ataxin-3, and Vms1. These proteins each play a part—some large, some small—in the molecular underpinning of mitochondrial deficits in PD and related conditions. So sit back and let this molecular biology revue pass before your eyes as Alzforum closes its Keystone coverage.
First, consider parkin. Though this gene was one of the first to be linked to Parkinson’s, what the protein does in the disease has remained a bit of a puzzle. It is a ubiquitin ligase with so many substrates that it is uncertain which is most important to pathology, said Ted Dawson, Johns Hopkins University, Baltimore, Maryland. Now, there is good reason to think that PARIS might be one of the important ones, he suggested. Originally isolated by yeast two-hybrid analysis using parkin as bait, PARIS co-immunoprecipitates with parkin, colocalizes with it in cortical neurons, and binds to the parkin RING domains that are crucial motifs in many ubiquitin ligases. Parkin ubiquitinates PARIS in neuroblastoma cells, and familial parkin mutations that cause PD block this modification, said Dawson. Furthermore, his group rescued PARIS ubiquitination deficits in parkin knockdown cells by overexpressing wild-type parkin in the cells.
Merely being a parkin substrate does not mean a protein plays a role in PD pathology, but Dawson reported high levels of PARIS in brain tissue from both familial and sporadic PD. In the former, the scientists saw this in the cingulate cortex, but did not have enough material to test brain regions most susceptible to pathology, such as the striatum and substantia nigra. In sporadic cases, PARIS was up twofold in both of these locations, but was normal in the cerebellum and frontal cortex, areas that stay relatively unscathed long into this disease. The pattern of change suggests a link between PARIS and PD at the protein level, Dawson said, as PARIS messenger RNA was unchanged.
PARIS sports zinc finger domains as well as a Kruppel-associated box, and proteins containing both are usually transcriptional repressors, said Dawson. It turns out PARIS binds to a specific form of insulin response sequence. One gene that shares this sequence is PGC1α, a mitochondrial protein that may protect against Parkinson’s (see ARF related news story) and even Alzheimer’s (see ARF related news story). In Taos, Dawson showed that PARIS repressed expression of PGC-1α, and that parkin rescued this suppression. In parkin knockouts, on the other hand, PARIS levels rise and PGC-1α drops, leading to dopaminergic cell death, which is prevented by knockdown of PARIS. These findings suggest that excess PARIS contributes to the demise of dopaminergic neurons when parkin is mutated, suggested Dawson. This work appeared in the March 4 issue of Cell (Shin et al., 2011).
You Say Fission—I Say Fusion
Whether PARIS links parkin to mitochondrial defects is not yet clear, but perturbation of mitochondrial morphology has been a topic of great interest among Parkinson’s researchers. Both parkin and the product of another PD gene, the kinase Pink1, somehow toggle the balance between mitochondrial fission and fusion (see ARF related news story and ARF news story). To work out the mechanism, Ming Guo, University of California, Los Angeles, uses fruit fly muscle as a model system. Muscle cells are highly dependent on mitochondria, and in Taos, Guo showed that they have a similar phenotype to neurons when Pink1 or parkin are mutated or absent.
Guo described genetic screens to identify components of the fission/fusion pathways that might interact with parkin and Pink1. From this exercise emerged the mitochondrial anchored protein ligase (MAPL). Recently discovered elsewhere (see Braschi et al., 2009), MAPL is one of the family of ligases that tag other proteins with small ubiquitin modifiers (SUMOs). Guo reported that MAPL seems to work similarly to the protein Drp1, in that it promotes fission and attenuates Pink1 mutant phenotypes in both muscles and dopaminergic neurons. MAPL seems to work through Drp1, said Guo, and there is evidence that the ligase SUMOylates Drp1. MAPL also rescues Pink1 mutants in the absence of SUMOs, however, so SUMOylation may not be the full story, Guo said.
On the other side of the coin—that is, fusion—evidence is emerging that parkin/Pink1 perturbs that as well. The mechanism involves ubiquitinating mitofusin, which, as the name suggests, promotes mitochondrial fusion. Scientists in Richard Youle’s lab at the National Institute for Neurological Disorders and Stroke, Bethesda, Maryland, reported last year that small compounds that uncouple mitochondrial electron transport from ATP synthesis cause a buildup of Pink1 in mitochondria. The theory is that uncoupling reduces the voltage potential across mitochondria that fuels transport of Pink1 from the outer to the inner mitochondrial membrane, where normally it is quickly degraded. Stuck on the outer membrane, Pink1 then recruits parkin. In Taos, Youle added some of the missing pieces to the puzzle, namely, the protease that normally degrades Pink1 and a potential substrate for parkin. The former is PARL, or presenilin-associated rhomboid-like protein. When the researchers take PARL out of the picture, they prevent the normally constitutive degradation of the Pink 1 kinase (see Jin et al., 2010). The parkin substrate is mitofusin. When mitochondria are uncoupled, extracts contain mitofusin protein of progressively larger sizes, indicative of ubiquitination.
As these studies show, researchers are beginning to unravel manifold links among parkin/Pink1, mitochondrial dynamics, and Parkinson’s pathology. The findings could have broader ramifications, since mitochondrial damage is implicated in other neurodegenerative diseases. For example, Machado-Joseph disease (MJD), one of the most common forms of ataxia, is caused by a polyglutamine (PolyQ) expansion in ataxin-3. In Taos, Thomas Durcan, who works in Edward Fon’s lab at McGill University, Montreal, Canada, noted that many people with MJD have symptoms of Parkinson’s. Could this be because ataxin-3, a de-ubiquitinating enzyme, interferes with parkin, the ubiquitin ligase? That is exactly what Durcan and colleagues reported earlier this year. Ataxin-3 clips ubiquitin off parkin and promotes its clearance from cells via autophagy. In effect, ataxin-3 could be acting as a parkin loss-of-function mutation, Durcan said. Interestingly, it is not the wild-type form of ataxin-3 that de-ubiquitinates parkin, but PolyQ expanded forms (see Durcan et al., 2011). In Taos, Durcan added another twist. PolyQ-ataxin-3 increases oxygen consumption in cells without boosting ATP synthesis. This is just what mitochondrial uncouplers do, bringing to mind Youle’s work. PolyQ ataxin might be trapping Pink1 on the mitochondria and recruiting parkin, which then drives mitochondrial autophagy (see ARF related news story).
One of the challenges of studying mitochondria in disease, noted Jin-Mi Heo, is that a fourth of mitochondrial proteins are characterized either poorly or not at all. Working in Jared Rutter’s lab in the University of Utah School of Medicine, Salt Lake City, Heo reported on one protein that might be part of a novel mitochondrial degradation pathway. Vms1, which stands for VCP/Cdc48-associated mitochondrial stress responsive 1, partners Cdc48 and Np14, components of the ubiquitin-proteasome system. The two are known to help clear proteins associated with the endoplasmic reticulum. Do they help dispatch mitochondrial proteins as well? Heo found that when mitochondria are stressed, Vms1 recruits Cdc48/Np14 to the organelles. When Vms1 is absent, ubiquitinated proteins accumulate in mitochondria, the organelles lose function, and the cell soon dies. It seems that Vms1 is necessary for maintaining a healthy mitochondrial pool. Whether it may be important for Pink1/parkin-induced mitophagy, or figure in any neurodegenerative disease, remains to be seen.
As this parade of presentations revealed, a growing list of proteins is beginning to fill the nexus between neurodegeneration and mitochondria. Most of them drive (e.g., Pink1, parkin, mitofusin, Drp1, MAPL, polyQ-ataxin-3) or are driven by (e.g., PARIS, PGC-1α) fission- or fusion-related events. At the end of Casablanca, of course, a police roundup of the usual suspects failed to nab the culprit. Have researchers done better, or are the major players yet to be discovered?—Tom Fagan.