Mutations in parkin can cause familial Parkinson disease (see ARF related news story), but how normal parkin may contribute, if at all, to non-familial PD is more of a mystery. In last week’s Neuron, Andres Lozano and colleagues at the University of Toronto report that the protein BAG5 inhibits parkin and may contribute to degeneration of dopaminergic neurons. This development suggests new ways in which parkin may be involved in the etiology of sporadic Parkinson disease.

BAG5 is just one member of a family of proteins that can, among other things, bind to heat shock protein 70 (Hsp70), a chaperone that has also been implicated in PD (see ARF related news story). First author Suneil Kalia and colleagues identified BAG5 as being potentially involved in neurodegeneration when they examined changes in gene expression elicited by transecting axons in the nigrostriatal pathway of rats. The substantia nigra (SN), of course, is where dopaminergic neurons are most affected in PD. They found BAG5 was one of the transcripts upregulated in transected neurons as opposed to normal neurons. In subsequent experiments, Kalia and colleagues found that BAG5, the only uncharacterized member of the BAG family, could be found in various regions of both rat and human brain, including the cortex and hippocampus, but particularly in dopaminergic neurons.

Using immunoprecipitation assays, Kalia and colleagues found that this BAG, like its five siblings (BAG1-4, and BAG6), binds to and inhibits the chaperone function of Hsp70. When the authors added either BAG1 or BAG5 to cells expressing a reversibly denatured luciferase, the BAG proteins prevented Hsp70-mediated renaturation of the reporter. Immunoprecipitation assays also revealed that BAG5 binds to parkin, too, and that all three proteins—BAG5, Hsp70, and parkin—form a complex.

To examine the consequences of these interactions, Kalia and colleagues tested how BAG5 might impact parkin function. As parkin is an E3 ubiquitin ligase, capable of self-ubiquitination, the authors measured this activity in the presence and absence of BAG5. In test tube experiments, the protein prevented parkin self-ubiquitination, and this inhibition could not be relieved by addition of Hsp70. The results suggest that any positive effect the chaperone may have on Parkinson pathology (for example see Muchowski, 2002) could be offset by BAG5. The authors also found that BAG5 could prevent ubiquitination of synphilin, a protein that interacts with α-synuclein, which can also cause PD when mutated (see ARF related news story).

How might BAG5 contribute to sporadic Parkinson disease? Intracellular protein aggregates, or Lewy bodies, are a hallmark of PD. To see if BAG5 might have any impact on these bodies, the authors examined parkin aggregates that form when the proteasome degradation machinery is deliberately weakened (see Junn et al., 2002). They found that when the proteasome inhibitor MG132 was added to HEK293 cells expressing parkin and BAG5, both proteins—not just parkin—turned up in perinuclear aggregates. And while coexpression of Hsp70 can normally keep parkin out of such melees, (it reduced the number of cells with aggregates from 40 to about 5 percent, for example), in the presence of BAG5 about 30 percent of cells still had such aggregates.

But how might these interactions relate to neurodegeneration? To test this, the authors targeted adenoviruses harboring BAG5 genes to the substantia nigra of rats. The authors found that seven days after injuring the neurons by axotomy, only 35 percent of neurons in adenovirus-infected rats survived, as opposed to 66 percent in control animals. Furthermore, in a similar experiment where the animals were given MPTP, a chemical that induces PD-like symptoms, adenoviruses harboring BAG5 genes exacerbated the effect of the toxin, reducing the survival of dopaminergic neurons by about 25 percent. In contrast, MPTP-treated animals infected with viruses carrying a gene for a BAG5 mutant that cannot bind Hsp70 had about 50 percent more dopaminergic neurons than did controls.

“Based on our findings, we propose a mechanism for neurodegeneration in which BAG5 can interact with both parkin and Hsp70, resulting in decreased parkin and Hsp70 function, two outcomes that are deleterious to cell survival,” write the authors. They also suggest that BAG5 may serve as an important link among the chaperone system, the ubiquitin proteasome system, and protein aggregation.—Tom Fagan


  1. The experiments presented by the Lazano group are solid because the authors have been careful to cover as many bases as they can to test their major hypothesis: that BAG5 is a natural inhibitor of parkin function. But as well as being careful science, the study is interesting for two reasons. Firstly, along with recent data from several other groups, this reinforces the idea that parkin is a generally good neuroprotective protein for nigral neurons. In this particular case, the link to the E3 ligase function of parkin is strengthened because BAG5 appears to decrease parkin’s enzyme activity. That this is specific is demonstrated by the use of the DARA-BAG5 construct that is expressed but cannot bind to its targets, parkin and Hsp70. Secondly, the role of Hsp70 is highlighted by the observation that BAG5 antagonizes the previously identified interaction of this chaperone with parkin (Imai et al., 2002).

    This makes me wonder about the exact role of Hsp70. Hsp70 could be unfolding substrates for parkin and/or moving them from insoluble to more soluble fractions, enhancing their ability to be recognized by parkin. Alternatively—and the Kalia et al. paper provides evidence for this—Hsp70 may be important in maintaining a stable, soluble pool of active parkin. The distinction between these two ideas may seem minor: In either case, the presence of Hsp70 makes for more parkin activity and BAG5 would prevent this. But it has some implications for how we interpret studies in which we rely on the overexpression of parkin to protect cells. Is the effect really due to parkin being an E3 ligase, or is it sequestering some of its toxic proteins in the insoluble fraction because parkin itself is prone to misfolding and tends to make inclusion bodies? The paper by Kalia et al. provides the impression that one has to have active, soluble, and folded parkin for neuroprotection, even if sequestration of damaging substrates into the insoluble fraction is a beneficial mechanism by itself.


    . CHIP is associated with Parkin, a gene responsible for familial Parkinson's disease, and enhances its ubiquitin ligase activity. Mol Cell. 2002 Jul;10(1):55-67. PubMed.

    View all comments by Mark Cookson

Make a Comment

To make a comment you must login or register.


News Citations

  1. New Parkinson’s Gene
  2. Controlling Huntingtin with Chaperones and Intrabodies
  3. Parkin and α-synuclein Linked Through Synphilin-1

Paper Citations

  1. . Protein misfolding, amyloid formation, and neurodegeneration: a critical role for molecular chaperones?. Neuron. 2002 Jul 3;35(1):9-12. PubMed.
  2. . Parkin accumulation in aggresomes due to proteasome impairment. J Biol Chem. 2002 Dec 6;277(49):47870-7. PubMed.

Further Reading

Primary Papers

  1. . BAG5 inhibits parkin and enhances dopaminergic neuron degeneration. Neuron. 2004 Dec 16;44(6):931-45. PubMed.