Over the past dozen years, great strides have been made in identifying mutations that cause Alzheimer (AD) and Parkinson (PD) diseases. But while familial AD cases result from mutations that impinge on a single biological pathway—the cleavage of the amyloid-β precursor protein (AβPP) by γ-secretase—mutations that cause familial PD have no such nexus. The puzzle, then, lies in how mutations in proteins as disparate as parkin, α-synuclein, DJ-1, and Dardarin converge to destroy only dopaminergic neurons of the substantia nigra (SN), the part of the brain devastated by Parkinson disease. One theory suggests that dopamine itself is the key. This view is bolstered by an advanced online publication in the October 16 Nature Medicine. Matt LaVoie and colleagues, including Michael Schlossmacher and Dennis Selkoe, at Boston’s Brigham and Women’s Hospital, report that dopamine can covalently modify parkin, rendering parkin inactive and insoluble.

Scientists have long suspected dopamine, which can readily convert to the oxidizing agent dopamine quinone, of being involved in Parkinson disease pathology. Four years ago, researchers showed that in test-tube experiments dopamine quinone could bind covalently with α-synuclein (see ARF related news story). More recently, research showed that ablating dopamine protects cells from the effects of PD-causing α-synuclein mutations (see ARF related news story). Now, in what may be the strongest link between dopamine and PD yet, LaVoie and colleagues discovered not only that dopamine can covalently modify and inactivate parkin, but also that this reaction may be relevant to pathology, as the authors were able to isolate dopamine-derived parkin from the substantia nigra of human brain.

LaVoie first tested if dopamine has any effect on parkin in neuronal cell cultures. When the researchers exposed dopaminergic MES cells expressing myc-tagged human parkin to exogenous dopamine (50-100 μM), they found that some of the parkin was rendered insoluble. The catecholamine had a similar, dose-dependent effect on endogenous parkin, reducing the amount of soluble protein by more than 50 percent and increasing the amount of high-molecular-weight, insoluble parkin.

These experiments do not prove that dopamine reacts with parkin. To test that, the scientists treated cells with radiolabeled dopamine. When they immunoprecipitated parkin from these cells, they found that it was radioactive, indicating that dopamine, or a derivative thereof, had bound covalently to the protein.

These findings led the authors to examine human tissue samples to see if the dopamine-parkin interaction has pathological meaning. They found insoluble parkin in caudate nuclei isolated from PD brain (the caudate receives input from the dopaminergic neurons in the substantia nigra). Furthermore, using a phenyl boronate-based affinity purification method to isolate proteins that have reacted with dopamine quinone, they were able to isolate dopamine-bound parkin from normal human brain extracts. Significantly, the only tissue tested that yielded the modified protein was the substantia nigra. The caudate-putamen; the red nucleus, which lies adjacent to the SN; and the cerebellum, which contains little dopamine, all tested negative. In summary, the authors propose that “dopamine-induced loss of parkin activity represents a possible mechanism contributing to the selective degeneration of nigral neurons over time.”

How these findings tie in with the other proteins that have been implicated in PD pathology remains unclear at present. That said, redox chemistry is surfacing time and again as a key factor in PD pathology. Just last week, two papers in press at the Journal of Biological Chemistry strengthened the link between oxidative stress and PD. In one, Wenbo Zhou and Curt Freed at the University of Colorado Health Sciences Center, Denver, reported that DJ-1 upregulates expression of heat shock protein 70, inhibiting the aggregation and toxicity of the A53T mutant of α-synuclein. But the authors also report that in cells subjected to hydrogen peroxide or 6-hydroxydopamine, DJ-1 upregulates synthesis of glutamate cysteine ligase. This enzyme catalyzes the rate-limiting step in the synthesis of glutathione, a major cellular antioxidant. DJ-1 has previously been suggested to protect cells from oxidative attack (see ARF related news story).

In the second JBC paper, Benjamin Wolozin and colleagues at Boston University School of Medicine, together with collaborators in the U.S. and Portugal, make a strong case that mitochondria, cellular hotbeds of oxidative chemistry, are particularly vulnerable to altered expression of α-synuclein, parkin, or DJ-1. Using the roundworm Caenorhabditis elegans as a research model, joint first authors Rina Ved, Shamol Saha, and colleagues deleted the parkin homolog, knocked down the DJ-1 homolog, and expressed wild-type or mutant α-synuclein in the worms (roundworms have no α-synuclein homolog). They then challenged the animals with mitochondrial inhibitors and toxins. While both wild-type and genetically modified worms reacted similarly to oxidants like paraquat and divalent metal ions, or to the enzyme inhibitor sodium azide—which attacks the last three of the four complexes that make up the respiratory chain in the inner mitochondrial membrane—the PD model worms were much more sensitive to rotenone and other chemicals that specifically attack complex I of the respiratory chain (see ARF related news story). The results suggest that complex I of the mitochondria could be a focal point for PD pathology.—Tom Fagan

Q&A with Mathew LaVoie. Questions by Tom Fagan.

Q: Why were the statistically significant increases in insoluble parkin detected only in the caudate?
A: We have not yet conducted an exhaustive examination of numerous regions from these brains. We speculate that the observed increase in insoluble parkin within normal and PD caudate may be related to the dopaminergic innervation and dopamine content of this brain region. The more subtle loss of dopaminergic input to the caudate (compared to putamen) may provide a window into earlier events in the disease process, whether it’s an effect of metabolic impairment, oxidative stress, or dopamine, but again, we have not yet formally tested these ideas.

Q: Why do you think you detected insoluble catechol-modified parkin only in the substantia nigra?
A: The proportion of total protein in the substantia nigra lysates that actually originates from dopaminergic cells is much greater than the contribution of protein from the dopaminergic axons that project into the caudate and putamen. We believe that there might be some catechol-modified parkin present in the caudate/putamen, but since only a very small percentage of the total parkin in this brain region originates from dopaminergic axons and only a portion of parkin is likely modified by dopamine, we were unable to detect modified parkin from the insoluble (parkin-rich) fraction in caudate/putamen. However, we did find evidence of very low levels of catechol-modified parkin from nigra and caudate/putamen from the soluble brain fractions.

On a general note regarding the insoluble nature of parkin in human brain, Pawlyk et al. (2003) demonstrated that, in the aged human brain, parkin is predominantly in an insoluble state. We found this to be true in all brain areas we examined, as well. However, we also found readily detectable levels of soluble parkin, albeit much lower than in the insoluble fractions. Since parkin solubility can be affected by direct oxidant exposure (Winklhofer et al., 2003) or dopamine (our work), it may not be surprising that insoluble parkin is found throughout the brain. I imagine that in the near future, many other stressors may also be found to affect parkin solubility, as was first discovered for hydrogen peroxide. This may or may not be related to the observation of insoluble parkin throughout the brain (including cortex).

Q: You found covalently modified parkin in the SN of normal brain. Did you also try to detect modified parkin in PD tissue?
A: We have not yet analyzed PD tissue, and if the hypothesis is correct, dopamine-modified proteins should be extremely hard to detect from a brain where most of the dopamine-producing neurons have degenerated and the relevant, modified proteins degraded. Nonetheless, this is an experiment that should be attempted. Nothing is known about the handling of dopamine-modified proteins, how they are degraded, or how rapidly they are turned over, so we don’t really know what to expect. For example, are dopamine-modified proteins present in nigral Lewy bodies or within neuromelanin? These are the types of questions that would be very relevant for examination in PD tissue.

Q: Can you speculate on why parkin is so sensitive to reaction with dopamine?
A: Having compared other RING-domain E3 ligases, we found that some are subject to hydrogen peroxide-induced insolubility, but their solubility is not affected by dopamine. We are now examining whether these other E3 ligases (and other control proteins) are covalently modified by dopamine in living cells, and how readily. We are also working to obtain mass spectrometry data on the actual cysteine residues attacked by dopamine. Such structural information might shed light on why parkin is more sensitive than related proteins. For example, we could compare which cysteines are vulnerable to nitrosylation (i.e., Yao, et al., 2004) and which to dopamine quinone attack.

Q: What new experiments are suggested by your data? You mentioned in the paper, for example, that it is unclear how the findings related to the use of L-dopa as a treatment. Are there other burning questions that need to be addressed?
A: There are a few reasons why we mention in the paper that it is unclear how our data relate to the use of L-dopa in PD. First, the definitive clinical study has yet to be completed to determine whether L-dopa administration, which represents the basis for neurotransmitter replacement therapy of nearly all Parkinson disease patients, hastens disease course. Second, I wonder whether the small population of nigral neurons that survives in a late-stage PD patient remains due to some intrinsic resistance to the neurodegenerative process. If this is the case, these remaining neurons might be relatively immune to whatever increased oxidative stress might be caused by L-dopa treatment. It seems reasonable to expect the most vulnerable neurons to die first and those that remain at later stages of disease might be less affected by further stress. This speculation may also be relevant to the interpretation of the clinical studies, as well, since such a high percentage of the total nigral neurons are already lost when a typical patient is brought to the attention of a neurologist.

There are certainly some additional questions raised by this work. The first is, “Why are the cells which are most dependent on parkin for survival also most vulnerable to a biochemical loss of parkin?” Those of us who study the adverse effects of dopamine on neuronal function might also ask, “What are the most relevant targets of dopamine in vivo?” To answer this, we hope to make use of the methods developed in this paper to screen, in an un-biased manner, catechol-modified proteins from various cell culture and animal model systems and determine a rank-order of vulnerable proteins. We would ultimately then look for confirmation in the human brain. This will require a lot of work, material, and collaborators. Hopefully we can make some headway on this project in the future.

Comments

Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. Dopamine did it again. LaVoie and colleagues presented some compelling new evidence suggesting that dopamine may contribute to the demise of the same neurons that produce this essential neurotransmitter, this time by inactivating parkin. The authors analyzed the parkin species in dopaminergic cell lines and human tissues from normal and PD brains. In each case, they have convincingly demonstrated the presence of parkin species that are covalently modified by oxidized dopamine. The modification of parkin by dopamine quinone was specific, since several other cysteine-rich proteins and PD-related proteins were not similarly modified. In addition, this modification correlated with the accumulation of Triton-insoluble parkin. Furthermore, using an in vitro ubiquitination assay, the authors demonstrated that the highly reactive dopamine quinone inactivated parkin’s E3 ligase activity. Therefore, the authors proposed that parkin is a direct molecular target of oxidized dopamine, and the dopamine-induced loss of parkin activity represented a possible mechanism of the selective neurodegeneration in PD. This report echoes two earlier reports from the Lansbury’s lab (1) and us (2) suggesting that dopamine contributes to the α-synuclein toxicity. It also provides an interesting and exciting link between dopamine and another protein genetically linked to PD. In addition, this report strengthens the roles of reactive oxygen species in PD pathogenesis.

    Like many other interesting studies, this paper raises some new questions and will certainly stimulate additional discussions and future studies. Although the in vitro experiments in this study clearly demonstrated the covalent modification of parkin by dopamine quinone and the dopamine-induced decrease of soluble parkin, the results from the human brain tissues are less consistent. First, it is surprising that there was no detectable DA-conjugated parkin in the caudate-putamen, where dopamine is abundant, especially given that the insoluble parkin species accumulated in this region in both control and PD brains. Second, unlike in cultured dopaminergic cell lines, the presence of endogenous dopamine did not lead to deceased levels of soluble parkin in either the control or PD caudate-putamen, while dopamine presumably resulted in higher insoluble parkin species in this region.

    Interestingly, it is known that the dopamine content significantly decreases in PD brains. Therefore, the role of dopamine in parkin solubility in vivo, and the functional consequence caused by the change in parkin solubility remain to be clarified. Certainly, the accumulation of insoluble parkin may occur over a long period of time. Even so, the equivalent amount of a soluble (presumably the functional) pool of parkin among cortical and caudate tissues in normal and PD brains suggests a disease mechanism more complex than just the regulation of parkin solubility by DA. The DA-induced loss of E3 ligase activity, on the other hand, is a very interesting mechanism. It will be of great interest to pinpoint the site of the modification in parkin in vitro and in vivo using mass spectrometry, and to confirm the effects of DA modification on the solubility and the E3 ligase activity using a mutagenesis approach. In addition, identifying the modification site will also shed light on the specificity and selectivity of the parkin-DA conjugation, and potentially facilitate the design of inhibitors to alter the modification of parkin by DA.

    Another interesting observation presented in this paper is that DJ-1 and α-synuclein do not covalently bind oxidized dopamine. Like parkin, DJ-1 is a neuroprotective protein. The loss-of-function mutations in the human DJ-1 gene cause early-onset PD. Several groups have suggested that cysteine 106 in the DJ-1 protein is sensitive to oxidative damage/modification, and is important for its normal function (3,4). This report has excluded the possibility that DA is the culprit potentially inactivating DJ-1 at C106. In addition, since DA theoretically modifies a cysteine residue within parkin, it will be of interest to determine why certain cysteine residues are particularly vulnerable to DA modification. A previous report (1) using recombinant α-synuclein indicates that DA modifies α-synuclein and stabilizes the toxic protofibrillar α-synuclein. However, DA-modified α-synuclein species were not detected in human brain tissues in this study. Therefore, the presence of potentially toxic DA-modified α-synuclein in vivo needs to be further confirmed.

    Overall, this paper provides an intriguing potential mechanism for selective parkin inactivation in PD. Given the unequivocal evidence that parkin is a ubiquitin E3 ligase, it is important to identify the authentic in vivo parkin substrates to fully understand their roles in PD.

    View all comments by Jin Xu

References

News Citations

  1. Dopamine and α-Synuclein Are Found to Be Chemically Linked
  2. Dopamine Renders α-Synuclein Toxic to Neurons
  3. <i>Drosophila</i> Define DJ-1’s Defensive Role
  4. Loss of Parkin in Mammals Takes Steam Out of Mitochondria

Paper Citations

  1. . Novel monoclonal antibodies demonstrate biochemical variation of brain parkin with age. J Biol Chem. 2003 Nov 28;278(48):48120-8. PubMed.
  2. . Inactivation of parkin by oxidative stress and C-terminal truncations: a protective role of molecular chaperones. J Biol Chem. 2003 Nov 21;278(47):47199-208. PubMed.
  3. . Nitrosative stress linked to sporadic Parkinson's disease: S-nitrosylation of parkin regulates its E3 ubiquitin ligase activity. Proc Natl Acad Sci U S A. 2004 Jul 20;101(29):10810-4. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . Dopamine covalently modifies and functionally inactivates parkin. Nat Med. 2005 Nov;11(11):1214-21. PubMed.
  2. . DJ-1 up-regulates glutathione synthesis during oxidative stress and inhibits A53T alpha-synuclein toxicity. J Biol Chem. 2005 Dec 30;280(52):43150-8. PubMed.
  3. . Similar patterns of mitochondrial vulnerability and rescue induced by genetic modification of alpha-synuclein, parkin, and DJ-1 in Caenorhabditis elegans. J Biol Chem. 2005 Dec 30;280(52):42655-68. PubMed.