*This website was produced as an assignment for an undergraduate course at Davidson College.*

Caroline Bennett's Review Paper

paper being reviewed:

Mitochondrial pathology and apoptotic muscle degeneration in Drosophila parkin mutants

By Jessica C. Greene, Alexander J. Whitworth, Isabella Kuo, Jaurie A. Andrews, Mel B. Feany, and Leo J. Pallanck

From PNAS, Volume 100, Number 7, 1 April 2003: 4078-4083

Edited by Kathryn V. Anderson


Summary / Critique

Abstract and Introduction

This April 1, 2003 article from PNAS focuses on Parkinson’s Disease (PD), particularly an early-onset familial form of the disease called autosomal recessive juvenile parkinsonism (AR-JP). The Abstract and Introduction provide adequate background regarding previous research done to understand the function and mechanism of AR-JP. PD in general is a neurodegenerative disorder that causes loss of dopaminergic neurons in the substantia nigra pars compacta and the accumulation of proteinaceous intraneuronal inclusions called Lewy bodies. The molecular mechanism for these defects thus far remains unclear, but scientific evidence points to mitochondrial dysfunction as the major causative factor. In the Introduction, Greene et al. describe the previous identification of loci responsible for rare monogenic forms of PD as insightful to mechanistic understanding of the disorder; this approach identified the gene parkin. A paper published in 1998 illustrated that loss-of-function mutations in parkin result in early onset of AR-JP and lack of Lewy body formation. The parkin gene encodes a polypeptide with an N-terminus ubiquitin-domain, two C-terminus ring-finger domains, and an in-between ring-finger domain. In accordance with this structure, parkin has previously been found to act as an ubiquitin protein ligase; in other words, dopaminergic neuron loss in AR-JP results from the failure to label specific cellular targets with ubiquitin.

Greene et al. recognized the need for an animal model of this disorder in order to investigate the relevance of these Parkin substrates to dopaminergic neuron loss. They applied a functional approach to further the scientific understanding of the molecular mechanism responsible for selective cell death in AR-JP by creating a Drosophila model for the disease. This Abstract and Introduction give the reader a clear general understanding of Parkinson’s Disease and the research that has been done on AR-JP thus far. It also explains that the null mutants for parkin generated in this study exhibit reduced life span, locomotion defects, and male sterility, indicating that the defects result from mitochondrial dysfunction. In the Abstract, the authors write that these results “raise the possibility that similar mitochondrial impairment triggers the selective cell loss in AR-JP” (4078). Greene et al. chose these words carefully, explaining what the data “indicate” and “suggest” rather than “prove.”


Initially, the researchers probed the Berkeley Drosophila Genome Project Database using a human Parkin polypeptide query sequence. They found the DNA sequence encoding Drosophila parkin ortholog and fully sequenced it. Parkin mutants were generated by inducing transposition of a P element insertion, and a single mutant line, parkEP(3)LA1, was recovered via PCR analysis of genomic DNA from the offspring of approximately 5,500 flies. ParkEP(3)LA1 has an insertion 71 bp upstream of the parkin start codon. They introduced modified cDNA (parkin cDNA clone SD01679) into the Drosophila germ line after being ligated into the P{UAST} vector. Northern blot analysis followed.

Greene et al. then conducted three behavioral assays on the mutant flies to analyze the functional differences between flies with mutant parkin and wild-type flies. The first assay tested the longevity of flies. It was conducted in triplicate for each genotype, and the researchers recorded the number of dead flies when 0-24 hr old flies were transferred to new vials every 2-3 days. The test for flight was a bit more complicated, and used the Benzer apparatus. At least 100 flies of each type (1-2 days old) were tested for flight. In general, flies were released into a graduated cylinder and then stuck to an acetate sheet coated with vacuum grease; with this method, flies stick wherever they alight and flight efficiency can be calculated based on where they stick. To test climbing efficiency in mutant vs. wild type flies, at least 60 flies of each genotype were involved in climbing assays using a multi-chambered apparatus. By counting the number of flies able to climb to various chambers in the apparatus during timed trials, a climbing index was calculated in the same manner as the flight index. It is important to note that these behavioral results are based on averages of many flies; the researchers were conscious of the fact that large sample sizes provide more accurate results. However, the statistical values (indexes) calculated seem to have significant ranges of error, based on the graphs in Figure 3. This issue will be discussed again in the analysis of the results, but it illustrates the difficulty of behavioral tests.

Frontal brain and muscle tissue were sectioned and stained with hematoxylin and eosin or immunostained with a polyclonal antibody to tyrosine hydroxylase. Also, tissues for electron microscopic analysis were prepared by dissecting testes from 6-hr old males, thoraces from 1- to 2-day old adults, and aged pupae.


When probing the Drosophila genome for parkin, a single was gene identified. The fact that the human homolog recognized this gene and that this gene is the only one in the Drosphila genome encoding a polypeptide bearing a ubiquitin-like domain, ring-finger domains, and IBR domain indicate that the correct gene was isolated in this process. This 468 aa gene shares 59% overall similarity with human Parkin, based on sequence analysis. Description of Figure 1B explains that “embryos, larvae, and adults detected transcripts at all developmental stages with particularly high abundance in adults” (4079). While the figure does illustrate that the parkin transcript is detected in each of the three stages of development shown, I am skeptical whether this data can actually be quantified in this way. The blot shows no loading controls to indicate that the same amount of RNA was loaded in each lane. I would have appreciated seeing mw markers as well.

After identifying and isolating the fly Parkin ortholog, Greene et al. generated a disruption of the parkin gene. They did this using a transposon mutagenesis screen using a P element mapping close to parkin. This process yielded the single mutated line ParkEP(3)LA1 with an insertion 71 bp unstream from the parkin start codon. They further generated more severe alleles via the screening process. Of the collection of deletion alleles via the transposon mutagenesis screen method, some were null alleles lacking all of the parkin coding sequence. Figure 1C accurately illustrates various mutants and the portions of the parkin coding sequence deleted with each. A vital component of their work is the control chromosome generated as a standard of comparison (parkrvA).

Greene et al. then studied the physical effects of replacing the wild type parkin allele with the null allele (the Df(3L)Pc-MK deletion chromosome was responsible for removing the wt parkin gene).
Observed physical effects in flies with the null allele include:

1) Flies live to adulthood, but average lifespan is significantly shorter (27 days with a max. of 50 days vs. 39 days with a max. of 75 days).
2) Complete male sterility was observed in flies tested. With this knowledge, the sterile lines were screened for additional parkin mutations. Sequencing two of the mutants recovered from this screen revealed missense and premature stop codon mutations; this illustrates that the observed phenotype does in fact result from loss of parkin function. Analysis of testes of parkin mutants compared to wt testes (shown in Figure 2) illustrates that male sterility is caused from a late defect in spermatogenesis. The defect occurs at the level of individualization; germ-line cysts that typically separate into individual sperm cells fail to do so. Figure 2 clearly shows this structural defect. The inset in Panel 2B shows that spermatids do form in the mutant cells, but they fail to individualize like the sperm shown in panel 2A (wt). Panels 2E and 2F illustrate conservation of axoneme structure, while the Nebenkern structure is disrupted. Some spermatids have multiple Nebenkern (which are specialized mitochondrial derivitives), and others have only a reduced Nebenkern component. Since this specific structural change is observed and the axoneme structure is not disrupted, it is logical for Greene et al. to suggest that defective Nebenkern formation and/or function may underlie the failure of spermatid individualization.
3) All flies with null parkin alleles also exhibit abnormal wing posture. The images in Figure 3 illustrate this “partially penetrant downturn wing phenotype,” a condition that progresses significantly with age (4080). Locomotor ability and climbing ability was prompted as a result of this finding; panels 3C and 3D show the significant defects in both of these functions. Both of these two graphs base their analysis on the control phenotype, and the climbing index is significantly lower for the mutant flies compared to wt; also, the proportional difference between wt and mutant is consistent at each varying age level. In panel 3C, the mutants are again significantly lower in flight index, and one mutant flight index is lower than the other two. The authors do not address this slightly lower average index. I am concerned the significant error bars for each bar graph in Figure 3. This error range could alter the difference in index values between wt and mutant flies, and therefore cause the data to be less consistent that it appears in this graph.

Panels 3E and 3F show an important technique used to support the role of parkin. A GAL4-UAS system was used to as a “rescue” mechanism to return parkin expression to defined tissues. The bar graphs in 3E and 3F provide compelling evidence that Parkin function is required in the musculature. Two GAL4 lines were found to rescue wing posture, flight, and climbing phenotypes of mutants to the level of wt or near-wt levels. The fact that expressing parkin in mesoderm cells can fully restore gene function is positive, convincing evidence for gene function.

Figure 4 further demonstrates the physical consequences of the null allele through histological analysis of the major flight muscles in Drosophila. As expected, the parkin mutants have severe disruption of muscle integrity; acute degeneration is characterized by vacuole formation and accumulation of cellular debris (Figure 4B). Amazingly, transgenic expression of parkin again rescues this phenotype (although occasional vacuoles are still seen). This raises one question, however. Why isn’t recovery always 100% complete if you are restoring the exact gene that encoded the muscle structure seen in the control (4A)? This observation suggests that perhaps the recovery process used is not a completely efficient method. The occasional occurrence of vacuoles in the transgenic tissue could also be a result of experimental error. I think it is unlikely that this observation undermines the role of the parkin gene. It is curious that the authors fail to include data to fully support their claim that “only a subset of muscles are affected by loss of parkin function” (4080). The paper is not very clear about exactly which muscles are and are not affected by parkin, and I don’t know whether they know this information and simply did not include it or whether the specific muscles are still being investigated.

As far as their analysis of the indirect flight muscles (IFMs), Figure 4 shows grossly swollen mitochondria with “severe disruption and disintegration of the cristae” in parkin mutants (4080). Also, overall density of myofibrils decreases, the myofibril Z-line broadens, and the sarcomere shortens in mutants. As expected, transgenic parkin expression results in recovery of myofibril integrity and mitochondrial morphology.

The next question Greene et al. asked was: does muscle degeneration proceed through an apoptotic mechanism?

To investigate this issue, IFM in parkin mutants and controls were subjected to TUNEL staining. A significant increase in TUNEL-positive nuclei appeared in the IFM of 1-day-old adult parkin mutants (Figure 5), suggesting that the muscle mitochondrial defects do indeed ultimately result in cell death through an apoptotic mechanism.

Greene et al. further investigated the role of parkin in the brain by histologic analysis. No clear neuronal loss was observed when brain sectioned were immunostained with tyrosine hydroxylase. However, as illustrated in Figure 6, cells of the dorsomedial dopaminergic cell cluster reliably showed shrinkage of the cell body and decreased immunostaining in proximal dendrites in aged mutants relative to controls. Why is this observed? The authors mention the fact that this specific cluster of neurons has enhanced toxicity to a protein implicated in familial PD as an interesting connection. I did not find this connection very striking, but that may be because I am not familiar with the toxin.

Discussion and Future Experiments

Greene et at. successfully create a Drosophila model for AR-JP by disrupting a highly conserved Drosophila parkin ortholog. Their extensive analysis of null mutant flies illustrates the phenotypical effects of loss of parkin function. They also resaonably hypothesize the mechanisms through which these physical effects occur. Based on histological analysis, they propose that male sterility results from a spermatid individualization defect in the male germ line, and that locomotion defects arise from apoptotic muscle degeneration.

Although loss of parkin function in Drosophila mutants and in AR-JP targets differnt tissues, it is likely that AR-JP operates via a similar molecular mechanism as the Drosophila model studied. Mitochondrial defects are a common characteristic of the pathology of male germ lines and IFM in parkin mutants (under ultrastructural examination). But the possible mechanistic relationship between mitochondrial pathology and lack of spermatid individualization is still unknown. Greene et al. target initial mitochondrial disfunction as the common origin of the Drosophila parkin mutant phenotypes. The logical next step for understanding this mechanism is to investigate exactly how loss of parkin can trigger mitochondrial pathology and ultimately cell death at the molecular level.

More in depth structural analysis of the protein that parkin encodes is a good start. We know that Parkin functions as a ubiquitin protein ligase, tagging cellular targets with ubiqutin. In Parkin mutants, failure to tag the targets causes dopaminergic neuron loss in AR-JP. The Introduction explains that it was recently found that Parkin functions this way in a pathway with ubiquitin-activating (E1) and ubiquitin-conjugating (E2) enzymes. What exactly does the ubiquitin protein ligase do in order to cause neuron loss and what significance is there that it acts in a pathway with two other enzymes? And does the Drosophila parkin operate in a pathway as well? How conserved is the exact mechanism through which Parkin operates? It might be helpful to stain the protein in wt fly cells from various tissues and follow its path and mechanism over a time period. You could use Green Flourescence Protein labeling in order to stain the protein while the cell remains alive and functional. You could use GFP in human wt cells and conpare the observations.

Another logical next step following this paper is to create a mammalian model for the disease in mice. You could test mice in a similar manner as Greene et al. tested flies. You would isolate the parkin ortholog in the mouse genome by probing it with the human sequence and then sequence the mouse version of parkin. You could also analyze phenotypical defects in mutant mice, although this procedure would be more difficult in mammals with longer gestation periods. If mutant mice show muscle degeneration and locomotive defects, this is strong evidence that the gene is highly conserved in mice as well. You would analyze muscle tissue for degeneration and the mitochondrial pathology within muscle tissue, looking out for swollen mitochondria. You would also investigate the sterility of male mice and then analyze the testes on a cellular level for the mechansim for spermatid individualization. You would expect to see Nebenkern abnormalities. A mouse model for AR-JP would be incredibly imformative and a fascinating. It would be another key to understanding evolutionary mechanistic conservation in relation to PD.

Althoght this field still requires extensive research in order to gain the ability to more effectively treat and/or cure the disease, this study represents a valuable step toward understanding the mechanism of AR-JP and PD in general. The findings presented by Greene et al. certainly have greater implications for a potential mechanism for idiopathic PD. In the Discussion, the authors point out that "a substantial body of evidence suggests that mitochondrial disfunction and apoptosis are important functions underlying neurodegeneration inidiopathic PD" (4083). If we can uncover the complete mechanism for Parkinson's at the level of Drosophila (or mice), we will have reason to believe that the mechanism of the disease in humans is highly conserved. Undertanding the mechanism of Parkinson's Disease will inform drug research in the direction of what to target in order to treat the diesease effectively.


back to Caroline's homepage

Molecular Biology Homepage

Davidson College Homepage

Please direct questions or comments to Caroline Bennett at cabennett@davidson.edu.