Human Molecular Genetics, 2012, Vol. 21, No.

12 2698–2712
doi:10.1093/hmg/dds096
Advance Access published on March 7, 2012

Alternative oxidase rescues mitochondria-mediated

dopaminergic cell loss in Drosophila
Dickon M. Humphrey1, Richard B. Parsons2, Zoe N. Ludlow1, Thomas Riemensperger3,
Giovanni Esposito4, Patrik Verstreken4, Howard T. Jacobs5, Serge Birman3 and Frank Hirth1,∗
1
MRC Centre for Neurodegeneration Research, King’s College London, Institute of Psychiatry, Department of
Neuroscience and 2Institute of Pharmaceutical Science, King’s College London, UK, 3Genetics and Physiopathology
of Neurotransmission, ESPCI Neurobiology Unit, Centre National de la Recherche Scientifique, Paris, France, 4VIB,
Department of Molecular and Developmental Genetics, KU Leuven, Belgium and 5Institute of Biomedical Technology,
University of Tampere, Finland

Received January 20, 2012; Revised and Accepted March 2, 2012

Mitochondrial dysfunction is commonly observed in degenerative disorders, including Alzheimer’s and

Parkinson’s disease that are characterized by the progressive and selective loss of neuronal subpopulations.
It is currently unclear, however, whether mitochondrial dysfunction is primary or secondary to other patho-
genic processes that eventually lead to age-related neurodegeneration. Here we establish an in vivo
Drosophila model of mitochondrial dysfunction by downregulating the catalytic subunit of mitochondrial
DNA (mtDNA) polymerase in cholinergic, serotonergic and dopaminergic neurons. The resulting flies are
characterized by lowered respiratory chain activity, premature aging, age-related motor deficits as well as
adult onset, progressive and cell-type-specific, dopaminergic neurodegeneration. Using this model, we
find that associated lethality can be partially rescued by targeting PINK1/parkin signaling or Drp1, both of
which have been implicated in mitochondrial dynamics and Parkinson’s disease. Bypassing mitochondrial
complex III/IV deficiencies with Alternative oxidase (AOX), however, fully restores ATP levels and prevents
dopaminergic neurodegeneration. In contrast, ATP levels and neurodegeneration are not rescued when mito-
chondrial complex I deficiencies are bypassed with NADH-Q oxidoreductase. Our results demonstrate that
mtDNA-mediated mitochondrial dysfunction can cause age-related and cell-type-specific neurodegeneration
which AOX is able to alleviate and indicate that AOX or its surrogates may prove useful as a therapeutic tool
for limiting respiratory chain deficiencies caused by mtDNA decline in healthy aging and neurodegenerative
disease.

INTRODUCTION of neurodegenerative disease cases are sporadic where the

cause of pathogenesis remains elusive. Environmental factors,
Adult-onset neurodegenerative diseases are characterized by the for example pesticides in Parkinson’s disease (3), and genetic
accumulation of protein-based deposits and the progressive loss risk factors such as apolipoprotein E in Alzheimer’s disease
of neuronal subtypes in specific regions of the nervous system (2), have been implicated in idiopathic cases.
(1). Of individuals over 65 years of age, typically 1–2% are Despite differences in disease symptoms, comorbidities and
diagnosed with Parkinson’s disease, whereas 13% suffer from affected neuronal sub-populations, a large number of neurode-
Alzheimer’s disease (2). Heritable forms of the diseases, generative disorders are characterized by defective mitochon-
which account for only "5–10% of cases, are often associated drial remodeling and/or mitochondrial dysfunction (4,5),
with mutations of pleiotropic genes that are expressed ubiqui- including Alzheimer’s disease (6), Parkinson’s disease (7) as
tously but only affect specific cell populations. The majority well as Huntington’s disease (8) and amyotrophic lateral

∗
To whom correspondence should be addressed at: Department of Neuroscience, Centre for Neurodegeneration Research, Institute of Psychiatry,
King’s College London, PO Box 37, 16 De Crespigny Park, SE5 8AF London, UK. Tel: +44 2078480786; Fax: +44 2077080017; Email: frank.
[email protected]

# The Author 2012. Published by Oxford University Press. All rights reserved.
For Permissions, please email: [email protected]
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2699

Figure 1. Knockdown of POLGa causes mitochondrial dysfunction in Drosophila. (A) tamas codes for POLGa. POLGa RNA-interference (RNAi) targeted to
either exon 3 (POLGa-IR2 shaded region) or exon 2 (POLGa-IR1 shaded region). Primers to detect POLGa RNA knockdown: tF, forward primer; tR reverse
primer. (B) daughterless-driven POLGa-RNAi (da.POLGa-IR) decreases POLGa RNA levels (n ¼ 3); d-actin is unaffected. Data are shown for 5-day-old
adult flies. (C) Tubulin-Gal4-driven POLGa-RNAi (Tub.POLGa-IR) decreases DNA levels of the mitochondrially encoded genes: cytochrome c oxidase sub-
units 1 and 3 (COX I and COX III) and cytochrome b (Cyt b). Genomic DNA control is ribosomal protein L32 (RPL32). Data are shown for pupae. (D)
Actin-Gal4-driven POLGa-RNAi (Act.POLGa-IR) decreases mtDNA-encoded proteins NADH-ubiquinone oxidoreductase subunit 5 (ND5) and cytochrome
c oxidase subunit I (COXI). Nuclear-encoded adenosine triphosphate (ATP) synthase subunit a is included for comparison and b-tubulin is a loading control.
Data are shown for 2 –5-day-old adult flies. (E) Protein abundance of mtDNA-encoded ND5 and COXI (normalized to tubulin loading control) are significantly
decreased, but ATP synthase subunit a (CxVa) is unchanged in Act.POLGa-IR flies compared with controls. Data are shown for 2– 5-day-old adult flies. (F)
Complex 1 activity (n ¼ 2, 4 replicates) is decreased in Act.POLGa-IR flies. Data are shown for 5 –10-day-old adult flies. (G) ATP levels (n ¼ 22) are
decreased in Act.POLGa-IR flies compared with controls. Data are shown for larvae. n/s, not significant; ∗ P,0.05; ∗∗∗ P,0.001. Error bars, SEM.

sclerosis (9). For example, in the case of Parkinson’s disease, mtDNA for correct mitochondrial function and cell survival,
several lines of experimental evidence suggest that defective it is currently unclear whether mitochondrial alterations and
mitochondrial dynamics is one of the pathogenic mechanisms dysfunction are a direct cause or consequence of age-related
leading to motor deficits and dopaminergic neurodegeneration neurodegeneration.
(7). Significantly, mutations in Parkinson’s disease-related Here we establish a Drosophila model of mitochondrial
genes parkin, PINK1 and DJ-1 are all associated with mito- dysfunction using a combination of genetics, behavioral ana-
chondrial alterations and/or deficits (10), suggesting that mito- lysis and integrative physiology and show that downregulation
chondrial dysfunction is a major contributing factor to of the catalytic subunit of mtDNA polymerase results in mito-
neurodegeneration and disease progression. chondrial dysfunction and causes progressive motor deficits as
Functional mitochondria are critical for cell homeostasis and well as adult onset, age-related neurodegeneration specific to
survival which is exemplified by animal models. These models dopaminergic cells. Moreover, we provide evidence that Alter-
reveal that targeted mutation or knockout of genes involved in native oxidase (AOX) is able to compensate reduced adenosine
mitochondrial DNA (mtDNA) replication and repair phenocopy triphosphate (ATP) levels and mitochondria-mediated neuro-
mitochondrial dysfunction and other physiological features of degeneration in Drosophila. Our data suggest that AOX may
neurodegenerative disorders (11). For example, knockout of prove useful for limiting respiratory chain deficiencies in
mitochondrial transcription factor TFAM in mice abolishes healthy aging and age-related neurodegeneration.
mtDNA expression, causing a severe respiratory chain defi-
ciency which is embryonic lethal (12), and can result in progres-
sive parkinsonism when targeted to dopaminergic neurons (13). RESULTS
Expression of a proof-reading-deficient catalytic subunit of
DNA polymerase g (POLGa) can lead to accumulation of Reduced mitochondrial gene expression impairs
mtDNA alterations and premature aging and death (14,15). respiration and causes premature aging
However, these mice are also characterized by unusually high To model senescence-related mitochondrial dysfunction, we
levels of mtDNA mutations that are neither seen in normal used Gal4/UAS-mediated RNA interference (RNAi) to
aged wild-type mice (16) nor in patients with recognized knock down the catalytic subunit of the Drosophila mtDNA
mtDNA diseases (17). While illustrating the importance of polymerase g (POLGa or tamas) utilizing two separate
2700 Human Molecular Genetics, 2012, Vol. 21, No. 12

hairpin-loop UAS-RNAi lines that either target exon 2 or exon

3 of the POLGa transcript (Fig. 1A). POLGa-IR driven by the
daughterless-Gal4 driver, which shows ubiquitous but moder-
ate Gal4 activity during development, caused "80% knock-
down of POLGa transcript level when compared with
controls (Fig. 1B). This reduction in POLGa transcript
levels was independent of the POLGa-RNAi strain used (Sup-
plementary Material, Fig. S1A) and hence interchangeable.
For subsequent experiments and analysis, we used line
POLGa-IR1 unless stated otherwise.
Knockdown of POLGa decreased DNA levels of the mito-
chondrially encoded genes cytochrome c oxidase subunits I
and III (COXI, COXIII) and cytochrome b (Cytb) (Fig. 1C)
and protein levels of the mitochondrially encoded genes
NADH-ubiquinone oxidoreductase chain subunit 5 (ND5)
and COXI (Fig. 1D and E) and COXII (Supplementary Mater-
ial, Fig. S1B). In all cases, a significant decrease in protein
abundance was detected in adult POLGa-RNAi flies when
compared with controls (Fig. 1D and E; Supplementary Mater-
ial, Fig. S1B). Nuclear-encoded ATP synthase, subunit a
(CxV), however, was unaffected (Fig. 1D and E). Together,
these data show that partial reduction of POLGa is sufficient
to decrease mitochondrial gene expression.
To assess the physiological effect of decreased mitochon-
drial gene expression, we measured the respiratory activity
and total ATP content of POLGa-RNAi flies. In Act.-
POLGa-IR flies, complex I activity (Fig. 1F) and ATP
levels (Fig. 1G) were significantly decreased when compared
with controls. These data confirm that POLGa-RNAi directly
affects mitochondrial function by decreasing expression of
mtDNA-encoded respiratory chain subunits.
Next we determined whether decreased mitochondrial gene
expression, and in turn decreased mitochondrial function,
could directly affect aging. We studied the effects of
POLGa-RNAi on survival using either Tubulin-Gal4 or Actin-
Gal4, drivers that are ubiquitously and constitutively active
during development, adulthood and aging. Tub.POLGa-IR
mimicked tamas mutations (18) and caused larval/pupal le-
thality (Fig. 2A), whereas Act.POLGa-IR caused partial
larval/pupal lethality (Fig. 2B) with adult escapers that were Figure 2. POLGa-RNAi affects development and survival dependent on
characterized by reduced longevity (Fig. 2C; Table 1). To phenotypic threshold effects. (A) Tubulin-Gal4-driven POLGa-RNAi (Tub.-
exclude developmental effects, we knocked down POLG POLGa-IR) mimics tamas mutations and causes larval/pupal lethality. (B)
only in adult flies using the Gal80 system (19). These flies Actin-Gal4-driven POLGa-RNAi (Act.POLGa-IR) causes larval/pupal le-
(Tub.TubP-Gal80ts .POLGa-IR) were also characterized thality but also leads to adult flies. (C) Actin-Gal4-driven POLGa-RNAi
(Act.POLGa-IR) results in substantially shorter lifespan compared with
by a significantly shortened lifespan, with a median lifespan Act/+ controls. Kaplan–Meier data fitted to a Weibull distribution (dashed
of 59 days and a maximum lifespan of 70 days, compared lines). Mean lifespans: Act.POLGa-IR ¼ 25.4, Act/+ ¼ 69.7. (D)
with controls that showed a median lifespan of 70 days and POLGa-IR knockdown in adult flies (Tub.TubP-Gal80ts .POLGa-IR)
a maximum lifespan of 87 days (Fig. 2D; Table 1). These results in significantly reduced lifespan compared with Tub.TubP-Gal80ts/+
controls. Kaplan– Meier data fitted to a Weibull distribution (dashed lines).
data demonstrate that adult-specific knockdown of POLGa Mean lifespans: Tub.TubP-Gal80ts .POLGa-IR ¼ 56.9, Tub.-
is sufficient to decrease mitochondrial gene expression and TubP-Gal80ts/+ ¼ 69.3.
to shorten lifespan in Drosophila.

Cell-type-specific mitochondrial dysfunction causes (Cha-Gal4) and tyrosine-hydroxylase-Gal4 (TH-Gal4) to

progressive behavioral deficits drive POLGa-RNAi and measured climbing as a basic
Alzheimer’s disease is associated with age-related degener- readout of neuronal function. Cha.POLGa-IR flies showed
ation of cholinergic neurons, whereas loss of dopaminergic an age-related decline in climbing performance similar to con-
neurons is observed in Parkinson’s disease (3). To assess the trols (Fig. 3A), with a steep decline that we attribute to the
effect of POLGa knockdown on these neuronal subpopula- presence of Dcr2 which is used to enhance RNAi efficacy
tions in aging flies, we used choline acetyltransferase-Gal4 (20). In contrast, TH.POLGa-IR flies showed a significant
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2701

Table 1. Lifespan statistics

Figure Genotype N Median Change Gehan–Wilcoxon 90% max. Change Log rank
x2 P-value Sig. x2 P-value Sig.

1C Act5C-GAL4/+ 94 72 80
Act.POLGa-IR1 49 24 267% 166 ,0.001 ∗∗∗
53 234% 184 ,0.001 ∗∗∗

1D Tub-Gal4/TubP-Gal80ts .+ 87 70 87
Tub-Gal4/ 45 59 216% 36.6 1.45 × 1029 ∗∗∗
70 220% 40.5 2 × 10210 ∗∗∗

TubP-Gal80ts .POLGa-IR1
4B TH.Dcr2, GFP 63 62 79
TH.Dcr2, GFP + POLGa-IR1 74 62 – 2.62 0.105 ns 71 210% 5.71 0.0169 ∗

4D TH.Dcr2, GFP 182 66 81

TH.Dcr2, GFP + POLGa-IR2 112 69 +4.5% 2.60 0.107 ns 83 +2.5% 4.62 0.0317 ∗

Dcr2/+, POLGa-IR2/+ 97 74 83
TH.Dcr2, GFP + POLGa-IR2 112 69 26.8% 3.36 0.0666 ns 83 – 0.214 0.644 ns
Dcr2/+, POLGa-IR2/+ 97 74 83
TH.Dcr2, GFP 182 66 211% 18.2 2.02 × 1025 ∗∗∗
81 22.4% 7.47 0.00628 ∗∗

∗
Significant at P , 0.05.
∗∗
Significant at P , 0.01.
∗∗∗
Significant at P , 0.001.

Figure 3. Mitochondrial dysfunction causes progressive motor deficits in aging Drosophila. (A) POLGa knockdown in cholinergic neurons (Cha.POLGa-IR)
does not affect climbing during aging, whereas (B) POLGa-IR in dopaminergic neurons (TH.POLGa-IR) causes significant and progressive decline in climbing
performance (Wilcoxon’s signed rank test). A decrease in climbing is already detectable by Day 10 and further declines by Day 40 and later. For detailed pro-
gression analysis, see Figure 4. (C) POLGa knockdown in serotonergic neurons (TRH.POLGa-IR) does not affect climbing performance. (D) Open field
motion tracking of aged, 62-day-old TH.POLGa-IR flies reveals (E) decreased activity, (F) reduced velocity and (G) a shorter distance traveled. Ten
minutes of tracking data are shown in trajectories (D). Thirty minutes of tracking data for control and experimental conditions were compared using Mann–
Whitney U-test. ∗ P , 0.05, ∗∗ P , 0.01, ∗∗∗ P , 0.001.

age-related decline in climbing performance when compared second, independent POLGa-IR2 strain showed a 2.4-fold
with controls (Fig. 3B) which was 4.3-fold greater than controls greater decline than controls (P , 0.001; Fig. 4C; Supplemen-
(P , 0.01; Fig. 4A; Supplementary Material, Table S1); a tary Material, Table S1). These effects were not due to
2702 Human Molecular Genetics, 2012, Vol. 21, No. 12

Figure 4. POLGa knockdown in dopaminergic cells causes progressive motor deficits but does not affect lifespan. (A) Linear regression fit of TH.Dcr2, GFP,
POLGa-IR1 compared with control (TH.Dcr2, GFP) shows significant and progressive decline in climbing performance. (B) Lifespan analysis of TH.Dcr2,
GFP, POLGa-IR1 reveals no significant changes versus control (Gehan–Wilcoxon test, P ¼ 0.105, Table 1). (C) A second, independent POLGa-RNAi strain
(TH.Dcr2, GFP, POLGa-IR2) shows similar age-related decline in climbing performance compared with control (TH.Dcr2, GFP). (D) Lifespan comparison
of flies expressing POLGa-IR2 in dopaminergic neurons shows no alterations compared with controls (Gehan– Wilcoxon test, P ¼ 0.107, Table 1). (A, C) Linear
regression slopes compared using ANOVA. ∗∗ P , 0.01, ∗∗∗ P , 0.001; see also Supplementary Material, Table S1 for statistical analysis.

premature aging, as adult TH.POLGa-IR flies showed normal mitochondrial function. These data demonstrate that decreas-
lifespan (Fig. 4B and D; Table 1). ing mitochondrial gene expression can cause age-related and
We reasoned, however, that climbing could be specifically cell-type-specific behavioral deficits, indicating that dopamin-
sensitive to neuronal dysfunction of dopaminergic but not cho- ergic neurons, but not cholinergic or serotonergic neurons, are
linergic cells. To address this, we used an open-field paradigm specifically vulnerable to POLGa-RNAi-mediated decrease in
(21). When compared with controls, analysis of Day 62 mitochondrial gene expression.
Cha.POLGa-IR flies (Supplementary Material, Fig. S2A –
D) did not reveal any differences in spontaneous activity,
speed of walking or total distance traveled. In contrast, TH.- Decreased mitochondrial gene expression sensitizes specific
POLGa-IR flies showed decreased activity (Fig. 3D and E) neurons to degeneration
and, when moving, they walked at a slower speed (Fig. 3F) We next investigated whether POLGa-RNAi can lead to
and covered less total distance (Fig. 3G) compared with con- age-related neurodegeneration. We examined mid-aged (Day
trols, consistent with the observed decline in climbing per- 31) and old-aged (Day 70) brains of flies with or without
formance of TH.POLGa-IR flies. POLGa-RNAi targeted to dopaminergic, serotonergic or cho-
To address the possibility that cell-type-specific phenotypes linergic neurons. To visualize and number targeted neurons,
might be due to neurotransmitter-related effects, we used we expressed a membrane-bound form of green fluorescent
tryptophan-hydroxylase-Gal4 (TRH-Gal4) to target POLGa- protein (GFP) (UAS-mCD8::GFP) alone or together with
RNAi to serotonergic neurons and analyzed both climbing UAS-POLGa-RNAi in a cell-type-specific manner (Supple-
and open field behavior in aging TRH.POLGa-IR flies. mentary Material, Fig. S3). The number of GFP-positive
When compared with age-matched controls, TRH.- cells was counted to determine the neuron survival at Day
POLGa-IR flies did not show decreased climbing performance 31 and Day 70 (Fig. 5A – D).
(Fig. 3C); they moved as fast as control flies and covered as Analysis of Cha.POLGa-IR flies did not show any neuron
much distance as controls (Supplementary Material, loss (Fig. 5A), whereas TRH.POLGa-IR flies revealed sero-
Fig. S2E– H). However, a small but significantly decreased ac- tonergic neuron loss only in old-aged flies (Fig. 5B), suggest-
tivity was detectable in aged TRH.POLGa-IR flies (Supple- ing that serotonergic neurons are partially vulnerable to
mentary Material, Fig. S2F), suggesting that serotonergic decreased mitochondrial gene expression but only at old age
neurons of old-aged flies become vulnerable to decreased in Drosophila. Conversely, TH.POLGa-IR flies showed a
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2703

Figure 5. Mitochondrial dysfunction causes age-related and cell-type-specific neurodegeneration. (A) Cholinergic neuron numbers with or without POLGa-IR
are unaltered at Day 31 and Day 70. (B) POLGa knockdown in serotonergic neurons only affects cell number at Day 70. (C) TH-Gal4-driven POLGa-IR affects
neuron numbers at Day 31, with progressive loss until Day 70. Analysis of TH.POLGa-IR at Day 5 revealed no significant neuron loss compared with Day 30
(P ¼ 0.45), demonstrating adult-onset neurodegeneration. (D) Progressive neurodegeneration affects individual dopaminergic cell clusters. For all three geno-
types, the numbers of GFP-labeled cells were recorded per brain hemisphere. ∗ P , 0.05, ∗∗ P , 0.01, ∗∗∗ P , 0.001. Error bars, SEM.

significant and age-related decrease in dopaminergic neurons. POLGa-RNAi-mediated decrease in mitochondrial gene ex-
When compared with age-matched controls, knockdown of pression can cause adult-onset and progressive dopaminergic
POLGa caused an average loss of 14/80 dopaminergic neurodegeneration.
neurons by Day 31 and an average loss of 20/80 dopaminergic Dopaminergic neurons in the adult fly brain can be grouped
cells by Day 70 (Fig. 5C). Two-way analysis of variance into clusters that show stereotypical axonal projections to
(ANOVA) and post-hoc analysis of Day 31 and Day 70 data target regions that represent distinct functional areas (21)
revealed that TH.POLGa-IR caused a significant 16.5% (Supplementary Material, Fig. S3A – D). We therefore charac-
loss of dopaminergic neurons at Day 31 and a larger 23.4% terized cluster-specific dopaminergic neuron loss in TH.-
loss of dopaminergic neurons at Day 70 compared with con- POLGa-IR flies (Fig. 5D). In controls, neuron numbers at
trols. This phenotype was confirmed by significantly increased mid-age and late-life remained the same in each cluster ana-
dopaminergic neuron loss between Day 31 and Day 70 lyzed, as observed previously (21). In contrast, TH.-
(Fig. 5C; Supplementary Material, Fig. S4), suggesting pro- POLGa-IR flies showed cluster-specific and age-related
gressive neurodegeneration. In comparison, controls neuron loss. A significantly increased loss of dopaminergic
(TH.Dcr2, mCD8::GFP) showed no cell loss between age neurons was seen in the PPL1 and PPM1/2 clusters when com-
groups (Fig. 5C), as observed previously (21). Moreover, paring mid-aged and old-aged TH.POLGa-IR flies, whereas
neuron numbers were unaltered in 5-day-old TH.POLGa-IR neuron loss occurred early in PPM3 and PAL clusters, but did
flies (Fig. 5C). Together, these data demonstrate that not increase significantly with age (Fig. 5D).
2704 Human Molecular Genetics, 2012, Vol. 21, No. 12

Figure 6. Mitochondria-mediated neurodegeneration is cell-type-specific. (A, B) Confocal images of whole-mount adult brain immunolabeled with anti-TH visu-
alizing dopaminergic neurons. (C, D) Anti-5HT immunolabeling of serotonergic neurons whole-mount adult brain. (E–G) Pan-neuronal Elav-Gal4-driven
UAS-POLGa-RNAi knockdown specifically affects dopaminergic neurons (E) but not serotonergic neurons (F). (G) Loss of anti-TH labeled cells indicates
loss of dopaminergic neurons in aging Elav.POLGa-IR flies; the observed degenerative cell loss affects all dopaminergic cell clusters except PAL. For
neuron counts, numbers of anti-5HT and anti-TH labeled cells were recorded per brain hemisphere, respectively. Scale bars: 100 mm. ∗ P , 0.05,
∗∗
P , 0.01, ∗∗∗ P , 0.001. Error bars, SEM.

Table 2. Genetic rescue of lethality caused by POLGa knockdown

Responder Actin-Gal4 CyO Total % with Actin-Gal4 % Expected versus w1118 % Expected versus genetic control

w1118 57 61 118 48.31 100.0

UAS-POLGa-IR 12 283 295 4.07 8.4
UAS-DTH.g4 69 241 310 22.26 46.1 100.0
UAS-POLGa-IR; UAS-DTH.g4 4 152 156 2.56 5.3 11.5
UAS-Drp1WT 319 308 627 50.88 105.3 100.0
UAS-POLGa-IR; UAS-Drp1WT 140 290 430 32.56 67.4 64.0
UAS-parkin2.1 155 326 481 32.22 66.7 100.0
UAS-POLGa-IR; UAS-parkin2.1 38 245 283 13.43 27.8 41.7
UAS-dhd 80 58 138 57.97 120.0 100.0
UAS-POLGa-IR; UAS-dhd 84 130 214 39.25 81.3 67.7
UAS-Ndi1A46 233 314 547 42.60 88.2 100.0
UAS-POLGa-IR; UAS-Ndi1A46 70 155 225 31.11 64.4 73.0
UAS-AOXF24-3 122 265 387 31.52 65.3 100.0
UAS-POLGa-IR; UAS-AOXF24-3 93 154 247 37.65 77.9 119.4

Number of flies eclosed with Actin-Gal4-driven UAS-POLGa-IR either alone or in combination with selected genes (responder). Actin-Gal4/CyO crossed to w1118
served as control. ‘Actin-Gal4’ shows the number of flies eclosed with Actin-Gal4 driver; ‘CyO’ shows the number of flies eclosed with CyO balancer; ‘Total’
shows the total number of flies eclosed; ‘% with Actin-Gal4’ shows the number of Actin-Gal4 flies divided by the total number of eclosed flies. The 48.3% eclosed
flies from control cross Actin-Gal4/CyO to w1118 without CyO were set as 100% expected w1118 eclosion. ‘% Expected versus w1118’ shows the percentage of
Actin-Gal4 flies eclosed divided by 48.3%. ‘% Expected versus genetic control’ shows percentage expected versus w1118 divided by the respective genetic control
data in ‘% Expected versus w1118’. TH, tyrosine hydroxylase; Drp1, dynamin-related protein 1; dhd, deadhead; Ndi1, yeast-derived NADH dehydrogenase Ndi1;
AOX, C. intestinalis-derived Alternative oxidase.

To exclude the possibility that differences in neuron loss Parkinson’s disease-related genes partially rescue POLGa
were dependent on Gal4 drivers used, we utilized the pan- knockdown effects
neuronal Elav-Gal4 driver. We counted the number of
neurons using antibodies specific to tyrosine hydroxylase In Parkinson’s disease, mitochondrial dysfunction and
(anti-TH) (Fig. 6A and B; Supplementary Material, Fig. S3A age-related loss of dopaminergic neurons has been linked to
and C) or serotonin (anti-5HT) (Fig. 6C and D; Supplementary pathogenic PINK1/Parkin signaling and altered fission/fusion
Material, Fig. S3E and G) and observed a significant loss of (10). We therefore wondered whether targeting Parkinson’s
dopaminergic neurons (Fig. 6E), but not serotonergic disease-related genes implicated in PINK1/Parkin signaling
neurons (Fig. 6F) in 60-day-old Elav.POLGa-IR flies when or mitochondrial fission/fusion could ameliorate Act.-
compared with age-matched controls. In addition, we also POLGa-IR-induced lethality (Fig. 2B, Table 2) and
observed cluster-specific dopaminergic neuron loss in Elav.- co-expressed in an Act.POLGa-IR genetic background
POLGa-IR brains (Fig. 6G). These data confirm that either parkin that promotes mitophagy or dynamin related
decreased mitochondrial gene expression can trigger protein 1 (Drp1) which induces mitochondrial fission (10).
age-related and cell-type-specific neurodegeneration in Analysis of Act.POLGa-IR, park flies revealed that parkin
Drosophila. was able to partially rescue POLGa-RNAi-mediated lethality
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2705

Figure 7. AOX restores ATP levels and rescues dopaminergic neurodegeneration. (A) Strong knockdown of POLGa causes lethality, also in the presence of
tyrosine hydroxylase; compare with Act.POLGa-IR control in Figure 2B. Act.POLGa-IR-induced lethality can be partially rectified by simultaneous expres-
sion of (B) Parkin involved in mitophagy; (C) dynamin-related protein 1, Drp1, involved in mitochondrial remodeling; (D) the thioredoxin homolog and anti-
oxidant deadhead, dhd; and (E) yeast-derived NADH dehydrogenase Ndi1. (F) In contrast, sea squirt-derived AOX fully rescued Act.POLGa-IR-induced
lethality. (G) POLGa-RNAi-driven dopaminergic neurodegeneration at Day 31 can be rescued with AOX but not Ndi1 expression. Control and POLGa-IR
(Day 31) neuron counts from Figure 5B and D are shown for clarity. (H) AOX expression restores ATP production in adult POLGa knockdown flies (n ¼ 5
in duplicate). ∗ P , 0.05, ∗∗ P , 0.01, ∗∗∗ P , 0.001. Error bars, SEM.

by 41.7% when compared with Act.park control (Fig. 7B). by 67.7% when compared with control (Act.dhd; Fig. 7D).
Co-expression of Drp1 with POLGa-RNAi (Act.POLGa-IR, These data show that dhd partially rescues
Drp1) rescued lethality by 64.0% when compared with control POLGa-RNAi-mediated lethality.
(Act.Drp1; Fig. 7C). These data suggest that the Parkinson’s
disease-related genes parkin and Drp1 partially rescue
POLGa-RNAi-mediated lethality. Bypassing respiratory chain deficiency rescues
Oxidative stress has been related to mitochondrial dysfunc- cell-type-specific neurodegeneration
tion and dopamine-specific neurodegeneration (4). We there- Knockdown of POLGa causes decreased mitochondrial gene
fore wondered whether targeting the antioxidative response expression and in turn decreases both respiratory activity
might be able to rescue Act.POLGa-IR-induced lethality. and ATP levels (Fig. 1). We therefore tested whether compo-
We chose to target thioredoxin, an evolutionarily conserved nents of the respiratory chain have impact on Act.-
antioxidant and redox enzyme (22), that is able to block Par- POLGa-IR-induced lethality. For this, we used the NADH
kinson’s disease-related MPP+-toxicity affecting mitochon- dehydrogenase Ndi1 from Saccharomyces cerevisiae and
drial complex I activity in human cells (23) and can rescue AOX from the sea squirt, Ciona intestinalis. Ndi1 consists of
behavioral deficits and selective loss of dopaminergic a single polypeptide chain that acts as the main entry point
neurons in Drosophila (24). Expression of the Drosophila to the respiratory chain similar to mammalian mitochondrial
thioredoxin homolog deadhead (dhd) alone was benign and complex 1 (25). Ndi1 can bypass Complex I deficiencies
had no effect on viability (Table 2). Co-expression of dhd and, when expressed in Drosophila, is able to mitigate the
with POLGa-RNAi (Act.POLGa-IR, dhd) rescued lethality age-related decline in respiratory capacity and the
2706 Human Molecular Genetics, 2012, Vol. 21, No. 12

accompanying increase in mitochondrial reactive oxygen alleviate. These findings have implications for understanding
species (ROS) (26). AOX is a single-subunit ubiquinol mitochondria-mediated mechanisms of aging and
oxidase found in plant mitochondria, as well as in the cell-type-specific neurodegeneration and they identify AOX
mitochondria of fungi, protists and animals, but not in arthro- as a potential therapeutic tool for respiratory chain deficiencies
pods or vertebrates (27). AOX from C. intestinalis is able to in mtDNA disorders and neurodegenerative diseases.
complement cytochrome c oxidase deficiency in human cells
(28) and in Drosophila can bypass Complex III and IV defi-
ciencies (29). A Drosophila model of mtDNA-mediated mitochondrial
Ectopic expression of Ndi1 partially rescued POLGa- dysfunction
RNAi-mediated lethality as 73.0% of Act.POLGa-IR, Ndi1
flies eclosed (Fig. 7E) when compared with controls We used targeted knockdown of POLGa to model mitochon-
(Act.Ndi1, P ¼ 0.003). The proportion of eclosed Act.- drial dysfunction in Drosophila. POLGa encodes the catalytic
POLGa-IR, Ndi1 flies was significantly greater than Act.- subunit of mtDNA POLGa which is the sole mitochondrial
POLGa-IR flies (Fig. 2B). Analysis of Act.POLGa-IR, polymerase required for mtDNA transcription (18).
AOX flies showed that ubiquitous co-expression of AOX In Drosophila, as in other eukaryotes, mtDNA encodes
with POLGa-IR completely rescued POLGa-RNAi lethality several components of mitochondrial complexes I, III, IV
(Fig. 7F). Significantly, all Act.POLGa-IR, AOX flies and V of the electron transport chain. Our Drosophila model
eclosed when compared with control. This AOX-mediated of mitochondrial dysfunction decreases the availability of
rescue of POLGa-RNAi-induced lethality was unlikely to be functional wild-type POLGa and in turn mtDNA-encoded pro-
caused by Gal4 dilution as ubiquitous co-expression of TH teins required for mitochondrial respiration. Previous studies
with POLGa-IR (11.5%, Fig. 7A) closely resembled Act.- suggested that a minimal threshold level of mitochondrial
POLGa-IR (8.4%, Fig. 2B). damage or decline is required to cause respiratory chain defi-
Next we determined whether AOX and Ndi1 were able to ciency and mitochondrial dysfunction (17) that can impact on
rescue dopamine-specific neurodegeneration caused by aging (30). In our model, the resulting phenotypic effects of
POLGa knockdown. As a control, we ectopically expressed mitochondrial dysfunction depend on transgene activity and
Ndi1 or AOX by TH-Gal4, and neither resulted in any loss hence Gal4 driver and UAS responder efficacy. Ubiquitous
of dopaminergic neurons (Supplementary Material, Fig. S4). POLGa knockdown causes either larval/pupal lethality or
When expressed together with POLGa-RNAi, analysis of shortened lifespan. Our data therefore provide experimental
TH.POLGa-IR, Ndi1 flies revealed that yeast Ndi1 was not evidence that phenotypic threshold effects of mitochondrial
able to prevent neurodegeneration in any of the dopaminergic dysfunction (31) impact on aging and survival in Drosophila.
clusters (Fig. 7G). In contrast, analysis of 31-day-old TH.- Aging in Caenorhabditis elegans and Drosophila (32,33), fish
POLGa-IR, AOX flies showed that expression of AOX was (34) as well as in mice (35), rats (36) and humans (37) has
sufficient to protect against neurodegeneration in all dopamin- been characterized by downregulation of mitochondrial
ergic neuron clusters that are normally affected in age- genes, including those encoded by mtDNA (38). Our results
matched TH.POLGa-IR flies (Fig. 7G). demonstrate that adult-onset knockdown of Drosophila
When expressed in Drosophila, AOX can complement POLGa is sufficient to shorten lifespan, indicating that de-
defects in mitochondrial oxidative phosphorylation (29). We creasing mitochondrial gene expression can recapitulate an
therefore wanted to know whether the neuroprotective poten- evolutionarily conserved molecular aging mechanism (35,39).
tial of AOX expression (Fig. 7G) might be mediated by a res- In addition to an aging phenotype, our POLGa-IR experi-
toration of mitochondrial metabolism in POLGa knockdown ments establish that cell-type-specific decreases in mitochon-
flies. Hence we measured total ATP content of adult Act.- drial gene expression can trigger progressive motor deficits
POLGa-IR, AOX flies and compared them with adult age- and dopaminergic neurodegeneration that resemble major
matched Act.POLGa-IR and Act.AOX flies. Although manifestations of familial and sporadic Parkinson’s disease
AOX expression on its own had no significant effect on ATP (40). These progressive phenotypes do not occur in young
levels (P ¼ 0.307), co-expression of AOX with POLGa-RNAi flies, but require advancing age before cell loss is observed.
significantly improved ATP levels and returned them to It is conceivable that the progressive phenotypes observed in
control levels when compared with Act.POLGa-IR flies our fly model are the combined consequence of
(Fig. 7H). Together, these data show that AOX expression in POLGa-IR-induced low base-line levels of mitochondrial
Drosophila is able to compensate for POLGa-RNAi-mediated gene expression together with the natural age-related
lethality, decreased ATP levels and cell-type-specific neurode- decline. We previously showed that under normal wild-type
generation. conditions, at least in Drosophila, age-related dopaminergic
neurodegeneration is not typically observed (21), suggesting
that senescence-related decrease in mitochondrial gene expres-
sion of wild-type Drosophila is insufficient to cause neurode-
DISCUSSION
generation. Together with our previous data, our POLGa-IR
Our results demonstrate that decreasing levels of mitochon- findings therefore suggest that phenotypic manifestation of
drial gene expression in Drosophila results in premature mitochondrial dysfunction occurs only when a threshold
aging and age-related as well as progressive and level is exceeded which in turn can cause age-related and
cell-type-specific neurodegeneration which AOX is able to cell-type-specific neurodegeneration.
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2707

Dopaminergic neurons are selectively vulnerable to effects include respiratory chain deficiency, lowered ATP
mitochondrial dysfunction levels and age-related dopaminergic neurodegeneration, sug-
gesting that defective respiration is causally related to
Mitochondrial dysfunction is commonly observed in neurode-
adult-onset and cell-type-specific neurodegeneration in
generative disorders that target different neuronal subpopula-
Drosophila.
tions (4,5). For example, cholinergic neurons preferentially
Previous studies identified Ndi1 and AOX as potential tools
degenerate in Alzheimer’s disease (41), whereas loss of dopa-
to counteract respiratory chain deficiencies in human cells and
minergic neurons is a pathological hallmark of Parkinson’s
disease (7). This selective, disease and cell-type-specific neu- animals (28,50,51). We have used both enzymes for heterol-
rodegeneration raises the question how particular neuronal ogous expression in Drosophila in an attempt to counteract
sub-populations are selectively vulnerable to disease forma- phenotypic effects of POLGa-IR. Yeast Ndi1 is a single-
tion. We show that in aging flies, decreasing mitochondrial subunit enzyme of the mitochondrial matrix catalyzing
gene expression specifically targets dopaminergic neurons NADH-quinone oxidoreduction similar to complex I but
for progressive neurodegeneration. POLGa-IR did not affect without proton pumping function. Application of Ndi1 is
the survival of cholinergic neurons when targeted to this neur- able to ameliorate phenotypes in acute models of Parkinson-
onal subpopulation, and when POLGa-IR was targeted to ser- ism where MPP+ or rotenone treatment directly affect mito-
otonergic neurons, phenotypic defects were only observed in chondrial complex I (51), and recent data suggest that it can
old aged flies. These data suggest that serotonergic and espe- rescue, at least to some extent, altered mitochondrial morph-
cially dopaminergic neurons are specifically vulnerable to de- ology and synaptic dysfunction in PINK1-mutant flies (52).
fective mitochondria. Interestingly, however, Ndi1 was not able to rescue lethality
Dopaminergic neurons are characterized by high- and dopaminergic neurodegeneration in our POLGa-IR-
intracellular calcium levels and high-energy demands that mediated model of mitochondrial dysfunction. POLGa-IR
are partly due to widely ramified axonal branches and dense affects the expression not only of mitochondrial complex I
dendritic arborizations (40)—also seen in flies (21)—and components, but also of components of complexes III, IV
require effective axonal transport of mitochondria (42). De- and V that are encoded by mtDNA, indicating that bypassing
fective sequestration of dopamine into vesicles can lead to complex I with Ndi1 is insufficient to compensate for the
cytoplasmic accumulation and auto-oxidation which ultimate- phenotypic effects of mtDNA-mediated mitochondrial dys-
ly leads to the generation of reactive oxygen and other toxic function. Moreover, this also indicates that respiratory chain
species that can damage nucleic acid, proteins and lipids deficiencies other than complex I, such as complex IV defi-
(40). Our data show that POLGa-IR-mediated mitochondrial ciencies also seen in Parkinson’s disease (53), significantly
dysfunction specifically sensitizes dopaminergic neurons to contribute to age-related neurodegeneration.
We identified AOX as a potent alternative enzyme to coun-
degenerative cell death in aging flies. Time-of-onset and pro-
teract mtDNA-mediated respiratory chain deficiency and mito-
gression of dopaminergic neurodegeneration suggest that the
chondrial dysfunction. Previous studies have shown that AOX
pathological process depends on a combination of age-related
can partially replace the electron transfer chain by directly re-
mitochondrial dysfunction and at least another cell-type-
ceiving electrons from ubiquinol to reduce molecular oxygen
specific stressor (43). Lowering levels of tyrosine hydroxylase,
to water, thus bypassing complexes III and IV of the oxidative
the rate-limiting enzyme for dopamine synthesis, has been
phosphorylation system; it can prevent metabolic acidosis and
shown to ameliorate a-synuclein- or rotenone-induced pheno-
types in Drosophila models of Parkinson’s disease (44). This overproduction of the quinone pool and complements respira-
suggests that in combination with POLGa-RNAi, dopamine tory deficiencies in human cells (28,50). When expressed in
or dopamine-related stressors might be sufficient to trigger Drosophila, AOX decreases mitochondrial ROS production
phenotypic threshold effects in aging Drosophila that cause and compensates phenotypes caused by mutations affecting
age-related and progressive degeneration of dopaminergic complex IV or the Parkinson’s disease-related gene DJ-1b
neurons. (29,54). Our findings demonstrate that AOX expression in
Drosophila is able to restore POLGa-IR-mediated phenotypes,
including reduced ATP levels and adult-onset, age-related
dopaminergic neurodegeneration in the presence of defective
Respiratory chain deficiency can cause age-related mitochondria. Together with previous studies, this suggests
neurodegeneration that AOX may prove useful for limiting respiratory chain defi-
Respiratory chain deficiencies are frequently observed in neu- ciencies caused by mtDNA decline as seen in healthy aging
rodegenerative diseases (4,5) and several lines of evidence (38,55) and age-related neurodegenerative diseases (7,56).
suggest that defective respiration specifically contributes to
dopaminergic cell loss in Parkinson’s disease. For example,
mutation of the Parkinson’s disease-associated gene PINK1 MATERIALS AND METHODS
has been associated with complex 1 deficiency (45– 48) and
genetically induced respiratory chain deficiencies in mice are Drosophila strains and genetics
associated with fragmentation of the mitochondrial network Flies were obtained from the Bloomington Stock Centre and
and dopaminergic cell loss (49). We used targeted knockdown raised at 258C in a 12 h light/dark cycle using a standard/
of POLGa to model mitochondrial dysfunction and our data agar diet unless stated otherwise. For POLGa knockdown,
show that POLGa-IR results in decreased expression of we used UAS-POLGa-IR lines from the VDRC (#106955;
mtDNA-encoded genes and proteins. The resulting phenotypic mRNA: 2176 $ 2852; and #3133; mRNA: 3075 $ 3359).
2708 Human Molecular Genetics, 2012, Vol. 21, No. 12

w,UAS-Dcr2 (VDRC) was used to enhance the efficiency of Amplified DNA was separated by electrophoresis in 1.1%
RNA interference (20). agarose. DNA was visualized using ethidium bromide and
For time- and tissue-specific knockdown, we used da-Gal4 digitally imaged by a computer-mounted camera. Grey value
(a kind gift from J. Bateman) active throughout development; of the DNA bands captured in the digital gel images were
Tub-Gal4 (a kind gift from G. Technau), a strong Gal4 driver, measured using plot profile in FIJI image processing
and Act5C-GAL4, a medium Gal4 driver, both active through- package (http://pacific.mpi-cbg.de).
out development, adulthood and aging; w,TubP-Gal80ts/(FM7)
(Bloomington) was combined with Tub-Gal4 to drive
temperature-sensitive knockdown restricted at 188C and mtDNA analysis
active at 258C; w,UAS-Dcr2; P{UAS-mCD8::GFP.L}LL5; Total DNA was extracted from experimental and control flies
TH-Gal4/TM6B,Sb specific to dopaminergic neurons (57); using lithium chloride/potassium acetate. Ten whole bodies
w,UAS-Dcr2; P{Cha-GAL4.7.4}19B, P{UAS-GFP.S65T}T2, per extraction were homogenized in 400 ml of buffer
specific to cholinergic neurons (58); w,UAS-Dcr2; [100 mM Tris– HCl, pH 7.5, 100 mM NaCl, 100 mM ethylene-
TRH-Gal4/CyO; UAS-mCD8::GFP, specific to serotonergic diaminetetraacetic acid (EDTA), 0.5% w/v sodium dodecyl
neurons (S. Birman). sulfate (SDS)] using a pellet pestle in a 1.5 ml tube and incu-
To rescue POLGa-RNAi phenotypes, we used: bated at 658C for 30 min to denature protein. Eight hundred
UAS-DTHg.4 for Tyrosine hydroxylase overexpression (59); microliters of 2:5 5 M KAc 6 M LiCl solution was added, the
UAS-parkin2.1 for Parkin overexpression (60) (a kind gift samples were incubated on ice for 15 min and centrifuged at
from Brian Staveley); UAS-drp1WT for Drp1 overexpression 16 000g for 15 min. The resultant supernatant was transferred
(61) (a kind gift from Ming Guo); UAS-dhd for Deadhead to a fresh tube and 600 ml of ice-cold isopropanol was added,
overexpression (24) (a kind gift from T. Aigaki); inverted to mix and centrifuged once again for 17 min. Isopro-
UAS-Ndi1A46 and UAS-AOXF2423 (29) for Ndi1 and AOX panol was discarded and the pellet was washed twice with
overexpression, respectively. 70% ethanol. The DNA pellet was re-suspended in 75 ml of
ddH2O. mtDNA-encoded genes were amplified using PCR.
Primers used were:
cytochrome c oxidase subunit I (COXI):
Reverse transcription-polymerase chain reaction forward primer: 5′ -GGTGCTCCTGATATAGCATTCCCA
Four flies per genotype were homogenized using a pestle CGA-3′ ;
(Fisher) in 125 ml of Trizol (Invitrogen) per fly until no reverse primer: 5′ -CTCCTCCTCCCGCTGGGTCA-3′ ;
body structures were identifiable. RNA extraction using cytochrome c oxidase subunit III (COXIII):
Trizol was performed following the manufacturer’s instruc- forward primer: 5′ -TGACCATTAACAGGAGCTATCGGA
tions. RNA was resuspended in nuclease-free GC-3′ ;
(diethylpyrocarbonate-treated) H2O using 1 ml of water per reverse primer: 5′ -TGATGCTCCTAATTCAATAGCGGG
10 ml of Trizol. RNA content was measured using a NanoDrop TGA-3′ ;
(Thermo Scientific); concentrations were typically 150 – cytochrome b (Cytb)
300 ng RNA/ml H2O. Isolated RNA was treated with
forward primer: 5′ -CGAACTTTACATGCTAACGGTGC
DNA-free (Ambion) to remove contaminating DNA following
ATCA-3′ ;
the manufacturer’s instructions.
reverse primer: 5′ -GCAGGTGTTACTAAAGGAGTTGCT
For the reverse transcription reaction, 1 mg of DNase-
GGA-3′ .
treated RNA was amplified for 60 min at 378C using mouse
Subsequent analysis was carried out as for RT-PCR.
megalovirus reverse transcriptase (M-MLV RT; Promega)
and random hexamer oligonucleotide primers (Fermentas) fol-
lowing the manufacturer’s instructions. M-MLV RT was inac- Western blotting of mitochondrially encoded proteins
tivated by heating to 708C for 15 min. cDNA was stored at
Flies were homogenized into 200 ml of lysis buffer [radioim-
2208C for later use.
munoprecipitation assay (RIPA) buffer: 50 mM Tris – HCl,
To measure POLGa RNA levels, cDNA was amplified
pH 7.5, 150 mM NaCl, 1% v/v NP-40, 5 mM EDTA, 0.5%
using polymerase chain reaction (PCR) in a series of increas-
w/v sodium deoxycholate, 0.1% SDS] with protease and phos-
ing cycle numbers to obtain the linear phase of amplification.
phatase inhibitor cocktails (1 cOmplete Mini (Roche) and 1
PCR reactions were carried out to compare POLGa transcripts
PhosSTOP (Roche) tablet in 7 ml of RIPA). Particulate
in experimental conditions; as a control, equal amounts of
matter was pelleted by centrifugation, and protein content in
cDNA present in the starting reaction was confirmed by meas-
the supernatant was measured using the BioRad Dc protein
uring d-actin.
assay (BioRad). Supernatant samples were prepared by dilut-
Primers used were:
ing samples 5:1 in 5× loading buffer (312.5 mM Tris – HCl,
polymerase g subunit a (POLGa) pH 6.8, 10% w/v SDS, 250 mM dithiothreitol, 50% v/v gly-
forward primer: 5′ -TCCATAACGGCAAAGGCGGTCG-3′ ; cerol) followed by boiling for 3 min. Samples were electro-
reverse primer: 5′ -TGCGGGACACACAAAAGGAGCG-3′ ; phoresed using a NuPage 4 – 12% Bis – Tris Gel and
d-actin transferred to HyBond ECL nitrocellulose membrane (GE
forward primer: 5′ -ACTTCTGCTGGAAGGTGGAC-3′ ; Healthcare). After blocking for 2 h using phosphate-buffered
reverse primer: 5′ -AATCCGCAAGGATCTGTATGC-3′ . saline supplemented with 0.05% Triton X-100 and 5%
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2709

non-fat milk (M-PBST), proteins of interest were detected as a movement per frame above a velocity of 0.5 mm/s.
using anti-ND5 (1:100; Abcam), anti-COXI (1:5000; MitoS- Average activity was the percentage of frames where the fly
ciences), anti-ATP synthase subunit a (1:20 000; MitoS- was active (.0.5 mm/s velocity). Mean velocity was the
ciences) and anti-b-tubulin (1:200; Developmental Studies average of velocities in each frame of the recording only
Hybridoma Bank) incubated overnight at 48C. Horseradish when the fly was active.
peroxidase-conjugated secondary antibodies were used at dilu-
tions optimized for each primary antibody and detected using
electrochemiluminescence. Images were collected digitally Immunohistochemistry
and band intensities were quantified by densitometry using Immunolabeling of adult brains was carried out as described
the GeneTools image analysis software (Syngene) and normal- previously (21). Primary antibodies were: 1:50 mouse anti-
ized for protein loading using b-tubulin. tyrosine hydroxylase (a-TH; Immunostar, 22941) to stain
dopaminergic neurons, 1:500 rabbit anti-serotonin (a-5HT;
Sigma-Aldrich, S5545) to stain serotonergic neurons and
ATP measurements 1:100 mouse anti-choline acetyltransferase (a-ChAT; Devel-
Four to five pupae were homogenized in 50 ml of lysis buffer opmental Studies Hybridoma Bank [DSHB]) to stain choliner-
(100 mM Tris, 4 mM EDTA) with 6 M guanidine. Five adult gic neurons. Secondary antibodies were goat anti-mouse or
flies were homogenized into 200 ml of lysis buffer without anti-rabbit 568 Alexa fluorochromes at 1:150 (Invitrogen).
guanidine and snap-frozen in liquid nitrogen. Samples were
boiled for 3 min followed by centrifugation for 5 min at
8000g. Supernatant from adult samples was diluted 1:50 Confocal microscopy and image processing
using extraction buffer. ATP levels were determined from Confocal microscopy and image acquisition were performed
5 ml of lysate using ATP Determination Kit (Invitrogen) in a with a Leica TCS SP5 laser confocal microscope and Leica
Wallac Victor2 Multilabel photon counter (ThermoFisher) at Application Suite Advanced Fluorescence (LASAF) version
288C as described previously (62) and normalized to protein 2.0.2 software. Images were scanned at 1024 × 1024, with a
content using a Bradford assay (Bio-Rad). slice thickness of 1.75 mm and a line and frame average of
2. Different channels were scanned sequentially to avoid
bleed-through. Confocal stacks were loaded into FIJI image
Complex I activity measurements processing package (http://pacific.mpi-cbg.de) using LOCI
Complex I activity was measured using a 96-well colorimetric Plugins for ImageJ version dev-4.2.7503 (http://www.loci.
assay as described previously (62). wisc.edu/software). Z-projections were created and analyzed
using FIJI image processing package.
Startle-induced negative geotaxis assay
Neuron counting
Climbing performance was analyzed using a startle-induced
negative geotaxis assay as described previously (21,63). For cell counting, ImageJ Cell Counter Plugin (Kurt De Vos,
http://rsbweb.nih.gov/ij/plugins/cell-counter.html) was used.
For counting of DA neurons, whole-mount heterozygous
Motion tracking TH.Dcr2, mCD8::GFP females with or without POLGa-IR
Open-field locomotion was analyzed using an assay described were analyzed as described previously (21) with the exception
previously (21) with the following modifications: the arena for of PPL1, where we did not include the more lateral cells in
video motion tracking was constructed from the lid of a 9 cM that cluster. For serotonergic neurons, whole-mount heterozy-
plastic Petri dish. Twelve flies were briefly anesthetized with gous TRH.Dcr2, mCD8::GFP females with or without
CO2 and placed in an arena, which was then placed above POLGa-IR were counterstained for serotonin and those cells
an array of white light-emitting diodes within a temperature- overlapping GFP fluorescence with anti-serotonin were
controlled incubator (Stuart Scientific). Flies were allowed to counted. For cholinergic neurons, two regions of whole-mount
recover for 1 h before recordings were made at 25 frames heterozygous Cha.Dcr2, mCD8::GFP female brains with or
per second for 30 min with Virtaldub (http://www.virtua without POLGa-IR were analyzed. Anteriorly, the cell cluster
ldub.org). Videos were compressed as they were being immediately adjacent to the mushroom body calyx contained
recorded with an MPEG-4-compatible Xvid codec from within one Z-stack slice was counted. Posteriorly, a 20 mm
ffdshow tryouts (http://ffdshow-tryout.sourceforge.net/). cube was selected and the cell bodies contained within that
Recorded videos were converted to fly movie format from region were counted. For all three genotypes, the numbers of
the motmot package (64) which was loaded into Ctrax soft- cells per hemisphere were recorded. For brains from Elav.
ware (65) to analyze the walking positions of the flies Dcr2 flies with or without POLGa-IR, the number of cell
through the video. Position data for 30 min was exported as bodies stained with anti-TH for DA neurons or anti-5HT for
a Matlab-compatable (Mathworks) matrix file. Errors in the serotonergic neurons were counted at Day 61.
tracking were fixed using FixErrors GUI (65) in Matlab
(Mathworks). Fixed trajectories were loaded into GNU
Lifespan analysis
Octave 3.2.4 (http://www.gnu.org/software/octave/) and
custom scripts were used to determine mean velocity, mean Lifespan analysis was performed essentially as described pre-
activity and mean cumulative distance. Activity was defined viously (21). Virgin females from the driver lines were crossed
2710 Human Molecular Genetics, 2012, Vol. 21, No. 12

with UAS-POLGa-RNAi males. Synchronized egg collection ‘+’ indicates a data point within 3× the interquartile range;
was carried out on apple agar plates (2.13% w/v agar, ‘W’ indicates data outside 3× the interquartile range. Due to
1.25% w/v sugar, 25% v/v apple juice, 0.2% w/v Nipagen), the non-normal distribution of behavioral data, significance
which were swapped twice a day. Lifespans were performed was assessed with Mann – Whitney U-test using the u_test()
on 15% sugar/yeast medium at a maximum density of 15 function. Significance is indicated unless otherwise stated:
∗
flies per tube at 258C. For temperature-sensitive repression P , 0.05; ∗∗ P , 0.01; and ∗∗∗ P , 0.001. All statistical
of Gal4 activity using TubP-Gal80ts transgene to measure tests used and their results are shown in Supplementary Mater-
the effect of adult-only POLGa-IR (Tub.TubP-Gal80ts.- ial, Table S1.
POLGa-IR), developmental stages were reared at 188C. Two
days after eclosion, adult flies were transferred to 258C to
carry out lifespans. The probability of surviving to time t, SUPPLEMENTARY MATERIAL
S(t) ¼ Pr(T . t), was calculated by the Kaplan – Meier sur-
vival function. Differences in lifespan were only considered Supplementary Material is available at HMG online.
if they were above 5% greater or lower than the median life-
span of the control group.
ACKNOWLEDGEMENTS

Survival assay (fatality/eclosion) We thank J. Bateman, M. Guo, B. Staveley, G. Technau, the

Bloomington Stock Centre, the Vienna Drosophila RNAi
Female virgin UAS-POLGa-RNAi flies were crossed with Centre and the Developmental Studies Hybridoma Bank at
male Actin5C-Gal4/CyO driver flies to generate heterozygous the University of Iowa for stocks and reagents. We also
Act.POLGa-IR expressor flies or POLGa-IR/CyO controls. thank A. Pooler and R. Humphrey for comments on the
Parent flies were left to lay eggs for 2 days and then trans- manuscript.
ferred to a fresh vial. The vial with laid eggs was allowed to
develop for 20 days and the number of progeny of each geno- Conflict of Interest statement. None declared.
type was counted. The proportion of flies eclosing was calcu-
lated as the number of Act.POLGa-IR flies eclosed divided
by the total number of flies eclosed, i.e. including the CyO in- FUNDING
ternal control. For the control genotypes, this was normalized
to 100%. For additional genotypes tested, e.g. POLGa-IR This work was supported by the UK Medical Research
in combination with UAS-Drp1WT, the experimental Council (G-0701498 to F.H.), the Royal Society (Hirth/2007/
condition (Act.POLGa-IR, Drp1) was compared against the R2 to F.H.), the Motor Neurone Disease Association (Hirth/
control condition without POLGa-IR, e.g. Act.Drp1 (see Oct07/6233 to F.H.), the Fondation Thierry Latran (2/2011/
Table 2). DrosAls to F.H.) and Parkinson’s UK (G-0714 to F.H.).

Statistical analysis REFERENCES

For analysis of variance (ANOVA), two-way ANOVA of 1. Winklhofer, K.F., Tatzelt, J. and Haass, C. (2008) The two faces of
neuron counts, lifespan analysis and linear regression compar- protein misfolding: gain- and loss-of-function in neurodegenerative
isons R Project 2.9.2 (http://www.r-project.org/) was used. To diseases. EMBO J., 27, 336–349.
2. Nussbaum, R.L. and Ellis, C.E. (2003) Alzheimer’s disease and
assess whether the slopes in the climbing experiments were Parkinson’s disease. N. Engl. J. Med., 348, 1356–1364.
statistically different between TH.POLG-IR and controls 3. Mayeux, R. (2003) Epidemiology of neurodegeneration. Annu. Rev.
flies, the lm() function was used with coincidence. Linear Neurosci., 26, 81– 104.
model data were then compared with the anova() function to 4. Lin, M.T. and Beal, M.F. (2006) Mitochondrial dysfunction and oxidative
generate the P-values. Kaplan –Meier estimator survfit() func- stress in neurodegenerative diseases. Nature, 443, 787 –795.
5. Schon, E.A. and Przedborski, S. (2011) Mitochondria: the next
tion from the R Project Survival package (http://cran.r-project. (neurode)generation. Neuron, 70, 1033–1053.
org/web/packages/survival/index.html) was used to generate 6. Galindo, M.F., Ikuta, I., Zhu, X., Casadesus, G. and Jordán, J. (2010)
lifespan trajectories. Log-rank and Gehan – Wilcoxon tests Mitochondrial biology in Alzheimer’s disease pathogenesis.
were used to compare the lifespans. Weibull distributions J. Neurochem., 114, 933–945.
k 7. Winklhofer, K.F. and Haass, C. (2010) Mitochondrial dysfunction in
(−e−(x/l) ) were fitted to the data where indicated using the Parkinson’s disease. Biochim. Biophys. Acta, 1802, 29–44.
survreg() function in R, where l is the scale and k is the 8. Damiano, M., Galvan, L., Déglon, N. and Brouillet, E. (2010)
shape parameters. Mean lifespans were determined by calcu- Mitochondria in Huntington’s disease. Biochim. Biophys. Acta, 1802,
lating lG(1+1/k) from the fitted Weibull distributions. For 52– 61.
9. Cozzolino, M. and Carrı̀, M.T. (2011) Mitochondrial dysfunction in ALS.
Chi-squared tests to calculate significance in the pupal lethal- Prog. Neurobiol. [Epub ahead of print August 2].
ity experiments, OpenOffice.org 3.1 Calc (http://www. 10. Büeler, H. (2010) Mitochondrial dynamics, cell death and the
openoffice.org/product/calc.html) CHITEST function was pathogenesis of Parkinson’s disease. Apoptosis, 15, 1336– 1353.
used (Pearson’s chi-squared test). For the motion tracking 11. Park, C.B. and Larsson, N.G. (2011) Mitochondrial DNA mutations in
experiments, data were analyzed and plotted, using box-plots, disease and aging. J. Cell Biol., 193, 809– 818.
12. Larsson, N.G., Wang, J., Wilhelmsson, H., Oldfors, A., Rustin, P.,
in GNU Octave 3.2.4 (http://www.gnu.org/software/octave/). Lewandoski, M., Barsh, G.S. and Clayton, D.A. (1998) Mitochondrial
In the box-plots, the boxes show median, upper and lower transcription factor A is necessary for mtDNA maintenance and
quartiles; whiskers contain data 1.5× the interquartile range; embryogenesis in mice. Nat. Genet., 18, 231–236.
 Human Molecular Genetics, 2012, Vol. 21, No. 12 2711

13. Ekstrand, M.I., Terzioglu, M., Galter, D., Zhu, S., Hofstetter, C., and gene expression of Drosophila melanogaster mitochondria.
Lindqvist, E., Thams, S., Bergstrand, A., Hansson, F.S., Trifunovic, A. Biochimie, 89, 988– 1001.
et al. (2007) Progressive parkinsonism in mice with 34. Hartmann, N., Reichwald, K., Wittig, I., Dröse, S., Schmeisser, S., Lück,
respiratory-chain-deficient dopamine neurons. Proc. Natl Acad. Sci. USA, C., Hahn, C., Graf, M., Gausmann, U., Terzibasi, E. et al. (2011)
104, 1325– 1330. Mitochondrial DNA copy number and function decrease with age in the
14. Trifunovic, A., Wredenberg, A., Falkenberg, M., Spelbrink, J.N., Rovio, short-lived fish Nothobranchius furzeri. Aging Cell, 10, 824– 831.
A.T., Bruder, C.E., Bohlooly-Y, M., Gidlöf, S., Oldfors, A., Wibom, R. 35. Bishop, N.A., Lu, T. and Yankner, B.A. (2010) Neural mechanisms of
et al. (2004) Premature ageing in mice expressing defective mitochondrial ageing and cognitive decline. Nature, 464, 529– 535.
DNA polymerase. Nature, 429, 417 –423. 36. Blalock, E.M., Chen, K.C., Sharrow, K., Herman, J.P., Porter, N.M.,
15. Kujoth, G.C., Hiona, A., Pugh, T.D., Someya, S., Panzer, K., Foster, T.C. and Landfield, P.W. (2003) Gene microarrays in hippocampal
Wohlgemuth, S.E., Hofer, T., Seo, A.Y., Sullivan, R., Jobling, W.A. et al. aging: statistical profiling identifies novel processes correlated with
(2005) Mitochondrial DNA mutations, oxidative stress, and apoptosis in cognitive impairment. J. Neurosci., 23, 3807–3819.
mammalian aging. Science, 309, 481– 484. 37. Lu, T., Pan, Y., Kao, S.Y., Li, C., Kohane, I., Chan, J. and Yankner, B.A.
16. Vermulst, M., Bielas, J.H., Kujoth, G.C., Ladiges, W.C., Rabinovitch, (2004) Gene regulation and DNA damage in the ageing human brain.
P.S., Prolla, T.A. and Loeb, L.A. (2007) Mitochondrial point mutations do Nature, 429, 883– 891.
not limit the natural lifespan of mice. Nat. Genet., 39, 540–543. 38. Larsson, N.G. (2010) Somatic mitochondrial DNA mutations in
17. Zeviani, M. and Di Donato, S. (2004) Mitochondrial disorders. Brain, mammalian aging. Annu. Rev. Biochem., 79, 683–706.
127, 2153– 2172. 39. Zahn, J.M., Poosala, S., Owen, A.B., Ingram, D.K., Lustig, A., Carter, A.,
18. Iyengar, B., Roote, J. and Campos, A.R. (1999) The tamas gene, identified Weeraratna, A.T., Taub, D.D., Gorospe, M., Mazan-Mamczarz, K. et al.
as a mutation that disrupts larval behavior in Drosophila melanogaster, (2007) AGEMAP: a gene expression database for aging in mice. PLoS
codes for the mitochondrial DNA polymerase catalytic subunit Genet., 3, e201.
(DNApol-g125). Genetics, 153, 1809– 1824. 40. Lotharius, J. and Brundin, P. (2002) Pathogenesis of Parkinson’s disease:
19. McGuire, S.E., Le, P.T., Osborn, A.J., Matsumoto, K. and Davis, R.L. dopamine, vesicles and alpha-synuclein. Nat. Rev. Neurosci., 3, 932 –942.
(2003) Spatiotemporal rescue of memory dysfunction in Drosophila. 41. Kása, P., Rakonczay, Z. and Gulya, K. (1997) The cholinergic system in
Science, 302, 1765–1768. Alzheimer’s disease. Prog. Neurobiol., 52, 511–535.
20. Dietzl, G., Chen, D., Schnorrer, F., Su, K.-C., Barinova, Y., Fellner, M., 42. De Vos, K.J., Grierson, A.J., Ackerley, S. and Miller, C.C. (2008) Role of
Gasser, B., Kinsey, K., Oppel, S., Scheiblauer, S. et al. (2007) A axonal transport in neurodegenerative diseases. Annu. Rev. Neurosci., 31,
genome-wide transgenic RNAi library for conditional gene inactivation in 151– 173.
Drosophila. Nature, 448, 151–156. 43. Saxena, S. and Caroni, P. (2011) Selective neuronal vulnerability in
21. White, K.E., Humphrey, D.M. and Hirth, F. (2010) The dopaminergic neurodegenerative diseases: from stressor thresholds to degeneration.
system in the aging brain of Drosophila. Front. Neurosci., 4, 205. Neuron, 71, 35– 48.
22. Maulik, N. and Das, D.K. (2008) Emerging potential of thioredoxin and 44. Bayersdorfer, F., Voigt, A., Schneuwly, S. and Botella, J.A. (2010)
Dopamine-dependent neurodegeneration in Drosophila models of familial
thioredoxin interacting proteins in various disease conditions. Biochim.
and sporadic Parkinson’s disease. Neurobiol. Dis., 40, 113–119.
Biophys. Acta, 1780, 1368–1382.
45. Gegg, M.E., Cooper, J.M., Schapira, A.H. and Taanman, J.W. (2009)
23. Andoh, T., Chock, P.B. and Chiueh, C.C. (2001) The roles of thioredoxin
Silencing of PINK1 expression affects mitochondrial DNA and oxidative
in protection against oxidative stress-induced apoptosis in human
phosphorylation in dopaminergic cells. PLoS One, 4, e4756.
SH-SY5Y cells. J. Biol. Chem., 277, 9655–9660.
46. Morais, V.A., Verstreken, P., Roethig, A., Smet, J., Snellinx, A.,
24. Umeda-Kameyama, Y., Tsuda, M., Ohkura, C., Matsuo, T., Namba, Y.,
Vanbrabant, M., Haddad, D., Frezza, C., Mandemakers, W.,
Ohuchi, Y. and Aigaki, T. (2007) Thioredoxin suppresses
Vogt-Weisenhorn, D. et al. (2009) Parkinson’s disease mutations in
Parkin-associated endothelin receptor-like receptor-induced neurotoxicity
PINK1 result in decreased Complex I activity and deficient synaptic
and extends longevity in Drosophila. J. Biol. Chem., 282, 11180– 11187. function. EMBO Mol. Med., 1, 99–111.
25. de Vries, S. and Grivell, L.A. (1988) Purification and characterization of a 47. Liu, W., Acı́n-Peréz, R., Geghman, K.D., Manfredi, G., Lu, B. and Li, C.
rotenone-insensitive NADH:Q6 oxidoreductase from mitochondria of (2011) Pink1 regulates the oxidative phosphorylation machinery via
Saccharomyces cerevisiae. Eur. J. Biochem., 176, 377–384. mitochondrial fission. Proc. Natl Acad. Sci. USA, 108, 12920–12924.
26. Sanz, A., Soikkeli, M., Portero-Otı́n, M., Wilson, A., Kemppainen, E., 48. Abramov, A.Y., Gegg, M., Grunewald, A., Wood, N.W., Klein, C. and
McIlroy, G., Ellilä, S., Kemppainen, K.K., Tuomela, T., Lakanmaa, M. Schapira, A.H. (2011) Bioenergetic consequences of PINK1 mutations in
et al. (2010) Expression of the yeast NADH dehydrogenase Ndi1 in Parkinson disease. PLoS One, 6, e25622.
Drosophila confers increased lifespan independently of dietary restriction. 49. Sterky, F.H., Lee, S., Wibom, R., Olson, L. and Larsson, N.G. (2011)
Proc. Natl Acad. Sci. USA, 107, 9105– 9110. Impaired mitochondrial transport and Parkin-independent degeneration of
27. McDonald, A.E., Vanlerberghe, G.C. and Staples, J.F. (2009) Alternative respiratory chain-deficient dopamine neurons in vivo. Proc. Natl Acad.
oxidase in animals: unique characteristics and taxonomic distribution. Sci. USA, 108, 12937–12942.
J. Exp. Biol., 212, 2627–2634. 50. Rustin, P. and Jacobs, H.T. (2009) Respiratory chain alternative enzymes
28. Dassa, E.P., Dufour, E., Goncalves, S., Paupe, V., Hakaart, G.A.J., Jacobs, as tools to better understand and counteract respiratory chain deficiencies
H.T. and Rustin, P. (2009) Expression of the alternative oxidase in human cells and animals. Physiol. Plant., 137, 362–370.
complements cytochrome c oxidase deficiency in human cells. EMBO 51. Marella, M., Seo, B.B., Yagi, T. and Matsuno-Yagi, A. (2009) Parkinson’s
Mol. Med., 1, 30– 36. disease and mitochondrial complex I: a perspective on the Ndi1 therapy.
29. Fernández-Ayala, D.J., Sanz, A., Vartiainen, S., Kemppainen, K.K., J. Bioenerg. Biomembr., 41, 493– 497.
Babusiak, M., Mustalahti, E., Costa, R., Tuomela, T., Zeviani, M., Chung, 52. Vilain, S., Esposito, G., Haddad, D., Schaap, O., Dobreva, M.P., Vos, M.,
J. et al. (2009) Expression of the Ciona intestinalis alternative oxidase Van Meensel, S., Morais, V.A., De Strooper, B. and Verstreken, P. (2012)
(AOX) in Drosophila complements defects in mitochondrial oxidative The yeast complex I equivalent NADH dehydrogenase rescues pink1
phosphorylation. Cell Metab., 9, 449– 460. mutants. PLoS Genet., 8, e1002456.
30. Lanza, I.R. and Nair, K.S. (2010) Mitochondrial function as a determinant 53. Benecke, R., Strümper, P. and Weiss, H. (1993) Electron transfer
of life span. Pflugers Arch., 459, 277–289. complexes I and IV of platelets are abnormal in Parkinson’s disease but
31. Rossignol, R., Faustin, B., Rocher, C., Malgat, M., Mazat, J.P. and normal in Parkinson-plus syndromes. Brain, 116, 1451–1463.
Letellier, T. (2003) Mitochondrial threshold effects. Biochem. J., 370, 54. Sanz, A., Fernández-Ayala, D.J., Stefanatos, R.K. and Jacobs, H.T. (2010)
751–762. Mitochondrial ROS production correlates with, but does not directly
32. McCarroll, S.A., Murphy, C.T., Zou, S., Pletcher, S.D., Chin, C.S., Jan, regulate lifespan in Drosophila. Aging, 2, 200– 223.
Y.N., Kenyon, C., Bargmann, C.I. and Li, H. (2004) Comparing genomic 55. Greaves, L.C. and Turnbull, D.M. (2009) Mitochondrial DNA mutations
expression patterns across species identifies shared transcriptional profile and ageing. Biochim. Biophys. Acta, 1790, 1015– 1020.
in aging. Nat. Genet., 36, 197 –204. 56. Reeve, A.K., Krishnan, K.J. and Turnbull, D. (2008) Mitochondrial DNA
33. Dubessay, P., Garreau-Balandier, I., Jarrousse, A.S., Fleuriet, A., Sion, B., mutations in disease, aging, and neurodegeneration. Ann. NY Acad. Sci.,
Debise, R. and Alziari, S. (2007) Aging impact on biochemical activities 1147, 21–29.
2712 Human Molecular Genetics, 2012, Vol. 21, No. 12

57. Friggi-Grelin, F., Coulom, H., Meller, M., Gomez, D., Hirsh, J. and inhibit fusion in Drosophila. Proc. Natl Acad. Sci. USA, 105, 4503–
Birman, S. (2003) Targeted gene expression in Drosophila dopaminergic 14508.
cells using regulatory sequences from tyrosine hydroxylase. J. Neurobiol., 62. Parsons, R.B., Aravindan, S., Kadampeswaran, A., Evans, E.A., Sandhu,
54, 618–627. K.K., Levy, E.R., Thomas, M.G., Austen, B.M. and Ramsden, D.B. (2011)
58. Salvaterra, P.M. and Kitamoto, T. (2001) Drosophila cholinergic neurons Expression of nicotinamide N-methyltransferase increases ATP synthesis
and processes visualized with Gal4/UAS-GFP. Brain Res. Gene Expr. and protects SH-SY5Y cells against the toxicity of Complex I inhibitors.
Patterns, 1, 73– 82. Biochem. J., 436, 145– 155.
59. Friggi-Grelin, F., Iché, M. and Birman, S. (2003) Tissue-specific 63. Rival, T., Soustelle, L., Strambi, C., Besson, M.T., Iché, M. and Birman, S.
developmental requirements of Drosophila tyrosine hydroxylase isoforms. (2004) Decreasing glutamate buffering capacity triggers oxidative stress and
Genesis, 35, 175–184. neuropil degeneration in the Drosophila brain. Curr. Biol., 14, 599–605.
60. Todd, A.M. and Staveley, B.E. (2008) Pink1 suppresses alpha-synuclein- 64. Straw, A.D. and Dickinson, M.H. (2009) Motmot, an open-source toolkit
induced phenotypes in a Drosophila model of Parkinson’s disease. for realtime video acquisition and analysis. Source Code Biol. Med., 4, 5.
Genome, 51, 1040– 1046. 65. Branson, K., Robie, A.A., Bender, J.A., Perona, P. and Dickinson, M.H.
61. Deng, H., Dodson, M.W., Huang, H. and Guo, M. (2008) The Parkinson’s (2009) High-throughput ethomics in large groups of Drosophila. Nat.
disease genes pink1 and parkin promote mitochondrial fission and/or Methods, 6, 451 –457.