Mitochondrial Mislocalization Underlies Aβ42-Induced Neuronal Dysfunction in a Drosophila Model of Alzheimer's Disease

The amyloid-β 42 (Aβ42) is thought to play a central role in the pathogenesis of Alzheimer's disease (AD). However, the molecular mechanisms by which Aβ42 induces neuronal dysfunction and degeneration remain elusive. Mitochondrial dysfunctions are implicated in AD brains. Whether mitochondrial dysfunctions are merely a consequence of AD pathology, or are early seminal events in AD pathogenesis remains to be determined. Here, we show that Aβ42 induces mitochondrial mislocalization, which contributes to Aβ42-induced neuronal dysfunction in a transgenic Drosophila model. In the Aβ42 fly brain, mitochondria were reduced in axons and dendrites, and accumulated in the somata without severe mitochondrial damage or neurodegeneration. In contrast, organization of microtubule or global axonal transport was not significantly altered at this stage. Aβ42-induced behavioral defects were exacerbated by genetic reductions in mitochondrial transport, and were modulated by cAMP levels and PKA activity. Levels of putative PKA substrate phosphoproteins were reduced in the Aβ42 fly brains. Importantly, perturbations in mitochondrial transport in neurons were sufficient to disrupt PKA signaling and induce late-onset behavioral deficits, suggesting a mechanism whereby mitochondrial mislocalization contributes to Aβ42-induced neuronal dysfunction. These results demonstrate that mislocalization of mitochondria underlies the pathogenic effects of Aβ42 in vivo.


Introduction
Alzheimer's disease (AD) is a progressive neurodegenerative disease without effective therapies. Pathologically, AD is defined by an extensive loss of neurons and by formation of two characteristic protein deposits, extracellular amyloid plaques (APs) and intracellular neurofibrillary tangles (NFTs). The major components of APs and NFTs are the 40 or 42 amino acid amyloid-b peptides (Ab40 or Ab42) and the hyperphosphorylated microtubule associated protein tau, respectively [1].
Molecular genetic studies of early-onset familial AD patients have identified causative mutations in genes encoding APP and presenilins (PS1 and PS2), and these mutations increase Ab42 production and/or Ab aggregation [2]. Ab42 is highly toxic to cultured neurons and causes memory deficits and neurodegeneration in animal models overproducing human Ab42 [3]. Thus, Ab42 is thought to play a causative role in the pathogenesis of AD [4].
Several lines of evidence indicate that mitochondrial function is impaired in the brains of AD patients [5,6,7,8]. Markedly reduced levels of mitochondrial proteins and activities and increased abnormal and damaged mitochondria have been reported in AD brains [8,9,10,11]. Whether mitochondrial dysfunctions are merely a consequence of AD pathology or are early seminal events in AD pathogenesis remains to be determined.
In order to identify genes and pathways that are involved in Ab42-induced toxicity in vivo, we are utilizing Drosophila as a model system. To produce human Ab42 in the secretory pathway of fly brain neurons, the Ab42 peptide sequence is directly fused to a secretion signal peptide at the N-terminus. Using a GAL4-UAS transgene expression system [12], Ab42 peptide was expressed in the fly brain. Mass spectrometry analysis revealed that this construct produces the intact Ab42 peptide in the fly brain [13,14], and immuno-electron microscopy analysis showed that expressed Ab42 was distributed in the secretory pathways in neurons in the fly brains [14]. These Ab42 flies show late-onset, progressive short-term memory defects, locomotor dysfunctions, neurodegeneration, and premature death, accompanied by formation of Ab42 deposits [13,14]. This or similar Drosophila models have been used to study mechanisms underlying neurotoxicity of Ab42 in vivo [3,15,16,17,18,19,20,21,22,23]. Using this Drosophila model [13,14], here we have demonstrated that mitochondrial mislocalization underlies the pathogenic effects of Ab42 in vivo.

Mitochondria Are Reduced in the Axons and Dendrites in Ab42 Fly Brain
Using mito-GFP transgene, a reporter construct in which GFP is fused to a mitochondrial targeting signal [24], we analyzed the distribution of mitochondria in the Ab42 fly brain. For this purpose, we focused on the mushroom body structure, where axons, dendrites, and cell bodies can be easily identified in the fly brain [25] ( Figure 1A). Mito-GFP and Ab42 were expressed in all neurons by the pan-neuronal elav-GAL4 driver.
The mito-GFP signal in the axons and dendrites of the mushroom body structure was significantly decreased in the Ab42 fly brains ( Figure 1B). In contrast, the mito-GFP signal was increased in the cell bodies of neurons ( Figure 1B). These results suggest that Ab42 does not cause global reduction of mitochondria, but rather induces mitochondrial mislocalization. A significant reduction in mitochondria was observed in the axons at 5 days after eclosion (dae), while a reduction in mitochondria in the dendrites was detected by 21 dae. Thus, Ab42-induced reduction of mitochondria in the axons occurs earlier than in the dendrites. Similar results were obtained from four independent Ab42 transgenic fly lines using the pan-neuronal elav-GAL4 driver ( Figure 1B), or the cholinergic neuron-specific driver, Cha-GAL4 ( Figure S1). Reduced mito-GFP signals in neuropil in Ab42 fly brains were also observed in other brain structures including the central complex ( Figure 1C), which is required for the maintenance of locomotor activity in flies [26].
Mitochondrial mislocalization observed in the Ab42 fly brains is not due to overexpression of exogeneous protein, since neuronal expression human a-synuclein [27], which is thought to play a critical role in Parkinson's disease, did not induce mislocalization of mitochondria in the fly brains at 21 dae ( Figure S2).
The reduction in mitochondria in the axons and dendrites is unlikely to be due to degeneration of the mushroom body structure, since neurodegeneration in the Ab42 fly brain is not prominent at 5 dae [13,14]. To confirm that the mushroom body structure has not degenerated, and to test whether Ab42 expression non-specifically alters the protein distribution in axons and dendrites, we analyzed the distribution of the membrane protein CD8 fused to a GFP reporter (CD8-GFP). Ab42 did not cause a noticeable morphological change of the mushroom body structures or significant reduction in the CD8-GFP signal in axons or dendrites at 21 dae ( Figure 1D).
Mitochondria are transported along microtubules by the motor proteins. To test whether Ab42-induced mitochondrial mislocalization is due to an overall disruption of microtubule-based transport in neurons, we analyzed distributions of tubulin fused to GFP (tub-GFP) and the presynaptic protein synaptotagmin fused to GFP (syt-GFP) in axons and dendrites. Ab42 expression did not result in any significant difference in the distributions of tub-GFP in axons and dendrites ( Figure 1E) or syt-GFP in axons and cell bodies ( Figure 1F) at 21 dae. These data suggest that Ab42-induced mislocalization of mitochondria is not due to disorganization of microtubule or global disruption of axonal transport in neurons.

Mitochondria Are Not Severely Damaged in Young Ab42 Fly Brains
Mitochondrial damage and dysfunction have been shown to alter mitochondrial localization. We examined whether Ab42 caused severe mitochondrial damage at the ages at which we observed mitochondrial mislocalization. We compared the amount of mitochondrial genomes and the levels of ATP in the brains dissected from control and Ab42 flies, and found that they were not significantly different (Figure 2A and B). Electron microscopic (EM) analysis did not detect noticeable alterations in mitochondrial morphology in the neuropil or cell bodies of Kenyon cell region in the Ab42 fly brain ( Figure 2C). These data suggest that mitochondrial mislocalization is not due to severe damage to the mitochondria in the Ab42 fly brain.
Apoptosis can cause mitochondrial fragmentation and fission/ fusion defects, which can result in mitochondrial mislocalization [28]. Apoptosis was not detected in the Ab42 fly brain by EM analysis [13] or TUNEL staining ( Figure 2D), suggesting that the Ab42-induced reduction in mitochondria in neurites is not due to cellular responses associated with apoptosis.

Ab42-Induced Locomotor Deficits Are Enhanced by Genetic Reduction of Mitochondrial Transport
To test whether mitochondrial mislocalization contributes to Ab42 toxicity, we examined the effect of a genetic reduction in mitochondrial transport on Ab42-induced locomotor defects. Ab42 flies show age-dependent, progressive locomotor dysfunction starting from 14 dae, which can be detected by climbing assay [13,14]. In this assay, flies were placed in an empty plastic vial and tapped to the bottom. The number of flies at the top, middle, or bottom of the vial was scored after 10 seconds. Mitochondria are linked to motors by the mitochondrial membrane GTPase Miro, which is linked to kinesin by milton to allow transport in axons and dendrites [29]. Null mutations in milton and Miro have been reported to disrupt axonal and dendritic transport of mitochondria in neurons [30,31]. Expression of milton RNAi in neurons with the pan-neuronal elav-GAL4 driver reduced the mRNA levels of milton in fly heads ( Figure 3A), and resulted in 60% reduction in milton protein levels in dissected fly brains ( Figure 3B). We analyzed mitochondrial localization in the mushroom body structures to confirm that milton RNAi expression caused a significant reduction in the mito-GFP signal in axons and an accumulation in somata ( Figure 3C). Using this transgenic RNAi flies, we found that neuronal knockdown of milton enhanced Ab42-induced locomotor defects, while milton knockdown itself did not cause locomotor defects at this age ( Figure 3D, left). Similar results were obtained with the independent transgenic UAS-milton-RNAi fly line ( Figure 3D, right).

Ab42-Induced Locomotor Deficits Are Modified by cAMP Levels
cAMP is generated from ATP, and depletion of mitochondria in axons has been shown to disrupt cAMP/PKA signaling, which limits mobilization of the synaptic vesicle reserve pool in presynaptic terminals, and reduces synaptic strength [32]. We tested whether a reduction in the cAMP level by a genetic reduction of the rutabagaencoded type I Ca 2+ /CaM-dependent adenylyl cyclase (rut) enhanced neuronal dysfunction in Ab42 flies. Since the rutabaga mutation (rut [1]) is X-linked, we used the cholinergic neuron-specific Cha-GAL4 driver on the second chromosome, instead of the pan-neuronal elav-GAL4 driver on X choromosome, to drive Ab42 expression in male flies in the rut [1] background. Expression of Ab42 in cholinergic neurons using the Cha-gal4 driver caused locomotor defects by 17 dae (Figure 4A, left). In contrast, in the rutabaga mutant background (rut [1]), Ab42 caused locomotor dysfunctions by 7 dae (Figure 4A, right). Thus, reduced cAMP levels result in an earlier onset of Ab42-induced locomotor defects.
Next, we tested whether an increase in the cAMP level by a genetic reduction of the dunce-encoded phosphodiesterase (PDE), an enzyme that degrades cAMP, ameliorated neuronal dysfunction in Ab42 flies. Since the dnc mutation (dnc [1]) is also X-linked, we used the cholinergic neuron-specific Cha-GAL4 driver on the second chromosome to drive Ab42 expression in male flies in dnc [1] mutant background. We found that Ab42-induced locomotor defects were suppressed in flies with a hypomorphic mutation of dnc (dnc [1]). In contrast, dnc [1] flies show similar locomotor function as the control flies (See the ''material and methods'' section for genetic background for dnc [1] and control flies) ( Figure 4B).

Ab42-Induced Locomotor Defects Are Modified by Neuronal PKA Activity
Since PKA activity is regulated by cAMP levels, we examined whether PKA activity is involved in Ab42-induced toxicity. Knockdown of the catalytic subunit of PKA (PKA-C1) in neurons using UAS-PKA-C1-RNAi driven by the pan-neuronal elav-GAL4 driver enhanced Ab42-induced locomotor defects, while neuronal knockdown of PKA-C1 by itself did not cause locomotor defects at this stage ( Figure 4C). PKA activity is suppressed by binding of the regulatory subunits (PKA-R) to the catalytic subunit, and overexpression of PKA-R decreases, while knockdown of PKA-R increases, PKA activity. The transgenic fly lines EP2162 and EY11550 overexpress PKA-R2 in neurons when combined with the pan-neuronal elav-GAL4 driver. We found that neuronal overexpression of PKA-R2 significantly enhanced Ab42-induced locomotor defects, while overexpression of PKA-R2 by itself did not affect locomotor function ( Figure 4D).
We further examined the effects of a reduction in neuronal PKA-R2 expression on Ab42-induced locomotor dysfunctions. Knockdown of PKA-R2 in neurons using an RNAi transgene with the pan-neuronal elav-GAL4 driver suppressed the locomotor defects in Ab42 flies, while PKA-R2 knockdown by itself did not affect locomotor function ( Figure 4E). Similar results were observed using an independent Ab42 transgenic fly line ( Figure  S3).
Because rut, dnc, and the PKA complex is enriched in the axons and dendrites in fly neurons [33], these results suggest that neuronal dysfunctions in Ab42 flies may be attributable to reduced cAMP/PKA signaling in the axons and dendrites.
Neuronal knock-down of PKA-C1 or PKA-R2 did not affect the accumulation of Ab42 ( Figure S4), the number of Ab42 aggregation detected as Thioflavin S-positive deposits ( Figure  S5), or neurodegeneration ( Figure S6) in the Ab42 fly brain.

The Levels of Putative PKA Substrate Phosphoproteins Are Reduced in the Ab42 Fly Brain
To test whether PKA activity is reduced in the Ab42 fly brain, we compared PKA-C1 and PKA-R2 protein levels and total PKA activity in extracts from dissected brains from Ab42 and control flies. These parameters were not significantly different ( Figure 5A and 5B). We next examined whether the cellular distribution of PKA is altered in the Ab42 fly brain by immunostaining. A strong PKA-C1 signal was detected in the axons and dendrites, with less staining in the cell bodies of mushroom body structure. We did not detect obvious differences between Ab42 and control fly brains ( Figure 5C). We also compared that cAMP levels in head extracts of Ab42 and control flies, and found they were not significantly different ( Figure 5D and Figure S7).
To further investigate whether the levels of putative PKA substrate phosphoproteins are reduced in the Ab42 fly brain, we performed Western blot analysis using a phospho-PKA substrate  antibody (anti-RRxpS/T). We first identified the signals whose reductions were correlated with the decreased PKA activity in the dissected fly brains. Neuronal knockdown of PKA-C1 markedly reduced the signal intensities of phosphoproteins migrating at 24 kDa and 38 kDa ( Figure 5E, arrows). Because the identity of the 24 kDa and 38 kDa proteins is currently under investigation, it is not clear whether phosphorylation of these proteins is decreased without changes in the steady-state levels, and these proteins may be phosphorylated by kinases other than PKA. Nevertheless, since neuronal knockdown of PKA-C1 markedly reduced these signals, the levels of 24 kDa and 38 kDa phosphoproteins are correlated with PKA activity in the fly brains. We also found that some of the signals detected by anti-RRxpS/T, including a protein migrating at 30 kDa, were not affected by PKA-C1 knockdown in fly brains ( Figure 5E, arrowhead).
The signals of the 24 kDa and 38 kDa phosphoproteins were significantly reduced in the brains dissected from three independent Ab42 fly lines ( Figure 5F, arrows). In contrast, the signal of the 30 kDa protein was not affected by Ab42 expression (Figure 5F, arrowhead). Although we did not detect a change in overall PKA activity or cAMP levels, these data suggest that cAMP/PKA signaling is disrupted in the Ab42 fly brain.
In mammals, PKA activates cAMP-response element binding protein (CREB) via direct phosphorylation at Ser133, and Drosophila CREB (dCREB) Ser231 is equivalent to the mammalian Ser133 [34]. dCREB migrates around 38 kDa, and a reduction in CREB phosphorylation has been reported in cellular and animal models of AD [35,36]. We tested whether anti-RRxpS/T recognized phopshorylated dCREB, and whether phosphorylation of dCREB was affected by Ab42 expression by immunoprecipi- tation with anti-RRxpS/T followed by Western blotting with anti-dCREB antibody. Anti-RRxpS/T recognized phosphorylated dCREB ( Figure 6A), while we did not detect a significant reduction in dCREB phosphorylation in the Ab42 fly brain ( Figure 6B-C). This result indicates that a reduction of the level of 38 kDa phosphoprotein in the Ab42 fly brain is not due to a reduction in phosphorylation of dCREB.

Disruption of Mitochondrial Transport Causes Age-Dependent Behavioral Deficits and Reduces the Levels of Putative PKA Substrate Phosphoproteins
We have shown that mitochondria are reduced in the axons and dendrites in the Ab42 fly brain (Figure 1) and that a genetic reduction in mitochondrial transport enhances Ab42-induced behavioral deficits (Figure 3). We examined whether a disruption in mitochondrial transport is sufficient to cause late-onset behavioral deficits. Neuronal knockdown of milton by UASmilton-RNAi driven by the pan-neuronal elav-GAL4 driver did not affect locomotor function up to 10 dae ( Figure 7A, left), but caused locomotor dysfunctions after 17 dae ( Figure 7A, left). Similar results were obtained with the independent UAS-milton-RNAi transgenic fly line ( Figure 7A, right). In addition, we found that the levels of the 24 kDa and 38 kDa phosphoproteins were reduced in the brains dissected from flies with neuronal knockdown of milton ( Figure 7B, arrows), suggesting that mitochondrial mislocalization causes disruption of cAMP/PKA signaling.

Discussion
Elucidation of mechanisms underlying Ab42-induced toxicities is crucial to understanding the complex pathogenesis of AD. An altered distribution of mitochondria has been reported in the brains of AD patients and in cellular and animal models of Ab toxicity [5,6,7,8]. Using a transgenic Drosophila model, we have demonstrated that mislocalization of mitochondria is induced by Ab42 without severe mitochondrial damage or neurodegenera-tion, and that mitochondrial mislocalization underlies neuronal dysfunction induced by Ab42. Our findings suggest that mitochondrial mislocalization may contribute to the pathogenesis of AD.

Mechanisms Underlying Ab42-Induced Mitochondrial Mislocalization
In Ab42 fly brains, Ab42 is accumulated intraneuronally and extracellulary [14]. Although it is not yet clear whether intracellular and/or extracellular Ab42 causes mitochondrial mislocalization, several possible mechanisms could underlie mitochondrial mislocalization in Ab42 fly brain neurons.
Some reports have shown that Ab is present within mitochondria and induces mitochondrial damage [7]. Damaged mitochondria are normally transported from neurites to the cell body for repair or degradation in autophagosomes, and persistent mitochondrial damage induced by Ab42 may cause a reduction in mitochondria in neurites as a result [37]. Indeed, in neurons in the AD brain, damaged mitochondria accumulate in autophagosomes in the neuronal cytoplasm [10,11]. In addition, mitochondrial fragmentation occurs during apoptosis [28], which could be induced by Ab.
In the Ab42 fly brain, mitochondrial mislocalization occurred without severe mitochondrial damage ( Figure 2). An immunoEM analysis did not detect Ab42 accumulation in mitochondria in neurons in the Ab42 fly brain [14]. Moreover, our EM analysis [13] and TUNEL staining ( Figure 2) did not detect apoptosis in the Ab42 fly brain. These results suggest that severe mitochondrial damage or apoptosis is not likely to be the primary mechanism underlying mitochondrial mislocalization in the Ab42 fly brain.
In neurons, mitochondria undergo fission perinuclearly in the cell body and are transported along microtubule or actin bundles [24], and global disruption of microtubule-dependent transport may cause reduced mitochondria in the axons and dendrites. Axonal swellings that potentially block transport have been observed in AD mouse models and human AD brains [38], and Figure 6. Phosphorylation of dCREB is not reduced in head extracts from Ab42 flies at 25 dae. (A) Anti-RRxpS/T detects phosphorylation of Drosophila CREB (dCREB) at Ser231. dCREB Ser231, the site equivalent to Ser133 of mammalian CREB, is the only RRxpS/T site in dCREB. Head extracts were subjected to immunoprecipitation using anti-RRxpS/T, followed by Western blotting with anti-dCREB. The specificity of the antibody was confirmed using loss-of-function dCREB mutant flies (CREB S162 ). A low level of expression of a dCREB transgene was used to rescue lethality of CREB S162 (CREB S162 +hs-dCREB) [59]. The signal detected by anti-dCREB was reduced in these flies. (B) Head extracts from control flies or from Ab42 flies at 25 dae were subjected to immunoprecipitation using anti-RRxpS/T, followed by Western blotting with anti-dCREB. The phosphorylated CREB levels were normalized by the CREB level detected by Western blotting of the crude extract and are shown as ratios relative to controls. No difference in phosphorylated dCREB signal was detected (mean6SD, n = 4; p.0.05). (C) The total CREB level is not altered in Ab42 fly brains. The CREB levels were normalized by the tubulin levels detected by Western blotting and are shown as ratios relative to controls. (mean6SD, n = 4; p.0.05). doi:10.1371/journal.pone.0008310.g006 disruption of axonal transport increases Ab generation [38,39,40]. Ultrastructural studies reveal a loss of the normal microtubular architecture near intracellular Ab oligomers, which would impair movement of vesicles and mitochondria [41]. In contrast, it has also been reported that Ab rapidly impair mitochondrial transport without affecting mitochondrial function or the cytoskeleton in hippocampal neurons [42]. In the Ab42 fly brain, mitochondrial mislocalizaton was observed without significant alterations in microtubule assembly (Figure 1). In addition, we did not detect significant changes in the distribution of synaptotagmin-GFP, a marker for synaptic vesicles (Figure 1). These results suggest that mitochondrial mislocalization in the Ab42 fly brain is not due to an overall disruption of microtubule-based transport but may be due to an altered transport specific to mitochondria.
Mitochondrial transport is regulated by several intracellular signals. Elevation of intracellular Ca 2+ , which occurs in regions of high metabolic demand such as nerve terminals and postsynaptic specializations, arrests microtubule-based mitochondrial movement. Mitochondria are linked to motors by the mitochondrial membrane GTPase Miro [29], and a recent study shows that Miro mediates the Ca2 2+ -dependent arrest of mitochondria [43,44,45]. Since altered Ca 2+ homeostasis is observed in AD neurons [46,47], Ab42 may impair mitochondrial movement by disruption of signaling that regulates mitochondrial transport to the axons and dendrites.
The motility of mitochondria is thought to be interrelated with the fission-fusion machinery and an imbalance between mitochondrial fission and fusion induced by Ab42 may result in reduced mitochondria in the axons and dendrites [32,48,49]. In AD, mitochondrial size is increased and mitochondrial numbers are decreased in neurons [10], suggesting that the normally strict regulation of mitochondrial morphology is impaired. The dynamin-like GTPase, Drp1, influences mitochondrial density in axons and dendrites [32,48], and reductions in Drp1 levels or increases in Drp1 activity by S-nitrosylation are induced by Ab and detected in AD brains [49,50,51]. Very recent studies showed that mitochondria are reduced in neuronal processes in AD neurons [49], presumably due to alterations in the mitochondrial fission/fusion [49,50,51]. Although we did not detect severe changes in the morphology of mitochondria in the young Ab42 fly brain (Figure 2), further morphometric analysis will be required to determine whether defects in fission and fusion occur and contribute to the mislocalization of mitochondria in the Ab42 fly brain.

Dysregulation of the cAMP/PKA Pathway in AD
Impaired regulation of the cAMP/PKA pathway has been reported in the brains of AD patients. Decreases in levels of specific adenylyl cyclase (AC) isoforms and disruption of AC/ cAMP signal transduction have been detected in AD brains [52]. Decreased levels of the catalytic and regulatory subunits of PKA, as well as PKA activity, have also been observed in the brains of AD patients [53], although other reports did not detect widespread changes in PKA levels and activity in AD brains [52]. In cellular and animal models, Ab causes an accumulation of PKA-R by reducing proteosomal degradation, which leads to a reduction in PKA activity [36]. In the Ab42 fly brain, levels of putative PKA substrate phosphoproteins were significantly reduced, however, no overall changes in PKA protein levels, PKA activity, cAMP levels, or intracellular localization of the PKA complex were detected ( Figure 5). cAMP is generated from ATP, and it has been shown that depletion of mitochondria in axons disrupt cAMP/PKA signaling [32].
Our results suggest that Ab42-induced mitochondrial mislocalization causes local, but not global, alterations in cAMP/PKA activity, such as in the axons and dendrites.
In addition to cAMP/PKA signaling, the loss of ATP caused by the reduction of mitochondria in neurites could disrupt many biological processes and lead to neuronal dysfunction. Other major functions of mitochondria in neurons includes the regulation of Ca 2+ , which is important for synaptic plasticity, and cell survival [37]. Indeed, disrupting mitochondrial transport diminishes neuronal resistance to NMDA (N-methyl-D-aspartic acid)-induced excitotoxicity [45]. Thus, mitochondrial mislocalization may contribute to multiple aspects of Ab42 toxicity in the brain.

Concluding Remarks
Our study demonstrates that mislocalization of mitochondria underlies Ab42-induced toxicity in vivo. Several reports show that the loss of mitochondria from axons and dendrites is associated with defective synaptic transmission [30,31,32,48]. AD begins as a disorder in synaptic function [54], which is believed to be associated with increased levels of Ab42 in the brain [55]. Studies in animal models show that these functional deficits predate the onset of irreversible neurodegenerative damages [3], and restoration of the activities of certain signaling pathways could suppress Ab42-induced neuronal dysfunctions. For example, rolipram, the most widely used PDE4 inhibitor, ameliorates memory impairments in APP-PSEN1 double transgenic mice [56]. Thus, interventions that rescue mitochondrial function, mitochondrial localization, and associated defects may maintain synaptic plasticity and neurological function [57]. Further studies of the physiological and pathophysiological mechanisms that affect mitochondrial localization may lead to novel approaches for the prevention and treatment of AD.

Fly Stocks and Antibodies
Transgenic fly lines carrying the human Ab42 was established in the background of the Canton-S w 1118 (isoCJ1) genotype as described in [14]. The elav-GAL4 c155 line was outcrossed with the isoCJ1 flies for 5 generations. The X-linked dnc [1] allele, which was crossed into a background containing the iso1CJ autosomes, was a kind gift from Dr. T. Tully (Cold Spring Harbor Laboratory). A control cross to iso1CJ also was used.

GFP Analysis in Fly Brains
Fly brains were dissected in cold PBS, fixed in PBS containing 4% paraformaldehyde (Electron Microscopy Sciences), and then placed under vacuum in PBS containing 4% paraformaldehyde and 0.25% Triton X-100. The fluorescence intensity in the mushroom body regions was analyzed using a confocal microscope (Carl Zeiss LSM 510) and quantified using NIH image.

Genomic DNA Extraction and Quantitative Real Time PCR Analysis
Fly brains were dissected in cold PBS and frozen on dry ice, and genomic DNA was extracted. 20 brains were homogenated in 100 mM Tris-HCl pH 7.5, 100 mM EDTA, 100 mM NaCl, and 0.5% SDS, and incubate at 65uC for 30 min. Samples were treated with 1.5 M potassium acetate and 4 M LiCl, and incubated for 65uC for 30 min, and centrifuged. Supernatant was treated was phenol/chloroform, added isoprophanol, and centrifuged. Precipitated gemonic DNA was rinsed with 70% ethanol and subjected to quantitative real time-PCR (Applied Biosystems). The average threshold cycle value (Ct) was calculated from five replicates per sample. Levels of Co I, Co III and CytB DNA were standardized relative to that of rp49. Relative expression values were determined by the deltaCt method according to quantitative PCR Analysis User Bulletin (Applied Biosystems). Primers were designed using NIH primer blast as follows: Co I, CTGGAATTGCTCATGGTGGA (forward) and CTCCCGCTGGGTCAAAAA (reverse); Co III, CCCGCTATT-GAATTAGGAGCA (forward) and ATTCCGTGGAATC-CTGTTGC (reverse); CytB, TGAGGTGGATTTGCTGTTGA (forward) and TGGTTGAATATGGGCAGGTG (reverse); rp49, GCTAAGCTGTCGCACAAATG (forward) and GTTCGA-TCCGTAACCGATGT (reverse).

ATP, PKA, and cAMP Assays
ATP contents in dissected brains without eye pigments were analyzed using ATP Bioluminescence Assay Kit CLSII (Roche, Mannheim). PKA activity in dissected brains without eye pigments was measured with MESACUP Protein Kinase Assay Kit (MBL, Woburn, MA) in the presence or absence of 2 mM cAMP. cAMP levels was measured with cAMP-screen system (Applied Biosystems, Foster City, CA). ATP, PKA and cAMP levels were calculated by standard curves and normalized by protein levels.

Transmission Electron Microscopy
Probosces were removed from decapitated heads, which were then immersion-fixed overnight in 4% glutaraldehyde and 2% paraformaldehyde in 0.1 M PBS. Samples were post-fixed 1 hr in ferrocyanide-reduced osmium tetroxide (1% osmium tetroxide and 1.5% potassium ferrocyanide in distilled water). Fixation was followed by dehydration in a graded ethanol series and infiltration with Epon-Araldite resin (2 hr in 50% resin in acetone and 24 hr in 100% resin) using constant rotation. After transferring the samples to flat-bottom BEEM capsules with fresh resin, the samples were polymerized overnight at 60uC. Cured blocks containing fly heads were examined with a dissection microscope and heads with a suitable orientation (posterior oriented flat to the block surface) were selected for thin sectioning. Semi thin sections stained with toluidine blue were examined by light microscopy to localize the mushroom body region. Thin sections (120 nm) of entire heads were collected on nickel grids (100 mesh, Veco-EMS). Thin sections were stained for 5 minutes in lead citrate stain. Sections were examined and micrographs collected using a Hitachi H700T TEM.

TUNEL Staining
Fly brains were fixed in PBS containing 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA), treated with 25 mg/ ml proteinase K for 30 min, and incubated with In Situ Cell Death Detection Kit, Fluorescein (Roche, Mannheim) for 1 hr at 37uC. The brains were analyzed using a confocal microscope (Carl Zeiss LSM 510).

RNA Extraction and Quantitative Real Time PCR Analysis
For each sample, 30-40 flies were collected and frozen. Heads were mechanically isolated, and total RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's protocol with an additional centrifugation step (11,0006g for 10 min) to remove cuticle membranes prior to the addition of chloroform. Total RNA was reverse-transcribed using Superscript II reverse transcriptase (Invitrogen), and the resulting cDNA was used as a template for PCR on a 7500 fast real time PCR system (Applied Biosystems). The average threshold cycle value (Ct) was calculated from five replicates per sample. Expression of milton was standardized relative to actin. Relative expression values were determined by the deltaCt method according to quantitative PCR Analysis User Bulletin (Applied Biosystems). Primers were designed using NIH primer blast as follows: milton, CAGGATCAGCTGAAG-CAACA (forward) and ACACGCTACCTCCCATTGTC (reverse); and actin5C, TGCACCGCAAGTGCTTCTAA G (forward) and TGCTGCACTCCAAACTTCCA (reverse).

Western Blotting
Dissected brains were homogenized in Tris-glycine sample buffer (Invitrogen) and centrifuged at 13,000 rpm for 10 min, and the supernatants were separated on 6% or 10% Tris-glycine gels (Invitrogen) and transferred to nitrocellulose membranes (Invitrogen). The membranes were blocked with 5% nonfat dry milk (Nestlé) and blotted with the primary antibody (anti-Drosophila milton (a gift from Dr. T. L. Schwarz), anti-PKA-C1 (a gift from Dr. D. Kalderon), anti-PKA-R2 (a gift from Dr. D. Kalderon), anti-RRxpS/T (Cell Signaling), or anti-tubulin (Sigma)), incubated with appropriate secondary antibody and developed using ECL plus Western Blotting Detection Reagents (GE Healthcare).

Climbing Assay
Climbing assay was performed as previously described [14]. Approximately 25 flies were placed in an empty plastic vial. The vial was gently tapped to knock the flies to the bottom, and the number of flies at the top, middle, or bottom of the vial was scored after 10 seconds. Experiments were repeated more than three times, and a representative result was shown.

Whole-Mount Immunostaining
Fly brains were dissected in cold PBS, fixed in PBS containing 4% paraformaldehyde (Electron Microscopy Sciences), and then placed under vacuum in PBS containing 4% paraformaldehyde and 0.25% Triton X-100. After permeabilization with PBS containing 2% Triton X-100, the brains were stained with rabbit polyclonal anti-PKA-C1 antibody (a gift from Dr. D. Kalderon) followed by detection with biotin-XX goat anti-mouse IgG and streptavidin-Texas Red conjugate (Molecular Probes). The brains were analyzed using a confocal microscope (Carl Zeiss LSM 510).

Thioflavin S Staining
For thioflavin S (TS) staining, the dissected brains were permeabilized and incubated in 50% EtOH containing 0.1% TS (Sigma) overnight. After washing in 50% EtOH and PBS, the brains were analyzed using a confocal microscope. The numbers of TS-positive deposits were quantified from four hemispheres from three flies per genotype. The fluorescence intensity in Kenyon cell regions was analyzed using a confocal microscope (Carl Zeiss LSM 510) and quantified using NIH image.

Quantification of Neurodegeneration
For the analysis of neurodegeneration in Kenyon cell region, heads were fixed in 4% paraformaldehyde, processed to embed in paraffin blocks, and sectioned at a thickness of 6 mm. Sections were placed on slides, stained with hematoxylin and eosin (Vector Laboratories), and examined by bright field microscopy. To quantify neurodegeneration, images of the sections were captured, and the areas of the vacuoles were measured using NIH Image.  Figure S1 Mitochondria are mislocalized in cholinergic neurons in the Ab42 fly brain. Mito-GFP in axon bundle tips, dendrites, and cell bodies of cholinergic neurons in the mushroom body in control and Ab42 fly brains. The Cha-GAL4 driver was used to express transgene in cholinergic neurons. Signal intensities in control and Ab42 flies at 35 dae were quantified and are shown as ratios relative to control (mean 6 SD, n = 6-10; *, p,0.05, Student's t-test). Representative images are shown at the top. Male flies were used. Found at: doi:10.1371/journal.pone.0008310.s001 (0.07 MB DOC) Figure S2 a-synuclein did not cause significant alteration of mitochondria localization in the fly brain. Mito-GFP in axon bundle tips, dendrites, and cell bodies in the mushroom body in control and a-synuclein fly brains. Transgene expression was driven by the pan-neuronal elav-GAL4 driver. Signal intensities in control and a-synuclein flies at 20 dae were quantified and are shown as ratios relative to control (mean 6 SD, n = 6-10; *, p,0.05, Student's t-test). Representative images are shown at the top. Male flies were used.