Fly lines for Ddc-Gal4 (expresses in dopaminergic and serotonergic neurons), 24B-Gal4 (muscle-specific driver), Elav-Gal4 (pan neuronal driver), yw, w¹¹¹⁸, UAS-LRRK2 G2019S, UAS-mito-GFP, parkin null flies (park1) and pink1 null flies (B9) were purchased from the Bloomington Drosophila Stock Center (Bloomington, IN, USA). UAS-AF-6 expressing flies lines were generated via microinjection of pUAST-AF-6 (derived from PRK5-HA-AF-6 via NotI/KpnI digestion) into fly embryos by BestGene Inc (Chino Hills, CA, USA). UAS-canoe RNAi was purchased from VDRC (Vienna Drosophila Stock Center, Austria).

Climbing assays were carried out according to previously described methods (Ng et al., 2009). Briefly, 20 female adult flies from each group were randomly selected after anesthetization and placed in a vertical plastic column (length 30 cm; diameter 1.5 cm). Age-matched normal flies were used as controls. After a 2-h recovery period from CO₂ exposure, flies were gently tapped to the bottom of the column and the number of flies that reached the top of column at 1 min was counted. Results are presented as mean ± SEM of the scores obtained from three independent experiments. In rotenone-treated flies, flies were fed with cornmeal-agar medium containing 50 μM rotenone (purchased from Sigma Aldrich) immediately after eclosion and during the entire experimental period, as described previously (Ng et al., 2009). For the flight assay, a 30 cm-tall bottle was filled with water to a height of 2 cm and a funnel was placed at the bottle opening. Flies from a vial were expelled into the bottle via the funnel. The flies found on the walls after their explusion indicated their flight ability whilst the flies on the water indicated their inability to fly. The number of flies for both outcomes was recorded accordingly. Flies were then retrieved from the bottle via anesthetizing with CO₂.

Immunohistochemical analysis of whole-mount adult fly brains were prepared according to published protocols (Whitworth et al., 2005). Brains were dissected from adult flies in phosphate buffered saline (PBS) and fixed in 0.1 M HEPES with 4% (w/v) paraformaldehyde overnight at 4°C. Subsequently, they were washed with 0.2% (v/v) Triton X-100 in PBS and blocked with 5% (v/v) normal goat serum (NGS) for 1 h at room temperature. This is followed by incubation with primary antibodies (1:300 rabbit anti-tyrosine hydroxylase, Pel-Freeze and 1:250 mouse anti-GFP, Sigma Aldrich) overnight at 4°C. Brain tissues were then washed with 0.2% Triton X-100 for 1 h and incubated with secondary antibodies (1:400 FITC-conjugated goat anti-mouse and 1:300 Rhodamine-conjugated goat anti-rabbit, Jackson ImmunoResearch) overnight at 4°C. The following day, specimens were washed with 0.2% Triton X-100 for 1 h and mounted with Vectashield onto glass cover-slips for confocal microscopy (Olympus). The quantification of the dopaminergic cluster and mitochondrial area (size of mito-GFP puncta expressed as mean ± SD with n ≥ 10 DA neurons per experimental group) using ImageJ software were as described previously (Ng et al., 2012).

For TEM analysis, indirect flight muscles were prepared according to previous protocol (Ng et al., 2012) and specimens were sent for commercial EM analysis at the EM Research Services, Newcastle University (UK). For DA measurement, 10 fly heads were homogenized in 0.5 N perchloric acid and the lysates were filtered through 0.1 μm Durapore Ultrafree Centrifugal filters (Milipore, USA) prior to HPLC measurement.

Unpaired Student’s t-test is used to calculate the statistical p value in comparison between groups (*p < 0.05, **p < 0.001).

To examine the effects of AF-6 overexpression in flies, we generated Drosophila lines that express human AF-6 cDNA as a transgene under the regulation of the GAL4/UAS system (Figure 1A). We selected line #3 that clearly expresses human AF-6 at a robust level for the majority of our investigations. When driven by the Ddc-GAL4 driver (that restricts the expression of the transgene largely to TH-positive dopaminergic neurons and serotonergic neurons), AF-6 expression has negligible effects on the survival of adult flies up to 60 days post-eclosion (Figure 1B). Further, Ddc-GAL4 driven AF-6 expression in flies also exert no apparent effects on their climbing score (Figure 1C) or dopaminergic neuronal number (not shown), particularly in the PPL1 cluster that is known to be selectively susceptible to degeneration in several genetic PD fly models (Figure 1D). Similarly, when driven by the muscle-specific 24B-GAL4 driver, AF-6 expression neither affects the survival nor locomotion ability of these flies (Figures 1E,F). Consistent with this, we did not detect prominent signs of mitochondrial abnormalities in the indirect flight muscles of 24B-GAL4 driven AF-6 expressing flies relative to their control counterparts (Figure 1G). Taken together, our results demonstrate that AF-6 overexpression alone does not produce any overt phenotypes in Drosophila.

Next, we examined the effects of AF-6 expression in parkin null flies. Notably, several groups including ours have reported that Drosophila parkin mutants display increased mortality and prominent mitochondrial pathology in their indirect flight muscles that is accompanied by marked locomotion defects (Greene et al., 2003; Wang et al., 2007). Interestingly, we found that AF-6 expression (via the 24B-GAL4 driver) in parkin null flies dramatically enhances their survival (Figure 2A). Moreover, the double mutant flies also exhibit significantly improved climbing scores that is comparable to the performance by their control counterparts (Figure 2B). Remarkably, the muscle mitochondrial pathology associated with parkin null flies is virtually rescued in the presence of AF-6 overexpression (Figure 2C). Similar results were obtained when we repeated the experiment with another AF-6 overexpressing line (i.e., line #8; Supplementary Figure S1). Thus, AF-6 could apparently compensate for the loss of parkin function, the normal activity of which is needed for optimal mitochondrial QC. As Parkin and PINK1 functions are intimately interwoven, we also examined the effects AF-6 expression in pink1 null flies. Like Drosophila parkin null mutants, pink1-deficient flies are plagued with prominent mitochondria abnormalities that severely impair their flight ability and increase their mortality (Clark et al., 2006; Park et al., 2006). Again, we found that AF-6 overexpression (via the 24B-GAL4 driver) enhances the survival and flight ability of pink1 null flies (Figures 3A,B) and markedly rescues their muscle mitochondrial pathology (Figure 3C).

We have previously demonstrated that LRRK2 G2019S expression in flies also result in mitochondrial abnormalities that leads to dopaminergic neuronal loss and climbing deficits, and that Parkin overexpression can ameliorate these pathological phenotypes in LRRK2 mutant flies (Ng et al., 2012). Given our demonstration above that AF-6 overexepression can rescue the mitochondrial pathology of parkin null flies, we sought to examine whether the same strategy might work in reducing the pathologies of LRRK2 mutant flies. For this purpose, we crossed AF-6 expressing flies with transgenic LRRK2 mutant flies to create a double transgenic Drosophila line that co-expresses AF-6 and LRRK2 G2019S. As previously reported (Ng et al., 2012), Drosophila expressing LRRK2 G2019S (via Ddc-GAL4) exhibit significant impairment in their climbing ability (Figure 4A). However, in the presence of AF-6 co-expression, we recorded a marked improvement in the climbing performance of LRRK2 mutant flies (Figure 4A). As the expression of AF-6 and LRRK2 G2019S is driven by the Ddc-GAL4 driver in this case, we also counted the dopaminergic neurons in the PPL-1 cluster that we and others have previously reported to be affected in LRRK2 mutant flies (Ng et al., 2012). Surprisingly, AF-6 overexpression does not appear to appreciably retard the loss of PPL-1 dopaminergic neurons in LRRK2 mutant flies (Figure 4B). However, HPLC analysis revealed that the DA level that is deficient in LRRK2 mutant flies is remarkably restored in the presence of AF-6 overexpression (Figure 4B), suggesting that AF-6 helps to improve the function of dopaminergic neurons in LRRK2 mutant flies that leads to their improved climbing performance (Figure 4C). Finally, we also examined the status of the neuronal mitochondria in LRRK2 mutant flies in the absence or presence of AF-6 co-expression via the mito-GFP assay. Consistent with our previous report, LRRK2 G2019S expression (via Ddc-GAL4) triggers the appearance of significantly enlarged mitochondria in TH-positive dopaminergic neurons (Figures 4D,E). Notably, this abnormality is significantly mitigated in the double transgenic AF-6/LRRK2 G2019S expressing flies. Taken together, our results demonstrated that AF-6 overexpression is an effective strategy in mitigating the pathologies of several Drosophila genetic models of PD that represent recessive and dominant parkinsonism.

To examine whether AF-6 deficiency might promote LRRK2-induced neurotoxicity, we knocked down the expression of canoe, the orthologous counterpart of AF-6 in flies, via a siRNA strategy. When driven by the pan-neuronal Elav-GAL4 driver, the extent of silencing of canoe expression is remarkable (Figure 5A), which illustrates the efficiency of the siRNA-based strategy. When the knockdown of canoe expression is restricted to TH-positive neurons via the Ddc-GAL4 driver, we did not observe any obvious effects on the mortality, climbing performance or dopaminergic neuronal number of these flies relative to their control counterparts (Figures 5B–D), suggesting that canoe expression is not essential for the physiological functions of dopaminergic neurons under normal conditions. However, in the presence of LRRK2 G2019S expression, the silencing of canoe expression aggravates the locomotion deficits of the mutant flies (Figure 5E). Again, the dopaminergic neuronal count does not appear to be significantly different between LRRK2 G2019S flies and those with reduced canoe expression (Figure 5F), although the DA level in the latter group is evidently lower than LRRK2 mutant flies (Figure 5G), which might explain their poorer performance in climbing. We also carried out experiments to silence the expression of canoe via the 24B-GAL4 driver, which is more relevant to our investigation on its effects on parkin and pink1 null flies as these mutant Drosophila exhibit prominent mitochondrial pathology in their muscles. However, 24B-mediated silencing of canoe expression alone results in significantly accelerated mortality that precludes our intended line of investigation (Supplementary Figure S2). Indeed, nearly 70% of the 24B-sicanoe flies died after a mere week post-eclosion (Supplementary Figure S2).

Given the apparent neuroprotective role of AF-6 overexpression in genetic fly models of PD that are associated with mitochondrial dysfunction, we wondered whether the strategy could also provide protection against exposure to PD-related mitochondrial toxins. To address this, we treated control and AF-6 transgenic flies with rotenone, a well established neurotoxin that induces PD via its inhibition of mitochondrial complex I (Johnson and Bobrovskaya, 2015). We and others have previously reported that rotenone-treated flies exhibit marked dopaminergic neurodegeneration and associated climbing defect (Coulom and Birman, 2004; Ng et al., 2009). Consistent with this, we found that control flies treated with rotenone display these parkinsonian phenotypes prominently (Figures 6A–C). In contrast, Drosophila overexpressing AF-6 (via the Ddc-GAL driver) that are treated with rotenone performed significantly better in their climbing score compared to their similarly treated control counterparts (Figure 6A). Indeed, they are as good a climber as untreated control flies. In good correlation with this, rotenone-mediated dopaminergic neuronal loss is significantly retarded in AF-6 overexpressing flies (Figures 6B,C). Thus, AF-6-mediated neuroprotection similarly works in a fly PD model induced by a toxin that is well known to inhibit mitochondrial function.