Experiments were conducted using D. melanogaster males and/or virgin females from w¹¹¹⁸, w¹ or mini-w⁺ on a w¹¹¹⁸ background (w¹¹¹⁸;{P[w[+mc] = UAS-GFP. S65T]}^IIT₁₀) stocks, and two wild-type stocks commonly used by the scientific community, i.e., Oregon R (http://flybase.org/reports/FBsn0000276.html) and Vallecas (Morata and Garcia-Bellido, 1973), named hereafter as w⁺. Flies were raised in standard conditions (25°C, 12:12 h light:dark cycle, standard food). Flies were anesthetized either with nitric oxide (Inject+Matic Sleeper) or CO₂ for sex identification under a stereoscopic microscope.

Histological sections of the retina were prepared from virgin female flies of w¹¹¹⁸, w¹, mini-w⁺ in w¹¹¹⁸ background, or w⁺ stocks, aged 5, 15, or 30 days. Five flies of each genotype and age were anesthetized and decapitated with a sharp needle. Heads were placed on a microscope slide within a droplet of physiological saline solution. The proboscis was cut off and the occipital cuticle was removed, using fine forceps and a sharp needle, to improve fixative penetration. Heads were fixed overnight in an ice-cold solution of 2.5% glutaraldehyde and 4% paraformaldehyde prepared in 0.1 M phosphate buffered saline pH 7.3. After rinsing in saline solution heads were post-fixed for 1 h in 0.5% osmium tetroxide, rinsed in water, dehydrated in 10 min steps (50, 70, 80, 90, and 100% ethanol and twice in acetone for 20 min), embedded in resin (AGAR 100, AGAR Scientific), and polymerized at 60°C for 48 h. Histological sections of 1 μm thickness were cut with a glass knife on a RMX MT-X ultramicrotome, stained with 0.1% boracic toluidine blue and mounted on microscope slides with DPX (AGAR Scientific) for observation with an Olympus IX81 microscope. Sections were carefully taken at about the same depth/region of the eye to allow proper comparison. Images were acquired with a digital microscope camera Olympus DP71 and processed with Adobe Photoshop.

Lacunae were quantified in three virgin female flies from w¹¹¹⁸ and w⁺ stocks, aged 5, 15 and 30 days and in three virgin female flies from w¹ stock aged 30 days. We registered lacunae number per genotype and age, measured lacunae area and calculated average lacunae area (μm²) for each genotype and age. w⁺ flies never showed lacunae. Statistical analyses were conducted using STATISTICA (7.0 Version, StatSoft, Inc.). The Shapiro–Wilk test (Shapiro et al., 1968) was used to check for normal distribution and Levene test (Brown and Forsythe, 1974) was used to check for homogeneity of variances. When both conditions were confirmed, One-Way ANOVA parametric test was used (lacunae number, w¹ 30 d vs. w¹¹¹⁸ 30 d). If not, non-parametric Kruskal–Wallis (Kruskal and Wallis, 1952) was used instead (lacunae number and area, w¹¹¹⁸ 5 d vs. w¹¹¹⁸ 15 d vs. w¹¹¹⁸ 30 d) and Mann–Whitney U-test (Mann and Whitney, 1947) was used for post-hoc analysis, or directly Mann–Whitney U-test (lacunae area, w¹¹¹⁸ 30 d vs. w¹ 30 d).

The size (diameter in cross section) of each rhabdomere in photoreceptors R1 to R7 was measured in three virgin female flies from w¹¹¹⁸ and w⁺ stocks, aged 5 and 30 days. Rhabdomeres were measured in retinal histological sections (one eye per fly, six equatorial located ommatidia per eye). Statistical analyses were conducted using STATISTICA (7.0 Version, StatSoft, Inc.). The Shapiro–Wilk test (Shapiro et al., 1968) was used to check for normal distribution and Levene test (Brown and Forsythe, 1974) to check for homogeneity of variances. After both conditions were confirmed, Two-Way ANOVA parametric test was used to check for significant differences in rhabdomeres R1-R7 diameter between different genotypes of the same age (w⁺ 5 d vs. w¹¹¹⁸ 5 d and w⁺ 30 d vs. w¹¹¹⁸ 30 d) and between different ages of the same genotype (w⁺ 5 d vs. w⁺ 30 d and w¹¹¹⁸ 5 d vs. w¹¹¹⁸ 30 d). The Fisher exact test or Bonferroni test were used for post-hoc analysis. The percentage of ommatidia with seven rhabdomeres (i.e., the total number that can be observed at this level of the retina in normal flies) was calculated from histological sections of three virgin female flies from w¹¹¹⁸ and w⁺ stocks, aged 5 and 30 days, from three virgin female flies from w¹ stock aged 30 days and from three virgin female flies from mini-w⁺ stock aged 30 days (about 200 ommatidia per genotype and age). The Shapiro–Wilk test (Shapiro et al., 1968) was used to check for normal distribution and Levene test (Brown and Forsythe, 1974) to check for homogeneity of variances. When both conditions were confirmed, Two-Way ANOVA parametric test was used to check for significant differences in the percentage of ommatidia with seven rhabdomeres between different genotypes of the same age (w⁺ 5 d vs. w¹¹¹⁸ 5 d and w⁺ 30 d vs. w¹¹¹⁸ 30 d) and between different ages of the same genotype (w⁺ 5 d vs. w⁺ 30 d and w¹¹¹⁸ 5 d vs. w¹¹¹⁸ 30 d). The Fisher exact test or Bonferroni test were used for post-hoc analysis. If not, non-parametric Kruskal–Wallis (Kruskal and Wallis, 1952) was used instead (w⁺ 30 d vs. w¹¹¹⁸ 30 d vs. w¹ 30 d). Mann–Whitney U-test (Mann and Whitney, 1947) was used for post-hoc analysis.

ERG assays were conducted in retinas of live virgin female flies from w¹¹¹⁸ and w⁺ stocks, aged 5 and 30 days (n = 4 per genotype and age). Flies were first immobilized by ice cooling and placed with their heads emerging from the tip of a disposable plastic micropipette, in order to manipulate their orientation under a microscope (Axioscope, Zeiss). After an adaptation period of at least 10 min, ERG recordings were obtained with glass electrodes filled with saline solution. The active electrode was placed on the cornea of the right eye. The reference electrode was a wire inserted in saline-soaked cotton and touching the fly body. The stimulus was a light pulse emitted by a 15 mA white LED placed at 5 cm from the cornea. A dim light background, generated by a computer monitor placed about 1.5 m apart was present during the experiments. The stimulus regime consisted of a train of 50 rectangular pulses of 130 ms each, separated by 5 s. Electrode voltage was amplified using an Axoclamp 2B (Axon Instruments) and continuously sampled at 20 kHz using Pclamp software (Axon Instruments). Post-stimulus recordings from each fly were averaged off-line per genotype and age, and their traces were overlapped for visual comparison. Statistical analyses were done using Willcoxon rank-sum test (Wilcoxon, 1945).

For climbing assays, 30–50 virgin male and female flies from w¹¹¹⁸ and w⁺ stocks (separated per genotype and sex in tubes of 10 flies each), were selected within 1 day after hatching. Flies were transferred every 3–5 days to tubes containing fresh food. Climbing ability was tested in these flies at four times along their life (5, 15, 25, and 30 days of age). Each tube was quickly tapped 8 consecutive times to make the flies fall to the bottom: this forces all flies to start climbing (negative geotaxis reflex, flies move opposite the Earth's gravitational vector when disturbed). Ten seconds later, we recorded the number of flies that have crossed a line drawn at 8 cm from the bottom of the tube. This procedure was repeated 10 times for each tube, leaving a 1 min interval between each measurement. The 10 measurements per tube were averaged for graphical representation and statistic comparisons. All assays were made under red light to avoid phototaxis effects. Data obtained were compiled into Excel tables and plotted per genotype, sex and age. Statistical analyses were conducted using STATISTICA (7.0 Version, StatSoft, Inc.). The Shapiro–Wilk test (Shapiro et al., 1968) was used to check for normal distribution and Levene test (Brown and Forsythe, 1974) to check for homogeneity of variances. When both conditions were confirmed, Two-Way ANOVA parametric test was used to check for significant differences in climbing ability between different genotypes of the same age (females analysis: w⁺ 5 d vs. w¹¹¹⁸ 5 d, w⁺ 15 d vs. w¹¹¹⁸ 15 d, w⁺ 25 d vs. w¹¹¹⁸ 25 d and w⁺ 30 d vs. w¹¹¹⁸ 30 d) and between different ages of the same genotype (females analysis: w⁺ 5 d vs. w⁺ 15 d vs. w⁺ 25 d vs. w⁺ 30 d and w¹¹¹⁸ 5 d vs. w¹¹¹⁸ 15 d vs. w¹¹¹⁸ 25 d vs. w¹¹¹⁸ 30 d). The Fisher exact test or Bonferroni test were used for post-hoc analysis. If not, non-parametric Kruskal–Wallis (Kruskal and Wallis, 1952) was used instead (males analysis: w⁺ 5 d vs. w¹¹¹⁸ 5 d, w⁺ 15 d vs. w¹¹¹⁸ 15 d, w⁺ 25 d vs. w¹¹¹⁸ 25 d, w⁺ 30 d vs. w¹¹¹⁸ 30 d, w⁺ 5 d vs. w⁺ 15 d vs. w⁺ 25 d vs. w⁺ 30 d, w¹¹¹⁸ 5 d vs. w¹¹¹⁸ 15 d vs. w¹¹¹⁸ 25 d vs. w¹¹¹⁸ 30 d). Mann–Whitney U-test (Mann and Whitney, 1947) was used for post-hoc analysis.

For stress assays we collected males and virgin female flies from w¹¹¹⁸ and w⁺ stocks during their first day of life (dextrose assays: n = 60 flies per genotype and sex; starvation assay: n = 60 flies per genotype and sex; paraquat assays: n = 50 flies per genotype, sex and paraquat concentration; hydrogen peroxide assays: n = 20 flies per genotype, sex and hydrogen peroxide concentration). Flies of each genotype and sex were kept in tubes containing ~10 flies each. For dextrose treatment, animals were placed into tubes having a filter paper in the bottom soaked in a 5% dextrose-water solution. For starvation experiments, adults were placed into tubes having a piece of paper soaked in water (to avoid thirst and desiccation) but without food. For paraquat and hydrogen peroxide treatments, animals were first starved overnight (Wang et al., 2008). Next morning, animals were placed into tubes having a filter paper soaked in a 5% sucrose-water solution containing either paraquat (2, 10, or 20 mM) or hydrogen peroxide (0.5 or 5%). Half-life was measured for each tube (each corresponding to 10 flies), and used for statistically comparing genotypes of the same sex. Two-Way ANOVA parametric test was used to check for significant differences in half-life between genotypes and treatment conditions. The Fisher exact test or Bonferroni test were used for post-hoc analysis.

For life span experiments we used males and virgin female flies from w¹¹¹⁸ and w⁺ stocks selected immediately after hatching. Twenty to hundred flies per genotype and sex were kept in groups of ~10 flies per tube. Flies were transferred every 3–5 days to tubes containing fresh food. For total life span measurement (25°C assays: n = 100 flies per genotype and sex; dextrose assays: n = 60 flies per genotype and sex; starvation assays: n = 60 flies per genotype and sex; paraquat assays: n = 50 flies per genotype, sex and paraquat concentration; hydrogen peroxide assays: n = 20 flies per genotype, sex and hydrogen peroxide concentration), the number of dead flies per tube was counted every day from day 1 until the last fly died. Half-life was calculated as the age where 50% of the flies of each genotype, sex and experimental condition died. Data obtained were compiled into Excel tables, plotted per genotype and sex, and separated by stress treatment. Statistical analyses were conducted using STATISTICA (7.0 Version, StatSoft, Inc.). The Shapiro–Wilk test (Shapiro et al., 1968) was used to check for normal distribution and Levene test (Brown and Forsythe, 1974) to check for homogeneity of variances. When both conditions were confirmed, Two-Way ANOVA parametric test was used to check for significant differences in half-life between genotypes and treatment conditions. The Fisher exact test or Bonferroni test were used for post-hoc analysis.

For the analysis of retinal organization in mini-w⁺ flies, in addition to histological sections, we used the method of optical neutralization of the cornea as previously described (Franceschini and Kirschfeld, 1971; Franceschini et al., 1981). After nitric oxide anesthesia and decapitation, the heads of 30 days-old mini-w⁺ and w⁺ flies were mounted on a microscope slide with a droplet of transparent nail polish. Illumination through the eye using an Olympus IX81 inverted microscope allowed to visualize the tips of the rhabdomeres with a 40 x objective. Images were acquired with a digital microscope camera (Olympus DP71) and processed with Adobe Photoshop.

Retinal degeneration can be precisely monitored in Drosophila by examination of histology sections. We prepared sections of the retina from w¹¹¹⁸ mutants and w⁺ flies aged 5, 15, and 30 days and from w¹ mutants aged 30 days. The extraordinarily regular array of ommatidia in the w⁺ fly eye (Figures 1A–C) allows detection of even small deviations from the normal pattern in the w¹¹¹⁸ fly eye (Figures 1D–F). A modest but clear phenotype was observed already in the eye of w¹¹¹⁸ mutants of the youngest age (5 d), comprising mild disorganization of the characteristic pattern of ommatidia (Figure 1D) and occasional lacunae probably representing missing ommatidia (Figure 1G). This w¹¹¹⁸ phenotype, never observed in the retina of w⁺ flies, became aggravated with age (15 and 30 d) and comprised greater disorganization and progressively larger lacunae (see black stars in Figures 1E,F and average lacunae area in Figure 1G). In the oldest w¹¹¹⁸ flies (30 d), part of the spaces devoid of ommatidia were filled with osmophilic material resembling a glial scar (see white stars in Figure 1F). Lacunae were also observed in 30 d flies from a second mutant w⁻ allele (w¹) (see black stars in Figures 1J,K).

Retinal degeneration in Drosophila is almost always associated with degeneration of the rhabdomeres, i.e., the microvilli-packed apical portion of the photoreceptor enriched in light-sensing proteins (Shieh, 2011). We found that the size of rhabdomeres decreased with age in w⁺ and w¹¹¹⁸ mutant retinas (5 vs. 30 d). Most importantly, we also found that rhabdomeres from w¹¹¹⁸mutants were smaller than those from age-matched w⁺ control flies at both ages (Figure 1H). We also quantified the number of ommatidium with seven rhabdomeres in w¹¹¹⁸ and w⁺ retinas (5 and 30 d), and in 30 d w¹ retinas, as seven is the number of rhabdomeres per ommatidium expected to be observed in histological sections taken at this level of the retina (Cagan, 2009). We found significant differences in the percentage of ommatidia with seven rhabdomeres between w¹¹¹⁸ and w⁺ flies at both ages analyzed. There were no significant age-dependent differences in the percentage of ommatidia with seven rhabdomeres within each genotype. At 5 days of age, all ommatidia contained the complete set of rhabdomeres in w⁺ flies but only 46% of ommatidia in w¹¹¹⁸ flies. In older flies (30 d) the percentage of ommatidia with seven rhabdomeres was 97% in w⁺ flies and 40% in w¹¹¹⁸ flies (Figure 1I). w¹ 30 d mutants showed a percentage of ommatidia with seven rhabdomeres similar to that of w⁺ flies of the same age and differed from that of w¹¹¹⁸ 30 d flies (Figure 1K).

Disorganization of the regular array of ommatidia, atrophied rhabdomeres, and progressive loss of photosensitive units will most probably have functional consequences for the retina of w⁻ mutants. This can be monitored by ERG, a robust assay applied to a variety of experimental conditions (Belušič, 2011). Previous studies showed that mutations in Drosophila w gene affect the ERG in several ways (Stark and Wasserman, 1972; Pak and Lidington, 1974; Wu and Wong, 1977; Belušič, 2011). However, those assays were done with flies of unreported age and/or sex, or carrying additional mutations, making difficult to discern the contribution of sex, age, and genotype to the reported ERG abnormalities.

Here we compared the ERG of w¹¹¹⁸ and w⁺ female flies of 5 and 30 days of age kept in standard laboratory conditions (Figure 2). The ERG of w⁺ Drosophila flies classically contains three components, the ON potential (on), the receptor potential (Rp), and the OFF potential (off). Rp (shaded gray area Figure 2A top-left) is produced by the activation of photoreceptors while the “on” and the “off” potentials are generated by the synaptic activation of structures present in the lamina (Trujillo-Cenóz, 1965; Heisenberg, 1971). The ERG of w¹¹¹⁸ and w⁺ flies differed in the shape and time course of the Rp [compare red (w⁺) and black (w¹¹¹⁸) traces in Figure 2A top right]. This difference is expressed as an initial corneal negativity that opposes (and consequently reduces) the “on” potential and increases the initial stages of Rp in w¹¹¹⁸ flies (Figure 2A right, black arrow). This phenomenon is associated with the lack of eye pigments in these mutants as compared to w⁺ (reviewed in Belušič, 2011) and is probably due to the massive recruitment of the photoreceptor population in w¹¹¹⁸ mutants. We confirmed that these differences between w¹¹¹⁸ and w⁺ in young (5 d) flies were significant (p < 0.05). The comparison of both genotypes at young age showed also that the “off” potential is significantly smaller and delayed in the mutant (Figure 2A, asterisk, compare red and black traces in right top panel; p < 0.05). There was no difference in Rp responses between young and old flies of w⁺ genotype. In w¹¹¹⁸ 30d flies, the Rp difference with w⁺ was reduced becoming not statistically significant [Figure 2A, compare red and blue (w¹¹¹⁸) traces, p = 0.1]. Similarly, the differences in “on” and “off” potentials between w¹¹¹⁸ and w⁺ were not observed in the 30 d flies (Figure 2A, open circle, compare red and blue traces in right bottom panel). In w⁺ flies, the age progression produced a decrease in the early deflection characteristic of the young mutants with a recovery of the “on” and “off” potentials (Figure 2B, compare black and blue traces, p < 0.05).

The majority of neurodegenerative conditions studied in fruitflies and other animals have been associated with deficiencies in locomotor ability (Lessing and Bonini, 2009; Hirth, 2010; Jaiswal et al., 2012). The most frequently used locomotion assay in Drosophila is the “climbing assay,” which measures a motor activity that requires a brain circuit including identified neurons expressing the biogenic amine dopamine (Riemensperger et al., 2013).

We tested the locomotion ability of w¹¹¹⁸ mutant flies using the climbing assay. Flies of either genotype (w¹¹¹⁸ or w⁺) of both sexes were tested at four different ages from 5 days post-hatching. Thirty days w¹¹¹⁸ females had a significant deterioration of their climbing ability relative to w⁺ females of the same age and relative to the 5 d w¹¹¹⁸ females (Figure 3A). Older w¹¹¹⁸ males (25 and 30 d) have a significant deterioration of their climbing ability relative to w⁺ males of the same age (Figure 3B). This tendency for lower climbing scores in w¹¹¹⁸ flies of both sexes was detected at younger ages, albeit differences with respect to w⁺ were not statistically significant (Figures 3A,B).

Shorter life span, relative to wild-type strains, is another key feature of all neurodegenerative phenotypes in Drosophila (Lessing and Bonini, 2009; Hirth, 2010; Jaiswal et al., 2012). We performed life span experiments on female and male flies of w¹¹¹⁸ and w⁺ strains at standard temperature (25°C), under normal raising conditions -ie. without stressors- (Figures 4A,B) or under different forms of stress (Figures 4C–J).

We found that w⁻ mutants had shorter life span relative to w⁺ at most conditions tested. Under normal raising conditions at 25°C, we observed a significant life shortening in w¹¹¹⁸ females compare to w⁺ females (w¹¹¹⁸ vs. w⁺ half-life was 47 vs. 65 days, p < 0.01; Figure 4A). In males, w¹¹¹⁸ vs. w⁺ half-life was 43 vs. 49 days (p > 0.05; Figure 4B).

We then tested the resistance of Drosophila w⁻ mutants to two different challenges, starvation and sugar-enriched diet (reviewed in Ristow and Schmeisser, 2011). Female or male flies from w⁺ and w¹¹¹⁸ strains were kept under standard laboratory conditions except that their food was either absent (starvation) or enriched with dextrose (Figures 4C–F). Our results confirmed previous reports in which, irrespective of their genotypes and sexes, flies kept under a high-sugar diet have shorter life span than flies raised on standard diet (compare Figure 4C to Figure 4A and Figure 4D to Figure 4B). Most importantly, we observed that w⁻ mutants were significantly less resistant to a sugar-enriched diet than w⁺ flies. In females w¹¹¹⁸ vs. w⁺ half-life was 18 vs. 26 days (p < 0.0001; Figure 4C) while in males was 20 vs. 26 days (p < 0.0001; Figure 4D). w¹¹¹⁸ flies were also significantly less resistant to starvation than w⁺ flies. In females, w¹¹¹⁸ vs. w⁺ half-life was 64 vs. 104 h (p < 0.0001; Figure 4E) while in males was 56 vs. 80 h (p < 0.0001; Figure 4F).

A strong association between neurodegenerative pathologies and oxidative stress is well-documented in humans and other mammals (Yan et al., 2013; Cobb and Cole, 2015; Kim et al., 2015) as well as in flies (Gruenewald et al., 2009). We compared the resistance of Drosophila w¹¹¹⁸ mutants to oxidative stress relative to w⁺ flies, with two widely used oxidative agents at various concentrations (Figures 4G–J). Flies of each genotype were kept under standard laboratory conditions except that their diet contained either paraquat (Figures 4G,H) or hydrogen peroxide (Figures 4I,J). Both treatments are reported to induce oxidative stress and to have consequences at the transcriptional level (Zou et al., 2000; Landis et al., 2004, 2012; Brown et al., 2009).

We found that w¹¹¹⁸ mutants of both sexes were significantly less resistant to paraquat than w⁺ flies. After paraquat treatment, the half-life of females was 56 h in w¹¹¹⁸ relative to 112 h in w⁺ (p < 0.05) at 20 mM; 72 vs. 128 h (p < 0.05) at 10 mM and 192 vs. 320 h (p < 0.001) at 2 mM (Figure 4G). In males, half-life was 40 vs. 88 h (p > 0.05) at 20 mM; 56 vs. 168 h (p < 0.05) at 10 mM and 80 vs. 296 h (p < 0.001) at 2 mM (Figure 4H). w¹¹¹⁸ females were also significantly less resistant to hydrogen peroxide than w⁺ at the 0,5% concentration (w¹¹¹⁸ vs. w⁺: 264 vs. 320 h (p < 0.05; Figure 4I).

Most of the transgenic stocks generated in a w¹¹¹⁸ mutant background carry a form of the gene w (mini-w⁺) that recuperates the eye color to a certain extent.

Here we found that expression of a mini-w⁺ Drosophila transgene in w¹¹¹⁸ flies rescued not only the eye pigmentation phenotype, but also the retinal-degeneration phenotype that we had detected in w⁻ mutant flies (Figure 5). Neither missing rhabdomeres nor lacunae (missing ommatidia) were observed in 30 days-old flies expressing mini-w⁺ across the eye in a w¹¹¹⁸ null genetic background (compare Figures 5A,B with Figures 5C,D; see Figure 5E, percentage of ommatidia with seven rhabdomeres).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.