Benjamin Eaton, Department of Physiology, UTHSCSA, 7703 Floyd Curl Dr., San Antonio, TX 78229, USA. Tel.: +1 210 567 4383; fax: +1 210 567 4410; e-mail:firstname.lastname@example.org
Synaptic dysfunction is considered the primary substrate for the functional declines observed within the nervous system during age-related neurodegenerative disease. Dietary restriction (DR), which extends lifespan in numerous species, has been shown to have beneficial effects on many neurodegenerative disease models. Existing data sets suggest that the effects of DR during disease include the amelioration of synaptic dysfunction but evidence of the beneficial effects of diet on the synapse is lacking. Dynactin mutant flies have significant increases in mortality rates and exhibit progressive loss of motor function. Using a novel fly motor disease model, we demonstrate that mutant flies raised on a low calorie diet have enhanced motor function and improved survival compared to flies on a high calorie diet. Neurodegeneration in this model is characterized by an early impairment of neurotransmission that precedes the deterioration of neuromuscular junction (NMJ) morphology. In mutant flies, low calorie diet increases neurotransmission, but has little effect on morphology, supporting the hypothesis that enhanced neurotransmission contributes to the effects of diet on motor function. Importantly, the effects of diet on the synapse are not because of the reduction of mutant pathologies, but by the increased release of synaptic vesicles during activity. The generality of this effect is demonstrated by the observation that diet can also increase synaptic vesicle release at wild-type NMJs. These studies reveal a novel presynaptic mechanism of diet that may contribute to the improved vigor observed in mutant flies raised on low calorie diet.
Synaptic dysfunction is believed to be responsible for the age-dependent declines in neural function observed during many adult-onset neurodegenerative diseases (Arendt, 2009; Murray et al., 2008). The age dependence of these diseases has led to experiments investigating the therapeutic effects of manipulations that extend lifespan on neurodegeneration, including dietary restriction (DR) (Duan et al., 2003; Halagappa et al., 2007; Chen et al., 2008; Steinkraus et al., 2008; Kerr et al., 2011). These studies have documented that diet can improve neural function in many disease models but direct evidence for reduced synaptic dysfunction in response to diet is lacking. It should be noted that the beneficial effects of DR on healthy synapses have been documented and include enhanced postsynaptic function in the hippocampus and improved neuromuscular junction (NMJ) morphology (Hori et al., 1992; Eckles-Smith et al., 2000; Valdez et al., 2010). It is currently unknown whether diet can modulate neurotransmitter release under conditions of neurodegenerative disease.
Mutations in the DCTN1 gene encoding the p150 subunit of the dynactin complex result in late-onset lower motor neuron disease in humans (Münch et al., 2004, 2005; Puls et al., 2005). Genetic disruption of dynactin complex function within the nervous system of mice results in animals that develop motor neuron disease with symptoms including progressive muscle weakness (LaMonte et al., 2002; Haferzparast et al., 2003; Laird et al., 2008; Chevalier-Larsen et al., 2008). Further analysis of these mouse models has revealed synapse pathologies including the deterioration of NMJ morphology and the loss of synapses, but neurotransmission has not yet been assayed (Chevalier-Larsen et al., 2008; Laird et al., 2008). Synaptic pathologies are also observed at the Drosophila larval NMJ when dynactin complex function is specifically inhibited within the motor neuron and include decreased synaptic stability and reduced neurotransmission (Eaton et al., 2002). Together, these data support the hypothesis that degeneration of synaptic morphology is an important and conserved pathological event during neurodegeneration in dynactin complex mutants, but direct evidence supporting that impaired neurotransmission is an important early event during neurodegeneration in dynactin mutants is lacking. The early loss of neurotransmission has previously been documented in other motor neuron disease models consistent with the hypothesis that impaired neurotransmission is an early etiology in motor diseases (Rich et al., 2002; Pun et al., 2006; Murray et al., 2008).
In this study, we have developed a novel adult NMJ preparation to investigate the effects of diet on neural function during neurodegeneration in dynactin complex mutants. Using the CM9 motor system, we characterize the synaptic phenotypes resulting from the inhibition of dynactin complex function within adult motor neurons. These phenotypes include decreased longevity, age-dependent declines in motor function, and synaptic pathologies, including impaired NMJ morphology and reduced neurotransmission. The reductions in neurotransmission and motor function appeared early during disease progression prior to significant changes in NMJ morphology are consistent with impaired neurotransmission contributing to the reduction in motor function observed early during pathogenesis. We find that dynactin complex mutants raised on a low calorie diet have improved motor function and extended of lifespan compared to mutants raised on a high calorie diet. Analysis of neurotransmission at mutant NMJs revealed that presynaptic function is significantly increased in the mutant flies raised on the low nutrient diet compared to flies raised on high nutrient diet. We also observed a similar effect of diet on neurotransmission at wild-type NMJs consistent with diet being a general modulator of neurotransmitter release.
Dynactin complex mutants have reduced lifespan
The Gl1 mutation is a well-characterized dominant negative mutation within the glued gene that has been shown to specifically inhibit dynactin complex function by generating a truncated version of the Glued protein (Fig. 1A; Harte & Kankel, 1982; Swaroop et al., 1986). We utilized a fly line harboring a UAS-bearing transgene encoding a truncated Glued mutant protein (UAS-DNGlued) to specifically inhibit dynactin complex function within motor neurons (Fig. 1A; Fan and Ready, 1997; Eaton et al., 2002). Survival data from virgin females generated on standard laboratory food demonstrate that Gl1 mutants and flies expressing the UAS-DNGlued transgene using the motor neuron-specific drivers D42-Gal4 (D42/DNGlued) or OK6-Gal4 (OK6/DNGlued) had significant reductions in survival compared to the driver (D42/+ and OK6/+) and transgene (DNGlued/+) control flies including changes in median (50%) and maximum (90%) lifespan (Table 1). The differences in lifespan between D42/DNGlued and OK6/DNGlued are most likely due to the strength of the OK6-Gal4. Analysis of hazard parameters for the D42/DNGlued and Glued1 survival data shows that these glued mutants have significantly elevated age-dependent (γ or b, respectively) components of their respective hazard functions compared to controls consistent with age-dependent neurodegeneration in mutants (Fig. 1B,C, respectively; Table 1). These data demonstrate that inhibition of dynactin complex function specifically within motor neurons is sufficient to reduce fly survival and consistent with previous studies demonstrating the importance of motor neuron health in determining lifespan in flies (Parkes et al., 1998; Phillips et al., 2000; Chan et al., 2002; Morrow et al., 2004).
Table 1. Survival and mortality analysis on dynactin complex mutants
log rank statistic§
All data represent survival data from virgin female flies of the indicated genotype and diet conditions. All insets and mutations used in these analyses were backcrossed 10 generations to our w1118 background and maintained in this background.
†,‡Values represent median (50%) and maximum (90%) lifespan. Significance determined versus control using quantile regression (Koenker, 2005). Upper (UB) and lower (LB) 95% confidence intervals are indicated. *P < 0.05, **P < 0.01.
§The log-rank statistics represent a comparison between the control and experimental groups.
¶Parameters reported correspond to the indicated mortality model. Significant differences are indicated between control and experimental groups for parameters of the hazard model that best fit the data from the respective experiments (GM = Gompertz–Makeham with parameters λ, γ, and c, L = Logistic with parameters λ, γ, and s. LM = Logistic-Makeham with parameters λ, γ, c, and s) as described in materials and methods.
††Food conditions are described in the materials and methods. Diet conditions used were pre-determined to maximize the effects of diet on lifespan in D42/DNGlued flies (see Table S3).
Normal laboratory food††
84 (84, 84)
46 (46, 48)**
Effects of diet on survival and mortality in glued mutants
To investigate the effects of diet on neuronal function in glued mutant motor neurons, we first sought to define the dietary conditions that result in the maximum effect of diet on lifespan in flies expressing the DNGlued transgene in all motor neurons (D42/DNGlued). A standardized sugar-yeast (SY) diet was developed and multiple diet dilutions were analyzed for effects on median lifespan in D42/DNGlued flies (Table S2; Supplemental Methods). This analysis revealed that the 0.5X and the 1.5X SY diet conditions resulted in the maximum difference in median lifespan in D42/DNGlued flies (Table S2). These diet conditions represent our low nutrient (0.5X) and high nutrient (1.5X) diet conditions for the subsequent analyses.
To determine the effects of diet on mortality and survival, lifespan data were collected from Gl1 flies and from D42/DNGlued and OK6/DNGlued flies raised on 0.5X or 1.5X diet (Fig. 1D,E, Table 1). We observed a robust effect of our diet on lifespan and mortality in D42/DNGlued animals including significant increases in median and maximum lifespan as well as reduced mortality in the D42/DNGlued female flies raised on the 0.5X diet (gray line) compared to flies raised on the 1.5X diet (black line; Fig. 1D and Table 1). Similar but reduced effects of diet were also observed male and female OK6/DNGlued flies (Table 1). Interestingly, we did not see a similar effect of dietary conditions on lifespan or mortality in Gl1 mutants raised under identical conditions suggesting that glued could be required outside of the motor neuron for the effects of diet on lifespan (Fig. 1E and Table 1).
glued mutants exhibit an accelerated decline in motor function with age
Previous studies have demonstrated that expression of antioxidants or protein chaperones within motor neurons is sufficient to extend lifespan indicating that motor neuron health is critical in promoting survival in flies (Parkes et al., 1998; Phillips et al., 2000; Morrow et al., 2004). Therefore, we hypothesized that the reduction in survival in D42/DNGlued flies may be the result of declining motor function. To investigate this possibility, the effects of glued mutations on motor function were investigated by analyzing a simple motor reflex in flies known as the Proboscis Extension Reflex (PER), a feeding behavior elicited in response to the stimulation of gustatory receptors located on the legs of the adult fly (Fig. 2A). PERs were visually evaluated in flies expressing the UAS-DNGlued in the CM9 motor neuron using the E49-Gal4 enhancer (E49/DNGlued; Gordon et al., 2009). The CM9 is one of at least nine muscle groups located within the fly head that are required for the extension of the proboscis during feeding, and genetic ablation of the CM9 muscle group has previously been reported to severely impair proboscis extension (Kimura et al., 1986; Demerec, 1994). The E49-Gal4 enhancer element expresses Gal4 in a pair of bilateral motor neurons that innervate CM9 and have been shown to be necessary and sufficient for normal proboscis extension (Gordon et al., 2009; see also Fig. 3). Thus, it is likely that changes in PER observed in E49/DNGlued flies would be reflective of changes in the function of the CM9 motor neuron.
To evaluate the effect of the expression of the DNGlued transgene within the CM9 motor neuron on PER, 8–10 female flies of each genotype were raised on normal laboratory food, loaded into pipette tips, given 5 tarsal stimuli and each resulting PER was scored on a 1–4 scale based on the extent of proboscis extension (Fig. 2A; Chabaud et al., 2006). Positive responses to stimuli were used to generate an average value (n = 40–50 events per genotype) referred to as the Proboscis Extension Index (P.E.I). Animals that fail to elicit at least two PER responses to the five stimuli were removed from analysis. This analysis demonstrated no significant differences in P.E.I. at 7 and 21 days between mutant and control flies (Fig. 2B,C). P.E.I. analysis from 21 days to 42 days showed an accelerated deterioration of proboscis extension in E49/DNGlued flies, which resulted in a highly significant reduction in P.E.I. in mutant flies compared to controls at 42 days (Fig. 2C).
glued mutants exhibit progressive deterioration of CM9 NMJ morphology
It is possible that the deterioration of motor function observed in E49/DNGlued flies at 42 days is a reflection of altered CM9 NMJ connectivity. Therefore, we analyzed the innervation pattern and synaptic morphology of the CM9 NMJs in mutant and control flies over time. We first performed extensive characterization of the CM9 NMJs. 7-day-old virgin female flies expressing a UAS-CD8-GFP transgene via the E49-Gal4 driver were dissected, fixed, and processed for immunofluorescence microscopy using antibodies against GFP and the PSD-95 homologue Discs-large (Dlg), a protein present postsynaptically at the larval NMJ (Koh et al., 1999). Analysis of bright field and DAPI staining demonstrated that the CM9 consists of 15.125 (±0.32) individual multi-nucleated muscle fibers (Fig. 3B, panels i and ii; Table S1). Analysis of Dlg and GFP staining revealed that every muscle group harbored an average of 32.9 (±1.1) separate innervations with each fiber receiving 2–3 individual Dlg-positive innervations (Fig. 3B panel iii and C). All Dlg-positive synapses were opposed by CD8-GFP expressed by the E49-Gal4 driver, and we did not observe any CD8-GFP innervations on the muscle surface that were not opposed by Dlg (Fig. 3B,C). Further immunofluorescent analyses found that many other larval synaptic antigens were also present at the CM9 NMJ including the presynaptic vesicular glutamate transporter (VGluT) (Figs 3D and S1C). These data are consistent with the CM9 motor neurons providing glutamatergic innervation to the CM9 that are similar in composition to the larval NMJ.
We next analyzed the innervation pattern and synaptic morphology of the CM9 NMJs as a function of age in E49/DNGlued and E49/+ control flies using immunofluorescent microscopy and antibodies against presynaptic VGluT and postsynaptic Dlg (Fig. 3D–H). We first determined that there are no differences in the number of CM9 innervations at 7, 21, or 42 days of age (Table S1). We then quantified the total presynaptic area of VGluT immunoreactivity from z-sections of CM9 NMJs at 7, 21, and 42 days and found that total synaptic area significantly increased by 15% between 7 and 42 days at E49/+ NMJs, but significantly decreased by 10% at E49/DNGlued NMJs over the same time period (Fig. 3F and Table S1). We also observe a significant reduction in synaptic area at 42 days when comparing E49/DNGlued NMJs to E49/+ NMJs (Fig. 3E). We do not see a difference in synaptic area at 42 days between the E49/+ and DNGlued/+ controls (Table S1). In addition, we observed a general disorganization of the presynaptic nerve terminal at 42-day-old CM9 NMJs in E49/DNGlued flies compared to control flies (Fig. 3D, panel iv).
We further characterized the glued mutant NMJ phenotype by assaying for the presence of sub-synaptic phenotypes previously used to characterize larval mutant NMJs. First, we quantified the presence of synaptic sprouts, defined as an event where the presynaptic VGluT extends beyond the postsynaptic Dlg staining (Fig. S1D, arrow; Marie et al., 2004; Koh et al., 2004). This analysis demonstrated significantly more presynaptic sprouts per CM9 NMJ in E49/DNGlued flies compared to control flies at both 21 and 42 days (Fig. 3G). We also quantified the number of synaptic retractions per CM9 NMJ, defined as regions of Dlg not opposed by presynaptic VGluT staining and previously shown to be a prominent synaptic phenotype at glued mutant NMJs in larvae (Fig. S1D, arrowhead; Eaton et al., 2002). We found that the number of synaptic retractions per CM9 NMJ was only increased at 42-day-old E49/DNGlued NMJs compared to 42-day-old E49/+ NMJs (Fig. 3H). Again, we found no differences in the number of sprouts or retractions at 42 days between E49/+ and DNGlued/+ controls (Table S1). Finally, we found that neither genotype exhibited any age-related change in CM9 fiber number, myonuclei number, or gross organization of T-tubules in the muscle between 7 and 42 days (Table S1 and data not shown). Taken together, these data are consistent with a progressive deterioration of the CM9 NMJ morphology in E49/DNGlued flies that correlates with the accelerated decline seen in motor function (P.E.I.) observed in 42-day-old mutant flies (Fig. 2C).
Effects of diet on the declining motor function in glued mutants
Although we did not observe a difference in motor performance at 21 days between mutant and control flies using the P.E.I. assay (Fig. 2C), we suspected that the rate of proboscis extension in E49/DNGued flies was already compromised at 21 days. Because we are interested in early pathogenic events, we developed a new method to more accurately measure the velocity of proboscis extension at 21 days. Bristles on the tip of the proboscis were tracked during an extension event using high-speed video microscopy and particle-tracking software (Fig. 4A). From these PER videos, bristle paths were constructed which allowed for the analysis of numerous metrics including average velocity, maximum velocity, and the net displacement of the proboscis tip.
Using the bristle-tracking method, we verified that E49/DNGlued flies had a significant reduction in average velocity compared to E49/+ control flies at 21 days (Fig. 4B). These experiments were performed under our dietary paradigm to determine the effects of diet on motor neuron function. We observed that the reduction in average velocity in E49/DNGlued flies compared to E49/+ controls was present in flies raised on either the 0.5X diet or the 1.5X diet (Fig. 4B). Maximum velocity was also significantly reduced in E49/DNGlued flies compared to E49/+ controls when raised on the 1.5X diet but not in flies raised on the 0.5X diet (Fig. 4C). Consistent with our P.E.I data, we did not observe any significant changes in the distance that the proboscis is extended in E49/DNGlued flies compared to E49/+ flies at 21 days of age regardless of diet (Fig. 4D). These data demonstrate that expression of the UAS-DNGlued transgene within the CM9 motor neuron results in a reduction in the velocity of extension regardless of dietary condition.
We next investigated the age dependence of the decline in PER between 7 and 21 days in E49/DNGlued and E49/+ flies raised on 0.5X diet (gray diamonds) or 1.5X diet (black squares). First, we observed that both average velocity (Fig. 4E) and maximum velocity (Fig. 4F) decline significantly between 7 and 21 days in E49/+ control flies raised on either food condition. Importantly, there were no significant differences between average and maximum velocity in E49/+ flies comparing dietary conditions at 7 or 21 days (Fig. 4E,F). We next investigated the effects of diet on motor function in mutant flies. E49/DNGlued flies fed the 1.5X diet also showed a significant decline in both average and maximum velocity between 7 and 21 days, similar to controls (Fig. 4G,H, black line). In contrast, E49/DNGlued flies fed the 0.5X diet did not exhibit a significant decline in either average velocity or maximum velocity between 7 and 21 days (Fig. 4G,H, gray line). This trend results in a significantly greater average and maximum velocity at 21 days in E49/DNGlued flies fed the 0.5X diet compared to E49/DNGlued flies fed the 1.5X diet (Fig. 4G,H, gray diamonds versus black squares; also see Fig. 4B,C. black bars). These data demonstrate that the 0.5X diet reduced the decline in motor function compared to the 1.5X diet in E49/DNGlued flies, but had no effect on the decline in motor function observed in E49/+ flies.
glued mutant CM9 NMJs have impaired neurotransmission
To further define the neuronal dysfunction in glued mutants, we analyzed synaptic transmission at CM9 NMJs from 21-day-old E49/DNGlued and E49/+ flies. We found that average amplitudes of evoked endplate junctional potentials (EJPs) are significantly reduced at NMJs from E49/DNGlued flies compared to E49/+ flies regardless of dietary condition (Fig. 5A,B). We also observed that the amplitudes of miniature EJPs (mEJPs), also referred to as quantal size, were found to be significantly reduced at NMJs in E49/DNGlued flies raised on the 0.5X diet compared to mEJP amplitudes at NMJs in E49/+ flies on the same diet (Fig. 5E,F). A similar, but not significant, trend was observed in mEJP amplitudes in flies raised on the 1.5X diet (P = 0.4). Analysis of quantal content, a measure of presynaptic release determined by dividing the EJP amplitudes by the mEJP amplitudes, revealed no significant difference in vesicle release in mutant flies compared to controls (Fig. 5C). In these recordings, we did not observe any significant differences in resting membrane potential or input resistance of the CM9 (Table 2). These data demonstrate that neurotransmission is impaired at the CM9 NMJ in 21-day-old E49/DNGlued flies compared to controls and suggest that the observed decrease in average EJP amplitude is at least partially the result of reduced mEJP amplitudes.
Table 2. Electrophysiological analysis of CM9 neurotransmission
Genotype and diet
All values were obtained from recordings made from CM9 fibers from virgin females of the indicated genotypes except for values obtained from muscle 6 of 3rd Instar larvae. Significance determined between 1.5X values versus 0.5X values of same genotype using a student’s t-test. *P < 0.05. **P < 0.01.
†Represents the number of individual animals from which recordings were obtained. Only one recording per individual was used.
‡Values represent the maximum negative muscle potential recorded.
§Values represent the average input resistance.
¶Values represent the average endplate junctional potentials (EJP) amplitude from CM9.
††Values represent the average miniature EJPs (mEJP) amplitude of 100 events/neuromuscular junction (NMJ).
‡‡Values represent the frequency of mEJPs in Hz.
§§Values represent the average quantal content.
3rd Instar larvae
The effects of diet on neurotransmission at mutant CM9 NMJs
We next performed electrophysiology on 21-day-old E49/DNGlued flies raised on either the 0.5X or the 1.5X diet and found that E49/DNGlued flies raised on the 0.5X diet had significantly larger evoked EJP amplitudes than E49/DNGlued flies raised on the 1.5X diet (Fig. 5A,B, black bars; Table 2). Importantly, analysis of quantal size reveals that the average amplitude of the mEJPs did not significantly change at mutant CM9 NMJs as a result of diet demonstrating that the effects of diet on neurotransmission are not because of changes in mEJP amplitude (Fig. 5E, black bars). We also did not observe any effects of diet on the passive properties of the E49/DNGlued CM9 including the resting membrane potential (RMP) or input resistance (IR; Table 2) suggesting that increased synaptic vesicle release is responsible for the difference in EJP amplitude observed between the 0.5X and 1.5X diets. In support, we found that quantal content is significantly larger at E49/DNGlued NMJs in flies raised on the 0.5X diet compared to the 1.5X diet resulting in almost twice as many synaptic vesicles being released per action potential (Fig. 5C, black bars; Table 2).
Because quantal content could be influenced by synapse size, we investigated the effects of diet on NMJ morphology in E49/DNGlued flies. CM9 NMJs from 21-day-old E49/DNGlued flies raised on either the 0.5X or the 1.5X diet were processed for immunofluorescent microscopy. We found no significant changes in any of our previously described measures of synaptic morphology at 21-day-old E49/DNGlued NMJs as a result of dietary conditions, including no change in synaptic area, synaptic retractions, or synaptic sprouts (Fig. S2C–E). We also did not observe a change in muscle fiber number, T-Tubule morphology or myonuclei density in mutant CM9 under either diet condition (Table S1; data not shown). These data support that the larger quantal content observed at NMJs in E49/DNGlued flies raised on the 0.5X diet is not because of increased synapse size.
Diet can alter quantal content at healthy CM9 NMJs
We also observed a strong effect of our diet on synaptic vesicle release at NMJs in E49/+ flies (Fig. 5A–C, gray bars). Analysis of EJP amplitudes at E49/+ CM9 NMJs determined that average evoked EJP amplitudes are significantly larger in flies raised on the 0.5X diet compared to flies raised on the 1.5X diet (Fig. 5B, gray bars). In these recordings, we did not observe significant differences in mEJP amplitudes (Fig. 5E, gray bars), resting membrane potential (RMP) or input resistance (IR) because of the diet conditions (Table 2). Analysis of quantal content supports that the increase in EJP amplitudes observed in flies raised on the 0.5X diet compared to flies raised on the 1.5X diet is because of increased synaptic vesicle release (Fig. 5C, gray bars). There was no difference in the average frequency of spontaneous mEJPs events in flies raised on 0.5X compared to 1.5X food (Table 2). To investigate whether the effects of diet on synaptic function were also present at the larval NMJ, we performed electrophysiology on 3rd instar larvae raised on either the 0.5X diet or the 1.5X diet. We observed no difference in EJP amplitude, mEJP amplitude, or quantal content comparing larvae raised on either diet condition (Table 2). These results demonstrated that the effects of diet on neurotransmission appear to be specific to the adult NMJ.
Despite the numerous observations of the therapeutic effects of dietary restriction during neurodegenerative disease, synaptic effects of diet are not well defined. In this study, we demonstrate a link between diet and the release of synaptic vesicles from the NMJ, revealing a novel synaptic mechanism for the effects of diet on nervous system function. For this investigation, we developed a novel motor system, the CM9 NMJ, which combines analyses of motor behavior with high-resolution analysis of synaptic structure and function. An important property of this system is that we are able to measure neurotransmission at a synapse that is necessary and sufficient for a defined motor behavior, proboscis extension, allowing for a strong correlation between changes in synaptic function and motor output (Gordon et al., 2009). In addition, the highly restricted expression of the E49-Gal4 supports that the observed changes in proboscis extension in E49/DNGlued flies are reflective of CM9 motor neuron dysfunction.
Diet, safety factor, and lifespan in glued mutants
To insure muscle contraction in the face of stress, the NMJ in vertebrates and invertebrates releases more neurotransmitter than is required to depolarize the muscle and initiate contraction (Marrus & DiAntonio, 2005; Wood & Slater, 2001). The excess synaptic vesicles released divided by the number of vesicles required to generate a muscle contraction is referred to as the ‘safety factor’ of the NMJ and can range from 2 to 10 at various NMJs (Wood & Slater, 2001). Importantly, there exist a number of diseases, such as the myasthenic syndromes, which are characterized by reduced safety factors and muscle weakness demonstrating a link between impaired safety factor and motor function (Shigemoto et al., 2010). The etiology of these diseases consists of reduced presynaptic release of neurotransmitter and/or reduced sensitivity of the muscle to released neurotransmitter.
We observed a nearly 2-fold difference in evoked EJP amplitude and quantal content at control CM9 NMJs when comparing flies raised on the 0.5X diet with flies raised on the 1.5X diet (Fig. 5B and C, gray bars). Despite this increase in vesicle release, we do not observe significant increases in either average or maximum velocity of proboscis extension in E49/+ flies that were raised on the 0.5X diet compared with flies raised on the 1.5X diet (Fig. 4E,F). This is consistent with the CM9 NMJ having a sufficiently large safety factor to insure normal muscle contraction in flies raised on the 1.5X diet. We also found that diet had no effect on the significant decline in extension velocities observed between 7 and 21 days in control flies supporting that this decline is not because of declining neuronal function (Fig. 4E,F). These data are consistent with previous studies reporting modest effects of diet on wall climbing behavior in wild-type flies (Bhandari et al., 2007).
In contrast, we find that the decline in motor function observed between 7 and 21 days in E49/DNGlued flies is reduced in flies raised on the 0.5X diet compared to flies raised on the 1.5X diet (Fig. 4G,H). This reduction in motor function decline resulted in a significantly greater extension velocity in 21-day-old mutant flies raised on the 0.5X diet compared to mutant flies raised on the 1.5X diet (also see Fig. 4B,C, black bars). One possibility is that the decline in motor function observed between 7 and 21 days in glued mutant flies is the result of diet-sensitive pathologies that are distinct from than those involved in the decline in motor function observed in wild-type flies during the same time period. In support of the existence of distinct pathological and nonpathological aging processes, we observe that synaptic area decreases with age at glued mutant NMJs, but increases at control NMJs over the same period (Fig. 3F). A second possibility is that the effects of diet on declining motor behavior are only manifest when neurotransmission falls below a certain threshold (e.g. the safety factor). In this context, the improvement of extension velocity in glued mutants raised on the 0.5X diet is because the glued mutant NMJ is operating below safety factor and thus sensitive to the effects of increased presynaptic release.
Regardless of the underlying mechanism, our data support that the larger EJP amplitude (approximately 2X) in glued mutant flies raised on the 0.5X diet compared to mutant flies raised on the 1.5X diet contributes to the improved proboscis extension in mutants at 21 days and suggests a correlation between EJP size and motor output in glued mutant flies. In support, we found that E49/DNGlued flies that were raised on the 0.5X diet had extension velocities and EJP amplitudes that were not significantly different to the values observed in E49/+ flies raised on the 1.5X diet (P > 0.2; Figs 4E–H and 5B). In combination with the lack of changes observed of the morphological or electrical properties of the muscle, our data support that the 0.5X diet can improve motor output in the E49/DNGlued flies by increasing presynaptic release. This effect on motor output is not manifest in wild-type because of the intact safety factor.
Given the established role of motor neuron health on influencing lifespan in flies (Parkes et al., 1998; Phillips et al., 2000; Chan et al., 2002; Morrow et al., 2004), it seems reasonable to speculate that motor function contributes to the determination of lifespan in D42/DNGlued flies. Our data in E49/DNGlued flies demonstrate a correlation between motor function and neurotransmission in response to diet. In addition, our synaptic analyses demonstrate that mutant synaptic pathologies, such as mEJP amplitudes and increased synaptic sprouts, are not reduced by rearing mutant flies on the 0.5X diet supporting that the increase in survival in D42/DNGlued mutants raised on the 0.5X diet is not because of the reduction of glued mutant pathology in motor neurons per se. This would support that the improved motor neuron function and the resulting increase in neurotransmission could be major contributors to the increase in lifespan observed in D42/DNGlued mutant flies reared on the 0.5X diet, although other nonautonomous factors cannot be ruled out. Because safety factor is intact in a wild-type fly, it is unlikely that this mechanism contributes to lifespan extension in control flies raised on 0.5X.
How does diet alter neurotransmission?
One possibility is that the 0.5X diet induces diet-responsive cellular processes within the motor neuron that enhance synaptic vesicle release. Because we observe this effect of diet in mutant flies, this would require that these diet-dependent processes function independently of the dynactin complex. This is not trivial because of the established requirement for the dynactin complex in a number of important cellular processes, such as autophagy, that have previously been implicated in the therapeutic effects of dietary restriction (Ravikumar et al., 2005; Di Bartolomeo et al., 2010; Gamerdinger et al., 2011). Dynactin complex function is also required during neurotrophic signaling in motor neurons, which is known to control synaptic efficacy throughout the nervous system, including at the larval NMJ (Delcroix et al., 2003; McCabe et al., 2003; Cosker et al., 2008). To date, a role for dynactin complex function during insulin or TOR signaling has not been documented making these pathways attractive candidates for mediating the effects of diet on synaptic efficacy.
It is interesting that we do not observe a similar effect on lifespan by diet in somatic Glued1 mutants compared to D42/DNGlued flies (Fig. 1E and Table 1). One possibility is that mortality in Glued1 mutants is the result of a distinct diet-insensitive morbidity compared to D42/DNGlued mutants. Although we cannot rule out this possibility, it should be noted that the nervous system is the tissue predominantly affected in dynactin complex mutations in humans, mice, and flies (Reddy et al., 1997; Martin et al., 1999; Eaton et al., 2002; Puls et al., 2003; Münch et al., 2004, 2005; Laird et al., 2008). Another possibility is that the dynactin complex has a nonautonomous role outside of the motor neuron during the dietary response. For example, dynactin complex function could be required within the median neurosecretory cells that are necessary for lifespan extension by diet, and Glued1 mutants have impaired signaling from these cells (Broughton et al., 2010). This would support that the effects of diet on motor neuron output are part of the ensemble of physiological changes that occur throughout the animal in response to the diet-responsive signal originating from the median neurosecretory cells.
In contrast, it is possible that neurotransmission at the CM9 NMJ is sensitive to the content of the diet and that increased dietary intake negatively impacts neurotransmission because of the increased production of metabolites and reaction products, such as reactive oxygen species (ROS) and ATP. In the nervous system, it has been shown that ROS can directly modify the activity of a number of ion channels and alter the lipid composition of the nerve terminal (Adibhatla & Hatcher, 2010; Sesti et al., 2010). Thus, it is possible that synaptic vesicle release is scaled by diet because of the resulting oxidative state of the nerve terminal. In addition, increased levels of ATP in neurons have been shown to activate ATP-sensitive potassium channels resulting in reduced neuronal activity (Ma et al., 2007). Therefore, increased ATP levels resulting from the 1.5X diet within the motor neuron could cause the activation of ATP-sensitive potassium channels leading to a reduction in excitability of the motor neuron. Importantly, these biochemical mechanisms might be predicted to effect synaptic release at numerous synapses producing a broad effect of diet on synaptic efficacy.
The D42-Gal4 and OK6-Gal4 lines fly lines were previously described (Eaton et al., 2002). The UAS-DNGlued line used was created using standard transposase mediated P-element mobilization of the UAS-DNGlued96 transgene located on the second chromosome to the third chromosome (Eaton et al., 2002). The E49-Gal4 line was obtained from the Kristin Scott laboratory (Gordon et al., 2009). All stocks used in this study were backcrossed ten generations to the w1118 strain. All lifespan analyses were performed on virgin female flies kept on a 12-h light/dark cycle at 25 °C and 60% humidity. Dietary restriction was performed using dilution of a standard SY diet (see Data S1). All food was made fresh every week and flies flipped every 2 days to minimize water loss for all diet conditions. The following genotypes were abbreviated in the text: D42-Gal4/UAS-DNGlued = D42/DNGlued; OK6-Gal4/+; UAS-DNGlued/+ = OK6/DNGlued; E49-Gal4/+; UAS-DNGlued/+ = E49/DNGlued; UAS-DNGlued/+ = DNGlued/+; E49-Gal4/+ = E49/+.
An Ellison Medical Foundation New Scholar Award (AG-NS-0415-07) to BAE supported this work. JMR is supported by grant T32-AG021890 from the NIA.
Joel M. Rawson-concept and design, data acquisition and interpretation, manuscript preparation. Holly Davison-data acquisition. Tabita Kreko-data acquisition. Rebekah Mahoney-data acquisition. Leo Chang-data acquisition. Alex Bokov-data interpretation. Jon Gelfond-data interpretation. Greg T. Macleod-data interpretation. Benjamin A. Eaton-concept and design, data acquisition and interpretation, manuscript preparation.