Alpha- and beta-adrenergic octopamine receptors in muscle and heart are required for Drosophila exercise adaptations

Summary

Endurance exercise has broadly protective effects across organisms, increasing metabolic fitness and reducing incidence of several age-related diseases. Drosophila has emerged as a useful model for studying changes induced by chronic endurance exercise, as exercising flies experience improvements to various aspects of fitness at the cellular, organ and organismal level. The activity of octopaminergic neurons is sufficient to induce the conserved cellular and physiological changes seen following endurance training. All 4 octopamine receptors are required in at least one target tissue, but only one, Octβ1R, is required for all of them. Here, we perform tissue- and adult-specific knockdown of alpha- and beta-adrenergic octopamine receptors in several target tissues. We find that reduced expression of Octβ1R in adult muscles abolishes exercise-induced improvements in endurance, climbing speed, flight, cardiac performance and fat-body catabolism in male Drosophila. Importantly, Octβ1R and OAMB expression in the heart is also required cell-nonautonomously for adaptations in other tissues, such as skeletal muscles in legs and adult fat body. These findings indicate that activation of distinct octopamine receptors in skeletal and cardiac muscle are required for Drosophila exercise adaptations, and suggest that cell non-autonomous factors downstream of octopaminergic activation play a key role.

Keywords:

Drosophila melanogaster – RNA interference – Fats – heart – exercise – Cardiac muscles – Climbing – Myosins

Introduction

Endurance exercise is a potent, low-cost intervention with broad healthspan-extending effects [1]. Chronic endurance training simultaneously promotes healthy physiology and prevents disease, improving function in heart, skeletal muscle and brain while reducing obesity, heart disease and cognitive decline [2–4]. These benefits are associated with adaptive changes to gene expression and metabolism [2,5–8].

Nonetheless, these benefits are inaccessible to much of the population that are unable to perform an endurance exercise regimen because of injury, illness, advanced age, or lifestyle. Dissecting the mechanisms underlying exercise adaptations in order to identify exercise mimetics remains a prominent research goal. Until recently, studies of life-long exercise effects were limited to rodent models and retroactive comparisons of human cohorts, making controlled, longitudinal analysis and large-scale genetic studies difficult. We and others have developed endurance training programs for Drosophila, taking advantage of their innate instinct for negative geotaxis, allowing for controlled training of large, genetically identical cohorts [9,10]. After our 3 week training protocol, male flies increase climbing speed, cardiac stress resistance [11], endurance [12], flight performance [5], and lysosomal activity in their fat body [7]. Trained male flies have increased mitophagy in cardiac and skeletal muscle [13], increased mitochondrial enzyme activity [11,14] and changes in transcript expression similar to those found in long-lived flies [5]. These genetic and physiological adaptations closely resemble benefits seen in both rodent models [15] and humans [1].

We have previously found that increased activity of octopaminergic neurons is both necessary and sufficient for exercise adaptations in Drosophila, even in sedentary flies where mobility is restricted by a foam stopper placed low in the vial [8]. Octopamine signals through conserved α and β-adrenergic receptors [16,17], similar to those that bind vertebrate norepinephrine. Both norepinephrine and octopamine have been associated strongly with the fight-or-flight response. In addition to increasing the drive to exercise, octopamine and norepinephrine are known to facilitate transient increases in endurance, lipolysis, and fatty acid metabolism [18–23]. As either daily exercise or daily short-term activation of octopaminergic neurons is sufficient to provide the benefits of endurance exercise in Drosophila [8], and chronic exercise stimulates increased norepinephrine secretion in humans [24], these mechanisms appear to be conserved.

Norepinephrine is produced neuronally and also released into circulation from the adrenal gland [25,26]. Its signal can be received by a family of alpha- and beta- adrenergic receptors that are expressed in complementary patterns [27]. The specific requirements for each of these receptors in various tissues in driving the response to chronic exercise is incompletely understood in any organism. While Drosophila do not have an adrenal gland, neuronally produced octopamine can signal to neighboring cells or be released into circulation, where it can potentially bind receptors in a variety of tissues [17]. Here, we utilize the Drosophila system to uncover tissue-specific requirements for each receptor in executing the adaptive response to chronic exercise.

Drosophila have one α-adrenergic octopamine receptor, OAMB, and 3 dedicated β-adrenergic octopamine receptors, Octβ1R, Octβ2R, and Octβ3R [17]. We have previously established that all 4 are required for at least some aspect of exercise adaptation in Drosophila, but only one, Octβ2R, was required for all of the characteristic adaptations when knocked down using global, inducible RNAi [8]. These results are consistent with studies in which β-blockers tend to inhibit exercise performance and adaptation in humans [28,29].

Here, we separately map receptor requirements for exercise adaptations using the UAS-Gal4 system and Gene Switch Gal4 to reduce adrenergic receptor expression in heart, adult fat body and adult skeletal muscle. We find that Octβ2R expression is required in skeletal muscle for improvements to endurance and speed, but also cardiac performance and fat-body autophagy, suggesting tissue non-autonomous effects. OAMB and Octβ3R have tissue-specific effects on exercise adaptations, with OAMB being more important for cardiac improvements and Octβ3R essential for flight. Intriguingly, we also find cell non-autonomous effects resulting from reductions in cardiac-specific OAMB and Octβ1R knockdown, as well as muscle-specific OAMB and Octβ1R knockdown.

Results

Octopamine receptor expression is tissue-specific

OAMB, Octβ1R, Octβ2R and Octβ3R transcript expression was measured in heart, muscle and fat body using qRT-PCR, in agreement with previously published reports [17] (S1A–S1C Fig). All 4 receptors were detected in adult muscle (S1B Fig), while only Octβ3R transcript expression was found in adult fat body (S1C Fig). We also detected for the first time OAMB and Octβ1R in hearts, supported by RT-PCR and subsequent physiological analyses. Octβ2R and Octβ3R transcripts were not detected in heart tissue (S1A Fig). Because skeletal muscle, heart and adipose tissue are the primary tissues where we have characterized changes following chronic exercise, and because OA receptors are expressed in each of these tissues, we set out to map which receptors are required in each of these key tissues to execute the effects of chronic exercise training.

We first tested knockdown efficiencies for each RNAi construct used with each driver, separately testing at 72 hours after induction by addition of RU486 (see methods), and again at 25 days after induction (S1A–S1C Fig). Unexercised and exercised cohorts were also separately measured. Knockdowns ranged from 50% to 95% and in most cases were more efficient at 25 days then at 72 hours. There was no consistent effect of exercise on knockdown efficiency.

OAMB is required in muscle and heart for improved climbing speed, endurance and cardioprotection with exercise

Runspan, a measure of endurance in which Drosophila time to fatigue is scored in real time and plotted similarly to a survival curve (see methods), was scored on day 5 post-eclosion, 72 hours after induction of RNAi expression through RU486 feeding. RNAi against OAMB (Fig 1) in neither adult muscles (MHC GS RU+, Fig 1A) nor adult hearts (hand GS RU+, Fig 1G) altered baseline endurance. The same cohort of flies was given 3 weeks of exercise training and assessed again for endurance on day 25. Typically, exercise-trained wild-type flies run longer than genetically identical, untrained siblings that have been placed on the machine on each day, but with a foam stopper to prevent climbing as a control for the exercise environment [9]. Genetic background controls were genetically identical but lacked the inducing drug for the RNAi (RU-). Flies with muscle- or heart- specific OAMB RNAi failed to increase endurance with exercise, while control flies responded to exercise with increased endurance as normal (Fig 1B and 1H).

Wild-type exercise-trained males retain greater negative geotaxis climbing speed across ages than age-matched control siblings [11]. We assessed negative geotaxis speed 5 times per week prior to the start of daily training, as described [30]. RNAi against OAMB in adult muscles prevented adaptation in trained flies, while RU- controls improved normally with exercise (Fig 1C). Heart-specific OAMB knockdown more severely impaired climbing speed, with exercised flies actually climbing slower than unexercised siblings. RU- control flies responded to exercise normally (Fig 1I).

Exercise training is cardio-protective in wild-type males, as measured by response to external electrical pacing stress [31]. Both muscle- and heart-specific OAMB knockdown prevented cardioprotective benefits of exercise training, while RU- controls responded normally to exercise (Fig 1E and 1K).

Exercise-trained wild-type male flies also have better flight performance as measured by recording landing height after ejection from a platform [8], and wild-type male flies increase autophagy in the fat body [8] during chronic exercise. Knockdown of OAMB in muscle or heart did not prevent exercise from increasing flight ability (Fig 1D and 1J) and fat body LysoTracker staining (Fig 1F and 1L and S9A and S9B Fig).

Octopamine β1 receptor is specifically required in muscle for adaptive response to chronic exercise

Muscle-specific knockdown of Octβ1R (MHC GS Octβ1R RU+) reduced baseline endurance in comparison to uninduced controls (Fig 2A). Importantly, Octβ1R reduction did not prevent repetitive climbing exceeding 400 minutes, meaning that it did not preclude these flies from performing our exercise regimen. Muscle-specific Octβ1R RNAi completely prevented exercise from increasing endurance, and Octβ1R RNAi flies ran shorter than untrained RU- controls whether exercised or not (Fig 2B). Octβ1R was also required in muscles for exercise-dependent improvements in climbing speed, flight, and cardiac performance (Fig 2C–2E). Fat body LysoTracker staining was abnormal in muscle-specific Octβ1R RNAi flies, with high lysosomal activity whether exercised or not, and unexercised flies actually showing higher activity (Fig 2F, S9C Fig). Heart-specific knockdown of Octβ1R (hand GS Octβ1R RU+) had similar detrimental effects on baseline and post-training runspan, climbing speed, and LysoTracker staining, with exercised RU+ flies all performing similar to or worse than untrained, uninduced RU- controls (Fig 2G–2I and 2L, S9D Fig). Octβ1R reduction in adult hearts did not affect adaptations in landing height or cardiac stress resistance, however (Fig 2J and 2K).

Octopamine β2 and β3 receptor expression in muscle separately coordinate exercise adaptations

Adult-specific reduction of neither Octβ2R nor Octβ3R significantly affected baseline endurance (Fig 3A and 3G). In contrast, knockdown of either receptor prevented adaptation to chronic exercise (Fig 3B and 3H). RNAi against Octβ2R in adult muscle caused unusual climbing phenotypes, with unexercised knockdown flies climbing faster than controls, but actually becoming slower when exercise-trained (Fig 3C). Muscle-specific Octβ3R knockdown flies had normal baseline performance, but exercise training significantly worsened their climbing speed (Fig 3I). Knockdown of Octβ2R or Octβ3R in adult muscle also reduced normal increases in LysoTracker staining in trained RU+ flies, (Fig 3F and 3L S9E and S9F Fig) while exercised RU- flies adapted with exercise normally in all assessments (Fig 3, RU- EX). Muscle specific Octβ2R RNAi did not block adaptations to flight performance (Fig 3D), but Octβ2R was required in muscles for the cardioprotective effect of exercise (Fig 3E). In contrast, muscle–specific Octβ3R RNAi did not block exercise-induced cardiac improvements (Fig 3K), but did prevent improvements to flight (Fig 3J).

Octopamine β3 receptor is important for fat body homeostasis

All 4 octopamine receptors tested here are known to be present in adult brain, with developmental and tissue-specific activities that are context dependent [17]. Among tissues tested here, Octβ3R transcript was only detected in adult fat body, and has been previously reported to be present at low levels in hindgut and Malpighian tubules [17]. Octβ3R RNAi in adult fat body had no effect on baseline or post-training endurance, exercise-induced climbing improvement, or resistance to pacing-induced cardiac stress (Fig 4A, 4B, 4C and 4E). The major effect of Octβ3R knockdown in adult fat body was a cell-autonomous block of LysoTracker staining accumulation after exercise (Fig 4F and 4G). Perhaps surprisingly, Octβ3R expression was also required in adult fat body for exercise-dependent increases in flight performance (Fig 4D).

Given that exercised male flies increase lysosomal activity and mitochondrial turnover in the fat body [8,13] and flies with defects in fatty acid metabolism increase lipolysis following exercise [7], we tested whether flies with reduced Octβ3R expression in adult fat body also failed to upregulate autophagy by examining dAtg8-II/I ratio via Western blot. Autophagy downregulation is observed by a decrease in the dAtg8-II/I ratio, indicating a reduction in activated dAtg8-II, while restoration or upregulation is represented by an increase in the activated form [32]. We also tested exercise trained octopamine-fed flies, as both exercise and octopamine have been previously shown to increase autophagy and lipolysis in multiple flying insect species [20,33]. dAtg8-II/I ratios trend toward an increase in exercise-trained flies and were significantly higher in OA-fed whole flies, and this effect was blocked when Octβ3R was knocked down in adipose tissue (Fig 4H and 4I).

OA feeding rescues phenotypes of some, but not all, octopamine receptor knockdowns

Either 5μM octopamine (OA) feeding or intermittent OA-ergic neuron activation is sufficient to replicate exercise adaptations in sedentary Drosophila [8]. To test whether surplus OA could overcome the effects of tissue-specific octopamine receptor depletion, we repeated the RNAi experiments above but fed OA to half the flies. We selected MHC GS>Octβ2R and MHC GS>Octβ3R RNAi flies for feeding tests since these lines were directly comparable (same driver) but blocked distinct and separable exercise adaptations. Runspan was tested on day 5 post-eclosion, after 3 days of feeding with 5μM OA or vehicle and/or RU486 if Gene-Switch Gal4 was employed. Drug/vehicle feeding continued until the end of experimentation. A summary of OA-feeding+exercise results is in S1 Table. As in the first repetition, knockdown of Octβ2R in skeletal muscle prevented exercise-induced improvements. OA-fed control flies had performance characteristic of exercised flies, as previously observed [Figs 8] (S8B Fig). (Compare 5A to S1B RU-EX). However, OA-feeding did not restore the exercise response to flies lacking Octβ2R in muscle, as measured by endurance (compare Fig 5A RU+ OA-fed flies to S8B RU+) or climbing speed (Fig 5B, S8C Fig).

As above, improvements to flight performance after exercise did not require Octβ2R in adult muscles (S8D Fig, Fig 3D). Both exercise training and OA feeding were able to improve flight to levels that were similar to trained control flies (Fig 5C, compare to S8D Fig). Confirming results shown above, Octβ2R was required in muscle for cardio-protective effects of exercise, and for increased adipose lysosomal activity (S8E and S8F Fig) OA-feeding was completely unable to rescue these effects, although it successfully mimics exercise in OA-fed RU- controls (Fig 5D–5F). These results indicate that Octβ2R is absolutely required in muscle for chronic exercise to increase endurance, cardiac performance and lysosomal activity, even if exogenous OA is supplied.

Octβ3R

Muscle-specific Octβ3R knockdown does not alter baseline endurance (S8G Fig). Following endurance training, Octβ3R knockdown flies again failed to increase endurance. However, they did increase performance when supplemented with OA, in both exercised and unexercised cohorts, suggesting that the requirement for Octβ3R can be partially circumvented by other receptors, if exogenous OA is present (Compare Fig 6A RU+ to S8H Fig RU+).

Exogenous OA was able to stimulate performance of muscle-specific Octβ3R RNAi flies, with both exercised and unexercised cohorts responding to OA feeding. OA feeding also improved heart performance in MHC GS>Octβ3R RNAi flies, with exercised and unexercised RU+ groups receiving as much cardioprotection as RU- controls (Fig 6D). By contrast, OA feeding did not alter the effect of muscle-specific Octβ3R knockdown on flight performance or lysosomal activity (Fig 6C, 6E and 6F; S8J–S8L Fig).

Thus, in general, exercise adaptations that were unaffected by a specific knockdown responded to OA feeding as normal, but adaptations that required a particular receptor were not rescued by OA feeding. This strongly suggests that some receptors are specifically required for particular exercise adaptations, and not all OA receptors are interchangeable in this context.

Discussion

Octopamine signaling is a vital mediator of behavior and metabolism, and is critical for exercise adaptation in Drosophila [8,30]. OA directly affects muscle contractility in larval body wall muscle [34,35], metabolism [36], mobility in response to starvation [37] and fat storage [36], all of which may be important mechanisms modulating exercise adaptations. OA signals through various receptors that have been found to regulate essential processes from egg-laying to sleep, metabolism, learning and memory, and social aggression [38–42].

Octopamine is analogous to vertebrate norepinephrine, and noradrenergic signaling is known to be important in the human exercise response. While transient increases in OA-ergic signaling are sufficient to replicate exercise adaptation in sedentary Drosophila [8], prolonged effects of increased catecholaminergic signaling in humans would have adverse effects on blood pressure and heart rate [43,44]. Indeed, in our studies combining OA-feeding and endurance exercise, we often see less benefit than OA-feeding or exercise alone, suggesting that activation of OA-receptors may become deleterious if activated at too high a level even in invertebrate systems. Here, we demonstrate that OA acts during exercise to stimulate autophagic flux in Drosophila. Taken together, our observations suggest that adrenergic signaling is an important mechanistic part of the conserved adaptive response to endurance exercise.

Although OA-receptors are thought to act through highly conserved canonical signaling pathways [45–47], we find strong evidence that their activity is not interchangeable in the context of exercise, as has been previously demonstrated in the context of female reproduction (41). Knockdown of any of the four receptors tested here (a recently discovered receptor that responds to both OA and serotonin was not examined here [48]) eliminates some portion of the response to chronic exercise, and, in most cases, the response cannot be rescued by stimulating other receptors with OA feeding. This is true even when knockdown takes place in tissues that express multiple receptors, strongly implying that downstream effects of these receptors are not identical.

Another key finding here is that several receptor knockdowns produced clear tissue-non-autonomous effects, particularly knockdowns in skeletal or cardiac muscle. These effects could result from improved muscle and heart performance that alters the overall metabolic environment during chronic exercise; for example, improved cardiac performance could improve circulation to other tissues. An intriguing alternative, albeit non-mutually exclusive hypothesis would be that OA-receptors promote release of circulating factors from muscle or heart that affect metabolism in the fat body. These hypotheses are currently under further investigation. These results are consistent with the prior observations that restoration of exercise adaptations to female flies required masculinization of all Tdc2-expressing neurons, no smaller subset was capable of this effect (8). This strongly implies that the effect of OA is not solely mediated by a particular circuit, but at least in part requires release of OA into circulation, where it can then be received by receptors in various tissues.

An unexpected finding was the requirement for OAMB in the heart, where it has not previously been reported to be expressed. Here, OAMB reduction in heart and muscle prevented exercise-dependent adaptations in endurance, climbing speed and heart performance but did not negatively affect flight or fat body lysosomal activity. OAMB has been implicated in reward signaling [49], but more recently has been directly linked to behavior and metabolism via insulin like signaling [50] and sugar overconsumption studies [51]. Taken together, those studies and ours suggest that OAMB is modulating exercise adaptations by responding to energy needs via cell- and non-cell autonomous mechanisms. In the myocardium, OAMB may also mediate increases in heart rate during a bout of exercise itself, as adrenergic signaling increases heart rate in multiple species, including humans [26], and OA can increase larval heart rate in Drosophila [52]. It is possible that increased stimulation of heart rate may have secondary effects in other tissues by changing the rate of circulation of nutrients, hormones, or OA itself.

We find that Octβ2R and Octβ3R are important in adult Drosophila muscle for exercise adaptations to endurance and climbing speed. Both muscle-specific knockdowns also have tissue non-autonomous effects on the fat body. When we combine OA-feeding with exercise, we see a partial rescue of both climbing and endurance in Octβ3R knockdowns, an improvement that is not seen during exercise alone. This suggests that, unlike the other receptors tested here, Octβ3R activity is partially redundant and can be supplemented by other receptors if ligand dose is high enough.

Adding further complexity to mapping the specific roles of each receptor, there are several cases where tissue-specific knockdowns produce effects that are not predicted by the results of ubiquitous knockdown. For example, we find that fat-specific knockdown of Octβ3R blocks increased Lysotracker activity during exercise, which is surprising because we previously reported that ubiquitous knockdown of Octβ3R does not (8). Under wild-type conditions, Octβ3R is the primary OA receptor in adipose tissue. It could be that disruption of Octβ3R in both muscle and fat induces compensatory effects to maintain lysosomal activity in the absence of Octβ3R. Alternately, it is possible that the highly efficient and consistent knockdown provided by the ubiquitous Tub5-Gal4 induces a compensatory response that is not induced by the gradually accumulating knockdown driven by S106-Gal4 (S1 Fig). This phenomenon was not limited to Octβ3R knockdown, as muscle-specific Octβ1R knockdown unexpectedly reduced endurance more than ubiquitous Octβ1R knockdown. Whereas the role of OA in stimulating lipolysis is thought to involve signaling through PKA/cAMP, OA also regulates muscle contractility through its effects on Ca²⁺/IP3/CaMK signaling [53]. It may be that knockdowns of different strength or in different tissue combinations affect signaling through these pathways differentially. Alternately, tissue-non-autonomous effects of OA receptors may contribute to differences in these phenotypes in complex ways. Further investigation of downstream factors activated through OA-ergic signaling in different tissues during exercise and how those change during various manipulations will be necessary to resolve these questions unambiguously.

Octβ1R reduction in muscle or heart produced the most broadly deleterious phenotypes of all, reducing endurance as early as day 5, and negatively affecting all parameters of exercise adaptation. It is worth noting that Octβ1R transcript was previously found to be upregulated in both endurance-exercised and longevity selected flies, indicating its importance in the preservation of healthy physiology [5].

We have successfully mapped several specific adrenergic receptor requirements for endurance exercise adaptations in Drosophila (Fig 7). Further understanding of tissue-specific requirements for adrenergic signaling moves us closer to comprehensive mechanisms that govern exercise responses and potentially contribute to genetic differences in individual exercise responses.

Materials and methods

Fly stocks and maintenance

All fly lines were reared and aged at 25°C; 50% humidity with a 12-hour light-dark cycle and provided with a standard 10% yeast/10% sucrose diet unless otherwise indicated. All RNAi lines were validated in combination with each driver before and after exercise. All Drosophila lines were from the Bloomington Drosophila Stock Center or Vienna Drosophila RNAi Center with the following exceptions: hand GS Gal4 and MHC GS were obtained from Rolf Bodmer (Sanford Burnham Medical Research Institute) and S106 GS was obtained from Marc Tatar (Brown University). BDSC lines were w¹¹¹⁸ (BDSC3605) and OAMB RNAi (BS31171). VDRC lines were Octβ1R RNAi (v47895), Octβ2R RNAi (v8486) and Octβ3R RNAi (v101189). All driver lines have been previously characterized [54–56]. All experiments were performed using Gene-switch Gal4. For Gene-switch experiments, genetic background effects were controlled for by using RU- flies of the same background as the negative control. Raw data from all experiments throughout the manuscript is provided in S10 Fig.

Drug treatment

For gene-switch experiments, adult progeny were age-matched by collecting within 2 hours of eclosion over a 72 hour time period and immediately transferred into vials containing 5mL standard medium. Populations were split into control RU- and experimental RU+ groups on the 2^nd day and transferred into vials containing 5mL medium containing either 70% ethanol vehicle or 100 μM mifepristone (RU486) (Cayman Chemical, Ann Arbor, MI), respectively. Experimental and control flies were then housed at 25ºC on either RU486 or vehicle until experimentation.

Flies fed octopamine were treated similarly to gene-switch experiments but were collected within a single 24-hour window immediately after eclosion and housed on SY10 food containing 5 μg/mL octopamine (Sigma-Aldrich, St. Louis, MO), or an equal volume of ddH₂O vehicle.

Exercise training

Cohorts of at least 800 flies were collected under light CO₂ anesthesia within 2 hours of eclosion and separated into vials of 20. Flies were then further separated into 2 large cohorts of at least 400 flies divided into exercised and unexercised groups. If OA-feeding was employed, cohort size was doubled, and flies were divided into 4 cohorts: OA/UN, OA/EX, Vehicle/UN, Vehicle/EX. The unexercised groups were placed on the exercise training device but were prevented from running by the placement of a foam stopper low in the vial. The stopper is returned to the top of the vial at the conclusion of daily training. The exercise device drops the vials of flies every 15 seconds, inducing a repetitive negative geotaxis response. Exercised flies are free to run to the top of the vial. Daily time of exercise followed the previously described ramped program [11].

All exercised and unexercised cohorts were assessed for speed, endurance, cardiac performance, flight, and fat body Lysotracker.

Endurance

Climbing endurance was measured using the fatigue assay described previously [9]. Eight or sixteen vials of flies from each cohort were subjected to the fatigue assay at two time points: once on day 5 and once on day 25 of adulthood. For each assessment, the flies were placed on the Power Tower exercise machine and made to climb until they were fatigued. Monitored at 15 min intervals, a vial of flies was visually determined to be fatigued when 20% or fewer flies could climb higher than 1 cm after four consecutive drops. A minimum of 8 vials containing 20 flies each was used for each fatigue assessment with each vial plotted as a single datum. Each experiment was performed in duplicate or triplicate, and runspans were scored blindly when possible. The time from the start of the assay to the time of fatigue was recorded for each vial, and the data analyzed using log-rank analysis in GraphPad Prism (San Diego, CA, USA).

Climbing speed

Adult flies were collected with light CO₂ anesthesia within 2 hours of eclosion and housed in appropriate fresh food vials. Negative geotaxis was assessed in Rapid Negative Geotaxis (RING) assays in groups of 100 flies as described [9]. Flies were transferred to individual polypropylene vials in a RING apparatus and allowed to equilibrate for 1 minute. Negative geotaxis was elicited by sharply rapping the RING apparatus four times in rapid succession. The positions of the flies were captured in digital images taken 2s after stimulus. Images were analyzed using ImageJ (Bethesda, MD). The distance climbed by each fly was converted into quadrants using Microsoft Excel. The performance of each vial of 20 flies was calculated as the average of four consecutive trials to generate a single datum. Flies were longitudinally tested 5 times per week for 4–5 weeks to assess decline in negative geotaxis speed with age. Data were further consolidated into weekly performance. Between assessments, flies were returned to food vials and housed until the following RING test. Negative geotaxis results were analyzed using two-way ANOVA analysis with post hoc Tukey multiple comparison tests in GraphPad Prism (San Diego, CA, USA). All negative geotaxis experiments were performed in duplicate or triplicate.

Flight performance

Flight was analyzed as in Sujkowski et al. 2017 [8]. Duplicate or triplicate cohorts of at least 50 flies were exercise trained in narrow vials housing groups of 20 age-matched siblings. Acrylic sheeting with paintable adhesive was placed in the flight tube, and fly cohorts were ejected into the apparatus to record flight performance and subsequent landing height after release. Fly cohorts were introduced to the flight tester one vial at a time using a gravity-dependent drop tube in order to reduce variability. After a full cohort of flies was captured on the adhesive, the sheeting was removed to a white surface in order to digitally record the landing height of each fly. Flies with damaged wings were censored from final analysis to control for mechanical stress not related to training performance. Images were analyzed using ImageJ. Landing height was averaged and compared in Prism using ANOVA with Tukey post-hoc comparison.

Cardiac pacing

25-day old flies were removed from appropriate experimental cohorts and subjected to electrical pacing as in Wessells et al. [57]. The percentage of fly hearts that responded to pacing with either fibrillation or arrest were recorded as “% failure”. Pacing-induced failure rate is a marker for stress sensitivity and characteristically declines with age [31,58]. Endurance exercise reduces cardiac failure rate across ages in trained male Drosophila [7,11,31]. Failing hearts are scored as “1” and hearts that respond to pacing stress with normal beating are scored as “0”. Averages are analyzed by Chi-squared test for binary variables.

Lysotracker

Lysotracker staining of adult fat bodies was performed as in Sujkowski et al. [8]. Adult flies separated by age, genotype, and or treatment were dissected, ventral side up, in room temperature PBS. Having exposed fat bodies, partially dissected flies were rinsed 1X in fresh PBS. Lysotracker green (Molecular Probes, Eugene, OR) was diluted to 0.01μM in PBS and applied to dissected preps for 30 seconds. Samples were washed 3 times in fresh PBS. Stained fat bodies were subsequently removed and mounted in Vectashield reagent (Vector Laboratories, Burlingame, CA, USA). Confocal imaging was done in the Department of Physiology Confocal Microscopy Core at Wayne State School of Medicine on a Leica DMI 6000 with a Crest X-light spinning disc confocal using a 63X oil immersion objective or widefield fluorescent 40X objective. Images were analyzed using ImageJ. A minimum of 10 samples were analyzed for each sample and duplicate or triplicate biological cohorts were assessed for each group. Data were subjected to ANOVA with Tukey post-hoc.

Western Blotting

Triplicate biological cohorts of 3 whole flies per genotype/treatment were homogenized in boiling lysis buffer (50 mM Tris pH 6.8, 2% SDS, 10% glycerol, 100 mM dithiothreitol), sonicated for 15 seconds, boiled for 10 min, and centrifuged at 13,300 × g at room temperature for 10 min. Samples were electrophoresed on 4–20% gradient gels (Bio-Rad). Western blots were developed using the ChemiDoc system (Bio-Rad). Direct blue staining was used for total protein loading: PVDF membranes were submerged for 5 min in 0.008% Direct Blue 71 (Sigma-Aldrich) in 40% ethanol and 10% acetic acid. PVDF membranes were then rinsed briefly in 40% ethanol and 10% acetic acid solvent, then ultrapure water, air dried, and imaged using the ChemiDoc system. Anti-dAtg8a antibody (ab109364) was from obtained from Abcam. Blots were quantified using ImageLab software (Bio-Rad).

qRT PCR

RNAi efficacy was confirmed tissue-specifically pre- and post-exercise for all Gal4-UAS RNAi combinations tested (S1A–S1C Fig). To control for non-specific effects of RNAi, physiological assessments for Gal4-UAS RNAi combinations in absence of receptor expression are included as S2–S7 Figs. cDNA was prepared using a Cells to CT Kit (Invitrogen) from 20 adult fly hearts, or indirect flight muscle (IFM) or fat body from 5 adult flies. Two independent cDNA extractions were prepared for each sample. Differences between genotypes were assessed by ANOVA. Primer sequences are listed below.

5’ OAMB- CGGTTAACGCCAGCAAGTG

3’ OAMB- AAGCTGCACGAAATAGCTGC

5’Octβ1R GGCAACGAGTAACGGTTTGG

3’ Octβ1R TCATGGTAATGGTCACGGGC

5’Octβ2R TTAGTGTGCAAGTAACTGGGC

3’ Octβ2R TGAGAAGTAGACATCGAGGCTG

5’Octβ3R TGTGGTCAACAAGGCCTACG

3’ Octβ3R GTGTTCGGCGCTGTTAAGGA

5’ act5C GGCGCAGAGCAAGCGTGGTA

3’ act5C GGGTGCCACACGCAGCTCAT

Relative message abundance was determined by amplification and staining with SYBR Green I using an ABI 7300 Real Time PCR System (Applied Biosystems). Expression of Actin5c and corresponding RU- control flies were used for normalization.

Supporting information

S1 Fig [a]
Confirmation of RNAi efficacy in target tissues.

S2 Fig [tiff]
Baseline endurance is unaffected by non-specific RNAi effects.

S3 Fig [tiff]
Post-training endurance is unaffected by non-specific RNAi effects.

S4 Fig [a]
Negative RNAi controls adapt to exercise with increases in climbing speed.

S5 Fig [b]
Flight performance is increased in exercise-trained RNAi negative control flies.

S6 Fig [tiff]
No non-specific RNAi effects on post-training adaptations to cardiac stress resistance.

S7 Fig [b]
Fat body LysoTracker staining is increased exercise-trained RNAi negative control flies.

S8 Fig [a]
Vehicle-fed RNAi flies have reductions in endurance, speed, cardiac stress resistance and fat body LysoTracker staining.

S9 Fig [tiff]
Representative 40X Fat Body Lysotracker Images

S10 Fig [xlsx]
Raw Data file.

S1 Table [docx]
Summary Statistics of combinatorial treatment of 5μM OA feeding plus exercise training in selected RNAi lines.