Figures
Abstract
Circadian rhythms in animals are regulated at the level of individual cells and by systemic signaling to coordinate the activities of multiple tissues. The circadian pacemakers have several physiological outputs, including daily locomotor rhythms. Several redox-active compounds have been found to function in regulation of circadian rhythms in cells, however, how particular compounds might be involved in regulating specific animal behaviors remains largely unknown. Here the effects of hydrogen peroxide on Drosophila movement were analyzed using a recently developed three-dimensional real-time multiple fly tracking assay. Both hydrogen peroxide feeding and direct injection of hydrogen peroxide caused increased adult fly locomotor activity. Continuous treatment with hydrogen peroxide also suppressed daily locomotor rhythms. Conditional over-expression of the hydrogen peroxide-producing enzyme superoxide dismutase (SOD) also increased fly activity and altered the patterns of locomotor activity across days and weeks. The real-time fly tracking system allowed for detailed analysis of the effects of these manipulations on behavior. For example, both hydrogen peroxide feeding and SOD over-expression increased all fly motion parameters, however, hydrogen peroxide feeding caused relatively more erratic movement, whereas SOD over-expression produced relatively faster-moving flies. Taken together, the data demonstrate that hydrogen peroxide has dramatic effects on fly movement and daily locomotor rhythms, and implicate hydrogen peroxide in the normal control of these processes.
Citation: Grover D, Ford D, Brown C, Hoe N, Erdem A, Tavaré S, et al. (2009) Hydrogen Peroxide Stimulates Activity and Alters Behavior in Drosophila melanogaster. PLoS ONE 4(10): e7580. https://doi.org/10.1371/journal.pone.0007580
Editor: Mark A. Frye, UCLA, United States of America
Received: September 9, 2009; Accepted: September 28, 2009; Published: October 28, 2009
Copyright: © 2009 Grover et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: CB was supported by the Genomic Research Experience for Undergraduates program of the USC Center for Excellence in Genomic Science (NIH P50 HG002790; http://cegs.usc.edu). This work was supported by a grant from the Department of Health and Human Services to JT (AG011833; http://www.hhs.gov/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Drosophila melanogaster exhibits numerous complex behaviors, including walking, flight, grooming [1], foraging [2], fighting [3], mating [4], [5] and egg-laying [6], and most of these behaviors are under circadian control. The central circadian pacemakers in the mammalian and fly brains involve cellular feedback loops regulated at the level of protein modification and turnover, transcription and translation, and can coordinate biological rhythms throughout the animal in response to stimuli such as heat and light [7], [8]. The mechanisms for the coordination of rhythms in multiple tissues are unknown, however in mammals circulating hormones such as glucocorticoids have been implicated.
Cell autonomous oscillators have been characterized in both yeast and mammalian cells [9]. The yeast oscillator regulates the expression of both metabolism and detoxification (Phase I/II response-like) genes, and creates a metabolic cycle consisting of distinct oxidative and reductive periods. This temporal separation of potentially antagonistic biochemical pathways may optimize cell function and repair processes [10]. These results extend to metazoans, where the central pacemaker and tissue pacemakers control circadian expression of similar metabolism and detoxification gene sets, as well as additional genes such as those of the innate immune response [11], [12]. Strikingly, these same gene sets are altered during aging [13], [14] and in response to aging interventions across species [15], [16], supporting a link between circadian rhythms, metabolism/detoxification cycles and life span regulation. Consistent with this link, both aging and the oxidative stressor paraquat have been shown to alter Drosophila sleep cycles [17], [18], and the toxic effects of sleep deprivation can be ameliorated by certain heat shock proteins (hsps) [19], which are in turn induced in response to oxidative stress and aging [14], [20]–[22]. In mice, when the circadian rhythm genes Period 1 and Period 2 were simultaneously knocked-out, in addition to disrupted rhythms, the animals displayed signs of premature aging, decreased ability to repair DNA damage, and an increase in the incidence of tumors [23] – all phenotypes associated with oxidative stress.
In addition to circadian pacemakers regulating metabolism, several mechanisms have been defined through which metabolism can in turn regulate circadian rhythms [24]. For example, in mammals, the NAD(P)/NAD(P)H ratio regulates clock proteins via conserved PAS domains. PAS domain proteins can also be regulated by additional redox-active compounds, including Heme, the Heme breakdown product CO gas, as well as NO gas.
One of the major stumbling blocks for detailed analysis of the behavioral effects of genetic and pharmacological manipulations in Drosophila has been the lack of methods capable of tracking flies and quantifying their behavior. To this end, several machine vision tracking systems have been developed recently to analyze behaviors such as walking movements [25]–[28] and flight trajectories in single [29], [30] and multiple [31] Drosophila flies in 2D. We have recently developed a tracking method involving multiple video cameras that allows for detailed analysis of the movement of groups of flies through 3D space in real-time [32], along with simultaneous assay of transgenic reporter constructs expressing GFP or DsRED [22], [33]. These methods provide an ideal way to analyze the effects of chemical and transgenic manipulations on fly behavior and rhythms. Hydrogen peroxide is a good candidate for a behavior regulator, as it is the most stable and diffusible of reactive oxygen species (ROS), and has been shown to function as a cellular signaling molecule in several other processes [34]. Here hydrogen peroxide was found to stimulate fly motion parameters and to alter daily locomotor rhythms, consistent with a normal role for hydrogen peroxide in the regulation of fly movement and behavior.
Results
The average activity of groups of 25 flies was measured using the Drosophila Activity Monitor (DAM) system (Trikinetics), in which three rings of infrared beams record fly movement in a small chamber. When hydrogen peroxide was fed to flies it increased activity in a dose-dependent manner across a period of several days (Figure 1A; Supplemental Figure S1; Table 1). Concentrations greater than 0.1% hydrogen peroxide also caused significant mortality (Figure 1C). One way that cells control ROS levels is with abundant cytoplasmic catalase enzyme that converts excess hydrogen peroxide to water and oxygen. Tertiary butyl hydroperoxide (TBHP) has some similarity in chemical properties with hydrogen peroxide, but cannot be degraded by catalase [35]. When fed to flies TBHP also caused increases in activity, however it was also more toxic (data not shown). At low concentrations TBHP appeared to suppress the normal diurnal activity pattern, such that flies were more active than controls during the sleep periods, and less active than controls during the active periods (Supplemental Figure S2). Treatment of flies with the catalase enzyme inhibitor 3-amino-1,2,4-triazol (AMT) [36] was toxic without causing detectable increase in fly activity (data not shown).
(A, B) Average activity of 25 young adult male flies was measured using the DAM. Alternating light and dark boxes above the charts indicate the 12 hr-light/12 hr-dark cycle in which the flies were housed during the experiment. (A) The dose-response for flies fed hydrogen peroxide (H2O2) is presented with data expressed as activity units per fly; the control data is the same in each panel. Drug treatment began at 21 hours. (B) Oregon-R wild-type male flies in one DAM (red) compared to a duplicate vial of flies in a second DAM (blue). (C) Survival curves for flies fed increasing amounts of hydrogen peroxide, and mock-fed controls. All experiments were repeated with similar results.
One limitation of the DAM system is that it measures average activity in discrete time intervals, and offers little information regarding the specific nature of the activity. In order to better study Drosophila movement behavior and the effects of hydrogen peroxide on it, the automated video tracking system was applied [22], [32], [33]. Multiple flies were detected and tracked simultaneously using an array of calibrated and synchronized digital video cameras operating at 60 frames per second, allowing for real-time analysis of fly movement. The resulting fly trajectories are typically cylindrical (for example, see Figure 2C), corresponding to the round shape of the vials in which the flies were housed. Drosophila flies tend to favor the periphery of their container space, a behavior termed “centrophobism” [22], [25], [28].
Multiple Oregon-R wild-type male flies, 5 days old, were placed in two vials and tracked simultaneously for 48 hours. The first vial contained four flies on normal food, two of which had been pre-fed 1.0% hydrogen peroxide (H2O2) for two hours before being transferred to the vial. The second vial contained two flies and food supplemented to 1.0% hydrogen peroxide. Black/white bars indicate the light/dark cycle in which the flies were housed. Activity is expressed as distance moved (cm) per hour. (A) Two control flies (black lines, one solid and one dashed) and two flies pre-fed 1.0% H2O2 (blue lines, one solid and one dashed). B. Two control flies (black lines, one solid and one dashed) and two flies continuously fed 1.0% H2O2 (red lines, one solid and one dashed). The control fly data is the same in (A) and (B). C. The 3D tracks of individual male flies were recorded for two-minute periods at one-hour intervals beginning 9.5 hours after first exposure to control food or food adjusted to 1.0% H2O2. Statistical analysis is presented in Table 2B. D. Effect of dietary hydrogen peroxide on fly heading changes. A control fly and a fly fed continuously with 1% hydrogen peroxide were tracked for 48 hours using the video tracking assay, and average heading changes per hour are presented. The data were modeled using ARIMA(5,0,0). Control mean = 0.3396, SEM = 0.0223. 1% H2O2 mean = 1.2836, SEM = 0.1359. Z statistic = 6.8546, p<<0.001 (highly significant). All experiments were repeated with similar results (data not shown).
The real-time 3D tracking assay was used to determine how fly movement and trajectories might be affected by hydrogen peroxide treatment. The four parameters measured were speed, average speed, heading, and average heading (see Materials and methods). Average values and their corresponding standard errors for multiple time points after hydrogen peroxide administration are presented (Table 2A). Flies exhibited an increase in activity beginning on average approximately 11 minutes after first hydrogen peroxide exposure (Figure S3; Materials and methods). These changes included an approximately 9-fold increase in speed (flying and walking) over time (Table 2A). Additionally, it was observed that the flight path trajectories of treated flies were considerably more erratic, as evidenced by greater heading and average heading values (Table 2A; Figure 2D). This erratic movement is apparent in the more jagged appearance of the trajectories for hydrogen peroxide-fed flies relative to controls (Figure 2C).
Strikingly, while hydrogen peroxide increased fly activity on average, continuous administration of hydrogen peroxide was found to suppress daily locomotor rhythms. Feeding flies with 1% hydrogen peroxide for two hours caused an increase in activity, as indicated by increased distance moved per hour (Figure 2A). The increased activity returned to normal levels within a few hours, and subsequent daily locomotor rhythms were normal. However, when flies were continuously fed hydrogen peroxide, activity was initially increased, and then after about 12 hours the daily locomotor rhythms became suppressed: flies were more active than controls during the sleep periods, and less active than controls during the wake periods (Figure 2B,C). Even during the time intervals when flies fed hydrogen peroxide were less active than controls, based on distance moved per hour, their behavior was abnormal, in that movement was more erratic (greater heading changes) than a control fly moving at a comparable speed (Table 2B; Figure 2D). Because the activity measured by the DAM is a combination of distance moved and amount of heading changes, distinguishing between these parameters was only possible using the video tracking assay.
Hydrogen Peroxide Injection and SOD Over-Expression Also Cause Increased Fly Activity
As described above, feeding flies hydrogen peroxide caused an increase in activity and alterations in behavior, beginning approximately 11 minutes after first exposure. One possibility is that after the fly ingests the hydrogen peroxide, it passes into the fly's circulatory system, and then causes chemical changes in cells that regulate behavior. Alternatively, the hydrogen peroxide might cause chemical changes in the cells that line the digestive tract, and this signal is then transduced to cells that regulate behavior via some signal other than hydrogen peroxide. Finally, it is conceivable that the fly senses hydrogen peroxide in its external environment, and this information is then transmitted to cells regulating behavior. To demonstrate that the behavioral changes were due to redox alterations taking place inside the fly's tissues and cells, two additional approaches were used: injection and transgenic manipulation.
Flies were injected in the open circulatory system (hemocoel) of the abdomen with ∼0.05 ul of PBS adjusted to 1% hydrogen peroxide, or with PBS alone as a control. Hydrogen peroxide injection caused an immediate increase in fly activity that persisted for approximately 6 hours, before returning to control levels (Figure 3A). Injection of hydrogen peroxide also caused the transient induction of a transgenic reporter that is sensitive to oxidative stress [21], [33], the hsp22 gene promoter driving expression of DsRED (Figure 3B). Therefore, similar to the effect of dietary hydrogen peroxide, direct injection of hydrogen peroxide into the fly's circulatory system caused a reversible increase in fly activity.
Three male flies, 10 days old, were placed in a vial and tracked simultaneously for 24 hours. The flies were transgenic for the hsp22-DsRED reporter, strain 1MI1. Black/white bars indicate the light/dark cycle the flies were cultured under prior to the beginning of the experiment. Activity is expressed as distance moved (cm) per hour (A), and DsRED fluorescence intensity is expressed as the average pixel intensity per hour (B). Uninjected (black line), buffer injected (blue line) and 1.0% hydrogen peroxide (H2O2) injected (red line). The experiment was conducted four times with similar results, and a representative trial is presented.
In cells the superoxide dismutase (SOD) enzymes convert superoxide anions to hydrogen peroxide and oxygen. MnSOD is located exclusively in the inner mitochondrial space, whereas Cu/ZnSOD is located in the outer mitochondrial space and cytoplasm. Together these SOD enzymes are thought to be responsible for generating the majority of hydrogen peroxide in the cell [37]. A conditional transgenic system called “Tet-on” [38], [39] was used to cause tissue-general, doxycycline (DOX)-dependent over-expression of MnSOD or Cu/ZnSOD cDNAs in adult flies. Previously, over-expression of these cDNAs using the FLP-out system was shown to cause a proportional increase in enzyme activity [20], and similarly, both RNA and enzyme over-expression were confirmed for these new Tet-on MnSOD strains [16], [39] and Cu/ZnSOD strains (D. Ford and J. Tower, in preparation).
To investigate the effects of SOD over-expression on behavior and locomotor rhythms, activity was assayed using the DAM across five weeks of adult life span of control and experimental flies, in groups of 25 each. Both MnSOD and Cu/ZnSOD over-expression were found to cause a time-dependent increase in fly activity (Figure 4), consistent with a role for hydrogen peroxide. Larger average increases in activity per fly were observed upon over-expression of MnSOD than with Cu/ZnSOD (Table 3), consistent with an effect on behavior of hydrogen peroxide produced within the mitochondria.
The Tet-on transactivator insertion strain rtTA(3)E2 was crossed to Or-R wild-type control flies to generate progeny containing only the driver (A), as well as to the indicated Cu/ZnSOD and MnSOD target construct strains to generate progeny containing both constructs (B–D). Young adult males (2–4 days old) were cultured on food ±DOX at 25 flies per vial, and assayed across five weeks of adult life span using the DAM. (A–D) The details of activity profiles across weeks one and five are presented, with control fly data (-DOX) indicated in blue, and experimental fly data (+DOX) indicated in red.
To analyze the patterns of activity in flies with SOD over-expression, a seasonal-trend decomposition approach [40], as implemented in the R function stl, was applied to the average activity per fly data. This procedure decomposed the raw time-series activity data into a “seasonal” component indicating rhythms, and a “trend” component indicative of changes in overall activity. The rhythms of control flies were relatively constant across five weeks, although some gradual and subtle reductions in relative peak size and number per day were observed (Figure 5A). The activity trend of control flies (indicated by blue dashed line) dropped quickly from an initial high level, followed by several weeks of sustained activity. DOX treatment had no detectable effect on either rhythms or activity in control flies. In contrast, DOX treatment of SOD transgenic lines had dramatic effects on both the locomotor rhythms and the trend in activity in each case (Figure 5B–D; Table 3). SOD over-expression significantly increased activity, particularly in the fourth and fifth weeks. To a much lesser extent these effects were apparent even in the –DOX flies, perhaps indicating some leaky expression of the transgenes. Interestingly, while both dietary hydrogen peroxide and SOD over-expression greatly increased fly activity across all the movement parameters, there were some differences in effects. For example, hydrogen peroxide feeding produced a relatively more erratic flight path (greater heading and average heading values) while MnSOD over-expression caused relatively faster-moving flies (greater speed and average speed values) (Table 2A, 2C).
Time series of average adult activity per fly across weeks one through five is presented for each set of flies, ±DOX, as indicated. A seasonal-trend decomposition procedure decomposed the raw time-series activity data (gray) into a “seasonal” component indicating rhythms, and a “trend” component indicative of changes in overall activity (blue). This was performed in R [63] using the function stl. A linear model was fit to the trend component of the raw activity time series (red) to demonstrate the trend of increasing activity for flies over-expressing SOD. Next, an ARIMA time series modeling approach (Material and methods) was utilized for further analysis, which revealed average activity values for -DOX and +DOX flies (with standard errors in parentheses). (A) Control -DOX 39.6 (2.05), +DOX 38.6 (1.66). (B) Progeny with Cu/ZnSOD(2)21 -DOX, 41.8 (2.07), +DOX 65.1 (1.13). (C) Progeny with MnSOD(2)12,22 -DOX 40.2 (1.07), +DOX 87 (1.10). (D) Progeny with MnSOD(2)12 -DOX 39.6 (1.56), +DOX 161 (1.59).
Discussion
The real-time 3D tracking system allowed the analysis of specific effects of hydrogen peroxide and SOD over-expression on fly behavior and daily locomotor rhythms. Hydrogen peroxide was found to affect the activity, behavior and locomotor rhythms of Drosophila, suggesting a role for hydrogen peroxide in the normal regulation of these processes. Continuous dietary hydrogen peroxide stimulated several types of movement, while at the same time it tended to suppress daily locomotor rhythms. MnSOD over-expression stimulated activity in a similar way to hydrogen peroxide, causing flies to be more active and erratic. In flies with increased MnSOD expression, the patterns of daily locomotor rhythms appeared largely intact, but were significantly altered. The pattern of behavior across the day was altered in both the size and number of peaks, and in general activity was increased, particularly at later time points. In the MnSOD lines, at later time points activity was increased at all hours of the day to the extent that sleep was nearly eliminated. Surprisingly these hyperactive and sleep-deprived flies can have either reduced life span, as with line MnSOD(2)12,22 [39], or increased life span, as found for line MnSOD(2)12 [16].
Several previous observations are consistent with a link between SOD, hydrogen peroxide and behavior regulation. An activity monitor has been used to show that flies over-expressing catalase maintain activity for a longer period when challenged with toxic concentrations of hydrogen peroxide [41]. Melatonin has been identified as a circadian rhythm regulator in mammals, and has been reported to act as a scavenger of several different ROS [42], including hydrogen peroxide [43] (but see also [44]). Circadian variation in SOD enzyme activity is observed in primate erythrocytes [45]. Interestingly, genetic selection of mice for increased wheel running is associated with decreased MnSOD activity in liver tissue [46].
The fact that both dietary hydrogen peroxide and SOD over-expression had similar effects on behavior is consistent with the idea that these manipulations act through a similar mechanism, which most likely involves alterations in redox physiology in one or more critical tissues. The products of SOD enzyme are hydrogen peroxide and oxygen, however simply because SOD enzyme activity was increased it cannot be assumed that the phenotypic effects are due to increased hydrogen peroxide levels, as SOD could also conceivably act by altering superoxide levels, oxygen levels, metal scavenging or other mechanisms [47]. However, two lines of evidence point towards increased hydrogen peroxide as the relevant effector of SOD over-expression phenotypes: First is the similar effects of dietary hydrogen peroxide, injected hydrogen peroxide, and SOD over-expression on fly behaviors, as presented here; second is recent data from our research group indicating that SOD over-expression causes changes in gene expression that are highly similar to hydrogen peroxide feeding, including the preferential induction of genes bearing hydrogen peroxide response elements in their promoters ([16] and unpublished data).
One direct way to test this model would be to measure hydrogen peroxide levels in the cells of the SOD over-expressing flies. However, using simple spectrofluorometric assays we were unable to detect reproducible changes in hydrogen peroxide levels in whole-fly extracts, or from mitochondria isolated from SOD over-expressing flies (data not shown). This result is perhaps not surprising as it might be expected that changes in hydrogen peroxide levels sufficient to mediate signaling could be quite small and limited to within a narrow physiological range. Moreover, altered hydrogen peroxide signaling might occur only in a subset of fly cells, and might be limited to localized and transient changes within the relevant cells, and thus would not be apparent in whole-fly extracts. One possible approach for the future might be the use of redox-sensitive fluorescent proteins [48]–[52] as reporters in the control and SOD-over-expressing flies; however since so far no chemical or transgenic reporter has been shown to be completely specific for hydrogen peroxide this will not be a simple undertaking. Strikingly, it has recently been reported that mitochondria isolated from Drosophila heads show a circadian variation in the production of hydrogen peroxide [53], and in zebrafish larvae hydrogen peroxide has been shown to act as an systemic signaling molecule that recruits immune cells to sites of tissue damage [52]. These results are consistent with the idea that hydrogen peroxide may act as both an intracellular and intercellular signaling molecule.
Daily locomotor activity in Drosophila is regulated by the central circadian pacemaker, and it is possible that the effects of dietary hydrogen peroxide and SOD over-expression were caused by alterations in the circadian machinery. Several studies have demonstrated that the conserved circadian oscillator is regulated by cellular metabolism [9]. In particular, redox-active signaling molecules such as NADPH, CO, NO and heme have been shown to regulate circadian transcription factors such as Clock, Cycle and BMAL1 through conserved PAS protein domains [54], [55]. Since the mitochondria are a key regulator of metabolism and a primary source of cellular ROS, the data suggest that a retrograde redox signal from the mitochondria may normally be involved in regulating cellular oscillators and circadian rhythms. Hydrogen peroxide has not previously been directly implicated in these processes, however it is interesting to note that exogenous hydrogen peroxide has been shown to be capable of advancing the period of the yeast metabolic cycle [10]. The fact that MnSOD is a mitochondrial enzyme, that the mitochondria are the primary cellular source of hydrogen peroxide, and that dietary hydrogen peroxide and MnSOD over-expression both altered a direct output of circadian rhythms in flies (daily locomotor activity), suggests that hydrogen peroxide may be an additional retrograde signal affecting circadian rhythms. Aging is associated with increased oxidative stress and a deterioration of behaviors and circadian rhythms in both flies and humans. One possible model is that the oxidative stress associated with aging results in part from a breakdown in normal mitochondrial-nuclear signaling pathways involving periodic (circadian) variations in redox signaling molecules, including hydrogen peroxide [34], [56], [57].
Materials and Methods
Drosophila Strains and Culture
The wild-type Oregon-R Drosophila melanogaster strain was obtained from Bloomington Drosophila stock center. The flies were maintained on a standard cornmeal, yeast, dextrose, and agar medium at 25°C [39]. Age-synchronized cohorts of flies were collected over 48 hours from culture bottles or vials, and drug-treatment experiments used young (4–6 day old) Oregon-R male flies while experiments with transgenic flies used young (2–4 day old) male flies of the indicated genotypes. All reagents were purchased from Sigma-Aldrich. The Tet-on MnSOD and Cu/ZnSOD expression constructs were created by cloning the corresponding cDNAs [58] into the USC1.0 vector [59] and generating multiple independent transformant strains. Line MnSOD(2)12 contains a single insert, while lines MnSOD(2)12,22 and Cu/ZnSOD(2)21 contain double inserts. These strains are additionally described and characterized elsewhere [16], [39] (D. Ford and J. Tower, in preparation). The rtTA driver construct consists of the rtTA transcription factor coding region downstream of the actin5C promoter, and the rtTA(3)E2 insertion supports high-level, DOX dependent expression of target constructs in all tissues except the germ line [38]. For experiments involving SOD over-expression, flies of the desired genotypes were generated by crossing strains homozygous for the indicated Cu/ZnSOD and MnSOD transgene insertions to a strain containing the rtTA(3)E2 driver to obtain progeny containing each construct. Control flies were generated by crossing the rtTA(3)E2 driver strain to Oregon-R wild type flies to generate progeny containing only rtTA(3)E2.
Drug Treatments
For the Drosophila Activity Monitor (DAM) experiments lasting 4 or 5 days (Figures 1, 2 and 4), control groups of flies received 1% sucrose in pure de-ionized H2O, while experimental groups received 1% sucrose solution plus the indicated concentration of H2O2 (0.1%, 0.5%, 1.0%, 2.0%), tertiary butyl hydroperoxide (TBHP) (0.05%, 0.1%), or 3-amino-1,2,4-triazol (AMT) (0.07%). For each solution 1.0mL was separately added to a Drosophila culture vial containing a single Kimwipe (Kimberly Clark) packed tightly at the bottom (∼10 mm thick layer) to absorb the solution. Professional standard polystyrene Drosophila culture vials (25×95 mm; Genesee Scientific) were used in all experiments. For the three-camera tracking assays, vials containing standard fly media stained with blue food coloring (Kroger brand) were used, as the blue colored background was required for efficient tracking. The food in the vial was adjusted to 1.0% H2O2 using a 30% H2O2 solution. The experiment began after the solution was given twenty-four hours to absorb into the media. For DAM experiments lasting 5 weeks (Figure 4), vials containing Drosophila media were adjusted to 64 µg/ml doxycycline (DOX) plus 64 µg/ml ampicillin, while control vials were adjusted 64 µg/ml ampicillin alone. Chemical adjustment of food was done by applying 100 µl of a 10X stock solution to the surface of food vials, which penetrates 1 ml of food after 48 hours incubation based on colored dye absorption. The ampicillin was present to help prevent any bacterial growth in the vials.
Injection of Hydrogen Peroxide into Drosophila
Flies were injected with 1% hydrogen peroxide in PBS, and with PBS alone as a control. Green food coloring (Kroger brand) was added to the solutions to aid in liquid handling and scoring of injections. Microneedles were made using the PN-30 puller (Narishige) and Brosil glass capillary tubing (1.0 mm OD×0.75 mm ID; FHC Inc.). The needles were graduated before use with a scale of 1/32 inch, and ∼3 ul of solution was added to the needle. The needles with solution were then assembled into the FemtoJet express microinjector (Eppendorf). The flies were anaesthetized using CO2 and positioned on the pad with the abdomen oriented towards the needle using brushes. The colored solution was then injected into the abdomen of the adult fly using the microinjector; the volume injected was ∼0.05 ul per fly, based on the gradation markings on the needle.
Drosophila Activity Monitor Assay
After the flies regained consciousness, the vials were placed into separate TriKinetics® Drosophila Population Monitors (DAM). These devices are each equipped with 3 rings that constantly emit an infrared beam through the vial. The interruptions of the beams, or activity units, caused by a fly passing through were then tallied in every hour by the DAM system. For the 4 day experiments (Figures 1, 2), flies were initially placed in control vials, and then at 21 hours the flies were transferred to control vials or vials adjusted to the indicated H2O2 or TBHP concentrations. For the 5-week experiments (Figure 4), flies were placed directly into food vials ±DOX. The vials were then monitored for the remainder of the experiments, with transfer to fresh vials every other day. The experiments were conducted in an incubator on a 12 h/12 h light/dark cycle at 25°C. A building power outage caused a several-hour gap in the data collection in one experiment (Figure 4). To allow further analysis of that time-course, the missing data were imputed using the theory of state space models with missing observations [60].
Time Series Analysis of DAM Data
Seasonal time series like the fly activity data measured by the DAM consist of trend and periodic components. There are several approaches to analyzing such data [61]. The first is to extract the periodic and trend components of the data and analyze the residual. The second is to model the data as a whole. Due to the non-stationary nature of the fly activity data, the second technique was used, and an ARIMA (p, d, q)s process was employed. The acronym ARIMA stands for “Auto-Regressive Integrated Moving Average”, where p refers to the autoregressive, d the integrated, and q the moving average parts of the time series model. In [62] methods for choosing a suitable ARIMA model are discussed. A suitable ARIMA model can be selected using Akaike's Information Criterion (AIC), the model with the smallest AIC being chosen. Fitting can be performed using the function arima in the package stats in R [63]. The model used here included both differencing and a seasonal AR term of the form ARIMA ((1, 0, 1)x(1, 1, 1)24). The fit of the model was verified via both residual analysis, and the auto-correlation function of the residuals and the portmanteau lack-of-fit test (Ljung-Box test) [64]. Once the model was fitted to the data, statistics such as means and standard errors of the means were calculated.
Multiple Camera 3D Tracking and DsRED Fluorescence Assay
Tracking of multiple fly locomotor activity was conducted using previously published methods [22], [32], [33]; detailed protocols are available for download from the laboratory website (http://towerlab.USC.edu/). All tracking experiments were done with young male flies (4-6 days old) and were initiated at the same time of day (4PM). Flies were placed in standard 25×75 mm polyethelene culture vials with food at the bottom, and stoppered with cotton at the top. The food was colored blue (Kroger brand food color) to facilitate tracking. The vials were placed in the center of a circular camera rig, 70 cm in diameter. Multiple calibrated and synchronized Flea digital cameras (Point Grey) were mounted on the camera rig, facing downward at a distance of 15 cm from the vials. Experiments involving only visible light tracking used three cameras [32], whereas all other experiments utilized six cameras [33]. Each camera was fitted with a 8 mm megapixel fixed focal lens (Edmund Optics). Tracking of gene expression using DsRED reporter transgenes in multiple flies was accomplished using published methods [33]. Briefly, the excitation light source was a 5W Luxeon V star 550 nm endura bright green lambertian LED (Optotech, Cat # OT16-5100-G). The LED was powered with a xitanium 700 mA LED driver (Optotech, AC converter Cat # OTMI-0060). 585 nm barrier filters (Edmund Optics, Cat # NT39-417) were placed between the sensors and the lenses of three cameras to detect DsRED expression. The fluorescence tracking assays were conducted in a dark room where the only source of illumination was the green LED. To track flies in alternating light/dark cycles, the light period was generated with a Luxeon III star white lambertian LED, and the dark period was illuminated using only a 5W Luxeon III star 630 nm endura red lambertian LED (Optotech, Cat # OT16-5100-R); the Drosophila photoreceptor pigments are not responsive to red light [65]. Prior to the tracking assays, flies were cultured to the indicated age by transfer to new food every other day, under a 12 hr/12 hr light/dark cycle. Activity is plotted as distance moved (cm) per hour.
Fly Movement Parameters
To allow comparison of the effects of dietary hydrogen peroxide and SOD over-expression on fly movement behavior, the data from the tracking assay was used to calculate the following four parameters [66]:
Average speed over an interval (cm/sec).
This is the average distance traveled by a fly over an interval of six seconds. This number was calculated by taking an interval width of±3 seconds around the current time instant.
Directional heading change (radians/sec). This is the angle between the two tangent vectors of the fly position in the 3D trajectory over a period of one second or 60 frames. The angle was calculated by taking the inverse cosine of the dot product of the two normalized vectors.
Directional heading change over an interval (radians/sec). This is the average directional heading change over an interval of six seconds. This number, like that of the Average Speed, was over an interval size of ±3 seconds around the current time instant. For the 48-hour plot of heading changes (Figure 2D), the value plotted is the change in heading of the fly in every second, averaged per hour.
Measuring Time to Increased Activity
To determine the amount of time required for H2O2 to have an effect on the fly, five young (4–6 day old) adult male Oregon-R flies were placed in individual vials with food adjusted to 1.0% H2O2, and their activity was tracked for ∼30 minutes. We estimated the time from the first contact of each fly with the food until an increase in its activity was detected, where activity was measured as distance traveled in cm per hundredth of a minute. To do this it was assumed that each activity series had a changepoint at an unknown time, t. The series before and after time t were modeled as ARIMA(p,d,q) time series with (for simplicity) common values of p, d and q, but their own parameters. For each value of t the AIC from the two segments was found, and these values were plotted as a function of t; the value of t giving the smallest AIC was chosen as the breakpoint. This approach is essentially a simplified version of the method of Davis and coworkers, in which an a priori unknown number of breakpoints is estimated [67]. We chose ARMA(5,0,0) models, although the results were little influenced by other models we tested (data not shown). Supplemental Figure S3 shows the original series and the fitted means in each segment. The estimated breakpoints were 5.8, 7.98, 10.13 and 11.37 and 20.85 minutes (average = 11.23 minutes, SE 5.78 minutes). Each of the fitted models was compared to one with no breakpoint; the models with the breakpoint showed highly reduced AIC values.
Supporting Information
Figure S1.
Change in fly activity due to dietary hydrogen peroxide. Time series of activity data from the hydrogen peroxide (H2O2) dose-response in Figure 1A is presented. A seasonal-trend decomposition procedure decomposed the raw time-series activity data (gray) into a seasonal component indicating rhythms, and a trend component indicative of changes in overall activity (blue). This was performed in R using the function stl. A linear model was fit to the trend component of the raw activity time series (red) to demonstrate the trend of increasing activity for flies fed H2O2. Next, an ARIMA time series modeling approach (Material and methods) was utilized for further analysis which revealed average activity (with standard errors in parentheses) as follows: 0.1% H2O2 36.2 (1.67); 0.5% H2O2 43.7 (2.30); 1.0% H2O2 71.1 (1.61); and 2.0% H2O2 78.8 (2.41); Control 20.5 (1.20).
https://doi.org/10.1371/journal.pone.0007580.s001
(3.37 MB TIF)
Figure S2.
Effect of dietary TBHP on adult fly activity. (A) Oregon-R wild-type male flies fed 0.05% TBHP (red) or 0.1% TBHP (blue) and mock-fed controls (black), with data expressed as activity units per fly. For the TBHP experiments triplicate vials of 25 flies each were assayed and averaged for each condition. Drug treatment began at 21 hours. (B) The data are the same as in (A), and here are plotted to show difference in average activity per fly between experimental conditions and control.
https://doi.org/10.1371/journal.pone.0007580.s002
(1.23 MB TIF)
Figure S3.
Analysis of activity breakpoints. Activity time series for five young (4–6 days old) adult male Oregon-R flies from a tracking experiment to determine the amount of time required for H2O2 to have an effect on the fly. Flies were placed in individual vials with food adjusted to 1.0% H2O2, and their activity was tracked for approximately 30 minutes. Each panel corresponds to a separate fly; the red lines show the mean activity (distance traveled (cm per hundredth of a minute)) prior to and after the increase in activity.
https://doi.org/10.1371/journal.pone.0007580.s003
(5.55 MB TIF)
Acknowledgments
We thank Isaac Cohen for help with the tracking algorithm, Richard Davis for helpful discussions on time series analysis, and Michelle Arbeitman for comments on the manuscript draft.
Author Contributions
Conceived and designed the experiments: JGT. Performed the experiments: DG DF CB AE JGT. Analyzed the data: DG DF ST JGT. Contributed reagents/materials/analysis tools: NH JGT. Wrote the paper: DG ST JGT.
References
- 1. Szebenyi AL (1969) Cleaning behavior in Drosophila melanogaster. Animal Behaviour 17: 641–651.
- 2. Meunier N, Belgacem YH, Martin JR (2007) Regulation of feeding behaviour and locomotor activity by takeout in Drosophila. J Exp Biol 210: 1424–1434.
- 3. Vrontou E, Nilsen SP, Demir E, Kravitz EA, Dickson BJ (2006) fruitless regulates aggression and dominance in Drosophila. Nat Neurosci 9: 1469–1471.
- 4. Kadener S, Villella A, Kula E, Palm K, Pyza E, et al. (2006) Neurotoxic protein expression reveals connections between the circadian clock and mating behavior in Drosophila. Proc Natl Acad Sci U S A 103: 13537–13542.
- 5. Nishinokubi I, Shimoda M, Ishida N (2006) Mating rhythms of Drosophila: rescue of tim01 mutants by D. ananassae timeless. J Circadian Rhythms 4: 4.
- 6. Howlader G, Sharma VK (2006) Circadian regulation of egg-laying behavior in fruit flies Drosophila melanogaster. J Insect Physiol 52: 779–785.
- 7. Schibler U, Naef F (2005) Cellular oscillators: rhythmic gene expression and metabolism. Curr Opin Cell Biol 17: 223–229.
- 8. Boothroyd CE, Wijnen H, Naef F, Saez L, Young MW (2007) Integration of light and temperature in the regulation of circadian gene expression in Drosophila. PLoS Genet 3: e54.
- 9. Wijnen H, Young MW (2006) Interplay of circadian clocks and metabolic rhythms. Annu Rev Genet 40: 409–448.
- 10. Chen Z, Odstrcil EA, Tu BP, McKnight SL (2007) Restriction of DNA replication to the reductive phase of the metabolic cycle protects genome integrity. Science 316: 1916–1919.
- 11. King-Jones K, Horner MA, Lam G, Thummel CS (2006) The DHR96 nuclear receptor regulates xenobiotic responses in Drosophila. Cell Metab 4: 37–48.
- 12. Green CB, Takahashi JS (2006) Xenobiotic metabolism in the fourth dimension: PARtners in time. Cell Metab 4: 3–4.
- 13. Pletcher SD, Macdonald SJ, Marguerie R, Certa U, Stearns SC, et al. (2002) Genome-wide transcript profiles in aging and calorically restricted Drosophila melanogaster. Current Biology 12: 712–723.
- 14. Landis GN, Abdueva D, Skvortsov D, Yang J, Rabin BE, et al. (2004) Similar gene expression patterns characterize aging and oxidative stress in Drosophila melanogaster. Proc Natl Acad Sci U S A 101: 7663–7668.
- 15. Gems D, McElwee JJ (2005) Broad spectrum detoxification: the major longevity assurance process regulated by insulin/IGF-1 signaling? Mech Ageing Dev 126: 381–387.
- 16. Curtis C, Landis GN, Folk D, Wehr NB, Hoe N, et al. (2007) Transcriptional profiling of MnSOD-mediated lifespan extension in Drosophila reveals a species-general network of aging and metabolic genes. Genome Biol 8: R262.
- 17. Driver C, Georgiou A, Georgiou G (2004) The contribution by mitochondrially induced oxidative damage to aging in Drosophila melanogaster. Biogerontology 5: 185–192.
- 18. Koh K, Evans JM, Hendricks JC, Sehgal A (2006) A Drosophila model for age-associated changes in sleep:wake cycles. Proc Natl Acad Sci U S A 103: 13843–13847.
- 19. Shaw PJ, Tononi G, Greenspan RJ, Robinson DF (2002) Stress response genes protect against lethal effects of sleep deprivation in Drosophila. Nature 417: 287–291.
- 20. Landis GN, Tower J (2005) Superoxide dismutase evolution and life span regulation. Mech Ageing Dev 126: 365–379.
- 21. Yang J, Tower J (2009) Expression of hsp22 and hsp70 transgenes is partially predictive of Drosophila survival under normal and stress conditions. J Gerontol A Biol Sci Med Sci 64: 828–838.
- 22. Grover D, Yang J, Tavare S, Tower J (2008) Simultaneous tracking of fly movement and gene expression using GFP. BMC Biotechnol 8: 93.
- 23. Lee CC (2006) Tumor suppression by the mammalian Period genes. Cancer Causes Control 17: 525–530.
- 24. Tu BP, McKnight SL (2006) Metabolic cycles as an underlying basis of biological oscillations. Nat Rev Mol Cell Biol 7: 696–701.
- 25. Besson M, Martin JR (2005) Centrophobism/thigmotaxis, a new role for the mushroom bodies in Drosophila. J Neurobiol 62: 386–396.
- 26. Williams AE, Rose MR, Bradley TJ (2004) The respiratory pattern in Drosophila melanogaster selected for desiccation resistance is not associated with the observed evolution of decreased locomotory activity. Physiol Biochem Zool 77: 10–17.
- 27. Martin JR (2004) A portrait of locomotor behaviour in Drosophila determined by a video-tracking paradigm. Behav Processes 67: 207–219.
- 28. Valente D, Golani I, Mitra PP (2007) Analysis of the trajectory of Drosophila melanogaster in a circular open field arena. PLoS ONE 2: e1083.
- 29. Budick SA, Dickinson MH (2006) Free-flight responses of Drosophila melanogaster to attractive odors. J Exp Biol 209: 3001–3017.
- 30. Tammero LF, Dickinson MH (2002) The influence of visual landscape on the free flight behavior of the fruit fly Drosophila melanogaster. J Exp Biol 205: 327–343.
- 31. Branson K, Robie AA, Bender J, Perona P, Dickinson MH (2009) High-throughput ethomics in large groups of Drosophila. Nat Methods 6: 451–457.
- 32. Grover D, Tower J, Tavare S (2008) O fly, where art thou? J R Soc Interface 5: 1181–1191.
- 33. Grover D, Yang J, Ford D, Tavare S, Tower J (2009) Simultaneous tracking of movement and gene expression in multiple Drosophila melanogaster flies using GFP and DsRED fluorescent reporter transgenes. BMC Res Notes 2: 58.
- 34. Balaban RS, Nemoto S, Finkel T (2005) Mitochondria, oxidants, and aging. Cell 120: 483–495.
- 35. Chance B, Sies H, Boveris A (1979) Hydroperoxide metabolism in mammalian organs. Physiol Rev 59: 527–605.
- 36. Margoliash E, Novogrodsky A (1958) A study of the inhibition of catalase by 3-amino-1:2:4:-triazole. Biochem J 68: 468–475.
- 37. Fridovich I (1995) Superoxide radical and superoxide dismutases. Annu Rev Biochem 64: 97–112.
- 38. Bieschke ET, Wheeler JC, Tower J (1998) Doxycycline-induced transgene expression during Drosophila development and aging. Mol Gen Genet 258: 571–579.
- 39. Ford D, Hoe N, Landis GN, Tozer K, Luu A, et al. (2007) Alteration of Drosophila life span using conditional, tissue-specific expression of transgenes triggered by doxycycline or RU486/Mifepristone. Exp Gerontol 42: 483–497.
- 40. Cleveland RB, Cleveland WS, McRae JE, Terpenning I (1990) STL: A seasonal-trend decomposition procedure based on Loess. Journal of Official Statistics 6: 3–73.
- 41. Kaneuchi T, Togawa T, Matsuo T, Fuyama Y, Aigaki T (2003) Efficient measurement of H(2)O(2) resistance in Drosophila using an activity monitor. Biogerontology 4: 157–165.
- 42. Allegra M, Reiter RJ, Tan DX, Gentile C, Tesoriere L, et al. (2003) The chemistry of melatonin's interaction with reactive species. J Pineal Res 34: 1–10.
- 43. Tan DX, Reiter RJ, Manchester LC, Yan MT, El-Sawi M, et al. (2002) Chemical and physical properties and potential mechanisms: melatonin as a broad spectrum antioxidant and free radical scavenger. Curr Top Med Chem 2: 181–197.
- 44. Grzelak A, Macierzynska E, Bartosz G (2004) Melatonin does not react rapidly with hydrogen peroxide. Free Radic Res 38: 1155–1158.
- 45. Goncharova ND, Shmaliy AV, Bogatyrenko TN, Koltover VK (2006) Correlation between activity of antioxidant enzymes and circadian rhythms of corticosteroids in Macaca mulatta monkeys of different age. Exp Gerontol 41: 778–783.
- 46. Thomson SL, Garland T Jr, Swallow JG, Carter PA (2002) Response of Sod-2 enzyme activity to selection for high voluntary wheel running. Heredity 88: 52–61.
- 47. Liochev SI, Fridovich I (2007) The effects of superoxide dismutase on H2O2 formation. Free Radic Biol Med 42: 1465–1469.
- 48. Hanson GT, Aggeler R, Oglesbee D, Cannon M, Capaldi RA, et al. (2004) Investigating mitochondrial redox potential with redox-sensitive green fluorescent protein indicators. J Biol Chem 279: 13044–13053.
- 49. Dooley CT, Dore TM, Hanson GT, Jackson WC, Remington SJ, et al. (2004) Imaging dynamic redox changes in mammalian cells with green fluorescent protein indicators. J Biol Chem 279: 22284–22293.
- 50. Bjornberg O, Ostergaard H, Winther JR (2006) Measuring intracellular redox conditions using GFP-based sensors. Antioxid Redox Signal 8: 354–361.
- 51. Meyer AJ, Brach T, Marty L, Kreye S, Rouhier N, et al. (2007) Redox-sensitive GFP in Arabidopsis thaliana is a quantitative biosensor for the redox potential of the cellular glutathione redox buffer. Plant J 52: 973–86.
- 52. Niethammer P, Grabher C, Look AT, Mitchison TJ (2009) A tissue-scale gradient of hydrogen peroxide mediates rapid wound detection in zebrafish. Nature 459: 996–999.
- 53. Krishnan N, Davis AJ, Giebultowicz JM (2008) Circadian regulation of response to oxidative stress in Drosophila melanogaster. Biochem Biophys Res Commun 374: 299–303.
- 54. Rutter J, Reick M, McKnight SL (2002) Metabolism and the control of circadian rhythms. Annu Rev Biochem 71: 307–331.
- 55. Rutter J, Reick M, Wu LC, McKnight SL (2001) Regulation of clock and NPAS2 DNA binding by the redox state of NAD cofactors. Science 293: 510–514.
- 56. Fridovich I (2004) Mitochondria: are they the seat of senescence? Aging Cell 3: 13–16.
- 57. Kondratov RV (2007) A role of the circadian system and circadian proteins in aging. Ageing Res Rev 6: 12–27.
- 58. Sun J, Folk D, Bradley TJ, Tower J (2002) Induced overexpression of mitochondrial Mn-superoxide dismutase extends the life span of adult Drosophila melanogaster. Genetics 161: 661–672.
- 59. Allikian MJ, Deckert-Cruz D, Rose MR, Landis GN, Tower J (2002) Doxycycline-induced expression of sense and inverted-repeat constructs modulates phosphogluconate mutase (Pgm) gene expression in adult Drosophila melanogaster. Genome Biol 3: research0021.
- 60.
Brockwell PJ, Davis RA (1993) Time Series: Theory and Methods.: Springer-Verlag New York, Inc.
- 61. Levine JD, Funes P, Dowse HB, Hall JC (2002) Signal analysis of behavioral and molecular cycles. BMC Neurosci 3: 1.
- 62. de Gooijer JG, Abraham B, Gould A, Robinson L (1985) Methods for determining the order of an autoregressive-moving average process: A survey. International Statistical Review 53: 301–329.
- 63.
RDevelopmentCoreTeam (2006) R: A language and environment for statistical computing.: R Foundation for Statistical Computing, Vienna, Austria.
- 64. Trapletti A, Hornik K (2007) tseries: Time Series Analysis and Computational Finance. R package version 010-11:
- 65. Montell C (1999) Visual transduction in Drosophila. Annu Rev Cell Dev Biol 15: 231–268.
- 66. Feldman A, Balch T (2004) Representing honey bee behavior for recognition using human trainable models. Adaptive Behavior - Animals, Animats, Software Agents, Robots, Adaptive Systems 12: 241–250.
- 67. Davis RA, Lee TCM, Rodriguez-Yam GA (2006) Structural break estimation for nonstationary time series models. Journal of the American Statistical Association 101: 223–239.