Reduced branched-chain aminotransferase activity alleviates metabolic vulnerability caused by dim light exposure at night in Drosophila

Abstract The rhythmic pattern of biological processes controlled by light over 24 h is termed the circadian rhythm. Disturbance of circadian rhythm due to exposure to light at night (LAN) disrupts the sleep-wake cycle and can promote cardiovascular disease, diabetes, cancer, and metabolic disorders in humans. We studied how dim LAN affects the circadian rhythm and metabolism using male Drosophila. Wild-type flies exposed to the dim light of 10 lux at night displayed altered 24 h sleep-wake behavior and expression patterns of circadian rhythm genes. In addition, the flies became more vulnerable to metabolic stress, such as starvation. Whole-body metabolite analysis revealed decreased amounts of branched-chain amino acids (BCAAs), such as isoleucine and valine. The dim light exposure also increased the expression of branched-chain amino acid aminotransferase (BCAT) and branched-chain α-keto acid dehydrogenase (BCKDC) enzyme complexes that regulate the metabolism of BCAAs. Flies with the Bcat heterozygous mutation were not vulnerable to starvation stress, even when exposed to dim LAN, and hemolymph BCAA levels did not decrease in these flies. Furthermore, the vulnerability to starvation stress was also suppressed when the Bcat expression level was reduced in the whole body, neurons, or fat body during adulthood using conditional GAL4 and RNA interference. Finally, the metabolic vulnerability was reversed when BCAAs were fed to wild-type flies exposed to LAN. Thus, short-term dim light exposure at night affects the expression of circadian genes and BCAA metabolism in Drosophila, implying a novel function of BCAAs in suppressing metabolic stress caused by disrupted circadian rhythm.


Introduction
A circadian rhythm refers to a repetitive physiological cycle with a periodicity of 24 h. A variety of physiological phenomena display circadian rhythms, including sleep, liver metabolism, hormone secretion, blood pressure, and body temperature (Aoyama & Shibata, 2020). Light, as the primary entrainment signal (Zeitgeber), regulates the circadian rhythm and induces changes in various physiological events depending on the exposure time and intensity. Contemporary people are exposed to extended light sources at night. Such a condition disrupts the rhythmic expression of the circadian clock genes (Fonken & Nelson, 2014) and the secretory patterns of hormones such as melatonin and cortisol (Gooley et al., 2011;LeGates et al., 2014;Rahman et al., 2019).
Disturbance of the circadian rhythm due to nocturnal light exposure is associated with cancer and metabolic diseases including diabetes and obesity (Huang et al., 2011). Disturbed circadian clocks affect carbohydrate or lipid metabolism. For example, oscillation of triacylglyceride (TAG) and protein concentrations disappeared in timeless (tim) mutants (Seay & Thummel, 2011). The Clock (Clk) gene maintains blood glucose levels by regulating glycogen and lipid metabolism (Doi et al., 2010;Gooley, 2016). In Drosophila, inhibition of Clk expression in PDF neurons that control circadian rhythm increases the accumulation of TAG in the body (DiAngelo et al., 2011). In tim mutants, the circadian rhythm of trehalose and glycogen levels disappeared, and TAG levels in the hemolymph decreased (Seay & Thummel, 2011). Further research is needed to elucidate the effect of circadian rhythm on the metabolism of amino acids and proteins.
Even exposure to light with low illuminance at night disrupts the circadian rhythm and metabolism. In fruit flies, dim light, such as moonlight (0.6-0.8 lux), also disturbs the circadian rhythm (Bachleitner et al., 2007;Kempinger et al., 2009). In mice exposed to 5 lux light at night (LAN), expression levels of the period circadial regulator 2 (PER2) and cryptochrome circadian regulator 1 (CRY1) genes reportedly decrease in white adipose tissue and liver, and period circadian regulator 1 (PER1), PER2, and CRY1 expression levels decrease in the hypothalamus (Fonken et al., 2013). In addition, the weekly eating behavior pattern changes, resulting in increased body weight, fat mass, and glucose tolerance (Fonken et al., 2010). Exposure to 2 lux LAN was shown to promote hepatic TAG accumulation and increase glucose and fatty acid uptake (Okuliarova et al., 2020). In humans, body weight and basal metabolic index were significantly higher in those exposed to 3 lux LAN, with increased TAG, low-density lipoprotein, and cholesterol levels (Fonken & Nelson, 2014). Furthermore, patients with metabolic diseases appeared more sensitive to dim LAN and suffered from disrupted circadian rhythm when exposed to 5 lux LAN (Bedrosian & Nelson, 2017). Most humans maintain a normal daily lifecycle despite circadian rhythm and metabolic disturbances that may occur due to frequent exposure to dim LAN. Thus, unknown mechanisms that maintain circadian rhythm and metabolic homeostasis in response to these stimuli may exist.
Branched-chain amino acids (BCAAs), including leucine (Leu), isoleucine (Ile), and valine (Val), are essential amino acids. In cells, BCAAs are transaminated by branched-chain amino acid aminotransferase (BCAT) and decarboxylated by branched-chain a-keto acid dehydrogenase enzyme complex (BCKDC). BCAAs are finally converted to acetyl-CoA or succinyl-CoA to generate ATP in the tricarboxylic acid (TCA) cycle (Nie et al., 2018). Feeding BCAAs to middleaged mice increases mitochondrial biosynthesis in cardiac and skeletal muscles. It improves the healthy lifespan by activating the mammalian target of rapamycin (mTOR) and sirtuin 1 (SIRT1) signaling and reactive oxygen species (ROS) defense systems, such as superoxide dismutase 1 (SOD1) and glutathione peroxidase 1 (GPx1) (D'Antona et al., 2010). In addition, BCAAs are effectively metabolized in immune cells and play a role in immune cell growth, proliferation, and activation (Monirujjaman & Ferdouse, 2014). However, side effects occur when excess BCAAs are accumulated. For example, maple syrup urine disease (MSUD) can develop when BCAA levels in the blood are abnormally high due to decreased activity of BCKDC. Patients with MSUD are more sensitive to oxidative stress and have impaired brain function, showing increased levels of BCAAs, ketonic acids, and free radicals in the blood (Sitta et al., 2014;Strauss et al., 2020). The suppression of BCAT activity by overexpressing miR-277 in Drosophila leads to mTOR activation, BCAA accumulation, and a shortened lifespan (Esslinger et al., 2013).
In this study, we observed that circadian disturbances induced by short-term nocturnal exposure to 10 lux light increased the vulnerability to metabolic stress in Drosophila, which was associated with decreased levels of BCAAs. Reducing the Bcat expression level to inhibit the metabolism of BCAAs or feeding BCAAs to maintain the blood concentration of BCAAs reduced the vulnerability to metabolic stress induced by dim light exposure at night. These results demonstrate the role of BCAA metabolism in relieving metabolic stress caused by mild disturbances of circadian rhythm.

Methods and materials Drosophila strains
Wild-type male flies (w CS10 ) were cultured in a standard medium under conditions of 12:12 h light (L)/dark (D) cycle, temperature of 25 C, and 40-60% relative humidity. Actingene switch (GS)-GAL4, elav-GS-GAL4, S106-GS-GAL4, Bcat RNA interference (RNAi), Oregon-R, and the circadian mutants including tim, Clk, and cry were obtained from the Bloomington Drosophila Stock Center at Indiana University (Bloomington, IN, USA). Oregon-R wild-type flies were used as controls. The transgenic flies were fed 150 mM of RU486 (RU, mifepristone; Sigma-Aldrich, St. Louis, MO, USA) to activate GS-GAL4 drivers after eclosion.

Generation of Bcat mutant flies
Target sequence selection The genomic target sequence for the Bcat gene was designed using the flyCRISPR optimal Target Finder tool (www.flycrispr.org). The sequence was 20 nucleotides in length with a three-nucleotide protospacer adjacent motif (PAM) NGG sequence driven by the U6 promoter. The Cas9 nuclease was used to cut the three-nucleotide region upstream of the PAM site. The 12-nucleotide region upstream of the PAM site (i.e. the seed sequence) is the most critical determinant of cleavage specificity.

Cloning
Oligo primers were designed for the sense strand (5 0 -CTTC) and anti-sense strand (CAAA-5 0 ) of the Bcat gene. These oligos were annealed, ligated with the pU6-BbsI-chiRNA plasmid, and transformed into Escherichia coli. For annealing, 100 mM of each oligo along with T4 ligation buffer were run on a thermal cycler at 95 C for 5 min, followed by a rampup to 25 C at a rate of 0.1 C/s. After the run, the annealed oligos were used for transformation by the pU6-BbsI-chiRNA plasmid. The plasmid was linearized by cutting using the BbsI restriction enzyme. After digestion, the linearized plasmid was mixed with annealed oligo, the oligos were ligated into the plasmid using T4 ligation buffer and T4 DNA ligase, and the reaction mixture was incubated at 37 C overnight. The insertion was confirmed by sequencing with T3 or T7 oligos.

Selection of Bcat mutant flies by PCR analysis
After the microinjection, the flies were crossed with X chromosome balancers to detect mutants of Bcat in the F1 offspring. The deletion of the Bcat was confirmed by PCR analysis. The primer sequences are listed in Supplementary Table 3.

Circadian rhythm
For locomotor activity experiments, freshly eclosed males were collected and transferred onto standard glass tubes containing standard Drosophila medium. These flies were monitored for 6 days using the Drosophila Activity Monitoring System (Trikinetics, Waltham, MA, USA) in an incubator with a 12 h LD (500 lux: 0 lux) cycle. They were exposed to 5, 10, 100, or 500 lux at night after 3 days of LD. Locomotor activity was calculated at 30 min intervals (Chiu et al., 2010). The entrainment index (EI), the ratio of peak activity to total activity for the day, morning anticipation (MA), and the ratio of pre-morning activity to midnight activity, were calculated from the locomotor activity data on day 3 of LL. The estimated intensity of circadian oscillations was calculated (Cho et al., 2016).

Stress resistance assay
To determine resistance to superoxide-induced stress, 20 males (w CS10 ) sampled 3 days after eclosion were kept at 25 C and entrained in 12 h-LD cycles followed by 3 days in LL (n ¼ 60, each condition). The flies were transferred to new food vials containing 20 mM paraquat (PQ, Sigma-Aldrich, St. Louis, MO, USA) after 6 h of food deprivation. To determine heat resistance, male flies were placed in a vial and exposed to a temperature of 40 C. To assess resistance to metabolic stress in the form of starvation, male flies or another wild-type male flies (Oregon-R) were cultured in a vial containing 1% agar. The number of dead flies was counted every 10 min until all the flies died. To confirm the effects of BCAAs on metabolic stress, the male flies were cultured in 1% agar containing 0.1 M or 0.05 M Val, Leu, Ile, Methionine (Met), or Glycine (Gly). Dead flies were counted every 4 h. The survival rate (% survival) was calculated at each time point.

Quantitative reverse-transcription PCR (qRT-PCR)
The mRNA levels of circadian clock genes (cry, tim, per, Clk), glucose-6-phosphatase alpha (G6PC), and phosphoenolpyruvate carboxykinase (PEPCK) were analyzed with fly heads by qRT-PCR in triplicate using the SYBR Green Master Mix and an ABI 7500 instrument (Applied Biosystems, Waltham, MA, USA). On the other hand, the expression levels of the Bcat (CG1673), bckdhA (CG8199), and bckdhB (CG17691) genes were analyzed using the whole body. The mRNA level of the ribosomal protein 49 (rp49) gene was used as the internal control (Benito et al., 2010, Bae et al., 1998. The expression levels of the circadian genes at ZT 6 was used as reference to compare the daily expression levels of these genes. Approximately 1 mg of RNA was used for cDNA synthesis. The primer sequences used in this study are listed in Supplementary Table 3.

Thin-layer chromatography (TLC)
Thirty flies exposed to 0 or 10 lux at night were collected and homogenized in a 2:1 v/v mixture of chloroform and methanol. The homogenates were directly spotted on TLC glass plates (20 Â 20 cm) covered with a Silica matrix. TAG was separated with a 70:30:1 v/v/v mixture of hexane, diethyl ether, and acetic acid as the mobile phase. The sample volumes loaded on the TLC plates were standardized to be equivalent to 100 lg of protein. Lipid bands were visualized by submerging the plate in iodine or vanillin solution for 10 min, followed by charring at 100 C. TAG bands from this plate were scanned and quantified using ImageJ software (NIH, Bethesda, MD, USA).

Carbohydrate measurement
After exposure to LL (at zeitgeber time, ZT 0), whole w CS10 flies were collected after exposure to 0 lux or 10 lux and homogenized using Tris-Phosphate-Triton X-100 buffer. The homogenates were centrifuged at approximately 3,600 Â g for 15 min at 4 C. According to the manufacturer's instructions, glucose levels were determined in the supernatant using a glucose assay kit (Sigma-Aldrich). Trehalose and glycogen levels were assessed in the same sample by incubating with trehalase and amylase enzymes at 37 C overnight. Similarly, hemolymph was collected by centrifugation at 2,500 Â g for 5 min from the whole body of 40 flies with a pierced thorax after exposing them to 0 lux or 10 lux.

Bodyweight measurement
Freshly eclosed flies were collected and exposed to LD conditions for 3 days. The flies were then exposed to 10 lux at night for 3 days in the LL condition. At the end of the LL period, body weight was measured in a batch of 10 flies in each sample.

Feeding assay
The amount of food consumed by flies was estimated by the CAF E (Capillary Feeder) assay (Ja et al., 2007). After flies were exposed to 0 lux or 10 lux at night for 3 days, the amount of food containing blue dye (Edentown FNB, Incheon, Korea) in a capillary tube was measured.

Gas chromatography/mass spectrometry (GC/MS)
At the end of LL, the whole bodies of 20 flies in triplicate were homogenized in 80 mL of water and 300 mL of chloroform: methanol (1:2) for 1 min. One hundred microliters of chloroform was added, followed by addition of 100 mL of water and thorough mixing. The samples were centrifuged at 6,000 Â g for 10 min at 4 C. The upper phase (polar) was transferred to new tubes and dried for 3-4 h using a ScanSpeed 40, SpeedVac concentrator (Thermo Fisher Scientific, Waltham, MA, USA). The concentrated samples were dissolved in 100 mL N,O-Bis trifluoroacetamide (BSTFA) and 10% trimethylchlorosilane (TMCS) (final concentration 100 mg/mL) for GC/MS.

Statistical analyses
Data were collected and analyzed using Excel (Microsoft, Redmond, WA, USA) and Prism 6 (GraphPad, La Jolla, CA, USA). A log-rank test was performed to analyze the lifespan trends of survival curves obtained from starvation assays. The expression patterns of circadian genes measured by the qPCR data were analyzed by two-way analysis of variance (ANOVA) with ZT and light intensity as the main effects. The EI was analyzed by one-way ANOVA. The student's ttest was used in statistical analyses of the body weight, carbohydrate assays, and expression of amino acids. Statistical significance was considered at p < 0.05.

Results
Disruption of circadian rhythm by short-term nocturnal exposure to 10 lux dim light Flies were exposed to 0-500 lux LAN for 3 days (from day 7 to day 9) to examine whether the dim LAN affects the circadian rhythm after exposure to light (LD) of the normal cycle for 3 days (from day 4 to day 6) (Figure 1(A)). Wild-type flies (w CS10 ) exposed to 5 lux (Figure 1(B)b) or 10 lux (Figure 1(B)c)) light intensity for 3 days showed an abnormal rhythm with decreased morning locomotor activity and increased night locomotor activity (Figure 1(B)a-(B)c)) compared to the control group not exposed to LAN (0 lux, LD condition, Figure 1(B)a). The circadian rhythm of flies exposed to higher light intensity (100 lux) at night was more disturbed (Figure 1(B)d). When exposed to light equivalent to daytime light (500 lux) at night, the rhythmic locomotion disappeared within 3 days (LL condition, Figure 1(B)e). Similar results that the rhythmicity of locomotor activity disappears in constant light have been previously reported (Azevedo et al., 2020;Cho et al., 2016).
The degree to which flies are entrained by light was compared by calculating the entrainment index (EI) ( Figure  1(C)). When exposed to the intense light of 100 lux or 500 lux at night, the EI decreased significantly (Figure 1(C); 100 lux, p ¼ 0.0152; 500 lux, p < 0.0001). Also, the EI decreased significantly when flies were exposed to dim light of 5 lux or 10 lux (Figure 1(C); 5 lux, p ¼ 0.0404; 10 lux, p ¼ 0.0179). Drosophila raised in the laboratory typically display increased activity before the light is turned on, a phenomenon termed morning anticipation (MA) behavior. We observed that short-term light exposure at night affected MA behavior. This behavior decreased significantly in flies exposed at night to a short period of intense light (100-500 lux) and dim light (5-10 lux) (Figure 1(D); 5 lux, p ¼ 0.0247; 10 lux, p ¼ 0.0335). These results imply that The experimental design used to study the perturbation of circadian rhythm by short-term nocturnal exposure to dim light. After hatching, male flies were cultured in a normal photoperiod from day 4 to day 6 (LD, day 500 lux/night 0 lux) and then exposed to light with different illuminances (0-500 lux) at night for 3 days from day 7 to day 9 (LL). The effects of light exposure at night, including effect on starvation stress, during normal photoperiod (LD), were evaluated by various methods. (B) During the LL period, flies were exposed to the light of different illuminances (a, 0 lux; b, 5 lux; c, 10 lux; d, 100 lux; e, 500 lux) at night. The locomotor activity of the flies was measured. (C,D) During the LL period, the entrainment index and morning anticipation degree of flies exposed to different illuminances at night (exposure to dim light of 5 lux or more) significantly decreased (C, one-way ANOVA). (E-H) The expression of circadian clock genes (E, cry; F, tim; G, per; H, Clk) in the head was confirmed by RT-PCR on day 3 after night exposure to 10 lux light at 4 h intervals. Although the oscillations of expression of each gene were not shifted, a significant difference in expression level was observed over time (two-way ANOVA). Expression levels were normalized using rp49 as a standard, and data are presented as mean ± SEM. Ã p<0.05, ÃÃ p<0.001, ÃÃÃ p<0.0001.
short-term dim light exposure at night affects the rhythmic behaviors of Drosophila.
Disruption of circadian rhythm by short-term exposure to 10 lux light is associated with increased vulnerability to starvation stress Since short-term dim light exposure at night affected the molecular and behavioral circadian patterns of Drosophila (Figure 1), we investigated the effects of these conditions on the responsiveness to stresses like heat, PQ, and starvation. Wild-type flies exposed to 10 lux at night showed no significant difference from the control group in resistance to heat stress (Figure 2(A); P ¼ 0.9010). The group exposed to 10 lux at night showed a more vulnerable response to PQ (Figure 2(B); p < 0.0001) and starvation stress (Figure 2(C); p < 0.0001). Other wild-type flies (Oregon-R) exposed to the same conditions showed a similar response to starvation stress (Figure 1(D); p ¼ 0.0002). These results suggest that nocturnal 10 lux dim light exposure affects the expression levels of circadian genes and rhythmic locomotion in wildtype Drosophila.
We were interested in the increased vulnerability to metabolic stress in the flies briefly exposed to dim LAN and tried to find the possible causes. As circadian genes play an essential role in regulating genes involved in catabolism (Bailey et al., 2014), we measured glucose levels and the expression levels of genes involved in carbohydrate metabolism to confirm whether the sensitivity of metabolic stress to dim light is due to increased glycogen breakdown. Flies exposed to 10 lux LAN for 3 days displayed significantly increased food consumption (Figure 2(E); p < 0.01), but there was no difference in body weight (Figure 2(F); p ¼ 0.0638). Also, while glucose levels in the hemolymph were similar between the two groups, trehalose and glycogen levels were significantly decreased (Figure 2(G); trehalose, p < 0.05, glycogen, p < 0.001). The level of stored fat, TAG, was not significantly different (Figure 2(H); p ¼ 0.301). In addition, the hemolymph glucose level immediately after exposure to the nocturnal dim light for 3 days (ZT 0 or ZT 24) was not different from that of the control group. After 24 h of starvation, the glucose level of the flies exposed to 10 lux was significantly decreased (Figure 2(I); p < 0.01). We also measured the mRNA expression levels of G6PC and PEPCK to determine whether exposure to dim light affects Figure 2. Fruit flies exposed to 10 lux LAN for 3 days become sensitive to metabolic stress, such as starvation. (A-D) Resistance to various stresses (A, 40 C heat; B, paraquat, PQ; C and D, starvation, with only water provided) of flies exposed to 10 lux LAN for 3 days. Wild-type flies (w CS10 ) exposed to the LAN showed no change in heat resistance (A) but were more sensitive to PQ and starvation stress (B and C, p<0.0001, log-rank test). Other wild-type flies (Oregon-R) were also more sensitive to starvation after short-term exposure to the LAN (D, p<0.001 by log-rank test). Exposure to LAN significantly increased food consumption (E, p<0.01, unpaired t-test), but there was no change in body weight after exposure (F, p¼0.064, unpaired t-test). (G,H) Glucose and TAG levels in the whole body of flies after short-term exposure to dim LAN were not different from those of the control group (G, p¼0.7759; H, p¼0.301, unpaired t-test), but trehalose and glycogen levels were significantly lower in flies exposed to 10 lux light (G, unpaired t-test). (I) Short-term nocturnal exposure to 10 lux light followed by 24 h starvation significantly decreased blood glucose levels in flies (p<0.01, unpaired t-test). (J) After short-term nocturnal exposure to 10 lux light, the mRNA levels of G6PC were significantly increased (p<0.01, unpaired t-test). Data are presented as mean ± SEM. Ã p<0.05, ÃÃ p<0.001, ÃÃÃ p<0.0001.
In summary, exposure of w CS10 flies to dim LAN increases activity but decreases sleep, blood sugar, and stored energy levels. The nocturnal dim light exposure also promotes the decomposition of stored sugars and gluconeogenesis. As a result, blood glucose decreases further after starvation, while feeding behavior increases to compensate for the reduced energy storage.

Enhanced metabolism of BCAAs by perturbation of short-term circadian rhythm
To find the cause of vulnerability to starvation due to exposure to dim LAN, we examined changes in amino acid levels, which are another energy source. Amino acid levels in whole body-homogenates of flies were determined using GC/MS. In the flies exposed to 10 lux LAN, BCAAs, such as Ile and Val, decreased significantly compared to the control group (Figure 3(A); Ile, p < 0.0001; Val, p < 0.001). The other BCAA, Leu, was not detected. The findings reveal that exposure to short-term nocturnal dim light is associated with the metabolism of BCAAs.
BCAT and BCKDHA/B complexes convert BCAAs to branched-chain acyl-CoA used for ATP generation through the TCA cycle (Figure 3(B)). The decrease in BCAA levels in Drosophila exposed to dim LAN may be due to the increased activity of these enzymes. To assess this possibility, we checked the mRNA levels of Bcat, bckdhA, and backhB, which encode BCAA metabolizing enzymes. The expression levels of these genes increased significantly in fly bodies at ZT 24 after exposure to dim light compared with those in the control group (Figure 3(C); CG1673, p < 0.05; CG8199, p < 0.01; CG17691, p < 0.05). In addition, Bcat expression showed a daily oscillation without LAN but increased significantly at ZT 18 and ZT 22 with LAN (Figure 3(D); ZT 18 and ZT 22, p < 0.05).
To validate that BCAA-metabolizing enzymes mediate the circadian disturbance by dim light exposure, Bcat mutant flies in which a part of the Bcat gene initiation site was deleted were constructed using CRISPR/Cas9 technology. The deletion of the Bcat gene was confirmed by PCR of genomic DNA (Figure 3(E,F)). Bcat is located on the X chromosome, and homozygous mutants are lethal. Thus, heterozygous males (Bcat 171/þ ) were subjected to starvation stress after the same circadian perturbation. Bcat 171/þ and another mutant, Bcat 172/þ showed no vulnerability to the starvation stress (Figure 3(G), p ¼ 0.6096; Supplementary   Figure 3. Short-term exposure to dim LAN promotes BCAA metabolism, reduces levels of BCAAs, and sensitizes flies to metabolic stress. (A) The levels of amino acids Ile, Val, and Met in the whole body of Drosophila exposed to short-term nocturnal 10 lux light were significantly reduced (unpaired t-test). (B) BCAT and BCKDH complex enzymes metabolize BCAAs to branched-chain acyl-CoA that enters into the TCA cycle. (C) Expression levels of Bcat (CG1673), bckdhA (CG8199), and bckdhB (CG17691) genes measured in the whole body exposed to LAN significantly increased (unpaired t-test). (D) The expression level of Bcat was high during the day and low at night. Exposure to 10 lux LAN significantly increased Bcat expression level at . (E-H) The Bcat gene in the Drosophila X chromosome was removed by the CRISPR/Cas9 method (E), and the deletion site was confirmed by PCR (F). Small PCR products (0.4 kb) were detected in Bcat deletion mutants. Bcat heterozygous flies (Bcat 171 /þ) showed no sensitivity to starvation stress even when exposed to 10 lux LAN (G, p¼0.6096, log-rank test). BCAA levels in the whole body of Bcat heterozygous flies were not different from those of the control flies (H, unpaired t-test). Data are presented as mean ± SEM. Ã p<0.05, ÃÃ p<0.001, ÃÃÃ p<0.0001. Figure 1(A), p ¼ 0.3599). Also, when the whole body homogenate of Bcat 171/þ flies exposed to 10 lux LAN was analyzed by GC/MS, the relative amounts of Leu, Ile, and Val did not differ from the control group (Figure 3(H); Leu, p ¼ 0.0921; Ile, p ¼ 0.4474; Val, p ¼ 0.4726). These results indicate that short-term exposure to dim LAN increases the activity of BCAA metabolizing enzymes and lowers the concentration of BCAAs in the body, which may cause vulnerability to starvation stress.
Suppressing Bcat expression or feeding BCAAs alleviates metabolic vulnerability caused by circadian rhythm disturbances Transgenic flies in which Bcat expression is knocked down during the adult stage in specific tissues were produced using inducible GAL4 and RNAi. To confirm whether Bcat RNAi is working with the GAL4 driver, we conducted qRT-PCR with the flies (actin GS-GAL4 > Bcat RNAi) fed with RU. The expression of Bcat was significantly reduced in the whole body of these flies (Supplementary Figure 3). The flies with normal Bcat expression showed vulnerability to starvation after exposure to 10 lux LAN (RU-with 10 lux LAN), whereas the normal flies that were not exposed to LAN (0 lux) and not fed with RU (RU-) and flies with suppressed Bcat expression in the whole body (RU þ with 0 lux or 10 lux LAN) showed significantly higher survival (actin GS-GAL4 > Bcat RNAi; Figure 4(A); Supplementary Table 2). Similarly, knocking down Bcat expression in all neurons (elav GS-GAL4 > Bcat RNAi; Figure 4(B)) or fat bodies (S106 GS-GAL4 > Bcat RNAi; Figure 4(C)) during adulthood significantly increased the survival, compared with that of control flies exposed to 0 lux LAN (Supplementary Table 2). These results show that the decrease in Bcat expression in several types of cells or tissues alleviates the susceptibility to metabolic stress after exposure to dim LAN.
We carried out a starvation test after the dim light exposure at night to determine whether feeding various amino acids during starvation rescues the vulnerability to metabolic stress reflected by decreased BCAAs in the wild-type flies (w CS10 ) (Figure 4(D-I)). When only water was provided, the survival rate of flies exposed to 10 lux LAN decreased significantly compared with that of the control group ( Figure  4(D); p < 0.0001). Methionine (Met)-which was detected as a low-level amino acid by GC/MS analysis after short-term dim light exposure (Figure 3(A))-or glycine (Gly), as another negative control, were fed during starvation. However, neither of them improved susceptibility to metabolic stress (Figure 4(D), p < 0.0001; Figure 4(E) and F, p < 0.0001). In contrast, feeding Val (Figure 4(G)), Leu (Figure 4(H)), or Ile (Figure 4(I)) significantly decreased vulnerability to metabolic stress. Therefore, the short-term disturbance of circadian rhythm due to dim light exposure at night reduces the concentration of BCAAs in the body by enhancing the expression of BCAA metabolic enzymes, making flies more susceptible to metabolic stress (Figure 4(J)). The findings show that BCAAs alleviate the metabolic vulnerability caused by disturbances in circadian rhythm.

Discussion
The circadian rhythm is synchronized by 24-h oscillations of environmental changes, such as daily light and temperature cycles. The circadian rhythm regulates the daily rhythms of physiology and behavior. Drosophila exhibits two distinct peaks in locomotor activity in response to external light: a morning peak and an evening peak (Helfrich-F€ orster, 2000). Flies have approximately 150 clock neurons distributed along the lateral and dorsal regions of the brain (Yao et al., 2016). These clock neurons regulate the daily locomotor activity of Drosophila. The small lateral ventral neurons (sLNvs) expressing PDF and the posterior dorsal neuron 1 (DN1p) that receive inputs from PDF-expressing sLNvs regulate the MA behavior (Zhang et al., 2010). In contrast, the CRY-positive dorsal lateral neurons (LNds) and 5th sLNv that do not express PDF participate in regulating the evening anticipation behavior (Grima et al., 2004;Stoleru et al., 2004). We observed that the flies exposed repeatedly to dim LAN decreased the MA behavior but increased nocturnal locomotor activity (Figure 1(B,D)). Thus, though confirmation at the cellular level is required, it could be hypothesized that exposure to dim LAN reduces the activity of PDF-positive sLNvs in the nighttime and increases the activity of the CRY-positive LNds and 5th sLNv in the daytime.
We performed a starvation test to determine whether LAN affects metabolism and confirmed that the sensitivity of starvation significantly increases in wild-type flies when exposed to LAN (Figure 2(C,D)). Interestingly, the metabolic vulnerability to starvation after short-term exposure to dim LAN was not found in the null mutants of tim, Clk, and cry genes ( Supplementary Fig. 1B-D). In addition, the gene expression levels of BCAA-metabolizing enzymes, including Bcat, bckdhA, and backhB, in wild-type flies significantly increased at ZT 24 by exposure to 10 lux LAN (Figure 3(C)). But such induction of BCAA metabolizing genes was not found in tim mutants (Supplementary Figure 2). Surprisingly, Bcat expression showed an oscillating pattern. Its expression levels increased significantly at ZT 18 and ZT 22 when exposed to 10 lux LAN, compared to those at 0 lux LAN (Figure 3(D)). The amount of BCAAs throughout the body was inversely proportional to the expression level of Bcat. The increased activity of BCAT increased the metabolism of BCAAs, resulting in a decrease in BCAAs in the body of fruit flies. Similarly, the concentration of BCAAs in the muscle, liver, and blood showed a circadian pattern, increasing at night and decreasing during the day, in mice (Dyar et al., 2018;Oishi et al., 2014). Thus, it is likely that the metabolism of BCAAs is associated with the circadian regulatory system. To further confirm this hypothesis, whether the oscillation of Bcat expression disappears under DD conditions should be examined.
Several possibilities are considered for explaining how the circadian system regulates BCAA metabolism. First, the expression of Bcat can be controlled by binding of circadian factors to the E-box present around Bcat. We investigated Ebox distribution, a target site of the BMAL1/CLOCK complex, inside and outside Bcat, bckdhA, and bckdhB genes (Supplementary Table 1). While more than 15 E-boxes were present around per and tim genes, 8 were found in Bcat, none in bckdhA, and 3 in bckdhB. If BCAA metabolism is under the direct control of the circadian regulatory system, Bcat should be the primary target. However, verifying whether the CLOCK/CYCLE complex binds to these E-boxes and regulates Bcat expression is necessary. Second, the circadian system can regulate BCAA metabolism through Kr€ uppel-like factor 15 (KLF15). KLF15 is a C2H2-type zinc finger transcription factor and regulates the expression of enzymes involved in nitrogen homeostasis in a clockdependent manner (Jeyaraj et al., 2012). KLF15 controls the circadian rhythm of BCAA metabolism in the mouse liver and muscle (Fan et al., 2018;Jeyaraj et al., 2012). In the fasting state, KLF15 upregulates BCAT2 to inhibit lipogenesis and promote gluconeogenesis in the liver (Gray et al., 2007). Also, when the circadian rhythm is disrupted, KLF15 expression increases. The expression levels of KLF15 mRNA and protein increased in per3 null mutant mice and decreased in a per3 overexpression model (Aggarwal et al., 2017). It should be interesting to determine whether metabolic sensitization appears after exposure to LAN in Klf15 (CG2932) mutant flies. Third, the circadian system can regulate BCAA metabolism through microRNAs. MiR-277, activated by dCLOCK/CYCLE, interacts with PAR-domain protein 1 (Pdp1) mRNA (Xia et al., 2019). Overexpression of Drosophila miR-277 inhibits BCAT activity and induces mTOR activity and BCAA accumulation (Esslinger et al., 2013). Therefore, it is necessary to confirm the role of microRNAs in regulating BCAA metabolism.
Drosophila exposed to dim LAN for a short period were vulnerable to metabolic stress, such as starvation ( Figure  2(C)). We found that the cause of vulnerability to metabolic stress is the metabolism of specific essential amino acids, BCAAs. The concentration of BCAAs in the body fluid decreased upon exposure to dim LAN (Figure 3(A)). Resistance to starvation was restored by supplementation with BCAAs (Figure 4(J-L)). Furthermore, Bcat Figure 4. BCAAs alleviate the vulnerability to metabolic stress caused by short-term dim light exposure. (A-C) Suppression of Bcat expression blocks the vulnerability to starvation stress that occurs after exposure to LAN. By using the inducible GAL4/UAS system, Bcat expression was knocked down during the adult stage in the whole body (A, actin GS-GAL4 >Bcat RNAi), neurons (B, elav-GS-GAL4 >Bcat RNAi), or fat body (C, S106 GS-GAL4 >Bcat RNAi). After exposure to 10 lux LAN for 3 days, the response to starvation stress was determined by measuring the survival of these flies. Compared with the survival rate of flies exposed to 10 lux LAN in which Bcat expression increased (RU-with 10 lux LAN), that of the flies that were not exposed to night light (RU-with 0 lux LAN) or the flies in which Bcat expression was suppressed in the whole body, neuron, or fat body with Bcat RNAi (RU þ with 0 lux or 10 lux LAN) was significantly improved (statistical summary in Supplementary Table 2). (D) Short-term nocturnal exposure to 10 lux light sensitized the wild-type flies (w CS10 ) to starvation stress (p<0.0001). (E,F) Feeding 0.1 M Met or 0.1 M Gly resulted in increased sensitivity to starvation stress in the wild-type flies (E and F, p<0.0001). (G-I) Such vulnerability to starvation stress was not observed in the wild-type flies fed with 0.1 M Val (G, p¼0.9888), 0.1 M Leu (H, p¼0.2310), or 0.1 M Ile (I, p¼0.2647) during starvation after short-term nocturnal exposure to 10 lux light. (J) Normally, without LAN, the activity of BCAA metabolizing enzyme genes is not increased to maintain the concentration of BCAAs in the hemolymph, which makes the flies resistant to metabolic stress like starvation. However, when flies are exposed to 10 lux LAN, this light affects the circadian genes and increases the expression of BCAA metabolizing enzyme genes, resulting in decrease in BCAA in the hemolymph and thus increasing sensitivity of flies to the metabolic stress. The log-rank test (A-I) was used to compare the data, and the results are presented as mean ± SEM. heterozygous mutants did not show sensitivity to metabolic stress after exposure to dim LAN (Figure 3(G)). How BCAAs lower susceptibility to metabolic stresses, such as starvation, is an interesting study concept. First, as mentioned above, BCAAs are metabolized and converted to acetyl-CoA or succinyl-CoA, and they can be used for ATP generation via the TCA cycle. In addition, supplementation of BCAAs extended the chronological lifespan of yeast (Alvers et al., 2009), increased mitochondrial biosynthesis, and activated mTOR and SIRT1 signaling and the ROS defense system involving superoxide dismutase 1 and glutathione peroxidase 1 in the heart and skeletal muscle of mice (D'Antona et al., 2010). In addition, Leu directly activates SIRT1 and increases mitochondrial biosynthesis and fatty acid oxidation in adipocytes and myotubes (Liang et al., 2014). Finally, mTOR signaling was activated in mice fed BCAAs, possibly through RHEB-dependent activation of the mTORC1 complex (Jewell et al., 2015). Thus, BCAAs seem to confer resistance to metabolic stress by being used as an energy source and affecting metabolic control factors, ROSrelated defense system, and the immune system.
When the circadian rhythm is disturbed by exposure to LAN, the incidence of metabolic diseases including cancer, diabetes, and obesity increases (Huang et al., 2011). Circadian rhythm is especially disturbed in night workers exposed to excessive LAN. In these individuals, melatonin production is suppressed, resulting in sleep disturbance (Gooley et al., 2011). In addition to light, changes in eating and sleeping schedules, and consumption of stimulants, such as caffeine, can also affect circadian rhythm (Lee & Kim, 2019). Nocturnal caffeine intake has similar effects to changes in circadian rhythm caused by 3 h exposure to 3,000 lux bright light (Burke et al., 2015). Since disturbance of the circadian rhythm in humans is generally associated with sleep disturbance, most treatments for disturbed circadian rhythm consist of improving sleep quality. Improved sleep quality treatment focuses first on restoring the regularity of life, followed by melatonin supplementation and the short-term use of various sedatives and hypnotics (Pagel and Parnes, 2001). Antioxidants, including proanthocyanidins and polyphenols, can improve circadian rhythm. When mice were fed proanthocyanidins, the expression levels of Clk and per2 circadian genes increased in the liver (Pagel and Parnes, 2001). Also, when obese rats were fed polyphenols, the disruption of the circadian rhythm that occurred with obesity was counteracted (Rodr ıguez et al., 2022). In our study, even 5-10 lux dim LAN perturbed circadian locomotor activity and sleep rhythm. Furthermore, this disturbance induced Bcat expression and enhanced BCAAs metabolism, thereby reducing BCAAs in the hemolymph and inducing sensitivity to metabolic stress, such as starvation. The sensitivity was alleviated by suppressing the expression of Bcat (Figure 4(A-C)) or via BCAA supplementation (Figure 4(G-L), Supplementary Figure 4(D-F)). Therefore, BCAAs may have value as defense factors against metabolic changes caused by circadian rhythm disturbances.
We found that short-term exposure to dim LAN in wildtype flies increased the Bcat expression level and decreased the amount of BCAAs in the body, making the flies more susceptible to metabolic stress. Flies with Bcat heterozygous mutations maintained the concentration of BCAAs in the body, even when exposed to the same light, and did not show sensitivity to metabolic stress. These results highlight the relationship between light-induced circadian rhythm regulation and amino acid metabolism and suggest a protective effect of BCAAs against metabolic stress.