Clock‐controlled arylalkylamine N‐acetyltransferase (aaNAT) regulates circadian rhythms of locomotor activity in the American cockroach, Periplaneta americana, via melatonin/MT2‐like receptor

Melatonin (MEL) orchestrates daily and seasonal rhythms (eg, locomotion, sleep/wake cycles, and migration among other rhythms) in diverse organisms. We investigated the effects of pharmacological doses (0.03‐1 mM) of exogenous MEL intake in the cockroach, Periplaneta americana, on locomotor activity. As per os MEL concentration increased, cockroach locomotor rhythm in light‐dark (LD) cycles became more synchronized. The ratio of night activity to 24‐h activity increased and the acrophase (peak) slightly advanced. MEL application also influenced total activity bouts in the free‐running rhythm. Since MEL slightly influenced τ in the free‐running rhythms, it is not a central element of the circadian pacemaker but must influence mutual coupling of multi‐oscillatory system components. Arylalkylamine N‐acetyltransferase (aaNAT) regulates enzymatic production of MEL. aaNAT activities vary in circadian rhythms, and the immunoreactive aaNAT (aaNAT‐ir) is colocalized with the key clock proteins cycle (CYC)‐ir and pigment‐dispersing factor (PDF)‐ir These are elements of the central pacemaker and its output pathway as well as other circadian landmarks such as the anterior and posterior optic commissures (AOC and POC, respectively). It also partially shares immunohistochemical reactivity with PER‐ir and DBT‐ir neurons. We analyzed the role of Pamericana aaNAT1 (PaaaNAT1) (AB106562.1) by injecting dsRNAaaNAT1. qPCR showed a decrease in accumulations of mRNAs encoding PaaaNAT1. The injections led to arrhythmicity in LD cycles and the arrhythmicity persisted in constant dark (DD). Continuous administration of MEL resynchronized the rhythm after arrhythmicity was induced by dsRNAaaNAT1 injection, suggesting that PaaaNAT is the key regulator of the circadian system in the cockroach via MEL production. PaaaNAT1 contains putative E−box regions which may explain its tight circadian control. The receptor that mediates MEL function is most likely similar to the mammalian MT2, because injecting the competitive MT2 antagonist luzindole blocked MEL function, and MEL injection after luzindole treatment restored MT function. Human MT2‐ir was localized in the circadian neurons in the cockroach brain and subesophageal ganglion. We infer that MEL and its synthesizing enzyme, aaNAT, constitute at least one circadian output pathway of locomotor activity either as a distinct route or in association with PDF system.


| INTRODUCTION
McCord and Allen first observed that bovine pineal extracts injected into Rana pipiens tadpoles caused lightening of their skin color. 1 Lerner and his coworkers 2 isolated and identified the bioactive factor as melatonin (N-acetyl-5methoxytryptamine; MEL) from bovine pineals, initiating the modern era of pineal research. The pineal organ in some avian species contains a circadian pacemaker, because the surgical removal renders the normal rhythms arrhythmic; however, rhythmicity was restored by transplanting a donor pineal into the pinealectomized bird, which drove the phase of the donor bird. 3 The pineal gland secretes MEL both in vivo and in culture, either from q dispersed cell culture or organ culture of chicken pineal, exhibiting a circadian rhythm. 4 Exogenous MEL administered periodically restores locomotor rhythmicity to pinealectomized birds. 5 Arylalkylamine N-acetyltransferase (aaNAT) is the final step in MEL biosynthesis, and it regulates seasonal reproduction in hibernating mammals, reptiles, and songbirds, and it controls migration behavior. 6,7 aaNAT in vertebrates catalyzes the N-acetylation of serotonin (5-hydroxytryptamine; 5-HT) to produce N-acetyl serotonin (NAS) for the synthesis of MEL by hydroxy indole O-methyltransferase (HIOMT). aaNAT is the key enzyme to generate a MEL rhythm 8,9 and a rate-limiting step in circulating MEL levels. 10 MEL is a chemical token for night, and the accumulation of MEL depends on the length of the dark period aaNAT is exposed to. The latter is very sensitive to light, 11,12 which delayed its chromatographic isolation for a long time.
The suprachiasmatic nucleus (SCN) serves as the circadian pacemaker (CPM) in mammalian and some avian systems. SCN receives photoperiodic signals from the retina along the retinohypothalamic tract via glutamate and sympathetically regulates pineal MEL biosynthesis via the supercervical ganglion. 13 Both avian and mammalian aaNATs are clock-controlled genes (CCGs). 14,15 The rhythmic activity in the MEL metabolic pathway is highly conserved among vertebrates. MEL shows a stable rhythmicity of about 24 h, with the peak at night. MEL is a neurohormone that regulates many physiological processes including locomotory activity, sleep/wake rhythms, body temperature, blood pressure, feeding, and oncogenesis. 16 MEL also has been reported in various nonvertebrate taxa, including insects. [17][18][19] The aaNAT family has been found in Gram-positive bacteria, fungi, algae, placozoans, annelids, cephalochordates, and vertebrates 9,20 as well as in higher plants. 21 Based on sequence similarity, it can be classified as either vertebrate (VT-aaNAT) or nonvertebrate (NV-aaNAT). 20 Structural and functional aspects of insect aaNAT (insect-type iaaNAT) have been reviewed in Hiragaki et al 22 iaaNAT occurrence is not limited to arthropods but it occurs also in lower deuterostomes. Evolutionary lineage analyses indicate that iaaNATtype is more basally placed, that is, apomorphic, and that VT-type show more recent derivation, since some protochordates such as tunicates and hemichordates had insect-type enzyme. 22 iaaNAT is involved in coloration and sclerotization of cuticles and suppression of melanin pigmentation. 23 This is the first documented role of aaNAT in insects, and it was conventionally called dopamine N-acetyltransferase (DAT). aaNAT loss of function by mutation causes possible buildup of excess dopamine that is then diverted for melanization in Bombyx mori. 24 The aaNAT enzymes also control the inactivation and metabolic clearance of a number of neurotransmitter monoamines as insects do not express or have very low monoamine oxidase (MAO) activities. 25,26 This stands in contrast to mammals where MAOs are the major detoxification enzymes. 27 MEL has a neurohormonal releasing effect in the insect nervous system and induces the release of the prothoracicotropic hormone (PTTH) from the brain in the American cockroach, P americana. 28 Orally administrated MEL synchronized both entrained and free-running rhythms in locomotor activity of the house cricket, Acheta domesticus. 29 Both MEL content in the hemolymph and aaNAT enzymatic activity in the brain-subesophageal ganglion complex showed a circadian fluctuation in P americana, 30 B mori, 31 function is most likely similar to the mammalian MT2, because injecting the competitive MT2 antagonist luzindole blocked MEL function, and MEL injection after luzindole treatment restored MT function. Human MT2-ir was localized in the circadian neurons in the cockroach brain and subesophageal ganglion. We infer that MEL and its synthesizing enzyme, aaNAT, constitute at least one circadian output pathway of locomotor activity either as a distinct route or in association with PDF system.  32,33 We purified the enzymes from various tissues of P americana and obtained calculated amino acid sequences. 30,34 The sequences are true aaNAT, since the proteins expressed showed N-acetylation capabilities. 31,32 iaaNAT is the critical link between the circadian system and photoperiodism in A pernyi. 33 After injecting dsRNA iaaNAT into diapause pupae, the pupae failed to exit diapause under long days. aaNAT was indispensable for PTTH secretion. Immunohistochemical localization of various clock proteins indicates that PER-ir, Cyc-ir, etc, are colocalized with aaNAT-ir, MT-ir, and HIOMT-ir, while hMT2-ir and PTTH-ir are colocalized. Further, both groups of neurons are situated side by side. 33 Injection of dsRNA iaaNAT induced a large decline of MEL content over 48 h. In contrast, after knockdown of per, MEL levels were significantly higher than controls after only 24 h. We infer that pupal diapause of A pernyi is terminated by MEL via aaNAT activation and the iaaNAT is a circadian-controlled gene (CCG). 33 Our next goal was to clarify the possible role of iaaNAT in circadian rhythms in locomotor activity, using P americana as a model insect.
In this study, we posed the hypothesis that aaNAT functions as an essential connection between the circadian system and behavioral rhythm. Here, we report on the outcomes of experiments designed to test our hypothesis.

| Experimental insect
Stock colonies of American cockroaches (Periplaneta americana L.) were maintained in a walk-in constant temperature room set at 25°C and 60% relative humidity (RH) under 12:12 (LD) and provided with water and artificial diet (MF, Oriental Yeast Corp., Tokyo, Japan) ad libitum. Newly emerged white male adults were collected immediately and kept in a transparent plastic box (14 × 14 × 20 cm) to ensure the ages.

| Application of exogenous melatonin and related pharmaceuticals
Pharmaceutical grade MEL (Wako Pure Chemical Industries, Osaka, Japan) was dissolved in a minimal volume of ethanol and then diluted with Gibco distilled water (Thermo Fisher Scientific, MA, USA), 0, 0.03, 0.1, 0.3, and 1 mM MEL concentrations, with less than 0.01% EtOH. Cockroaches were treated with MEL per os. 29 The EtOH did not influence the amplitude, acrophase, and daily activity of the locomotor rhythm (data not shown). Luzindole, a competitive antagonist of MEL receptor type 2 (MT 2 ), was purchased from Sigma-Aldrich (St. Louis, USA). It was dissolved in 10 µl phosphate-buffered saline (PBS) at 100 nM (=30 pg). Luzindole was injected into each cockroach with a Hamilton 701 syringe equipped with a 26s gauge needle (Hamilton Co., Reno, Nevada, USA). The injury to the cuticle was sealed with an instant adhesive, Aron-alpha (Toagosei Co. Ltd., Tokyo, Japan). Locomotor activity was examined in 10 male cockroaches injected with luzindole, and an equivalent control group was injected with MEL at concentrations reported in Results. Locomotor activity in LD 12:12 was recorded for 10 days before injection and for 10 days after injection. The activity was recorded for a further 10 days in DD from the last lights-off.

| Behavioral analysis
Locomotor rhythms were recorded in a monitoring chamber (plastic box, 193 × 104 × 27 mm) with a shelter (U-shaped plastic of 50 × 40 × 20 mm upside down) flanked by an infrared light emitter and detector (GT2; Takenaka Electronic Industrial, Kyoto, Japan). 35 Water was provided from a bottle (35 mm D × 50 mm H) with a wire-meshed opening. Artificial diet was supplied as food. When a cockroach crossed the infrared beam, a signal was sent a computer (PC98; NEC, Tokyo, Japan) for a 6-min bin. Locomotor rhythms of individuals were recorded at 25°C under LD 12:12; light source was a 10 W fluorescent lamp providing more than 400 lux. Adult males aged 7 days were kept in LD 12:12 at 25°C and 60% RH for 10 days for entrainment and then were allowed to free-run in DD. Ten insects were used for each group. Statistics of t and F were used for the differences in activity level, acrophase, τ, and amplitude between the group drinking water control and that drinking different concentrations of MEL. Amplitudes and acrophases were analyzed by cosinor analysis and τ by chi-square periodogram analysis.

| Preparation and injection of dsRNA
The brain-subesophageal ganglion complex (Br-SOG) of P americana was dissected and immediately transferred to liquid N 2 . The total RNA was extracted by using the RNAiso Plus reagent (Takara Bio Inc, Kusatsu, Japan). Two hundred fifty ng samples of total RNA and ReverTra Ace ® qPCR master mix (Toyobo Co. Ltd., Osaka, Japan) were used to synthesize cDNA. Double-stranded DNA with 282-base pair (bp) fragments of aaNAT (GenBank accession number: AB106562.1, now designated as PaaaNAT1) was amplified by PCR. The sequence AB106562.1 used here for RNAi corresponds to the encoded aaNAT A that showed circadian fluctuation in the activity and mRNA content. 34 The primers used to generate templates for in vitro transcription contained a T7 promoter (TAATACGACTCACTATAGGGAGA) at each end. The primers for amplification were as follows: F: 5´-TAATACGACTCACTATAGGGAGATAATG GCAG TATCCAGAA-3′; and R: 5′-TAATACGAC TCACT ATAG G GAGAGATATTATGCGCACT TCTAC-3′. The PCR mixture (50 µL) included 4µL of 50X diluted cDNA template, 5 µL (10 pmol) of each primer, 5µL of 2 mM dNTPs, 25 µL of 2x buffer, and 1 µL of KOD FX Neo (Toyobo Co. Ltd.). PCR (35 cycles) was performed with denaturation at 98°C for 10 sec, annealing at 65°C for 30 sec, and extension at 68°C for 5 min. The aaNAT double-stranded RNA (dsRNA aaNAT1 ) was synthesized from a purified PCR product by using a MEGAscript T7 Transcription Kit (Thermo Fisher Scientific, MA, USA), following the manufacturer's protocol. For the experimental group, 32 µg of dsRNA aaNAT1 was injected by Hamilton syringe into the abdominal cavity of each cockroach. The control groups were injected with an equal volume of H 2 O. 36 We injected dsRNA GFP into cockroaches and observed no effect on the expression of iaaNAT transcription. After injection, each cockroach was returned immediately to the rearing container or placed in a locomotion monitoring chamber.

| Measurement of iaaNAT expression by qPCR
Total RNA was extracted and purified from individual adult male Br-SOGs with RNAiso Plus. Total RNA was treated with DNase I (Takara Bio Inc) followed by cDNA synthesis using Rever Tra Ace ® qPCR master mix. The qRT-PCR was performed with the SYBR ® Green and THUNDERBIRD ™ qPCR Mix (Toyobo Co. Ltd.), with the forward and reverse primers 5´-TGTGTTTCAACCAGCTCTGC-3´ and 5´-AACTTCCACTCGTAGTGGTTCC-3´ for iaaNAT; and 5´-TGAATCCTAAGGCCAACAGG-3´ and 5´-AC CG G AATCCAGCACAATAC-3´ for actin (GenBank: AY116670.1). Cycling parameters were 95°C for 1 min to activate DNA polymerase and then 40-cycle amplification programmed for 95°C for 15 sec and 60°C for 2 min. The acquisition of fluorescence data was performed at the end of the elongation step using thermal cycler dice real-time system software (Takara Bio Inc). An initial amount of template was calculated from cDNA, and then, a standard curve was generated for each PCR run. For expression levels of each transcript, actin mRNA was used as the internal control. 33 The primers used in qPCR were designed to correspond to outside the region of the dsRNA construct. The sizes of the PCR products were 130 bp for iaaNAT and 140 bp for actin.

| Immunohistochemistry
Specificities of all antibodies used in this study have been described. 33,37 The Br-SOGs were isolated at ZT6 and fixed in Bouin's solution for 24 h at 4°C. Artificial Zeitgeber time (ZT) is defined as the time after the light-on under LD (lightdark) cycle = 12:12. Standard histochemical methods were employed for tissue dehydration, embedding in Paraplast Plus ® (Sigma-Aldrich, MO, USA), sectioning (12 µm), deparaffinization, and rehydration. The sections were washed in distilled water and Tris-buffered saline (TBS; 135 mM NaCl, 2.6 mM KCl, 25 mM Tris-HCl, pH 7.6) containing 0.1% Tween 20 (TBS-Tw) at RT, blocked with antibody dilution buffer (TBS-Tw containing 1% BSA) for 30 min at RT. Double labeling followed using anti-Ap iaaNAT (rabbit) and anti-CYC (rat). The sections were incubated with a cocktail of both primary antibodies containing anti-iaaNAT (1:1000) and anti-CYC (1:200), rinsed 3 times in TBS-Tw, and then incubated with 7.5 µg/ml of biotinylated goat anti-rabbit IgG (Vector Laboratories, CA, US). After rinsing in TBS-Tw, the iaaNAT-like immunohistochemical reactivity (-ir) was visualized with green fluorophore using a TSA Labeling Kit #22 (Thermo Fisher Scientific). The sections were rinsed 3 times in TBS-Tw, treated with 1% H 2 O 2 in TBS for 30 min at RT to inactivate residual HRP, and washed again three times with TBS-Tw. Incubation with 0.025% (w/v) avidin in TBS for 30 min at RT was followed by rinsing 3 times with TBS-Tw. Subsequently, incubation with 0.001% (w/v) biotin in TBS for 15 min at RT was conducted to block reactive biotin and streptavidin. The slides were incubated with 7.5 µg/ml of biotinylated goat anti-rat IgG (1:2000) (Vector Laboratories). CYC-ir was visualized with red fluorophore using a TSA Labeling Kit #42 (Thermo Fisher Scientific).
Double labeling was performed using antibodies derived from the same animal donor according to Hiragaki et al. 38 In detail, anti-aaNAT (Rabbit) with anti-PDF (Rabbit) was used as follows. The slides were incubated with the antibody solution (anti-iaaNAT1:1000) overnight at 4°C. After rinsing with TBS-Tw, the slides were incubated with 7.5 µg/ml of biotinylated goat anti-rabbit IgG for 60 minutes in RT. After rinsing with TBS-Tw, the sections were treated with TSA Biotin System (PerkinElmer, MA, USA). After labeling, anti-aaNAT was stripped off the sections for 24 h at RT in stripping buffer (100 mM 2-mercaptoethanol, 50 mM glycine-HCl, pH2.2) in a 40 V horizontal electric field. The sections were then incubated with anti-PDF (1:2000). iaaNAT-ir was visualized with green fluorophore using a TSA Labeling Kit #22. Sections were treated with H 2 O 2 , avidin, and biotin, incubated with biotinylated goat anti-rabbit IgG. PDF-ir was visualized with red fluorophore using a TSA Labeling Kit #42.
The stained sections were washed in TBS and mounted in Aqua-Poly/Mount medium (Polysciences Inc, PA, USA) and examined using a BX50 microscope (Olympus, Tokyo, Japan) equipped with BX-FLA reflected light fluorescence, WIG and NIBA mirror/filter units for detection of Alexa Fluor ® 555 and Alexa Fluor ® 488, respectively and a CCD camera. Immunohistochemical specificities of all antibodies used in this study were confirmed previously. 33,39 The MEL-ir was investigated by employing the primary antibody Rabbit anti-MEL serum (Immunotech IM0615, purchased from MBL) that was used for immunohistochemical staining. Cockroach was preinjected with 10 µl of 1 mM colchicine in PBS to block neurosecretion and dissected 2 days later. A second anti-melatonin serum purchased from Stockgrand Ltd, product number AB/S/01, previously HP/S/704-6483 (also used for the Radioimmunoassay (RIA) procedure), had weaker reactivity in immunohistochemistry (IHC).
Colocalization of aaNAT-ir with clock proteins, PER and DBT, was made using antisera we produced against PamPER and BmDBT of that specificities were checked in Sehadova et al 37 Adjacent sections of 7 µm were mounted onto two sets of slides, and each set was exposed to either anti-PER or anti-DBT as primary antibody at 1:1000.

PaaaNAT1 cis-regulatory motifs related to circadian control
Periplaneta americana genome assembly sequence ASM293952v1 was retrieved from NCBI and blasted against the PaaaNAT1 cDNA sequence AB106562.1. The resulted blast hits were converted into exon and intron coordinates. The PaaaNAT 3kb upstream genomic sequence (from the transcription start side) was then submitted to JASPAR database (available at http://jaspar.gener eg.net) for identifying the putative transcription factors binding sites (TFBS). The resulted output was explored manually to identify a targeted set of conserved cis-regulatory motifs linked to circadian control that were subsequently highlighted as sequence annotations in Geneious 7.1.7 (Biomatters Ltd., Auckland, New Zealand). Corresponding figure was manually redrawn in Adobe Illustrator CS5 (Adobe Inc CA, USA), and the obtained putative elements are coded.

| Statistical analysis
Activity rhythm was analyzed by chi-square periodogram 35 at the significance level of 0.05. Mean differences were analyzed by one-way ANOVA followed by Tukey's HSD (Honestly significant difference) test using SPSS ver. 22 (IBM, NY, US). Differences were accepted as significant for P < .05. Significance of mRNA level was examined each time point by t test.

| Effects of exogenous melatonin on
circadian locomotor activity Figure 1 shows representative double-plotted actograms of insects drinking 0, 0.03, 0.1, 0.3, and 1.0 mM MEL, respectively, under LD12:12 at 25°C. The activity of 10 individuals per group was recorded in LD12:12 for 10 days and then for another 10 days in DD following the last lights-off. Figure 2 summarizes the compiled data in LD12:12. When exogenous MEL was applied, there was a ca. 10% arrhythmic pattern, in comparison to the control, which showed a ca. 30% arrhythmic pattern. The daily activity, acrophase, ratios of nighttime activity, N to entire activity, T (N/T, %), and amplitude were analyzed for each concentration of MEL ( Table 1).
The acrophase of the rhythm in individuals drinking MEL was slightly phase-lagged, and the circadian period was slightly lengthened by increased MEL, compared with the control (F = 3.61, P < .01), but the difference was not significant as shown in Table 2. Both the activity level (F = 3.76, P < .01) and amplitude (F = 3.03, P < .001) increased as MEL concentrations increased up to 1 mM. The daytime activity was suppressed with increased concentration of MELdrinking groups under LD cycle. Therefore, the N/T values were changed to square root of percentages and ANOVA statistical analyses were performed ( Table 1). The effect was highly significant (F = 20.1, P < .0001). A complete description of each circadian parameter is detailed in Table 1. MEL imbibing/drinking had clear effects on the circadian parameters.

PaaaNAT expression
To examine the effect of injection of dsRNA aaNAT1 on the PaaaNAT transcript levels, we measured the levels of PaaaNAT mRNA in the Br-SOG by qPCR (Figure 4). Br-SOGs were collected at ZT6, 7, 14, 21, and 28 days after injection. In dsRNA aaNAT1 -injected cockroaches, the PaaaNAT mRNA level was reduced significantly, by about 89% (F = 10.5, P < .01) from control cockroaches at each time point, suggesting that injection of dsRNA aaNAT1 suppressed the expression of PaaaNAT mRNA throughout the experimental period of 28 days.

| dsRNA aaNAT1 injection desynchronized locomotor rhythm
To investigate the effects of dsRNA aaNAT1 injection on the locomotor rhythm, we measured the locomotor activity in 20 male cockroaches injected with dsRNA aaNAT1 . We also used 10 intact and 10 water-injected cockroaches as controls. Locomotor activity in LD 12:12 was recorded for 10 days before injection and 10 days after injection. It was then recorded for 10 days in DD following the last lights-off. The representative actogram records of locomotor activity are shown in Figure 5. Injection of dsRNA aaNAT1 significantly reduced locomotor activity at night ( Figure 6 and Table 3). Seventy percent of the treated cockroaches (n = 20) became arrhythmic, in comparison to the control cockroaches (n = 20), which remained rhythmic ( Figure 5) after injection in the LD cycle. The result indicates the treated cockroaches lost their ability to synchronize with the Zeitgeber due to silencing of their PaaaNAT1 gene. Under DD conditions, 70% of dsRNA aaNAT1 -injected insects expressed arrhythmia with suppressed activity over a 24-hour period, while only 10% insects in the control lost their free-running rhythm ( Figure 6). DsRNA aaNAT1 -injected cockroaches showed significantly shorter τ compared with control (Table 3). DsRNA aaNAT1 injection disrupted not only phases of circadian rhythm and synchrony in locomotor Ordinate indicates relative abundance of Pa aaNAT1 mRNA. Samples were collected at ZT6 in LD 12:12, 25°C. The Br-SOGs were collected 7, 14, 21, and 28 days after injection of dsRNA aaNAT1 . The abundance of Pa aaNAT1 mRNA was measured by qPCR. The data collected from three independent experiments were averaged and plotted as mean ± SEM. against water injected (control) and dsRNA aaNAT1injected insect (dsRNA aaNAT1 ). The abundance of actin was used as reference. **: highly significant difference (P < .01), as compared to the control group, using Student's t test rhythm, but also the activity level significantly ( Figure 6 and Table 3). Activities in 20 days of postinjection became less than a half that of control.

| Reestablishment of locomotor rhythm by melatonin following dsRNA aaNAT1 injection
Resynchronization of locomotor rhythm was observed in dsRNA aaNAT1 -injected cockroaches when MEL (1.0 mM) was supplied in the drinking water. DsRNA aaNAT1 -injected cockroaches were examined 10 days before MEL administration under LD cycle. Ten days after injection, the cockroaches were given MEL-infused water to drink ad libitum and then transferred into DD. The representative actogram records of locomotor activity are shown in Figure 7. Under LD conditions, 76% (n = 10) of the water-only insects were arrhythmic, whereas only 28.6% (n = 10) of MEL-supplied insects showed an arrhythmic pattern on the PaaaNAT-silencing background. The re-entrained rhythm suggested that MEL enhanced night phase activity, resulting in increased N/T ratio ( Figure 8E) against control group ( Figure 8B). Under DD, MEL supply significantly increased the amplitude (Table 4).

| Colocalization of aaNAT-ir, CYCir, and PDF-ir in the Br-SOGs of P americana
Involvement of aaNAT in the regulation of the circadian system was investigated by double labeling for aaNAT-ir and CYC-ir and PDF-ir ( Figure 9A-l). NAT-ir disappeared when aaNAT was reabsorbed with 1 µg/ml of antigen (data not shown). Both small and large aaNAT-ir neurons in the proximal frontoventral (Pfv) cluster in the optic lobe (OL) shared CYC-ir ( Figure 9A-F). aaNAT-ir and PDF-ir in this region partially overlapped: PDF-ir occurred at least in several small NAT-ir neurons while the large aaNAT-ir neuron was not PDF-positive ( Figure 9G-I). The colocalization of aaNAT-ir and PDF-ir was also observed in some of the landmarks for circadian structures such as the posterior optic commissure (POC) and anterior commissures (AOC) (Figure 9J-l). Hence, the aaNAT system is located on the circadian neural system. Figure 10A-D shows MEL-ir in the protocerebrum of P americana. Prominent large neurons in the pars intercerebralis showed MEL-ir and axonal branch was traced to the ocellus ( Figure 10E 2-8). Figure 10F,G shows PER-ir and DBT-ir in adjacent sections of 7 μm. These reactivities are at identical neurons at the apical optic chiasma. The same neurons at the dorsolateral protocerebrum showed no change in PER F I G U R E 5 Effect of RNA interference against aaNAT1. Double-plotted actograms (left) and χ 2periodograms (right) of representative cockroaches injected with dsRNA aaNAT (B). White and black bars above actograms indicate light and dark phases, respectively. The solid arrow indicates the time of dsRNA aaNAT injection. The open arrow in the actograms indicates the time of water injection (A). Rhythmicity was suppressed by the injection of the dsRNA aaNAT antigen localization between cytoplasm and nucleus, unlike D melanogaster. 40 aaNAT-ir was not completely observed in the same neurons as clock proteins-ir but it occurred at the close vicinities to PER-ir and DBT-ir as well as MEL-ir particularly in the protocerebrum.

| Effect of luzindole on locomotor rhythm
Whether the MEL action on locomotor activity is receptormediated or not was investigated by using luzindole, a competitive antagonist of MT 2 receptor. Representative double-plotted actogram records of locomotor activity are shown in Figure 11. After injection, 90% of the luzindoletreated cockroaches (n = 10) became arrhythmic in LD ( Figure 11B), while all the control cockroaches (n = 10) remained rhythmic ( Figure 11A). The luzindole injection significantly reduced locomotor activity during the night (Figures 12 and 13). In DD, 80% of the luzindole-injected insects showed no rhythmicity, with continuing reduced activity. Only 10% of the control insects lost their rhythm. The luzindole injection disrupted the circadian rhythm of locomotor rhythm and the activity level (Figures 12 and 13).  Note: Student' t test was used to compare the daily activity before and after the injection (P < .05).

| Cis-acting elements of aaNAT for circadian regulation
Putative Cis-acting elements involved in circadian regulation and metamorphic influence were explored within the genomic sequence of PaaaNAT1 available through analyzing the publicly accessible draft genome of P americana. The analyzed 3kb region upstream from PaaaNAT1 transcription start site (TSS), following Zhang et al, 41 contains numerous instances of the putative circadian E-box motif CANNTG, 42 PER repeat (PERR), and Pdp1-and Dboxes ( Figure 14; Table 5). Nine putative E-box motifs, with the first two, CAGCTG and AACGTG, are present within the first 200 bp of the TSS (Table 5). Four transcription binding sites for ecdysone receptor (ECR) are also detected.

F I G U R E 9
Colocalization of immunoreactivities of aaNAT and CYC in the brain including one of the circadian landmark structure, posterior optic commissure (POC) in P americana. iaaNAT-ir and immunoreactivities of circadian clock-related gene products were visualized using Alexa Fluor ® 488 (green fluorophore) and Alexa Fluor ® 555 (red fluorophore), respectively. Pictures on the right show merged images.

| DISCUSSION
The data put forth in this paper strongly support our hypothesis that aaNAT functions to integrate the circadian system and a behavioral rhythm via MEL. Several points are germane. First, MEL drinking synchronized locomotor rhythm both in LD and DD, enhancing the total bouts, amplitude, and nocturnality. Second, knocking down of aaNAT1 blocked the rhythmicity which was restored by MEL drinking. Similarly, the injection of a competitive MT2 antagonist, luzindole, abolished the rhythmicity but the injection of MEL after luzindole injection restored the rhythmicity. We infer that these points make up a convincing argument supporting that circadian system controls locomotor activity in P americana via MEL receptor 2 (MT2). As in A permyi, 33 PaaaNAT probably is a CCG. Transcriptional rhythmicity probably most strongly contributes to the generation of overt rhythm in P americana. Pigment-dispersing factor is a strong candidate as an output messenger of the circadian pacemaker in D melanogaster 43,44 and L maderae. 45,46 PDF peaks early in the day and troughs at early night in axon terminals in D melanogaster. 47 When examined in a per 0 and tim 0 mutant background, the expression of PDF was detected at constant high levels at axon terminals. 47 Although the molecular gating mechanism of the neurosecretion of PDF is not fully understood, we can pull in guidance from other works. Accumulations of mRNA encoding pdf occurred in D melanogaster heads 48 and were recorded by in situ hybridization in the CPM 47 and they remained constant in LD cycles. We suggest that PDF was not solely involved in mutual synchronization among multiple circadian neurons in this species because RNAi against aaNAT1 alone desynchronized the locomotor rhythm and could be resynchronized by MEL administration alone. Unlike PDF, MEL levels show a circadian rhythm with a peak at night in P americana. 30 Some other insect species also show rhythmic fluctuation in the hemolymph MEL concentrations, recorded in the cricket, Gryllus bimaculatus 49 and the damselfly, Ischnura verticalis. 50 RNAi knockdown studies showed that aaNAT1 is controlled by a negative feedback loop via regulatory binding of the E-box elements in A pernyi. 33 Hemolymph MEL concentrations and accumulations of mRNAs encoding iaaNAT1 occur in rhythms, as does iaaNAT1 enzyme activity occurred in rhythmicity in the Br-SOG of P americana 30 and A pernyi. 33 aaNAT-ir, HIOMT-I, and MEL-ir were colocalized with PER-ir and Cyc-ir (or Bmal1-ir) while PTTH-ir and MT2-ir colocalized in the brain of A pernyi. 33 Coincubation of the brain-suboesophageal ganglion-prothoracic gland complex with MEL led to release of PTTH. 28 The same neural network was confirmed here. aaNAT-ir was colocalized in Cyc-r(Bmal1-ir). Colocalization of aaNAT-ir occurred not only with CYC-ir (Bmal1-ir) but also with PDF-ir In L maderae, dissevering the connection between the accessory medulla (AMe) of the optic lobe (OL) and the central brain-induced arrhythmic locomotor activity, but subsequent regeneration of PDF-neural fibers between the Pfv cluster of the OL and median protocerebrum restored rhythmicity. 45 Similarly, transplantation of PDF-positive neurons in the Pfv of the OL into the AMe of an arrhythmic individual restored rhythmicity after restoration of the neural network from the transplant was attained. 51 Thus, PDF-neural fibers projected from the Pfv cluster connect the CPM and motor controller. These results demonstrate the occurrence of aaNAT-ir in the PDF-positive neural structure in the Pfv region of P americana and colocalization of aaNAT-ir with CYC-ir This supports the hypothesis that aaNAT underlies the output pathway connecting CPM with the motor controller.
There is evidence that the mutually coupled CMPs regulate the overt circadian rhythm, through either dawn/dusk oscillator coupling or bilateral coupling. [52][53][54] In digital assistant simulations, the coupled oscillator process is affected by the output of the rhythm. 55 Our observations revealed that MEL strengthens the coupling between oscillators because as the concentration of MEL increased, the amplitude increased.
The neural commissure coupling of bilateral CPMs has been demonstrated by rhodamine backfill into the AMe, immunostaining using positive fibers running through the anterior optic commissure (AOC), and the POC in L maderae. 56 These commissures are PDF-ir 46 but aaNAT-ir also occurs here. The effects of MEL on circadian rhythms of locomotor activity in P americana were observed in distinct ways: (1) exogenous MEL synchronized locomotor activity under both LD cycle and DD as MEL affected the locomotor rhythm in A domesticus 29 ; (2) rhythmicity in locomotor activity was disrupted by injection of dsRNA aaNAT1 ; (3) arrhythmia was reduced by per os MEL treatments after dsRNA aaNAT1 injections, confirming that the enzyme is an endogenous MEL synthesizing enzyme; (4) the enzyme activity is controlled by the circadian pacemaker, because aaNAT-ir and CYC-ir are colocalized in the same neurons; and (5) the N-terminal promotor sequence, the 3kb region upstream from PaaaNAT1 TSS, contains cis-acting elements, notably E-boxes, Pdp1, D-boxes, and PERR which may enhance the modulation of circadian system for transcription. In the latter context, direct coupling of the circadian processes to the E-boxes is an essential feature of rhythmic transcription. For example, the noncanonical E-box enhancer CACGTT drives mouse Period2 circadian oscillations in vivo. 57 The PER repeat (PERR) plays a role in transcriptional activation of the Drosophila period. 58 Other detected elements include homeobox transcription factors that contribute to intracellular timekeeping mechanisms responsible for daily rhythms in Drosophila 59 and cAMP response elements (CRE) which may function for a feedback regulation by MEL as in vertebrate systems. 60 Cis-regulatory elements linked to metamorphic influence like EcR are also detected. This is reasonable since aaNAT is mobilized upon metamorphosis and cuticle formation occurs with a circadian periodicity. 24 Two possible mechanisms for MEL's action enhancing synchronization exist. (1) MEL synchronizes rhythms via suppressing light phase activity through the photo-receptor pathway, or (2) MEL enhances coupling among constituent oscillators of the circadian system. The first possibility is however refuted, because MEL synchronized free-running rhythm.
The MEL-ir was located in the protocerebrum and ocellar nerve in P americana ( Figure 10). In most insect species, the  ocelli often suppress nocturnal activity. 61 5-HT-ir, HIOMT-ir, and iaaNAT enzyme activity have been characterized in the P americana brain. 62,63 The NAT activity occurs in the brains of fruit flies and cockroaches 30,64 and the head of the fruit fly synthesizes MEL. 65 We speculate that changes in NAT activity in the cockroach brain may regulate the cycle of MEL synthesis. aaNAT is a multi-substrate enzyme but the destruction and subsequent restoration of rhythmicity after dsRNA aaNAT1 injection and MEL drinking suggest to us that this enzyme regulates locomotor activity rhythm via MEL. We suggest that MEL in the cockroach may function as a circadian regulator. Evidence of MEL affecting τ has come from the work of Diez-Noguera, 66 who presented a model showing that τ depends both on "neutral elements" (NEs) and internal coupling. With substantial values of NEs, strong coupling tends to increase τ. The observed longer τ after dsRNA aaNAT1 injection and drinking higher MEL concentrations in P americana may reflect this situation. Higher levels of MEL in the brain may be responsible for rhythmic regulation of circadian activity. A competitive presynaptic MT antagonist, luzindole disrupts circadian rhythms in vertebrates, 67,68 and here we showed that luzindole inhibits the presynaptic receptor to control locomotor activity.
In conclusion, our results appear to provide the first evidence for the receptor-mediated effect of MEL action on circadian rhythms of locomotor activity in the insect brain. The insect aaNAT/MEL pathway appears to be similar to the CPM output pathway found in vertebrates and merits further study.