Evidence for a role of orcokinin-related peptides in the circadian clock controlling locomotor activity of the cockroach Leucophaea maderae

Orcokinins are a family of arthropod neuropeptides that have been identified in decapod crustaceans and, more recently, in two species of insects (Stangier et al.,1992; Bungart et al.,1995; Huybrechts et al.,2003; Pascual et al.,2004; Hofer et al.,2005). All orcokinins are highly conserved between species and, in the investigated crustaceans, occur in multiple isoforms within a given species (Bungart et al., 1995; Yasuda-Kamatani and Yasuda,2000; Skiebe et al.,2002; Huybrechts et al.,2003; Fu et al.,2005). The insect orcokinins identified to date, the tetradecapeptide Scg-OK1 from the locust Schistocerca gregaria and the dodecapeptide Blatella-OKL from the cockroach Blatella germanica,are closely related to crustacean orcokinins by identical N-terminal NFDEIDRSGF-sequence with crustacean Asn¹³-, Val¹³- and Ala¹³-orcokinins. Immunocytochemistry with an antiserum against Asn¹³-orcokinin revealed widespread distribution of orcokinins in the brain of polyneopteran insects, such as the cockroach Leucophaea maderae and the locust Schistocerca gregaria, but complete lack of immunostaining in the brain of endopterygote insects(Hofer et al., 2005). Labeling of brain interneurons and processes in the neurohemal retrocerebral complex in the cockroach and locust suggests that orcokinin-related peptides function both as hormones and as neuromodulators in these insects, as described for crustaceans.

In the present study, we have analyzed the role of orcokinin-related peptides in the circadian system of the cockroach Leucophaea maderae. As in the fruitfly Drosophila melanogaster, neurons with arborizations in the accessory medulla (AMe), a small neuropil at the anterior base of the medulla, constitute the master circadian clock in the brain of the cockroach (Reischig and Stengl,2003a) (reviewed by Homberg et al., 2003; Helfrich-Förster, 2005). The cockroach AMe consists of a core of dense nodular neuropil, embedded in and surrounded by coarse neuropil (Petri et al., 1995; Reischig and Stengl, 1996; Reischig and Stengl, 2003b). Six groups of somata near the AMe send neurites into the AMe neuropil (Reischig and Stengl, 2003b). Neurons of the distal tract connect the medulla to the AMe (Reischig and Stengl,1996; Reischig and Stengl,2003b; Petri et al.,2002) and are candidates for mediating light entrainment of the clock through photoreceptors of the compound eye(Roberts, 1965; Nishiitsutsuji-Uwo and Pittendrigh,1968). Neural connections between both brain hemispheres of L. maderae serve for bilateral coupling of the clocks and, in addition, for light entrainment of the clock through the contralateral compound eye(Page, 1978; Page, 1981; Page, 1983a). Commissural neurons between both AMae, which might mediate these functions, have been identified (Loesel and Homberg,2001; Reischig and Stengl,2002; Reischig et al.,2004). Immunocytochemical studies and injections of neuroactive substances suggest that several neuropeptides and γ-aminobutyric acid(GABA) play roles as neuromediators in the circadian system of the cockroach(Petri et al., 1995; Petri et al., 2002; Petri and Stengl, 1997; Schneider and Stengl, 2005). Neurons immunoreactive for β-pigment-dispersing factor (PDF) may serve as pacemakers of the clock; some of these neurons with projections to the lamina and selected parts of the midbrain might also serve as outputs of the clock(Reischig et al., 2004), while others with fibers in the anterior and posterior optic commissure transmit coupling information into the contralateral clock(Petri and Stengl, 1997; Reischig et al., 2004). GABA-immunoreactive neurons of the distal tract and allatotropin-related peptides in local interneurons of the AMe are part of the ipsilateral light entrainment pathway, as suggested by injection experiments(Petri et al., 2002). Finally,both GABA and PDF contribute to synchronize electrical activity among clusters of neurons in the circadian clock of the cockroach(Schneider and Stengl,2005).

Orcokinin-related peptides were detected immunocytochemically in the AMe of the cockroach (Hofer et al.,2005; Hofer and Homberg,2006). Detailed mapping showed that staining was present in about 30 neurons of the AMe (Hofer and Homberg,2006). Staining in AMe neurons with axonal fibers in the posterior optic commissure was particularly prominent, suggesting that orcokinin-ir neurons participate in coupling of the bilateral clocks. The present study shows that four pairs of orcokinin-ir neurons connect both AMae. Microinjections of Asn¹³-orcokinin into the vicinity of the AMe result in phase-dependent phase shifts in circadian wheel-running activity resembling the phase-shifting effects of light. These experiments are the first to demonstrate a physiological role of orcokinins in insects and suggest that these peptide(s) plays a role in light entrainment of the clock via the contralateral compound eye.

Male adult cockroaches Leucophaea maderae Fabricius 1792 were taken from crowded colonies at the University of Marburg, Germany. Animals were reared under 12 h:12 h light:dark (LD) photoperiod, at about 60% relative humidity, and a temperature of 28°C.

Animals (N=22) were anesthetized with CO₂ and fixed in a mounting device. A small window was cut into the head capsule above the left optic lobe to expose the brain. An amount of 1-3 nl Texas Red-dextran solution(TRed-D, dextran conjugated with Texas Red, 3000 kDa, lysine fixable,Molecular Probes Inc., USA; 0.1 mg ml^-1 in water) was pressure-injected with a microinjector (Microinjector 5242, Eppendorf,Germany) under stereomicroscopic control into one AMe with a glass capillary(Clark, Pangbourne Reading, England). The capillary was pulled to a pipette as used for patch clamp experiments, with a tip diameter of 1-3 μm. After the injection, the head capsule was closed with wax to allow intracellular transport of the dye in those insects that survived overnight.

The next day the injected brains of the cockroaches were removed and fixed for 4 h or overnight in 4% paraformaldehyde/7.5% saturated picric acid in sodium phosphate buffer (0.1 mol l^-1, pH 7.4) at room temperature. The brains were embedded in gelatine/albumin (4.8% gelatine and 12% ovalbumin in demineralized water) and postfixed in 8% formalin in sodium phosphate buffer (0.1 mol l^-1, pH 7.4). The brains were sectioned in the frontal plane at 40 μm thickness using a vibrating blade microtome (Leica,Nussloch, Germany). The brain sections were washed in Tris-buffered saline(TBS; 0.1 mol l^-1 Tris-HCl/0.3 mol l^-1 NaCl, pH 7.4)containing 0.1% Triton X-100 (TrX). They were preincubated in TBS with 0.5%TrX and 5% normal goat serum (NGS; Dako, Hamburg, Germany). Primary antiserum,anti-Asn¹³-orcokinin (provided by Dr H. Dircksen, Department of Zoology, Stockholm), was diluted at 1:4000 in TBS containing 0.5% TrX and 1%NGS and was applied to the sections for 18-20 h at room temperature. The sections were washed in TBS containing 0.1% TrX and incubated with Cy2-conjugated goat anti-rabbit antiserum (GAR; Dianova Hamburg, Germany,diluted 1:300) in TBS containing 0.5% TrX and 1% NGS for 1 h. Afterwards, the sections were thoroughly washed, and mounted on chromalum/gelatine coated microscope slides.

The anti-Asn¹³-orcokinin antiserum has been characterized(Bungart et al., 1994) by testing HPLC-fractions of different astacidean crustaceans with an enzyme-linked immunosorbent assay (ELISA). On cockroach brain sections,specificity of the antiserum was determined by liquid-phase preadsorption of the diluted primary antiserum with various concentrations (10^-4 mol l^-1 up to 10^-11 mol l^-1) of Asn¹³-orcokinin [NFDEIDRSGFGFN-OH(Stangier et al., 1992);Bachem, Heidelberg, Germany], before adding the combined solution to the preparation. Immunostaining was abolished after preadsorption with 1 nmol l^-1 Asn¹³-orcokinin for 18-20 h at room temperature.

Microscopic images were captured with a Zeiss microscope equipped with a 2-megapixel digital camera (Polaroid, Cambridge, MA, USA). A Leica TCS SP2 confocal laser scan microscope equipped with a spectrophotometric emission light detection system was used to evaluate the double-staining experiments. All scans were performed using a Leica HPX PL apochromate 40×/1.25 oil immersion objective. To exclude crosstalk artifacts, the specimens were scanned sequentially, and the detection ranges were separated as far as possible. Cy2 fluorescence was excited with the 488 nm line of an argon laser and detected between 505 and 525 nm. Texas Red fluorescence was excited with the 543 nm line of a helium/neon laser and detected between 585 and 625 nm.

All manipulations were performed in dim red light using a microinjector(see above). The experimental animals were removed at different circadian times from their running wheels and mounted in metal tubes. The animals were anesthetized with CO₂. A small window was cut in the head capsule,and one optic lobe was injected with 2 nl of Asn¹³-orcokinin solution or saline in the vicinity of the AMe, following the procedure described above for dextran injections. After the injection, the excised piece of cuticle was waxed back and the animal was returned to the running wheel. The time of injections did not take more than 15 min. The injection volume(150 fmol in 2 nl saline with blue food dye [FD+C blue No. 1; McCormick,Baltimore, MD, USA]) was controlled before and after injection with test injections in mineral oil. The concentration of 10^-4 mol l^-1 was chosen because similar doses had been effective in previous peptide injection experiments (Petri and Stengl, 1997; Petri et al.,2002). Concentrations of 10^-8 mol l^-1 and 10^-12 mol l^-1 orcokinin were tested at circadian time(CT) 13-15 of the circadian cycle, corresponding to the peak of the phase-response curve, to investigate dose-dependency of the response. Control injections consisted of 10% blue food dye in saline without orcokinin.

Circadian behavior was analyzed from cockroaches kept in constant darkness(DD) and constant temperature (28°C) and humidity (60%). Locomotor activity was recorded with running wheels(Wiedenmann, 1977) equipped with a magnetic reed switch. One revolution of the running wheel resulted in one impulse. Impulses were continuously counted by a computer over 1 min intervals and condensed and processed by a custom-designed PC-compatible software (developed by H. Fink, University of Konstanz). The data were plotted in double plot activity histograms. The free-running period τ and the induced phase shifts were estimated by converting the raw data into ASCII format. They were then merged into 30 min intervals and analyzed with Chrono II software (provided by Till Roenneberg)(Roenneberg and Morse, 1993)on a Macintosh computer. The χ²-periodograms were calculated with Tempus 1.6 (Reischig,2003), an add-in for Microsoft Excel, on an IBM-compatible PC. Data were evaluated from 110 of the 157 animals used. The remaining 47 cockroaches died after the operation. The free-running periods before and after injection were calculated by linear regression through daily ctivity onset. Changesa in τ(Δτ_after=τ-τ_before) were calculated,with periods estimated by regression through activity onsets and byχ ²-periodogram analysis(Enright, 1965; Sokolove and Bushell, 1978). Phase shifts were determined as time differences between the regression lines before and after injection extrapolated to the day after treatment. Phase delays were plotted as negative values and phase advances as positive values. Time on the x-axis of the resulting phase-response curve is shown as CT, with CT 12:00=activity onset= beginning of the subjective night. Daily activity onsets were determined by using Chrono II(Roenneberg and Morse,1993).

The behavioral data were merged into 2 h time intervals and the means and standard deviations (s.d.) were calculated for each bin. Changes of phases and periods in a given time interval were considered to be significantly different from zero, if the calculated 95% confidence interval of phase shifts and periods in a respective time interval did not contain the value zero. The differences of phase and period changes were tested for each pair of orcokinin and control injections with a two-tailed Student's t-test. In addition, phase shifts caused by either orcokinin or control injections were tested separately aginst all other respective values applying ANOVA with Tukey's post-hoc test. Significant differences in all cases were assumed at P<0.05. The statistical analyses were performed with SPSS 11.0 (Superior Performing Software Systems; SPSS Inc.) and Excel XP(Microsoft). Smoothed phase-response curves were produced with Excel.

To determine whether subsets of orcokinin-ir accessory medulla (AMe)neurons qualify for direct, monosynaptic coupling of both AMae, we injected Texas Red-conjugated dextran (TRed-D) as neuronal tracer into one AMe. Brains were immunostained with anti-orcokinin antiserum (N=21), and the AMe contralateral to the injection site and commissures in the midbrain were examined for double fluorescence (Figs 1, 2). These experiments showed that one bilateral pair of orcokinin-ir ventral neurons (VNe) and three pairs of ventro-median neurons (VMNe) provide direct connections between both AMae. To examine whether orcokinin influences circadian locomotor activity of the cockroach, we injected Asn¹³-orcokinin into the vicinity of one AMe of the cockroach. These experiments resulted in phase-dependent phase shifts in circadian locomotor activity resembling the phase-shifting effects of light.

After injection of approximately 2 nl TRed-D into one AMe(Fig. 1A), we found up to four colabeled orcokinin/TRed-D somata in the contralateral AMe(Fig. 1B-D, Fig. 3). These somata could be identified as one orcokinin/TRed-D colabeled VNe and three colabeled VMNe. Prominent orcokinin/TRed-D colabeled fibers projected via the lobula valley tract into the AMe and arborized preferentially in the internodular neuropil (Fig. 1C,D). Colabeled orcokinin/TRed-D fluorescent fibers were visible in the posterior optic commissure (Figs 2B, 3). These fibers connected both AMae without extending sidebranches to the midbrain, but the number of colabeled fibers could not be clearly determined. In contrast, no colocalization of orcokinin/TRed-D was found in the anterior optic commissure(Fig. 2A).

To investigate whether orcokinin plays a role as an input signal to the circadian clock, we examined whether the peptide influences circadian locomotor activity of the cockroach. Asn¹³-orcokinin was injected into the vicinity of one AMe at different circadian times, and locomotor activities of the free-running cockroaches were recorded before and after the injections (Fig. 4). Control injections with carrier solution alone (blue food dye in saline) did not cause significant phase shifts in circadian locomotor rhythm, except for a small but significant phase delay from CT21-23 (Table 1, Fig. 5B). Injections of orcokinin resulted in large and significant phase-dependent phase shifts in circadian activity at several circadian times. Maximal phase delays (-3.8 h_ct) occurred when orcokinin was injected at CT13, and maximal phase advances (2.2 h_ct) were observed at CT18(Fig. 5A). As judged from the 95% confidence intervals, significant orcokinin-dependent phase delays occurred from CT7-15 and significant phase advances from CT17-19(Table 1, Fig. 5). Phase shifts during the rest of the cycle were not significantly different from zero. Orcokinin-induced phase shifts from CT7-15 and from CT17-19 were also significantly different from phase shifts induced by control injections. Finally, orcokinin-induced phase shifts at CT21-23 were significantly different from control injections, but not from zero(Table 1, Fig. 5C). Orcokinin-induced phase shifts at CT13-15 were not significantly different from orcokinin-induced phase shifts from CT7-13, but from all other orcokinin-induced phase shifts. Similarly, orcokinin-induced phase shifts between CT17 and CT19 were not significantly different from orcokinin-induced phase shifts at CT19-7 and CT15-17.

The orcokinin-dependent phase shifts at CT13-15 were positively correlated with the dose of orcokinin-injections (Fig. 6). The phase delays decreased with decreasing amounts of injected peptide. Significant phase delays were caused by injection of 150 fmol[-2.92±0.81 h, mean ± s.d., CI=(-1.84, -4.00), N=5] and 1.5×10^-2 fmol [-1.70±0.99 h, mean ± s.d.,CI=(-2.73, -0.66), N=6] orcokinin. Phase shifts induced by injections of 1.5×10^-6 fmol orcokinin [-0.32±0.38 h, mean± s.d., CI=(-0.64, 0.00), N=8] were neither significantly different from zero nor from control injections(Fig. 6).

We did not find significant changes in the free-running periods at any CT before and after injection within the saline or peptide injected groups, nor between these groups. The observed effects were always small, included both lengthening (by maximally 0.48 h) and shortening (by maximally -0.63 h) of the period, and were independent of the time of injection during the circadian cycle. On average, the mean period (23.58±0.23 h, mean ± s.d., N=96) was altered neither by orcokinin [difference in period lengths before and after injection 0.00±0.19 h, mean ± s.d., CI=(-0.05,0.06), N=50] nor by control injections [-0.04±0.22 h, mean± s.d., CI=(-0.11, 0.02), N=46].

The present study provides strong evidence for a physiological role of orcokinin-related peptides in the circadian system of the cockroach L. maderae. Injections of Asn¹³-orcokinin into the vicinity of the AMe, a master circadian pacemaker in the cockroach brain, resulted in phase-dependent phase shifts in circadian locomotor activity that closely match the phase-response curve obtained for light pulses(Page and Barrett, 1989). These data suggest an involvement of orcokinin-related peptides in circuits relaying photic information to the circadian pacemaker. In a previous study,we showed that up to 30 neurons distributed in five cell groups of the AMe of L. maderae are immunoreactive with an antiserum against Asn¹³-orcokinin (Hofer and Homberg, 2006). Dextran injections into one AMe combined with orcokinin immunostaining revealed that four pairs of orcokinin-ir AMe neurons provide direct connections between the bilateral pacemakers. The additional arborizations of some of these neurons in the medulla proper suggest that these neurons may be responsible for the light-like phase shifts seen in the peptide injection experiments (see below).

For a well-synchronized circadian rhythm in behavior, the bilaterally distributed pacemakers in the insect brain have to be mutually coupled. In crickets, bilateral coupling of the clocks is relatively weak, but in the cockroach L. maderae, bilateral coupling is strong(Page et al., 1977; Wiedenmann and Loher, 1984; Ushirogawa et al., 1997), and is assumed to be mediated by direct neuronal connections between the AMae in the right and left brain hemisphere (Page,1983a; Page,1983b). Tracing studies showed that anterior neurons with cell bodies in two clusters termed MC I (4 cells) and MC II (35 cells) near the AMe connect both AMae directly (Reischig et al., 2004). The MC I neurons correspond to four ventral neurons(VNe), and the MC II cells are identical with the ventromedian neurons (VMNe)of Reischig and Stengl (Reischig and Stengl, 2003b). Three of the four VNe that connect both AMae directly are PDF-ir (Reischig et al.,2004). The neurons project via the anterior (two neurons)and posterior (one neuron) optic commissures, innervate the internodular and shell neuropils of the AMe, and send a fan of fibers along the distal surface of the medulla toward the first optic chiasm and lamina. Neurons of this morphological type were found to be unresponsive to light stimuli in intracellular recordings (Loesel and Homberg, 2001). Together with the non-photic phase-response curve obtained by PDF injection, Petri and Stengl(Petri and Stengl, 2002) and Reischig et al. (Reischig et al.,2004), therefore concluded that the three PDF-ir VNe transmit phase information to the contralateral pacemaker. We show here that one contralaterally projecting VNe is orcokinin-ir. Since colocalization of PDF and orcokinin occurs neither in the anterior nor in the posterior optic commissure (Hofer and Homberg,2006), this neuron has to be the fourth non-PDF-ir VNe with projections to the contralateral AMe.

Page (Page, 1978; Page, 1983a; Page, 1983b) proposed that the bilateral optic lobe pacemakers of L. maderae not only exchange phase information, but also receive entraining light signals from the contralateral eye. The pathway for contralateral light entrainment is most likely provided by the second group of commissural neurons, the VMNes. Neurons of this group have contralaterally projecting fibers in the posterior optic tract; they invade the internodular and shell neuropil of the AMe and, in contrast to VNes, have tangential arborizations in a median layer of the medulla(Reischig et al., 2004). Interestingly, neurons of this type are highly sensitive to light stimuli(Loesel and Homberg, 2001). Ensemble reconstructions of orcokinin-ir VMNes(Hofer and Homberg, 2006) and,in this study, tracer injections combined with immunocytochemistry, clearly show that three of the 35 contralaterally projecting VMNes are orcokinin-ir. We, therefore, suggest that these neurons transmit light information to the contralateral AMe, and that release of an orcokinin-like substance is the output signal of these neurons. The neurotransmitter of the remaining 32 VMNes is not known.

The colabeling of dextran-injected and orcokinin-immunostained fibers in the posterior but not in the anterior optic commissure suggests that the orcokinin-immunolabeled VNe, like the VMNes, traverse the brain midline via the posterior optic commissure. Since dextran transport through the anterior optic commissure was often only of low intensity (see also Reischig et al., 2004),however, we cannot completely exclude the possibility, that the orcokinin-ir VNe projects via the anterior optic commissure but was not detected in the dextran injections.

Microinjections of orcokinin into the vicinity of the AMe resulted in phase delays and phase advances of the circadian wheel running activity. The effects of peptide injections were dose dependent and were significantly different from control injections. The phase-response curve observed after orcokinin injections is similar to the biphasic phase-response curve after light-pulses(Page and Barrett, 1989)(Fig. 7), as well as to phase-response curves after GABA- and allatotropin injections(Petri et al., 2002).

Our recent immunocytochemical study revealed widespread occurrence of orcokinin-related peptides in different cell types, in addition to the commissural neurons, in ventroposterior neurons (VPNe), in distal frontoventral neurons (DFVNe), in ventral neurons (VNe), and in median neurons(MNe) of the AMe (Hofer and Homberg,2006) [for nomenclature see(Reischig and Stengl, 2003b)]. Therefore, we suggest that orcokinin serves a multitude of functions in the circadian system, possibly including a role in light entrainment, output pathways and internal synchronization.

Interestingly, the maximal phase advance observed after orcokinin injection was not as high as that observed for GABA- and allatotropin injections. Furthermore, the peaks of the orcokinin-induced phase-response curve were broader (with respect to the range of significant phase shifts) than the peak phase shifts after light pulses, GABA, allatotropin and PDF injections(Page and Barrett, 1989; Petri and Stengl, 1997; Petri et al., 2002). These differences might be explained as follows. The AMae appear to be connected by two types of orcokinin-ir neurons, which are part of functionally different circadian coupling pathways: a pathway transmitting phase information (one orcokinin-ir VNe) and a pathway transmitting light information (three orcokinin-ir VMNe). Phase information appears to be carried by PDF-ir VNe, and if the orcokinin-ir VNe plays a similar role, a mixture of light-like (three neurons) and non-light-like (one neuron) phase responses should occur after orcokinin injections, which would then lead to smaller and broader peaks in the phase-response curve compared to those observed after light pulses, GABA or allatotropin injections (Page and Barrett, 1989; Petri et al.,2002). This assumption is underlined by the fact that orcokinin immunoreactivity occurs in five of the six morphologically distinguishable and obviously also functionally divergent neuron groups associated with the AMe. Nevertheless, after injection of orcokinin, the biphasic light-like phase-response effects appear to dominate over an additionally expected overlapping non-photic phase-response curve.

Combining the results of this work with previous immunocytochemical studies(Hofer et al., 2005; Hofer and Homberg, 2006), we propose that orcokinin has the following functions in individual AMe neurons:(1) Orcokinin-ir neurons (VMNe, VPNe and MNe) might play a role in the light entrainment pathway; (2) orcokinin-ir neurons (VNe) might form output pathways to different effectors of the clock; (3) one pair of orcokinin-ir neurons(VNe) may transmit coupling information to the contralateral AMe. In summary,our work supports earlier studies suggesting that orcokinin plays multiple roles in the cockroach circadian system(Hofer et al., 2005; Hofer and Homberg, 2006), and presents the first direct evidence for a physiological function of this peptide in insects.

List of abbreviations
 
• AMe

  accessory medulla

 
• CT

  circadian time

 
• DFVNe

  distal frontoventral neurons of the AMe

 
• ELISA

  enzyme linked immunosorbent assay

 
• GABA

  γ-aminobutyric acid

 
• GAR

  goat anti-rabbit antiserum

 
• MNe

  median neurons of the AMe

 
• NGS

  normal goat serum

 
• PDF

  β-pigment-dispersing factor

 
• TBS

  Tris-buffered saline

 
• TRed-D

  Texas Red-conjugated dextran

 
• TrX

  Triton X-100

 
• VMNe

  ventro-median neurons of the AMe

 
• VNe

  ventral neurons of the AMe

 
• τ

  free-running period

 
• Δφ

  phase delay

 
• h_ct

  circadian hours

We are grateful to Dr Franz Grolig for advice in confocal laser scan microscopy and to Sandra Söhler for introductions into the behavioral assays and into the peptide injection technique. We thank Dr H. Dircksen for providing the anti- orcokinin antiserum and Dr Thomas Reischig for providing the 3D-model of the cockroach brain. This work was supported by Deutsche Forschungsgemeinschaft (DFG) grant HO/950-9.