Perfusion with cGMP analogue adapts the action potential response of pheromone-sensitive sensilla trichoidea of the hawkmoth Manduca sextain a daytime-dependent manner

Males of the nocturnal hawkmoth Manduca sexta detect the pulsatile release by female moths of the sex pheromone blend using specialized trichoid sensilla. The pheromone-sensitive sensilla are located in large numbers at the leading edge of the antenna and are each innervated by two olfactory receptor neurons (ORNs) (Sanes and Hildebrand,1976; Keil, 1989; Lee and Strausfeld, 1990). Both ORNs are sensitive to different components of the pheromone blend. Only one of the ORNs, which elicits the larger action potentials (APs), responds to the main sex pheromone component bombykal (BAL)(Starratt et al., 1979; Tumlinson et al., 1989; Kaissling et al., 1989; Dolzer et al., 2001; Kalinová et al.,2001).

Much is already known about the peripheral signal transduction processes underlying the detection of pheromone by the male moth. Pheromone-dependent rises in inositol triphosphate (IP₃)(Breer et al., 1990) cause an influx of Ca²⁺ into insect ORNs(Stengl, 1994). This triggers the opening of Ca²⁺-dependent ion channels involved in the generation of pheromone-dependent receptor potentials(Stengl, 1993; Stengl, 1994; Zufall and Hatt, 1991; Zufall et al., 1991). In contrast, little is known about the modulation of the pheromone transduction cascade leading to adaptation and sensitization. Previous studies in Manduca sexta distinguished desensitization from short-term and long-term adaptation (Dolzer et al.,2003). Desensitization is the decline in excitation, as seen during a phasic-tonic response to a stimulus of long duration, while short-term adaptation is the rapidly (within seconds to minutes) reversible reduction of sensitivity due to prior stimulation. Long-term adaptation is the more slowly occurring and longer persisting (for minutes to hours) reduction of sensitivity due to previous strong stimulation(Ziegelberger et al., 1990; Marion-Poll and Tobin, 1992; Boekhoff et al., 1993; Stengl et al., 2001; Dolzer et al., 2003). Increasing evidence supports the involvement of cyclic guanosine monophosphate(cGMP) in long-term olfactory adaptation and regulatory processes in moths. After strong pheromone stimulation, cGMP levels rise slowly in antennae and antennal homogenates of Antherea polyphemus and Bombyx mori,remaining elevated for nearly 30 min and, thus, matching the time course of long-term adaptation (Ziegelberger et al.,1990; Boekhoff et al.,1993). In the silkmoth Bombyx mori the perfusion of the sensillar lymph with dibutyryl guanosine 3′,5′-cyclic monophosphate caused a suppression of the sensillar potential and a reduction of the AP response in pheromone-dependent sensilla, mimicking a state of adaptation (Redkozubov,2000).

In tip recordings of trichoid sensilla of Manduca sexta, adapting pheromone stimulation shifted the dose-response curves to higher stimulus intensities (Dolzer et al.,2003). In other moths, as for Manduca sexta, this shift is also larger for the AP than for the sensillar potential response,suggesting the presence of more than one adaptation mechanism(Zack, 1979; Kaissling et al., 1986; Kaissling et al., 1987; Dolzer et al., 2003). At least one adaptation mechanism acts on the first step of the olfactory transduction cascade, the generation of the receptor potential, which is assumed to occur in the outer dendritic segment of the ORNs(Kaissling and Thorson, 1980; Stengl et al., 1998; Dolzer et al., 2003). An additional adaptation mechanism apparently acts on the second step of the transduction cascade, the transformation of the amplitude-modulated receptor potential into frequency-modulated APs. The APs are apparently elicited in the soma- or axon-hillock region, as indicated by the polarity of APs and the current necessary to elicit them, and not at a dendritic location, as suggested by other experiments (Kaissling and Thorson, 1980; de Kramer,1985; de Kramer et al.,1984; Dolzer et al.,2001). The molecular basis and the behavioral consequences of these different mechanisms of insect olfactory adaptation are still not fully understood.

Previous studies have revealed that the responsiveness of male moths to pheromone is daytime-dependent. Males of the nocturnal moths Trichoplusia ni and Agrotis segetum showed a distinct daily rhythm in their response to pheromone, with a maximum response during the scotophase(Linn et al., 1996; Rosén et al., 2003). Also, daytime-dependent changes in responsiveness to pheromone persist under constant conditions so are controlled by an endogenous circadian clock(Baker and Cardé, 1979; Rosén et al., 2003). Not only male responsiveness, but also calling behavior and pheromone release of female moths, show a diurnal distribution with a maximum at the end of the scotophase for Manduca sexta and the middle of the scotophase for Agrotis segetum (Itagaki and Conner, 1988; Rosén,2002). For Manduca sexta it was shown that peaks in the calling behaviour of female moths and male flight activity are correlated during the scotophase (Sasaki and Riddiford, 1984). In addition, Spodoptera littoralismoths mated significantly less when males and females were raised in different light:dark cycles out of phase (Silvegren et al., 2005). This indicates that circadian clocks rule rhythmic mating preference in male and female moths and that photoperiod and pheromones synchronize the mating behavior of both sexes(Silvegren et al., 2005).

To determine whether daytime-dependent changes in pheromone sensitivity already occur at the periphery in male Manduca sexta moths and to determine whether cGMP-dependent mechanisms of long-term adaptation might be employed, we applied the membrane-permeant cGMP analogue 8-bromo guanosine 3′,5′-cyclic monophosphate (8bcGMP) in extracellular tip recordings from trichoid sensilla at two Zeitgebertimes (ZTs; ZT 1-4 and ZT 8-11). ZT 1 is at the beginning of the day (lights on at ZT 0) when the nocturnal moths switch from their active to their inactive phase, and ZT 8 is at the middle of the day when the moths are resting. Pheromone responses to the main pheromone component BAL and spontaneous APs of unstimulated ORNs of the hawkmoth Manduca sexta were investigated.

Hawkmoths Manduca sexta (Johannson) (Lepidoptera: Sphingidae) were raised from eggs, feeding on an artificial diet [modified after Bell and Joachim (Bell and Joachim,1976)]. Male pupae were kept under a long-day photoperiod (17 h:7 h L:D) at 24-27°C and 40-60% relative humidity and were isolated 1 day before emergence, gently cleaned with 70% ethanol, and allowed to hatch without contact with pheromone. After their second dark phase, 30 min prior to the start of each recording, the adults were taken out of the isolation boxes and fixed in a Teflon™ holder. Thus, moths at both ZTs spent the same time in the Teflon™ holder. The flagellum of the right antenna was immobilized with dental wax (Boxing wax, Sybron/Kerr, Romulus, MI, USA), and the 15-16 most apical annuli were clipped off. A glass electrode filled with haemolymph Ringer (Kaissling,1995), which served as the indifferent electrode, was inserted into the flagellar lumen. To stop the first annuli from drying out the end of the antenna was sealed with ECG electrode gel (PPG, Hellige, Freiburg,Germany) afterwards. The tips of long trichoid sensilla from the apical row on the second remaining annulus were clipped off using sharpened forceps. The recording electrode, filled with sensillum lymph Ringer(Kaissling, 1995), was slipped over one sensillum. To minimize contributions of the electroantennogram, we recorded from an annulus close to the tip of the haemolymph electrode. The connection to the amplifier inputs was established with Ag/AgCl wires immersed in the electrolytes. Signals were amplified about 200-fold in a custom-built amplifier (0 Hz-2 kHz, input impedance 10¹² Ω), and passed through an anti-aliasing filter with a cut-off frequency of 2 kHz (900C/9L8L,Frequency Devices, Haverhill, MA, USA). For data acquisition, a Digidata 1200 B digitizer and pCLAMP software (version 8) (Axon Instruments, Union City, CA,USA) were used. The electrophysiological signal and a high-pass filtered equivalent (cut-off frequency 5 Hz) were continuously recorded on a strip chart recorder (EasyGraf, Gould, Valley View, OH, USA). Voltage polarity is given with the sensillum lymph electrode in reference to the haemolymph electrode. To be able to record in the photophase we compared the beginning of the day with the middle of the day to search for ZT-dependent effects of 8bcGMP. All experiments were performed with room lights switched on, starting either at the beginning of the day, 1 h after lights on at ZT1, or at the middle of the day, at ZT8.

Drug was applied to pheromone sensilla during tip recordings by perfusion with water-soluble agents, as first suggested by Kaissling et al.(Kaissling et al., 1991). Long-term recordings from intact animals revealed no damage of the sensilla when continued for up to several days(Dolzer et al., 2001). So we allowed the sensillum lymph ringer to passively perfuse the sensillum lymph cavity and did not apply any pressure or suction. In recordings of stimulated trichoid sensilla, 8bcGMP diluted in sensillum lymph ringer was applied over the recording electrode. Therefore, we used an altered sensillum lymph ringer with a concentration of 10 mmol l^-1 8bcGMP. This application method is further referred to as sensillum lymph perfusion. Biogenic amines injected into the head capsule near the base of the antenna influenced oscillations of the transepithelial potential (TEP) of trichoid sensilla(Dolzer et al., 2001). Thus,the agents were transported into the antenna after injection. In recordings of the spontaneous activity of ORNs, 8bcGMP diluted in haemolymph ringer was injected through a hole in the head capsule, which was pierced with a syringe needle approximately 1 mm dorsocaudal to the right antennal base. We injected a minimum of 3 μl of 10 mmol l^-1 8bcGMP solution and a maximum of 5 μl of 100 mmol l^-1 solution, resulting in a concentration of 30-500 nmol l^-1 8bcGMP. Because an adult moth contains approximately 1 ml of haemolymph the final 8bcGMP concentration in the haemolymph was between 30-500 μmol l^-1. This application mode is further referred to as haemolymph injection. The ringers used for sensillum lymph perfusion and the respective controls were prepared with N-(2-hydroxyethyl)-piperazine-N′-(2-ethane sulfonic acid) (Hepes; all chemicals from Sigma, Deisenhofen, Germany); the ringers used for haemolymph injection were prepared with a phosphate buffer(monopotassium phosphate). The pH was adjusted to 6.5, and osmolality was adjusted to 475 mosmol l^-1 with mannitol for sensillum lymph ringer and to 450 mosmol l^-1 for the haemolymph ringer.

All recordings were performed at room temperature (18-23°C). Charcoal-filtered and moistened air was permanently blown over the preparation through a glass cartridge (13 l min^-1). The air stream could be redirected through cartridges containing a piece of filter paper (about 1 cm²) loaded with synthetic bombykal (E,Z-10,12-hexadecadienal; BAL)generously provided by T. Christensen (Tucson, AZ, USA). The air stream was switched between the cartridges using solenoid valves (JFMH-5-PK and MFH-5-1/8, Festo, Esslingen, Germany) controlled by the computer. Doses of 10μg BAL dissolved in n-hexane (Merck, Frankfurt, Germany) were applied to the filter papers (10 μl per paper), and the solvent was allowed to evaporate. Stimulus intensity is always given in terms of the BAL dose applied to the filter paper. The cartridges were placed in the outlet in a distance of 4.5-6 cm from the recording site about 25 s prior to stimulation. Stimuli of 10 μg BAL and 50 ms duration were applied every 5 min during a recording session of 3 h. A suction tube of 10 cm diameter was placed below the animal for rapid removal of the pheromone component after stimulation, and to avoid uncontrolled stimulation due to BAL leaking out of the stimulus cartridges. Between the recording sessions, the cartridges were stored at-20°C in individual glass scintillation vials. Control cartridges loaded with hexane alone were prepared and treated the same way. A set of stimulus cartridges was used for 5-10 recording sessions; the control cartridges were used for up to 30 recordings.

including 4-point smoothing. For analysis of the APs(Fig. 1C), the low pass-filtered trace was subtracted from the original response. This pseudo-high-pass filtering procedure, in contrast to actual high-pass filters,does not distort the shape of the APs, and therefore allows the analysis of their amplitude and waveform. The time and positive amplitude of every AP occurring during the 5161 ms-long sweep recorded during the stimulations was counted. When APs of both amplitude classes were observed during the response,they were analyzed separately. The AP response was characterized by: (1) the peak frequency computed from the first five interspike intervals (AP frequency; Fig. 1C); (2) the latency between the beginning of the DC response and the occurrence of the first AP (AP latency; Fig. 1B);and (3) the positive AP amplitude. The peak frequency represented the phasic part of the phasic-tonic AP response and was calculated according to f=N×1000/∑ISI [N=number of interspike intervals (ISI)]. Due to the high variability of the evaluated response parameters of each recording, the parameters were normalized to the first response of a recording. Subsequently the data for each parameter was binned to intervals of 5 min and the mean was plotted against the time of recording.

The data of the evaluated parameters of 8bcGMP and associated control recordings in each time slot were compared using the Student's t-test. For statistical analysis of changes in several parameters over the recording duration, its time course was divided into three intervals and then analyzed using a one-way ANOVA followed by the Tukey HSD post-hoc test. For analysis of the AP amplitude reduction, action potentials of three consecutive responses within an interval (from 1 min before to 2 min after the stimulus) were binned to 10 ms intervals and plotted against time. To measure changes in the amplitude reduction, the ratio of the minimal and maximal AP amplitudes was calculated and the normalized parameter for each recording plotted against the time of recording. To evaluate changes in the AP frequency or distribution in time, the AP response characteristics were analyzed using post-stimulus-time histograms; the AP responses were added up, binned to 10 ms intervals and plotted over time, with t=0 being the start of the DC response.

In extracellular tip recordings of trichoid sensilla of the hawkmoth Manduca sexta, we investigated daytime-dependent effects of 8bcGMP perfusion on stimulated ORNs. The recordings were performed over the course of 3 h. Recordings started either at the beginning of the day at ZT 1, when the nocturnal moths switch from their active to their inactive phase, or at the middle of the day at ZT 8, when the moths are resting. Slow,amplitude-modulated BAL-dependent sensillar potentials with superimposed,fast, frequency-modulated APs were quantified(Fig. 1A-C). BAL stimuli were applied in a non-adapting protocol with an interstimulus interval of 5 min(Fig. 1E). In addition, the effects of 8bcGMP on spontaneous APs of unstimulated trichoid sensilla were examined.

To search for cGMP-dependent modulation of BAL responses, we evaluated various parameters of the BAL-dependent sensillar potential and AP response(Fig. 1). Parameters describing the sensillar potential are the maximum sensillar potential amplitude, the initial slope characterizing the first half, and fit τ describing the second half of the rising phase of the sensillar potential(Fig. 1A,B). Parameters describing the phasic part of the AP response are the AP frequency computed over the first five interspike intervals, the AP latency in relation to the onset of the sensillar potential, and the stimulus-dependent amplitude reduction (Fig. 1B-D). The beginning and end of the recordings were compared(Fig. 2). In both time slots from ZT 1-4 and ZT 8-11, tip recordings with perfusion of the sensillar lymph with 10 mmol l^-1 8bcGMP applied over the recording electrode were compared to control recordings without cyclic nucleotides. In control recordings at both ZTs, no changes in the evaluated parameters could be observed throughout the 3 h of recordings (Figs 3, 5). In the presence of 8bcGMP the AP frequency was continuously decreased, and in a few recordings the latency between the onset of the sensillar potential and the occurrence of the first AP increased (Figs 2, 4). All other parameters that describe the AP and the sensillar potential response remained constant over the time of recording (Fig. 4). For recordings from ZT 8-11 (Fig. 5), significant differences were found in both the comparison of the AP frequency of 8bcGMP and control recordings (Student's t-test for independent samples, P<0.01) and also within the time course of the AP frequency (ANOVA and Tukey HSD post-hoc test; α=0.01, P<0.01). No significant difference was found between intervals 2 and 3 of the time course. In contrast no significant differences were found for the sensillar potential amplitude. Due to the high variability for the AP latencies no significant difference for recordings with or without 8bcGMP was found (data not shown).

As in recordings from ZT 8-11, the AP frequency of recordings from ZT 1-4(Fig. 5) showed a significant decrease when 10 mmol l^-1 8bcGMP was included in the recording electrode (Students t-test for independent samples, P<0.01), but no significant difference was found between groups 2 and 3 of the time course (ANOVA and Tukey HSD post-hoc test;α=0.01, P<0.01). The sensillar potential amplitude and the evaluated parameters of control recordings(Fig. 5) did not change over time (N=9). Comparison of the AP frequency depletion between both time slots revealed a 12% stronger decrease for recordings at ZT 8-11.

It was next determined whether there are any daytime-dependent differences in the distribution of APs in response to stimulation with BAL in controls or in the presence of 8bcGMP. To analyse the distribution of APs occurring during the first 1000 ms, for responses from the beginning (0-20 min), middle (80-100 min) and end (160-180 min) of the recordings, the counted APs were binned to 10 ms intervals and plotted as post-stimulus-time histograms (Figs 6, 7). In control recordings from ZT 1-4 during the first 100 ms of the responses the response peak occurred later and the number of APs in the first 100 ms decreased significantly over the 3 h recording, leading to a slightly less phasic characteristic of the responses (Fig. 6, insert in C). Under the influence of 8bcGMP a weak decline in the number of APs over the course of the 3 h recording was also recognizable(Fig. 6A-C); owing to high variability, however, the decrease over the first 100 ms of the responses was not significant (Fig. 6C,insert).

In control recordings from ZT 8-11 the distribution of APs changed to a more tonic spike pattern (Fig. 7), which was much more pronounced than in control recordings at ZT 1-4. Also, the number of APs over the first 100 ms of the BAL responses significantly decreased over the course of the recordings. The numbers of APs were much more strongly decreased in control recordings at ZT 8-11 than at ZT 1-4 (Fig. 6C, Fig. 7C, inserts). In the presence of 8bcGMP the change to a more tonic spike pattern, as well as the decline in the number of APs in the first 100 ms of the BAL responses, were even more pronounced than in the controls at ZT 1-4 and at ZT 8-11.

Characteristic for responses to BAL stimuli of higher dose is the reduction of the AP amplitude (Fig. 1C,D), which lasts for seconds to minutes. Especially in responses to strong stimuli with BAL concentrations higher than 1 μg per filter paper, the reduction of the peak-to-peak amplitude in the phasic portion is very strong and, in many cases, after about 150 ms the APs cannot be distinguished from noise. When comparing the beginning and end of a recording(Fig. 2), or the positive amplitude of APs in responses during an interval of 1 min before to 2 min after stimulation, several effects of 8bcGMP on the amplitude reduction can be observed (Fig. 8A). At the beginning of the recordings the amplitude reduction is very strong, but it gradually attenuates over the time of the recording until the reduction is only very weak and transient in the presence of 8bcGMP(Fig. 8A). In control recordings no strong changes in the amplitude reduction are recognizable(Fig. 8A). In addition to the cGMP-dependent attenuation of the decline of the AP amplitude, the kinetics of the amplitude attenuation also changed (data not shown). At the beginning of the recording the APs did not return to their prestimulus amplitude until 2 min after the stimuli had been applied. Under the influence of 8bcGMP the recovery phase was gradually attenuated until it disappeared. Also the minimal positive AP amplitude linearly increased over the course of the recording(data not shown). Furthermore, the normalized ratio of the minimal and maximal positive AP amplitude (Fig. 8B)(as a mean of the strength of the amplitude reduction) shows that 8bcGMP perfusion attenuates the AP amplitude reduction more strongly than in the controls at both ZTs. In addition, there are differences between both ZTs in the control recordings and in those with 8bcGMP perfusion. At ZT 8-11 with application of 8bcGMP, an attenuation of the amplitude reduction was recognizable in all of the recordings, whereas in the associated controls the strength of the amplitude reduction was reduced in only one recording over the entire time course of the recordings. At ZT 1-4 the recordings were more unequally distributed. In the control recordings there was higher variance compared to recordings at ZT 8-11, which also increased with the duration of the recording. Under the influence of 8bcGMP 2 populations of recordings were recognizable. In one population that describes the majority of the recordings,the amplitude reduction increased only slightly and this increase did not start until 90 min after the beginning of the recording. In the second population a very strong attenuation of the amplitude reduction was present from the beginning. Changes in AP amplitudes of spontaneous APs between BAL responses were not found.

In addition to the effects of 8bcGMP on pheromone responses, 8bcGMP also had an effect on the waveform of spontaneous APs of unstimulated sensilla trichoidea of isolated male moths that had never experienced pheromone. Changes in the waveform of both AP classes were observed after 6 of 8 haemolymph injections of 30-500 nmol 8bcGMP for several hours. In one recording (Fig. 9) the time course of the AP waveform could be monitored almost continuously for 6 h after the injection at the base of the antenna. The waveforms of both AP classes were influenced independently, suggesting that the 8bcGMP effect took place within the individual ORNs (N=6). After injection of 8bcGMP, two aspects of the waveform were altered: the peak-to-peak amplitude increased and the negative phase of the APs was prolonged. These effects then reverted and reappeared later, suggesting the presence of feedback-coupled mechanisms that were triggered by 8bcGMP injection. A consistent change in the frequency or burst behaviour of the APs was not observed. In recordings of stimulated olfactory neurons no effects of 8bcGMP on the waveform of spontaneous APs between the stimuli were found (data not shown).

Using long-term tip recordings from pheromone-sensitive trichoid sensilla of the hawkmoth Manduca sexta, we tested whether cGMP affects responses to stimulation with the main pheromone component BAL in a daytime-dependent manner. Perfusion of the sensillum lymph with 10 mmol l^-1 of the membrane-permeable cGMP analogue 8bcGMP reduced the AP frequency of ORNs in response to BAL stimulation in a daytime-dependent manner. The reduction was 12% stronger at the middle of the day (ZT 8-11) than in recordings starting at ZT 1. Also, in control recordings daytime-dependent differences in the AP distribution of the pheromone responses were observed. Pheromone responses at ZT 8-11 became increasingly tonic and fewer APs were elicited within the first 100 ms of the responses. At ZT 8-11 only, 8bcGMP also strongly decreased the number of APs occurring during the first 100 ms of the response, leading to a more tonic AP response. The 8bcGMP perfusion also attenuated the reduction of the AP amplitude in most of the recordings from ZT 8-11 and in a subpopulation of recordings from ZT 1-4. Furthermore, the injection of 8bcGMP into the haemolymph altered the waveform of spontaneous APs of unstimulated ORNs. The sensillar potential in contrast remained unaffected at all ZTs.

It was known from extracellular tip recordings from different moth species that strong, long or high frequency pheromone stimuli can cause desensitization, short- or long-term adaptation to pheromone responses depending on stimulus strength and time course(Zack, 1979; Dolzer et al., 2003; Marion-Poll and Tobin, 1992). In Manduca sexta it was shown that, depending on BAL stimulus, length and strength parameters of the sensillar potential and AP response adapt differentially (Dolzer et al.,2003). After adapting to pheromone stimuli, the time of rise of the sensillar potential did not increase after short-, only after long-adapting stimuli, whereas both stimulation schemes shifted the dose-response curves of the sensillar potential amplitude, as well as the initial slope of its rising phase, to higher stimulus concentrations. The shift in the dose-response curve of the AP response was larger compared to the shift in dose-response curve of the sensillar potential response, indicating that depending on the properties of the stimulus unknown mechanisms of adaptation occur at different levels of the transduction process(Dolzer et al., 2003).

Biochemical experiments implicated that rises in cGMP concentration are involved in at least one mechanism of long-term adaptation, because strong and long pheromone stimuli caused delayed and sustained rises in intracellular cGMP concentrations in antennal homogenates of Antherea polyphemusand Bombyx mori. These increased levels of cGMP concentration matched the time courses of long-term adaptations(Ziegelberger et al., 1990; Boekhoff et al., 1993). In immunocytochemical experiments and in situ hybridizations it could be shown that at least subpopulations of pheromone-sensitive ORNs of male Manduca sexta upregulate cGMP concentrations after minute-long (but not seconds-long) exposure to female pheromones(Stengl et al., 2001). These cGMP rises were augmented via exposure to NO donors in the presence of a pheromone-inducible NO-synthase-like enzyme in ORNs, but not viastimulation of a soluble guanylyl cyclase(Stengl and Zintl, 1996; Stengl et al., 2001). Thus,adapting, long and strong pheromone stimuli, as occur in close vicinity to the calling female, cause long-lasting rises in cGMP concentrations. But it remained unknown which antennal targets are affected via cGMP and whether rises in cGMP concentration underlie at least one mechanism of long-term olfactory adaptation.

In the present study we have shown in tip recordings of trichoid sensilla of the hawkmoth Manduca sexta that brief but strong pheromone stimuli applied in a non-adapting protocol with interstimulus intervals of 5 min cause no changes in the pheromone response, either in the sensillar potential or in the AP response. Only during long-term exposure to the cGMP analogue 8bcGMP was the AP frequency decreased over the 3 h of the recordings; the sensillar potential response remained unchanged. This selective decrease in the AP frequency distinguishes one of the different forms of adaptation in Manduca sexta for the first time. Interestingly, injection of dibutyryl guanosine 3′,5′-cyclic monophosphate into the antennae of Bombyx mori decreased the AP frequency as well as the sensillar potential amplitude (Redkozubov,2000). The recording methods and the cGMP analogues employed differed between both moth species, as well as the stimulation scheme. While only very brief strong pheromone stimuli were applied in our recordings,Redkozubov employed either 10 times longer and stronger pheromone stimuli, or presented continuous pheromone stimuli of a lower dose. Most likely, these differing pheromone stimulation protocols were responsible for the differences observed.

The decrease in the AP frequency, the decrease in the number of APs in the first 100 ms of the response, and the observed cGMP-dependent slowdown of the negative phase of spontaneous APs, could all be explained by a closure of K⁺-channels. Also, the increase in the peak-to-peak amplitude of spontaneous APs could be explained by the closure of K⁺-channels resulting in an increased resistance. This increase in the resistance of the preparation could also be responsible for the 8bcGMP-dependent attenuation of the AP amplitude reduction. In Manduca sexta, a pheromone-activated,cGMP-blockable K⁺-channel was described without its function being understood (Zufall et al.,1991; Stengl et al.,1992). Based on our findings and on FURA-measurements with cultured ORNs from Manduca sexta (M.S., unpublished observations), we assume that strong or long pheromone stimulation resulted in long-term rises of intracellular Ca²⁺ levels. These Ca²⁺ elevations then appeared to cause rises of intracellular cGMP levels, which then close the fast BAL-dependent cGMP-dependent K⁺-channel. Whether other cGMP-gated ion channels that were described in patch clamp recordings in Manduca sexta (Dolzer,2002) are also involved in the observed adaptations of the action potential response remains to be shown.

In addition to daytime-dependent differences in the control recordings,several daytime-dependent effects of cGMP were found. 8bcGMP-dependent distribution of APs became more tonic and the number of APs during the first 100 ms of the BAL response only decreased at ZT 8-11. Furthermore, the 8bcGMP-dependent decrease of the AP frequency at ZT 8-11 was 12% stronger than in recordings at ZT 1-4. Also, the 8bcGMP-dependent attenuation of the reduction of the positive AP amplitude was more prominent at ZT 8-11. In addition to daytime-dependent differences in the effects of 8bcGMP, a different distribution in the timing of APs in control recordings between both ZTs were observed. In recordings from ZT 8-11 the pheromone-dependent AP response became less phasic and more tonic, and also fewer numbers of APs were elicited during the first 100 ms of the BAL response, than at ZT 1-4. Thus,the main differences between the responses of the recordings at ZT 1-4 and ZT 8-11 are the distribution and numbers of APs elicited in the phasic portion of the response. An ORN with more tonic responses would very likely be less able to resolve pheromone filaments of high frequencies.

The effects of the plume structure on behavioural and flight responses of male Cadra cautella to pheromone were investigated in wind tunnel experiments (Mafra-Neto and Cardé,1995a; Mafra-Neto and Cardé, 1995b). Faster frequencies of pulses were followed by a higher percentage of males responding by shorter latencies and less time spent in the search behaviour. Because the variation of the fine structure of the plume had more influence on the flight pattern of males than a 1000-fold increase in the pheromone dosage, the temporal resolution of ORNs appears to be very crucial. In addition, it was shown that the antennal lobe network is strongly modulated by the temporal pattern of the stimulus(Christensen et al., 1998a; Christensen et al., 1998b). The antennal lobe is tuned to fast temporal discrimination of pheromone pulses,which appears to be necessary for odor blend discrimination(Christensen and Hildebrand,1997; Stopfer et al.,1997). Whether the decrease of the numbers of APs within the first 100 ms represents a further adaptation mechanism that decreases the ability of the ORNs to encode pulsed pheromone signals and therefore blend discrimination remains to be shown in future investigations.

Our results suggest for the first time that there might be daytime-dependent differences in the sensitivity of moth ORNs to pheromone. Whether the observed differences are controlled by an endogenous circadian clock that regulates the sensitivity of ORNs remains to be studied. So far, it has been assumed that daytime-dependent rhythms in male responsiveness to pheromone occur at the level of the antennal lobe and midbrain(Payne et al., 1969; Worster and Seabrook, 1988; Rosén et al., 2003). In addition, investigations of circadian sensitivity changes to food odours in Leucophaea maderae indicated that these rhythms were driven by a circadian pacemaker in the optic lobes but not in the ORNs(Page and Koelling, 2003),while circadian changes in the electroantennogram (EAG) in Drosophila melanogaster suggested that they depend on PER-dependent endogenous circadian pacemakers within the antenna(Krishnan et al., 1999; Tanoue et al., 2004; Zhou et al., 2005). The main difference between our experiments and the studies mentioned is that we performed tip-instead of EAG-recordings, which allowed us to analyse the AP response of a single ORN with high temporal and spatial resolution. The shift from a phasic-tonic to a tonic response pattern as an effect of 8bcGMP application and circadian differences has never been shown before because it cannot be resolved by the EAG-recording technique.

It is still unknown whether Manduca sexta males are maximally sensitive to their intermittently pulsed pheromone signals during the late night when the females are calling. Also, it is unresolved whether there is a circadian rhythm that might adapt the ORNs in the photophase when no pheromone is released by the females and when the moths are inactive. To investigate rhythms in the sensitivity of ORNs to pheromone and to challenge the hypothesis that the blend discrimination is affected via a cGMP-dependent mechanism, long-term recordings starting at the late scotophase, which comprises the shift from scoto- to photophase, will be performed. In addition, it will be investigated in behavioral studies whether male Manduca sexta show rhythmic changes in their responsiveness to pheromone. Furthermore, current biochemical experiments examine whether there are differences in cyclic nucleotide concentrations in the moth antenna at different ZTs depending on differing stimulation schemes (K. Riedinger and M.S., unpublished).

List of abbreviations
 
• AP

  action potential

 
• BAL

  bombykal

 
• cGMP

  cyclic guanosine monophosphate

 
• EAG

  electroantennogram

 
• IP₃

  inositol triphosphate

 
• ISI

  interspike interval

 
• ORN

  olfactory receptor neuron

 
• SP

  sensillar potential

 
• SPA

  SP amplitude

 
• TEP

  transepithelial potential

 
• ZT

  Zeitgebertime

The authors would like to thank Sandy Fastner, Miriam Hock, Klaus Isselbächer, Corinna Struwe and Sandra Utz for insect rearing, Thomas Christensen for the generous gift of BAL and Horst Schmidt for excellent technical solutions. Thanks also to the referees, whose comments helped to considerably improve the manuscript. This work was supported by DFG grant STE 531/13-1 to Monika Stengl.