Olfactory receptor neurons from antennae of developing male Manduca sexta respond to components of the species-specific sex pheromone in vitro

Male-specific olfactory receptor neurons, dissociated from developing antennae of the moth Manduca sexta and grown in long-term primary cell culture, can respond to at least one component of the female moth's sex- pheromone blend with the opening of a nonspecific cation channel. This response does not require the coapplication of pheromone-binding protein.

The transduction processes that underlie these adapting receptor potentials in moth sensilla are assumed to occur in the dendrites ofthe ORNs, in an unusual ionic environment. Within the hollow, hairlike cuticular sensillum, the dendrites are bathed by a receptor lymph containing circa 200 mM K+, circa 20 mM Na+, and presumably several micromolar Ca2+ (Keil, 1984;Griinert, 1985;Kaissling, 1986) as well as abundant soluble proteins, the pheromone-binding proteins (PBPs) (Vogt and Riddiford, 1981;Kaissling et al., 1985;Vogt et al., 1985;Klein, 1987;Gyorgyi et al., 1988;Raming et al., 1989Raming et al., , 1990. It has been proposed that these PBPs (in M. sex&, ca. 16 kDa; Gyijrgy et al., 1988) carry the lipophilic pheromone components through the aqueous receptor lymph and may be required for the recognition of the pheromone components by presumptive receptors in the dendritic membranes of the ORNs (Vogt et al., 1985;Vogt, 1987;van den Berg and Ziegelberger, 199 1). Alternatively, the PBP may scavenge, and thus rapidly inactivate, the pheromone components after their interactions with receptor sites (Kaissling, 1986).
Because the ORNs are not readily accessible to intracellular or patch-clamp recording techniques in situ (Zufall and Hatt,199 l), we developed a primary-cell-culture system for studies of the mechanisms underlying the responses of male-specific ORNs to pheromone (Stengl and Hildebrand, 1990). Cultures are derived from antennae of stage 3 male M. sexta pupae, within 2 d after the mitotic birth of the ORNs and their associated cells.
Among the diverse cells found in these cultures, one particular type with a 5 pm soma and fine processes has been correlated morphologically and immunocytochemically with ORNs in vivo (Hishinuma et al., 1988a,b;Stengl and Hildebrand, 1990). In situ morphological, immunocytochemical, and physiological studies indicate that about 35% of cells of this type are recognized by the monoclonal "male-olfactory-specific antibody" (MOSA) and are responsive to pheromone components Hildebrand, 1976a,b, Hishinuma et al., 1988a,b;Kaissling et al., 1989;Lee and Strausfeld, 1990). Patch-clamp studies show further that ORNs differentiate physiologically in vitro and acquire various types of ion channels in their soma membranes. After 3 weeks in culture, the ORNs express at least one type of Na+ channel and at least three types of K+ channels (Zufall et al., 1991b).
an eight pole Bessel filter and digitally sampled at 25 kHz using a Hewlett-Packard P 9802 computer equipped with a Hewlett-Packard Multiprogrammer 11 interface. More detailed descriptions of the methods used can be found elsewhere (Dude1 and Franke, 1987;Zufall et al.. 1991b).
Here we demonstrate that cultured ORNs can respond specifically to extracts of the female sex-pheromone gland or to the synthetic pheromone component bombykal with opening of a nonspecific cation channel. Furthermore, addition of PBP is not required for these responses.
Preliminary accounts of some of this work have been presented elsewhere (Stengl et al., 1989, 199 1).
Extraction of pheromone glands and delivery of pheromonal stimuli. Sex-pheromone extracts were prepared by dipping female abdominal tips, isolated from adult virgins, in dimethyl sulfoxide (DMSO), dimethyl formamide (DMF), ethanol, or n-hexane. Freshly prepared for each experimental session, the "stock extract" consisted of 100 J of solvent in which five female abdominal tips had been dipped for 2 min each (Tumlinson et al.. 1989). These stock extracts were diluted 1: 10' to 1:105 in 1 ml of extracellular saline solution. The hexane stock solution (which resulted in the purest pheromone extracts) was mixed with 10% DMSO, to facilitate aqueous solubility, before dilution in extracellular saline. The concentration of pheromone in dilutions of the stock extracts was estimated (as "bombykal equivalents") by means of electroantennogram (EAG) recordings, obtained with adult male antennae by established procedures (Christensen et al., 1989).

Materials and Methods
Unless otherwise specified, all culture media were purchased from GIB-CO (Grand Island. NY). and all chemicals and biochemicals. from Sigma Chemical Co. (St:'Louis, MO).
Animals. Manduca sexta (Lepidoptera: Sphingidae) were reared from eggs on artificial diet (modified from Bell and Joachim, 1976) on a longday photoperiod regimen (17 hr light/7 hr dark) at 25-26°C and 50-60% relative humidity. Pupae were staged as previously described (Sanes and Hildebrand, 1976a;Tolbert et al., 1983). They were usually selected for dissection between 2:00 and 5:00 AZT (Arbitrary Zeitgeber Time, with lights on at 0O:OO AZT) and anesthetized by chilling on ice for lO-15 min before dissection of the antenna1 flagellum.
Cell cultures. A detailed description of the culture techniques has been reported elsewhere (Stengl and Hildebrand, 1990). Briefly, antenna1 flagella from male M. sexta pupae (stage 3 of the 18 stages of adult development) were disrupted by a combination of mechanical and enzymatic treatments. The dispersed cells were plated in concanavalin A-coated or uncoated Falcon plastic dishes, in Leibowitz L15 medium, supplemented with 5% fetal bovine serum (Hyclone) and 20-hydroxyecdysone (ca. 1 &ml) or conditioned medium (supematant fluid from cultures of a non-neural M. sexta cell line, generously provided by Drs. J. Hayashi and L. Oland, of the ARL Division of Neurobiology, University of Arizona, or extracellular fluid from antennae) (Stengl and Hildebrand, 1990). The cultures were maintained for 2-4 weeks at about 30°C at high humidity in an incubator.
Patch-clamp technique and data analysis. Patch-clamp recording experiments closely followed the methods described by Hamill et al. (198 1). Patch pipettes were made from borosilicate glass capillaries (World Precision Instruments, GC. 150 TlO; or Clark Electromedical Instruments, Reading, UK) with a two-stage electrode puller (DMZ, Zeitz Instruments, Augsburg, Germany) or a Sutter Instruments micropipette Duller (model P8O/PC). The Dinettes were coated with Svlgard (Dow Coming, Midland, MI) and then fire polished. The tip resistance was 6-12 MB when the electrodes were filled with physiological saline solution. The cells were viewed at 320 x magnification with a Zeiss Axiovert 10 inverted microscope equipped with phase-contrast optics or at 400 x with an Olympus inverted microscope equipped with phasecontrast or Hoffmann modulation contrast optics. After formation of a seal between the pipette and the cell membrane, the electrode capacitance was compensated. During application ofthe pheromone, the membrane patches were kept at 0 mV holding potential (at the cells' resting potential) in the cell-attached configuration.
Single-channel currents were measured at room temperature with an EPC-7 amplifier (List Electronic, Darmstadt, Germany) or an Axopatch-1C patch-clamp amplifier (Axon Instruments Co., Burlingame, CA). The signals were recorded on videotape with a modified Sony PCM-501-ES video recorder or with an instrumentation FM tape recorder (A. R. Vetter Co., Repersburg, PA) or acquired on line with an 80386based microcomputer (Dell Computer Corp., Austin, TX) using ~CLAMP software (Axon Instruments Co.). The data were analyzed with ~CLAMP software (Axon Instruments) or with custom programs (Dude1 and Franke, 1987). Single-channel currents were low-pass filtered at 2-5 kHz with Diluted pheromone extract was applied via a glass-capillary pipette with a tip opening of < 10 pm, driven by a Picospritzer (General Valve Corp., Fairfield, NJ). The pheromone was puffed onto the cell's soma (2 20 pm from the pipette tip) before or after formation of a seal between the pipette and the cell membrane, respectively, to allow or prevent direct access of the pheromone to the membrane patch. The final concentration of pheromone in the gland extracts used ranged between 3 fg and 30 pg (bombykal equivalents) per ml (as estimated by means of EAG recordings), and the final concentration of the stock solution of synthetic bombykal (prepared from a 10 &ml DMSO solution of bombykal kindly provided by Prof. H. Bestmann) ranged between 100 fg and 1 ng per ml. Usually the cells were first stimulated with solvent alone (control) and then tested with the pheromone solutions in the same solvent. Because the pheromone adsorbed to plastic or glass surfaces, the first application of pheromone contaminated the preparation dish. Once pheromone had been introduced into the dish, its physiological effect (thus possibly its effective concentration) could be reduced by addition of bovine serum albumin (BSA) and extensive washing, but the physiological responses of the cells indicated that the pheromone could not be completely removed from a contaminated dish.
Preparation ofpheromone-bindingprotein. PBP was extracted by previously described methods (Klein, 1987;Vogt, 1987;Gyijrgyi et al., 1988). Antenna1 flagella (n = 44) of adult male M. sexta (1 d posteclosion) were collected over solid CO,. The frozen flagella were vortexed in pulverized solid CO, to break off the sensory hairs. After the solid CO, had fully sublimed away, the flagellar shafts were removed. The detached sensory hairs and scales were collected with diethyl ether. After evaporation of the ether, the fractions were taken up in 1 ml of buffer solution containing 250 mM sucrose, 2 mM EDTA, and 50 mM Tris (pH 8) and homogenized on ice. The homogenates were centrifuged at 2000 rpm for 5 min in a Sorvall RC-5B centrifuge (Du Pont Instruments, Wilmington, DE), and the supematant fractions for 10 min at 9000 rpm. The resulting supematant fractions were centrifuged at 27,000 rpm for 1 hr. The membranous and cytoplasmic fractions were taken up in buffer, and the protein concentrations were determined with a modified Lowry assay (Peterson, 1977; data not shown). The fractions were electrophoresed on SDS-PAGE gels or on native 12% acrylamide gels (Vogt and Riddiford,198 1) and stained with Coomassie blue. The cytoplasmic fraction of the sensilla and scales contained two bands with a molecular mass between 14 and 2 1 kDa, corresponding to the published molecular mass of PBP (18 kDa). This fraction was used as PBP in physiological experiments.

Results
To test the ability of cultured ORNs to respond specifically to pheromone, we obtained patch-clamp recordings from cells in antennal-cell cultures 2-4 weeks old. Except for control recordings (see Control experiments, below) on various types of an-tennal cells, only ORN-like cells (i.e., with a 5 pm soma and fine processes) were chosen for these experiments (Stengl and Hildebrand, 1990). The number of ORNs from which recordings were obtained successfully is given as n.

cell attached
To obtain responses from intact cells without dialysis of cytoplasmic components such as second messenger systems, we recorded in the cell-attached mode from the cell soma. The pheromonal stimulus was applied either before or after formation of the seal between the patch pipette and the cell membrane. The pheromone was applied via a puffer pipette, either 5 PA with or without PBP, to test whether addition of PBP is nec-Iessary to obtain pheromone-dependent responses. The stimulus 50 Ins concentrations varied between about 3 fg and 30 pg of bombykal per ml for the pheromone-gland extracts, as estimated by EAG recordings (Christensen et al., 1989) (see Materials and Methods), and varied between about 100 fg and 1 ng per ml for the B synthetic bombykal solutions.
Control experiments Cultured antenna1 cells were tested for responses to stimulation with the solvents alone. We also sought to determine whether the responses to pheromone were specific to pheromone components. Furthermore, we tested whether the cells respond at a specific time in vitro and whether the responses are specific to certain types of cells.
In more than 50 cell-attached patch-clamp experiments, cultured ORNs were stimulated first with the solvents (DMSO, DMF, ethanol, hexane-DMSO) alone, before application of the pheromone sample. In none of these control experiments were responses elicited by the solvents alone. Cells that responded to pheromone did not respond to 1 PM citral (n = 8) a potent plant-derived odorant. These findings suggested that the responses might be cell-type-specific.
None of the 15 nonreceptor cells (judged by morphological criteria) responded to pheromonal stimulation. To demonstrate that the cells we tested were male-specific ORNs, we processed cells (n = 2) that responded specifically to pheromonal stimulation for MOSA immunocytochemical reactivity. These pheromone-responsive cells expressed the male-specific antigen recognized by MOSA (data not shown), which is expressed in situ by mature male-specific ORNs (Hishinuma et al., 1988a,b). Because most tested cells detached from the bottom of the culture dish or shriveled up and died after prolonged stimulation with pheromone, we did not succeed in processing all pheromone-responsive cells for MOSA immunocytochemistry.

Specific responses to pheromone
To investigate the responses of 2-3-week-old cultured ORNs to pheromone, we stimulated the ORNs with extracts of female pheromone glands while recording in the cell-attached configuration at the resting potential. The pheromone was applied either before or after formation of the patch-clamp seal. Specific responses to the female pheromone-gland extracts (n = 48 out of 127 ORNs from which recordings were obtained successfully) could be accounted for fully by the response to synthetic bombykal (n = 5). This finding suggested that the cells under study responded mainly or exclusively to bombykal in the extracts.
In these experiments, 38% (48 of 127) of ORN-like cells in 2-3-week-old cultures responded specifically to extracts of fe-I ' I 5 PA 50 ms Figure 1. Pheromone-dependent currents in cultured ORNs. A, After application of pheromone (marked with a solid circle) to cultured ORNs (which had been in vitro at least 10 d) in the cell-attached configuration (at 0 mV holding potential), with 156 mM KC1 and lo-' M Ca*+ in the pipette, a large inward current appeared. B, After several minutes of washing the cell with extracellular solution, single-channel events could be resolved that showed at least three channels with superimposed openings, all reversing around 0 mV cell potential. At least two different channel types could be distinguished according to small differences in the amplitudes, and according to the presence of subconductance states in only one of the channels present. One type of channel (circles) that showed amplitude substates resembled in its kinetics and conductance the pheromone-dependent cation channel (Fig. 2). The other channel type (squares) opened in bursts with rapid flickers between the open and closed states.
male pheromone glands or to synthetic bombykal. The cultured cells exhibited specific pheromone-evoked responses whether or not they had been exposed to the development-regulating hormone 20-hydroxyecdysone. After application of higher concentrations of pheromone (about 1 &ml), a depolarizing inward current of 35-60 pA occurred (Fig. 1A) in 23% (11 of 48) of the pheromone-elicited responses. The current started after delays of < 100 msec to several seconds following application of pheromone (at different distances from the cell). The duration of the inward current was at least 40 set while the pheromone concentration was reduced by extensive washing. In 4 of the 11 experiments in which pheromone-dependent inward current was observed, single-channel openings could be distinguished after several rinses with extracellular saline solution (Fig. lB), which diminished the concentration of pheromone. In three cases channels could be discerned that shared properties with the cation channel described in Figure  2. These channels, with a reversal potential around 0 mV, opened in bursts and exhibited amplitude substates. In at least one case, another channel was recognized that shared properties with the previously described Ca2+-dependent K+ channels (Zufall et al., Stengl et al. l  , The I-Vcurve shows that the channel reversed around 0 mV. Its conductance was nonlinear, with about 20 pS estimated at 25 mV and 50 pS at -100 mV. The channel did not discriminate among K+, Na+, and Cs+. The cell's membrane potential was estimated in cell-attached recordings with symmetrical K+ concentrations from the reversal potential of the previously described delayed-rectifier channels (Zufall et al.,199 1 b) and confirmed in the current-clamp mode after establishing a whole-cell configuration. The Z-C curve was obtained from amplitude histograms at different pipette potentials, taking only the main amplitude state into account. 1991 b). This channel opened in bursts, with rapid flickers between open and closed states, and showed no amplitude substates ( Fig. 1 B, squares).
After stimulation with lower concentrations of pheromone (pg-fg/ml), in 54% (26 of 48) of the pheromone-specific responses, openings of nonspecific cation channels could be noted in single-channel records ( Fig. 2A). This type of channel was never observed in cell-attached experiments in the absence of applied pheromone (n > 200), whereas K+ channels resembling the delayed rectifier could be observed regularly. The current through the cation channels reversed around 0 mV cell potential (i.e., between -10 mV and 20 mV, no channel activity was observed) in the cell-attached configuration with Ca2+-buffered intracellular saline solutions containing different principal cations (150 mM K+, Na+, or Cs+) in the patch pipette (Fig. 2B). This indicated that the channels do not discriminate among K+, Na+, and Cs'. The conductance of the channels was nonlinear: circa 50 pS at -100 mV, 36 pS at -25 mV, and 18 pS at 25 mV. Amplitude histograms at -70 mV holding potential revealed additional smaller-amplitude substates (Fig. 3). The burst length of channel openings (of the largest-amplitude state, at -70 mV holding potential) can be fitted by one exponential with a time constant ofabout 43 msec (Fig. 4A). The distribution of open states of the largest-amplitude state can be described by two exponentials, with time constants of about 0.6 msec and 2.3 msec at -70 mV holding potential (Fig. 4B). The pheromone-dependent cation channel closed after prolonged (several minutes) or strong (pg/ml) pheromonal stimulation (cell attached) (see Fig. 7B).
To determine whether the pheromone opened the pheromone-dependent cation channels via mechanisms confined to the membrane, we stimulated outside-out patches with pheromone. Patch excision to the outside-out configuration, even in the absence of pheromone, activated cation channels. These cation channels reversed around 0 mV, did not discriminate among Na+, K+, and Cs+; and expressed conductance substates (Fig. 5). Thus, these channels closely resembled the pheromonedependent cation channels that were observed in the cell-attached configuration. The conductance of the excision-activated channels was also nonlinear, but showed less inward rectification than the pheromone-dependent cation channels. At -100 mV cell potential, the main conductance state of these channels was about 50 pS, while it was about 40 pS at -30 mV and 37 pS at 30 mV (Fig. 5B). The cation channels opened readily and continuously after patch excision even in the absence of pheromone but were never observed in the cell-attached configuration under control conditions. Thus, we concentrated on cell-attached patch recordings in our search for opening of pheromone-specific ion channels.
The cation channels that were opened by patch excision were tested with a variety of substances in a search for agents that could close the channels. In outside-out patches, this type of channel was unaffected by 1O-6 M TTX, 20 mM 4-aminopyridine, 10 mM ATP, 1 mM bromo-cGMP, 1 mM bromo-CAMP, 100 FM amiloride, or Ca2+-channel blockers including Co2+, Sr2+, and Ni*+ (applied outside). The channels were blocked in most cases by 20 mM tetraethylammonium chloride (applied to outside-out patches; data not shown). Replacement of Cl-by aspartate, glutamate, or acetate did not affect the currents through the cation channels, but no currents through cation channels were observed if the cations were replaced by choline.
also contained responses of several cation channels or were too short, we did not analyze them further. In at least 6% (3 of 48) of all pheromone-dependent responses, channels opened (Fig.  lB, squares) that resembled the previously described Ca2+-dependent K+ channels (Zufall et al.,199 lb) in kinetics and amplitude (data not shown).
Coapplication of PBP (Fig. 7A) was not required to obtain specific pheromone-dependent responses. After lo-60 min of incubation with high concentrations of pheromone (ca. 30 pg/ ml), the pheromone-dependent cation channels and the delayedrectifier channels closed. If PBP or BSA was then applied (at a concentration of about 10 mM), channels with a reversal potential of about 0 mV opened (Fig. 7B). Channel activity that reversed around 0 mV could be suppressed again with application of more pheromone, or thereafter elicited again with application of more PBP (n = 3).

Discussion
The findings presented here demonstrate that cultured ORNs from male M. sexta antennae can respond specifically to at least one component of the species-specific sex-pheromone blend with changes in the opening of nonspecific cation channels, Ca*+dependent K+ channels, and cGMP-dependent K+ channels. Furthermore, this action does not require addition of PBP. The development of primary cell cultures of differentiating ORNs that are morphologically and immunocytochemically distinguishable (Stengl and Hildebrand, 1990) has facilitated studies of the cellular mechanisms underlying olfactory transduction. This approach now makes possible a detailed analysis of pheromone-dependent ion channels in all patch-clamp configurations, as well as in intracellular recordings, and also opens the way to molecular studies of ORNs.
Because the ORNs express different ion channels in their soma membranes at specific times in their development in vitro (Zufall et al.,199 1 b), we assumed that pheromone-dependent ion channels would also be expressed in their somata, before becoming localized to the dendritic sites of their final deployment. This assumption was supported by findings in other neuronal culture systems in which Caz+ or Na+ channels are first expressed in soma membranes and then became localized to the cells' processes (Fukuda and Kameyama, 1979;Roederer and Cohen, 1983). To search for pheromone-evoked responses, we recorded from soma membranes using the patch-clamp technique in the cell-attached configuration. We intended to prevent loss of cytoplasmic constituents, including second messengers that might be crucial to the mechanisms of pheromone transduction.
Female pheromone-gland extracts were chosen as the main olfactory stimuli to maximize the stimulation of all pheromonesensitive ORNs. From in vivo studies, it is known that there are at least three types of pheromone-sensitive antenna1 ORNs, which respond to different components of the pheromone-gland extract (Kaissling et al., 1989). To assess the specificity of the responses to the gland extract, we compared different solvent extracts, and in some of our experiments the synthetic pheromone component bombykal was employed. The similarities in the responses to bombykal and to the pheromone-gland extracts and 21 kDa of the cytoplasmic fraction of the sensilla and scales were used as "PBP" in physiological experiments. Lane 3 contained the cytoplasmic fraction of the antenna1 shafts and exhibits a protein band of about 14 kDa and another between 3 1 kDa and 2 1 kDa. Proteins of higher molecular mass remained in the stacking gel. Lane 4 was loaded with the membrane fraction (A4) of the antenna1 shafts and shows proteins of higher molecular mass that still remained in the stacking gel. B, After prolonged exposure to high concentrations of pheromone (ng/ ml), the delayed-rectifier channels as well as the nonspecific cation channels had closed (star). After subsequent application of PBP at the cell's resting potential, single-channel openings again could be observed. Two channels with a reversal potential around 0 mV and an amplitude resembling that of the pheromone-dependent cation channel opened with very long open times.
suggest that the responses could have been due mainly or entirely to bombykal in the gland extracts. In support of this interpretation, solvents alone did not elicit any response, the pheromone concentrations used were in the physiological range (see Ma: terials and Methods), and cells that responded to pheromone did not respond to the plant odorant citral. Moreover, the responses were specific to the ORN cell type and could be elicited only late in development in vitro. There was a good correlation between the proportion of responding cells (38%) and the expected occurrence of pheromone-sensitive cells (35%) among the odor-sensitive neurons (recognized on the basis of their 5 pm somata and their fine bipolar processes in vitro; Stengl and Hildebrand, 1990). Finally, we showed that cells that responded with opening of cation channels were MOSA immunoreactive. Because MOSA immunoreactivity is characteristic of male-spe-cific ORNs in situ (Hishinuma et al., 1988a,b), it provided further confirmation of the specificity of the responses to pheromone. Although dose-response curves usually provide further indication of response specificity, they were not determined in our experiments because we could not ascertain accurately the concentration and arrival time of the pheromonal stimulus at the cell. In a more general way, however, concentration dependence was exhibited: at higher concentrations of pheromone (picogram of bombykal), large inward currents obscured singlechannel events. Single-channel events could be resolved only after reduction of the pheromone concentration by extensive washing or addition of BSA to the medium. Even before application of the pulse of pheromone solution, diffusion of the pheromone out of the puffer pipette appeared to be sufficient to open the cation channel in cell-attached recordings. This extraordinary sensitivity is consistent with possible involvement of second-messenger cascades in the opening of the pheromonedependent cation channels. The relatively long response delays (from < 100 msec up to several seconds) might also be accounted for by interposed second-messenger actions. Because the cultured ORNs responded to the pheromone, however, even when formation of the seal preceded delivery of the pheromonal stimulus, second-messenger control of the cation channel appears to be likely. Considerable evidence for involvement of second-messenger systems in olfactory transduction, in vertebrates as well as in invertebrates, is accumulating (Pace et al., 1985;Nakamura and Gold, 1987;Breer et al., 1988Breer et al., , 1990Firestein and Shepherd, 1989;Firestein and Werblin, 1989;Boeckhoff et al., 1990;Zufall and Hatt, 1991). After exposure to pheromone, insect ORNs exhibit relatively long-lasting increases in cGMP (Ziegelberger et al., 1990). These increases in cGMP occur in the somata rather than the dendritic regions of the ORNs. G-Protein-dependent increases of inositol 1,4,5-triphosphate (IP,) occur in insect ORNs within milliseconds after pheromonal stimulation, but in contrast to the situation in vertebrates, no increases in CAMP could be detected (Breer et al., 1988Boekhoff et al., 1990;Ziegelberger et al., 1990). Thus, a G-protein-dependent phospholipase C (Nishizuka, 1984;Berridge, 1987), which might cause an IP,-dependent rise in internal Ca2+, may play a role in the production of odor-dependent potential changes in insect ORNs. Furthermore, a guanylate cyclase or possibly a phosphodiesterase might be involved in the olfactory transduction mechansims. The involvement of diffusible second messengers in olfactory transduction processes thus makes it likely that channels not only in the outer dendritic segment but also in the inner dendritic segment and the soma may underlie odor-dependent potential changes that influence the firing of action potentials in ORNs. Therefore, more than one type of secondmessenger-mediated channel might be involved in the generation of the receptor potentials in ORNs.
The pheromone-dependent cation channel described here is a good candidate to be one of the second-messenger-gated channels involved in the mechanisms that underlie the depolarizing receptor potentials in pheromone-sensitive ORNs. Because this type of channel is not voltage gated and is observed only in cellattached recordings after pheromonal stimulation, its activation apparently depends upon pheromone. The opening of this channel leads to inward currents carried by K+, Na+, and possibly also Ca2+ at negative membrane potentials. Thus, these cation channels depolarize the cell if opened at the cell's resting potential. The suggestion that the opening ofcation channels might be one of the events that underlie the pheromone-dependent generator currents is consistent with previous observations that the depolarizing receptor potentials are accompanied by a decrease in membrane resistance (Kaissling and Thorson, 1980;Kaissling, 1987). This suggestion is further supported by recent findings of pheromone-dependent, second-messenger-mediated (diacylglycerol, cGMP) cation channels on extruded dendrites of moth ORNs in situ and second-messenger-dependent (Ca*+, cGMP) cation channels on dendritic segments of ORNs in Antheruea polyphemus (Zufall and Hatt,199 1;Zufall et al.,199 la). The pheromone-dependent cation channel of cultured ORNs in M. sexta has a conductance different from both cation channels observed in A. polyphemus. Current experiments examine whether the observed pheromone-dependent cation channel in M. sexta, which otherwise shares reversal potential, ion selectivity, and the property of expressing multiple conductance states with both of the cation channels observed in A. polyphemus, also depends on second messengers (Stengl et al., 199 1, 1992).
The M. sexta cation channels activated during excision of membrane patches appear not to be blockable by amyloride or cGMP, both of which block the Ca2+-dependent cation channels activated by patch excision in A. polyphemus. It seems likely, however, that increased Caz+ concentrations (via influx from the extracellular medium or damage to internal stores) during the process of patch excision activated these channels in both species. Ion channels activated by patch excision have also been found in other systems, but their functions and mechanisms of activation are unknown, although it is also assumed that Ca2+ and ATP may play a role in their activation (Yazejian and Byerly, 1989;McClintock and Ache, 1990).
Whether the cation channels activated via patch excision belong to the same type of cation channel as the pheromonedependent cation channels measured cell-attached remains to be demonstrated. Experiments in progress on M. sexta show that cultured ORNs are equipped with Ca*+-dependent cation channels as well as with protein kinase C-dependent cation channels that are affected differently by cGMP (Stengl and Hildebrand,199 1;Stengl et al.,199 1,in press). These data will be presented in a separate paper (M. Stengl and J. G. Hildebrand, unpublished observations). Further studies should ascertain whether the apparently different cation channels in both species of moths are truly different types of cation channels or whether they belong to the same channel type, which changes its properties via phosphorylation by protein kinases during prolonged pheromone exposure.
The opening of a channel that resembles in its kinetics and amplitude the previously described Ca2+-dependent K+ channels (Zufall et al.,199 lb) implies a possible increase of at least lo-fold in internal Ca2+ after application of pheromone. The observed changes in the probability of opening of the nucleotidesensitive delayed-rectifier K+ channels could possibly be accounted for by changes in the levels of ATP or cGMP or by changes in the cell's potential (Zufall et al., 1991 b). Because increases in cGMP after pheromonal stimulation have been found in insect ORNs (Ziegelberger et al., 1990) the closing of the delayed-rectifier K+ channel after prolonged exposure to pheromone might be caused by cGMP. Because the current amplitude of the delayed rectifier did not decrease in all recordings, the pheromone-dependent increase in activity is probably not caused by a depolarization of the ORN.
In view of the fact that application of PBP was not necessary to obtain pheromone-dependent responses in cultured ORNs as well as in recent in situ exneriments (Zufall and Hatt. 199 1).
Kaissling K-E (1986) Chemo-electrical transduction in insect olfactory II it seems that the pheromone need not be presented in the form of a pheromone-PBP complex to a presumptive receptor site. We cannot exclude the possibility, however, that PBP might receptors. Annu Rev Neurosci 9:121-145. Kaissling K-E (1987) RH Wright lectures on insect olfaction (Colbow K, ed). Burnaby, British Columbia, Canada: Simon Fraser University. Kaissling K-E, Thorson J (1980) Insect olfactory sensilla: structural, have been nroduced in vitro and therefore nresent in cultured ORNs in a-membrane-bound form in which it was not washed away by superfusion. The PBP could be used interchangeably with BSA (which unspecifically binds lipophilic substances), apparently reducing pheromone concentrations and thus possibly preventing channel closure. Thus, our preliminary results are consistent with the possibility that PBP may also serve to scavenge the pheromone and hence to prevent adaptation due to overstimulation by pheromone. However, this study did not attempt to analyze the function of the PBP. The influence of the PBP on the time course and adaptation of the response to pheromone remains to be investigated further.