Chemosensitivity of the Osphradium of the Pond Snail Lymnaea Stagnalis

The Journal of Experimental Biology 198, 1743Ð1754 (1995) 1743 Printed in Great Britain © The Company of Biologists Limited 1995

CHEMOSENSITIVITY OF THE OSPHRADIUM OF THE POND SNAIL LYMNAEA STAGNALIS

HEINER WEDEMEYER AND DETLEV SCHILD* Physiologisches Institut, Universität Göttingen, Humboldtallee 23, D 37073 Göttingen, Germany

Accepted 14 April 1995

Summary The osphradium of the pond snail Lymnaea stagnalis was group of 15 neurones that lay next to the issuing osphradial studied to determine the stimuli to which this organ nerve, to determine whether ganglion cells were involved responds. The following stimuli were tested: hypoxia, in olfactory signal processing. All neurones tested hypercapnia, a mixture of amino acids, a mixture of responded to at least one of the three mixtures of odorants. citralva and amyl acetate and a mixture of lyral, lilial and Both excitatory and inhibitory responses occurred. ethylvanillin. Our results indicate that the osphradium of the pond The mean nerve activity consistently increased with snail Lymnaea stagnalis is sensitive to elevated PCO∑ as well elevated PCO∑, whereas hypoxia produced variable effects. as to three different classes of odorants. In addition, at least The nerve activity became rhythmic upon application of some neurones within the osphradium are involved in the citralva and amyl acetate, but it increased in a non- processing of olfactory information. rhythmic way upon application of the other two odorant mixtures tested. Key words: Lymnaea stagnalis, osphradium, chemosensitivity, Whole-cell patch-clamp recordings were made from a olfaction, PCO∑.

Introduction Since its ﬁrst description by Lacaze-Duthier (1872), the involvement of the osphradium in respiratory behaviour. function of the osphradia of snails has been a topic of debate. Kamardin (1976a) and Sokolov and Kamardin (1977) reported, Sensitivity to mechanical stimuli was proposed for marine on the basis of recordings from the right internal pallial (RIP) prosobranchs by Copeland (1918), Hulbert and Yonge (1937), nerve, that the osphradium of Lymnaea stagnalis responded to Yonge (1947), Kohn (1961) and Welsch and Storch (1969), but low partial pressures of oxygen (PO∑). This could possibly the osphradia of Buccinum undatum and Paludina vivipara parallel the function of the glomus I cells of the vertebrate have been shown to respond to chemical and not to mechanical carotid body (López-López et al. 1989), which directly activate stimuli (Wölper, 1950; Bailey and Laverack, 1963, 1966; the respiratory network in the brainstem via the Welsch and Storch, 1969). Comparative and evolutionary glossopharyngeal nerve. However, morphological studies aspects of gastropod osphradia are discussed in detail by Croll (Spengel, 1881; Benjamin, 1971; Benjamin and Peat, 1971; (1983) and Haszprunar (1985a,b, 1987a,b). Studies of the Kamardin, 1976b; Zaitseva, 1982) showed the presence of transduction mechanisms of the receptor neurones have not yet sensory neurones in the osphradial epithelium, although the been undertaken. odorants that were attractive in Y-maze experiments In opisthobranchs, Stinnakre and Tauc (1969) reported (Michelson, 1960) did not change the spike rate in intracellular evidence for the presence of osmoreceptors within the recordings from osphradial ganglion cells (Bailey and osphradium and Jahan-Parwar et al. (1969) recorded central Benjamin, 1968). Nor did a functional osphradium seem to be effects after chemical, osmotic and mechanical stimulation of essential for directional movement towards the source of a food the osphradium of Aplysia californica. A role for the extract (Townsend, 1973). Kamardin (1983, 1988) studied the opisthobranch osphradium in the control of respiratory homing behaviour of different snails, including two pumping was suggested by Croll (1985), Levy et al. (1989) and pulmonates, and suggested that the osphradia might be Levy and Susswein (1993). These authors found the organ to involved in this kind of behaviour. Recently, Nezlin et al. be sensitive to hypercapnia and hypoxia. (1994) described leucine-enkephalin- and methionine- In pulmonate snails, Lankester (1883) suggested an enkephalin-immunoreactive cells within the osphradial

*Author for correspondence. 1744 H. WEDEMEYER AND D. SCHILD ganglion that project their neurites into the sensory part of the (Diamond General, Ann Arbor, Michigan, USA) placed in the epithelium. The physiological roles of these cells are not yet Petri dish bath next to the osphradium. During hypoxia known. Taken together, the evidence regarding the function of experiments, PO∑ was between 7 and 11 mmHg. For patch- osphradia in aquatic pulmonate snails has remained highly clamp recordings, the patch pipette was filled with a solution controversial. containing (in mmol lϪ1): NaCl, 2; potassium aspartate, 37; In this study we have reinvestigated this question. First, we KCl, 5; MgCl2, 2; Hepes, 10; EGTA, 1; K2ATP, 1; Li2GTP, tried to determine the type of stimulus that changes the activity 0.5; cyclic AMP, 0.01. The combination of cyclic AMP and of the RIP nerve, by recording the activity of this nerve under ATP largely prevented the so-called ‘calcium wash-out’ hypoxia and hypercapnia and upon application of olfactory (Kostyuk et al. 1986). stimuli. The results show that the osphradium of L. stagnalis Three different solutions (A, B and C) of odorant stimuli is unique, in that it is sensitive to odorants as well as to changes were prepared by adding odorants to solution LR1. Solution A in PCO∑. Second, we used the whole-cell configuration of the contained the following L-amino acids (final concentrations in patch-clamp technique to determine whether osphradial mmol lϪ1): arginine, 0.6; cysteine, 0.2; isoleucine, 0.4; leucine, ganglion neurones were involved in the signal processing of 0.4; lysine, 0.4; methionine, 0.1; phenylalanine, 0.2; threonine, the responses to odorants. As this turned out to be the case, the 0.4; tryptophan, 0.1; tyrosine, 0.1; and valine, 0.4. Solution B osphradium of L. stagnalis could be unique also because both contained the odorants citralva and amyl acetate (100␮mol lϪ1 primary and secondary olfactory signal processing may take each). Solution C contained the odorants lyral, lilial and place in the same organ. ethylvanillin (100 ␮mol lϪ1 each). Stock solutions of citralva, amyl acetate, lyral, lilial and ethylvanillin were made by dissolving the odorants in methanol. The final concentration of Materials and methods methanol in solutions B and C was 0.01 %. All stimulus Preparation solutions were applied to the Petri dish bath by superfusion Specimens of Lymnaea stagnalis L., collected originally in (exchange time was about 1 s). The application of methanol a local pond and maintained for many generations under alone had no effect upon membrane current, membrane voltage laboratory conditions, were kept in aquaria at 15Ð17 ûC and fed or RIP nerve activity. on lettuce supplemented with tropical fish food. Electrophysiological experiments were performed on Nerve activity recordings isolated osphradia from snails weighing approximately 2 g. The activity of the RIP nerve was recorded using suction After anaesthesia in iced water, the tissue above the central electrodes with a tip diameter of about 75 ␮m, amplified with nervous system was removed and the central ganglion ring was an EPC7 amplifier (List, Darmstadt, Germany) in current- extirpated by cutting all nerves close to the ganglia except the clamp mode. Data were recorded on video tape after pulse code RIP nerve. The mantle was then folded back and the modulation (Sony PCM501) and on a chart recorder. The osphradium, with a small section of the RIP nerve attached to activities were off-line low-pass-filtered at 1 kHz (eight-pole it, was extirpated together with a small piece of adjacent tissue, Bessel filter, Frequency Devices) and sampled at 8 kHz using large enough to fix the preparation with minute needles on a an A/D converter (Burr & Brown, MPV952) in a VME silicon layer within a Petri dish. The connective tissue sheet computer (Eltek, Darmstadt, Germany). The digitized data above the neurones was removed by microdissection. were then analysed using a SUN workstation. For the detection of spikes, a custom-built program written in ‘C’ language was Solutions used. The algorithm assigned a spike event to the first of four The isolated osphradia were continuously superfused successive bins of an activity histogram (bin width, 1 ms), (4 ml minϪ1, gravity feed) with standard Lymnaea Ringer whenever 32 successive data points (corresponding to a period Ϫ1 (LR1) containing (in mmol l ): NaCl, 42; KCl, 1.5; CaCl2, 4; of 4 ms) were above a threshold value. For the purposes of this MgCl2, 2; glucose, 5; Hepes, 10. Two other solutions were paper, we call these histograms ‘raw data histograms’. The used where indicated: LR2 (NaCl, 29; KCl, 1.5; MgCl2, 15; threshold was chosen to lie just above the peak noise level of glucose, 5; Hepes, 10; in mmol lϪ1) and LR3 (NaCl, 36; KCl, the nerve activity signal. The validity of the algorithm, i.e. its 1.5; CaCl2, 4; MgCl2, 2; NaHCO3, 7.5; glucose 5; Hepes, 10; ability to detect all spikes above the threshold value, was in mmol lϪ1). Solutions were adjusted to pH 7.8 with NaOH. determined by comparing the original nerve activity traces and Solution LR2 contained an elevated concentration of Mg2+ and the raw data histograms. Finally, the the raw data histograms no added Ca2+, corresponding to an actual calcium were transformed into histograms with a broader bin width concentration of about 20 ␮mol lϪ1, measured with a Ca2+- (5 s) by adding the events of 5000 adjacent bins from the raw sensitive electrode (Ingold, Steinbach, Germany). data histograms. The resulting histograms thus gave the nerve To increase PCO∑ without changing PO∑ or pH, solution LR3 activity averaged within bins of 5 s. In some cases, we chose was bubbled with a mixture of 98 % O2 and 2 % CO2 and a bin width of 10 s (e.g. Fig. 6). As details of the discharge adjusted to pH 7.8 with NaOH. In order to decrease PO∑ at behaviour of single units, easily detectable in the original ambient PCO∑, solution LR1 was bubbled with 100 % N2. PO∑ records, are lost in histograms with a broad bin width, the was continuously measured with an O2-sensitive electrode results are presented either in the form of original traces or as Chemosensitivity of pond snail osphradia 1745

A Epithelium

Osphradial Suction nerve pipette

200 ␮m

B EC

RN RN

Sens

Fig. 1. Schematic diagrams of the osphradium of Lymnaea stagnalis. (A) Section through the GC osphradium taken from a digitally scanned histological tranverse section of the organ, showing the recording sites, i.e. the whole nerve and a group of neurones arranged as a ring in the vicinity of the issuing nerve (outlined). The sections were stained with Neutral Red. Epithelial tubes are indicated by arrows. u, voltage. (B) Schematic view of the ␭-shaped epithelial tube within the osphradium. EC, opening of the epithelial canal; RN, receptor neurones; Sens, sensory region of the epithelium; GC, ganglion neurones; ON, osphradial ON nerve. 1746 H. WEDEMEYER AND D. SCHILD histograms, depending on whether representations of single- A unit behaviour, mean activity or both were appropriate.

Patch-clamp recordings Patch-clamp experiments were carried out in the whole- cell configuration (Hamill et al. 1981) and the procedures were similar to those reported previously (Schild and 100 s Bischofberger, 1991). In this study, we limited recording to a pool of about 15 neurones, located close to the issuing B osphradial nerve. The connective tissue above these neurones was carefully removed. Electrodes of 7 M⍀ resistance were fabricated from borosilicate glass (1.8 mm o.d., Hilgenberg, Malsfeld, Germany) using a custom-built two-stage electrode puller. The patch-clamp arrangement consisted of a patch-clamp amplifier (EPC7, List, Darmstadt, Germany); a PCM unit (Instrutech 100 s VR100), a video recorder and a computer (VME/Eurocom5, ELTEK, Darmstadt, Germany) were used for data recording. C LR2 The command pulse programs were written in ‘C’ language and ran on a VME computer, which delivered the command signals to a 12-bit D/A converter. The resulting pulses were fed to the patch-clamp amplifier. Seal resistances were in the range 1Ð20 G⍀. Whole-cell breakthrough was achieved by brief suction pulses. Data were filtered and sampled at 3 and 100 s 10 kHz, respectively, and evaluated on a SUN workstation using custom-built programs. Fig. 2. Resting patterns of nerve activity in three osphradia. (A) Non- rhythmic nerve activity; 45 spikes sϪ1. (B) Slow rhythm with a mean interburst interval of 90 s. (C) Desynchronizing effect of solution LR2 Results upon resting rhythmic activity. Stimulus application is indicated by Fig. 1 shows two schematic diagrams of the osphradium of the solid bar. L. stagnalis. Fig. 1A was digitized using a scanner, from a photograph of a section stained with Neutral Red, and shows pattern reversibly changed to a non-rhythmic one (observed in the nerve recording pipette and the recording sites: the whole all 12 osphradia tested). nerve and a group of neurones in the vicinity of the issuing nerve. This group of neurones consisted of about 15 cells that Stimulation with odorants were arranged as a ring around the issuing nerve; seven of them Application of stimulus solution A (amino acids) are shown in the section. In situ, their diameters lay between consistently led to a recruitment of single units (Fig. 3A) as 35 and 65 ␮m. Fig. 1B shows the structure of the osphradium well as to an increase in nerve discharge activity (Fig. 3B) with a ␭-shaped epithelial tube, consisting of a ciliated, a (found for 32 out of 33 stimulations in 15 osphradia). In some sensory and a secretory region, the subepithelial receptor cells, osphradia, the post-stimulus nerve activity decreased to values ganglion neurones and the osphradial nerve. The unknown lower than the pre-stimulus activity before recovering to that synaptic connections are omitted. level. Each stimulus solution was applied at least twice to the same preparation, and each application resulted in a virtually Resting activity patterns identical response (Fig. 3C,D). The response characteristics of 22 extirpated osphradia of L. Application of the odorants citralva and amyl acetate stagnalis were studied using suction electrode recordings from (stimulus solution B), which have been reported to increase the the RIP nerve. The patterns of resting activity had mean rates concentration of cyclic AMP in olfactory receptor neurones of of 15Ð65 spikes sϪ1. The resting activity of some specimens other species (Sklar et al. 1986; Boeckhoff et al. 1990; Breer was rhythmic and that of others was random. Fig. 2A shows and Boeckhof, 1991), in all cases induced rhythmic responses an example of a recording of non-rhythmic activity, while (Fig. 4A). Irrespective of the pre-stimulus activity pattern, the Fig. 2B exemplifies a recording of rhythmic discharge. activity became rhythmic with a burst period of about 10 s Thirteen out of 22 osphradia showed rhythmic activity, with (range 3Ð13 s). In addition, stimulus solution B caused an an interburst interval of approximately 80 s, five changed their increase in mean nerve activity (Fig. 4B). The effects of activity pattern occasionally without any obvious cause and stimulus solution B lasted for between 180 and 600 s. These four did not show spontaneous rhythmic activity. When long delays might, at least partly, be due to the slow diffusion solution LR2 was applied (Fig. 2C), the rhythmic activity and extrusion of the stimulus out of the epithelial tube of the Chemosensitivity of pond snail osphradia 1747

A C

120 s 60 s

B D )

1 150 −

100

Nerve activity (spikes s 60 s 0 100 200 300 400 Time (s) Fig. 3. Nerve activity in response to stimulus solution A (amino acids). Application of solution A caused an increase in nerve activity followed by a transient decrease in the activity of some units. (A) Record of nerve activity. (B) Activity histogram (bin width, 5 s) calculated from A. In this and all other histograms, the nerve activity (in spikes sϪ1) is represented as a function of time (in s). Stimulus applications are indicated by solid bars above the traces. (C,D) Two responses of the same preparation to stimulus solution A. Between the two responses, two different stimuli had been applied, but the amplitude and kinetics of the responses are identical. Stimulus application is indicated by the solid bars above the traces.

osphradium. However, compared with the other two odorant either by recruitment (Fig. 6A) or by suppression (Fig. 6B) of mixtures used, the recovery time was consistently longer after single units or by an increase (Fig. 6C) or decrease (Fig. 6D) stimulation with solution B. in mean activity. Thus, the response to hypoxia was variable. Application of a mixture of ethylvanillin, lyral and lilial This may indicate that processes related to hypoxia, rather than (stimulus solution C), which are known to increase the hypoxia itself, caused the observed changes in activity, a concentration of inositol-1,4,5-trisphosphate (InsP3) in hypothesis supported by the consistent increase in nerve olfactory receptor neurones of other species (Boeckhoff et al. activity that was observed when PCO∑ was increased from 0.04 1990; Breer and Boeckhoff, 1991), resulted in a marked to 2 %, (Fig. 7, 51 out of 57 stimulations, 16 osphradia). The increase of nerve activity (Fig. 5) in 28 out of 32 stimulations effect of hypercapnia upon single units, however, could be (in 14 out of 18 osphradia). In nine recordings, we observed a either excitatory or inhibitory (Fig. 7B). In some cases, the suppression of the post-stimulus spiking activity, before activity increase showed a typical phasic component (Fig. 7C) recovery to pre-stimulus levels (Fig. 5A). A rhythmic activity and suppression of activity after the end of hypercapnic pattern was never observed in response to application of conditions (Fig. 7A,C). stimulus solution C. As the responses to hypercapnia could have been mediated by a decrease in tissue pH, we investigated the effect of Responses to hypoxia and hypercapnia solutions of differing pH at constant PCO∑ and PO∑. Fig. 8A,B To investigate the sensitivity of the osphradia to changes in shows a typical single-unit record together with the PO∑ or PCO∑ in the bath solution, we changed PO∑ from corresponding activity histogram before, during and after the 130 mmHg to approximately 7 mmHg (hypoxia) or PCO∑ from application of a solution of pH 7.0 (LR1, adjusted with NaOH). 0.2 to 15 mmHg (hypercapnia). No response to hypoxia was In no cases (15 experiments) were changes in single-unit found in nine out of 20 osphradia (23 of 50 stimulations). In discharges or nerve activity observed, suggesting that changes 11 osphradia (27 stimulations), the nerve activity was changed in pH between 6.8 and 7.8 did not inﬂuence nerve activity. 1748 H. WEDEMEYER AND D. SCHILD

A A

100 s 60 s

100 B ) 1 − 150 ) 1 −

100 50

50 Nerve activity (spikes s

0 Nerve activity (spikes s 0 100 200 300 400 0 Time (s) 0 100 200 300 Time (s) Fig. 4. Nerve activity in response to stimulus solution B. (A) Application of citralva and amyl acetate (both 100 ␮mol lϪ1) resulted Fig. 5. Nerve activity in response to stimulus solution C. Lilial, lyral in a rhythmic activity pattern. (B) Activity histogram showing an and ethylvanillin were applied to the bath, all at ﬁnal concentrations increase in spiking frequency (bin width 5 s). Decreased post-stimulus of 100 ␮mol lϪ1, over the period indicated by the solid bar. (A) Record nerve activity occurred in this case, lasting for approximately 10 min of nerve activity; a high-amplitude unit is recruited and then before complete recovery. Stimulus application is indicated by the suppressed during 30 s of the post-stimulus phase. The recruitment of solid bar above the traces. other units of lower amplitude is shown in the corresponding activity histogram (bin width 5 s) (B).

Responses of single ganglion cells to stimulation with odorants neurones tested were responsive to all three stimulus solutions Zaitseva (1982) suggested that the epithelial sensory (Fig. 9). neurones projected directly and exclusively into the osphradial Fig. 10 gives representative examples of ganglion cell nerve; accordingly, involvement of ganglion neurones in responses to different stimulus solutions. In response to olfactory signal processing was not expected. Bailey and stimulus solution A, after a delay of approximately 5 s, a Benjamin (1968) did not observe changes in activity of single depolarisation with superimposed spikes occurred (Fig. 10A). osphradial ganglion cells after stimulation with odorants in The response lasted for considerably longer than the stimulus Planorbarius corneus. However, the observation that the application, presumably because the sensory cells are situated activity of the osphradial nerve became rhythmic upon within the epithelial tube of the osphradium (Fig. 1) and olfactory stimulation (Fig. 4A) suggested a possible because of the slow wash-out time. involvement of ganglion cells in the generation of the rhythm. Fig. 10B shows an example of the response to stimulus B; To investigate this further, we performed whole-cell recordings the rate of production of action potentials increased markedly from a group of osphradial ganglion cells located next to the upon stimulation of the organ. In some cases, the action of the issuing nerve. stimulus solutions A, B and C was inhibitory. Fig. 10C is an Figs 9 and 10 show the membrane voltage of ganglion cells example of a record where solution B caused a suppression of recorded under current-clamp conditions. Resting membrane activity. The excitatory and inhibitory effects of the stimulus potentials of the neurones ranged between Ϫ45 and Ϫ74 mV. solutions are summarized in Table 1. Twelve out of 75 ganglion cells showed no response to any of These results clearly indicate that odorants differentially the stimulus solutions presented. 63 ganglion cells responded inﬂuence the membrane properties and discharge behaviour of to at least one of the stimulus solutions, and six out of eight ganglion cells in the osphradium of L. stagnalis. Chemosensitivity of pond snail osphradia 1749

A C 75

25 ) 1

100 s −

0 0 100 200 300 400 500

75 B Nerve activity (spikes s

100 s 0 0 200 400 600 800 Time (s)

Fig. 6. Nerve activity of four osphradia in response to hypoxia. The solid bars indicate the duration of hypoxia. PO∑ was decreased within 15 s to 9 mmHg. (A) Recording of nerve activity showing recruitment of single units. (B) Recording of nerve activity showing suppression of single units. (C, D) Activity histograms showing an increase (C) and decrease (D) in mean activity. Bin widths in C and D were 5 s and 10 s, respectively.

Table 1. Effects of the three odorant stimulus solutions on the membrane voltage of ganglion cells Total number Number of cells Number of cells Number of cells of cells tested showing excitation showing suppression showing no response Solution A 36 31 5 0 Solution B 24 20 2 2 Solution C 17 9 4 4 Solutions B+C 16 11 0 5

Sixty-three ganglion cells were used. Amino acids (solution A) affected the membrane potential in all cases. All three cocktails sometimes induced hyperpolarisation of the neurone or suppression of spiking. Depolarisations ranged between 3 and 40 mV, hyperpolarisations ranged between 3 and 12 mV. Changes smaller than 3 mV were not considered to represent a response.

Discussion non-rhythmic (Fig. 2A) or rhythmic (Fig. 2B) so that stimulus- In the present paper, we have attempted to determine the induced changes in resting activity could be easily detected. stimuli to which the osphradium of the pond snail L. stagnalis However, in some preparations, only a few units appeared to be responds. spontaneously rhythmically active, and in ‘non-rhythmic’ The resting activity of the RIP nerve was either constant and recordings, the activity of some rhythmic units might have been 1750 H. WEDEMEYER AND D. SCHILD

A C

150

100

50 ) 1 −

0 60 s 0 100 200 300

B D

Nerve activity (spikes s 50

0 0 100 200 300 400 60 s Time (s)

Fig. 7. Nerve activity in response to hypercapnia. The solid bars above the traces indicate periods during which PCO∑ was raised from 0.04 % to 2 %. (A) Recording of nerve activity showing recruitment of single units with subsequent post-stimulus transient suppression. (B) Recording of nerve activity showing transient recruitment with tonic suppression of single units. (C) Activity histogram of the record shown in A; there is a phasic increase of nerve activity followed by a transient suppression. (D) Activity histogram of the record shown in B showing a tonic increase of nerve activity. Bin widths 5 s. concealed within the noise. We did not attempt to study the The responses of a particular ganglion cell to a given odorant origin and mechanism of generation of the rhythm, because were excitatory, inhibitory or absent (Table 1). Generally, a nerve recordings are not the appropriate method for this purpose. given stimulus solution produced different responses in The resting rhythm was in every case suppressed by solution different cells. Stimulus solutions A (amino acids) and C LR2 (Fig. 2C), possibly because neurotransmitter release was (ethylvanillin, lyral and lilial) consistently led to marked blocked (Goldschmeding et al. 1981; Syed et al. 1991). increases in the mean nerve activity, although single units were The stimuli used were hypoxia, hypercapnia, amino acids occasionally suppressed, while cocktail B (citralva and amyl (solution A), a mixture of citralva and amyl acetate (solution acetate) induced rhythmic activity, the period of which was B) and a mixture of lyral, lilial and ethylvanillin (solution C). about an order of magnitude shorter than that of the resting Owing to the diffusion constants of O2 and CO2 in living tissue, rhythm. Odour-modulated oscillations have also been it is likely that hypoxic and hypercapnic conditions affected described in the protocerebral lobe of Limax maximus not only the sensory epithelium but the whole organ. The (Gelperin and Tank, 1990; Gelperin et al. 1993). Owing to the location of hypercapnia-sensitive cells remains to be marked anatomical differences between the protocerebral lobe determined. The diffusion of larger molecules, such as the of Limax maximus and the osphradium of L. stagnalis, a applied odorants, will be considerably slower, so it seems comparison of the response characteristics is difficult. highly improbable that they penetrated the connective tissue Amino acids are known to act as stimuli for olfactory covering the osphradium. Presumably they entered the receptor cells of aquatic animals (Ivanova and Caprio, 1993; epithelial canal, partly by diffusion and partly facilitated by the Zippel et al. 1993; Miyamoto et al. 1992; Dionne, 1992; beating cilia. Restrepo and Boyle, 1991; Restrepo et al. 1990; Schmiedel- Chemosensitivity of pond snail osphradia 1751

A A

Stimulus A 25 mV 30 s 100 s

75 B B ) 1 −

Stimulus B 25 mV 25 10 s Nerve activity (spikes s C 0 0 100 200 300 400 500 600 Time (s) Fig. 8. Nerve activity in response to a change in pH from 7.8 to 7.0. (A) Recording of nerve activity. (B) Activity histogram (bin width 5 s). Neither in the original recording nor in the activity histogram Stimulus B 25 mV were significant changes observed. Application of solution of pH 7.0 10 s is indicated by the solid bar. Fig. 10. Responses of osphradial ganglion cells to application of stimulus solutions. Activity was recorded in the whole-cell Stimulus A Stimulus B configuration and in the current-clamp mode. Stimulation of the osphradium led to different responses of the osphradial ganglion cells. Stimulus C (A) Depolarization and spiking of an osphradial ganglion cell upon stimulation with stimulus solution A (amino acids). The resting membrane potential of this cell was Ϫ62 mV. There is a complex stimulus-induced spiking pattern, which outlasts the stimulus application by about 60 s. (B,C) Opposite responses to stimulus 10 mV solution B (citralva and amyl acetate). In both cases, the neurones show spontaneous activity. This is enhanced in B and suppressed in C. 30 s Membrane potentials of the cells were Ϫ58 mV (B) and Ϫ48 mV (C). Fig. 9. Responses of an osphradial ganglion cell to successive stimulation with stimulus solutions A, B and C. The cell was held in the whole-cell configuration in the current-clamp mode. Resting not only to a solution of amino acids but also to the odorants membrane potential was Ϫ64 mV. Each stimulus led to a of the stimulus solutions B and C, which are known to act as depolarization, although with different kinetics for the three stimuli for olfactory receptor neurones in non-aquatic animals applications. The example shown is a neurone for which spike (Breer and Boekhoff, 1991; Boekhoff et al. 1990) and generation could not be detected at the soma. Stimulus applications amphibians (Schild and Lischka, 1994). The osphradium thus are indicated by the solid bars above the trace. appears to have the capability of perceiving both aquatic and non-aquatic olfactory stimuli. The opening of the epithelial tube is located in the proximity of the pneumostome; it is Jakob et al. 1989). Given the presence of subepithelial therefore under water when the animal is diving, whereas it is olfactory receptor neurones in the osphradium of the aquatic covered only by a very thin film of water while the animal is pulmonate snail (Benjamin and Peat, 1971), responses to breathing. The respiratory air flow or the fluid film above the amino acids were, therefore, not unexpected. opening of the osphradium might contain molecules that are It was, however, surprising that the osphradium responded relevant cues for the animal. 1752 H. WEDEMEYER AND D. SCHILD

The finding that odorants of the ‘cyclic AMP class’ and stimulus solution B were never rhythmic, unlike those of the odorants of the ‘InsP3 class’ appear to be processed differently nerve recordings, suggesting that ganglion cells other than in the osphradium of the pond snail L. stagnalis is interesting, those from which we recorded were responsible for the since the odorants of solution B have been shown to increase observed rhythm. the concentration of cyclic AMP in receptor neurones of other The evidence that ganglion cells responded to odorants does species (Sklar et al. 1986; Boeckhoff et al. 1990; Breer and not necessarily contradict Zaitseva’s (1982) finding that Boeckhoff, 1991), while the odorants of solution C are known receptor neurones project directly into the RIP nerve, because to increase the concentration of InsP3 in olfactory receptor the ganglion cells could possess ‘free nerve endings’ in the neurones (Boeckhoff et al. 1990; Breer and Boeckhoff, 1991). epithelium such as the enkephalin-positive cells of L. stagnalis InsP3 is thought to depolarize the receptor neurones by reported by Nezlin et al. (1994). However, it is also possible triggering the opening of plasma membrane calcium and cation that some receptor neurones or axon collaterals of receptor channels (Fadool and Ache, 1992; Schild et al. 1995; Restrepo cells make synaptic contacts with ganglion cells, so that both et al. 1994). It is thus tempting to hypothesize that a correlation primary and secondary olfactory signal processing could take between certain odorants and second messenger pathways is place within the osphradium. conserved among phyla; however, there are currently too few data available for too few species to ascertain the generality of References such a correlation. Yet another parallel between the osphradium of the pond BAILEY, D. F. AND BENJAMIN, P. R. (1968). Anatomical and electrophysiological studies on the gastropod osphradium. Symp. snail and the olfactory mucosae of vertebrates is the presence zool. Soc. Lond. 23, 263Ð268. of NO synthetase. NO appears to play a role in the transduction BAILEY, D. F. AND LAVERACK, M. S. (1963). Central nervous processes in olfactory receptor neurones of the rat and frog responses to chemical stimulation of a gastropod osphradium. (Breer et al. 1992; Lischka and Schild, 1993), and NO Nature 200, 1122Ð1123. synthetase may also be expressed in the epithelium of the BAILEY, D. F. AND LAVERACK, M. S. (1966). Aspects of the osphradium, in particular in the region where the receptor neurophysiology of Buccinum undatum L. (Gastropoda). I. Central neurones are located (Elofsson et al. 1993). responses to stimulation of the osphradium. J. exp. Biol. 44, Our results on the responses of the osphradium to changes 131Ð148. ENJAMIN in PO∑ confirm Kamardin’s (1976a) evidence that PO∑ can B , P. R. (1971). On the structure of the pulmonate influence the activity of the RIP nerve. However, there was osphradium. I. Cell types and their organisation. Z. Zellforsch. mikrosk. Anat. 117, 485Ð501. considerable variability in the nerve responses to low PO∑, and BENJAMIN, P. R. AND PEAT, A. (1971). On the structure of the 45 % of the osphradia did not respond to hypoxia at all. Mean pulmonate osphradium. II. Ultrastructure. Z. Zellforsch. mikrosk. activity consistently increased with elevated PCO∑ (Fig. 7C,D), Anat. 118, 168Ð189. although individual units were also suppressed (Fig. 7B). BOEKHOFF, I., TAREILUS, E., STROTMANN, J. AND BREER, H. (1990). Changes in pH between 6.8 and 7.8 did not induce changes in Rapid activation of alternative second messenger pathways in nerve activity, so the effect of hypercapnia did not appear to olfactory cilia from rats by different odorants. EMBO J. 9, be mediated by a change in pH. The osphradium seems to 2453Ð2458. contain cells sensitive to PCO∑, and it is the only organ known BREER, H. AND BOEKHOFF, I. (1991). Odorants of the same odor class that is sensitive to both PCO∑ and to different classes of activate different second messenger pathways. Chem. Senses 16, odorants. 19Ð29. BREER, H., KLEMM, T. AND BOEKHOFF, I. (1992). Nitric oxide The responses of osphradia to elevated PCO∑ could possibly serve as a sensory input to the respiratory central pattern mediated formation of cyclic GMP in the olfactory system. NeuroReport 3, 1030Ð1032. generator (Syed et al. 1990, 1991). This could be in accordance COPELAND, M. (1918). The olfactory reactions and organs of the with the observations of Jones (1961) that the diving duration marine snails Alectrion obsoleta (Say) and Bysicon canaliculatum of L. stagnalis depends on water oxygen concentration. (Linn.). J. exp. Zool. 25, 177Ð227. However, Janse et al. (1985), recording from the nervus analis, CROLL, R. P. (1983). Gastropod chemoreception. Biol. Rev. 58, showed that other peripheral and central neurones are involved 293Ð319. in oxygen perception. The possible effects of PCO∑-sensitive CROLL, R. P. (1985). Sensory control of respiratory pumping in afferents upon central neurones remain to be studied. Aplysia californica. J. exp. Biol. 117, 15Ð27. Our recordings clearly show that the ganglion cells of the DIONNE, V. E. (1992). Chemosensory responses in isolated olfactory osphradium respond to three different classes of odorants receptor neurons from Necturus maculosus. J. gen. Physiol. 99, (Table 1). Although we did not attempt to classify the 415Ð433. responsive ganglion cells with respect to their odorant DUCHAMP, A. AND SICARD, G. (1984a). Influence of stimulus intensity on odour discrimination by olfactory bulb neurons as compared specificities, they clearly showed overlapping sensitivities to with receptor cells. Chem. Senses 8, 355Ð366. the applied odorant solutions. This sensitivity overlap is a DUCHAMP, A. AND SICARD, G. (1984b). Odour discrimination by feature well-known in vertebrate olfactory receptor neurones olfactory bulb neurons: statistical analysis of electrophysiological (Sicard, 1990) and mitral cells in the olfactory bulb (Duchamp responses and comparison with odour discrimination by receptor and Sicard, 1984a,b). The responses of ganglion cells to cells. Chem. Senses 9, 1Ð14. Chemosensitivity of pond snail osphradia 1753

ELOFSSON, R., CARLBERG, M., MOROZ, L., NEZLIN, L. AND SAKHAROV, Blainv. (Polyplacophora), Siphonaria grisea L. et Siphonaria sp. D. (1993). Is nitric oxide (NO) produced by invertebrate neurones? (Gastropoda, Pulmonata). Mesogèe 48, 125Ð130. NeuroReport 4, 279Ð282. KOHN, A. J. (1961). Chemoreception in gastropod molluscs. Am. Zool. FADOOL, D. A. AND ACHE, B. W. (1992). Plasma membrane inositol 1, 291Ð308. 1,4,5-trisphosphate-activated channels mediate signal transduction KOSTYUK, P. G., DOROSHENKO, P. A., MARTYNYUK, A. E., in lobster olfactory neurons. Neuron 9, 907Ð918. VESELOVSKY, N. S. AND FEDULOVA, S. A. (1986). Calcium channels GELPERIN, A., RHINES, L. D., FLORES, J. AND TANK, D. W. (1993). and cyclic AMP metabolism in nerve cells. In Calcium Coherent network oscillations by olfactory interneurons: Electrogenesis and Neuronal Functioning (ed. U. Heinemann, M. modulation by endogenous amines. J. Neurophysiol. 69, Klee, E. Neher and W. Singer), pp. 9Ð16. Berlin: Springer. 1930Ð1939. LACAZE-DUTHIERS, H. (1872). Du système nerveux des mollusques GELPERIN, A. AND TANK, D. W. (1990). Odour-modulated collective gastéropodes pulmonés aquatiques et d’un nouvel organne network oscillations of olfactory interneurons in a terrestrial d’innervation. Arch. Zool. exp. gén. 1, 97Ð168. mollusc. Nature 345, 437Ð440. LANKESTER, E. R. (1883). Mollusca. In Encylopaedia Britannica, GOLDSCHMEDING, J. T., VAN DUIVENBODEN, Y. A. AND LODDER, J. C. edition 9, vol. 16, pp. 632Ð695. London: Encyclopaedia Britannica. (1981). Axonal branching pattern and coupling mechanisms of the LEVY, M., ACHITUV, Y. AND SUSSWEIN, A. J. (1989). Relationship cerebral giant neurones in the snail, Lymnaea stagnalis. J. between respiratory pumping and oxygen consumption in Aplysia Neurobiol. 12, 405Ð424. depilans and Aplysia fasciata. J. exp. Biol. 141, 389Ð405. HAMILL, O. P., MARTY, A., NEHER, E., SAKMANNN, B. AND SIGWORTH, LEVY, M. AND SUSSWEIN, A. J. (1993). Separate neural pathways F. J. (1981). Improved patch-clamp techniques for high-resolution respond to different noxious stimuli affecting respiratory pump current recording from cells and cell-free membrane patches. frequency in Aplysia fasciata. Brain Res. 616, 218Ð229. Pflügers Arch. 391, 85Ð100. LISCHKA, F. W. AND SCHILD, D. (1993). Effects of nitric oxide upon olfactory receptor neurones in Xenopus laevis. NeuroReport 4, HASZPRUNAR, G. (1985a). The fine morphology of the osphradial sense organs of the Mollusca. I. Gastropoda, Prosobranchia. Phil. 582Ð584. Trans. R. Soc. Lond. B 307, 457Ð496. LÓPEZ-LÓPEZ, J., GONZÁLES, C., UREÑA, J. AND LÓPEZ-BARNEO, J. (1989). Low PO∑ selectively inhibits K channel activity in HASZPRUNAR, G. (1985b). The fine morphology of the osphradial chemoreceptor cells of the mammalian carotid body. J. gen. sense organs of the Mollusca. II. Allogastropoda Physiol. 93, 1001Ð1015. (Architectonicidae, Pyramidellidae). Phil. Trans. R. Soc. Lond. B MICHELSON, E. H. (1960). Chemoreception in the snail Australorbis 307, 497Ð505. glabratus. Am. J. trop. Med. Hyg. 9, 480Ð487. HASZPRUNAR, G. (1987a). The fine morphology of the osphradial MIYAMOTO, T., RESTREPO, D. AND TEETER, J. H. (1992). Voltage- sense organs of the Mollusca. III. Placophora and Bivalva. Phil. dependent and odorant-regulated currents in isolated olfactory Trans. R. Soc. Lond. B 315, 37Ð61. receptor neurons of the channel catfish. J. gen. Physiol. 99, HASZPRUNAR, G. (1987b). The fine morphology of the osphradial 505Ð529. sense organs of the Mollusca. IV. Caudofoveata and Solenogastres. NEZLIN, L., MOROZ, L., ELOFSSON, R. AND SAKHAROV, D. (1994). Phil. Trans. R. Soc. Lond. B 315, 63Ð73. Immunolabeled neuroactive substances in the osphradium of the HULBERT, G. C. E. B. AND YONGE, C. M. (1937). A possible function pond snail Lymnaea stagnalis. Cell Tissue Res. 275, 269Ð275. of the osphradium in the Gastropoda. Nature 139, 840Ð841. RESTREPO, D. AND BOYLE, A. G. (1991). Stimulation of olfactory IVANOVA, T. T. AND CAPRIO, J. (1993). Odorant receptors activated by receptors alters regulation of [Ca]i in olfactory neurons of the amino acids in sensory neurons of the channel catfish Ictalurus catfish (Ictalurus punctatus). J. Membr. Biol. 120, 223Ð232. punctatus. J. gen. Physiol. 102, 1085Ð1105. RESTREPO, D., MIYAMOTO, T., BRYANT, B. C. AND TEETER, J. H. JAHAN-PARWAR, B., SMITH, M. AND VON BAUMGARTEN, R. (1969). (1990). Odor stimuli trigger influx of calcium into olfactory Activation of neurosecretory cells in Aplysia by osphradial neurons of the channel catfish. Science 249, 1166Ð1168. stimulation. Am. J. Physiol. 216, 1246Ð1257. RESTREPO, D., TEETER, J. H., MIYAMOTO, T., LISCHKA, F. W. AND JANSE, C., VAN DER WILT, VAN DER PLAS, J. AND VAN DER ROEST, M. SCHILD, D. (1994). Do vertebrate olfactory neurons respond to (1985). Central and peripheral neurones involved in oxygen InsP3 by activating different conductances? In Chemical Signals in perception in the pulmonate snail Lymnaea stagnalis (Mollusca, Vertebrates, vol. VII (ed. R. Apfelbach). Oxford: Elsevier (in Gastropoda). Comp. Biochem. Physiol. 82A, 459Ð469. press). JONES, J. D. (1961). Aspects of respiration in Planorbis corneus L. SCHILD, D. AND BISCHOFBERGER, J. (1991). Ca2+ modulates an and Lymnaea stagnalis L. (Gastropoda: Pulmonata). Comp. unspecific cation conductance in olfactory cilia of Xenopus laevis. Biochem. Physiol. 4, 1Ð29. Expl Brain Res. 84, 187Ð194. KAMARDIN, N. N. (1976a). Response of osphradium of mollusk SCHILD, D. AND LISCHKA, F. W. (1994). Amiloride-insensitive cation Lymnea stagnalis to different oxygen concentrations in water. J. conductance in Xenopus laevis olfactory neurons: a combined patch Evolut. Biochem. Physiol. 12, 427Ð428. clamp and calcium imaging analysis. Biophys. J. 66, 299Ð304. KAMARDIN, N. N. (1976b). The structure and cellular organization of SCHILD, D., LISCHKA, F. W. AND RESTREPO, D. (1995). InsP3 causes 2+ the osphradium in Lymnea stagnalis. Archs Anat. Histol. Embryol., an increase in apical [Ca ]i by activating two distinct current Moscow 71, 87Ð90. components in vertebrate olfactory receptor cells. J. Neurophysiol. KAMARDIN, N. N. (1983). Investigation of the homing behaviour of 73, 862Ð866. the lung snail Siponoria grisea L. Vestnik Leningradskovo SCHMIEDL-JAKOB, I., ANDERSON, P. A. AND ACHE, B. W. (1989). Universiteta 15, 101Ð104. Whole cell recordings from lobster olfactory receptor cells: KAMARDIN, N. N. (1988). Le rôle probable de l’osphradium dans le responses to current and odor stimulation. J. Neurophysiol. 61, homing des mollusques marins littoraux Acanthopleura gemmata 994Ð1000. 1754 H. WEDEMEYER AND D. SCHILD

SICARD, G. (1990). Receptor selectivity and dimensionality of odours analysis and the identiﬁcation of motor neurons. J. comp. Physiol. at the stage of the olfactory receptor cells. In Chemosensory A 169, 541Ð555. Information Processing (ed. D. Schild), pp. 21Ð32. Berlin: TOWNSEND, C. R. (1973). The role of the osphradium in Springer. chemoreception by the snail Biomphalaria glabrata (Say). Anim. SKLAR, P. B., ANHOLT, R. R. H. AND SNYDER, S. H. (1986). The Behav. 21, 549Ð556. odorant-sensitive adenylate cyclase of olfactory receptor cells. J. WELSCH, U. AND STORCH, V. (1969). Über das Osphradium der biol. Chem. 261, 15538Ð15543. prosobranchen Schnecken Buccinum undatum L. und SOKOLOV, V. A. AND KAMARDIN, N. N. (1977). The relation of impulse Neptunea antiqua (L.). Z. Zellforsch. mikrosk. Anat. 95, frequency in the osphradial nerve to the concentration of oxygen 317Ð330. and inulin in liquid passing over the osphradium of the pond snail. WÖLPER, C. (1950). Das Osphradium der Paludina vivipara. Z. vergl. Vestnik Leningr. Univ. Biol. 1, 87Ð90. Physiol. 32, 272Ð286. SPENGEL, J. W. (1881). Die Geruchsorgane und das Nervensystem der YONGE, C. M. (1947). The pallial organs in the aspidobranch Mollusken. Z. wiss. Zool. 35, 333Ð383. Gastropoda and their evolution throughout the Mollusca. Phil. STINNAKRE, J. AND TAUC, L. (1969). Central neuronal response to the Trans. R. Soc. Lond. B 232, 443Ð518. activation of the osmoreceptors in the osphradium of Aplysia. J. ZAITSEVA, O. V. (1982). Structural organization of the exp. Biol. 51, 347Ð361. chemoreceptory organs in the snail (Lymnaea stagnalis). Archs SYED, N. I., BULLOCH, A. G. M. AND LUKOWIAK, K. (1990). In vitro Anat. Hist. Embryol. 23, 55Ð69. reconstruction of the respiratory central pattern generator of the ZIPPEL, H. P., LAGO-SCHAAF, T. AND CAPRIO, J. (1993). Ciliated mollusk Lymnaea. Science 250, 282Ð285. olfactory receptor neurons in goldﬁsh (Carassius auratus) partially SYED, N. I., HARRISON, D. AND WINLOW, W. (1991). Respiratory survive nerve axotomy, rapidly regenerate and respond to amino behaviour of the pond snail Lymnaea stagnalis. I. Behavioural acids. J. comp. Physiol. A 173, 537Ð547.