Sensory innervation of the ovotestis in the snail Helix aspersa

Oviposition in land snails requires the successful completion of several metabolically costly processes, including oogenesis, albumen production, egg calcification and nest excavation (Tompa,1984). Furthermore, oviposition is strongly influenced by environmental factors such as climate, photoperiod, competition and parasitism(Chase, 2002; Wayne, 2001). Taken together,these considerations suggest that egg laying is a resource-limited process that should occur relatively infrequently. Indeed, even under the optimized conditions of a breeding farm, only 13% of snails Helix aspersaoviposit more than once every month, while the majority (56%) oviposit no more frequently than once every three months(Daguzan, 1981). Clutch sizes in H. aspersa vary considerably between populations and under different conditions, with reported means ranging from 59(Koene and Chase, 1998) to 108(Daguzan, 1981).

Obviously, as clutch size increases, so too does the energetic cost to the female function. It is important to note, however, that while clutch sizes vary, each oviposition event is coupled to a fixed minimum cost. Factors contributing to the fixed cost of egg laying include the search for adequate soil conditions (travel distances of up to 15 m in Helix pomatia; Pollard, 1975), nest excavation and the generation of hydrostatic pressure to expel the egg mass(Perrot, 1938). Since the fixed costs are high, it is reasonable to attribute the infrequency of oviposition to the fixed costs. The cost per egg can be reduced, however, by producing large clutches. Furthermore, as clutch size increases, so too does the offspring survival rate because large clutches are less prone to desiccation (Bayne, 1968). On the other hand, predation risks may favour more frequent depositions of smaller clutches at multiple sites. Ultimately, the number of eggs laid by an individual is determined by balancing energetic costs against reproductive benefits. We propose that oviposition events occur when the clutch size surpasses a minimum value for which the reproductive benefit exceeds the total of fixed and variable costs. Therefore, a prerequisite for egg laying should be the availability of the threshold number of mature oocytes. In the present study, we address the question of how H. aspersa is able to monitor its oocyte production to ensure the presence of this minimum number of gametes before it initiates the costly cascade of physiological events that results in oviposition.

In 1914, Schmalz described the innervation of the ovotestis (OT) in H. pomatia by a branch of the intestinal nerve. The functional significance of this innervation, however, remains unknown. Similarly, innervation of the mammalian ovary has been described in several species, including pigs(Majewski et al., 2002), rats(Burden and Lawrence, 1977) and humans (Hill, 1962), but information regarding the function of such innervation is sparse. Anatomical clues and lesion studies suggest a sensory role possibly related to follicular recruitment (Aguado, 2002),steroidogenesis (Kawakami et al.,1981) or blood flow regulation(Ojeda and Lara, 1989). Here,we describe further details of the innervation of the ovotestis in H. aspersa, and we demonstrate a sensory function. We conclude that sensory endings in the gonad respond to the volumetric increase that occurs with oocyte maturation and that the neural signal is relayed to the central nervous system (CNS) via a low-frequency, tonic discharge of afferent spikes.

Adult specimens (curled shell lips) of the land snail Helix aspersa Müller were obtained from Mary's Garden (Strathmore, CA,USA). Snails were kept in large Lucite boxes (36 cm×36 cm×8 cm) at 18-21°C under a reversed 16 h:8 h light:dark photoperiod. They were cleaned every 2-3 days and fed a diet of powdered grains and chicken feed supplemented with calcium.

The innervation of the gonad was examined after labelling the ovotestis branch of the intestinal nerve with neurobiotin (Vector Labs, Burlington, ON,Canada). The nerve branch was cut at its point of entry into the ovotestis. The distal nerve stump was sucked into a glass capillary that was then filled with 8% neurobiotin in 0.1 mol l^-1 Tris buffer (pH 7.6). Infiltration was allowed to continue for 24 h. After fixation in 4%paraformaldehyde, the preparation was incubated for 16 h in 4% Triton X-100 in 0.1 mol l^-1 phosphate-buffered saline (pH 7.4). For visualization,we used the Vectastain ABC kit (Vector Labs). Preparations were initially examined as wholemounts, then cut as 20 μm frozen sections.

For electron microscopy, a nerve segment was fixed in 5% glutaraldehyde in 0.1 mol l^-1 sodium cacodylate (pH 7.2) and dehydrated in acetone. Epon sections were stained with 4% uranyl acetate and Reynolds lead citrate(Reynolds, 1963). Photomicrographs were digitized, then analysed using Sigma Scan software (SPSS Inc., Chicago, IL, USA). The perimeter of every fibre profile was individually traced using a mouse-driven cursor. For each profile, we calculated the diameter (feret diameter) of a hypothetical circular object with an area equivalent to the measured profile.

Nerve recordings were obtained from reduced preparations consisting of the CNS, the albumen gland, the seminal vesicle and the OT embedded in the digestive gland (Fig. 1). The entire preparation was pinned down in a Sylgard-coated recording dish comprising a CNS chamber, an OT chamber and a central gap(Fig. 1). The intestinal nerve was desheathed and passed through small slits in the Sylgard walls of the central gap that were subsequently sealed with Vaseline. During recordings,either saline or an isotonic sucrose solution was perfused through the central gap at a flow rate of approximately 10 ml min^-1, while the recording chambers were always filled with snail saline(Kerkut and Meech, 1966).

To avoid possible artefacts associated with cut nerves, we recorded from the intact intestinal nerve using en passant glass suction electrodes. To assess whether recorded spikes were afferent or efferent, it was necessary to record simultaneously from at least two locations on the nerve (Fig. 1). In the OT recording chamber, one distal electrode and one proximal electrode (hereafter referred to as SV-D and SV-P, respectively) were placed on the nerve adjacent to the seminal vesicle portion of the hermaphroditic duct. There was no branching of the nerve between SV-D and SV-P, and nearly all recorded action potentials were detected at both electrode sites. A third central electrode was placed on the nerve in the CNS recording chamber (CNS electrode). In some experiments, a fourth electrode was placed in the OT chamber to monitor activity from the proximal stump of the pericardial branch of the intestinal nerve (PC electrode). The amplified signals were digitized using the Digidata 1322A converter and viewed using Axoscope 8.0 software (Axon Instruments,Union City, CA, USA).

A naive participant identified afferent spikes in the records. SV-D and SV-P traces were viewed simultaneously on a computer screen with an expanded time scale. A moveable vertical line provided by Axoscope was used to assess the intervals between pairs of spikes when both members of the pair were judged to be from the same unit based on waveform similarities and inter-spike intervals. A spike that appeared first at SV-D and second at SV-P was identified as afferent, whereas a spike appearing first at SV-P followed by SV-D was identified as efferent (see Fig. 4B). Since the number of afferent spikes was much less than the number of efferent spikes, the latter was determined by subtraction. First,the total number of spikes with amplitudes exceeding 100% of the noise level was counted using Mini Analysis software (Synaptosoft, Decatur, IL, USA). Then, the number of identified afferent spikes was subtracted from the total to give the number of efferent spikes. To facilitate the identification of afferent spikes, most of the efferent activity was suppressed during recording periods devoted exclusively to assessing afferent discharges; this was accomplished by perfusing sucrose through the gap. The directional nature of spikes in the pericardial nerve branch was not considered because these recordings were obtained from the proximal nerve stump; Mini Analysis was used to obtain the total spike count.

To test for sensory responses, a wooden applicator stick was held by a micro manipulator, and its cut end (2 mm diameter) was lowered onto the centre surface of the OT. The probe covered approximately 50% of the OT surface area. Responses were evoked when the probe was moved laterally across the surface of the OT using a knob on the manipulator. A back-and-forth movement was employed, the total duration of which was 3 s. Punctate stimuli of shorter duration were ineffective.

After completing the electrophysiological recordings, fine forceps were used to break apart the connective tissue covering the OT. Suction from a 1 ml micropipette was used to collect the OT homogenate and transfer it to a test tube. By using coded tubes and examining homogenates in groups of two or more,the oocytes were counted without bias. Small volumes (1.4 ml) of the OT extract were successively transferred to a shallow glass dish and viewed under a dissecting microscope at 20× magnification. A grid system was used to ensure that each oocyte was counted only once. Each oocyte was measured on two perpendicular axes and the resulting mean was taken as the cell's diameter. Only mature oocytes at the vitellogenic stage (diameter >100 μm) were counted (Griffond and Bolzoni-Sungur,1986). Sperm were not counted because they are stored in the seminal vesicle not the OT.

To ensure an adequate representation of snails with high oocyte counts,some individuals were selected after they exhibited nest-excavating behavior(a prelude to egg laying), since oviposition is performed only by snails with a large store of mature oocytes. To obtain nest-excavating snails (hereafter referred to as `diggers'), a group of approximately 50 snails was transferred to a Lucite box that contained 5-7 cm of moist potting soil and sand in a 1:1 mixture. Diggers were removed before they could oviposit, and they were used for electrophysiological experiments within 3-5 days.

Counts of oviposited eggs were obtained after placing snails on soil as described above. Once a snail had departed from its nest, the deposited eggs were removed, washed and counted.

The major features of the OT, as illustrated in Fig. 2A, are the acini, in which the gametes develop, and the ductal system through which the gametes exit the gonad. The organ itself is embedded in the digestive gland, with only the top surface exposed after dissection. Oocytes appear dark in Fig. 2A and can be seen lying at the distal ends of the exposed acini. A fine branch of the intestinal nerve travels along the portion of the hermaphroditic duct known as the seminal vesicle and terminates within the OT, as originally described by Schmalz(1914). As the nerve extends distally along the duct, fibres branch laterally to innervate structures within the OT (Fig. 2B). These fine branches disperse throughout the organ, and some are visible within the acinar walls (Fig. 2C). In none of our five successful neurobiotin backfill preparations did we observe any peripheral cell bodies.

We examined the OT branch of the intestinal nerve in cross sections at the level of the seminal vesicle or just before it enters the OT. Here, the nerve measures only 20 μm in diameter but it contains 3025 fibre profiles(Fig. 3A). Two types of fibres appear to be partitioned in distinct regions of the nerve. The smaller of these regions is approximately 5 μm×5 μm, contains approximately 2250 fibre profiles and is surrounded by darkly stained glial processes(Fig. 3B). Nearly all of these profiles measure <0.25 μm in diameter. Larger fibres comprise the remaining 80% of the cross-sectional area. Because many of the larger profiles are irregularly shaped, the dimensions of all fibres were measured as feret diameters. Overall, 57% of the fibres have diameters of <0.20 μm; the largest diameters are 2.1 μm. It is likely, based on numerous studies in other animals, that the smaller fibres serve sensory functions, while the larger fibres have motor functions.

To address a possible sensory role of OT innervation, we first examined the effect of OT stimulation on spike activity in the intestinal nerve. Using SV-D, SV-P and CNS electrodes (Fig. 1), we recorded a significant increase in spike frequency following tactile stimulation of the OT(Fig. 4A). Quantification of afferent spikes was performed on the basis of spike timing at SV-P and SV-D electrodes (Fig. 4B). Our reason for preferring this laborious procedure over the alternative of recording from a distal nerve stump is that early recordings of the latter type exhibited prominent bursting patterns and much greater spike rates than did recordings from an intact nerve. We concluded that the nerve lesion caused a persistent increase in excitability, yielding anomalous results. With the intact nerve, OT stimulation resulted in a 20-fold increase in afferent spike frequency (Fig. 4C). Concomitantly, we observed a 40-fold increase in overall spike frequency at the CNS electrode. These results indicate a sensory function for the OT innervation.

To further elucidate the sensory role of innervation we examined spontaneous afferent spike rates in relation to the number of mature oocytes present in the OT. The frequency of afferent spikes varied from 1.0 spikes min^-1 with 31 oocytes present to 79.2 spikes min^-1 with 118 oocytes present. In a sample of 14 snails, the afferent spike rate was strongly correlated with the number of oocytes in the OT(Fig. 5A). A linear regression analysis on log-transformed data revealed a highly predictive relationship(r²=0.812, P<0.001; Fig. 5B). This result suggests that information regarding oocyte number in the OT reaches the CNS via the intestinal nerve. Furthermore, we found a significant relationship between log afferent spike frequency and log efferent spike frequency (r²=0.623, P<0.001; Fig. 5C), suggesting that afferent spikes originating in the OT reach the CNS and cause increased firing in neurons that project back to the OT. Fig. 5A,B shows that animals selected as nest diggers had significantly higher oocyte counts than did non-diggers. Diggers had a mean (±s.e.m.) oocyte count of 116.2±4.3 while non-diggers had a mean oocyte count of 53.8±14.2. These numbers are in line with the 86.9±2.7 fertilized eggs deposited during 104 oviposition events observed in our laboratory colony of snails. Notably, we observed a jump in the frequency of afferent activity when the number of oocytes approached 87(Fig. 5A). The mean afferent spike frequency in all animals with fewer than 87 oocytes was 4.23 spikes min^-1, whereas the mean frequency in animals with oocyte counts exceeding 87 was 44.5 spikes min^-1. Thus, the data suggest the presence of a threshold oocyte value below which afferent spike activity is low and above which spike frequency increases substantially.

Because oocytes grow to large sizes (up to 250 μm), at which point their diameters exceed that of the acini and ducts within which they develop, it can be assumed that they will eventually exert pressure on the innervated walls of the acini and ducts. Therefore, to examine the mechanism of sensory excitation, we artificially inflated the OT to simulate the expected volumetric expansion (N=3). First, we cut the hermaphroditic duct at the junction of the OT and the seminal vesicle. We then inserted a cannula into the distal portion of the duct and injected fast green dye (0.01 g per 100 ml saline) into the OT. For the experiment illustrated in Fig. 6, we used an animal that had very few (17) oocytes. Injections of 10 μl or 20 μl were made incrementally at 5 min intervals. As can be seen by the traces shown in Fig. 6A, the combined afferent and efferent activity increased significantly following the injections. When afferent spikes were analysed separately (raster display in Fig. 6A; Fig. 6B), the effect of OT inflation was even more apparent. While the afferent spike frequency prior to OT inflation was 4.5 spikes min^-1, this value increased 10-fold to 47 spikes min^-1 after a total of 40 μl saline had been injected. These data are consistent with the idea that the high levels of afferent spiking observed in animals with high oocyte counts are caused by a volumetric expansion of the OT associated with oocyte growth.

Given that the OT branch of the intestinal nerve responds to sensory stimulation at the periphery, we looked for an associated reflex output from the CNS. The intestinal nerve branches twice prior to its innervation of the OT, and the largest of these early branches innervates the pericardium(Schmalz, 1914). We recorded from the central stump of the severed pericardial branch while mechanically stimulating the surface of the ovotestis (N=6). Fig. 7A shows that stimulation induced a surge in spike frequency detectable at the CNS, seminal vesicle and pericardial electrodes. In the OT branch, the afferent activity increased 4-fold after stimulation, while the concurrent efference increased 6.7-fold(Fig. 7B). In the central stump of the pericardial nerve branch, the spike rate increased 3.8-fold(Fig. 7B). These results demonstrate that OT sensory stimulation effectively elicits a reflexive response from the CNS. Moreover, it is noteworthy that the efferent response was sustained for more than 10 min following the brief stimulus.

Knowledge that the ovotestis is innervated immediately raises questions of function, and our study has begun to provide some answers. Our results show that the innervation has at least a sensory function and probably a motor function as well. The most direct evidence of a sensory function is the finding that brief tactile stimulation of the organ causes a discharge of action potentials that is conducted centrally from a point near the nerve's insertion into the OT. However, the anatomical placement of the OT deep within the apical whorls of the hard shell makes it unlikely that the OT would ever experience mechanical stimulation in vivo. Rather, we propose that the natural stimulus for the nerve's sensory endings is the oocytes growing within the OT. Consistent with this idea is the strong correlation between afferent spike rates and the number of ripe oocytes. In fact, the log-linear relationship between these measures is in accord with the assumption that eggs are the natural stimulus. As we explain below, our results suggest that eggs generate at least two types of sensory signals for integration in the CNS.

It is remarkable that the OT branch of the intestinal nerve, which measures only 20 μm in diameter, should contain 3025 axon profiles. This is made possible by the inclusion of many fibres that approach the smallest size reported for any animal (0.05-0.1 μm). Another intriguing aspect of the morphology is that the total number of profiles far exceeds that expected from the size of the nerve, in relation to the sizes and number of other nerves,and the number of neurons in the CNS. The central neuronal population in Helix aspersa is estimated to be ∼40 000 neurons(Chase, 2001; this value excludes the small olfactory interneurons in the procerebrum). Thus, as a first approximation, the number of axon profiles in the OT nerve is approximately 8% of the number of neurons in the CNS. However, because the circumference of the OT nerve is much less than 8% of the total circumference of all nerves combined, there is an apparent paradox. Possibly, the majority of the axons belong to peripheral sensory cells that, inexplicably, were not labeled when the OT nerve was back-filled with neurobiotin. Another possible explanation is provided by Pin and Gola(1984), who demonstrated, by Lucifer Yellow injections and novel electrophysiological experiments, that certain central neurons in H. pomatia send as many as 25 axons, in parallel, into the intestinal nerve. If this were generally true, then the 3025 axons in the OT would represent only 121 central neurons. However, this would raise the question, why do the axons branch so profusely before reaching their target organ?

For the reasons presented in the Introduction, it would appear that H. aspersa delays laying eggs until it has a sufficient store of ripe oocytes. Our results demonstrate a close association between the number of ripe oocytes and the rate of spontaneous afferent discharge in the OT nerve branch (Fig. 5). Furthermore,there is a sharp increase in afferent activity when the oocyte count exceeds 87, which is the mean number of eggs actually laid by snails in our laboratory colony. We conclude that one function of the OT innervation is to signal the availability of a minimum quantity of ripe oocytes. Since oviposition requires that certain other conditions are also satisfied (e.g. soil composition and environmental moisture), we view the OT afferent activity as providing a permissive signal for oviposition, not a trigger.

The results from our OT inflation experiment(Fig. 6), together with the anatomical data, suggest that the fine terminals of the OT nerve branch contain stretch-sensors capable of monitoring oocyte growth. The oocytes within the OT grow substantially as they mature from the early oogonia stage to the final vitellogenic stage, increasing 10-fold in diameter and 1000-fold in volume (Griffond and Bolzoni-Sungur,1986). It is likely that this growth, which can occur concurrently in over 100 oocytes, results in a stretching of the acinar walls and the neuronal processes that lie within. From Fig. 2, it can be seen that the diameter of the largest oocytes (250 μm; Fig. 2A) is greater than the diameter of a typical acinus (∼170 μm; Fig. 2C). However, the oocytes are nonetheless able to move in the acini and in the hermaphroditic duct because their shape is plastic. Similar examples of volumetric expansion causing sensory discharges in gastropod nerves have been reported for the anterior gut of Aplysia (Susswein and Kupfermann, 1975), the pro-oesophagus of Lymnaea(Elliott and Benjamin, 1989)and the prostate gland of Lymnaea(De Boer et al., 1997).

The frequency of afferent spikes is strongly associated with the frequency of efferent spikes conducted towards the OT(Fig. 5C), suggesting that, as the oocytes mature, the heightened afferent signal mediates the increased efferent signal. Here, we are discussing only the spike activity that occurs continuously and spontaneously; in the section below, we discuss the phasic responses following mechanical stimulation. If the efferent activity is continuous in vivo, it is probably not related to an imminent oviposition event. Rather, the efferent fibres might release a trophic factor that mediates the final stages of oocyte maturation. Or, the efference could provide a motor command to maintain tonus in the radial muscles surrounding the hermaphroditic duct, thereby retaining the oocytes until ovulation.

While the presence of mature eggs causes a tonic afferent discharge in the OT branch of the intestinal nerve, tactile stimulation causes a phasic response (Figs 4, 7). Just as the tonic afferent discharge causes a tonic efferent discharge, the phasic afferent response causes a phasic efferent, or reflexive, discharge(Fig. 7). In both cases, the physical isolation of the CNS in our recording dish allows us to rule out mediation by diffuse chemical signals. We speculate that the natural stimulus for the phasic response is ovulation or, specifically, the passage of oocytes from their sites of attachment on the walls of the acini into distal segments of the hermaphroditic duct. We further propose that the sensory signal generated by the movement of oocytes is responsible for orchestrating subsequent events in the multi-stage process of oviposition, in a manner similar to that proposed for the control of egg laying in Aplysia(Cobbs and Pinsker, 1982; Ter Maat and Ferguson,1996).

Oviposition is a process that requires nearly 48 h of heightened metabolic activity (Perrot, 1938). A snail spends the majority of this time with its head beneath the soil surface in an extended, forward position. Thus, in addition to meeting the metabolic load, an ovipositing snail must increase its cardiac output to elevate its internal hydrostatic pressure. Cardiac output also plays an essential role in the mobilization of calcium for eggshell calcification(Tompa, 1984). Just prior to oviposition, bound and unbound fractions of blood calcium increase by approximately 60% (Tompa and Wilbur,1977). As in vertebrate and other molluscan species, the heart of H. aspersa is myogenic but subject to central nervous control(Chase, 2002). Therefore, to achieve the elevated heart rate necessary to accommodate the hydrostatic and metabolic demands of oviposition, CNS excitatory output to the heart is expected to increase. Consistent with this view, our experiments detected a dramatic acceleration of activity in the pericardial branch of the intestinal nerve following brief mechanical stimulation of the OT(Fig. 7). Also, electrical stimulation of the OT branch of the intestinal nerve produces a robust increase in heart beat amplitude (D. Weatherill, unpublished).

The OT of H. aspersa is another target of the efferent reflex elicited by mechanical stimulation. Studies in marine bivalves indicate that gonadal innervation is implicated in the maturation of oocytes and the induction of ovulation (Matsutani and Nomura, 1987; Ram et al.,1996). Studies in our laboratory are attempting to elucidate the role of efferent activity in the OT nerve branch of Helix. Thus far,neurobiotin labelling has revealed dense, varicose innervation of the distal hermaphroditic duct. Also, results indicate that efferent spikes, elicited by stimulation of the intestinal nerve, increase cilia beat frequencies and cause radial contractions in the hermaphroditic duct (E. Geoffroy, unpublished). These effects are consistent with a motor role for OT innervation in promoting the movement of oocytes from the OT to the fertilization chamber during the early stages of egg laying.

As we briefly summarized in the Introduction, there exists a substantial body of literature describing innervation of the mammalian ovary (reviewed in Aguado, 2002). A handful of such reports has characterized the innervation as sensory on the basis of anatomical clues, immunohistochemical results and lesion studies(Kummer et al., 1990;Majewski, 2002; Price and Mudge,1983). However, to our knowledge, the results presented here constitute the first physiological evidence for sensory innervation of the gonad in any species. The sensory innervation of the OT in H. aspersamay provide insights into the function of analogous innervations in humans and other mammalian species.

We thank E. Geoffroy, A. Reynolds and D. Weatherill for unpublished data,K. Dedovic for axon profile measurements and spike record analyses, E. Geoffroy and D. Weatherill for comments on an earlier version of the manuscript, and NSERC for its grant to R. Chase.