The Location of Circa-Tidal Pacemakers in the Estuarine Amphipod Corophium Volutator Using A Selective Chilling Technique

In recent years it has become increasingly apparent that the rhythmic behaviour and physiology of many animals may result from the interaction of a number of anatomically discrete pacemakers (see reviews by Aréchiga & Naylor, 1976; Aréchiga, 1977; Page, 1981b).

The eyes and optic tract are especially implicated in the control of circadian periodicity, but pacemakers have also been located elsewhere in the nervous system. In the opisthobranch mollusc Aplysia for example, each eye has been found to function as an independent pacemaker (e.g. Jacklet, 1969) capable of driving the circadian rhythm of locomotor activity shown by this animal (Lickey et al. 1977), but in the closely related Bursatella the eyes play only a minor role, the rhythm being under the control of extra-ocular pacemakers (Block & Roberts, 1981).

In the arthropods, pacemakers have been located in the paired optic lobes of insects (see review by Page, 1981a; Fleissner, 1982), but there is also evidence of extra-ocular pacemakers, possibly in the suboesophageal ganglion (Harker, 1974; Page, 1981a), and the persistence of rhythmic deposition of endocuticle in decapitated locusts (Neville, 1965) and cockroaches (Lukat, 1978) provides further indication of at least one other endogenous clock.

Amongst crustaceans the rhythmic migration of retinal pigments in the crayfish Procambarus is regulated by the protocerebrum (Barrera-Mera, 1976) but the centre part of the optic tract must remain intact to carry information from the brain to the effectors (Larimer & Smith, 1980). The protocerebrum is also responsible for controlling the circadian rhythm of locomotor activity in the crayfish (Page & Larimer, 1975) although secondary oscillators may also be involved (Gordon, Larimer & Page, 1977).

The search for a circa-tidal oscillator has been confined to another decapod crustacean, the shore crab Carcinus, in which chilling the eyestalk alone rephases the rhythm of locomotor activity (Naylor & Williams, 1968) presumably by affecting the release of neuroendocrine material from the eyestalk itself (Aréchiga, Huberman & Naylor, 1974). Many other malacostracans show equally well defined activity rhythms of tidal periodicity (e.g. Enright, 1965; Jones & Naylor, 1970; Fincham, 1970; Fish & Fish, 1972), including the estuarine amphipod Corophium volutator (Morgan, 1965; Holmstrom & Morgan, 1983a). In this animal the ebb-tide rhythm of swimming activity is phase shifted by pulses of low temperature, the phase delay induced being proportional to the duration of the cold pulse (Holmstrom & Morgan, 1983b).

The work described below utilizes this hypothermal sensitivity in an attempt to locate the anatomical site of the clock by selectively chilling different parts of the animal’s body. In Corophium the compound eyes are sessile structures and do not appear to be associated with the neuro-endocrine system, but in view of the pacemaking properties attributed to the eyes of certain marine gastropods (Jacklet, 1969; Lickey et al. 1977; Block & Davenport, 1982), experiments on eye ablation have also been performed.

Large numbers of Corophium volutator were collected by sieving mud at Tites Point, Purton on the Severn Estuary (O.S. 692047) and transported to a constant temperature room (12 ± 1 °C) in the Zoology Department at Birmingham. Here the animals were kept under constant illumination (3·45 μεm⁻²s⁻¹) in gently aerated sea water (Tropic Marin neu, Tropicarium Buchslag, W. Germany) of 30 ‰ held in clear plastic aquaria. All experiments were initiated within 18 h following collection and the animals were denied food and substrate during this period.

Low temperature pulses were applied by transferring the animals to pre-cooled sea water held at the appropriate temperature, as described by Holmstrom & Morgan (1983b). Selective chilling was achieved by applying a cooled wire to selected body sites. Corophium were harnessed across loops of tinned copper wire (0·711 mm diameter) that projected through a 10 mm thick plastic covered, expanded polystyrene block into a bath of ethylene glycol cooled to sub-zero temperatures by continuous circulation through a Grant waterbath (Grant Instruments Ltd., Cambridge, England) (Fig. 1A).

Preliminary experiments had shown Corophium to be intolerant of the large temperature differential which developed along the body following localized chilling at comparatively high ambient temperatures (approximately 18–22 °C). The glycol bath was thus mounted in a box of expanded polystyrene and the air temperature regulated by controlling the flow of coolant through a glass coil arranged immediately above the to which the animals were attached. A glass plate placed over the top of the box completed the insulation. The animals were covered with filter paper moistened with sea water for the duration of the cold period. In this way up to 20 animals could be treated simultaneously with a localized cold pulse, while a duplicate arrangement, without the cooling apparatus, allowed a further twenty to be harnessed as controls. Holmstrom & Morgan (1983b) demonstrated that cold pulses given at low tide caused arrhythmicity, so low temperature pulses of 3 h duration were restricted to the high tide period to induce phase delays.

The thermal gradient along the body away from the wire was monitored directly using a fine thermocouple attached to a Comark electronic thermometer (Comark Electronics Limited, Littlehampton, England) and by a thermoscanning technique using a Barnes (Barnes Electronics, U.S.A.) infrared image analyser. In the latter case it was necessary to use a small unit for selective chilling fitted to the stage of the analyser. This consisted of a small Perspex box (54×30×20 mm) containing antifreeze and insulated with 4 mm expanded polystyrene. Two side arms provided the support for a Perspex plate (74 × 14 × 1 mm) with a centrally placed hole (2·25 mm radius). The box was pre-cooled in a deep freeze and a single animal fastened ventral side down across the hole by two strips of adhesive tape. A silver wire (0·5 mm diameter pared down to 0·3 mm at the tip) transferred the cold from the antifreeze to a selected point on the head or body.

The eyes of Corophium, which appear to be the typical gammarid type as defined by Hallberg, Nilsson & Elofsson (1980), were ablated using a silver wire (0·5 mm diameter) heated to red heat by means of a nickel wire connected to a 12 V supply. The success of eye ablation was tested by observing the response of treated individuals to lateral unidirectional light. Since actively swimming Corophium are strongly photopositive (Meadows & Reid, 1966) the removal of the eyes precludes this response.

Swimming activity was monitored by time-lapse photography using the apparatus described by Beeston & Morgan (1979) while animals were observed in groups of ten. The single frame advance of a Bolex-Paillard ciné camera was triggered by an intervalometer programmed to allow at least four frames, separated from each Qther by a 10-s delay, to be taken at intervals of 12 min. The film was analysed with the aid of a photographic enlarger and a high intensity spot lamp. The percentage of animals swimming in all four frames was taken as the activity level at that time.

The gross morphology of the anterior nervous system was investigated by dissecting Corophium, freshly killed in 5% sea water formalin solution, using a stereomicroscope. Methylene blue was used occasionally to stain the nervous tissue.

In any experiment on localized chilling it is necessary to determine the extent to which the cold permeates through the body away from the point of application. Thus in an initial series of experiments the animals were secured with their heads resting on the chilled wire and the thermal gradient along the surface of the body was measured with a thermocouple. Observations were made on at least five animals in each of three different combinations of air and wire temperature and the results are shown in Fig. 2.

The temperature rises sharply with increasing distance from the cold source in all three experiments, the gradient being proportional to the difference in temperature between the wire and the air immediately above it. However under none of the experimental conditions investigated do sub-zero temperatures extend more than 1 mm beyond the point of application of the cold, and the temperature may be seen to have risen beyond the values (−4·51 to −1·5 ± 0·5 °C) known to reset the rhythm (Holmstrom & Morgan, 1983b) within about 0·5 mm of the wire. The gradient the found to be stable for periods of at least 3 h, the longest cold pulse used in these experiments.

These observations were confirmed in a second series of experiments in which a point source of cold was applied with the wire tip positioned on the animal’s head and the thermal gradients determined with an infrared image analyser. The temperature of the wire, 9 °C, was higher than that used in the chilling experiments but at a room temperature of 20 °C an approximately comparable gradient existed about the cold source. Fig. 3 shows a typical computer enhanced thermoscan of Corophium chilled in this way. Compared to the surrounding air the thermal conductivity of animal tissue is relatively high, as may be seen from the steep gradient around the periphery of the animal. To look at the gradient more closely a temperature profile along a line bisecting the head, A—B, was constructed by the computer (Fig. 4A). The temperature rises (between 3–4 °C mm⁻¹) with increasing distance from the cold source and, assuming similar thermal conductivity under the conditions of the experiment, temperatures at which rephasing would normally be expected to occur (i.e. − 1·5 ± 0·5 °C) would be limited to an area extending 0·5 mm from the wire.

In a further experiment, the head was chilled in the region of the brain and the temperature recorded with thermocouples implanted in the brain and sub-oesophageal ganglion as the wire rewarmed after chilling to −6 °C (Fig. 4B). The sub-oesophageal ganglion was warmer by 1·5–2·5 °C throughout and at a wire temperature of −3 °C, that used in the current rephasing experiments, the temperature of the sub-oesophageal ganglion did not fall below zero.

In comparison with the decapods and some of the other Malacostraca, the nervous anatomy of amphipods has been poorly studied, and the nervous organization of Corophium appears unrecorded. Examination of methylene blue stained preparations however suggests that its anterior nervous system corresponds in most respects to the general amphipod organization described by Stephensen (1940) and Ingle (1969). Five specimens of body length 7 ± 0·5 mm were used to obtain measurements of the distances between the anterior ganglia. The supra-oesophageal ganglion lies immediately below the dorsal surface and, on average, was found to be 0·35 ± 0·02 mm wide by 0·20 ± 0·03 mm long. The sub-oesophageal ganglion was found to be slightly smaller (0·23 ± 0·02 by ·26 ± 0·03 mm) being located approximately 0·5 mm posteriorly and about 0·3 mm below. The sub-oesophageal ganglion is distinguished from the other segmental ganglia by increased dorso-anterior development, its larger size, and by a third pair of nerve roots. The organization of the anterior nervous system and the positions of the brain and sub-oesophageal ganglion in relation to the cooling wire are shown in Fig. 5.

The thermal tolerance of Corophium varies seasonally (Holmstrom, Grout & Morgan, 1981) and the optimum conditions for selective chilling were found to fluctuate similarly. The experiments reported here were carried out during the summer months (June), when the head region alone of freshly collected Corophium was subjected a 3-h cold pulse starting at the time of the predicted high tide. The animals were harnessed dorsal surface down so that the region of the supra-oesophageal ganglion lay directly across the wire at −3 °C (see Fig. 5), the ambient temperature of the experimental chamber being 7 °C. At the end of the pulse the animals were transferred to the observation chamber and their swimming behaviour recorded. The results are shown in Fig. 6, together with the activity rhythm of animals completely chilled to −2 °C for a similar period, a control group similarly harnessed but to an uncooled wire, and freshly collected Corophium.

Both groups of restrained animals show a lower level of activity than their counterparts but a rhythmic pattern is still apparent in their swimming activity. Chilling the whole animal and chilling the supra-oesophageal region alone delays the times of onset of activity and of maximum activity relative to those of untreated Corophium but no such delay is evident in the activity of restrained animals which had not been subjected to localized chilling.

A statistical comparison (Anovar) of the mean phase delays (Table 1) of the experimental and control groups relative to the rhythm shown by untreated freshly collected animals indicated significant differences between the four groups, peak values being taken as a phase reference point. Chilling the whole animal and chilling the supra-oesophageal ganglion significantly delay the time of maximum swimming relative to the untreated animals and to the unchilled controls.

There is some evidence for a change in the wave form of the rhythm following selective chilling of the brain alone. The first activity peak shown on recovery from such treatment is clearly bimodal (Fig. 6D) although the rhythm subsequently reverts to a unimodal pattern. This transitory phase-splitting was also evident in a replicate experiment in which the animals were similarly chilled, although in both cases animals chilled in toto did not show such a split.

In a second series of experiments ten Corophium were secured ventral surface down and the region of the sub-oesophageal ganglion chilled to —2 °C for 3 h at an ambient temperature of 7 °C. The swimming behaviour of this and the three other control groups is shown in Fig. 7.

All four groups of animals show a significant tidal rhythm although this is now more distinct after both chilling treatments. Chilling either the whole body or the ventral surface of the head again rephases the rhythm by a period equivalent to the duration of the pulse.

Single factor analysis of variance on the phase delays indicates significant variation between the treatments, and the mean delays are shown in Table 2.

In a further series of experiments cold pulses were applied to areas of the head a short distance away from each of the cephalic ganglia. These experiments were carried out later in the year (September), using larger animals. Initially the ventral area 15 mm posterior to the sub-oesophageal ganglion, i.e. in the region of the third segmental ganglion, was chilled to −3 °C at an ambient temperature of 8 °C. The result of this experiment together with the activity pattern obtained from the three controls are shown in Fig. 8.

The activity level shown by the experimental animals is low but ebb-tide swimming is still evident. The phase of the rhythm shown by the selectively chilled group is now not significantly different from that of the untreated or restrained animals although the rhythm of the group chilled in toto is significantly rephased (P < 0·001). Similarly, no phase delay was observed to local chilling in a second experiment in which the dorsal area 1·5 mm posterior to the supra-oesophageal ganglion was chilled to −3 °C, as before. Single factor analysis of variance comparing the phase relative to high water in each of the four groups of animals used in this experiment indicated no significant (P>0·05) difference between the four groups. In a further experiment, instead of chilling whole animals to −3 °C, a temperature known to rephase the rhythm, an equivalent group of Corophium was chilled to just 0 °C, the minimum temperature the body tissues would be expected to reach at a distance from the cooling wire equivalent to that between the two cephalic ganglia. Again no phase delay was observed in the rhythmic swimming activity of animals thus treated.

In a further series of experiments a 3-h cold pulse (−2°C) was applied to either the mid-body region (4th and 5th segments of the pereon) or to the last two segments of the pleosome.

In both these series of experiments, rephasing of the rhythm was only observed in completely chilled animals. Statistical analysis of the phase delays relative to high water seen in untreated animals and in animals chilled in the mid-body or pleosome regions revealed no significant difference between these groups (P< 0·05, Anovar).

Selective cold pulses applied to the supra-oesophageal ganglion include the eyes their radius of effective chilling, and in view of the pacemaking properties demonstrated in the eyes of various gastropod molluscs (Jacklet, 1969; Lickey et al. 1977; Block & Davenport, 1982), a further series of experiments was carried out to investigate the effect of eye ablation on the activity rhythm of Corophium. Eye ablation and local cautery were performed with a hot wire (see Materials and Methods) and the animals allowed to recover before recording cinematographically as before.

The activity of 10 eyeless Corophium is shown in Fig. 9, along with records obtained from Corophium following the ablation of only one eye, and with animals in which the cuticle was cauterized just behind both eyes, the eyes themselves being left intact. For comparison, an activity record of an untreated group of Corophium is also included.

All three treatments resulted in a decrease in the overall level of activity relative to the untreated group. Rhythmic, ebb-tide swimming activity is evident in all four groups however, with little difference in phase or period. On subjecting the data to periodogram analysis using the method of Harris & Morgan (1983), significant rhythmicity was evident in all four groups (P < 0·01 in each case). These observations were confirmed in a second series of identical experiments.

The results described here suggest that, as with circadian rhythms in other arthropods (Page, 1981a), the circa-tidal rhythm of swimming activity shown by Corophium is under multiple clock control. Bilateral oscillators located in the eyestalks are implicated in the regulation of the circa-tidal rhythm of Carcinus (Naylor & Williams, 1968; Naylor, Smith & Williams, 1973; Williams et al. 1979), but in Corophium the eyes do not seem to be involved in the mechanism of rhythm control. This is perhaps to be expected for a sessile-eyed crustacean.

However, the rhythm can be rephased by cold pulses applied locally to selected parts of the body. Animals subjected to cold pulses applied in the regions of the supra- and sub-oesophageal ganglion alone are rephased to the same extent as those chilled in their entirety, whereas chilling elsewhere on the body is without effect.

Similar techniques have been used in clock location by earlier workers (Brown & Harker, 1960; Naylor & Williams, 1968) but if the results of such experiments are to be interpreted with confidence some estimate of the thermal conductivity of the body tissues must be made (cf. Page, 1981a). In the present study, measurement of the temperature gradient along the body indicated that effective low temperatures were confined to within 0·5 mm of the cold source. The cephalic ganglia were at least this far apart in even the smallest animals used in the present study and it is unlikely that chilling one ganglion to −3 °C would cool the other sufficiently to rephase the endogenous swimming rhythm. This is confirmed by the absence of rephasing on chilling the whole body to 0°C, the minimum temperature the tissues would be expected to reach at a distance from the cooling wire equal to that separating the cephalic ganglia, and is supported to a lesser extent by the results of selectively chilling points on the head approximately 1·5 mm away from the cephalic ganglia.

It seems probable, therefore, that the rhythm of swimming activity shown by Corophium is controlled by two anatomically discrete oscillators, one in each of the cephalic ganglia.

The precise interrelationship of the two pacemakers is unknown but, from the data presented here, it would appear that they are functionally equivalent, with that in the chilled ganglion acquiring a temporary dominance. Page’s (1981b) selective chilling experiments on the cockroach optic lobe also indicate a dominance of the chilled lobe and the results are consistent with the suggestion that each oscillatory centre consists of a number of oscillators (Winfree, 1975) all slightly out of phase. Chilling presumably realigns these clocks and thus imparts dominance to that centre by virtue of its greater internal synchrony.

Further evidence in support of a coupled oscillator mechanism is the bimodal pattern sometimes seen in the ebb-tide swimming of freshly collected untreated animals (Holmstrom & Morgan, 1983b; Harris, 1983). This bimodality is most clearly seen when observing groups of animals, but the results of other experiments (Harris, 1983) suggest that the activity of individuals may be similarly organized, such bimodality may therefore be comparable to the phase splitting reported for the circadian rhythms of many organisms (Pittendrigh, 1974; Roberts, 1960, 1962;Hoff man, 1971; Pohl, 1972; Gwinner, 1974; Daan & Aschoff, 1975; Eriksson, 1978; Koehler & Fleissner, 1978; Wiedenmann, 1980) and which is generally, although not universally, attributed to a two oscillator system. [For example Weidenmann (1980) has shown that the phase of the two peaks in the activity rhythm of the cockroach Leucophaea, is unaltered by light pulses and interpreted this as evidence of a single oscillator.] The reversion of the activity pattern to a unimodal wave form following whole body chilling of otherwise bimodally active animals (see Fig. 6, also Holmstrom & Morgan, 1983b) is thus consistent with the concept of entrainment of a two oscillator system. In this context it is interesting to note that selective chilling of either of the two main oscillators invariably results in the appearance of a bimodal activity peak approximately 12 h after rewarming, irrespective of the form of the original rhythm. Subsequent peaks revert to a unimodal form which remains phase delayed, suggesting that the coupling has been subtly altered by chilling.

The concept of a multiplicity of pacemakers is now generally accepted in chrono-biological literature and both hierarchial and equipotential systems have been postulated (e.g. Moore-Ede & Sulzman, 1981). Both types of organization appear to be involved in the endogenous control of ebb-tide swimming in Corophium. Equal pulses of low temperature do not induce an equal phase shift at all stages of the tide. Pulses applied during the flood tide cause a phase delay, while those applied during the ebb advance the rhythm, and Holmstrom & Morgan (1983b) suggested that this change in responsiveness could be accounted for in terms of the two oscillator hierarchial model postulated by Pittendrigh & Bruce (1959). The results of the experiments described here, however, indicate that although selective chilling can change their relative importance, the pacemaking centres of the two cephalic ganglia are normally of equal dominance. Thus both oscillators presumably correspond to the more fundamental type in the hierarchical-multioscillator model, a development of the Pittendrigh & Bruce (1959) model. This being so, the driven element(s) has still to be located.