ABSTRACT
The axon of the larger of the two motoneurones in the optic tract of Carcinus which bring about the rapid eye withdrawal is 30–50 µ in diameter and has an unusually thin sheath compared with crab leg neurones of similar diameter.
Within the brain the axon of the large eye-withdrawal motoneurone tapers fairly rapidly to 5–10µ. It extends diagonally across one side of the brain and branches in two near the mid line but does not pass the contralateral side of the brain. The axon terminals and the cell body have not been seen.
A single stimulus to the afferent inputs of the motoneurone produces a characteristic burst of impulses followed by an irregular train of impulses.
The excitatory effects of all tested afferent inputs to the motoneurone are summed before affecting a common spike-initiating locus on the motoneurone.
Intracellular recordings from the motoneurone within the brain show two types of subthreshold activity to follow a single pre-synaptic volley : a smooth graded depolarization and small superimposed depolarizations of constant amplitude and duration.
The synaptic link between afferents and the eye-withdrawal motor neurone is thought to be via a direct, possibly monosynaptic pathway and also by way of interneurone collaterals.
INTRODUCTION
The motor nerves causing the fast withdrawal of the crab’ s eye fire characteristic bursts of impulses following mechanical stimulation of the carapace. This reflex nerve discharge is not affected by the position of, or even presence of, the eye itself and can be recorded from the motor nerves of an isolated brain following electrical stimulation of the appropriate afferent inputs. The preparation provides an opportunity to study the link between the afferent and efferent pathways of a simple reflex system and, since the reflex is independent of peripheral feedback, the results of such a study should be relevant to the behaviour in the whole animal.
Previous study of this reflex system has shown that the afferent neurones causing the withdrawal of one eye lie in each of the ipsilateral brain nerves (Sandeman, 1967). Only two motoneurones, one larger than the other, bring about the fast eye withdrawal (Burrows, 1968) and both lie in the optic tract. Both give a burst of spikes following afferent nerve stimulation, but as the extracellular response of the large motoneurone is about 10 times the amplitude of any other recordable activity and is easily recognizable it is the only one to have been studied so far.
A pacemaker is involved in the reflex system. This causes a burst of spikes in both motoneurones to appear usually about every 13 sec., although interburst intervals may vary from 3 to 50 sec. Electrical stimulation of the inside half of the ipsilateral oesophageal connective can inhibit the spontaneous firing of the motoneurones.
This paper describes part of the anatomy of the large motoneurone and attempts to answer the following questions about the reflex system. (1) Is there a simple summation of excitatory effect from all the afferent inputs? (2) Do the excitatory inputs converge on a common spike-initiating locus in the motoneurone? (3) Does the link between the sensory and motor neurones involve interneurones?
MATERIALS AND METHODS
Brains of Carcinus maenas were isolated by cutting the anterior portion of the carapace containing the brain, eyes and statocysts away from the rest of the animal. The muscles lying over the brain were removed and a perfusion cannula was inserted into the cerebral artery. All the brain nerves were cut peripherally.
Anatomy
The axon of the large motoneurone, often clearly visible in fresh preparations, was impaled with a low-resistance glass micropipette filled with a mixture of 2-5 M-KCI and 0-5 M potassium ferrocyanide (Kerkut & Walker, 1962) and the intracellular responses were compared with extracellular recordings to ensure penetration of the correct axon. Negative pulses of 1 sec. duration were applied to the electrode at about 40 pulses/min. and the intensity of the pulse was adjusted until the maximum current passing through the electrode was about 2 µA.., measured with a microammeter in series with the electrode. After 15–20 min. all electrodes and the perfusion cannula were removed from the preparation and the brain was allowed to dry out for 20-30 min. This enhances the spread of the potassium ferrocyanide in the axon. The brain was then de-sheathed and immersed in Carnoy’ s fixative (Humason, 1962) mixed 10:1 with a saturated solution of ferric chloride. The fixative penetrates the tissue very rapidly, allowing the ferric chloride to mix with the potassium ferrocyanide and precipitate the Prussian blue dye (potassium ferric ferrocyanide). Tissue lying over the stained axon was lifted away with a fine needle, and the brain was transferred directly to absolute alcohol, cleared in methyl benzoate and benzene and mounted in xylene Damar.
Physiology
Glass micropipettes, 20–40 MΩ, filled with 3 M-KCI were used to penetrate the motoneurone in de-sheathed brains and optic tracts. Intracellular spikes were recorded on an oscilloscope via a capacity-compensating unity-gain amplifier (Bak, 1958).
The sensory nerves and the oesophageal connectives were stimulated through bipolar suction electrodes with pulses from a Tektronix 161 pulse generator and radiofrequency isolating probes. Extracellular recordings from the motoneurone were made with a suction electrode and a Tektronix 122 low-level pre-amplifier.
RESULTS
Anatomy
The crab’ s brain with its nerves is shown in Text-fig. 1. A cross-section of the optic tract shows an axon which is significantly larger than the rest (Plate 1), and intracellular recordings from this axon identified it as the large eye-withdrawal motoneurone. The axons of large motoneurones in Crustacea often have thick sheaths around them and are difficult to impale with micro-electrodes, but the axon of the eyewithdrawal motoneurone can be easily penetrated with fairly large micropipettes and the sheath surrounding it is unusually thin compared with that around the motor axons in the leg nerves of the same animal (Plate 2b, c). The significance of the different sheath thicknesses around motor axons in the leg nerves and in the optic tract is not clear, for the motor axons from these two bundles give essentially similar responses to direct electrical stimulation. The thick sheath of the leg nerve axons may be simply of structural importance and prevents the motor axons from being too severely deformed during leg movements. The optic tract moves very little, is never bent back upon itself and needs much less structural reinforcement.
The diameter of the axon of the eye-withdrawal neurone in the optic tract is between 30 and 50 µ but tapers on entering the brain. In the brain it crosses diagonally to a point near the mid line where it bifurcates. A superficial branch extends laterally and dorsally over the tract of axons of the oesophageal connective, and a deeper branch passes ventro-laterally (Plate 2a). The cell soma and axon terminals have not been seen and it is not yet known from the anatomy if the motoneurone can be included in the general arthropod pattern of having a cell body uninvolved in the transmission of signals. The study does, however, confirm the course taken by the axon through the brain previously plotted by electrophysiological means. In addition, no axon branch to the contralateral side of the brain is revealed, which agrees with the physiological finding that no excitatory effect follows normal electrical stimulation of the contralateral brain nerves.
The narrowing of the axon within the brain is interesting. The relatively large diameter of the peripheral part of the axon is presumably important for fast conduction in common with most protective reflex systems, but this advantage would be offset in the crab by the narrow slow conducting central portion, unless the spike-initiating locus was close to the broadened region. A narrow neck at the beginning of the axon is quite a common feature in vertebrate neurones (Cajal, 1933; Chu, 1954; Matthews, Phillips & Rushworth, 1958), and corresponds in the Mauthner cell to a region of low membrane resistance which passes a high current during spike activity (Furshpan & Furukawa, 1962). The withdrawal neurone in the crab, however, has a very thin sheath and there is no evidence yet of saltatory conduction with nodal regions of high current density during spike activity. Also spike activity has been recorded from as far along the axon as the bifurcation, which is well within the thin portion.
Physiology
The usual response of the large eye-withdrawal motoneurone following a single shock to the afferent nerves was a short high-frequency burst of impulses followed by an irregular train of impulses lasting for about 100 msec. The latency and number of spikes in the short burst depended upon the intensity of the stimulus and also upon the temporal relationship between the stimulus and the spontaneous bursts of the motoneurone. A stimulus which closely preceded a spontaneous burst sometimes produced a spike train of 80-100 impulses instead of the normal 1–10 impulses. The tegumentary nerve often produced a longer initial burst of spikes than the oculomotor nerve but latencies were similar, ranging from 4.5 to 7 msec.
Repetitive stimulation of either tegumentary or oculomotor nerves initially increased The number of long bursts but these rapidly adapted, the number of motor axon spikes dropping to one or two. Single spikes consistently followed with constant latency a stimulus frequency of between 5 and 10 impulses/sec. (Text-fig. 3/) but quickly dropped out at frequencies in excess of 50 or 60 impulses/sec. Failure of one afferent pathway due to repetitive stimulation did not result in failure of the others and spontaneous bursts still occurred in preparations in which both afferent pathways had failed.
(1) Summation of the excitatory inputs
Dividing the tegumentary nerve into thinner bundles and stimulating each of these bundles separately resulted in fewer spikes following maximal stimulation of the nerve bundle (Text-fig. 2). Stimulation of very thin bundles of tegumentary axons did not evoke motoneurone spikes but simultaneous stimulation of two such bundles would do so. Intracellular stimulation of single axons in the tegumentary nerve never produced a response in the motoneurone but there is no evidence that the axons tested in this way were connected to the motoneurone. Also behavioural studies have shown that the eye withdrawal is not produced by large hair-receptor organs on the carapace but by sensilla responsive to mechanical deformation of the carapace although both types of receptor have their axons in the tegumentary nerve (A. P. Scott, personal communication).
A subthreshold stimulus to one of the afferent nerve bundles would produce spikes in the motoneurone if it was preceded by a suprathreshold shock to the other (Textfig. 3). The interval between the two stimuli could be as long as 100 msec. Gradually decreasing the interval between two suprathreshold sequential stimuli initially shortened the latency and increased the number of spikes following the second stimulus. Eventually, when the two stimuli were close together, the two separate bursts blended into one even train. Pulses applied simultaneously to two different afferent inputs often gave a shorter but higher-frequency burst than two stimuli separated by 4 or 5 msec. (Text-fig. 3).
(2) Common spike-initiation point
Stimulation of the afferent nerves during a spontaneous burst produced a similar smooth summation of excitatory effect. The frequency of the spike train was raised but not accompanied by any break in the even spike pattern (Text-fig. 4). The absence of a break in the spike train is indicative of a common spike-initiation point on the motoneurone shared by the three inputs, but two other possibilities are (1) separate spike-initiation points one behind the other, and (2) separate spike-initiation points on different branches of the main axon trunk. The spike occlusion which would be expected to occur if either these two situations prevails was demonstrated by intracellularly stimulating the motoneurone near the periphery in the following way:
(1) Two sequential spike-initiating sites on the motoneurone could be reproducibly demonstrated if the axon was driven repetitive with a depolarizing pulse of 50 msec, duration. Extracellular records from the periphery show the spontaneously generated burst to be interrupted by a high-frequency burst followed by a break in the spike train (Text-fig. 4 c).
(2) To simulate different spike sites on different branches of the axon and the consequent mixing of two separately generated spike trains, motor axon spikes were initiated with single short depolarizing pulses through the intracellular electrode. The spontaneously generated train was interrupted by the additional spikes and an irregular pattern appeared at the periphery (Text-fig. 4 f). Neither of the above two irregularities follows afferent nerve stimulation combined with the spontaneous discharge.
(3) The sensory/motor link
The results of recording from the peripheral end of the axon suggest: (1) a long-lasting post-synaptic excitatory state within the motoneurone following a single afferent volley, (2) the excitatory effects of the afferent inputs and the spontaneous burst generator are summed before affecting a common spike-initiating locus on the motoneurone.
Evidence obtained from intracellular recordings from the motoneurone within the brain (A in Plate 2a) supports the above conclusions. Stimulation of the tegumentary nerve resulted in the depolarization of the proximal regions of the motoneurone and the size of the depolarization depended upon the stimulus intensity; increasing the stimulus intensity eventually yielded a spike which was conducted along the axon and was recorded at the periphery. The excitatory depolarization was long-lasting and had small humps of short duration and relatively constant amplitude superimposed on its falling phase. Increasing the intensity of the stimulus shortened the latency of the humps but did not alter their amplitude or duration (Text-fig. 5). These small unitary fluctuations in the membrane potential occurred irregularly for up to 100 msec, after a single pre-synaptic stimulus and, in some cases when the excitatory depolarization was large enough, erupted into conducted spikes. The separate excitatory depolarizations in the motoneurone caused by repetitive stimulation of one afferent nerve, or by stimulation of both afferents together, combined to initiate a motoneurone discharge.
Occasionally intracellular electrodes recorded smaller spike-like potentials of constant amplitude and short duration, interspersed between the full-sized action potentials of a spontaneous burst (Text-fig. 6). These were recorded when the motoneurone was impaled distally within the brain (B in Plate 2 a), and could represent all-or-none activity since the amplitude of the potentials decreased at high frequencies, a characteristic of action potentials. The small potentials were, however, never conducted as spikes to the periphery.
DISCUSSION
The response of the motoneurone following a single pre-synaptic stimulus is characterized by an initial short spike burst followed by an irregular tail of spikes. This could indicate two parallel pathways joining the sensory nerves with the motoneurone; a direct, possibly monosynaptic link, and an indirect link involving an intemeurone chain.
A direct monosynaptic link between all the sensory fibres and an extensive motoneurone arborization would explain the temporal spread of initially synchronous pre-synaptic action potentials and the production of a smooth 10–20 msec, long post-synaptic depolarization in the motoneurone. Spike latencies of 5–7 msec, may seem too long for a single synaptic delay but intracellular recordings show the excitatory potential in the motoneurone to begin about 2 msec, after a subthreshold pre-synaptic stimulus. If the electrode was relatively distant from the fine synaptic terminals, true synaptic delays are shorter than 2 msec., which would not exclude a monosynaptic link. The slow rise time of the excitatory post-synaptic depolarization, and thus comparatively long spike latencies, argue for a diffuse synaptic field some distance from the spike-initiation point. The system would act as a high-frequency filter and prevent regular one-to-one following of the post-synaptic neurone to a high rate of pre-synaptic firing, in spite of the ability of the post-synaptic axon to fire at 400 impulses/sec. to a maintained depolarization.
An indirect link between sensory and motor axons through a number of interneurone collaterals would explain the prolonged excitatory state of the motor axon following a single pre-synaptic volley. The activity of one of these interneurones appears as small fluctuations on the falling phase of the post-synaptic excitatory potential. The constant amplitude and duration of these fluctuations suggests that each one is a post-synaptic potential generated by a single pre-synaptic spike in the interneurone which has a relatively localized contact with the motoneurone. The eruption of these potentials into spikes making up the irregular tail of the motoneurone burst confirms their excitatory function and it is unlikely that they represent the activity of the primary afferents, because repetitive activity of the afferents would not normally follow a single 0·1 msec, shock.
The final output of the motoneurone is the result of a simple summation of activity within the axons of any one afferent nerve bundle. Where two afferent nerves are stimulated together during a spontaneous burst a general summation of excitatory effect occurs prior to the initiation of action potentials within the motoneurone at a single locus. The presence of all-or-none spikes in fine terminals of the motoneurone cannot be excluded as no recordings were made from these areas, but in view of the very smooth integration of excitatory effects from different inputs, terminal spikes if present, must be prevented from propagating to the main axon spike initiating locus.
Action potentials of two different amplitudes are occasionally recorded by an electrode which is almost certainly within one nerve cell. This phenomenon has been previously recorded in crayfish central nervous systems and shown to be the result of non-conducted action potentials in a side branch of the main axon in which the electrode is lodged (Takeda & Kennedy, 1965). In the crab the small-amplitude spikes are most likely not from a side branch of the main axon because they are recorded from the distal portion of the axon where there are no branches. Instead the sub-amplitude spikes may be caused by the broadening of the axon near the point from which the recording was taken. If the axial resistance of the fibre decreases at the transition between the thick and thin portions of the axon the density of the excitatory current passing through the membrane ahead of the spike would be less. A decrease in the safety factor and temporary slowing down of the spike conduction would follow and at high frequencies the central spikes would begin to pile up behind the peripheral ones. This is seen in the records, the central spikes (which are conducted electrotonically to the electrode and appear smaller than the peripheral ones) invariably riding high on the falling phase of the peripheral spikes. Also, in contrast with the crayfish, the small spikes in the crab motoneurone are always separated one from the other by at least one full-sized spike. It is therefore suggested that the small spikes are action potentials in the proximal portion of the motoneurone which fail to propagate past the recording electrode. A similar frequency barrier caused by the change in diameter of a neurone has been reported to occur in the cuticular mechanoreceptors of crustaceans (Mellon & Kennedy, 1964; Pabst & Kennedy, 1967). It is conceivable, however, that the slowdown in conduction velocity in the crab axon is aggravated by the presence of the electrode and that normally there is not such a low limit on the frequency of the motoneurone discharge.