Noise in a Locust Mechanoreceptor Synapse 129
Total Page:16
File Type:pdf, Size:1020Kb
The Journal of Experimental Biology 204, 127–138 (2001) 127 Printed in Great Britain © The Company of Biologists Limited 2001 JEB3071 INTRINSIC NOISE AT SYNAPSES BETWEEN A WING HINGE STRETCH RECEPTOR AND FLIGHT MOTOR NEURONS IN THE LOCUST PETER J. SIMMONS* Department of Neuroscience, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, UK *e-mail: [email protected] Accepted 9 October; published on WWW 5 December 2000 Summary Variability in postsynaptic potential (PSP) amplitude motor neurons showed that the amount of transmitter due to intrinsic noise limits the reliability of released varied independently between different output communication between neurons. I measured PSP sites from the stretch receptor. Histograms of excitatory variability at synapses between a forewing stretch receptor postsynaptic potential amplitude were normal and wing depressor motor neurons in locusts, a pathway distributions that lacked separate peaks. I estimate that that is important in the control of flying. The intrinsic noise quantal amplitude is significantly less than 0.1 mV and that in the stretch receptor output synapse was measured by several hundred quanta are released for each presynaptic subtracting the background noise, originating in other spike. This accords well with a previous estimate of synaptic pathways onto the motor neuron, from the the number of discrete anatomical synapses and would variability in the amplitudes of PSPs evoked by the stretch facilitate modulation of output from the stretch receptor. receptor. Intrinsic synaptic noise caused successive PSPs to vary by 4–10 % in basalar and subalar flight motor Key words: insect, locust, Schistocerca gregaria, quantum, neurons. Recordings from pairs of these wing depressor mechanoreceptor, reliability, postsynaptic potential. Introduction Synaptic noise could place a major constraint on information systems has also supported the idea that a functional processing in nervous systems. This has been demonstrated connection consists of a number of discrete anatomical in the early stages of visual processing (Ashmore and contacts, each of which has the same probability of releasing Copenhagen, 1983; Laughlin et al., 1987), but the impact of a single quantum of neurotransmitter when a presynaptic spike synaptic noise in other neuronal pathways has not been arrives (e.g. Kuno, 1964; Gulyás et al., 1993). In general, this extensively investigated. A major reason why synaptic idea has gained acceptance as a model for the way that transmission varies from one signal to the next is that synapses in vertebrate central nervous systems operate, neurotransmitter is released in a stochastic manner from although more complex models in which, for example, the presynaptic terminals, as was first established for probability of transmitter release varies between terminals, neuromuscular junctions in vertebrates (del Castilo and Katz, may describe particular synapses more accurately (Walmsley 1954; Boyd and Martin, 1956). At the neuromuscular junction, et al., 1998). The synapses at which the simpler model applies the postsynaptic response to a spike in the motor neuron is well have relatively few discrete contacts, between normally large relative to its degree of variability. In contrast, approximately 5 and 20, and the probability of release for a in the central nervous system, most synapses operate with quantum from any one is generally approximately 0.5. much smaller postsynaptic potentials, and their amplitudes In insects, discrete postsynaptic responses to individual can vary significantly from one presynaptic spike to the next. quanta of neurotransmitter have been described for output Where detailed analysis has been possible, postsynaptic synapses made by a local spiking interneurons in the locust potential (PSP) amplitude fluctuates in a manner that is thorax (Laurent and Sivaramakrishnan, 1992) and for the described well by a binomial distribution. synapses between cercal sensory hairs and giant interneurons One synapse that has been extensively studied is a sensory in crickets (Davis and Murphy, 1993) and in cockroach input to the Mauthner neuron in goldfish. Here, the number of nymphs (Sosa and Blagburn, 1995). At these synapses, peaks anatomical contacts between the pair of connected neurons, in the distribution of PSP amplitudes indicate that individual which are boutons at this synapse, corresponds well with the quanta produce PSPs 0.2–0.3 mV in amplitude, with each number of quanta that are available for release when a PSP containing 5–20 quanta. However, in contrast to these presynaptic spike arrives (Korn et al., 1981; Korn et al., 1982). physiological measurements showing that a PSP is composed Other work on synapses in vertebrate central nervous of relatively few quanta, anatomical studies have shown that 128 P. J. SIMMONS most insect synapses consist of hundreds or even thousands local pet shop. The thoracic ganglia were exposed, and their of discrete contacts. The anatomical data come from lateral nerves were cut, with mesothoracic nerves 3A and 4 examinations of series of electron micrographs in which of being cut near to their target muscles. The thoracic ganglia pairs of synaptically connected neurons can be identified in were then removed from the animal and placed in a dish flies (Nicol and Meinertzhagen, 1982) and locusts (Simmons containing 1–1.5 ml of saline. Connective nerves between and Littlewood, 1989; Burrows et al., 1989; Killman et al., ganglia were cut to interrupt inputs from intersegmental 1999). The smallest number of contacts reported for a synapse interneurons to the mesothoracic flight motor neurons, in the insect central nervous system is 50 for an inhibitory especially the large and continuous rhythmical PSPs from a connection between a pair of second-order ocellar pair of metathoracic interneurons (Burrows, 1975b). The neurons (Littlewood and Simmons, 1992). However, mesothoracic ganglion was secured by pins inserted into a electrophysiological measurements show that quantal layer of dental wax on the bottom of the dish. In most amplitude for this synapse, which conveys graded potentials, experiments, in which recordings were to be made from the is well below 0.1 mV (Simmons, 1999). neuropile, the ganglion was mounted dorsal surface up, but for The subject of this study was a synapse between the recordings from cell bodies it was mounted ventral surface up. forewing hinge stretch receptor and flight motor neurons in the The layer of fat surrounding the ganglion was removed, and locust. The aims were to measure by how much the PSP the fibrous neurilemma was softened by application of a 1 % fluctuates from spike to spike and to estimate the amplitudes solution of Protease (Sigma, Type XIV) in saline for 3 min, of the responses to individual quanta of neurotransmitter. The after which it was washed with saline. In some experiments, stretch receptor synapse was chosen for study because it plays the ganglion was perfused at a rate of 1–2 ml min−1 with saline an important role in controlling flight movements, and there is that could be replaced with saline solution containing an extensive literature on its physiology, pharmacology and 5 mmol l−1 cobalt chloride. In these experiments, the ganglion structure. In addition, it is possible to stimulate the stretch sheath was carefully torn to aid access of the cobalt to the receptor axon selectively without stimulating other axons neuropile. because of the way that its axon branches into two separate For stimulating or recording from nerves, paired platinum nerves (Burrows, 1975a). hook electrodes (diameter 0.1 mm) were used. The stretch A stretch receptor consists of a single sensory neuron that is receptor was stimulated by 0.1 ms shocks delivered through excited when its wing is elevated (Gettrup, 1962), and it extracellular electrodes placed under prothoracic nerve 6 at a produces 15–20 spikes each wing beat with inter-spike frequency of 2 Hz unless stated otherwise. A second pair of intervals of 2 ms or less during flight by a tethered locust electrodes, placed under mesothoracic nerve 1, recorded the (Möhl, 1985). Each stretch receptor makes short-latency, spikes evoked in the stretch receptor axon. Pairs of hook excitatory connections with wing depressor motor neurons and electrodes placed under nerve 3A or 4 were used to record from longer-latency, inhibitory connections with wing elevator or to stimulate axons of basalar or subalar motor neurons to motor neurons (Burrows, 1975a) and it excites and inhibits identify motor neurons from which intracellular recordings various local interneurons (Reye and Pearson, 1987). A stretch were made (the first and second basalar motor neurons were receptor has been shown to be part of the normal pattern- not distinguished). generating mechanism for flight because stimulating it can To make the intracellular recordings, glass capillary increase the frequency of the flight rhythm (Reye and Pearson, microelectrodes, filled with 2 mol l−1 potassium acetate and 1988) and bring forward the time of spikes in wing depressor having direct current resistances of 30–60 MΩ were used. The motor neurons (Pearson et al., 1983). In ultrastructural studies, electrodes were connected to an Axoclamp 2A microelectrode direct synapses between a stretch receptor and a stained wing amplifier. Frequent checks for the stability of an intracellular depressor motor neuron have been found, and it has been recording were made by monitoring changes in the resting estimated that the functional