ABSTRACT: Extracellular recordings of Mauthner (M-) cell activity was carried out in the Senorita wrasse, Ovyjulis californica Criteria for identifying the Mauthner cell have been satisfied in our experimentation. We observed a C-start tail flip behavior caused by both antidromic and orthodromic stimulation. Electrophysiological recordings from the medullary area thought to contain the M-cells reveal a short-latency, all-or-nothing spike of large amplitude. Preliminary experiments on the effects of an as yet uncharacterized cone snail toxin on M-cell activity were performed. INTRODUCTION: A dramatic and universal behavior pattern seen in teleost fish is the characteristic "tail flip" escape response. This two phase startle response consists of a fast body bend, where the fish assumes a"C"-like shape, followed by a straightening of the tail which can propel the fish a full body length (Eaton, 1976.) A single pair of cells in the medulla is responsible for the coordination of spinal motor neurons to produce this behavior. These cells, called the Mauthner (M-) cells, are the most conspicuous neurons in the brain. In goldfish, the two main dendrites can measure up to half a millimeter in length and the large myelinated axon extends down the length of the spinal cord (Furshpan and Furukawa, 1962.) The M-soma receives input from afferent neurons in the ear and lateral line. Thus a sudden vibrational stimulus can activate one of the M-cells and cause the stereotypic body contraction, resulting in rapid acceleration through the water. Most behavioral research conducted on Mauthner cells has been directed at a single behavior, the startle response. Although, the stereotypic escape behavior has appealed to researchers because of its simplicity and predictiblity, these responses are generally not all-or-none. Closer examination of the startle response in fish reveal that the rapid escape movements can be graded in amplitude (Aljure et al, 1980.) Thus it seems that the execution of escape movements require some finer control. It is conceivable that Mauthner cells may be involved in other fast movements with a directed orientation, such as predatory and feeding behavior. In hopes of understanding the M-cell in conjunction with activities other than the tail flip, we have chosen the intertidal wrasse as the subject. The Senorita, ovyjulis californica displays two separate "escape" reactions when it senses danger, an extraordinary burrowing behavior as well a more typical tail-flip response. It appears to use the same set of muscles for both escape responses. Presumably, the tail-flip response is, as in other teleosts, controlled by the M-cells. How then are the Mauthner cells involved in escape burrowing and in the process of deciding between the two behaviors? The following paper seeks to explore the above questions by first developing a workable method to record electrophysiologic behavior from the Senorita. We show by satisfying four criteria that we have indeed recorded from the Mauthner cell. First, we demonstrated that antidromic and orthodromic stimulations produce a c-start tail flip. Secondly, we have measured large, single-unit spikes in an area of the medulla corresponding to that where M-cells are located in goldfish. Third, the very short latency of these responses are similar to those known to accompany antidromic conduction along the large M-cell axons. Finally, we see that these short latency responses are act ivated with the lowest threshold observed and bear an all-or-nothing relationship to the strength of stimulus (Furshpan and Furukawa, 1962.) Our setup was applied to preliminary analysis of the effects of cone snail toxins. The toxin used in this study is a peptide abstracted from Conus snails of the Philippines that prey on fish. This as yet uncharacterized toxin is thought to be the single component that allows snails to paralyze their prey within seconds, but its mode of action has not been defined. (Olivera, personal communication.) Given the speed of its effectiveness, it was hypothesized that this "paralytic toxin" sends a train of rapidly firing action potentials to the M-cells. The M-cells could then stimulate the contraction of tail muscles on both sides thereby paralyzing the prey in a seizure of uncoordinated activity. We show that our setup was adequate in measuring M-cell activity as a response to toxin injection, though further study is needed to clarify the effects. HETHODS Preparation of fish and operation procedure Experiments were performed on adult Senorita (15.5-18)cm in length, Fish were kept in fresh sea water at 14 C, and fed chopped squid. Fish were anesthetized by a 5-10 minute imersion in sea water containing 0.0052 quinaldine and .028 isopropyl alcohol. Anesthetized fish were first wrapped in gauze to reduce slipping from mucous and were then secured by means of adjustable, plastic cable ties to a plexiglass platform fitted to a small tank. The mounted fish was placed in the experimental tank and, the head of the fish was immobilized by fixing the mouth over a disposable pipette tip. Cooled (14 C) and oxygenated bath solution was perfused at rate of 10Oml/min through the pipette tip during the entire experiment. The body was kept completely submerged at all times in a solution of 50% seawater and 50% deionized water which served as a simple and satisfactory saline solution. The water pump circulating the bath solution was powered by a 12V DC battery to avoid 60 cycle noise (Fig. 1). Äfter lowering the water level to expose the skull, a small amount of local anesthetic (1.52 Dibucane) was applied to the area with cotton swab before incision. A 1.Ox0.5 cm section of skin was removed from a central area approximately 1 cm posterior to line drawn between the fish's pupils (Fig .2). Muscle and connective tissues were scraped away to reveal the cranial ridge. Using a scalpel, a small slot was made through the skull and forceps were Used to peel away the bone. A Amm diameter hole was finally made through which an electrode could be inserted. Electrophysiological Observations Conventional methods were used for electrophysiological recordings and for electrode micromanipulation. M-cell activity was recorded with tungsten microelectrodes insulated with 0-dope varnish except at the tip. The recording electrode was first positioned over the center of the cerebellum slightly off the midline (Eaton, 1981). The trajectory of the electrode was at à 90 angle to the dorsal surface of the cerebellum. Data was recorded on a digital oscilloscope and photographed with Polaroid film. Since it was known that the M-cells in goldfish are situated approximately 1.5mm below the medulla oblongata (Zottoli,1978), and we measured the Senorita cerebellar thickness to be 2.6mm, we predicted that Senorita M-cells should be approximately 4.Omm below the surface of the cerebellum. Thus in most experiments for recording M-cell activity, the M-cell was found by lowering the electrode to a depth of 3-Amm below the surface of the brain while monitoring responses to antidromic stimulation. A large negative going extracellular action potential is generated at the M-axon in response to antidromic stimulation of the spinal cord (Faber and Korn, 1978; Diamond 1971.) (See figure 2 for area of insertion.) Care was taken not to lower the electrode beyond the region of M-spike focus. Survival of the animal appeared to be greatly impaired after electrode penetration into these deeper regions of the medulla. Antidromic stimuli were delivered to the spinal cord by means of stainless steel bipolar electrodes made from 25 gauge syringe needles insulated with nailpolish except at the tip. The negative electrode was inserted through the tail musclature down to the spinal cord at a location 6-8cm behind the cranium (Diamond 1971). The positive electrode was positioned just outside the skin. To ascertain the rate of repetitive stimulation useful for later experiments, we tested the limits of habituation on a fish that had not undergone surgery. By delivering a stimulus of 2 volts above threshhold value at regular intervals, it was determined that the animal quickly habituated to stimulus rates greater than 1shock per 5 seconds. Application of shocks was therefore kept at a frequency less than O.1Hz. RESULTS Orthodromic Stimulation Orthodromic depolarizations via a tungsten microelectrode in the medulla produced the characteristic all-or-nothing tail flip at a sharp threshold. Closing of both opercula and a slight bobbing of the head accompanied the tail flip. Reversing the stimulus polarity elicited an obviously different behavior best described as a twitch of the entire body of undef ined direction. At an electrode tip depth of 3mm below the surface of the cerebellum, we varied the stimulus duration and strength for depolarizing pulses to defined strength-duration relation for threshold activation of what appeared to be the all-or-none tail flip (Fig. 3). As expected, stronger stimull were necessary to elicit the behavior as the duration grew shorter. Stimulus strength lower than 2 volts did not produce a response even when the stimulus was very long. In general, threshold values were distinct within one experiment at any time but usually increased as the fish expired. Antidromic Stimulation Stimulation at the spinal cord with a bipolar electrode produced contraction of tail muscles ipsilateral to the stimulus. In addition, the stimulus also caused simultaneous closing of both opercula. Both behaviors appeared to be similar to those produced by orthodromic stimulation. Stimulus strength necessary to produce antidromic responses varied from as low as 1.5V at a duration of O.06ms to as much as 40 volts depending on the condition of the fish and the location of the electrode. Reversing the stimulus polarity (i.e. making the internal electrode positive) gave rise to a tail bob in the vertical direction and, at a higher voltages, caused the tail to curve in the opposite direction. Depending on the extent that experimentation worsened the health of the animal, the tail flip varied from strong, sweeping motions to only slight flicks. Although the behavior were not as well defined as those following orthodromic stimulation, both probably involve M-cells as supported by results below. Electrophysiological recordings made in a region dorsal to the M-cells typically reveal two fast negative spikes and one broader wave, 1.5ms in duration (Fig 44). As measured from the start of the shock artifact to the beginning of the initial positivity, the latencies were very short, approximately 1.2, 1.6 and 3.Oms (Fig. 4B). The two early components begin to appear at a depth of 1.5-2.Omm below the surface of the brain and their amplitudes increased with depth. From the last trace in figure 4C, we see that the one of the early components focuses to a large spike of 0.6V at a depth of 4.8 where the electrode is now in the medulla. The late potential change occuring at 3.Oms can be measured from the within the cranial cavity above the brain and increases as the electrode enters the cerebellum. However, as the electrode is lowered to the medulla, this late component disappears. In sum, the two early spikes seem to originate from the medulla while the slowest spike appears to originate from the cerebellum. The two early spikes and later single component vary differently as a function of voltage. Two separate experiments, figure 5A and 5B, confirm that both sets of spikes appear to have the same threshhold value (4.0 and 4.2V respectively). Once breaching that value, it is clear that the amplitude of the two early spikes do not increase with stimulus strength while the late one does. Thus it can be said that the two early spikes bear an all-or-nothing relationship with strength of the stimulus, and the single later component can be graded with voltage. To assess the level of background "cranial noise," the recording electrode was inserted in a location believed to be distant from the Mauthner location. In this experiment, the M-spike was first located by methods described above. Then the electrode was placed 1.3mm anterior and approximately 0.75mm to the right of the location of the M-spike, and lowered to a depth of 3.Smm. Medullary recordings in this region show no spikes corresponding to any of the three responses measured in the M-cell region nor any brief latency responses. On a slower time base, it is apparent that electrodes in the non-M-cell region recorded spontaneous and stimulated brain activity not seen in the M-cell regions (Figs 6B1,B2,B3.) In addition to spinal cord electrical stimulation, a strobe light flash was used to trigger a startle response and action potentials in the M-cell were measured from the same two regions. The trace from the M-cell region (fig 7A) shows only one spike at a latency of 1.Oms. The same stimulus failed to produced any short latency mauthner-like spikes in the non M-cell region but elicited a broad peak at 33ms after the shock artifact (fig 7B). Cnus Pentides and M-cell activity Preliminary studies were also performed on M-cell involvement with paralysis of Senorita with a toxin from cone snails. The peptide was injected within Icm radius of stimulating electrode insertion while delivering superthreshhold shock to tail at a rate of 6 shocks/minute. Amplitudes of both the M-cell spikes and cerebellar components decreased until disappearing. The trunk of the fish became very stiff as if muscles both sides were completely contracted. Its caudal and dorsal fins were fanned out like that of a the tail in mid-flip response. It was confirmed that the fish was partially paralyzed when the released fish could not move its tail but could flap its pectoral fins. In another experiment where less than the required dosage to paralyzed was injected, the fish was seen to twitch convulsively in frregular sessions while electrophysiological data is presently being analysed. DISCUSSION Observations of behavior induced by anti- and orthodromic stimulations, and electrophysiological data have shown we located and recorded from the Mauthner cell. Data from cranial recordings support the hypothesis of direct influence by M-cells on neurons in the cerebellum. In the best preparations of Senorita which were studied, behaviors produced by anti- and orthodromic stimulations of the medulla are in agreement with those expected to be produced by excitation of the Mauthner cell, based on published work. The trunk, gill and eye movements observed in our studies are all characteristics of M-initated responses in goldfish (Diamond, 1971.) The optically activated Mauthner activity produced the same behavior as that evoked by electrical stimulus. Detwiler (1927) has also demonstrated that visually-evoked startle responses are mediated by the M-cell in Amblystoma. This fact was sufficient to show that a "naturally" occuring stimulus would produce not only the same mechanical behavior but also the same electrical discharge characteristic of what we identify as M-cell related. From observations of antidromic stimulated behavior, it appeared that the tail flick response was sometimes graded with stimulus strength. The weaker tail flicks did not correspond in form to that of the naturally occuring startle response though they were accompanied by an M-spike. Eaton and Bombardieri (1978) recognized that the Mauthner response is not simply an explosive and random reflex contraction of the animal's body musclature. Rather, it is a highly organized motor pattern involving the M-cell and probably many other neurons as well. Studies conducted by Aljure (1980) indicate that the gradations result from inhibition onto motorneurons. Clearly, the ambiquity and speed of behavior renders unaided visual observation inadequate for adequately document ing M-cell related behavior. The above behaviors were observed to have a well-def ined threshold consistent with that of M-cell stimulated behavior in other studies. The strength-duration curve for tail-flick threshold obtained from orthodromic stimulation (fig3) indicates that the behavior observed is probably due to stimulation of a small number of neurons. The sharpness of the curve would be consistent with selective stimulation of the large M-cell in the medulla and supports the proposal that we are in fact stimulating M-cells to produce the tail flip. It has been shown that recordings from the brain concurring with Mauthner tail flip arise from two different depths. We believe the two early components measured in the medulla are the result of the asymmetrical activation of the two M-cells. This is confirmed by their all-or-nothing character and very short latencies of 1.2 and 1.5ms. The mean spike latency measured in a similar manner in goldfish is 0.78ms (Eaton 1981.) The apparent conduction velocity in our experiments was calculated by dividing the negative spike latency into the distance from the stimulation cathode to recording electrode. The result, a conduction velocity of 80m/s, agrees well with published values for the goldfish M-, ranging from 55 (Dorsett, 1980)to 10Oms (Eaton, 1981; Furshpan 1962.) Although the diameter of the M-axon in Senorita is not yet known, it would appear to be very close to that in goldfish. Further data which support that these are M-cell initiated spikes comes from inspection of the relative amplitudes of the two spikes. In Figure 48, the first of the two early spikes is larger, wheras in Fig 4C the second spike is very much larger. Since the M-cell axons are known to cross in the medulla (Diamond, 1971), one would expect stimulation on the one side of the fish to excite the contralateral M-cell first. In this case of Fig. 4B the recording electrode in the medulla was positioned to the left of the midline, and stimulation was on the right-hand side of the fish. Thus, we expect the first early spike to be the larger of the two, because the electrode is closer to the left M-cell. Exactly the opposite result is seen( l.e., the second early spike is larger) in Fig. 4C, when the recording and stimulating electrodes were plac on the same side of the fish. Since the slower component can be measured from outside the brain and essentially disappears once the electrode enters the medulla, we believe that neurons excited in the cerebellum caused this signal. The slow, i11-defined time course of the signal suggests the massive recruitment of many cerebellar neurons firing at slightly different times. In addition, we did not find any changes in the tail-flip behavior when the cerebellum was completely removed, but the slow component was no longer recordable. This lends further evidence to our hypothesis that the third component does originate from the cerebellum and does not play a necessary role in the M-cell activated tail flip. The facts that 1) both M-spikes and the cerebellar discharge appear to have the same threshold value and 2) there exists a consistant delay between the two early and third components, may indicate that M-cell activity can quite directly cause or influence the firing of the cerebellar neurons. We observed the appearance of the first two spikes without the cerebellar components only once in our entire study. This recording was made precisely at threshold in the first two spikes at a depth where both early and late components were measured in all other instances. The M-cell coupling to cerebellum discharge would thus appear to be powerful, but still susceptible to inhibition from other sources. The exact relationship between the cerebellum and the M-cells is still unclear. The most recent research, conducted by Bartelmez in 1915, suggested the cerebellum as a possible source of afferent input to M-cells. Experiments relating the maintainance equilibrium and M-cells were conducted in Ambystoma by Detwiler (1947) who showed that animals missing M-cells could hold normal posture at rest and during swimming. Thus M-cellsare not necessary for functions of the cerebellum pertaining to these behaviors. The Senorita proved to be an excellent subject for study of Mauthner initiated behaviors. Senorita are surprisingly hearty animals; most fish were able to live 1- 3 hours after the operation. Their brains, large for their body mass, are easily accessible through their thin skull. Electrophysiological recordings from freely mobile Senorita using techniques pioneered by Zottoli (1977) might help to elucidate the role of M-cells in the decision-making process such as that between the two escape responses, the C-start and the burrowing. Acknowlegement: would like to sincerely thank Dr. William Gilly for all his guidance and pat ience throughout the project. BIBLIOGRAPHY Aljure E, Day JW, Bennett MVL(1980) Postsynaptic depression fo Mauthner cell-mediated startle reflex, a possible contributor to habituation, JPhysiol, 45:533-564. Detwiler SR (1947) Quantitative Studies on the Locomotor Capacity of Larval Amblystoma (A. jetfersoniamun) Lacking Mauthner's Neuron or the Ear, J Exp Zool, 104.343-351. Diamond J(1971) The Mauthner cell. In: Hoar WS, Randall Do(eds) Fish Physiology vol 5. Academic Press, New York, pp 265-346. Dorsett DA(1980) Design and Function of Giant Fiber Systems, Trends in Neurosciences, 3:205-208. Eaton RC, Bombardieri RA, and Meyer DL (1976) The Mauthner¬ Initiated Startle Response in Teleost Fish, J Exp. Biol. 66:65¬ 81. Eaton RC, Hackett JT (1984) The Role of the Mauthner Cell in Fast Starts Involving Escape in Teleost Fishes. In: Neura/ Mechanisms of Startle Behavior, Eaton RC (ed) Neural Mechanisms of Startle Behavior, Plenum, New York,pp213-260. Eaton RC, Lavender WA, and Wieland CM(198 1)Identif ication of Mauthner-Initiated Response Patterns in Goldf ish: Evidence from Simultaneous Cinematography and Electrophysiology, Comp. Physiol. 144.521-531. Faber DS, Korn H (1978) Electrophysiology of the Mauthner cell: Basic Properties, Synaptic Mechanisms, and Associated Network. In: Faber DS, Korn H(eds) Neurobiology of the Mauthner cell Raven Press, New York, pp 47-131. Furshpan EJ, Furukawa T(1962) Intracellular and Extracellular Responses of the Several Regions of the Mauthner cell of the Goldfish, JNeurophysiology 25:732-771. Zottoli SJ (1978) Comparative Morphology of the Mauthner Cell in Fish and Amphibians. In: Faber DS, Korn H (eds) Neurobiology of the Mauthner Cell Raven Press, New York, pp 13-45. FIGURE LEGENDS Fig 1. Schematic of Experimental Setup. Fig 2. A. Drawing of Senorita brain and electrode insertion at place marked X. B. Drawing of location of M-cells as seen from above from goldfish. Cerebellum retracted to expose medulla (Furshpan and Furukawa, 1962) Fig 3. Strength-Duration Relationship of Medullary Shock to Produce Escape Response. Fig 4. Three components of brain activity measured at varying depths. Fig 5. Medullary components as a function of voltage. Fig 6. Comparisons of brain activity in two regions of the medulla. Fig 7. Response to strobe light stimulus measured in two regions of the medulla. — P d — V — — + Volts aa oC U g Z 0 U 0 Z L