Abstract Pannychia moseleyi exhibits waves of bioluminescence under neural control. The mean speed of propagation of the luminescent waves is 16.8 +- 8.8 cm/s. Excitation propagates both in the radial nerves and in the nerve net, with respective conduction velocities of 31.9 cm/s and 7.1 cm/s. The bioluminescence acts as an excitable medium; however, the dispersion relation predicted by excitable media theory between conduction velocity and period is not observed. Pacemakers controlling the bioluminescence have periods between 0.33 and 0.69 s. The activity of these pacemakers may be controlled by an inward calcium current because the period of the pacemaker increases in high concentrations of magnesium. Introduction Pannychia moseleyi is a benthic holothurian that lives between 500 and 3000 meters of depth. Complex waves of bioluminescence propagate swiftly across the organism's dorsal surface. The simplest patterns within these waves were studied using excitable media theory. An excitable medium is a refractory system with spatial dimension in which adjacent sections are coupled so that a wave of excitation can propagate (Winfree, 1986). Excitable media theory successfully explains oscillations in chemical media, but extension of the theory to inhomogeneous biological systems such as bioluminescence in Pannychia may strengthen the possible applications to cardiac contractions and nervous systems. The easily-detected bioluminescence serves as a useful reflection of neural activity that controls the luminescent waves. The non-invasive studies, although limited, may give important insights into neural activity as a naturally occurring process rather than as a result of experimental manipulation. The pacemaker neurons mediating the repetitive waves of stimulation may eventually be isolated by examining the innervation of the light-producing cells; however, for now, intact animals provide initial insights into the properties of the pacemaker neuron. Materials and Methods Pannychia were collected by the remotely operated vehicle Ventana aboard the ship Point Lobos in the Monterey Bay from 550 to 900 meters in depth. The organisms were transferred to Rubbermaid containers along with the mud that was grasped in the collection process. We stored the containers in holding tanks maintained at 5 C for the duration of the experiments. At least once a week we replaced the sea water in the containers. We fed the Pannychia nothing and exposed them only to red-filtered light. Approximately once every three days, we filmed each animal. We transferred the Pannychia to butter dishes full of sea water at least 1/2 hour before filming. We used a Cohu Silicon Intensified Tube (SIT) camera to capture both images illuminated obliquely with an unfiltered flashlight and images of luminescence with no background lighting. We stimulated the animals mechanically with a glass rod. We transferred images from videotape to optical disk using an OMDR. We then analyzed the images using Megavision. Images included in this paper were captured by scanning single frames from optical disk or overlaying and averaging up to 20 frames. We measured speed of propagation, space constant, and period of the waves. In some experiments, we altered the concentration of magnesium chloride in the environment by mixing ratios of normal sea water and isotonic magnesium chloride in the butter dish. We ran experiments at 1 to 2, 1 to 5, and 1 to 7 isotonic magnesium chloride to normal sea water ratios. We let the animal acclimate in the high magnesium for 45 minutes before filming. In other experiments, we transected the radial nerve in living Pannychia. Using scissors, we snipped through the outer layer of the animal about five centimeters from the caudal tip. In cutting the nerve cord, we also cut epithelial layers around the nerve. In one animal, we made two transections in a single radial nerve spaced one centimeter apart. Results General appearance of bioluminescence When prodded gently with a glass rod, Pannychia display a complex pattern of luminescence. I use the word wave to describe the propagation of a patch of light on the body surface. An area of excitation consists of a section of the body that contains any number of waves of luminescence. One response refers to all the excitations caused by a single prodding. Only the dorsal surface of the animal luminesces. The caudal end of the animal responds more frequently regardless of stimulus location. The papillae on the dorsal surface have no light-producing capabilities (Fig 1). Animals show three distinct classes of response. The most common class is defined waves of light moving across darkness. Often, one portion a few square centimeters in area remains excited for several seconds before this excitation migrates gradually, saltates to another area, or ceases entirely. However, in some animals, excited areas luminesce continuously with waves of greater intensity moving within the glowing area. One Pannychia has primarily small spots flashing rather than patches of light. Wave front shapes Whatever the class of response, the wave fronts form characteristic shapes as well as complex patterns (Fig. 2). One shape appears in the same location up to 15 times before the excitation evolves and the pattern disrupts. In general, the direction of propagation remains constant during repetitive waves; however, on one occasion we observed a reversal of wave direction while the wave shape and location remained unchanged. All wave shapes annihilate upon collision (Fig. 3). Annihilation is the abrupt loss of the intensity that occurs when two wave fronts approach each other's refractory region. Circular waves spread out from a single pulsing point. Circular waves sometimes evolve into spiral waves with period 0.42 +- 0.04 s. Propagating waves consist of wave fronts moving in a defined direction. These waves move most often along the radial nerve cords and have curved or straight wave front shapes. Propagating waves occasionally jump from one nerve cord to the other. Some waves display dark bands perpendicular to the direction of propagation. These dark bands consist of dark spots located at the base of the papillae as well as patches connecting these spots. In some waves with dark bands, a colorized frame from optical disk shows that patches connecting the dark papillae spots are in fact faintly illuminated (Fig 4). Some propagating waves are what I term decomposable. In decomposable waves, the overall direction of propagation appears to be longitudinal; however, the excitation is actually a series of small waves transverse to the radial nerve (Fig 5). Quantitative analysis of propagating waves The mean conduction velocity is 16.8 + 8.8 cm/s. Measurements of the conduction velocity of individual waves range from 5.18 to 40.3 cm/s. Even within a single animal, the velocity typically varies 10 cm/s or more between different locations on the body. For each measurement, we print a line on the monitor and watch the waves in slow motion until both observers agree that the line accurately reflects the direction of propagation of the wave. If the line is not acceptable, either we correct the line and re-run the analysis or we find alternate waves to measure. The image processor plots the intensity at points along the line through time. We also reject any readings whose scans through time do not have a definite slope of maximum intensity. We aim for a total of ten or more velocity readings along at least four different wave fronts for each stimulation transferred to optical disk, but sometimes only one or two velocity readings are acceptable. Velocities from decomposable waves are not included in averages and other analysis because their direction of wave propagation is not clearly defined. For one such excitation, the radial portion has speed 26.8 +10.7 cm/s. The transverse wavelets have a mean conduction velocity of 7.06 +2.1 cm/s. The period between repetitive waves ranges from 0.33 to 0.69 s with a mean of 0.50 + 0.09 s. Figure 6 shows the dispersion relation, defined as the dependence of conduction velocity on period. The points consist of averages of all the velocity and period readings from a single stimulation. Waves that do not repeat at least three times are not included in this analysis. The space constant l, defined as the distance between the maximum intensity and the leading edge of the wave front, averages 0.037 + 0.016 cm. The space constant is greater when the wave front is curved rather than straight (Fig. 7). The curved and straight velocities are statistically different (Mann-Whitney U-test, p«0.01). Magnesium chloride vs. normal sea water experiments Adding magnesium chloride at a ratio of 1/3 isotonic magnesium chloride to 2/3 normal sea water to the Pannychia's environment blocks all luminescence. Lower ratios increase the threshold of the organism so that only stimuli of greater magnitudes cause luminescence. The conduction velocity is not conclusively altered by magnesium chloride (Fig. 8). The period between waves increases in the high magnesium chloride experiments (Fig. 9). The normal and magnesium periods are significantly different (Mann¬ Whitney U-test, p«0.05). Trends and velocities in transection studies Transection of the radial nerve can block the spread of luminescence (Fig. 10). When a nerve cord is cut once and the animal is stimulated ipsilateral and rostral to the cut, most of the luminescence remains rostral to the cut, often with a marked wall at the site of the lesion. Large responses can sneak around the cut, but do not usually propagate far. A double transection of the same nerve cord effectively blocks all transmission of luminescence caudal to the lesion. The response to an ipsilateral stimulus in the tail does move rostral to the double cut. Contralateral rostral stimuli do propagate caudal to the single and double transections. The mean velocity of responses in transected animals is 15.5 +1.3 cm/s. Discussion The patterns of response in Pannychia suggest control by nerve nets as well as signaling by radial nerves. The average diffusion constant Dç=vA for the Pannychia luminescence was 5.8E+9 u2/s. Molecules diffuse at a maximum rate of 1000 u2/s in aqueous solutions (Meyer, 1991); hence, the luminescent waves cannot be account for by molecular diffusion. Modeling neural activity as the diffusion of charge, Meyer calculated a diffusion coefficient of 6.7E+10 u2/s for a giant squid axon (1991). Because long-distance voltage diffusion occurs only in the nervous system, the quickly diffusing bioluminescence in Pannychia must be under neural control. The tendency of the caudal end to luminesce more readily also suggest neural control, for molecular diffusion generally would not display such polarity. The tail of the Ptychodera flava (Baxter and Pickens, 1964), a luminescent system controlled by nerve nets, is also more sensitive to stimulation than the rest of the body. The decomposable waves along the radial nerve involve transmission of an impulse along the nerve cord that excites the nerve net locally. Hence the signal has an overall longitudinal direction as the radial nerve triggers sequential impulses to spread transversely through the nerve net. This hypothesis is supported by the relative velocities of the longitudinal and transverse parts. The 26.8 cm/s speed measured for the longitudinal part approaches the conduction velocity of 55 cm/s in an ophiuroid radial nerve (Yee et al., 1987). However, the 7.06 cm/s velocity of the transverse waves is similar to nerve net conduction speeds in other organisms. Pantin reported velocities of 10-20 cm/s for lateral conduction in coelenterate nerve nets (1935). The complexity of bioluminescent waves is based on the interplay between pattern and randomness. Studying intensity in an excitation area through time often shows that each point retraces the same pattern of intensity with each repetition of the wave. However, one cannot predict whether a pattern will repeat twice or twenty times or where the excitation will migrate next. Waves tend to retrace their predecessor's paths because of facilitation. Facilitation occurs in nerve nets when each impulse eases the passage of subsequent impulses. The radial nerve is most likely responsible for the saltation of an excitation area, whereas migration of excitation areas suggests facilitation. Facilitation enables migration because impulses travel slightly farther than previous impulses, leading to an overall extension or movement in the direction of wave propagation. The transection experiments support the idea that radial and nerve nets mediate different modes of evolution of excitation. When excitation cannot travel through the radial nerve, luminescence migrates rather than saltates. A double cut presents a greater block to luminescence than a single cut by forcing the waves to travel through greater lengths of the decremental facilitated nerve net. The radial nerves work independently of each other to respond to local stimuli, since ipsilateral and contralateral stimulation had different effects. The nerve cords can communicate, however, for excitation in non-transected animals can jump from one nerve cord to the other. Using bioluminescence as a monitor of neural activity, we can describe some properties of the Pannychia nerve net. The range of conduction velocities between 5 and 40 cm/s is consistent with neural conduction speeds. The wide range of velocities is due to two main experimental errors: the decrease in normal speed for a curved wave front and the difference in conduction speeds in waves controlled by radial nerves and by nerve nets. Even so, Brehm reported a range of 10 to 85 cm/s for bioluminescent propagation in an ophiuroid (1977). An area of nerve net characterized more by interneural facilitation than by through-conduction should have a lower velocity (Pantin, 1935). Different paths of spread might involve more synapses; hence, synaptic delay may decrease velocity. Therefore, even if the two experimental errors limiting appropriate comparison were avoided, we would see a range of velocities. Because nerve nets act as excitable media, I thought the neurally controlled bioluminescence might also be an excitable medium. The existence of spiral waves and annihilation proves the luminescence has a refractory period. Spiral waves exist because a regular wave front such as a circular wave hits a region with variable refractory periods, breaking the circle into a regular spiral wave whose period depends on the refractory period. Waves annihilate rather than pass through each other only in refractory media. The absolute refractory period of the nerve net is necessarily less than the minimum period observed between bioluminescent pulses. Our maximum limit on the refractory period is 0.33 s, much greater than the estimates for refractory periods in nerve nets of 40 to 65 ms (Pantin, 1935). The diffusion constant relationship often applied to excitable media, De = vA, explains why A increases with curvature, for normal velocity decrease for a curved wave front. The dispersion relationship is a central equation describing excitable media. Excitable media generally display an increase in velocity with increasing period, for the medium is still in a relative refractory period if the new wave is initiated too soon. The opposite trend observed in this experiment suggest that the refractory period is short enough that such slowing is not seen. The absolute refractory period calculated by one model for calcium stores was 4.7 s, two orders of magnitude greater than the nerve net refractory period (Lechleiter et al., 1991). Hence, this dispersion relation predicted by theory (Gerhart et al., 1990) and verified experimentally using the Belousov-Zhabotinskii reaction (Foerster et al., 1989) may apply only to media excited by molecular diffusion. 10 At extremely short periods, the conduction velocity of Pannychia bioluminescence should decrease with decreasing period; however, the theory is not sufficient to describe the dispersion relation at naturally-occurring periods. The regularity of the impulses suggest control by a pacemaker neuron. Bullock proposes two theoretical mechanisms for endogenous rhythms, oscillating circuits and pacemaker neurons, but states that no rhythmic oscillatory circuits have been described in invertebrates; therefore, the neural rhythm is probably set by pacemaker neurons (1965). Most pacemakers have, when excited, an inward current that gradually depolarizes the cell to threshold. At threshold, the action potential is generated, followed by quick repolarization. The repolarization returns the cell to its initial state in which the inward current causes the cycle to recur. Because the pulsing points that initiate bioluminescence can be located anywhere on the dorsal surface of the animal and because several points may be pulsing independently at the same time, the pacemaker neurons are most likely diffusely spread throughout the nerve net. The period of the pacemaker depends on the magnitude of the inward current, the threshold voltage, and properties of the cell membrane. A nerve cell with a constant current can be modeled as a simple RC circuit (Jack et al., 1983). The equation V = IR(I-etR describes the voltage during the charging of a capacitor. The time required to reach threshold is therefore t =-RC*In(1 - Vth/IR). Increasing the threshold voltage will therefore increase the period as it takes longer for the cell to reach threshold voltage. 11 The conduction velocity in a nerve net is also related to the threshold voltage for action potentials. If the threshold voltage is higher, accumulating enough charge for the post-synaptic potential to reach the threshold voltage will take longer. This increase in synaptic delay will slow conduction through the nerve net. Since an increase in threshold corresponds to both a larger period and a slower speed, the dispersion relation determined experimentally agrees with the properties of the pacemaker neurons. The increase in period under high magnesium conditions suggests that the pacemaker neuron is depolarized by a current carried at least in part by calcium ions. Most pacemaker neurons studied in invertebrates have currents carried by calcium ions. Most pacemaker neurons studied in Aplysia (Gorman et al., 1982), Helix (Gerschenfeldet al., 1986), Tritonia (Smith and Thompson, 1985), and calf Purkinje cells (Kass, et al., 1978) all have calcium-dependent pacemakers. In Aplysia, the channels are permeable to magnesium, but the mean open time is shorter when magnesium replaces calcium in the environment. Other calcium channels are impermeable to magnesium. Since magnesium blocks or alters calcium channels, the change in period at high magnesium concentrations indicates that calcium channels are involved in the oscillation of the pacemaker neuron. 12 Acknowledgments Thank you to Edge, for studying the silent scream with me, for measuring my kneecaps, and for making all those long nights fun. to Stuart, for both giving advice and holding back so I could think for myself. to Sam, for teaching me those things a real physics major should know. to Molly, for advice, friendship, and good luck head-rubs. to Chuck, for the Pannychia and the wise words. to Chris, for all-around help and for good advice on my presentation. to Erin, Jay, Ildiko, and Adawia, for making me feel welcome whenever I actually appeared during the day. to Michelle and Ravi, for our silly, eccentric, amazing home. Works Cited Baxter, Charles H. and Pickens, Peter E. (1964). Control of luminescence in hemichordates and some properties of a nerve net system. Journal of Experimental Biology 41. Brehm, Paul (1977). Electrophysiology and luminecence of an ophioid radial nerve. Journal of Experimental Biology 71, 215. Bullock, T.H., Mechanisms of integration. In Bullock, T.H. and Horridge, G.A. Structure and Function in the Nervous system of Invertebrates. (W.H. Freeman and Company, San Francisco, 1965) 314 - 322. Foerster, P., Muller, S. C., and Hess, B. (1989). Critical size and curvature of wave formation in an excitable chemical medium. Proceedings of the National Academy of the Sciences of the USA 86, 6831-6834. Gerhart, M., Schuster, H., and Tyson, John J. (1990). A cellular automaton model of excitable media including curvature and dispersion. Science 247, 1563-1566. Gerschenfeld, H.M., Hammond, C., and Paupardin-Tritsch, D. (1986). Modulation of the calcium current of molluscan neurones by neurotransmitters. Journal of Experimental Biology 124, 73-91 Gorman, A.L.F., Hermann, A., and Thomas, M.V. (1982). Ionic requirements fo rmembrane oscillations and their dependence on the calcium concentration in a molluscan pace-maker neurone. Journal of Physiology 327, 185-217. Jack, J.J.B., Noble, D., and Tsien, R.W. Electrical current flow in excitable cells (Clarendon Press, Oxford, 1983). Kass, R.s., Lederer, W.J., Tsien, R.W., and Weingart, R. (1978). Journal of Physiology 281, 187-208. Lechleiter, J., Girard, S., Peralta, E., and Clapham, D. (1991) Science 252, 123-126. 14 Meyer, Tobias (1991). Cell signaling by second messenger waves. Cell 64, 675-678. Pantin, C.F.A. (1935). The nerve net of the Actinizoa. Journal of Experimental Biology 12. Smith, S., and Thompson, Stuart H. (1987) Slow membrane currents in bursting pace-maker neurones of Tritonia. Journal of Physiology 382, 425-448. Winfree, Arthur T. When Time Breaks Down. (Princeton University Press, Princeton NJ, 1986). Yee, A., Burkhardt, J., and Gilly, W.F. (1987). Mobilization of a coordinated escape response by giant axons in the ophiuroid Ophiopteris papillosa. Journal of Experimental Biology 128. 15 Figure Captions Figure 1: The waves have dark spots because the papillae do not luminesce. Figure 2: Common shapes of waves. a) A circular wave on the left-hand side of the luminescence, single frame. b) A spiral wave, single frame. c) A propagating wave, accumulation of 20 frames. Figure 3: Annihilation a) The wave fronts in the upper right approach each other, single frame. b) Three frames later, the wave fronts are just about to collide. c) Four frames later, the intensity of hte colliding wave fronts has almost disappeared. Figure 4: Dark bands, accumulation of 20 frames. Patches between papillae sometimes show faint lighting (purple dots). Figure 5: The large arrow represents the overall direction of decomposable waves. The small arrows represent the individual wavelets that create the decomposable effect. Wavelet 1 fires just before wavelet 2, followed by 3 then 4 as the excitation propagates in the direction of the large arrow. Figure 6: The dependence of speed of propagation on period. Each point represents the average speed and average period for a single bioluminescent response. Figure 7: Comparison of the space constant X for curved and straight wave fronts. Figure 8: Comparison of the speed of propagation of bioluminescence in normal sea water and in high magnesium chloride. Figure 9: Comparison of the period between waves in normal sea water and in high magnesium chloride. Figure 10: Luminescence stops abruptly at the first cut in a double transection of the radial nerve, single frame. 16 2 — 100 — 0.1 Period (s) 30 tv o 8 ooo o Lambda (cm) oo Lambda (cm) 18 Conduciton Vel ter (cm/s) HI 10 15 20 25 30 35 40 Conduciton Velocity in High Magnesium Chloride (cm/s) 2.5 0.5 O Period in Normal Sea Water (s) e. 1 Period in High Magnesium Chloride (s)