Temperature Effects On Neuromuscular Transmission (Opener Muscle of Crayfish, Astacus Leptodactylus)

Three preceding publications described the effect of temperature on neuromuscular transmission in the closer muscle of walking legs of the crab Ocypode ceratophthalma (Florey & Hoyle, 1976) and of the crayfish Astacus leptodactylus (Harri & Florey, 1977,1979). The stenotherm Ocypode is incapable of co-ordinated movement outside a narrow temperature range. The eurytherm crayfish, on the other hand, shows normal locomotion over the entire temperature range from o° to 26 °C. Correspondingly, muscle tension reached during repetitive nerve stimulation shows a narrow temperature optimum in Ocypode, but in Astacus neuromuscular transmission becomes more and more effective as the temperature declines.

Apart from the oecophysiological interest of a comparison of the performance of motor control systems in stenotherm and eurytherm animals, the phenomenon of a physiological system which is capable of functioning over a temperature span of almost 30 °C presents a fascinating problem: how can a process like neuromuscular transmission, which obviously involves a series of chemical reactions (e.g. transmitter synthesis, transmitter action, calcium release, actin-myosin interaction), remain functional during temperature changes which ought to alter reaction rates by a factor of perhaps 8 (Q₁₀ of 2) or even 27 (Q₁₀ of 3)? As we will show, some of the events involved in neuro-muscular transmission of the crayfish have a Q₁₀ as high as 7!

Although one of the preceding investigations had already shown how the interaction of positive and negative temperature coefficients of different parameters can achieve a kind of temperature compensation (Harri & Florey, 1977), it is clearly necessary to elucidate the underlying mechanisms further.

We therefore decided to supplement the observations made on the crayfish closer muscle with data obtained from another muscle, the opener, and to extend the investigation to include measurements of membrane resistance and excitation-contraction coupling.

A preliminary account of part of this investigation has already been published (Fischer & Florey, 1978).

The experimental animals, mature male specimens of Astacus leptodactylus and, for supplementary experiments, Procambarus clarkii and Pacifastacus trowbridgii were maintained in aquaria as already described (Harri & Florey, 1977) at a temperature of 10 °C. Data were collected throughout the years of 1978, 1979 and 1980.

Third walking legs were always used. They were amputated at the autotomy joint. The opener muscle was exposed by removing part of the ventral exoskeleton of the propodite, the closer muscle, and the connective tissue covering the opener muscle. The dactylopodite was trimmed to a small stump and this was freed from its attachment to the propodite but left connected with the opener tendon. The stump was later used as a handle on the tendon for attachment to the transducer (see below). The leg nerve was exposed by removing part of the exoskeleton and all the muscles from the meropodite. It was separated into small bundles with glass needles. The preparation was then transferred to a double walled perspex chamber and mounted in the inner compartment with small pieces of wax. The experimental set-up is shown in Fig. 1.

Water of the desired temperature was circulated between the inner and the outer wall of the chamber. A constant-temperature circulating bath (Lauda, type K2R) permitted precise regulation of temperature anywhere between o and 25 °C. The bath temperature was monitored with the aid of a thermoelement (reference point in o °C ice water) connected to an indicating d.c. amplifier (Knick, type SV). The inner muscle bath (volume 1·5 ml) was perfused with saline which passed first through a heat exchanger coil placed in the thermostated water jacket of the chamber. The saline had the following composition: NaCl 205, KC1 5·4, CaCl₂ 13·5, MgCl₂ 2·6 mm. The perfusion rate was 2 ml/minute. In most experiments this saline was buffered to was 7·2 by the addition of 10 mm TRIS (tris-(hydroxymethyl)-aminomethane maleate). We also tried TES (tris-(hydroxymethyl)-methyl-2-aminoethane sulphonic acid), bicarbonate/CO₂ as well as unbuffered solutions. No obvious differences were noted when these other salines were used.

The nerve bundles containing the motor axon or the inhibitory axon supplying the opener muscle were taken up into suction electrodes. Nerve stimulation was applied using a programmable square wave stimulator and isolation units (Digitimer, type D4030).

For the recording of muscle tension the stump of the dactylopodite, which had been left attached to the opener muscle tendon, was grasped by the tips of fine forceps attached to an RCA 5734 transducer tube. After suitable amplification, the signal was displayed on one channel of a two channel oscilloscope (Tektronix R 5103) and stored on magnetic tape (8-FM-channel tape recorder, Hewlett Packard, type 3468). It was also recorded on a dual-channel chart recorder (Gould Brush, Mark 220). For recording membrane potentials we used 3 M-KCl-filled glass microelectrodes (resistance 5–15 MΩ). In order to avoid effects of temperature on the electrode potential of the AgAgCl reference electrode, this electrode was placed in a beaker filled with 3 M-KC1 solution (held at room temperature) which was connected with the muscle bath by way of an agar bridge (5 % agar in saline). The recording electrode was connected to a high-impedance balanced DC-amplifier provided with capacity neutralization (Electronics Division, University of Konstanz). The output of the amplifier was displayed on the second channel of the oscilloscope, stored on magnetic tape and registered on the chart recorder. For current injection, rectangular pulses delivered by a constant-current device (Electronics Division, University of Konstanz) were applied to a second microelectrode (3 M-KCI or 2 M-K citrate) inserted into the same muscle fibre. Current was monitored on a separate channel of the tape recorder. The recorded junction potentials were fed from magnetic tape into a desktop computer system (Hewlett Packard 9810, later Nicolet MED 80) to evaluate amplitudes and time constants.

When measured at 10 °C the resting potential of distal fibres of the opener muscle was found to have a mean value of 76·6± 1·15 mV (mean±S.E.M. ; n = 32; different preparations). The resting potential is clearly temperature dependent; it increases (to more negative values) with increasing temperature. As already observed for the closer muscle (Harri & Florey, 1977, 1979) this temperature dependence usually change around 15 °C : in a series of 11 experiments covering the entire temperature range considered in this investigation we found that below this temperature the potential changes by 1·13±0·52 mV/°C, but that above 15 °C the change is much smaller, namely 0·5 ± 0·5 mV/°C (mean ± S.E.M. ; n = 11). Similar slopes were found in other, less complete experiments. In a few fibres the slope of the temperature dependence was more or less the same throughout the temperature range, and in some experiments an inversion of the temperature dependence occurred at higher temperatures so that between 26 and 34 °C the membrane potential decreased with increasing temperature. Four examples of the temperature dependence of the membrane potential are shown in Fig. 2 to illustrate extremes of the variability of this relationship. In a few experiments in which the temperature was raised above 26 °C the membrane potential decreased. We did not persue this inverse relationship between temperature and membrane potential because these high temperatures were found to be deleterious to the preparations.

It has been shown for crayfish muscle fibres (Orkand, 1962) that in order to initiate tension development the fibres must be depolarized to the threshold for excitation-contraction coupling. We measured this threshold potential as well as the resting potential in opener muscle fibres at several temperatures, using two microelectrodes and a force-displacement transducer as described in the methods section (see Fig. 1).

In five successful experiments the mean threshold potential at 10 °C was 53·6 ± 2·2 mV and the mean resting potential was 77·5 ±3·9 mV (mean ± S.E.M.). AS can be seen in Fig. 3, the threshold potential is relatively independent of temperature: between 10 and 26 °C the threshold is nearly constant and when the temperature was lowered from 10 down to 2 °C the threshold decreased only by about 3 mV.

As already observed by Orkand (1962) the greater the suprathreshold depolarization, the faster increases the tension and the higher is the tension reached after a given time span. Fig. 4 shows the tension reached after 2 s of depolarization to different membrane potentials at 11 and 23 °C. The experiments clearly show that for a given supra-threshold depolarization the tension reached is highly temperature dependent. The curves which describe this relationship between depolarization and tension (see Fig. 4) intersect with the potential axis at the threshold potential; their slope defines the transduction of supra-threshold potential to tension. This slope can be expressed in N/V and will be referred to as the coupling factor. Particularly at higher temperatures (above 20 °C) the slope may increase with increasing suprathreshold depolarization. Fig. 5 shows typical examples of the temperature dependence of the coupling factor. It can be seen that at 2 °C a 1 mV supra-threshold depolarization of 2 s duration elicits a tension of only 12 μN, but at 32 °C the same depolarization level causes a five fold greater tension of 58 μN. It is obvious that excitation-coupling becomes more effective with increasing temperature. The Q₁₀ values are around 2.

In the case of nerve evoked tension development the difference between resting potential and threshold potential has to be overcome by the depolarization due to facilitated and summated e.p.s.p.s. Total depolarization during a train of e.p.s.p.s is the result of four factors: (1) the amplitude of the first, unfacilitated e.p.s.p., (2) the facilitation ratio (the quotient of the amplitude of the last or nth e.p.s.p. and the amplitude of the first e.p.s.p. of the train), (3) the time course of facilitation, and (4) the time constant of e.p.s.p.-decay (which determines the degree of summation). All these factors are temperature dependent.

1. The amplitude of unfacilitated e.p.s.p.s varies from fibre to fibre. We found maximal values of 4 mV in some fibres, smaller ones, down to less than 100 μV, in others. The e.p.s.p.s of very small amplitudes might be the result of injury to nerve twigs supplying some or many of the particular fibre’s nerve terminals. For this reason fibres giving small e.p.s.p.s were rejected.

   The unfacilitated e.p.s.p.-amplitude is temperature dependent. Fig. 6 A presents the results of 12 experiments carried out between August and October of 1979, in which the preparations were progressively but slowly cooled from 26 to 2 °C and then warmed again. Starting from a low value of 0 2 mV at 26 °C, the mean e.p.s.p.-amplitude increased with decreasing temperature and reached 14 mV at 2 °C. Above 26 °C no more e.p.s.p.s could be detected in 8 out of these 12 experiments. This effect of high temperature was reversible by cooling the preparations below 26 °C. In earlier experiments e.p.s.p.s could be elicited up to and even above 30 °C; the tern-perature dependence of e.p.s.p.-amplitudes was, on the whole, similar to that shown in Fig. 6 A except that in a few preparations the e.p.s.p.-amplitude had a temperature optimum between 20 and 10 °C. The observed differences may be the result of seasonal variations.

2. The facilitation ratio for consecutive e.p.s.p.s is interval-dependent: the shorter the interval between stimuli ( = e.p.s.p.s) the greater is the facilitation ratio. For any given stimulus interval, the facilitation ratio is remarkably independent of the temperature, as is shown in Fig. 6B. The absolute amount of facilitation, however, is highly temperature dependent since the amplitude of the first e.p.s.p. varies with temperature (see Fig. 6A); thus the difference between first and second e.p.s.p. changes with the same temperature coefficient as the first (unfacilitated) e.p.s.p. amplitude.

   Fig. 6.

   View largeDownload slide

   (A) Effect of temperature on the amplitude of unfacilitated e.p.s.p.s from 12 different preparations. From each experiment the average value of 40 consecutive measurements made at each of the indicated temperatures was used. The points give the mean of these values, the bars represent the S.E.M.

   (B) Effect of temperature on the facilitation ratio (the amplitude of the second e.p.s.p. resulting from a stimulus pair divided by the amplitude of the first). Three stimulus intervals were used: 1000, too, and 25 ms. Pooled data from 12 experiments (same as represented in 6 A). For each experiment the average value of 10 measurements made at each selected temperature was entered in the computation of the mean values shown here. Vertical bars = S.E.M.

   Fig. 6.

   View largeDownload slide

   (A) Effect of temperature on the amplitude of unfacilitated e.p.s.p.s from 12 different preparations. From each experiment the average value of 40 consecutive measurements made at each of the indicated temperatures was used. The points give the mean of these values, the bars represent the S.E.M.

   (B) Effect of temperature on the facilitation ratio (the amplitude of the second e.p.s.p. resulting from a stimulus pair divided by the amplitude of the first). Three stimulus intervals were used: 1000, too, and 25 ms. Pooled data from 12 experiments (same as represented in 6 A). For each experiment the average value of 10 measurements made at each selected temperature was entered in the computation of the mean values shown here. Vertical bars = S.E.M.

3. Although the facilitation is maximal between first and second e.p.s.p., facilitation proceeds during continued repetitive stimulation of the motor axon until the e.p.s.p.s reach a maximal value characteristic of the particular stimulus frequency and temperature. This maximal amplitude is conspicuously temperature dependent as shown in Fig. 7. In the example shown the extent of facilitation increases as the temperature is lowered from 26 to 2 °C. Since the pattern of facilitation varies somewhat from fibre to fibre we did not attempt to evaluate the temperature effect on the time course of facilitation.

   Fig. 7.

   View largeDownload slide

   (A) Typical e.p.s.p. sequences obtained during repetitive stimulation of the motor axon at different temperatures to show the progressive summation of e.p.s.p.s as the temperature is lowered. At the lowest temperatures the contraction artefact disturbs the records ; dotted lines indicate the configuration of the potentials as it would be seen in the absence of the artifact. Note the different time scales for recordings obtained with the two stimulation frequencies of 10/s (left) and 20/s (right).

   (B) Temperature effect on e.p.s.p. amplitude (same experiment as in A). Lowest curve: unfacilitated e.p.s.p.s; middle curve: 20th e.p.s.p. at 10/s; upper curve: 20th e.p.s.p. at 20/s.

   Fig. 7.

   View largeDownload slide

   (A) Typical e.p.s.p. sequences obtained during repetitive stimulation of the motor axon at different temperatures to show the progressive summation of e.p.s.p.s as the temperature is lowered. At the lowest temperatures the contraction artefact disturbs the records ; dotted lines indicate the configuration of the potentials as it would be seen in the absence of the artifact. Note the different time scales for recordings obtained with the two stimulation frequencies of 10/s (left) and 20/s (right).

   (B) Temperature effect on e.p.s.p. amplitude (same experiment as in A). Lowest curve: unfacilitated e.p.s.p.s; middle curve: 20th e.p.s.p. at 10/s; upper curve: 20th e.p.s.p. at 20/s.

4. Whenever the interval between successive e.p.s.p.s is shorter than the decay time of the individual e.p.s.p.s, summation occurs. A train of stimuli of such frequency, therefore, causes each successive e.p.s.p. to take off from the decay phase of the e.p.s.p. immediately preceding it. This inflexion potential reaches a constant level when facilitation has reached its maximum and when facilitation rate was dropped to zero. This level is the ‘plateau depolarization’ already discussed in a preceding publication (Harri & Florey, 1977). During each train the amplitudes of the e.p.s.p.s appear to sum with this developing plateau depolarization and give rise to peak depolarizations which may exceed the maximal e.p.s.p. amplitudes severalfold as can be seen in Fig. 7 in the examples of e.p.s.p. trains recorded at low temperatures.

Clearly, the extent of this plateau depolarization is temperature dependent. The reason for this is the negative temperature coefficient of the time constant of e.p.s.p. decay already described in earlier publications (Florey & Hoyle, 1976; Harri & Florey, 1977). This time constant, τ, can be defined as the time it takes for the potential to decline to 37% (actually i/e) of the e.p.s.p. amplitude measured at the beginning of the exponential decay. A computer program was written to calculate τ from a leastsquare fit applied to the logarithmized decay phase.

The temperature dependence of τ was determined in eight experiments, using the averaged decay of the last e.p.s.p.s of ten consecutive trains recorded at any given temperature. The normalized results are presented in Fig. 8. The average value of τ at 10 °C was 114·8 ± 29 ms (mean ± S.E.M., n = 8); the actual values obtained in the different experiments varied between 70 and 300 ms but, as the normalized data demonstrate, the temperature dependence found in these different experiments is remarkably similar. In the range from 26 to 15 °C, τ increases only slightly but as the temperature declines further the increase becomes progressively steeper and between 10 and o °C reaches a Q₁₀ of – 5 to nearly – 10. At o °C it was not uncommon to record time constants greater than 1 s. In some of the experiments it was possible to lower the temperature to – 2 °C. The negative temperature coefficient of τ had about the same values as that measured between 5 and 0 °C.

The data for the opener muscle resemble those obtained from the closer muscle of the same species (Harri & Florey, 1977) and from the closer muscle of the ghost crab Ocypode (Florey & Hoyle, 1976). It was of interest to extend the comparison further and to investigate the temperature dependence of e.p.s.p. decay in the opener muscle of other species of crayfish. Specimens of Pacifastacus trowbridgii and Procambarus clarkii were maintained under the same conditions as Astacus leptodactylus. As shown in Fig. 9, the temperature dependence of τ is practically identical with that observed in muscle fibres of Astacus leptodactylus. It may be concluded, therefore, that this type of temperature dependence of the time course of e.p.s.p. decay is indeed a general property of crustacean skeletal muscle.

Three mechanisms might explain the observed effects of temperature on the time constant r of the e.p.s.p. decay: (a) a change of membrane capacity, (b) a change of membrane resistance, and (c) a prolongation of transmitter action. The first mechanism may be ruled out since Fatt & Katz (1953) had already shown that membrane capacity of crustacean muscle fibres is not altered by a change of temperature. A change of membrane resistance, however, is a very likely cause of the observed change of τ. Colton & Freeman (1975) had already shown for muscle fibres of Homarus that the temperature dependence of membrane resistance follows a pattern which conspicuously resembles the temperature-induced change of τ found in our own experiments (see Fig. 8), and the same has recently been reported for Astacus opener muscle fibres (Florey & Rathmayer, 1981). It remained now to be shown to what extent the change of e.p.s.p. decay can be explained by a change of passive membrane properties (membrane resistance).

Estimates of membrane time constant based on the decay of electrotonic potentials generated by a current passing electrode showed that this has about the same value as the time constant of e.p.s.p. decay. An identical value cannot be expected because of the cable properties of the muscle fibre. However, the experimentally determined length constant in these fibres is about equal to the fibre length (1–2 mm). This means that more than 80 % of an electrotonic potential can be recorded even at the ends of a muscle fibre when the current electrode is inserted into the middle of the fibre Because of the small distance of 100 μm between current injecting and potential recording electrodes the muscle fibre can be treated as a very short cable. Thus the input resistance (measured as the quotient of recorded potential and injected current) is mainly determined by the membrane resistance and the resistance of the myoplasma can be considered negligible. Records of electrotonic potentials and of e.p.s.p.s obtained from the same muscle fibres, therefore, permit a comparison of membrane resistance with the time constant of e.p.s.p. decay, τ. A typical example of the temperature dependence of these two variables is shown in Fig. 10. The curves are nearly identical.

The possibility that the observed resistance change is due, at least in part, to nonlinearity of the current-voltage relation (the membrane potential changes with temperature) can be excluded. Measurements conducted over the entire temperature range from 2 to 32 °C have shown that the current-voltage relation is nearly linear over a wide range of currents (and membrane potentials) and that there is no anomalous rectification that might account for the observed temperature dependence of input resistance and of τ.

These results permit some important conclusions:

1. Passive membrane parameters (input resistance) and their temperature dependence fully account for the effect of temperature on e.p.s.p. decay time constant.

2. Membrane capacity is temperature independent (otherwise the two curves elating input resistance and τ to temperature would diverge), and

3. the duration of transmitter action has no direct influence on e.p.s.p. decay since the resistance change fully accounts for the observed temperature dependence of τ.

To analyse the temperature dependence of the membrane resistance (or of its equivalent, the time constant of e.p.s.p. decay) further, we made an Arrhenius plot of normalized time constants (Fig. 11). The data points can phenomenologically be fitted by the sum of two exponentials: τ/ τ₁₀ = C₁ exp (– E₁/RT) + C₂ exp (– E₂/RT), where C₁ and C₂ are proportionality constants, E₁ and E₂ are the apparent activation energies, and R is the universal gas constant (1 ·987 cal. °K⁻¹. mol⁻¹). The constants E₁, C₁, E₂, C₂ in this highly non-linear function were evaluated by a Fortran-written optimization program on a PDP 11/34 computer. Both exponentials are presented in the Arrhenius plot as two straight lines. The sum of these yields a curve which fits the actual data points very well. Assuming that the relevant parameter affected by temperature is the membrane conductance, activation energies can thus be computed from the slope of the straight lines. These activation energies are 46 kcal/mol for the steep line and 3 kcal/mol for the flat one. The low value is similar to the activation energies found for diffusion processes (e.g. K⁺-diffusion in aqueous solution; Landolt-Börnstein 1969) and the high value corresponds to activation energies reported for membrane bound proteins (for literature see Alexandrov, 1977). It seems more likely, however, that it is one and the same factor, namely the transport proteins representing the membrane conductance which is altered with temperature in such a way that its activation energy changes from a high value (as measured at 2 °C) to a low value observed at 26 °C.

It is well known that the opener muscle generates tension only when it is repetitively activated by several e.p.s.p.s. The tension reached depends on the frequency and duration of stimulation of the motor axon, but for any given frequency and duration it also depends on the temperature. For the closer muscle it has already been shown that at low temperature very much lower stimulus frequencies are required for the generation of a certain tension than at higher temperatures (Harri & Florey, 1977). We have observed a similar relationship in the case of the opener muscle. Fig. 12 presents the tensions produced by 2 s stimulation of the motor axon at different frequencies and different temperatures. It can be seen that the temperature effects are particularly prominent when the motor axon is stimulated at higher frequencies. It is also obvious that the tension increase for a 1 °C decrease of temperature is most prominent in the temperature ranges of 31 to 24 °C, and from about 12 to 4 °C.

The process of tension development in contractile systems is more effective at high than at low temperatures. This was shown for muscle models of such diverse origin as barnacle muscle (Ashley & Moisescu, 1977) and bovine skeletal and uterine muscle (Hasselbach & Ledermair, 1958). It is not surprising, therefore, that in crayfish opener muscle too the whole process of excitation-contraction coupling becomes more effective with increasing temperature. The temperature coefficient, Q₁₀, between 10 and 20 °C is around 2. For the contractor epimeralis muscle of crayfish, Dudel & Ruedel (1968) had also reported a Q₁₀ of 2 for a similar range of temperatures and membrane potentials.

In contrast to the positive temperature coefficient of excitation-contraction coupling, the nerve-evoked tension decreases with increasing temperature. Since the threshold potential for excitation-contraction coupling has been shown to be nearly independent of temperature (this investigation; see also Dudel & Ruedel, 1968), the main reason for this behaviour must be sought in the temperature dependence of other factors, of which the summation of e.p.s.p.s is the most important one.

As is illustrated in Fig. 13, the tension reached during a train of e.p.s.p.s is proportional to the product of suprathreshold depolarization and coupling factor: although the latter has a positive Q₁₀, the mean depolarization has a negative Q₁₀ value ; it is proportional to the membrane resistance (and time constant). Especially at temperatures below about 15 °C this negative Q₁₀, which may reach a value as high as – 7, overcompensates the positive Q₁₀ of the process of excitation-contraction coupling.

The effect of temperature on the resting potential is greater than can be predicted from the temperature dependence of diffusion processes (as described by the factor 1/T in the Goldman-Hodgkin-Katz equation). However, preliminary experiments using K-free saline gave no indication of a contribution of an electrogenic sodium pump. It is quite possible that temperature affects the permeability to potassium and/or chloride ions as has already been shown for barnacle muscle (Fischbarg, 1972; Dipolo & Latorre, 1972).

Of particular importance is the temperature dependence of membrane resistance (conductance): it fully accounts for the change of time course of postsynaptic potentials (see Figs. 8 and 10); no temperature effect on the duraction of transmitter action need be invoked. To what extent the temperature dependence of e.p.s.p.-amplitude can be accounted for by changes of amplitude and time course of synaptic currents is presently under investigation.

In the intact animal, the forces acting on legs and appendages involved in locomotion and posture are subject to external physical forces which are practically independent of temperature. The negative temperature coefficient of the performance of the peripheral neuromuscular system must, therefore, be compensated by a positive temperature coefficient of the motor output from the central nervous system. In a separate investigation (Harri & Florey, in preparation) it was indeed possible to show that the motor output from the central nervous system to the leg muscles of crayfish walking on a treadmill perfectly compensates for the temperature-dependence of the peripheral nerve-muscle mechanism.

Part of this investigation represents the substance of the doctoral thesis of K. L. Fischer (‘Temperaturabhängige Parameter der neuromuskulären Ü bertragung unter-sucht am Öffnermuskel des Flusskrebses’, Universität Konstanz, 1978). We whould like to thank Miss Birgit Bremer for excellent technical assistance.

This investigation was supported by Sonderforschungsbereich 138 of the Deutsche Forschungsgemeinschaft.