ABSTRACT
The peripheral dendritic arborizations of sensory units X, Y and Z of the oval organ have similar branching patterns. All three permeate the whole array of connective tissue strands without apparent regionalization or specialization.
The analogue components of sensory responses elicited in fibres X, Y and Z when the connective tissue array is stretched show considerable diversity: fibre Z has a higher threshold than X and Y; the dynamic peak values of X and Y saturate at pulls mid-range for Z; X, Y and Z form a spectrum of increasing adaptation.
Application of TTX abolishes impulse generation in fibre X earlier than in fibre Y, indicating diversity in spike initiating mechanisms from one fibre type to another.
Fibre X only spikes between certain limits of membrane depolarization. Usually the response includes one to five spikes which occur during the dynamic phase of a trapezoidal stretch stimulus.
Fibre Y fires throughout the stimulus duration for pulls of moderate amplitude and velocity. Spiking inactivation and a low maximum firing frequency (approximately 80s−1) limit the range of length sensitivity in fibre Y.
Fibre Z attains higher firing frequencies than either X or Y (approximately 110 s−1). The initial burst frequency (velocity dependent) may equal the firing rate of the dynamic peak.
INTRODUCTION
In the preceding paper (Pasztor & Bush, 1983) we have shown that each of three sensory afferents of the oval organ is capable of transmitting both graded potentials and spikes. Thus each fibre has available two methods of information transfer, effectively doubling the number of information channels. Since the oval organ is the only proprioceptor which has so far been identified in the second maxilla as being capable of providing sensory feedback about respiratory beating (Pasztor, 1969), any increase in signalling capacity is likely to be of significance to the animal.
In this paper we investigate the differences which exist between the sensory responses recorded from oval organ afferents X, Y and Z. We will show that the compound response of each fibre has distinctive characteristics which are consistent from one preparation to another. This analysis forms the basis for a current study on the functional role of the oval organ in the control of gill ventilation.
MATERIALS AND METHODS
An isolated preparation of the oval organ was dissected from the second maxilla of Homarus as described in the preceding paper (Pasztor & Bush, 1983). The oval base of cuticle was anchored to the Sylgard lining of the bath, while the apex of the organ was attached to the puller assembly by which stretch stimuli were presented. Standard trapezoidal (ramp-and-hold) stimuli were presented to the organ at 1-min intervals. Pull amplitude, ramp duration and total stimulus duration were controlled by a function generator. During series of graded amplitude stimuli, ramp duration was held constant and, as a consequence, ramp velocity increased concurrently with amplitude.
Intracellular recordings were taken from pairs of 3 M-KCl-filled micropipettes (15–25 MΩ) inserted into adjacent afferents as close together as possible. Responses were recorded on-line or stored on tape for subsequent analysis, using an Apple-Isaac microprocessor system.
RESULTS
Distribution of sensory terminals
In a previous paper (Pasztor, 1979) the sensory dendrites of the oval organ arborization were shown to end in naked bulbous terminals, anchored between epidermal cells at the base of an array of connective tissue strands. That ultrastructural study suggested that the terminals of all three fibres, X, Y and Z, formed a single population indistinguishable from one another, and that each fibre gave rise to branches distributed throughout the whole oval organ.
Cobalt backfills of the sensory arborization, through the cut distal stump of the scaphognathite nerve, have confirmed the pervasive nature of dendritic branching. The photograph of one such backfill, Fig. 1, shows closely similar (though not identical) branching of primary dendrites in two stained sensory units. Both penetrate all parts of the oval, and give rise to an even distribution of terminal branches. There is no indication of regionalization or specialization of the arborization of any one fibre.
Differences between X, Y and Z responses attributable to the receptor potential
The graded potentials of X, Y and Z fibres exhibited great diversity in waveform and amplitude (Figs 2, 3). Fig. 4 plots potential magnitude versus stimulus amplitude for representative X, Y and Z fibres. The peak dynamic and adapted static values were estimated, in the spiking responses of fibres Y and Z, from lines connecting the troughs of the post-spike afterpotentials. These are underestimates, which may displace the curves downwards, but we think they give an acceptable representation of the shape of the graded potentials, since they are comparable with data from tetrodotoxin-treated fibres (see Fig. 5). In the examples shown here in Fig. 4, X and Y responses were recorded concurrently. The Z responses were from an X-Z pair of a different preparation, chosen because the X responses (not shown) matched in threshold and amplitude the X data used in the graph. This ensures that the two preparations were undergoing comparable stimulation.
Three factors emerge from analysis of these and other similar examples. Firstly, fibre Z has a higher threshold to small stretch stimuli than X and Y. Both X and Y can discriminate between small increments of stretch with proportionate increases in graded potential amplitude, within the low part of the stimulus range, which is around threshold for Z (see Figs 3 and 12). Secondly, the dynamic peak values of X and Y saturate at mid-range pulls, leaving Z to encode length information at the upper end of the useful range of the whole organ. Thirdly, by comparing the dynamic and static curves for each fibre, it can be seen that the extent of graded potential adaptation differs significantly between fibres. There is minimal adaptation in X, moderate in Y and maximal in Z. Nakajima & Onodera (1969) found only slight generator potential adaptation in the paired crayfish MRO units, and ascribed their distinctive phasic and tonic firing patterns to differences in impulse initiation. This is not altogether the case with the oval organ as is clearly shown in responses of tetrodotoxin-treated fibres (Figs 5, 6). Once all spiking is abolished, the remaining graded potential in fibre Y (and Z, not illustrated) shows a distinct adaptive decline during the static phase of stretch, and this is reflected in the spike adaptation prior to tetrodotoxin (TTX) application. In contrast fibre X shows almost no receptor potential adaptation after TTX treatment. This in turn suggests that the initial transient peak in the graded potential underlying the brief burst of spikes in X may be a property of the spike generating neural membrane (see below).
Differences between X, Y and Z responses attributable to spiking properties
All three fibres have a peripheral spike initiating zone which is capable of responding to the depolarization of the receptor potential with active regenerative, overshooting spikes. Yet the spiking performances of the three are surprisingly diverse. This can be seen by comparing responses recorded concurrently from pairs of fibres as in Figs 2, 3, 11 and 12.
As described in the preceding paper, the spikes are progressively blocked by 10−7M-TTX. Of interest in the present context, the rate of development of TTX block is faster in fibre X than in fibre Y (Fig. 5). For example, the (dynamic) impulse response of fibre X in Fig. 5 decreased from three to two spikes within 1 min of TTX application, and was abolished altogether within 2 min, whereas the total number of spikes in the Y fibre response did not change for several minutes, although they began to decline in amplitude fairly soon. This suggests a difference in accessibility to the toxin of their spike-initiating zones, and/or a greater susceptibility of the fast inward current channels of the X fibre membrane to TTX block.
Fibre X
Although it has relatively large receptor potentials, fibre X showed a paucity of spikes. From extracellular evidence alone, one might be tempted to suppose that the X fibre represented a fast-adapting phasic receptor, similar to the Pacinian corpuscle, since it only gave 1–5 impulses per stimulus. Intracellular recordings showed otherwise, since the receptor potential adapted very little, and remained at an almost constant level of depolarization for the duration of the pull stimulus (Figs 2 and 3, also Figs 7 and 9 in the preceding paper). Thus, the brief burst may reflect accommodation of the spike initiating mechanism to a prolonged membrane depolarization. This is suggested by the response of the X fibre to depolarizing current injected into the fibre (Fig. 7A). The maintained depolarization brought the spike initiating mechanism to threshold, but after 1 s the mechanism inactivated and the active responses ceased. Accommodation may also be noted in the series of responses to slow ramp stretches at different velocities (Fig. 7B). In each case, spiking started at the same threshold level of depolarization, and then ceased although the membrane continued depolarizing. The cessation came at varying times after the initial impulse, but occurred at a particular level of depolarization (in this example 26mV from the resting potential). This suggests that impulse initiation in fibre X is only expressed between lower and upper limits of depolarization. A similar phenomenon has been noted in neurones of the lobster commissural ganglion (Robertson & Moulins, 1981).
Support for an upper spiking limit comes from the two experiments shown in Fig. 8. In Fig. 8A the prevailing membrane potential was manipulated by a combination of conditioning stretches and current injection. At rest length (l0), the first test pull elicited one spike, but injection of depolarizing current shifted the membrane potential outside the upper spiking limit and the response to the second test pull lacked a spike. Trials at lo + 0·35 mm show the converse situation, where conditioning stretch depolarized the membrane so that the test pull response was spikeless, then hyperpolarizing current restored the membrane potential to the spiking range, and the test pull again elicited a spike. At lo+0·4 mm, the depolarizing effect of stretch was greater and the hyperpolarizing effect of current injection was insufficient to counteract it, resulting in a spike-less response to the test pull. In Fig. 8B the resting length of the organ was increased in 0·1-mm increments (conditioning stretch) and the response was observed to a 0·3-mm pull (test pull) at each increment. As the membrane depolarized, in response to the conditioning stretches, spiking at the test pull became progressively impaired, and in the last example spikes were absent altogether.
Fibre Y
Under certain stimulus conditions fibre Y also shows limitations in its spiking capacity. In an amplitude series like the example shown in Fig. 9 a blockage of spiking was frequently observed at some critical increment in pull. The stimulus amplitude at which block occurred varied from one preparation to another, and in some fibres it was seen at pulls of 0·3 mm, while in others spiking was only blocked for pulls greater than 0·9 mm, depending partly upon the initial degree of stretch. Spiking always ceased at the dynamic peak of the response when firing frequency had attained 50–60Hz (see Fig. 11).
Preliminary experiments on this block indicate that it has a complex origin, but mechanical damage, as suggested by the original term ‘overstretch’, is not an underlying causal factor. As Eyzaguirre & Kuffler (1955) found with the lobster abdominal MRO, impulse block is not permanent and, under appropriate conditions, spiking will start again later during the static phase, at a firing frequency appropriate to the expected state of adaptation. In several respects the pause in spiking of the Y fibre resembles the ‘transition interval’ noted by Shepherd & Ottoson (1965) in the frog muscle spindle. The latter, however, rarely exceeded 25 ms, whereas in the oval organ Y fibre pauses of 500 ms and more were not uncommon.
The occurrence of spiking inactivation was governed by the velocity as well as the amplitude of the pull stimulus (Fig. 10). A blocking stretch of a certain amplitude and moderate velocity sometimes failed to induce block if the velocity was reduced, and a higher amplitude then had to be used to cause block. More surprisingly, high velocity pulls did not induce block, at least up to the maximum amplitude of stretch applied (1 mm). Unlike the situation with fibre X, it does not seem that level of membrane depolarization is the determining factor. As an alternative hypothesis, one might suppose that rate of depolarization could set limits on spiking ability. The spike initiating mechanism might be susceptible to inactivation by too rapid a change in membrane potential. The evidence of the high amplitude, high velocity pulls which do not produce block, however, suggests that both rate of depolarization and duration of dynamic stretch must reach critical values to bring about spiking inactivation.
Fibre Z
No such inactivation of the spiking mechanism was seen in Z fibre responses no matter what amplitudes or velocities of stretch were tested. As long as the depolarization was supra-threshold, spiking continued. The firing pattern was characteristic and readily distinguishable from that of fibre Y. During the dynamic phase, fibre Z could attain a much higher firing rate than Y, even though the underlying graded potential had a similar magnitude. As shown in the instantaneous frequency plots of Fig. 11, firing rates of over 100 Hz were not uncommon in Z whereas the maximum in Y was around 60 Hz. Adaptation during the static phase was, however, much more pronounced in Z, so the firing level dropped rapidly and, during long pulls, usually fell below that of Y (see also Fig. 8 in the preceding paper, Pasztor & Bush, 1983).
As shown in the stimulus-response curves plotted in Fig. 4, as pull amplitude reached the upper end of the range, receptor potential amplitude of fibre Y saturated before that of fibre Z. The effect of this on impulse coding in the two fibres is illustrated in Figs 11 and 12. Not only did Y dynamic frequencies show the saturation effect, but spiking inactivation limited the usefulness of fibre Y in coding amplitude information at the upper end of the range. Fibre Z was not so limited and continued to display increments in firing frequency for increases in stimulus. On the other hand, at small amplitude pulls, fibre Y had the greater sensitivity, as shown in the lower four pairs of responses in Fig. 12. Thus the two fibres have different, though overlapping, usable ranges, and can be said to exhibit a certain degree of range fractionation.
DISCUSSION
The lobster second maxilla performs one basic function, gill ventilation, and the beating of its scaphognathite is a relatively stereotyped behaviour. Water is expelled from the exhalant channel of the branchial chamber by the combined levation-depression and antero-posterior rocking of the appendage in a cyclical rhythm. Beat frequency, ranging from 50–120 min−1, is the dominant variable. The legs, by comparison, are multifunctional and have a wide repertoire of activities including slow postural stances, food gathering, grooming and rapid locomotory rhythms. As a corollary, the legs have a rich proprioceptive system with numerous information channels feeding back to the motor programme generators. (For reviews see Mill, 1976; Bush & Laverack, 1982; Evoy & Ayers, 1982.) The second maxilla has a much simpler array of mechanoreceptors, in which the oval organ predominates. It sends only three afferents to the suboesophageal ganglion, but each can carry two types of signal, graded and impulsive, and each has unique properties.
DISTINCTIVE FUNCTIONAL PROPERTIES OF FIBRES X, Y AND Z
X fibre responses have the largest graded potentials and the briefest burst of spikes. Since the graded potential is virtually non-adapting, X can be assigned the role of position detector. The analogue signal is continuously variable and well suited to transmitting information about extent of stretch, and hence position of the scaphognathite within the beat cycle. As discussed earlier (Pasztor & Bush, 1983), both phases of the cycle, levation and depression, could be encoded in the oscillation of membrane potential of the X unit. The interspike intervals within each burst are too few and too poorly modulated to encode stimulus parameters, but the interburst intervals provide a faithful measure of scaphognathite cycle period. Thus a second function for X is to act as a beat frequency marker.
Y fibres show a well-maintained impulse discharge superimposed upon a lengthdependent potential plateau during the static phase. They too can act as position detectors, but using both the analogue and the impulsive mode of signalling. Were the scaphognathite to be forced in to an unnatural position by an obstruction in the pumping chamber, prolonged stretch could be adequately encoded by firing frequency or membrane depolarization.
During normal gill ventilation the frequency of scaphognathite beating is the parameter showing the greatest variability. Thus, as the beat cycle shortens or lengthens, stretch of the oval organ occurs at higher or lower velocity. As shown in the previous paper, the initial part of the receptor potential waveform is a rapid depolarization where at least the initial phase is velocity dependent. It is this component which could be expected to show the greatest attenuation during the passive spread of the analogue signal into the ganglion, due to the additive effect of membrane capacity. Indeed such a distortion was seen in the transmission of the visual signal in non-spiking photoreceptors of barnacle (Hudspeth, Poo & Stuart, 1977). In Y and Z fibres the initial depolarization is encoded into an initial burst of impulses so that information about velocity is not lost. This makes both fibres candidates for movement detectors. In Z, sensitivity to movement is accentuated by the greater initial dynamic response, its high frequency firing and the rapidity of adaptive fall. As seen in Fig. 11 the early movement response of Z, the initial burst, sometimes elicits higher firing frequencies than the dynamic peak which, coming at the end of the ramp, signals the maximum extent of stretch.
IS THERE A MORPHOLOGICAL BASIS FOR PHYSIOLOGICAL DIFFERENCES?
Where individual afferents from a given sense organ have different physiological properties, it might be predicted that they would be differentiated anatomically. Fine structural differences have been demonstrated between S and T fibres of the crab T-C MRO (Whitear, 1965), fast and slow adapting crayfish abdominal MRO fibres (Euteneuer & Winter, 1979; Komuro, 1981), snake muscle spindles (Pallot & Ridge, 1972, 1973; Fukami, 1978, 1982) and primary and secondary endings in mammalian muscle spindles (Banks, Barker & Stacey, 1981). To what extent physiological differences between fast adapting and slowly adapting units can be ascribed to viscoelastic or other mechanical properties of the receptor organ is still open to discussion, and may differ from one receptor to another (see e.g. Bush & Laverack, 1982). In the crayfish abdomen, for example, the different degrees of adaptation of the two MROs have been attributed largely to differences in spike adaptation (Nakajima & Onodera, 1969). On the other hand, a modelling study on the T-C MRO suggests that the more pronounced differences in the anatomical relationship of the S and T fibres with the receptor muscle could underly their distinctive dynamic responsiveness (Berger & Bush, 1979).
In the oval organ we have so far been unable to reveal any significant differences in terminal structure, either in the distribution of the three fibres throughout the array of connective tissue strands or in the linkage between terminals and surrounding tissues (Pasztor, 1979; Fig. 1 of this paper). Yet we have shown clear differences between the waveforms of receptor potentials in fibres X, Y and Z, and differences in spiking performance over and above those attributable to the receptor potential.
In the absence of a mechanical explanation, a basis for the observed functional diversity must be sought in the peripheral membrane properties of the three fibres. Preliminary experiments indicate a clear difference in spike-generating capacity between fibre X on the one hand and fibres Y and Z on the other. Thus, prolonged depolarizing pulses injected into Y and Z elicit sustained repetitive firing at frequencies directly related to current intensity (at least up to 10 nA), whereas similar currents into fibre X never evoked more than one to five spikes at the beginning of the pulse (cf. Fig. 7A). Moreover the faster development of TTX block of impulses in the X fibre than in Y (Fig. 5) indicates a difference of some kind in their spike initiating or conducting membranes and the underlying ionic mechanisms. One possibility, for instance, might be that the X fibre membrane could contain fast outward current channels (probably for K+ ions) similar to those demonstrated in the crab T-C MRO afferents (Mirolli, 1981). The possible effect of such an outward current, by shunting the fast inward Na+ current, could be to limit spike generation in X, though not to suppress spiking altogether as in the T-C MRO.
ACKNOWLEDGEMENTS
This work was supported by research grants to BMHB from the Royal Society and the Science Research Council (U.K.) and to VMP from the Natural Sciences and Engineering Research Council of Canada. We thank Alison Walford for technical assistance.