Brain Research, 118 (1976) 45-62 © Elsevier/North-Holland Biomedical Press, Amsterdam - Printed in The Netherlands

45

F A C I L I T A T I O N A N D DEPRESSION OF SYNAPTIC TRANSMISSION IN AMPHIBIAN SYMPATHETIC GANGLIA

N. TASHIRO*, J. P. GALLAGHER** and S. NISHI*** Neurophysiology Laboratory, Department of Pharmacology and Therapeutics, Loyola University Medical Center, Maywood, IlL 60153 (U.S.A.)

(Accepted April 12th, 1976)

SUMMARY There have been few reports concerning facilitation and depression in sympathetic ganglia 9,17,4°. In the present investigation, pairs of excitatory postsynaptic potentials (EPSPs) were recorded intracellularly from bullfrog paravertebral sympathetic ganglia for an analysis of the site and mechanism responsible for the phenomena of facilitation and depression of ganglionic transmission. The ratio of the amplitude of the second of a pair of EPSPs divided by the first was compared to the time interval between each pulse. These ratios demonstrated two phases: an earlier phase of facilitation (20-500 msec pulse intervals) and a later phase of depression (500 msec-10 see), Additional parameters - - rate of rise of synaptic potentials (dV/dt), synaptic currents (EPSCs), and synaptic conductances (Gtr) - - were determined and all confirmed the results obtained with EPSPs. Furthermore, the degree of facilitation or depression could be modulated by altering the extraceUular concentration of calcium. On the other hand, comparison of the amplitude of pairs of presynaptic terminal spikes did not show any variability over similar stimulus intervals, nor were the amplitudes of miniature EPSPs significantly different before or after an evoked EPSP. Therefore, the processes of facilitation and depression of ganglionic transmission occur as a result of normal nerve terminal activity. The processes are occurring simultaneously, such that one or the other may predominate depending upon the interval between pulses, as well as the relative concentration of extracellular calcium. * Present address: Department of Neuropsychiatry, Faculty of Medicine, Kyushu University, Fukuoka, Japan. ** Present address: Department of Pharmacology and Toxicology, University of Texas Medical Branch, Galveston, Tex. 77550, U.S.A. To whom requests for reprints should be addressed. *** Present address: Department of Physiology, Kurume University School of Medicine, AsahiMachi, Karume-Shi 830, Japan.

46 The results from these experiments have been interpreted by a hypothetical equation which demonstrates that facilitation of ganglionic transmission is due to activation of a 'residual' calcium and depression is due to depletion of readily available stores of acetylcholine.

INTRODUCTION There are some reports concerning facilitation and depression of transmission in sympathetic gangliag,17,4°. In mammalian sympathetic ganglia, facilitation and depression of the postsynaptic responses following a conditioning presynaptic volley are complex due to the divergent and convergent supplies of presynaptic B and C fibers to the postsynaptic neurons 9. Because of this multiple innervation pattern and almost exclusively axodendritic type of synaptic arrangements 19, a direct assessment of the pre- and postsynaptic factors contributing to facilitation and depression of synaptic responses is extremely difficult. In the amphibian sympathetic ganglion, on the other hand, there is no convergence of presynaptic B and C fibers onto the postsynaptic neurons; the B and C ganglion cells are innervated, respectively, by the presynaptic B and C fibers 47. Moreover, the majority of B ganglion cells are innervated by single presynaptic fibers, and the synaptic boutons are confined chiefly to the lower hemisphere of the unipolar somata26,27,47,5a. Because of this simple, neuronal and geometric arrangement, the amphibian ganglion affords a suitable preparation for the investigation of the mechanism controlling facilitation and depression with conventional intracellular recording methods46 and the voltage-clamp technique. A disadvantage of bullfrog sympathetic ganglia as compared with other preparations is that the normal frequency of miniature EPSPs is very low. Since an accurate measurement of the mean miniature EPSP amplitude is necessary when using the quantum hypothesis to analyze synaptic transmision, a new hypothesis (see Methods) has been described which may prove useful in the analysis of synaptic transmission: especially to examine whether, in the ganglion, facilitation and depression of transmission are mainly a pre- or postsynaptic phenomenon. MATERIALS AND METHODS

Preparation Paravertebral sympathetic ganglia from bullfrogs (Rana catesbeiana) were isolated and prepared for intraceltular recording46.

Recording methods Intracellular recordings were made from B-type ganglion cells. A Wheatstone bridge circuit was used for passing current and recording intracellularly with the same electrode. All microelectrodes were filled with 3 M KCt and had tip resistances of 15-30 M~.

47 Synaptic currents (EPSCs) were obtained by a voltage-clamp technique35. After insertion of the recording electrode into a cell, a second - - the current-carrying electrode - - was inserted into the same cell. The voltage-clamp circuit included one of the operational amplifiers in a Tektronix Type 0 unit as a feedback amplifier, and the other amplifier monitored the feedback current. The current carrying electrode was connected to the amplifier through a variable resistor and capacitor which served to prevent oscillations in the signal 45. An estimation of the rate of rise of the synaptic potential was obtained from a first derivative RC circuit recorded simultaneously on the second beam of the oscilloscope. Extracellular recording of presynaptic activity was similar to that described by Ginsborg 22. Ganglia were superfused with nicotine (1 × 10-5 M) for at least 30 min to eliminate postsynaptic activity and isolate the field potentials that resulted from the excitation of individual presynaptic terminals.

Experimental measurement of facilitation and depression Pairs of stimuli were applied with equal supramaximal voltage and a duration of 0.1 msec. The interval between stimuli was varied (5 reset-10 sec) so that a broad range of paired pulses could be examined. An appropriate delay (30-50 sec) was allowed between successive pairs of pulses to establish a resting equilibrium condition. Paired stimuli were chosen rather than trains of stimuli to make the analysis of the transmitter releasing mechanisms as simple as possible, since repetitive stimulation has been suggested to alter the rate of mobilization of readily available stores of transmitterT M , in addition to changes in probability of release between pulses. The amplitude of each EPSP was corrected for non-linear summation43. To ensure that this correction was as accurate as possible only those EPSPs not greater than 15 mV were used throughout the experiments. The second of a pair of EPSP amplitudes (E2) was divided by the first (El) to yield the ratio (E2/E1). In those experiments in which the cell membrane was clamped at the resting or slightly hyperpolarized membrane potential level and the current monitored, the ratio, EPSC2/EPSC1 (C2/C1), was obtained from the amplitudes of the recorded currents. The maximum shunt conductance of the subsynaptic membrane (Gtr) could be calculated using the maximum rate of rise of the synaptic potential (dv/dt), the membrane potential at which the synaptic potential showed the maximum rate of rise (Vs), the equilibrium potential (Es), the effective membrane resistance (Re), the time constant of the postsynaptic membrane (rm) and the resting membrane potential (Vm) from the following relation (cf. ref. 44). All these parameters were obtained from the individual neurons used to calculate Gtr with the exception of Es, which was always assumed to be --10 mV. Gtr =

Vs--

Vm -]- Z'm dv/dt Re (Vs --

Es)

Knowing the maximum rate of rise of the synaptic potential (dv/dt), the ratio D2/D1 was calculated from the relationship dv2/dt :dvl/dt. Using the values of dv/dt obtained

48 for each EPSP, the maximum shunt conductance was calculated from the above relationship in order to determine the ratio Gtrg/Gtrl. Each point as plotted represents the mean from several cells of at least 10 pairs of pulses at each of the various pulse intervals. The ratios of the paired responses were plotted vs. the log of the pulse interval in msec, which has been chosen for convenience.

Theoretical measurement of facilitation and depression In the presynaptic terminal there are two kinds of stores of releasable acetylcholine (ACh): depot ACh and readily available ACh 5. The kinetics between these two compartments have been postulated by Liley and North 36 and can be summarized as follows: kl ~ k-1

Q

C

~p

where Q = amount of depot ACh; C = amount of readily available ACh; kl and k_l = rate constants; and p = fractional release of ACh. Assuming that the amount of depot A C h (Q) is constant, the amount of readily available ACh (C) at a time t, can be calculated from: dC -

dt

-

klQ-- k-lC

(Eqn. 1)

The boundary condition for this equation is that at t ---- 0, C ---- (l - - p)Cr, where Cr is the amount of ACh readily available for release at equilibrium, or Cr ---kl Q/k_1. Then, one can estimate the amount of readily available ACh at a time t after an initial impulse, Ct: Ct

-

-

Cr (1 - - pie -ht)

(Eqn. 2)

where p~ and l/h are the fractional release resulting from an initial impulse and the recovery time constant for readily available ACh, respectively. It has been suggested that the amplitude of an EPP or the quanta of transmitter released is proportional to the concentration of extracellular calciumS, 4s or the nth power of this concentration~5,25,zs, a3. Based on the assumption that intraceUular ionized calcium within the presynaptic nerve terminal reacts easily with specific sites on the inner surface of the terminal membrane to produce an 'active calcium', Caa, and that this Caa acts to release transmitter from the nerve terminal29, 3~, one can expect the following relations: [Ca]o c~ (Caa) n' c~ (ACh) where [Ca]0 is the concentration of extracellular calcium, (Caa) is the total amount of active calcium, and (ACh) is the amount of A C h released from the nerve terminal. And also the amplitude of an EPSP is assumed to be proportional to the n'th power of active calcium concentration. From the concepts of stores of A C h and the requirement that calcium is essen-

49 tial for evoked release the following relationship can be suggested: (ACh) ---- e C (Caa) n'

(Eqn. 3)

where ~ is a proportionality constant. Since the conductance of the subsynaptic membrane, Gtr, is thought to be proportional to the amount of transmitter released, the maximum Gtr produced by the first impulse is given by Gtrl: Gtrl ----y 0 CrX0 n'

(Eqn. 4)

where Cr is the readily available ACh in the resting state, Xo is the maximum amount of active calcium that is involved with the release resulting from an initial impulse, and y is a proportionality constant. Based on the hypothesis of residual calcium (see Discussion), the total residual calcium at a time t after the first impulse is expected to be Xoe-bt, where the boundary condition is that t = 0 when (Caa) = X0 and 1/b = the recovery time constant for residual calcium. Then, the maximum amount of calcium available for further release of transmitter immediately after a second impulse is given by (1 + e-bt)X0, which is based on the assumption that the second action potential carries the same amount of active calcium into the nerve terminal. The maximum Gtr produced by the second impulse, Gtr2 ,is given by the following equation: Gtr2 = y O Ct (1 -q- e-bt)n'X~'

(Eqn. 5)

where Ct is the amount of readily available ACh, as defined in Eqn. 2. From Eqns. 2, 4 and 5, the ratio of the second Gtr to the first can be expressed as;

Gtr~/Gtrl = (1 - - pie TM) (1 q- e-bt) n'

(Eqn. 6)

The amplitude of an EPSP corrected for non-linear summation is given by E = (Es - - Vm)Gtr/Gm, where Gm is the conductance of the postsynaptic membrane 43. Since Es, Vm and Gm are constant, E is proportional to Gtr, then: Ez/E1 ----- (1 - - pie -ht) (1 q- e-bt) n'

(Eqn. 7)

Solutions Preparations were maintained at room temperature (21-24 °C) in a frog Ringer's solution composed of NaCI, 112 raM; KCI, 2.0 mM; CaCls, 1.8 m M and NaHCOa, 2.4 raM. Atropine sulfate (Sigma) was used to block postsynaptic muscarinic sites. Whenever the composition of the Ringer solution was altered it was done so isosmotitally. Experimental observations were begun 10 min after the Ringer solution was altered. RESULTS

Pairs of stimuli The middle pair of potentials shown in Fig. 1 are examples of orthodromically

50

Fig. 1. An example of the 3 types of responses obtained from a 'B' type ganglion cell at equal stimulus intervals (30 msec). Antidromic (ANTI) and orthodromic (ORTHO) action potentials were elicited at the resting membrane potential of --55 inV. EPSPs were isolated from orthodromic responses by hyperpolarizing the membrane to a level of--100 inV. Calibration: vertical bars show 50 mV and the horizontal bar 20 msec.

induced action potentials. I f this pair is compared to the upper set of antidromically induced action potentials, a difference especially in the size of the afterpotentials is readily apparent. By hyperpolarizing the cell, one can observe the synaptic response as recorded in normal Ringer solution. The action potentials in normal Ringer solution were blocked when the membrane potential was hyperpolarized over 70 mV. The lowest tracing is a pair of EPSPs recorded from the same cell as that of the upper two tracings and at the same time interval. The second EPSP from this cell was about 1.5 x the amplitude of the first at an interval of 30 msec. Facilitation of the synaptic response can also be evidenced as a reduction in the size of the afterhyperpolarization in the second of a pair of orthodromicaUy induced action potentials, I f one monitors simultaneously the first derivative of the orthodromic action potentials (dv/dt) in normal Ringer solution, a later phase of depression can be observed as the pulse interval is increased to greater than 1 sec (Fig. 2). This depression becomes evident as a decrease in the second action potential's value of dv/dt. Experiments were performed both in the presence and absence of atropine (1 × 10 -6 M) added to the Ringer solution; no difference in the results was observed.

Hyperpolarized cells Synaptic responses were examined in 27 cells perfused with normal calcium

51 30

msec

I sec

I00

5

msec

sec

,.

,., S"

50 rnV

I 0 msec

2. Pairs of action potentials which have been differentiated to demonstrate facilitation and depression. The second response in each pair is due to the conditioning impulse, since in these recordings the time sequence has been reversed. The upper tracing of each inset represents the lower half of a pair of action potentials; the lower tracing represents the first derivative of the respective action potentials. The initial upward deflection of the lower tracing is an indicator of the magnitude of the rate of rise (dv/dt) of the synaptic potential. The upper left hand panel depicts a pulse interval of 30 msec; the upper right, 100 msec; the lower left, 1 sec; and the lower right, 5 sec. Calibrations: 50 mV and 10 msec. Fig.

(1.8 m M ) and low calcium (1/2 × normal, 0.9 m M ) Ringer solutions. Fig. 3a depicts the ratios of the paired stimuli obtained from these cells after sufficient hyperpolarization had been applied in order to isolate the EPSP. In the case of the normal calcium Ringer solution, an early phase of facilitation at intervals less than about 500 msec was followed by a later phase of depression lasting up to about 10 see. When the extracellular calcium concentration was lowered to one-half normal (0.9 m M ) , an earlier phase of facilitation, slightly increased in degree was obtained. This facilitatory phase was followed by a phase of depression that was a lesser degree than that observed in the normal calcium Ringer solution. A few cells were treated with a high calcium (5 × , 9 m M ) Ringer solution to compare the effects with those seen in the normal and lowered calcium solutions. However, it was technically very difficult to isolate the EPSPs from their action potentials, even with very high levels of hyperpolarization (over 150 mV), after the change to a high calcium Ringer solution. The few cells in which we did manage to isolate the EPSPs resulted in amplitudes too large and EPSPs too prolonged to use Martin's correction 4a such that these data did not appear to be reliable. However, the ratios - - E2/E1, D2/D1 and EPSC2/EPSC1 - - were analyzed in the presence of high calcium Ringer solution under 3 other circumstances: after the addition of magnesium; during the monitoring of the first derivative of synaptic potentials at normal resting potentials; and after the cells had been voltage clamped (Figs. 3b, c and 4).

High magnesium Ringer-treated cells In order to insure that the facilitatory and depressant phases were not the result

52 20 T

~

T

HYPERPOLARIZED C[ttS

I."

~IIT

2.0

I I 1 b

f 1

HIGH CMg.~ (9-1ernM)RINGER

I

,,,5, cc,~ ..i/2,cc.3

E%

I.(

1.5~

DIFFERENTIATED POTENTIALS

k

;o ;o ,o'

'i'~,a ,,0' Pulse

soo' ,g,, ,ooo' Interval

,,,~o ,,,k,,

( mite )

Fig. 3. Comparison of the time course of Es/E1 and D2/D1 values in media containing various concentrations of Ca z+. a: E2/E1 ratios taken from hyperpolarized cells in normal and 1/2 x Ca solutions, b: E~/Et ratios measured in high Mg solutions containing normal, 5 × or 1/2 × Ca. c: Da/D1 ratios obtained in normal and 5 × Ca media. Each point represents the mean from 12 to 15 cells of at least 10 pairs of pulses; the vertical bars indicate the standard error of the means. o f hyperpolarizing the membrane, magnesium, up to 18 m M , was added to 25 cells in order to block the action potential and isolate the synaptic potential at the normal resting m e m b r a n e potential. Magnesium has been demonstrated to little affect postsynaptic sensitivity to A C h but rather decrease the output o f A C h released by nerve stimulationG,7,sL The same time sequence employed for the recording o f hyperpotarized cells was again repeated in the presence o f magnesium (Fig. 3b). A pair o f curves was obtained which were similar to the two curves obtained in the absence o f magnesium but after

53 membrane hyperpolarization. A third curve, obtained in the presence of high calcium (5 ×, 9 mM) shows an early, less facilitatory phase when compared to the other curves (normal calcium-high magnesium and low calcium-high magnesium Ringer solutions). In a high calcium-magnesium-free Ringer solution, the EPSPs were increased in amplitude and prolonged in duration so that the ratios at intervals of less than 100 msec in hyperpolarized cells were not reliable. Presumably, the increased ratios of E2/E1 at a given pulse interval, obtained in the magnesium enriched solution, demonstrate the competitive action between calcium and magnesium for transmitter release x2' since a similar shift is observed in the hyperpolarized cell in the presence of a low calcium-magnesium-free Ringer solution.

Differentiated synaptic potentials at normal resting potential Due to the difficulty of separating the EPSP from its action potential in the presence of a high calcium Ringer solution it was concluded that another estimate of the paired ratios should be obtained. For this purpose, ratios of the rates of rise of paired pulses at normal resting membrane potentials and during perfusion with a normal and high calcium Ringer solution were obtained from 12 cells. As can be seen in Fig. 3c, the resultant ratios of D2/D1 produced similar curves in regard to facilitation and depression that had been obtained for the ratios of E2/E1 as observed in normal Ringer solution after hyperpolarization and elevated magnesium Ringer solution. Especially significant was the similarity between the ratios, D2/D1, during high calcium Ringer perfusion and the ratios, E2/E1, during the same treatment in the presence of a high calcium-magnesium Ringer solution.

Voltage-clamped cells In all the experiments wherein EPSP ratios have been compared, Martin's correction for non-linear summation 4a has been employed. In order to ascertain that the ratios, E~/E1, were not the result of a mathematical artifact, 14 cells were clamped at the resting or hyperpolarized level and the resultant pairs of synaptic currents (EPSCs) were examined at intervals less than 500 msec (Fig. 4a). These shorter intervals were specifically chosen because it was technically difficult to monitor a cell in good condition for the period of time it takes to record sufficient pairs of potentials beginning with the short stimulus intervals and extending through the longer pulse intervals. Low calcium Ringer solutions were not employed in these voltage-clamped preparations, because under these conditions the amplitude of the EPSC would be markedly decreased and have large variations. The importance of recording the EPSC was due to the fact that in the presence of an elevated calcium Ringer solution the falling phase of the synaptic potentials was greatly prolonged, which may have caused inaccuracies in the measurements of the second EPSP amplitude and of the membrane conductance during the falling phase. In a normal calcium Ringer solution, the EPSC ratios were quite similar to the ratios which were obtained by the EPSPs corrected for non-linear summation. In a high calcium solution, as expected, facilitation was observed at intervals

54 SYNAPTIC C ~ N T

a

Ix(Ca) •-, 5x(Ca)

1.5

c~,

1.0 b 5YNAPTIC CONDUCTANCE 1'5 f G2,"G,

1'01 (msec)

20 !00 200 . . 30 . . 50 ' . . .500 . 10002000 . Pulse Interval

10b00

Fig. 4. C o m p a r i s o n o f experimental data o f the synaptic current (a) - - C~/C~ - - a n d synaptic c o n d u c t ance (b) - - Gs/G~ - - f r o m ganglia perfused with n o r m a l Ringer a n d 5 x calcium Ringer solutions. T h e points represent the experimental values a n d the bars their s t a n d a r d error. See details in text.

less than 500 msec. The degree of facilitation in a high calcium solution, however, was smaller than that in normal solution and the time course of facilitation prolonged compared to that in a normal solution (Fig. 4). A similar observation has been reported at the neuromuscular junction 50.

Compar&on of synaptic conductances at normal resting potentials Since synaptic currents were only compared during intervals up to 500 msec, the calculated synaptic conductances (see Methods) were compared during the longer stimulus intervals. Since the measurements of dv/dt of the synaptic potential from the orthodromic response requires neither artificial hyperpolarization of the cell membrane nor insertion of two electrodes, the response could be recorded for a longer period of time (usually for a few hours) without any signs of deterioration. This method offered the most reliable measurement for comparisons of the EPSPs during the longer stimulus intervals. However, due to the prolonged conductance changes caused by the afterhyperpolarizations of the orthodromically induced action potentials, reliable data could only be collected from intervals greater than 200 msec (which was the approximate duration of the afterhyperpolarization). These data were plotted (Fig. 4b). A clear depression in the Gtr2/Gtrl ratio was observed between 500 msec and 10 sec. The recovery from the depression in normal Ringer solution followed an exponential time course having a time constant of 4.5 sec. When calcium was increased to 9 mM, depression was intensified and recovered exponentially with the same time constant of 4.5 sec.

Comparison of the amplitude of pairs of presynaptic spikes To be sure that at short intervals, especially, the alteration of the second EPSP

55

I.I

'

I

. . . . . .

I

o .01

. . . . . . .

I

. . . . . . . .

oJ

I

. . . . . . . .

I

I

Pulse Interval (sec)

Fig. 5. A comparison of the relative amplitudes of pairs of presynaptic spikes -- S~/S1. The left figure shows representative spikes at an interval of 12 msec; the right figure shows spikes at a 30-msec interval. The graph below depicts the ratio of the amplitude of the second spike (ordinate) as compared to the first at various pulse intervals (abscissa). The points represent the experimental values and the bars their standard error.

was not associated with a change in the amplitude of the second impulse, extracellular spikes were monitored at the various pulse intervals (Fig. 5). Under no circumstances could a second presynaptic spike be initiated at an interval less than 3 msec after the conditioning stimulus. In addition, the duration of the presynaptic spikes remained constant at about 1 msec. This indicated that the action potential of the presynaptic fibers was the result of a synchronous firing of the terminal arborization resulting from a descending orthodromic impulse. This concept has always been assumed, but the fact that the differentiated field potential has only one peak and such a brief duration supports this concept. As seen in Fig. 5, the ratios of the second of these responses compared to the first did not change significantly. Also, the presynaptic action potentials recorded by a sucrose-gap technique do not have an afterhyperpolarization but an afterdepolarization lasting for 200-500 msec 84; this fact suggests that if there were any changes in the second response which had not been detected by the field potential recording, it would be a small reduction in the amplitude of the second spike during the afterdepolarization of the first response. These findings indicate that the facilitation or depression of the second synaptic potential is probably not due to an alteration of the presynaptic spike amplitude.

Comparison of miniature EPSP amplitudes Miniature EPSPs (mEPSPs) were monitored before and after an orthodromic stimulus to determine if the activation of the subsynaptic membrane and the subsequent excitation of the postsynaptic membrane could alter the sensitivity of the subsynaptic membrane to the transmitter released from the presynaptic terminal. Since appearance of mEPSPs was very low in frequencye, high K + (11-12 m M ) solutions were used to observe mEPSPs more frequently. In Fig. 6, mEPSPs were collected at 1-sec intervals beginning 20 sec before and lasting 20 sec after a presynaptic stimulus.

56 MINIATURE

EPSP

AMPLITUDE

my

' mV

T

,-°

I '

I

1.5 I

I.O I

0.5

n II ~

Before

After

udl

20

15

I0

5

0

5

I0

15

20 see

Fig. 6. A comparison of m E P S P amplitudes up to 20 sec before and 20 sec after an orthodromically induced action potential. The m E P S P frequency was so low that a series o f 9 experiments from the same cell were pooled to yield the mean observations. The amplitudes o f the mEPSPs are represented in the graph as their means ± S.E. The figure gives an example o f a typical recording. Calibrations: 2 m V and 1 sec.

The results are depicted in Table I. As can be observed there is no significant difference between the amplitude before and after the stimulus was applied. The results of these experiments support the conclusion that the observed phases of facilitation and depression were due to presynaptic rather than postsynaptic effects.

Theoretical estimation of facilitation and depression An equation, E~/E1 = (1 - - pl exp (-- ht)) (1 + exp (-- bt)) n', was formulated to define facilitation and depression in terms of the probability of

TABLE I

Miniature EPSP amplitudes There was no statistical difference between the amplitude o f the miniature EPSPs before and after an orthodromic action potential at P < 0.10. D a t a expressed as mean ~ S.E. N is number of mEPSPs observed over a 20-see period.

Cell no.

Before

1

1.28 ± 0 . 0 8 ( N = 30)

2 3 4 5

1.48 1.24 1.52 1.35

-¢- 0.07 ± 0,05 4- 0.18 ± 0.06

After

(N (N (N (N

= = = =

80) 133) 21) 78)

1.50 1.51 1.28 2.13 1.53

4- 0.11 :k 0.08 i 0.07 4- 0.32 ~ 0.07

( N = 23)

(IN (N (N (N

= ~ = =

90) 158) 15) 57)

57

E~'E," ( I - p l o x p ( ' h t ) ) ( e + o x p ( - b t ) ) n'

o ix(ca! 1.S

n=l

p,.o.l~

0~'%.

h • 0.22 sec"l

O~,

E2/EI

_". . . .

b -6"30sK"

• s=(c,! ,-,

~'~

---,--; .... -

•

-IT "-~

p =0.33

•

i~,o.22 ,.c"

b -1"80see"

I.O

|

a

i

20 30 50

*

J

100 200

|

|

|

!

a

500

I

2

S

10

Fig. 7. Comparison of the e x ~ r ~ e n t a l data, same as in Fig. 4 with theoretically expected va]ues of the ratios which are derived from E~/E1 = (1 - - pie -ht) (1 + e-bt) n' where n' = 1, Pl = 0.13, h = 0.22 ~ - 1 and b = 6.3 s ~ -1 in normal calcium solution (open circles), and where n' = 1, pl = 0.33,

h = 0.22 see-1, and b = 1.80 sec-~ in high calcium solution (solid squares). See details in text.

ACh release by an initial impulse (pl), the recovery time constant for readily available ACh (l/h), and the recovery time constant for residual calcium in the presynaptic nerve terminal (l/b) (see Methods). By introducing appropriate values for pl, h, b and n'(see Discussion), the equation yielded the curves which fit the E2/Ez values obtained experimentally in normal Ringer solution and high calcium Ringer solution, as illustrated in Fig. 7. It is apparent from this figure that an elevation of [Ca]0 increases the value of 1/b and pl. The recovery time constant of readily available acetylcholine (l/h) is, however, independent of the elevation of [Ca]o. This equation has been expanded (Tashiro, Gallagher and Nishi, unpublished observations) in order to analyze how the various parameters (pl, h, b and n') are affected, and in turn determine the amplitude of each EPSP, during a train of impulses. DISCUSSION

These results at the amphibian sympathetic ganglion have revealed that a conditioning presynaptic impulse induces a period (up to 500 msec) during which a second EPSP will be facilitated. If a long enough period (1-10 see) is allowed to elapse, a second EPSP will be depressed. It has been concluded that these two temporal phases of facilitation or depression are solely presynaptic in nature and due to a change in the output of transmitter. There is much evidence for the relation between transmitter release and calcium: the presence of extracellular calcium is essential for the production of an evoked EPP al and influx of calcium through the axon membrane 1,2a is the first step in the process of transmitter release as quanta from the nerve terminala0. Experiments by

58 01× i~ 5~

1'5~

Pl 013

1'0 0"5

I

h= • 5X

xx

b = 6"30 sec-

I

1'51.0

' 20

. . . . 500 . . 1 5'0 100200 msec

2 5 1'0 sec

Fig. 8. A comparison of alterations in the parameters which comprise the phases of facilitation and depression. The uppermost family of curves demonstrates the effect of altering the prObability of release (pl) while maintaining both the time constant for residual calcium (l/b) and the time constant for ACh mobilization (l/h) constant: The middle family of curves alters "h' while keeping 'b' and 'p£ constant. The lowest family of curves demonstrates the effects of altering 'b' upon the facilitatory phase. Ordinates: ratios of E2/E1. Abscissae: pulse intervals (msec). Horizontal line in each graph indicates that ratio at which E2 - El, i.e., neither facilitation nor depression, is occurring.

Hogdkin and Keynes 23 and Baker et al. I have demonstrated that immediately after an impulse there is a rapid phase of calcium-efflux which is very long in comparison to the time course of calcium-influx. If one assumes that the process of facilitation develops during the long phase of calcium-efflux, then a residue of the 'recruited c a l c i u m ' which entered the nerve terminal and that calcium which was freed from the membrane as a result of a previous impulse, may act as a component in the process of facilitation. Included in this residue o f recruited calcium is that calcium proposed by Katz and Miledi 82 - - residue of 'active' calcium - - which they suggested would remain attached for some time to specific release sites o n the inner surface o f the axon membrane. The residue of the 'recruited calcium' is hereafter referred to as residual calcium. It is our assumption that this residual calcium decays exponentially with a time constant of 1/b and can be included with the recruited calcium that results from a second impulse to produce facilitation. The time course of facilitation and depression that resulted from these experiments on the ganglion was best described as the product of t w o exponential func-

59 tions: (1 - - pie -nt) (1 -q- e-bt) n' (Fig. 7). The former function is a depression factor and the latter function a facilitation factor. This kind of assumption has been proposed by Mallart and Martin4L The best fit of the experimental curve and theoretical curve resulted when 'n" was taken to be one (Fig. 7). As shown in Fig. 8, top, an increase in probability of release shifts the theoretical curve for facilitation to a lower level. This finding suggests that the decrease in facilitation is due to the marked depletion of available transmitter which resulted from a higher probability of release, i.e., facilitation appears virtually to be reduced due to an overriding increase in depression which masks the real increase in facilitation. The other factor which can influence the degree of facilitation, and apparently the only factor which does not affect the degree of depression, is the recovery time constant for residual calcium, 'l/b'. In normal Ringer solution 'l/b' equals 159 msec (Figs. 7, 8). This value is similar to that obtained for one of the time constants, 31 = 200 msec. used to describe the time course of decay of facilitation after a single impulse at the guinea pig superior cervical ganglion 40. Theoretically, as is demonstrated in Fig. 8, bottom, as the value of 'b' is decreased, i.e., residual calcium is available for a longer period of time (possibly due to an increased level of intracellular calcium), the facilitatory phase of the ratios, Ez/E1, is shifted upwards and to the right. Our experimental results in high (5 ×) calcium Ringer solution, however, produce values such that not only does the value for 'b' decrease, but the value for 'pl' increases, thus complicating the problem. The net result is a curve which shows only a slight prolongation of the facilitatory phase and a more pronounced decrease in the degree of facilitation (Fig. 7). This implies that although the residual calcium is increased in a high extracellular calcium medium, its facilitatory action on transmitter release is less effective due to the enhanced depletion of available transmitter by the first impulse as a consequence of a higher probability of release. EPSP or EPP amplitudes of the second response in a pair or train of impulses have been shown to be less than two times the amplitude of the first in many experiments 2,3,17,1s,42,~0. However, in low calcium or high magnesium containing solution, the facilitation ratios were more than two 15,32. In our preliminary experiments with low calcium (1/5-1/10 normal calcium), we observed a facilitation ratio of more than two (unpublished observations). But, we must emphasize that failure of response to the first impulse was frequently observed in spite of a clear response to a second impulse. In this case, the ratio, E2/E1, would be infinitely large. This phenomenon might be explained by assuming that there is a threshold level of intracellular calcium within the presynaptic terminal which is necessary for transmitter release, just as was observed that a threshold for calcium within muscle fibers is necessary for contraction (cf. ref. 16). If evoked transmitter release is not able to occur until intracellular calcium is accumulated up to a threshold level for release, the amplitude of an EPSP resulting from an initial impulse will be much smaller than its expected amplitude, especially in a low calcium solution. On the other hand, a second response should be close to or larger than its expected amplitude due to the presence of residual calcium. The resuiting ratio, E2/E1, will be much larger than expected.

60 Fig. 8, middle, demonstrates that changes in the recovery time constant for readily available A C h (l/h) only affect the phase of depression and not the phase of facilitation as measured by paired pulses. In normal Ringer solution '1/h' equals 4.5 sec (Figs. 7, 8). This value is not too dissimilar from that obtained for r2 - 13.3 sec 40 and used to describe the rate constant o f decay of the second c o m p o n e n t o f facilitation. Furthermore, ' h ' was independent o f increases in extracellular calcium. This latter result confirms the observation at the neuromuscular junctionZ9, ~1,~; where the time course o f depression was independent o f the extracellular calcium concentration. The value o f ' h ' seems dependent u p o n stimulus frequency (unpublished observations) as M a e n o 41 has suggested at the neuromuscular junction where repetitive stimulation altered the rate o f mobilization o f available ACh. In summary, these results demonstrate that during normal ganglionic transmission two processes, i.e., facilitation and depression, occur, which can serve to modulate the output o f transmitter. Earlier reportsa,13,20,37,3s,41,50, 51 have demonstrated that these processes play a significant role in normal transmission at the neuromuscular junction. We have concuded that facilitation and depression o f ganglionic tranmission are simultaneously occurring processes, which are solely the result of normal nerve terminal activity. One or the other process will predominate, depending upon the interval between impulses, as well as the relative concentration o f extracetlular calcium ion. Whether or not the resultant value o f pl is the same as the probability (P) which is given from the q u a n t u m hypothesis (m = nP) 13 remains to be resolved. ACKNOWLEDGEMENT This research was supported by USPHS, N I H Grants NS 06672 and N B 06544.

REFERENCES 1 Baker, P. F., Hodgkin, A. L. and Ridgway, E. B., Depolarization and calcium entry in squid giant axons, J. Physiol. (Lond.), 218 (1971) 709-755. 2 Balnave, R. J. and Gage, P. W., Manganese ions and transmitter release, Proc. ~lust. Physiol. Pharmacol. Soc., 2 (197I) 19. 3 Barrett, E. F. and Stevens, C. F., The kinetics of transmitter release at the frog neuromuscular junction, J. Physiol. (Lond.), 227 (1972) 691-708. 4 Betz, W. J., Depression of transmitter release at the neuromuscular junction of the frog, J. Physiol. (Lond.), 206 (1970) 629-644. 5 Birks, R. and Macintosh, F. C., Acetylcholine metabolism of a sympathetic ganglion, Canad. J. Biochem. Physiol., 39 (1961) 787-827. 6 Blackman, J. G., Ginsborg, B. L. and Ray, C., Spontaneous synaptic activity of sympathetic ganglion cells of the frog, J. Physiol. (Lond.), 167 (1963) 389--401. 7 Blackman, J. G., Ginsborg, B. L. and Ray, C., On the quantal release of the transmitter at a sympathetic synapse, J. Physiol. (Lond.), 167 (1963) 402-414. 8 Bracho, H. and Orkand, R. K., Effect of calcium on excitatory neuromuscular transmission in the crayfish, J. Physiol. (Lond.), 206 (1970) 61-71. 9 Brimble, M. J., Wallis, D. I. and Woodward, B., Facilitation and inhibition of cell groups within the superior cervical ganglion of the rabbit, J. Physiol. (Lond.), 226 (1972) 629-652. 10 Brown, G. L. and Harvey, A. M., Neuromuscular transmission in the extrinsic muscles of the eye, J. Physiol. (Lond.), 99 (1941) 379-399.

61 11 Christensen, B. N. and Martin, A. R., Estimates of probability of transmitter release at the mammalian neuromuscular junction, J. Physiol. (Lond.), 210 (1970) 933-945. 12 Del Castillo, J. and Katz, B., The effect of magnesium on the activity of motor nerve endings, J. Physiol. (Lond.), 124 (1954) 533-559. 13 Del Castillo, J. and Katz, B., Quantal components of the endplate potential, J. Physiol. (Lond.), 124 (1954) 560-573. 14 Del Castillo, J. and Katz, B., Statistical factors involved in neuromuscular facilitation and depression, J. Physiol. (Lond.), 124 (1954) 574-585. 15 Dodge, F. A. and Rahamimoff, P., Co-operative action of calcium ions in transmitter release at the neuromuscular junction, J. Physiol. (Lond.), 193 (1967) 419-432. 16 Ebashi, S. and Endo, M., Calcium ions and muscle contraction. In J. A. V. Butler and D. Noble (Eds.), Progress in Biophysics and Molecular Biology, Pergamon Press, New York, 1968, pp. 123-183. 17 Eccles, J. C., Facilitation and inhibition in the superior cervical ganglion, J. Physiol. (Lond.), 85 (1935) 207-238. 18 Eccles, J. C., Katz, B. and Kuffier, S. W., Nature of the 'end-plate potential' in the curarized muscle, J. Neurophysiol., 4 (1941) 263-287. 19 Elfvin, L. G., The ultrastructure of the superior cervical sympathetic ganglion of the cat. II. The structure of the preganglionic end fibers and the synapse as studied by serial sections, J. Ultrastruct. Res., 8 (1963) 441-476. 20 Elmqvist, D. and Quastel, D. M. J., A quantitative study of end-plate potentials in isolated human muscle, J. Physiol. (Lond.), 178 (1965) 505-529. 21 Feng, J. P., Studies on the neuromuscularjunction. XXVI. The changes of the endplate potential during and after prolonged stimulation, Chin. J. Physiol., 16 (1941) 341-372. 22 Ginsborg, B. L., On the presynaptic acetylcholine receptors in sympathetic ganglia of the frog, J. Physiol. (Lond.), 216 (1971) 237-246. 23 Hodgkin, A. L. and Keynes, R. D., Movements of labelled calcium in squid giant axons, J. Physiol. (Lond.), 138 (1957) 253-281. 24 Hubbard, J. I., Repetitive stimulation at the mammalian neuromuscularjunction and the mobilization of transmitter, J. Physiol. (Lond.), 169 (1963) 641-662. 25 Hubbard, J. I., Jones, S. F. and Landau, E. C., On the mechanism by which calcium and magnesium affect the release of transmitter by nerve impulses, J. Physiol. (Lond.), 196 (1968) 75-86. 26 Huber, B. C., A contribution on the minute anatomy of the sympathetic ganglia of the different classes of vertebrates, J. Morph., 17 (1899) 27-90. 27 Hunt, C. C. and Nelson, P. G., Structural and functional changes in the frog sympathetic ganglion following cutting of the presynaptic fibers, J. Physiol. (Lond.), 117 (1965) 1-20. 28 Jenkinson, D. H., The nature of the antagonism between calcium and magnesium ions at the neuromuscular junction, J. Physiol. (Lond.), 138 (1957) 434-444. 29 Katz, B. and Miledi, R., The effect of calcium on acetylcholine release from motor nerve terminals, Proc. roy. Soc. B, 161 (1965) 496--503. 30 Katz, B. and Miledi, R., The release of acetylcholine from nerve endings by graded electrical pulses, Proc. roy. Soc. B, 167 (1967) 23-28. 31 Katz, B. and Miledi, R., The timing of calcium action during neuromuscular transmission, J. Physiol, (Lond.), 189 (1967) 535-544. 32 Katz, B. and Miledi, R., The role of calcium in neuromuscular facilitation, J. PhysioL (Lond.), 195 (1968) 481-492. 33 Katz, B. and Miledi, R., Further study of the role of calcium in synaptic transmission, J. Physiol. (Lond.), 207 (1970) 789-801. 34 Koketsu, K. and Nishi, S., Cholinergic receptors at sympathetic preganglionic nerve terminals, J. Physiol. (Lond.), 196 (1968) 293-310. 35 Kuba, K. and Nishi, S., Membrane current associated with the fast e.p.s.p, of frog sympathetic neurons, Physiologist, 14 (1971) 176. 36 Liley, A. W. and North, K. A. K., An electrical investigation of effects of repetitive stimulation on mammalian neuromuscular junction, J. NeurophysioL, 16 (1953) 509-527. 37 Lipicky, R. J., Hertz, L. and Shanes, A. M., Ca45 transfer and acetylcholine release in the rabbit superior cervical ganglion, J. cell. comp. Physiol., 62 (1963) 233-242. 38 Lundberg, A. and Quiliscb, H., Presynaptic potentiation and depression of neuromuscular transmission in frog and rat, Acta physiol, scand., 30, Suppl. I!I (1953) 111-120.

62 39 Lundberg, A. and Quilisch, H., On the effect of calcium on presynaptic potentiation and depression at the neuromuscular junction, Acta physiol, scand., 30 Suppl. III (1953) 121-129. 40 McLachlan, E. M., An analysis of the release of acetylcholine from preganglionic nerve terminals, J. Physiol. (Lond.), 245 (1975) 447-466. 41 Maeno, T., Analysis of mobilization processes in neuromuscular transmission in frog, J. NeurophysioL, 32 (1969) 783-800. 42 Mallart, A. and Martin, A. R., The relation between quantum content and facilitation at the neuromuscular junction of the frog, J. Physiol. (Lond.), 196 (1968) 593-604. 43 Martin, A. R., A further study of the statistical composition of the endplate potential, J. Physiol. (Lond.), 130 (1955) 114-122. 44 Martin, A. R. and Pilar, G., Dual mode of synaptic transmission in the avian ciliary ganglion, J. PhysioL (Lond.), 168 (1963) 443-463. 45 Moore, J. W. and Cole, K. S., Voltage clamp techniques. In W. L. Nastuck (Ed.), Physical Techniques in Biological Research, Vol. 6, Academic Press, New York, 1963, pp. 263-321. 46 Nishi, S. and Koketsu, K., Electrical properties and activities of single sympathetic neurons of frogs, J. cell. comp. Physiol., 55 (1960) 15-30. 47 Nishi, S., Soeda, H. and Koketsu, K., Studies on sympathetic B and C neurons and patterns of preganglionic innervation, J. cell. comp. PhysioL, 66 (1965) 19-32. 48 Ortiz, C. L. and Bracho, H., Effects of reduced calcium on excitatory transmitter release at the crayfish neuromuscular junction, Comp. Biochem. Physiol., 41A (1972) 805-815. 49 Otsuka, M., Endo, M. and Nonomura, Y., Presynaptic nature of neuromuscular depression, Jap. J. Physiol., 12 (1962) 573-584. 50 Rahamimoff, R., A dual effect of calcium ions on neuromuscular facilitation, J. Physiol. (Lond.), 195 (1968) 471-480. 51 Takeuchi, A., The long-lasting depression in neuromuscular transmission of frog, Jap. J. Physiol., 8 (1958) 102-113. 52 Thies, R. E., Neuromuscular depression and apparent depletion of transmitter in mammalian muscle, J. Neurophysiol., 28 (1965)427-442. 53 Uchizono, K., On different types of synaptic vesicles in the sympathetic ganglia of amphibia, Jap. J. Physiol., 14 (1964) 210-219.