JOURNALOFNEUROPHYSIOLOGY Vol. 68, No. 4, October 1992. Printed

RAPID

in U.S.A.

PUBLICATION

Quanta1 Synaptic Transmission in Phrenic Motor Nucleus GUOSONG LIU AND JACK L. FELDMAN Systems Neurobiology Laboratory, Department of Physiological University of California, Los Angeles, California 90024-l 527 SUMMARY

AND

CONCLUSIONS

1. The quanta1 nature of excitatory synaptic transmission was studied in respiratory interneurons and phrenic motoneurons of intact neonatal rat brain stem-spinal cord preparations in vitro. Synaptic currents were recorded with whole-cell patch-clamp recording techniques. 2. Because the most important factor for quanta1 detection is the ratio of quanta1 size to quanta1 standard deviation, factors that influence this ratio were evaluated so that experimental techniques that enhance this ratio could be defined. 3. Under favorable conditions, we directly observed quanta1 amplitude fluctuations In spontaneous excitatory postsynaptic currents ( EPSCs) in spinal cord respiratory neurons. The quanta1 conductance size was 55-- 100 pS. With fast decay of these EPSCs, the charge reaching the soma for a single quantum is only - 15 fC (V = -80 mV). 2. We also studied miniature EPSC amplitude distributions. These were skewed, as previously reported; however, distinct quanta1 intervals were observed. Furthermore, in three cells tested, the quanta1 size in the miniature EPSC amplitude distribution was similar to the quanta1 size in the spontaneous EPSC amplitude distribution. 5. We conclude that excitatory synaptic transmission in the mammalian spinal cord is quanta1 and that the apparent skewness of miniature EPSC distributions results from summation of events with multiple quanta1 peak amplitudes.

INTRODUCTION

Chemical synaptic transmission is the principal means of neuronal communication in the CNS. A dominant hypothesis is that the synaptic transmission process is quantal. A powerful tool for understanding pre- and postsynaptic mechanisms at the neuromuscular junction has been quantal analysis ( Martin 1977 ) . However, the quanta1 nature of excitatory synaptic transmission in mammalian CNS has been difficult to demonstrate (Korn and Faber 199 1; Redman 1990). This requires demonstration of successive peaks of biological origin in histograms of peak amplitudes of synaptic events; ideally, these peaks will be evenly spaced. Previous reports (Jack et al. 198 1; Larkman et al. 199 1) that conclude excitatory synaptic transmission is quanta1 have been limited by unfavorable signal-to-noise ratios that preclude direct resolution of (small) quanta1 peaks (Clamann et al. 199 1; Clements 199 1; Kullmann 1989; Redman 1990). Under more favorable conditions (using whole-cell patch-clamp recording), quanta1 peaks have not been observed (Bekkers et al. 1990; Bekkers and Stevens 1990; Forsythe and Clements 1990; Malinow and Tsien 1990, but see Kullmann and Nicoll 1992). The matter is further complicated by recent reports of skewed am1468

Science,

plitude distributions of miniature EPSCs (mEPSCs) (Bekkers and Stevens 1990; Manabe et al. 1992; Silver et al. 1992), which question the validity of assumptions underlying application of quanta1 analysis. In the present study, we applied whole-cell patch-clamp techniques to record membrane currents in respiratory neurons in the neonatal rat spinal cord. Under favorable experimental conditions, quanta1 amplitude fluctuations of peak synaptic currents were observed. Preliminary results were published in abstract form (Liu and Feldman 1990, 199 1). METHODS

Whole-cell patch-clamp recordings were made from neurons within the phrenic motoneuron pool (~1 = 25) of neonatal rat brain stem-spinal cord preparations in vitro (Liu and Feldman 1990). The preparation was submerged in an in vitro chamber and superfused continuously with a modified Kreb’s solution [(in mM) 120 NaCl, 3 KCl, 1 CaCl,, 2 MgSO,, 26 NaHCO,, 1.25 NaH,PO, and 20 D-glucose], measured - 300 mOsM, and was equilibrated with 95% O,-5% CO, at temperature 24-26OC. Patch electrodes were pulled from aluminosilicate glass. (Resistance 3-4 MQ for potassium methanesulfonate ( KMeSO,) solution and tip size -2 pm). The pipette solution contained (in mM) 120 KMeSO, , 1 CaCl, , 5 NaCl, 10 IV-2-hydroxyethylpiperazineIV’-2ethanesulfonic acid (HEPES), 2 ATP (magnesium salt), 10 bis(o-aminophenoxy)-N,N,N’,iV’-tetraacetic acid (BAPTA) (tetrapotassium salt), pH 7.3, adjusted by KOH. The junction potentials for different intracellular solutions were corrected for all experimental records. Synaptic currents were low-pass filtered (2or 5-kHz Bessel filter), recorded, and digitized at 20 kHz onto a disk for computer-aided analysis. A maximum likelihood estimator (MLE) algorithm was used to determine location and variance of quanta1 peaks in each data set ( McLachlan and Basford 1988). The only assumption used for the MLE procedure was that each histogram was the sum of a few discrete Gaussian amplitude distributions, not necessarily equally spaced. When applied to a distribution of simulated synaptic events that were not evenly separated in peak amplitude, the MLE resolved correct peak locations. The likelihood values for MLE solutions served as indicators of goodness of fit. Our results from analyzing simulated quanta synaptic events suggest that, in general, the best fit is a good fit; when the best fit was obtained, the parameters resulting from the MLE procedure matched the parameters used to generate the simulated data sets (unpublished observation). RESULTS

Brain stem-spinal cord preparations retain functional respiratory motor circuitry. Inspiratory interneurons and phrenic motoneurons receive inputs that generate two

0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on January 11, 2019.

QUANTAL

ANALYSIS

OF

EXCITATORY

‘,‘I,,‘,I ,,“‘,“,’, ,, ,I,‘,‘II,’, Ill

-

expiration -

FIG. 1. A : representative whole-cell patch-clamp recording from neonatal rat spinal cord neuron in phrenic motoneuron pool with -500 pA inward drive current during inspiratory phase and many spontaneous single excitatory postsynaptic currents ( EPSCs) during expiratory phase (bottom). Simultaneously recorded C4 ventral root activity (top). B: EPSCs selected from expiratory period for quanta1 analysis. Trace is average of 59 events in 1 data set. This EPSC had a lo-90% rise time of 0.25 ms and decay time constant of 1.3 ms.

types of spontaneous EPSCs. One type, of large amplitude, occurs during the inspiratory period; a second type occurs infrequently during the expiratory period (Fig. 1). These EPSCs are generated by release of an excitatory amino acidlike (EAA) transmitter (Liu et al. 1990). The temporal isolation of EPSCs during expiration made them most suitable for quanta1 analysis. Optimal

conditions for detection ofsmall

quanta events

To establish optimal conditions for quanta1 detection, we analyzed factors that determine the shape of quanta1 amplitude distributions using simulated synaptic currents. Our results, consistent with previous investigations (Clamann et al. 199 1; Kullmann 1989; Redman 1990), indicate that detection of quanta1 peaks is affected by the ratio quanta1

quanta1 standard

size (4) deviation

(&,bs)

where o$,~ = quanta1 variance ( 0:) + noise variance ( ~2,). We found that unless q/o& > 2.2, direct resolution of quanta1 intervals, i.e., distinct peaks in the peak EPSC amplitude histogram, are unlikely (unpublished observation). For indirect detection with the MLE approach, q/gobs must be larger than 1.5. If q/ uobs< 1.5, the MLE may generate an inaccurate estimate of the quanta1 size or even incorrectly suggest that quanta1 peaks are present (Clamann et al. 199 1; Clements 199 1; Redman 1990). Thus, favorable experimental conditions for detection of small quanta1 intervals maximize q/ gObsby minimizing cn without affecting q.

SYNAPTIC

CURRENTS

1469

We red uced on as follows: I>C urrents were recorded bY low noise who le-cell patch-cl .amp . 2) Background than nel noise was minimized by selecting data from small neurons (as determined by membrane capacitance). 3) Data were filtered using an “optimal filtering” technique (Barkat 199 1; G. Liu and J. L. Feldman, unpublished data), which markedly reduced background noise with minimal effects on EPSC peak amplitude and time course. Further enhancement of q/ cobsrequires that attenuation of EPSC peak size due to imperfect voltage clamp be avoided. A fast-step response of the voltage-clamp system and a good space clamp is obligatory. Thus, we selected data from neurons with low membrane capacitance ( 15-20 pF) to minimize effects of series resistance and only analyzed recordings for which >95% of EPSCs bad rise times < 1 ms. We assumed these events were of (near to) somatic origin. Although recording EPSPs under current-clamp conditions can avoid problems of series resistance, our simulation studies indicate that the resolution of the recording system for quanta1 detection under current-clamp conditions is lower than that under voltage-clamp conditions (unpublished observation). Thus, only EPSC size was used for quanta1 analysis. Quanta1 nature ofexcitatory

synaptic transmission

A total of 25,000 EPSCs (from 10 cells) that satisfied the above criteria was selected for further quanta1 analysis. Rise time, decay time, half-amplitude width, peak amplitude, and total charge transfer for synaptic currents were measured. Each data set consists of l,OOO-5,000 consecutive EPSCs with mean rise time ( lo-90%) of 0.25-0.7 ms. The noise standard deviation gn after filtering (using an optimal filter) was 0.8-1.3 pA. Peak amplitude histograms of these EPSCs exhibited clear and regularly spaced peaks. Figure 2A shows EPSC amplitudes in one cell that ranged from 5 to 50 pA with multiple peaks at - 4-pA intervals ( -80 mV holding potential). Similar results were obtained from nine additional cells. The quanta1 conductance intervals were 55- 100 pS ( n = 10). The total charge transfer for a single quantum at -80 mV holding potential was only - 15 fC; this is a consequence of small quanta1 conductance and fast EPSC decay (decay time constant 1.2- 1.9 ms). We considered the possibility that peaks in amplitude histograms resulted from sampling error or contamination from recording noise (Clamann et al. 199 1; Clements 199 1; Redman 1990). On the basis of the following evidence, we conclude that they were not artifacts. ( 1) Data obtained at two different holding potentials (in 3 of 3 neurons tested) gave similar values for the quanta1 conductance intervals. For the neuron in Fig. 2, quanta1 current I, = 4.3 t 0.3 (SD) pA and oobs= 1.6 t 0.3 at -80 mV; Iq = 3.2 t 0.5 pA and oObs= 1.4 t 0.2 pA at -60 mV (Fig. 2, A and B, respectively). The ratio of quanta1 currents Iq,,/~qso was 0.74, a value very close to the ratio of drive potentials ( veO- vrc,)/ (J&- Vrev) - 0.74 ( vrevcalculated from intra- and extracellular solutions was -4 mV), indicating that calculated quanta1 current Iq was a linear function of holding potential with similar quanta1 conductance ( - 55 pS) . (2) Similar quanta1 peak locations remained when a whole data set (n

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on January 11, 2019.

G. LIU

AND

J. L. FELDMAN

= 3) was divided into two subsets (e.g., for the neuron in Fig. 2: 2,053 events; subsets: 1,249 and 8 14 events), suggesting that these quanta1 peaks were not an artifact due to small sampling size (Edwards et al. 1990). (3) For all data sets ( YI = 10) : i) the estimated peak locations were independent of bin size in the range 0.2-0.5 pA. ii) The largest likelihood value for solutions obtained by the MLE procedure was always achieved when peak locations from the MLE procedure coincided with peak locations determined by visual inspection. mEP.CC amplitude

Control

distribution 320 rnEPSC

Further evidence of quanta1 transmission is the close relationship between the quanta1 size of spontaneous EPSCs and of mEPSCs present after the application of 1 PM tetro-

480 amplitude

After

640 (PSI

TTX

* 50 v, = -80 mV

3 40 !

160

320 mEPsc

12

24 EPSC

B

T

60 CA z 44

36 amplitude

48

60

(PA)

v, =-60

mV

= 36 5 2 z= 24

12

24 EPSC

36 48 amplitude (PA)

60

FIG. 2. A : amplitude histogram of excitatory postsynaptic current (EPSC) peaks from 2,053 consecutive spontaneous synaptic events (vh = - 80 mV ) . Thick solid lines represent the individual Gaussian distributions determined by the maximum likelihood estimator (MLE) procedure ( Iq = 4.3 -t 0.3 pA and Gobs = 1.6 -t 0.3 pA, bin = 0.4 PA). The SD of baseline noise CT,was 1 .O pA. B: amplitude distribution from same neuron as in A but at -60 mV holding potential (Iq = 3.2 + 0.5 pA, q,bs = 1.4 -t 0.2 pA, and on = 0.94 pA; from 2,578 consecutive events). Note that peaks at -4-pA intervals are clearly visible in A, where mean q/a,b, = 2.6. As predicted (see text), this is not the case in B. This is because the reduction in Iq to -3 pA at -60 mV holding potential (as determined by MLE) reduces mean q/(~,b~ = 2.2. For this ratio, the peak separations should barely be visible if at all ( see text), but the MLE procedure is still valid kkb.s

>

l-5).

480 amplitude

640 (pS)

800

FIG. 3. Effect of tetrodotoxin (TTX) on peak excitatory postsynaptic current (EPSC) amplitude distribution. A : distribution of spontaneous EPSC amplitudes (first peak at 128 pS, yq = 94 pS, n = 446). B: distribution of miniature EPSC amplitudes after application of 1 PM TTX, collected over 1 h (first peak at 102 pS, yq = 80 pS, n = 348 ) . The fits to the last data peak were not plotted for both A and B. Note the similarity of peak locations, especially for first 3 peaks.

dotoxin (TTX) (n = 3). In the example of Fig. 3, the amplitude distribution was quanta1 before application of TTX (Fig. 3A; y4 = 94 pS). After TTX, the frequency and amplitude of EPSCs were reduced (the blockage of action potential generation by TTX was indicated by the inability to evoke fast Na+ currents with current injection); this suggests that most spontaneous EPSCs were action potentialdependent. The amplitude distribution was skewed with a range of peak amplitudes similar to previous studies of hippocampal neurons (Bekkers and Stevens 1990; Manabe et al. 1992) and cerebellar granule cells (Silver et al. 1992). However, in contrast to previous results, individual quanta1 peaks with intervals close to control (pre-TTX) values remained (Fig. 3B; y4 = 80 pS). Therefore, the apparent skewness of the mEPSC distribution results from the summation of events with multiple quanta1 peak amplitudes. DISCUSSION

Our results have several implications for studying the quanta1 nature of synaptic transmission in mammalian CNS. First, since a high 4 to (T,bsratio can avoid ambiguities associated with use of deconvolution techniques (e.g., MLE), it is important to achieve sufficiently high 4 to gobs ratio (>2.2) to reliably detect quanta1 peaks. Given a small

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on January 11, 2019.

QUANTAL

ANALYSIS

OF EXCITATORY

quanta1 EPSC size ( ~4 pA at vh = -80 mV), this ratio cannot be achieved without an ultralow noise, high-quality voltage-clamp recording, and good filteri ng techniques. Second, EAA-mediated synaptic transmission in the mammalian spinal cord is quantal. This quanta1 nature of synaptic transmission can facilitate studies of mechanisms of preand postsynaptic modulation. Last, because the range and shape of the mEPSC distributions were similar to those observed when single boutons are stimulated (Bekkers and Stevens 1990), our results suggest that single boutons can release more than one quantum. We are grateful to colleagues who commented on earlier drafts of this manuscript, especially C. F. Stevens, J. C. Smith, and A. Lev-Tov. We acknowledge the generous support of the National Institute of Neurological Disorders and Stroke Grant NS-24742. Address for reprint requests: G. Liu, Systems Neurobiology Laboratory, Dept. of Physiological Science, Univ. of California, 405 Hilgard Ave., Los Angeles, CA 90024- 1527. Received 8 May 1992; accepted in final form 2 1 July 1992. REFERENCES BARKAT, M. Signal Detection and Estimation. Boston: Artech House, 199 1, pp. 2 18-220. BEKKERS, J.M., RICHERSON,G. B., ANDSTEVENS, C.F.Originofvariability in quanta1 size in cultured hippocampal neurons and hippocampal slices. Proc. Natl. Acad. Sci. USA 87: 5359-5362, 1990. BEKKERS, J. M. AND STEVENS, C. F. Presynaptic mechanism for long-term potentiation in the hippocampus. Nature Land. 346: 724-729, 1990. CLAMANN, H.P., RIOULT-PEDOTTI, M.S., ANDLUSCHER, H.R.Theinfluence of noise on quanta1 EPSP size obtained by deconvolution in spinal motoneurons in the cat. J. Neurophysiol. 65: 67-75, 199 1. CLEMENTS, J. D. Quanta1 synaptic transmission? Nature Land. 353: 396, 1991. DEFELICE, L. J. AND CLAY, J. R. Membrane current and potential from single-channel kinetics. In: Single Channel Recording, edited by B. Sakmann and E. Neher. New Y .ork: Plenum, 1983, P- 323-342. EDWARDS, F. A ., K~NNERTH, A., ANDSAKMANN, B. Quanta1 analysis of

SYNAPTIC

CURRENTS

1471

inhibitory synaptic transmission in the dentate gyrus of rat hippocampal slices: a patch-clamp study. J. Physiol. Lond. 430: 2 13-249, 1990. FORSYTHE, I. D. AND CLEMENTS, J. D. Presynaptic glutamate receptors depress excitatory monosynaptic transmission between mouse hippocampal neurons. J. Physiol. Land. 429: I- 16, 1990. JACK, J. J. B., REDMAN, S. J., AND WANG, K. The components of synaptic potentials evoked in cat spinal motoneurones by impulses in single group Ia afferents. J. Physiol. Lond. 32 1: 65-96, 198 1. KORN, H. AND FABER, D. S. Quanta1 analysis and synaptic efficacy in the CNS. Trends Neurosci. 14: 439-445, 199 1. KULLMANN, D. M. Applications of the expectation-maximization algorithm to quanta1 analysis of postsynaptic potentials. J. Neurosci. Methods 30: 23 l-245, 1989. KULLMANN, D. M. AND NICOLL, R. A. Long-term potentiation is associated with increases in quanta1 content and quanta1 amplitude. Nature Land. 357: 240-244, 1992. LARKMAN, A., STRATFORD, K., AND JACK, J. J. B. Quanta1 analysis of excitatory synaptic action and depression in hippocampal slices. Nature Lond. 350: 344-347, 199 1. LIU, G. AND FELDMAN, J. L. Quanta1 analysis of endogenous excitatory post-synaptic currents. Sot. Neurosci Abstr. 17: 1324, 199 1. LIU, G. AND FELDMAN, J. L. Whole cell patch-clamp recording of endogenous synaptic currents in mammalian motoneurons in intact brainstem-spinal cord. Sot. Neurosci Abstr. 16: 1184, 1990. LIU, G., FELDMAN, J. L., AND SMITH, J. C. Excitatory amino acid-mediated transmission of inspiratory drive to phrenic motoneurons. J. Neurophysiol. 64: 423-436, 1990. MALINOW, R. AND TSIEN, R. W. Presynaptic enhancement shown by whole-cell recordings of long-term potentiation in hippocampal slices. Nature Lond. 346: 177- 180, 1990. MANABE, T., RENNER, P., AND NICOLL, R. A. Postsynaptic contribution to long-term potentiation revealed by the analysis of miniature synaptic currents. Nature Land. 355: 50-55, 1992. MARTIN, A. R. Junctional transmission. II. Presynaptic mechanisms. In: Handbook of Physiology. The Nervous System. Bethesda, MD: Am. Physiol. Sot., 1977, sect. 1, vol. I, part 1, chapt. 10, p. 329-355. MCLACHLAN, G. J. AND BASFORD, K. E. Mixture Models: Inference and Applications to Clustering. New York: Dekker, 1988. REDMAN, S. J. Quanta1 analysis of synaptic potentials in neurons of the central nervous system. Physiol. Rev. 70: 165- 198, 1990. SILVER, R. A., TRAYNELIS, S. F., AND CULL-CANDY, S. G. Rapid-timecourse miniature and evoked excitatory currents at cerebellar-synapses in situ. Nature Land. 355: 163-166, 1992.

Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (133.006.082.173) on January 11, 2019.

Quantal synaptic transmission in phrenic motor nucleus.

1. The quantal nature of excitatory synaptic transmission was studied in respiratory interneurons and phrenic motoneurons of intact neonatal rat brain...
969KB Sizes 0 Downloads 0 Views