Instantaneous inward rectification in the mauthner cell: A postsynaptic booster for excitatory inputs

INSTANTANEOUS INWARD RECTIFICATION IN THE MAUTHNER CELL: A POSTSYNAPTIC BOOSTER FOR EXCITATORY INPUTS D. S. FABER and H. KORN* Division of Neurobiology. Department of Physiology, SUNY at Buffalo. BulTalo. NY 142II. C.S.A. and *Laboratoire de Neurobiologie Cellulaire, INSERM UXI. Dept de ~i(~techn~l~~~jes.fwtitut P;rshx~r. 75713 Paris. France

Abstract-Single and double electrode voltage clamp techniques have been used to analyse the leakage conductance of the goldfish Mauthner cell. The results indicate that the input conductance of this neuron is maximal at the resting potential, having an average value of 4.68 /is. Approximately SW;, of thib conductance is voltage dependent and could be inactivated by large hyp~rp[~larlzin& command pulse< (30-50 mV magnitude) of 23-35 ms duration. The magnitude and apparent time constant 01‘ the inactivation are both functions of membrane potential, such that the relaxation of the in\lard currt’nt increased and occurred more rapidly for greater hyperpolarizdtions. The presence of instantaneous inward tail currents when the imposed voltage steps were terminated and membrane potential was returned to the resting level indicated that the conductance mechanism inactivated during hyperpolari7ation had generated a small outward current at the resting level. Thus. we propose that a major fraction of the Mauthner cell’s input conductance is a voltage-dependent K + conductance. Hyperpolarizing inactivatlon is a feature of an inward K+ rectifier in other cell types. and when the recording microelectrodc was located in the Mauthner cell lateral dendrite, it was possible to demonstrate characteristics consistent with rectification. namely a low conductance for outward depolarizing currents and a high conductance for inward currents. The inward rectifier of the Mauthn~r cell is different from that of other neurons in that it IS alrcadl; maximally activated at the resting potential and therefore is a major determinant of that parameter. In addition. Its activation and inactivation kinetics are quite fast. On the basis of evidence that the m%ard rectifier is found in the lateral dendrite and that the transition to the low conductance state occurs with about I@-20 mV depolarization, we postulate that this membrane property favors non-linear processing of excitatory postsynaptic potentials such that large excitatory postsynaptic potentials and those c~okcti in the presence of background dendritic depolarization are preferentially conducted to the soma.

Neuronai responsiveness to synaptic input is in part regulated by both the passive and active electrical properties of the soma-dendritic membranes, which may vary significantly from one cell to the next. For example. while in some instances these cellular regions are electrogenic and exhibit a number of different voltage-sensitive conductances, they are instead inexcitable in others. The teleost Mauthner (M-) cell is a medullary command neuron considered to be representative of the latter category and having membrane characterjstics seemingly well-matched with its function. That is, it has a high resting membrane potential. in the vicinity of -80 mV. and a quite low input resistance, ranging from I1 j to 200 kQ4.‘.Ii and the soma and dendrites are apparently passive.‘.‘.” all characteristics which would

Address correspondence to: Donald S. Faber. Division of Neurobiology, State University of New York. 313 Gary Hall. Buffalo. NY 14214, U.S.A. ~~~r~,~i~f~(~~.~:EPSP, excitatory postsynaptic potentiai; M-cell, Mauthner cell: SEVC and TEVC. single and two electrode voltage clamp. respectively.

contribute to a high threshold. Although the dominant ionic conductance underlying the high resting potential and low input resistance has not been identified directly, it is reasonable to assume that one or more K+-dependent mechanisms would be involved. We have carried out voltage clamp experiments to see whether this is the case and. furthermore, have examined whether a major component 01‘ the resting conductance has features consistent with those of the Inward or anomalous rcctilier studied extensively in many excitable and inexcitahlc cells. including skeletal” and hear? “’ muscle fibers. glia.“’ starfish eggs” and neurons.“i.2h

Experiments were performed on goldfish anesthetired with Flaxedil OI with tricainc and immobilized o-tubocurarine c1mg per g body wt). The preparatmn and basic electrophysiologicul techniques were similar to those described before.” The M-cell was identified on the basis ot its stereotyped response to antidromic stimulation ol’ its axon in the spinal cord.” and either the soma or laterai dendrite \+as penetrated with mIcroelectrodes filled with

A1

SEVC,3

I(nA)

1 KHz

--100

> E - -200 5

mSec

Fig. I. Current-voltage realtions of the Mauthner cell, obtained with single electrode voltage clamp (SEVC) techniques, and demonstrating hyperpolarizing inactivation of the instantaneous leakage current. (Al and A2) Averaged voltage clamp currents and corresponding current-voltage plot, respectively, from an experiment with a Cl--loaded cell. (Al) Family of currents (above) evoked by 20 ms command pulses (below) in the depolarizing (_+5, IS mV) and hyperpolarizing (- 5, - 15, -25, -35, -45, -55 mV) directions. Note that with the larger hyperpolarizations, there was a rapid decline in the inward currents from their initial peaks, and that there was a slow increase in the outward current with larger depolarizations, presumably due to the onset of delayed rectification at the axon hillock. (A2) Plots of the voltage clamp currents (ordinate I) as a function of membrane potential (abscissa, Vm). The peak currents (filled circles) were linearly related to membrane potential (input conductance = 4.5 PCS)while the currents measured after 20 ms (open circles) did not increase in direct proportion to the voltage steps. Note that corresponding symbols in (Al) designate times at which the currents were measured. RMP = resting membrane potentials. In this and subsequent figures, the indicated frequencies refer to SEVC chopping rates.

KC1 (3 M) or KAc (4 M). Both single (S) and two electrode voltage clamp (TEVC) techniques were used (Axociamp, Axon Instruments), with chopping rates being in the range of 16-33 kHz for the one electrode experiments. Electrode resistances were low, generally 2-5 MQ and, with appropriate shielding procedures, voltage control was 95% complete within 0.5 ms after the onset of a command pulse in the SEVC mode. This short response time was a consequence of the low resistance electrodes used and the brief M-cell membrane time constant (~400 ps, see Refs 7 and I I). In all cases, the M-cell was clamped at its resting potential, at which no steady clamp current was required, and its current-voltage relations were obtained during brief (2&35 ms) de- and hyperpolarizing voltage steps. Generally, somatic depolarizations more than 15-50 mV from the resting potential were not used, as they evoked a large allor-none inward current followed by an inhibitory postsynaptic response. Presumably, due to space clamp limitations, the larger pulses triggered an orthodromic action potential at the M-cell’s axon hillock and initial segment,‘O and a consequent activation of the recurrent collateral inhibitory network.‘.‘3.22 RESULTS

As shown in Fig. IA1 and A2 the M-cell’s instantaneous current-voltage relation obtained with a somatic electrode is quite linear in the voltage range studied. The input conductance of this Cl--loaded cell, measured at pulse onset, was 4.5 PS and in 24

experiments with KC1 electrodes this parameter averaged 4.68 + 2.1 PS (mean + SD), which is quite close to the previous estimate of 6.1 PS obtained with current clamp measurements.4 The records in Fig. lAl, however, also illustrate voltage- and timedependent changes in the “leakage” currents.5 In particular, during the larger hyperpolarizations (> - 15 mV) there was a pronounced and fairly rapid sag in the inward current. This partial inactivation developed more rapidly and was more marked during the larger hyperpolarizations, such that at the end of a 20 ms voltage step of - 55 mV, the remaining inward current was about 55% of the peak value (A2), and the apparent input conductance was reduced accordingly. It should be noted that during depolarization there was also a slight increase in the outward current, which is most likely due to the onset of delayed rectification at the axon hillock and initial segment and will not be analysed in more detail here. The observations in Fig. 1Al and A2, which were obtained consistently with both single and double electrode voltage clamp analyses, suggest that a major component of the M-cells’ input conductance is voltage-dependent, being inactivated by hyperpolarization, It, however, remained possible that this

M-cell instantaneous

A1

inward

rectilier

A2

KAc SEVC,

IhA)

19KHz a

-E

--.-__-d -I-

0 v

_--.------~--,-r_..__..

,.-_5---yA--

vm -110

-

I

I

-90 I

I

I

I

mV

1Omsec

l

/

-1

Fig. 2. Hyperpolarizing Inactivation of M-cell leakage currents is not dependent on Cl -loading of the M-cell. (Al and A3) Averaged SEVC (19 kHz) responses and corresponding I-V plot. respectively. obtained with a KAc electrode. (Al) Family of currents (upper) evoked by 27 ms command pulses (below). of amplitudes i5. ~ 15. -25 and -35 mV, from a resting potential (RMP) of -82 mV. Hyperpolarizing inactivation was apparent at - I5 mV and became faster and more pronounced with the larger steps. (A2) Plot of membrane currents (ordinate 1) as a function of membrane potential (abscissa, Vm). for a potenttal range of --62 to ~ I I7 mV. The peak currents (filled circles) were directly proportional to the voltage steps. yielding an instantaneous input conductance of 7.25 pS, while at the end of the pulse (open circles) the incremental increases in current were progressively smaller as the membrane hypcrpolarlration increased.

property was influenced by. or a consequence of. the recordrng conditions. namely a Cl -loading which shifted the Cl equilibrium potential (E,,) in the depolarizing direction. Two observations which rule out this possibility are that the voltage dependence of the inward currents was essentially the same (I) for 30 experiments in which E,, varied from + IO to +4X nrV. with respect to the resting potential, includrng variations within individual experiments, and (2) when KAc electrodes were used (n = 5). Figure 2 Al. A2 illustrate the voltage clamp currents and the current-voltage Again.

plot

the instantaneous

USC for a range

obtained

in the latter

relation

was

situation.

linear,

in this

of membrane potentials from -62 of the to - 122 mV. and there was a relaxation inward currents evoked by hyperpolarization. This rnactiv ation was minimal during small potential shifts. and it increased progressively as larger command pulses were applied. Presumably, the degree of mactivation and nonlinearity was actually underestimated rn these experiments, since the measuremcnts were obtained from a distributed cable. whrch has a linearizing effect on current-voltage relationships.‘” It should also be stressed that the observations reported here are not artefactual consequerrces of charging or discharging the distributed capacrtance of the somadendritic membrane. since the n,nlinearities were not seen for all command

pulses and the sag in the inward currents was not associated with an overshoot at pulse offset. i-c. wrth a fast outward tail current. Theoretically, the sag in the Inward currents observed during membrane hyperpolarizatron could bc due to an inactivation process or to the activation of a counteracting outward current. The latter is highly unlikely since at the larger membrane potentials (e.g. - 100 to - 130 mV) all relevant ionic equilibrrum potentials would be in the depolarizing direction and any conductance increase mechanism would then bc expected to rather generate an additional Inward current. Thus, the hyperpolarizations decrease membrane conductance to one or more ions. Inspection ol the tail currents remaining at the offset of the command pulses suggests more spccitically that .r K conductance is being inactivated. The argument le::ding to this conclusion is based on data such as 15 shown in Fig. 3Al and A2. In this example. whrch I\ characteristic of the results obtained In IO of I6 experiments where tail currents were analysed. there was an inward going current when membrane potential was returned to the resting level. and Its magnitude was proportional to the degree of the prcceding inactivation. In the other six cases. no significant tail current was detected. Since E, , and the sodium and calcium equilibrium potentials (E,,$ and E,,) are all positive with respect to the resting poten-

II S

I040

A1

I.ABEH

SEVC,29KHz

_ Torn” ~_________________________J_ -80

-

-90

-

-

“!

a c

1:

5msec

-8OmV’-

a c

-1

Icv In

n= 3

-100

/z

1 0 m S-G c

Fig. 3. Inactivation of inward currents during hyperpolarization is associated with a small inward tail current at resting potential. (Al and A2) Families of voltage clamp currents recorded from one cell, held at its resting potential of - 70 mV and hyperpolarized to the indicated levels for 20 ms. (Al) Current responses (upper) obtained during the command pulses, illustrating an inactivation which is slight at -80 and -9OmV, and greater at - 100 and - 1IOmV. Corresponding voltage traces are below. Arrow designates the superimposed inward tail currents at pulse offset. (A2) Averaged tail currents (n = 3) measured at -80 and - 100 mV, at higher gain and slower sweep. In the first case, the tail current was quite small and rapidly decayed to baseline, while, upon return from - 100 mV. there was a clear inward tail current which lasted longer than 40ms.

and

H. KoK’\

the initial segment and axon hillock and mtght LL\L masked any anomalous rectification. We thercf<)rc questioned whether better separation of the Jrtfercni conductances might be achrevcd by positlomng (1~. microelectrode in the lateral dendrite, aboui I()() /(r-n from the soma. Indeed. as >hoti n in Fig. GI i -12~111 B. under those conditions the inward rectifying pr”p-. erty could be clearly demonstrated. In (hc first example (At, A2). the currants evoked h\. y .(rril IO mV depolarizations from ;I resting potcnt~al iit’ - 77 mV were essentially the same. while with hypcrpolarization the instanraneous current ~c~i(age ILlation was linear. Thus, the rcctiher W;I~ !‘ull\ actsvated at the resting potential and the transrrron to ;I low conducting state was manifest ui(h II small depolarization. In the experiment of F(g -LB. (he family of inward currents dcmonstrate~ I bar 111~ rectifier was not totally activated at resting potential ( ---75 mV), as the inward-going currcn( mcreased slightly during the first 2 -10 tns after the onset of a hyperpolarizing pulse. The umc course of’ rhis additional activation and that of the subsequent pronounced inactivation were both accelerated as the magnitude of the imposed hyperpolaritauon increased. Finally, although the instantaneous current steps were apparently proportional to rhc ~oitage command pulse amplitude5 in both dtrecuons. rectification developed rapidly, within a few tns. when the dendrite was depolarized by 20 or 30 tnV. In summary then. instantaneouq or near instantaneous rectification can be readily demonstrated b;; ticndritic depolarizations to membrane potentials 01 60 IO -65 mV. DISCUSSION

tial, a resting conductance to one or more of these ions would produce a constant inward current, the inactivation and subsequent re-activation of which would be manifest as a transient outward tail current. In contrast, the potassium equilibrium potential (EJ is presumably hyperpolarized to or equal to the resting potential, and a high resting K+ conductance would contribute at most a small outward K+ current. Its inactivation and subsequent re-activation should result in a small transient inward tail current. such as shown in the records of Fig. 3, or in a negligible one. Thus, it appears that the M-cell has a high resting K+ conductance, at least 40-50% of which can already be inactivated by brief hyperpolarizations. A high conductance which is activated in the vicinity of E, and can be subsequently inactivated during large hyperpolarizations is characteristic of the inward K+ rectifier studied in other systems.“.‘.“.‘5.‘” However, the critical property which was not easily observed at the somatic level is that of rectification itself, that is, a low conductance for outward depolarizing currents. As mentioned already, somatic depolarizations often activated other voltage-dependent currents, which were generated at

The results presented here clearly demonstrate that a major component of the M-cell’s soma-dendritic input conductance generates an outward current at the resting potential and is inactivated during hyperpolarization, and that the lateral dendrite exhibits rectification when this neuron is depolarized. These features are consistent with a dominant K conductance, as expected. which is maximal at the resting level and has the properties of an inward K rectifier. Anomalous or inward rectifications have been observed previously in a number of nerve but those systems differ from that cells, I.:.li.lX.?O.25.26.28,31 described here in a number of ways. First. the inward rectifier of the M-cell is essentially fully activated at the resting potential, white in other neurons this conductance system is instead turned on by hyperpolarization. That is, there is no evidence that the inward rectifier of other nerve cells contributes significantly to their resting potential, in contrast with skeletal” and heartI muscle fibers, gliat cells.” and. apparently. the Mauthner cell. Second, the time constants of activation and inactivation of the inward rectifier in the hl-cell are significantly faster than in the other neurons studied thus far. Although we have

M-cell instantaneous

A1

inward

t5

SEVC,21KHz

rectifier

II)41

SEVC,35KHz

I”:c

--

-50

v:

-150

1 Omsec

1 Omsec

II

I(nA +25

-25 -50 -75 Fig. 4. Evidence that the resting conductance of the Mauthner ceil lateral dendrite cxhiblts Inuard-nl)lng rectification. (AI. A2 and B) Single electrode voltage clamp data obtamed from the lateral dendrytc in two experiments. (Al) Superimposed records of averaged clamp currents (above) produced by voltage steps (below) of i IO and + 15 mV, from a resting potential of -77 m\‘. The currents produced b> the two depolarizing pulses were the same. (AZ) Currents-voltage plot of the peak currents obtalned in the same experiment. There was a high input conductance around resting potential (RMP). which decreased with depolarizations of IOmV or more. (B) Data from a second experiment. demonstrating rapid Inactivation of outward currents produced by depolarizing command pulses > 5 mV in amplitude and inactivation of the inward currents produced by large hyperpolarizationa of 35 ms duration. .4bo\c and below are families of averaged current responses and voltage commands. respectively. CC) Diagram illustrating proposed role of dendritic (Dendr.) rectification in regulating M-cell responslvencss to excitatory inputs The soma and dendrite are outlined. with a represcntatlve excitatory input to the Iattcr bemg evoked near resting potential (- 80 mV. lower lines of current) and during dendritic depolarization ( m-50mV. upper lines of current). In the second case, the elevated input resistance of the dendrltc. represented by the thickened membrane. reduces dendritic current IN and channels more cr;cnator\

current to the soma

not rigorously quantified the response kinetics in this system. Inspection of records such as those in Figs I, 2 and 4 demonstrates that activation is complete within ;I few ms at most. and that the time course of inactivation is voltage-dependent with effective time constants in the range of 5-10 ms. In contrast. in the other neurons studied thus far. the time constan’ of activation of the inward rectifier has ranged t’rom 13 ms or more in cells of the olfactory cortex’ to 10@~700 ms in cultured hippocampal“ and splnal sensory” neurons. In current clamp experiments. anomalous rectification i5 sometimes manifest by an apparent increase in input resistance during depolarization. This phcnomcna need not be due to an inward K 1 rectifier but may rather be a consequence of non-inactivating inward Ca2’ and’or Na’ currents which activeI\ produce an additional depolarizing shift in membrane p,jtential,lh >a21I2Ii Such an alternative mechanism might contribute to the development of the rectification illustrated in Fig. 4B during depolal-

ization, and cannot be totally ruled out in that instance. although it is generally a slower process. and it would not account for the other dendritlc observations typified by the records of Fig. 4A. As stressed by Sakmann and Trube”’ hyperpolarizing inactivation may be a property associated with an inward K + rectifier. but it is not a characteristic of the rectification itself. Thus, m the case of the present study. demonstration of such an inactivation mechanism is but one line of cvidcncc suggesting that the rectifier contributes to the control of the M-cell‘s resting potential. On the other hand, inactivation itself probably has no functional significance since it is unlikely that this neuron ever undergoes an appreciable hyperpolariration under physiological condlions. The present voltage clamp analysis of the Mauthner cell’s somadendritic membrane propertIe confirms prc\ious measurements of this neuron’s input impedance. which were based upon double electrode penetrations and measurements of voltage

IO4

responses to constant current pulses.J.h,-.ii In contrast. however, anomalous rectification was not observed in the earlier studies. The most likely explanation fat this discrepancy is that the current clamp experiments typically were restricted to potential changes in the region of k 15 mV with respect to the resting potential and employed pulse durations of about 10 ms. conditions which would not be expected to reveal rectification. or its voltage-dependent inactivation. It is not possible to precisely determine the spatial distribution of an ionic conductance mechanism on the basis of voltage clamp data obtained from a neuron such as the Mauthner cell. with its large soma and two long primary dendrites. Nevertheless. the present observations suggest that an inward K’ rectifier is the major current carrier in the lateral dendrite, and that localization may have significant functional consequences in addition to the regulation and maintenance of resting potential. As illustrated in Fig. 4C, we postulate that the coupling coefficient between the dendrite and the soma is voltagcdependent, such that during a dendritic or soma-dendritic depolarization excitatory postysnaptic potentials (EPSPs) generated distally will be transmitted to the soma with less decrement than when evoked at the resting potential. That is, a greater fraction of the excitatory synaptic current will be channeled to the soma and to the initial segment--axon hillock region. This notion is based on the finding that the low conductance state appears with depolarizations of IO mV or more. Thus. as the M-cell approaches threshold for orthodromic spike initiation, the consequence of a dendritic depolarization will be an increase in the effective space constant of the lateral dendrite. This phenomenon will provide a mechanism

for enhancing the electrotomc conductlstn g.ii ihc excitatory postsynaptic responses. rclati\ v to ihc greater attenuation of those evoked in the .~bsencc ot background depolarization. Such a boostcl. Ihnctlon is similar to that proposed for a SIOU \,oltagedependent inward current generated by the soma dendritic membranes of cerebellar Purklnlc cells” although in the case of the Mauthncr cell. ~hc rapid kinetics of its rectifier suggests that background 01 steady-state depolarizations may not bt; ihc oni) means available to minimiLe the potentl~li gradlent from the dendrite to the borna. Rather. since the lo\+ conductance state is reached quickly with IO mV or more depolarization from the resting lc\cl. Iarpc composite EPSPs may be prcferentiallq tr‘tnsmittcd to the soma. This potentiation of a transient response might be less pronounced than in the steady-state. given the parallel increase in membrane time constant when the input resistance IS greater than a~ resting level, but, nevertheless, the clrect of a longer space constant would be dominant and there should bc a significant augmentation of the EPSP amplitude at the soma (see Fig. 7.32 in Ref. 19). In this manncl-. the M-cell would be rapidly shifted from :I high conductance, high threshold neuron. to ;I low conductance one which preferentially reinforces powerful excitatory synaptic inputs. Such a non-linear processing would be an important feature for a command neuron which at rest is subject to numerous powerful hyperpolarizing influences but nevertheless reliably triggers a rapid complex escape behaclor. Acknowledgemenrs-We are grateful to Juhe iakatos tar photographic services and Jean Seiler for secretarial abslstance. Supported in part by NIH grant NSl5375 and IX INSERM.

REFERENCES I. Barrett E. F., Barrett J. N. and Grill W. E. (1980) Voltage-sensitive outward currents in cat motoneurones. Lond. 304, 251-276.

J. PhJ,siol.,

2. Constanti A. and Galvan M. (1983) Fast inward-rectifying current accounts for anomalous rectification m olfactory cortex neurones. J. Physiol., Lond. 385, 153-178. 3. Faber D. S. and Korn H. (1978) Electrophysiology of the Mauthner cell: basic properties, synaptic mechanisms, and associated networks. In Neurobiology o//he Maufhner CeN (eds Faber D. S. and Korn H.). pp. 47 131. Raven Press, New York. 4. Faber D. S. and Korn H. (1982) Transmission at a central inhibitory synapse. I. Magnitude of the unitary postsynaptic conductance change and kinetics of channel activation. J. Neurophysiol. 48, 654678. 5. Faber D. S. and Korn H. (1985) Voltage dependent synaptic and non-synaptic conductances in the Mauthner cell. Sot Neurosci. Abst. 11,465. 6. Faber D. S. and Zottoli S. J. (1981) Axotomy-induced changes in cell structure and membrane excttability arc sustained in a vertebrate central neuron. Brain Res. 223, 436443. I. Fukami Y., Furukawa T. and Asada Y. (1965) Excitabilitv_ changes _ of the Mauthner cell during. collateral mhibition. J. gen. Physiol. 48, 58 l-600. 8. Fukushima Y. (1981) Sinale channel ootassium current of the anomalous rectifier. Nalure 294. 368 -371 9. Fukushima Y. (1982j Blocking kinetics of the anomalous potassium rectifier of tunicate egg studied by single channel recording. J. Physiol., Lond. 331, 31 I 33 1. 10. Funch P. G. and Faber D. S. (1982) Action potential propagation and orthodromic impulse initiation in the Mauthner axon. J. Neurophysiol. 41, 1214123 1. 1I. Furshpan E. J. and Furukawa T. (1962) Intracellular and extracellular responses of the several regions of the Mauthner cell of the goldfish. J. Neurophysiol. 25, 732-77 I 12. Furukawa T. (1966) Synaptic interaction at the Mauthner cell of goldfish. Prog. Brain Res. 21A,44 70. 13. Furukawa T. and Furshpan E. J. (1963) Two inhibitory mechanisms in the Mauthner neurons of goldfish. J. Neurophysiol. 26, 14G176.

M-cell instantaneous

inward

rectifier

lll-li

14. Hagtwara S., Miyazaki S., Woody W. and Pat!ak J. (1978) Blocktng effects of barium and hydrogen inn< on the potassium current during anomalous rectification in the starfish egg. J. Ph>~sio/.. Lond. 279, 167 185. IS. Hagtwara S., Miyazaki S. and Rosenthal N. P. (1976) Potassium current and the effect of cesium on thts current durmg anomalous rectification of the egg cell membrane of a starfish. J. ,cen. Ph~5iol. 67, 621 638. 16. Hall A. E., Hulter 0. F. and Noble D. (1963) Current voltage relations of Purkinje fibres in sodturn-deficient solutron\ J. Physiol., Lond. 166, 225-240. I7 Halliwell J. V. and Adams P. R. (1982) Voltage-clamp analysis of muscarinic cxcitatton in hippocampal ncurons. Nrr~l,r Rec. 25, 71-92. IX Hotson J. R.. Prince D. A. and Schwartckroin P. A. (1979) Anomalous Inward rectification tn htppocampal neur,,n\ J Ntwrophysiol. 42, 889 -895. 19. Jack J. J. B.. Noble D. and Tsien T. W. (1983) Elec,rrrc, C‘urrcnr Hoa, MI E.~~~cctrhic~ Cc~llc. Oxford Umverstty Prc\s. Oxford 20. Kandel E. R. and Taut L. (1966) Anomalous rectihcatton in the metaccrehral giant cells and Its con~cqucncc~ f;~r synaptic transmission. J. Physiol., Land. 183, 287-304. 21 Katz B. (1949) Les constantes Plectriques de la membrane du muscle. .drc,h.\ SC,;. PhJvrol. 3, 2x5 300 22 Korn H. and Faber D. S. (1975) An electrically mediated inhibttton III the goldfish medulla .I Vcrrr~~plrt \r,j/ 38, 452 371. 23 Llinas R. and Sugimori M. (1980) Electrophysiologtcal properties of ,,I rr/ro PurktnJe cell somata tn mammaltan cerehellar shces. J. PhJGol.. Lond. 305, l7l- 195. 24. Llinas R. and Sugimori M. (1984) Simultaneous intracellular somattc and dendritic rccordtng from Purktn)c ccII\ ot rl1ro. dynamic somadendritic coupling. Sot. Ne~ro.rci. .4h:;r. IO, 659. 25 Mayer M. L. and Westbrook G. L. (1982) ‘Anomalous’ rectification in mouse spinal sensory ncuroncs venerated hy a time- and voltage-dependent inward current. J. Ph,,.riol., Land. 332, 24 25P. 26. Nelson P. G. and Frank K. (1967) Anomalous rectification in cat spinal motoneuronrs and ctfcct of polart/tng currents on excitatory postsynaptic potential. J. Neuroph?siol 30, lO97- I I I?. 27. Newman E. A. (1985) Voltage-dependent calcium and potassium channels m rettnal glial cells. ~~rur(~ 317, 809 81 I. 2X. Purpura D., Prelevic S. and Santini M. (1968) Hyperpolarizing increase in membrane conductance tn h~ppocampal neurons. Bruin RPS. 7, 310~312. 29. Sakmann B. and Trube G. (1984) Conductance properties of single Inwardly rccttfyinp potasstum channel\ tn \cntrtcnIar cells from guinea-pig heart. J. Ph?siol.. Land. 347, 641 +57. 30. Sakmann B. and Trube G. (1984) Voltage-dependent inactivation of inward-rectifying ctnglc-channel current< tn the guinea-pig heart cells membrane. J. Phvxiol., Land. 347, 659~-683. 31 Segal M. and Barker J. L. (1984) Rat hippocampal neurons in culture: pota\stum conductances .I Vc~rvrq~/rr\/cJ/ 51, 1409 1433. 32 Stafstrom C. E., Schwindt P. C.. Flatman J. A. and C’rill W. E. (1983) Properttes of subthreshold response and actton potential recorded in Layer V neurons from cat sensorimotor cortex in rrr~o J .Ycurc~ph~~\io/. 52, 244 263. 33. Stafstrom C. E., Schwindt P. C. and Grill W. E. (19X2) Negative slope conductance due to a pcvistcnt ~ubthre\hold sodium current in cat neocortical neurons in r?rro. Bruin .fk\. 236. 721 226