Acta Physiol Scand 1992, 146, 221-232
Functional effects of a hyperpolarization-activated membrane current in the lobster stretch receptor neurone A. EDMAN, S. T H E A N D E R and W. GRAMPP Department of Physiology and Biophysics, University of Lund, Lund, Sweden
A.,
EDMAN, THEANDER, S. & GRAMPP,W. 1991. Functional effects of a hyperpolarization-activated membrane current in the lobster stretch receptor neurone. Acta PhysiolScand 1992,146,221-232. Received 4 March 1992, accepted 6 May 1992. ISSN 0001-6772. Department of Physiology and Biophysics, University of Lund, Lund, Sweden. The functional effects of a hyperpolarization-activated membrane current (Z,) in the slowly adapting lobster stretch receptor neurone were investigated. From comparisons between changes in membrane excitability due to blockage of I , by Cs+, in normally impaled and native unimpaled (Edman et al. 1987b) cells, it could be concluded that the resting voltage of native cells is distinctly more negative than -65 mV (average membrane voltage of impaled cells) and, therefore, under the control of an activated ZQ. Starting from this conclusion, impaled cells were polarized to holding (resting) voltages around - 75 mV and their polarization and excitability properties studied after tetanic impulse activity and variation of various external influences (K+, pH, temperature), both in control conditions and after blockage of Z, by 2 mM Cs+. It was found that an unblocked I , (a) greatly accelerates the initial (9076) decay of post-tetanic hyperpolarization, and (b) depresses distinctly any polarization and excitability alterations due to increases in extracellular K' concentration (from 2.5 to 10 mM), variations in extracellular pH (between 6.4 and 8.6), and changes in temperature (between 14 and 24 "C). It was inferred that in well polarized cells, I , plays a role as a stabilizer of membrane polarization and excitability in conditions of varying external influences. From a model study of ZQ it could be concluded that, with its slow dynamic responses, the current is well adapted to its functional purposes and to the rather slow homeostatic effects of the cell's Na-K pump. Key words: anomalous rectifier, excitability stabilization, hyperpolarization-activated membrane current, inward rectifier, lobster stretch receptor neurone, microelectrode artefacts, resting voltage, voltage stabilization.
I , is a depolarizing membrane current in the stretch receptor neurone of the lobster. T h e current is activated by membrane hyperpolarization beyond - 65 mV, and totally blocked by 1-2 m M Cs+ (Edman et al. 1987b). From previous publications (Edman 1987, Edman et al. 1987b, Edman & Grampp 1989) it is known: (1) that I , is carried by Na+ and Kf which are turned over at saturable binding sites in the interior of the conducting channel (QCorrespondence : Wolfgang Grampp, Department of Physiology and Biophysics, Solvegatan 19, S-223 62 Lund, Sweden.
channel); ( 2 ) that the current is controlled by means of a two-step gating mechanism; and (3) that the current is very little affected by changes in the cell's physical and chemical environment (see Discussion). Even though a number of facts are known about I,, which may assume considerable strengths in voltage regions below -65 mV, there is still no clear understanding of the function of this current. Much of this is because I , is activated in voltage regions in which, with the exception of post-tetanic hyperpolarizations, no electrical responses are normally taking place in cells examined by means of intracellular
22 1 8-2
222
.A". Ednzan et al.
techniques. There is, horn-ever, some evidence that in cells not disturbed by intracellulnr interventions the resting voltage might be more negative than -65 mV and, hence, under the control of an activated I , (Edman et a / . 1987b). The aim of the present study was to investigare this possible arrangement and its implications for an!- functional effects of I,. The results suggest that the resting voltage of cells, not disturbed by intraccllular measurements, lies well within the voltage range of 1, activation and that, as a consequence hereof, 1, normally exerts a stabilizing effect on the cell's membrane voltage and membrane ercitability in the presence of various disturbing influences (tetanic stimulation, changes in temperature and extracellular ion concentrations). From a model study of post-tetanic hyperpolarization, it also appears that the kinetic properties of IQ7 in particular its slow activation, are well adapted to the transport rate of the cell's Na-K pump. Some of the results to be discussed have been presented in a short communication (Theander 8i Grampp 1991).
toxin (8.3 phi) was added to all solutions in order to suppress spontaneous inhibitory post-synaptic potentials. In the case of some voltage clamp measurements in which the initiation of action currents was undesired, 480 nM tetrodotoxin (TTX), 10 mM tetraethylammonium (TEA), and 0.5 mhi 4-aminopyridine (4-,4P) were also added to the solutions. Mathematical reproductions of some of the empirical findings, and of conceivable I , functions, were performed using a previously de\-eloped FORTRAN-based (Soderlind 1980) model of the stretch receptor's transmembrane current control (Edman et al. 1986, 1987a, b, Edman 8( Grampp 1989). RESULTS T h e e f e c t of I , on electrical membrane properties in resting conditions
In previous experiments it was found that the blockage of I , b?- 1 mM Cs' causes a minor (1 1-12°0)1 but highly significant ( P < 0.001), decrease in membrane excitability in unimpaled cells (Edman et al. 1987b). From this it was inferred that normally in such cells the resting h l E 7'H 0 D S voltage is under the control of an activated I , The experiments were performed on slowly and, hence, more negative than in impaled cells, adapting stretch receptor neurones from the in which the average resting voltage was found to second and third abdominal segments of the be -65 mV. European lobster (Homarus gamnzarus). The For a validation of this inference, measurereceptors were prepared and intracellular micro- ments were made of how Cs+ affected membrane electrode measurements of electrical membranc resistance, membrane polarization, and memproperties were performed, as described in earlier brane excitability in impaled cells whose 'resting' investigations (Edman et at. 1986, 1987a, b). 1-oltagehad been set to about - 75 mV, i.e. near The temperature was kept constant at 18 "C, to a level ( - 77 mV) a t which the mechanism of unless stated otherwise. ZC4activation has its highest voltage sensitivity 'The standard bathing solution had the fol-- (cf. Edman et al. 1987b). T h e measurements are lowing composition (in mxt): NaCl 325, KC1 j3 illustrated in Figure 1 which, in (a), shows the CaCl, 25. MgC1,4, MgSO, 4, Tris HC1 26, Tris effect of 2 mM Cs+ on a cell's steady state I-V base 4,and glucose 5. The pH of the solutior: relationship in voltage clamp and, in (c), the as: set t o 7.3--7.4by first bubbling the solution effect of Cs' on the same cell's 'resting' with .iO,, CO, in 0, and then titrating it with polarization and relative membrane excitability O..i 31 NaOH. The concentration of HCO, was (the latter estimated from the inverse in height of thereby adjusted to about 1.5 mxi. Solutions intracellularly injected 300 ms long depolarizing with wrying K + concentrations (between 2.5 square current pulses just able to evoke one and 10 mar) were obtained by adding, to the action potential) in current clamp conditions. standard solution, varying amounts of KCI From the results of such measurements it without osmotic compensation. Solutions with appeared that, by eliminating inward (anomvarj-ing pH were prepared by titrating the alous) rectification in the voltage region of normal standard solution with either HCl (to I , activation (from about -65 mV to about pH 6.5~-6.7)or Tris base (to pH 8.2-8.4). Picro- -95 my), Cs' gives rise to three major effects.
Functional efects of I ,
h
+51
+5-
0-
0.
P -E -5-
223
Y
-5
-
-1Od
-10-100
-80
-60
V(mV)
V
I
40
NmV)
(d) 7
,
-80
-100
Ih block
I
4
1-75
I
I
m"
Fig. 1. (a, b) Relationships between membrane voltage, V, and membrane current, I,, in a slowly adapting receptor (a) and its mathematical model (b) after elimination of any spike current generation and adjustment of the resting voltage to a reference value of -76 mV, in control conditions and in presence of 2 mM Cs+ or following blockage of the I , system, as indicated. The dotted horizontal lines mark the reference (zero) value of the membrane current in resting conditions. (c, d) Membrane voltage, V , and relative membrane excitability, denoted by figures in a slowly adapting receptor beside arrows or vertical lines pertaining to the current traces (Iinj), (c) and its mathematical model (d) whose resting voltages had been set to -76 mV, in control conditions and during exposure to 2 mM Cs+ or blockage of the I , mechanism, as indicated. The relative excitability values were inferred from the inverse in height of 300 ms long depolarizing square current pulses [appearing as dots in (c) and vertical lines in (d)] just able to evoke one action potential
These are : (1) a n increase in the 'resting' input resistance by 107 & 2% (mean & SE, n = 5 ) (Fig. l a ) ; (2) an increase in 'resting' membrane polarization by 7.4k0.3 m V (Fig. 1 a, c) a n d ; (3) a decrease in steady state membrane excitability by 1 7 f 2 % (Fig. lc). The last of these effects is comparable to that previously observed in extracellular measurements on unimpaled cells (cf. Edman et d. 1987b). Hence, the above experiments d o support the notion that, in cells undisturbed by intracellular interventions, the resting voltage is distinctly more negative than -65 mV and, therefore, under the control of an activated I,.
I n addition, they indicate that, in native cells, the resting membrane resistance is kept quite low by an activated Q-channel system. Functionally, such an arrangement would provide for a certain degree of stabilization of the cell's resting polarization and, hence, membrane excitability in the presence of various disturbing influences. Such an effect seems to be particularly valuable for a receptor neurone whose proper function would be greatly dependent on a reasonably constant membrane excitability. T o test this possibility, cells were polarized to about -75 m V and then exposed to a number of disturbances (intense electrical stimulation evok-
(b) i
......
I 0”
a
AV
.-i2 mM Cs‘
4 rnin
e
.......
I
Fig. 2. Post-tetanic hyperpolarizations, AT irelati\-e to resting voltage at - 75 mV), in a slowly ‘idaping receptor following continuous impulse firing a t 40 Hz for 2.5 min in the absence and presence of 2 m u C s . (a), and in its mathematical model following a 15 mM increment of the intracellular \ a * concentration in the presence and absence of a normally functioning I , system (b), as indicated. (c) Recording of impulse activit!-, AV. due to a sawtooth stimulus, ISt,,”.(d) Posttetanic h!-perpolarizations following repetitiw saw-tooth stimulation, as shown in (c), at 1 Hz for 2 min in the absence and presence of 2 m u (AA, as indicated. In both cases, the average impulse fiequenc!- fell from 14 Hz, in the beginning, i o 9 Hz, at the end of the stimulation sequence. The horizontal dotted lines in (a) and (d) . . mark the terminal uhases of imuulse firing and the level of pre-tetanic membrane polarization, respectivel!- The arrows indicate the time for a 90°/, decq of the post-tetanic h!-perpolarizations, L
ing post-tetanic membrane hyperpolarizations, changes in the cells’ physical and chemical environment), while the effects of these disturbances on membrane polarization and/or excitability were measured both in control conditions and after the blockage of I , b!- 2 mx4
2.3 min (Fig. 2a) or prolonged saw-tooth pulses (meant to mimic stimulatory effects during movement-dependent load-compensating feedback acti\ation; cf. Kennedy & Davis 1977) at 1 Hz for 2 min (Fig. 2c, d), both in control conditions and after the blockage of 1, by 2 mM
Ca .
Cs’ .
The actkation on electrical membrane properties in post-tetanic conditions, experiments were performed as shown in Figure 2 . I n these experiments recordings were made of post-tetanic hyperpolarizations, following repetitive impulse firing e\-oked by stimulation \I-ith either short ( 3 ms) square pulses at 40 Hz for
From the results it became evident that, irrespecti\ely of the mode of stimulation, Cs+ causes (1) a 32 2 Z o o (mean SE, n = 4) increase in depth of the post-tetanic hyperpolarization and ( 2 ) a 2.4kO.j-fold lengthening of the first 90 O o of the decay of the hyperpolarization. Also, it became clear that Cs+ has essentially n o effect on the total length of post-tetanic hyperpolarizations. T h e latter observation is not surprising, considering that none of the factors determining the total length of the post-tetanic
Functional e f e c t s of I ,
Iinj -
225
0.8
1.0
14nA
1.o
1.1
_.
2 rnM C s *
lhj
-f
1.0
m.2
-
0.7
T1'o -1 -
4 nA
5 min
Fig. 3. Membrane voltage, V, and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Z1,,J, in a slowly adapting receptor at 14, 18 and 25 "C in the absence (a) and presence of 2 mM Cs+ (b), as indicated. The excitability values were obtained as described in Figure 1 . For a continuous assessment of the cell's input resistance, polarizing voltage transients due to 300 ms long intracellularly injected square current pulses (appearing as a line of dots in the current traces) were recorded throughout the experiment.
hyperpolarization, viz. the degree of intracellular Na+ accumulation during impulse firing and the rate of active Na+ extrusion after the impulse activity, is influenced by Cs' to any extent. The effects of Cs+ on the initial phase of the posttetanic hyperpolarization, on the other hand, are readily explainable on the basis of a suppressed I , action. I t can be concluded, therefore, that an activated IQ contributes to a fast restoration to near-resting values of the membrane polarization after various kinds of impulse firing and, in this way, minimizes the length of functionally dependent variations in receptor excitability (cf. Edman et al. 1987b). The effect of I , on electrical membrane properties in the presence of varying external influences
A number of external influences are able to affect cells with respect to their electrical membrane properties by producing changes in active ion transport or passive ion fluxes, or both. For a study of how such properties are dependent on
an activated IQ,experiments were performed as shown in Figures 3-5. Thus, cells were polarized to about -75 mV and then examined with respect to their membrane polarization, membrane resistance (indicated by the height of voltage transients produced by intracellularly injected 300 ms long hyperpolarizing square current pulses) and relative membrane excitability (reflected by the inverse in height of intracellularly injected depolarizing current pulses, as described above), during exposure to varying external influences (varying temperature, varying pH, varying K+ concentration), both in control conditions (Figs 3 , 4 & 5a) and in the presence of 2 mM c s + (Figs 3, 4 & 5 b). Consistently, it was found that Cs+ leads to increases of the resting input resistance through blockage of the Q-channel system. Also, it was found that, in the presence of Cs+, variations of the external influences give rise to changes in membrane polarization, and in relative membrane excitability, that are distinctly (by factors of 2, or more) greater than in control conditions. This is particularly true for the overshooting
226
.I. Edrnan et al.
V
-/------
5 min
Fig. 4. Xlembrane voltage. I ., and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Zini), in a slon.ly adapting receptor in solutions whose pH n a s set to 6 . 5 . 7.4, and 8.4 in the absence (a) and presence of 2 mM Cs' (b), as indicated. Estimates of the cell's membrane excitability and input resistance were performed as described in Figure 3.
fluctuations in membrane polarization which can be ascribed to temperature- (Fig. 3 ) or pH- (Fig. 4) dependent changes in pump current production (cf. Rakowski et a/'. 1989, Bonting 1970). In the case of estracellular K- (Fig. 3)' it was noted that lowering its concentration leads to increases in membrane polarization and decreases in relative membrane escitabilitv that are similar in size in the presence of Cs- and in control conditions. On the other hand, raising estracellular K- causes decreases in membrane polarization and increases in membrane excitability which amount to much higher !-alues (providing for the initiation of spontaneous impulse firing in three out of four investigated cells) in the presence than in the absence of Cs'. These differential effects are explicable on the ground that the Q-channel conductance and, hence, the Z, production increases with increasing concentrations of extracellular K (Edman & Grampp 1989). Consequently, blocking the laproduction by Cs' will give rise to changes in electrical membrane properties which are more pronounced in higher than in lower extracellular K'. '-
Thus, there is reason to conclude that, in an activated state, the I , system takes part in stabilizing the cell's membrane voltage and membrane excitability in the presence of varying external influences. I n particular, it may do so in the presence of influences that are giving rise to overshooting changes in membrane polarization and/or to increases in the degree of I , activation. .Ilathematiral model studies of the I , system
From the above findings it became evident that the resting membrane voltage is more negative (b! about 10mV) in cells not disturbed by intracellular interventions than in cells impaled n ith micro-electrodes. Not unlikely, this is because the impaling electrode gives rise to the appearance of an unspecific ion leak passing through an imperfect seal between the cell membrane and the electrode wall, and an intracellular accumulation of C1-, if KCl is used as an electrode electrolyte. For a test of this explanation, an investigation was made of nhether or not a previously developed mathematical receptor model (Edman et al. 1986,
Functional efects of I , (a) 5 mM K*
2.5 mM K+
10 mM K +
227
5 mM K'
Fig. 5. Membrane voltage, V , and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Iinj), in a slowly adapting receptor in solutions with 2.5 mM K+, 5 mM K+, and 10 mM K+ in the absence (a) and presence of 2 mM Cs+ (b), as indicated. Estimates of the cell's membrane excitability and input resistance were performed as described in Figure 3. Note that, in the presence of Cs+, high K+ elicits spontaneous impulse firing (only the bottom parts of the action potentials are shown), implying an infinitely high membrane excitability. 1987a, b, Edman & Grampp 1989) was able to predict a resting voltage of about -75 mV, if modified values were given to some of its parameters. It was found that a prediction of - 75.1 mV is in fact possible, if the value for the intracellular C1- concentration is lowered from its standard value (46 mM) to 20 mM (cf. Edman et al. 1983) and the leak permeabilities are corrected for the exclusion of an 'impalement'leak channel which has identical permeabilities for Na+, K+, and C1-, each amounting to 80% of the standard (impaled cell's) Naf leak permeability. The latter is achieved by multiplying the standard leak permeabilities for Na+, K+, and C1- by the factors 0.200, 0.974 and 0.627, respectively. The ability of the model to predict an ' unimpaled ' resting voltage of about - 75 mV is displayed in Figure 1(b), where it also appears that the model's I-V relationship in control
conditions and in the presence of a blocked IQ is quite similar to that of the living cell, and that a total blockage of 1, causes the resting voltage to fall by about 7 mV. The latter is also illustrated in Figure l ( d ) where, in addition, it is shown that the simulated membrane excitability (reflected by the inverse in height of depolarizing 300 ms long current pulses just strong enough to evoke one action potential; see above) is decreased by the same relative amount as the membrane excitability of the living cell following a total suppression of IQ. I n Figure 2(b) it is demonstrated that the modified model is able to simulate quite faithfully the time course (initial depth, rate of decay and total duration) of post-tetanic hyperpolarizations due to a 15 mM increment of the intracellular Na+ concentration (as has been seen to result from a 40 Hz impulse firing for 2.5 min; cf.
228
A?.Edninn et
al.
Fig. 6. hlathematicall!- simulated post-tetanic membrane h!-perpolarizations (due to a 15 mM increment of the intracelluiar S a . concentration), A I ., in control conditions (a), and after halving (h) o r doubling (c) the rate con~tantsfor I , ai:ti\-ation. The horizontal dotted lines mark the level of resting polarization a t -75.4 ml-, The arrow indicate the time for a 752; decay of the posttetanic hyperpolarizations For further details, see text.
Edman rt a / . 1986) both in control conditions and after blocking the Zci. T h u s there is reason to assume that the model adopted reproduces the effects of I , in the native stretch receptor neurone in a representative n.a!-. O n the basis of this assumption, it appeared appropriate to use the model for an investigation of the I,, system with respect to its d!-namic properties. These properties seem to be well adjusted, as far as the I , function of voltage and excitability stabilization in the presence of slowl!varl-ing external influences is concerned. However, it might be asked whether an increased rate in activation would provide for a faster deca! of post-tetanic hyperpolarizations and, thereb! , for a Faster restoration of pre-tetanic excitabilit) conditions. For an investigation of this possihilit!-, post-tetanic hyperpolarizations (in response to a 15 mhI increment of the intracellular Na' concentration) were simulated in control conditions and after, respectively, halving and doubling the rate constants of I , activation. T h e results are shown in Figure 6. It is seen that altering the rate constant has essentially no effect on the maximum depth of the post-tetanic hyperpolarization, but leads to distinct changes in the rate of- its initial (about 50°,) decal-. H o w v e r , for the subsequent decay, and the
gradual restoration of the membrane voltage to near pre-tetanic values, the alteration of the rate constants is without significant consequences. I n Figure 6 this is evident from the fact that, already at 7 j o o of the decay of the post-tetanic hyperpolarization (arrows), the time needed for the process is practically the same in the different experimental conditions. I n conclusion, therefore, the view m a j be adopted that the dynamic properties of the ZQ system are well adjusted also for achieving a maximum acceleration of the initial decay of post-tetanic hyperpolarizations which, for their total elimination, are dependent on the transport capacity of the Na-K pump. DISCUSSION T h e present study was undertaken in order to clarify the function of IQ in the lobster stretch receptor neurone. I n summary, it could be shown: (1) that in native unimpaled cells, the resting voltage is under the control of an activated I , and, hence, more negative than in impaled cells, where it is found to be around -65 mV; ( 2 ) that in native cells the presence of an activated I , gives rise to a decreased resting input resistance; (3) that, as a result hereof, IQis able to play a role as a stabilizer of membrane
Functional effects of’ I , polarization and excitability in conditions of varying external (physical and chemical) influences; and (4) that, with its slow dynamic responses, I , is well adapted to its functional purposes and to the rather slow homeostatic effects of the cells’ Na-K pump. Below these points will be discussed separately.
229
voltage occurs at values which provide for a significant activation of Z, and, therefore, must be under the influence of this current. Due to such an arrangement it seems possible that the Z, mechanism is functionally involved in the control of the cells’ membrane excitability. In the following this possibility will be discussed in further detail.
Adjustment qf the resling membrane voltage
’The finding that unimpaled cells have a distinctly more negative resting membrane voltage than impaled cells draws attention to the possibility that, in the latter cells, the resting membrane polarization is determined, not only by genuine membrane currents, but also by an unspecific leak current passing through an imperfect seal around the impaling electrode. In so-called successfully impaled cells, this short-circuiting leak may be quite small and produce no significant reduction of the cells’ input resistance. Nevertheless, it may allow a flow of current which, because of existing driving forces, is mainly carried by Na+ and which, because of the well maintained input resistance, can be shown to give rise to appreciable membrane depolarizations. At the same time, the inward leak is not larger than that it can be counterbalanced completely by the cells’ Na-K pump, whence the preparations are able to stay in seemingly perfect steady state for many hours of experimentation. In view of this apparent stability and the finding that C1- leaking from the impaling electrode affects the resting membrane polarization only marginally because of a rather small transmembrane C1- permeability (Edman et al. 1986), it was presumed in previous studies that the values of the resting membrane voltage recorded would not deviate much from those pertaining to native conditions. It was only with the discovery of Z,, and growing difficulties to identify the functional significance of this current which is activated at voltages apparently outside the cells’ normal working range, that it became necessary to reconsider whether or not measurements of the resting membrane voltage in impaled cells would reflect native conditions. As a result of such reconsiderations, there is now reason to assume that the resting membrane voltage of native unimpaled cells can be as much as 10 mV more negative than that of successfully impaled and apparently well adapted cells. This means that in native conditions the resting
Functional efects of I ,
In the lobster stretch receptor neurone there is essentially no overlap between the voltage regions of I,, and action current activation. Because of this, Z, is unable to play a functional role in the cell’s firing control, as similar currents are seen to do in, for example, heart muscle cells (DiFrancesco & Noble 1989, Denyer & Brown 1990, Bois & Lenfant 1990, van Ginneken & Giles 1991), smooth muscle cells (Hisada et al. 1991), and various mammalian central neurones (McCormick & Paper 1990a, Foehring & Waters 1991). Also, in the lobster stretch receptor, the activation of ZQ is not found to be (significantly) affected by : (1) stretch activation (unpublished observations); ( 2 ) a number of transmitters and hormones of which some are indigenous to crustaceans (Edman 1987), (3) a number of naturally occurring ions (Na+, Ca2+,H+)(Edman & Grampp 1989, 1991); and (4) by temperature (Edman & Grampp 1991). It is not likely, therefore, that the I , system should, as corresponding systems in other preparations [DiFrancesco & Tromba 1988 (heart muscle cells), Coleman & Parkinton 1990 (smooth muscle cells), McCormick & Pape 1990b, Takahashi 1990, Uchimura 1990, Gordon 1991 (mammalian central neurones and peripheral nerve)], be involved in mediating any effects that various external influences might have on the cell’s membrane excitability. I n consequence, there is reason to conclude that 1, becomes functionally involved primarily via its voltage sensitivity. The sensitivity is indeed quite high, such that the current needs only about 30 mV of potential change for a complete activation (Edman et al. 1987) whereas, for example, the Na+ current of the action potential mechanism needs about 60mV for a corresponding reaction (cf. Hille 1984). As a result hereof, I , is able to lower the cell’s input resistance significantly (by maximally 40-60~,,)
230
-4.Edmiin et al. ,
(a>
‘/
Y
I
-100
I
I
-80
-60
t
I
-80
-100
-60
VhV)
VhV)
Fig. 7 . hlathematicall!- simulated relationships between membrane voltage, V , and membrane currents, Ist3 (gated T a - current), I , (gated K* current), I,, (leak current), I , (Q-current), I , (pump current), and I,,, (total membrane current), as indicated, in control conditions (a) and after increasing the maximum Q-channel conductance by a factor of 3 (b). 0, marks the points where the total membrane current is zero at resting voltage levels of -75.0 m y in (a), and -72.2 m y in (b)
near native resting voltage levels and, thereby, to counteract effectivell- any fluctuations in membrane polarization and, hence, membrane escitability that might arise because of a number of disturbing influences. I n the present study such influences were achieved by varying the external concentrations of KT and H-, by changing the ambient temperature, and bj- provoking high-frequency impulse firing for prolonged periods of time. For all of these influences, it was assumed that the!might occur in the lil-ing animal, for example, in connection with feeding [giving rise to estracellular increases in K’ and H+ as a result of ;i less precise regulation of the ionic balance in thc bod\- fluids (cf. Robertson 1960, Parry 1960)], seasonal temperature variations or locomotion through lavers of water with different temperatures, and intense stimulation [giving rise to an increased pump current production (Edman r t i l l . 1987a), extracellular I(+ accumulation (Edman i’t a / . 1983), and intracellular acidification due to transmembrane ion shifts and an accelerated energj- metabolism (cf. Hanseri 198.5 11. From the results of this stud!-, it is clear that any changes in membrane polarization and membrane excitability due to the above kind of influences are reduced markedly by a functioning Zct. It can be concluded, therefore, that the purpose of this current is to stabilize the cells’
membrane excitability in the presence of a number of disturbing influences. This homeostatic effect appears to be of special value for a receptor neurone whose function is to respond to input signals in a consistent way, irrespectively of any varying external circumstances.
.-ldaptation of I ,
to
the Ku-K pump
From the above observations it might be inferred that a particularly important function of the IQ system is to counteract any changes in membrane polarization and membrane excitability that could be due to a n altered pump current production. It might seem conceivable, therefore, that the I , system is particularly well adapted to the cells’ &a-K pump. One indication hereof may be seen in the fact that, in this study, it has been possible to argue on the basis of computer simulations that the dynamics of I , are adjusted so that they provide for no faster and, hence, n o more energy consuming initial post-tetanic voltage restoration that what seems to be justifiable with respect to the transport capacity of the Na-K pump. Another indication is provided in Figure 7(a), which shows computer simulations of stationary relationships between membrane polarization and membrane currents in an unimpaled cell. From these it appears that, in resting conditions (in which the total membrane current, It,,, is
Functional e f e c t s of I , zero), the transmembrane current balance is dominated by an outwardly directed pump current and a somewhat larger inwardly directed I,. T h i s arrangement, in combination with the fact that both currents can be activated to similar maximum values in the presence of normally occurring intracellular Na+ concentrations (Edman et a/. 1986, Edman & Grampp 1989), may be in evidence of the I , system being balanced well against the Na-K pump. With a more powerful IQ system (represented by, for example, a larger maximum Q-channel conductance) it is true that the current could achieve a greater voltage stabilizing effect in resting conditions. This is illustrated by the current-voltage relationships in Figure 7 (b), which shows that a distinctly lower resting input resistance is obtained after a three-fold increase of the maximum Q-channel conductance. However, it is also true that this improvement in voltage stabilization entails a greater consumption of energy that is needed for maintaining a larger Q-channel conductance (a larger number of Q-channels) and a larger pump activity in the presence of a larger transmembrane ion leakage, as is evident in Figure 7(b). Apparently, in the living cell, this cost is too high for the benefits to be attained. Hence, it may be concluded that, in its observed form, the I , system is well adapted to the properties of the cells’ Na-K pump. This conclusion seems to be supported also by the observation that the production of ZQ is markedly smaller in the rapidly adapting lobster stretch receptor neurone which is equipped with a NaK p u m p that is clearly weaker than that of the slowly adapting cell (Edman et al. 1986). A fuller account of this functional arrangement will be given in a subsequent publication. The authors wish to express their sincere gratitude to Ms Kristina Borglid for expert technical and secretarial help. The work has been supported by the Swedish Medical Research Council (project No 2082), and by grants from the Medical Faculty of the University of Lund.
RE F E RE N CE S Bors, P. & LENFANT, J. 1990. Isolated cells of the frog sinus venosus: properties of the inward current activated during hyperpolarization. PJiigers Arch 416, 339-346. Bonting, S.L. 1970. Sodium-potassium activated adenosinetriphosphatase and cation transport. In :
231
E.E. Bittar, (ed.), Membranes and Ion Transport, pp. 257-363. Wiley-Interscience, John Wiley & Sons Ltd, London. Coleman, H.A. & Parkinton, H.C. 1990. Hyperpolarization activated channels in myometrium : A patch clamp study. In: Frontiers in Smooth Muscle Research, pp. 665-672. Alan R. Liss, Inc., New York. Denyer, J.C. & Brown, H.F. 1990. Pacemaking in rabbit isolated sino-atrial node cells during Cs+ block of the hyperpolarization-activated current if. J Physiol429, 401-409. DiFrancesco, D. & Noble, D. 1989. The current if and its contribution to cardiac pacemaking. In J.W. Jacklet (ed.), Neuronal and Cellular Oscillators, pp. 31-57. Dekker, New York. DiFrancesco, D. & Tromba, C. 1988. Inhibition of the hyperpolarization-activated current (i,) induced by acetylcholine in rabbit sino-atrial node myocytes. 3 Physi,o1405, 477-491. Edman, A. 1987. Analysis of a hyperpolarizationinduced membrane current in the crustacean stretch receptor. Thesis (LUMEDW/(MEFF-1019)/ 161). Reprocentralen, University of Lund, Lund, Swede!. Edman, A., Gestrelius, S. & Grampp, W. 1983. Intracellular ion control in lobster stretch receptor neuron!. Acta Physiol Scand 118, 241-252. Edman, A., Gestrelius, S. & Grampp, W. 1986. Transmembrane ion balance in slowly and rapidly adapting lobster stretch receptor neurones.JPhysio/ 377, 171-191. Edman, A., Gestrelius, S. & Grampp, W. 1987a. Analysis of gated membrane currents and mechanisms of firing control in the rapidly adapting lobster stretch receptor neurone. f Physiol 384, 649-664. Edman, A., Gestrelius, S. & Grampp, W. 1987b. Current activation by membrane hyperpolarization in the slowly adapting lobster stretch receptor neuron:. 3 Physiol384, 671-690. Edman, A. & Grampp, W. 1989. Ion permeation through hyperpolarization-activated membrane channels (Q-channels) in the lobster stretch receptor neurone. Pjiigers Arch 413, 249-255. Edman, A. & Grampp, W. 1991. Ion (H+, Ca2+,Co”) and temperature effects on a hyperpolarizationactivated membrane current in the lobster stretch receptor neurone. Acta PhysiolScand 141,251-261. Foehring, R.C. & Waters, R.S. 1991. Contribution of low-threshold calcium current and anomalous rectifier (I,) to slow depolarizations underlying burst firing in human neocortical neurones in aitro. Neurosci Lett 124, 17-2 1. Gordon, T.R., Kocsis, J.D. & Waxman, S.G. 1991. TEA-sensitive potassium channels and inward rectification in regenerated rat sciatic nerve. Nerve & Muscle 14, 640-646.
232
.$. Edmaii et al.
Hanscn, .A.J. 1985. Efect of anoxia on ion distribution in the brain. Phj~siolKtji. 6.5, 101-148. Hille, H, 1084. Iotiil Cha ti t i ~ l s F ~ l . i / [blr i .I Ilwihr(i t i t . ; . Sinauer .Associates Inc. Publishers Sunderland, llassachusetts. Hisada, T., Ordwy. R.\f-., Kirber. tl.T., Singer, J..i. 8; \\-alsh Jr, J.\-. 1991. Hyperpolarization-actii-ated cationic channels in smooth niusclc cells are stretch sensiti! e. Pfiiig17t.s . - l r ~ . / i417, 49.
Functional effects of a hyperpolarization-activated membrane current in the lobster stretch receptor neurone A. EDMAN, S. T H E A N D E R and W. GRAMPP Department of Physiology and Biophysics, University of Lund, Lund, Sweden
A.,
EDMAN, THEANDER, S. & GRAMPP,W. 1991. Functional effects of a hyperpolarization-activated membrane current in the lobster stretch receptor neurone. Acta PhysiolScand 1992,146,221-232. Received 4 March 1992, accepted 6 May 1992. ISSN 0001-6772. Department of Physiology and Biophysics, University of Lund, Lund, Sweden. The functional effects of a hyperpolarization-activated membrane current (Z,) in the slowly adapting lobster stretch receptor neurone were investigated. From comparisons between changes in membrane excitability due to blockage of I , by Cs+, in normally impaled and native unimpaled (Edman et al. 1987b) cells, it could be concluded that the resting voltage of native cells is distinctly more negative than -65 mV (average membrane voltage of impaled cells) and, therefore, under the control of an activated ZQ. Starting from this conclusion, impaled cells were polarized to holding (resting) voltages around - 75 mV and their polarization and excitability properties studied after tetanic impulse activity and variation of various external influences (K+, pH, temperature), both in control conditions and after blockage of Z, by 2 mM Cs+. It was found that an unblocked I , (a) greatly accelerates the initial (9076) decay of post-tetanic hyperpolarization, and (b) depresses distinctly any polarization and excitability alterations due to increases in extracellular K' concentration (from 2.5 to 10 mM), variations in extracellular pH (between 6.4 and 8.6), and changes in temperature (between 14 and 24 "C). It was inferred that in well polarized cells, I , plays a role as a stabilizer of membrane polarization and excitability in conditions of varying external influences. From a model study of ZQ it could be concluded that, with its slow dynamic responses, the current is well adapted to its functional purposes and to the rather slow homeostatic effects of the cell's Na-K pump. Key words: anomalous rectifier, excitability stabilization, hyperpolarization-activated membrane current, inward rectifier, lobster stretch receptor neurone, microelectrode artefacts, resting voltage, voltage stabilization.
I , is a depolarizing membrane current in the stretch receptor neurone of the lobster. T h e current is activated by membrane hyperpolarization beyond - 65 mV, and totally blocked by 1-2 m M Cs+ (Edman et al. 1987b). From previous publications (Edman 1987, Edman et al. 1987b, Edman & Grampp 1989) it is known: (1) that I , is carried by Na+ and Kf which are turned over at saturable binding sites in the interior of the conducting channel (QCorrespondence : Wolfgang Grampp, Department of Physiology and Biophysics, Solvegatan 19, S-223 62 Lund, Sweden.
channel); ( 2 ) that the current is controlled by means of a two-step gating mechanism; and (3) that the current is very little affected by changes in the cell's physical and chemical environment (see Discussion). Even though a number of facts are known about I,, which may assume considerable strengths in voltage regions below -65 mV, there is still no clear understanding of the function of this current. Much of this is because I , is activated in voltage regions in which, with the exception of post-tetanic hyperpolarizations, no electrical responses are normally taking place in cells examined by means of intracellular
22 1 8-2
222
.A". Ednzan et al.
techniques. There is, horn-ever, some evidence that in cells not disturbed by intracellulnr interventions the resting voltage might be more negative than -65 mV and, hence, under the control of an activated I , (Edman et a / . 1987b). The aim of the present study was to investigare this possible arrangement and its implications for an!- functional effects of I,. The results suggest that the resting voltage of cells, not disturbed by intraccllular measurements, lies well within the voltage range of 1, activation and that, as a consequence hereof, 1, normally exerts a stabilizing effect on the cell's membrane voltage and membrane ercitability in the presence of various disturbing influences (tetanic stimulation, changes in temperature and extracellular ion concentrations). From a model study of post-tetanic hyperpolarization, it also appears that the kinetic properties of IQ7 in particular its slow activation, are well adapted to the transport rate of the cell's Na-K pump. Some of the results to be discussed have been presented in a short communication (Theander 8i Grampp 1991).
toxin (8.3 phi) was added to all solutions in order to suppress spontaneous inhibitory post-synaptic potentials. In the case of some voltage clamp measurements in which the initiation of action currents was undesired, 480 nM tetrodotoxin (TTX), 10 mM tetraethylammonium (TEA), and 0.5 mhi 4-aminopyridine (4-,4P) were also added to the solutions. Mathematical reproductions of some of the empirical findings, and of conceivable I , functions, were performed using a previously de\-eloped FORTRAN-based (Soderlind 1980) model of the stretch receptor's transmembrane current control (Edman et al. 1986, 1987a, b, Edman 8( Grampp 1989). RESULTS T h e e f e c t of I , on electrical membrane properties in resting conditions
In previous experiments it was found that the blockage of I , b?- 1 mM Cs' causes a minor (1 1-12°0)1 but highly significant ( P < 0.001), decrease in membrane excitability in unimpaled cells (Edman et al. 1987b). From this it was inferred that normally in such cells the resting h l E 7'H 0 D S voltage is under the control of an activated I , The experiments were performed on slowly and, hence, more negative than in impaled cells, adapting stretch receptor neurones from the in which the average resting voltage was found to second and third abdominal segments of the be -65 mV. European lobster (Homarus gamnzarus). The For a validation of this inference, measurereceptors were prepared and intracellular micro- ments were made of how Cs+ affected membrane electrode measurements of electrical membranc resistance, membrane polarization, and memproperties were performed, as described in earlier brane excitability in impaled cells whose 'resting' investigations (Edman et at. 1986, 1987a, b). 1-oltagehad been set to about - 75 mV, i.e. near The temperature was kept constant at 18 "C, to a level ( - 77 mV) a t which the mechanism of unless stated otherwise. ZC4activation has its highest voltage sensitivity 'The standard bathing solution had the fol-- (cf. Edman et al. 1987b). T h e measurements are lowing composition (in mxt): NaCl 325, KC1 j3 illustrated in Figure 1 which, in (a), shows the CaCl, 25. MgC1,4, MgSO, 4, Tris HC1 26, Tris effect of 2 mM Cs+ on a cell's steady state I-V base 4,and glucose 5. The pH of the solutior: relationship in voltage clamp and, in (c), the as: set t o 7.3--7.4by first bubbling the solution effect of Cs' on the same cell's 'resting' with .iO,, CO, in 0, and then titrating it with polarization and relative membrane excitability O..i 31 NaOH. The concentration of HCO, was (the latter estimated from the inverse in height of thereby adjusted to about 1.5 mxi. Solutions intracellularly injected 300 ms long depolarizing with wrying K + concentrations (between 2.5 square current pulses just able to evoke one and 10 mar) were obtained by adding, to the action potential) in current clamp conditions. standard solution, varying amounts of KCI From the results of such measurements it without osmotic compensation. Solutions with appeared that, by eliminating inward (anomvarj-ing pH were prepared by titrating the alous) rectification in the voltage region of normal standard solution with either HCl (to I , activation (from about -65 mV to about pH 6.5~-6.7)or Tris base (to pH 8.2-8.4). Picro- -95 my), Cs' gives rise to three major effects.
Functional efects of I ,
h
+51
+5-
0-
0.
P -E -5-
223
Y
-5
-
-1Od
-10-100
-80
-60
V(mV)
V
I
40
NmV)
(d) 7
,
-80
-100
Ih block
I
4
1-75
I
I
m"
Fig. 1. (a, b) Relationships between membrane voltage, V, and membrane current, I,, in a slowly adapting receptor (a) and its mathematical model (b) after elimination of any spike current generation and adjustment of the resting voltage to a reference value of -76 mV, in control conditions and in presence of 2 mM Cs+ or following blockage of the I , system, as indicated. The dotted horizontal lines mark the reference (zero) value of the membrane current in resting conditions. (c, d) Membrane voltage, V , and relative membrane excitability, denoted by figures in a slowly adapting receptor beside arrows or vertical lines pertaining to the current traces (Iinj), (c) and its mathematical model (d) whose resting voltages had been set to -76 mV, in control conditions and during exposure to 2 mM Cs+ or blockage of the I , mechanism, as indicated. The relative excitability values were inferred from the inverse in height of 300 ms long depolarizing square current pulses [appearing as dots in (c) and vertical lines in (d)] just able to evoke one action potential
These are : (1) a n increase in the 'resting' input resistance by 107 & 2% (mean & SE, n = 5 ) (Fig. l a ) ; (2) an increase in 'resting' membrane polarization by 7.4k0.3 m V (Fig. 1 a, c) a n d ; (3) a decrease in steady state membrane excitability by 1 7 f 2 % (Fig. lc). The last of these effects is comparable to that previously observed in extracellular measurements on unimpaled cells (cf. Edman et d. 1987b). Hence, the above experiments d o support the notion that, in cells undisturbed by intracellular interventions, the resting voltage is distinctly more negative than -65 mV and, therefore, under the control of an activated I,.
I n addition, they indicate that, in native cells, the resting membrane resistance is kept quite low by an activated Q-channel system. Functionally, such an arrangement would provide for a certain degree of stabilization of the cell's resting polarization and, hence, membrane excitability in the presence of various disturbing influences. Such an effect seems to be particularly valuable for a receptor neurone whose proper function would be greatly dependent on a reasonably constant membrane excitability. T o test this possibility, cells were polarized to about -75 m V and then exposed to a number of disturbances (intense electrical stimulation evok-
(b) i
......
I 0”
a
AV
.-i2 mM Cs‘
4 rnin
e
.......
I
Fig. 2. Post-tetanic hyperpolarizations, AT irelati\-e to resting voltage at - 75 mV), in a slowly ‘idaping receptor following continuous impulse firing a t 40 Hz for 2.5 min in the absence and presence of 2 m u C s . (a), and in its mathematical model following a 15 mM increment of the intracellular \ a * concentration in the presence and absence of a normally functioning I , system (b), as indicated. (c) Recording of impulse activit!-, AV. due to a sawtooth stimulus, ISt,,”.(d) Posttetanic h!-perpolarizations following repetitiw saw-tooth stimulation, as shown in (c), at 1 Hz for 2 min in the absence and presence of 2 m u (AA, as indicated. In both cases, the average impulse fiequenc!- fell from 14 Hz, in the beginning, i o 9 Hz, at the end of the stimulation sequence. The horizontal dotted lines in (a) and (d) . . mark the terminal uhases of imuulse firing and the level of pre-tetanic membrane polarization, respectivel!- The arrows indicate the time for a 90°/, decq of the post-tetanic h!-perpolarizations, L
ing post-tetanic membrane hyperpolarizations, changes in the cells’ physical and chemical environment), while the effects of these disturbances on membrane polarization and/or excitability were measured both in control conditions and after the blockage of I , b!- 2 mx4
2.3 min (Fig. 2a) or prolonged saw-tooth pulses (meant to mimic stimulatory effects during movement-dependent load-compensating feedback acti\ation; cf. Kennedy & Davis 1977) at 1 Hz for 2 min (Fig. 2c, d), both in control conditions and after the blockage of 1, by 2 mM
Ca .
Cs’ .
The actkation on electrical membrane properties in post-tetanic conditions, experiments were performed as shown in Figure 2 . I n these experiments recordings were made of post-tetanic hyperpolarizations, following repetitive impulse firing e\-oked by stimulation \I-ith either short ( 3 ms) square pulses at 40 Hz for
From the results it became evident that, irrespecti\ely of the mode of stimulation, Cs+ causes (1) a 32 2 Z o o (mean SE, n = 4) increase in depth of the post-tetanic hyperpolarization and ( 2 ) a 2.4kO.j-fold lengthening of the first 90 O o of the decay of the hyperpolarization. Also, it became clear that Cs+ has essentially n o effect on the total length of post-tetanic hyperpolarizations. T h e latter observation is not surprising, considering that none of the factors determining the total length of the post-tetanic
Functional e f e c t s of I ,
Iinj -
225
0.8
1.0
14nA
1.o
1.1
_.
2 rnM C s *
lhj
-f
1.0
m.2
-
0.7
T1'o -1 -
4 nA
5 min
Fig. 3. Membrane voltage, V, and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Z1,,J, in a slowly adapting receptor at 14, 18 and 25 "C in the absence (a) and presence of 2 mM Cs+ (b), as indicated. The excitability values were obtained as described in Figure 1 . For a continuous assessment of the cell's input resistance, polarizing voltage transients due to 300 ms long intracellularly injected square current pulses (appearing as a line of dots in the current traces) were recorded throughout the experiment.
hyperpolarization, viz. the degree of intracellular Na+ accumulation during impulse firing and the rate of active Na+ extrusion after the impulse activity, is influenced by Cs' to any extent. The effects of Cs+ on the initial phase of the posttetanic hyperpolarization, on the other hand, are readily explainable on the basis of a suppressed I , action. I t can be concluded, therefore, that an activated IQ contributes to a fast restoration to near-resting values of the membrane polarization after various kinds of impulse firing and, in this way, minimizes the length of functionally dependent variations in receptor excitability (cf. Edman et al. 1987b). The effect of I , on electrical membrane properties in the presence of varying external influences
A number of external influences are able to affect cells with respect to their electrical membrane properties by producing changes in active ion transport or passive ion fluxes, or both. For a study of how such properties are dependent on
an activated IQ,experiments were performed as shown in Figures 3-5. Thus, cells were polarized to about -75 mV and then examined with respect to their membrane polarization, membrane resistance (indicated by the height of voltage transients produced by intracellularly injected 300 ms long hyperpolarizing square current pulses) and relative membrane excitability (reflected by the inverse in height of intracellularly injected depolarizing current pulses, as described above), during exposure to varying external influences (varying temperature, varying pH, varying K+ concentration), both in control conditions (Figs 3 , 4 & 5a) and in the presence of 2 mM c s + (Figs 3, 4 & 5 b). Consistently, it was found that Cs+ leads to increases of the resting input resistance through blockage of the Q-channel system. Also, it was found that, in the presence of Cs+, variations of the external influences give rise to changes in membrane polarization, and in relative membrane excitability, that are distinctly (by factors of 2, or more) greater than in control conditions. This is particularly true for the overshooting
226
.I. Edrnan et al.
V
-/------
5 min
Fig. 4. Xlembrane voltage. I ., and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Zini), in a slon.ly adapting receptor in solutions whose pH n a s set to 6 . 5 . 7.4, and 8.4 in the absence (a) and presence of 2 mM Cs' (b), as indicated. Estimates of the cell's membrane excitability and input resistance were performed as described in Figure 3.
fluctuations in membrane polarization which can be ascribed to temperature- (Fig. 3 ) or pH- (Fig. 4) dependent changes in pump current production (cf. Rakowski et a/'. 1989, Bonting 1970). In the case of estracellular K- (Fig. 3)' it was noted that lowering its concentration leads to increases in membrane polarization and decreases in relative membrane escitabilitv that are similar in size in the presence of Cs- and in control conditions. On the other hand, raising estracellular K- causes decreases in membrane polarization and increases in membrane excitability which amount to much higher !-alues (providing for the initiation of spontaneous impulse firing in three out of four investigated cells) in the presence than in the absence of Cs'. These differential effects are explicable on the ground that the Q-channel conductance and, hence, the Z, production increases with increasing concentrations of extracellular K (Edman & Grampp 1989). Consequently, blocking the laproduction by Cs' will give rise to changes in electrical membrane properties which are more pronounced in higher than in lower extracellular K'. '-
Thus, there is reason to conclude that, in an activated state, the I , system takes part in stabilizing the cell's membrane voltage and membrane excitability in the presence of varying external influences. I n particular, it may do so in the presence of influences that are giving rise to overshooting changes in membrane polarization and/or to increases in the degree of I , activation. .Ilathematiral model studies of the I , system
From the above findings it became evident that the resting membrane voltage is more negative (b! about 10mV) in cells not disturbed by intracellular interventions than in cells impaled n ith micro-electrodes. Not unlikely, this is because the impaling electrode gives rise to the appearance of an unspecific ion leak passing through an imperfect seal between the cell membrane and the electrode wall, and an intracellular accumulation of C1-, if KCl is used as an electrode electrolyte. For a test of this explanation, an investigation was made of nhether or not a previously developed mathematical receptor model (Edman et al. 1986,
Functional efects of I , (a) 5 mM K*
2.5 mM K+
10 mM K +
227
5 mM K'
Fig. 5. Membrane voltage, V , and relative membrane excitability, denoted by figures beside arrows pertaining to the current traces (Iinj), in a slowly adapting receptor in solutions with 2.5 mM K+, 5 mM K+, and 10 mM K+ in the absence (a) and presence of 2 mM Cs+ (b), as indicated. Estimates of the cell's membrane excitability and input resistance were performed as described in Figure 3. Note that, in the presence of Cs+, high K+ elicits spontaneous impulse firing (only the bottom parts of the action potentials are shown), implying an infinitely high membrane excitability. 1987a, b, Edman & Grampp 1989) was able to predict a resting voltage of about -75 mV, if modified values were given to some of its parameters. It was found that a prediction of - 75.1 mV is in fact possible, if the value for the intracellular C1- concentration is lowered from its standard value (46 mM) to 20 mM (cf. Edman et al. 1983) and the leak permeabilities are corrected for the exclusion of an 'impalement'leak channel which has identical permeabilities for Na+, K+, and C1-, each amounting to 80% of the standard (impaled cell's) Naf leak permeability. The latter is achieved by multiplying the standard leak permeabilities for Na+, K+, and C1- by the factors 0.200, 0.974 and 0.627, respectively. The ability of the model to predict an ' unimpaled ' resting voltage of about - 75 mV is displayed in Figure 1(b), where it also appears that the model's I-V relationship in control
conditions and in the presence of a blocked IQ is quite similar to that of the living cell, and that a total blockage of 1, causes the resting voltage to fall by about 7 mV. The latter is also illustrated in Figure l ( d ) where, in addition, it is shown that the simulated membrane excitability (reflected by the inverse in height of depolarizing 300 ms long current pulses just strong enough to evoke one action potential; see above) is decreased by the same relative amount as the membrane excitability of the living cell following a total suppression of IQ. I n Figure 2(b) it is demonstrated that the modified model is able to simulate quite faithfully the time course (initial depth, rate of decay and total duration) of post-tetanic hyperpolarizations due to a 15 mM increment of the intracellular Na+ concentration (as has been seen to result from a 40 Hz impulse firing for 2.5 min; cf.
228
A?.Edninn et
al.
Fig. 6. hlathematicall!- simulated post-tetanic membrane h!-perpolarizations (due to a 15 mM increment of the intracelluiar S a . concentration), A I ., in control conditions (a), and after halving (h) o r doubling (c) the rate con~tantsfor I , ai:ti\-ation. The horizontal dotted lines mark the level of resting polarization a t -75.4 ml-, The arrow indicate the time for a 752; decay of the posttetanic hyperpolarizations For further details, see text.
Edman rt a / . 1986) both in control conditions and after blocking the Zci. T h u s there is reason to assume that the model adopted reproduces the effects of I , in the native stretch receptor neurone in a representative n.a!-. O n the basis of this assumption, it appeared appropriate to use the model for an investigation of the I,, system with respect to its d!-namic properties. These properties seem to be well adjusted, as far as the I , function of voltage and excitability stabilization in the presence of slowl!varl-ing external influences is concerned. However, it might be asked whether an increased rate in activation would provide for a faster deca! of post-tetanic hyperpolarizations and, thereb! , for a Faster restoration of pre-tetanic excitabilit) conditions. For an investigation of this possihilit!-, post-tetanic hyperpolarizations (in response to a 15 mhI increment of the intracellular Na' concentration) were simulated in control conditions and after, respectively, halving and doubling the rate constants of I , activation. T h e results are shown in Figure 6. It is seen that altering the rate constant has essentially no effect on the maximum depth of the post-tetanic hyperpolarization, but leads to distinct changes in the rate of- its initial (about 50°,) decal-. H o w v e r , for the subsequent decay, and the
gradual restoration of the membrane voltage to near pre-tetanic values, the alteration of the rate constants is without significant consequences. I n Figure 6 this is evident from the fact that, already at 7 j o o of the decay of the post-tetanic hyperpolarization (arrows), the time needed for the process is practically the same in the different experimental conditions. I n conclusion, therefore, the view m a j be adopted that the dynamic properties of the ZQ system are well adjusted also for achieving a maximum acceleration of the initial decay of post-tetanic hyperpolarizations which, for their total elimination, are dependent on the transport capacity of the Na-K pump. DISCUSSION T h e present study was undertaken in order to clarify the function of IQ in the lobster stretch receptor neurone. I n summary, it could be shown: (1) that in native unimpaled cells, the resting voltage is under the control of an activated I , and, hence, more negative than in impaled cells, where it is found to be around -65 mV; ( 2 ) that in native cells the presence of an activated I , gives rise to a decreased resting input resistance; (3) that, as a result hereof, IQis able to play a role as a stabilizer of membrane
Functional effects of’ I , polarization and excitability in conditions of varying external (physical and chemical) influences; and (4) that, with its slow dynamic responses, I , is well adapted to its functional purposes and to the rather slow homeostatic effects of the cells’ Na-K pump. Below these points will be discussed separately.
229
voltage occurs at values which provide for a significant activation of Z, and, therefore, must be under the influence of this current. Due to such an arrangement it seems possible that the Z, mechanism is functionally involved in the control of the cells’ membrane excitability. In the following this possibility will be discussed in further detail.
Adjustment qf the resling membrane voltage
’The finding that unimpaled cells have a distinctly more negative resting membrane voltage than impaled cells draws attention to the possibility that, in the latter cells, the resting membrane polarization is determined, not only by genuine membrane currents, but also by an unspecific leak current passing through an imperfect seal around the impaling electrode. In so-called successfully impaled cells, this short-circuiting leak may be quite small and produce no significant reduction of the cells’ input resistance. Nevertheless, it may allow a flow of current which, because of existing driving forces, is mainly carried by Na+ and which, because of the well maintained input resistance, can be shown to give rise to appreciable membrane depolarizations. At the same time, the inward leak is not larger than that it can be counterbalanced completely by the cells’ Na-K pump, whence the preparations are able to stay in seemingly perfect steady state for many hours of experimentation. In view of this apparent stability and the finding that C1- leaking from the impaling electrode affects the resting membrane polarization only marginally because of a rather small transmembrane C1- permeability (Edman et al. 1986), it was presumed in previous studies that the values of the resting membrane voltage recorded would not deviate much from those pertaining to native conditions. It was only with the discovery of Z,, and growing difficulties to identify the functional significance of this current which is activated at voltages apparently outside the cells’ normal working range, that it became necessary to reconsider whether or not measurements of the resting membrane voltage in impaled cells would reflect native conditions. As a result of such reconsiderations, there is now reason to assume that the resting membrane voltage of native unimpaled cells can be as much as 10 mV more negative than that of successfully impaled and apparently well adapted cells. This means that in native conditions the resting
Functional efects of I ,
In the lobster stretch receptor neurone there is essentially no overlap between the voltage regions of I,, and action current activation. Because of this, Z, is unable to play a functional role in the cell’s firing control, as similar currents are seen to do in, for example, heart muscle cells (DiFrancesco & Noble 1989, Denyer & Brown 1990, Bois & Lenfant 1990, van Ginneken & Giles 1991), smooth muscle cells (Hisada et al. 1991), and various mammalian central neurones (McCormick & Paper 1990a, Foehring & Waters 1991). Also, in the lobster stretch receptor, the activation of ZQ is not found to be (significantly) affected by : (1) stretch activation (unpublished observations); ( 2 ) a number of transmitters and hormones of which some are indigenous to crustaceans (Edman 1987), (3) a number of naturally occurring ions (Na+, Ca2+,H+)(Edman & Grampp 1989, 1991); and (4) by temperature (Edman & Grampp 1991). It is not likely, therefore, that the I , system should, as corresponding systems in other preparations [DiFrancesco & Tromba 1988 (heart muscle cells), Coleman & Parkinton 1990 (smooth muscle cells), McCormick & Pape 1990b, Takahashi 1990, Uchimura 1990, Gordon 1991 (mammalian central neurones and peripheral nerve)], be involved in mediating any effects that various external influences might have on the cell’s membrane excitability. I n consequence, there is reason to conclude that 1, becomes functionally involved primarily via its voltage sensitivity. The sensitivity is indeed quite high, such that the current needs only about 30 mV of potential change for a complete activation (Edman et al. 1987) whereas, for example, the Na+ current of the action potential mechanism needs about 60mV for a corresponding reaction (cf. Hille 1984). As a result hereof, I , is able to lower the cell’s input resistance significantly (by maximally 40-60~,,)
230
-4.Edmiin et al. ,
(a>
‘/
Y
I
-100
I
I
-80
-60
t
I
-80
-100
-60
VhV)
VhV)
Fig. 7 . hlathematicall!- simulated relationships between membrane voltage, V , and membrane currents, Ist3 (gated T a - current), I , (gated K* current), I,, (leak current), I , (Q-current), I , (pump current), and I,,, (total membrane current), as indicated, in control conditions (a) and after increasing the maximum Q-channel conductance by a factor of 3 (b). 0, marks the points where the total membrane current is zero at resting voltage levels of -75.0 m y in (a), and -72.2 m y in (b)
near native resting voltage levels and, thereby, to counteract effectivell- any fluctuations in membrane polarization and, hence, membrane escitability that might arise because of a number of disturbing influences. I n the present study such influences were achieved by varying the external concentrations of KT and H-, by changing the ambient temperature, and bj- provoking high-frequency impulse firing for prolonged periods of time. For all of these influences, it was assumed that the!might occur in the lil-ing animal, for example, in connection with feeding [giving rise to estracellular increases in K’ and H+ as a result of ;i less precise regulation of the ionic balance in thc bod\- fluids (cf. Robertson 1960, Parry 1960)], seasonal temperature variations or locomotion through lavers of water with different temperatures, and intense stimulation [giving rise to an increased pump current production (Edman r t i l l . 1987a), extracellular I(+ accumulation (Edman i’t a / . 1983), and intracellular acidification due to transmembrane ion shifts and an accelerated energj- metabolism (cf. Hanseri 198.5 11. From the results of this stud!-, it is clear that any changes in membrane polarization and membrane excitability due to the above kind of influences are reduced markedly by a functioning Zct. It can be concluded, therefore, that the purpose of this current is to stabilize the cells’
membrane excitability in the presence of a number of disturbing influences. This homeostatic effect appears to be of special value for a receptor neurone whose function is to respond to input signals in a consistent way, irrespectively of any varying external circumstances.
.-ldaptation of I ,
to
the Ku-K pump
From the above observations it might be inferred that a particularly important function of the IQ system is to counteract any changes in membrane polarization and membrane excitability that could be due to a n altered pump current production. It might seem conceivable, therefore, that the I , system is particularly well adapted to the cells’ &a-K pump. One indication hereof may be seen in the fact that, in this study, it has been possible to argue on the basis of computer simulations that the dynamics of I , are adjusted so that they provide for no faster and, hence, n o more energy consuming initial post-tetanic voltage restoration that what seems to be justifiable with respect to the transport capacity of the Na-K pump. Another indication is provided in Figure 7(a), which shows computer simulations of stationary relationships between membrane polarization and membrane currents in an unimpaled cell. From these it appears that, in resting conditions (in which the total membrane current, It,,, is
Functional e f e c t s of I , zero), the transmembrane current balance is dominated by an outwardly directed pump current and a somewhat larger inwardly directed I,. T h i s arrangement, in combination with the fact that both currents can be activated to similar maximum values in the presence of normally occurring intracellular Na+ concentrations (Edman et a/. 1986, Edman & Grampp 1989), may be in evidence of the I , system being balanced well against the Na-K pump. With a more powerful IQ system (represented by, for example, a larger maximum Q-channel conductance) it is true that the current could achieve a greater voltage stabilizing effect in resting conditions. This is illustrated by the current-voltage relationships in Figure 7 (b), which shows that a distinctly lower resting input resistance is obtained after a three-fold increase of the maximum Q-channel conductance. However, it is also true that this improvement in voltage stabilization entails a greater consumption of energy that is needed for maintaining a larger Q-channel conductance (a larger number of Q-channels) and a larger pump activity in the presence of a larger transmembrane ion leakage, as is evident in Figure 7(b). Apparently, in the living cell, this cost is too high for the benefits to be attained. Hence, it may be concluded that, in its observed form, the I , system is well adapted to the properties of the cells’ Na-K pump. This conclusion seems to be supported also by the observation that the production of ZQ is markedly smaller in the rapidly adapting lobster stretch receptor neurone which is equipped with a NaK p u m p that is clearly weaker than that of the slowly adapting cell (Edman et al. 1986). A fuller account of this functional arrangement will be given in a subsequent publication. The authors wish to express their sincere gratitude to Ms Kristina Borglid for expert technical and secretarial help. The work has been supported by the Swedish Medical Research Council (project No 2082), and by grants from the Medical Faculty of the University of Lund.
RE F E RE N CE S Bors, P. & LENFANT, J. 1990. Isolated cells of the frog sinus venosus: properties of the inward current activated during hyperpolarization. PJiigers Arch 416, 339-346. Bonting, S.L. 1970. Sodium-potassium activated adenosinetriphosphatase and cation transport. In :
231
E.E. Bittar, (ed.), Membranes and Ion Transport, pp. 257-363. Wiley-Interscience, John Wiley & Sons Ltd, London. Coleman, H.A. & Parkinton, H.C. 1990. Hyperpolarization activated channels in myometrium : A patch clamp study. In: Frontiers in Smooth Muscle Research, pp. 665-672. Alan R. Liss, Inc., New York. Denyer, J.C. & Brown, H.F. 1990. Pacemaking in rabbit isolated sino-atrial node cells during Cs+ block of the hyperpolarization-activated current if. J Physiol429, 401-409. DiFrancesco, D. & Noble, D. 1989. The current if and its contribution to cardiac pacemaking. In J.W. Jacklet (ed.), Neuronal and Cellular Oscillators, pp. 31-57. Dekker, New York. DiFrancesco, D. & Tromba, C. 1988. Inhibition of the hyperpolarization-activated current (i,) induced by acetylcholine in rabbit sino-atrial node myocytes. 3 Physi,o1405, 477-491. Edman, A. 1987. Analysis of a hyperpolarizationinduced membrane current in the crustacean stretch receptor. Thesis (LUMEDW/(MEFF-1019)/ 161). Reprocentralen, University of Lund, Lund, Swede!. Edman, A., Gestrelius, S. & Grampp, W. 1983. Intracellular ion control in lobster stretch receptor neuron!. Acta Physiol Scand 118, 241-252. Edman, A., Gestrelius, S. & Grampp, W. 1986. Transmembrane ion balance in slowly and rapidly adapting lobster stretch receptor neurones.JPhysio/ 377, 171-191. Edman, A., Gestrelius, S. & Grampp, W. 1987a. Analysis of gated membrane currents and mechanisms of firing control in the rapidly adapting lobster stretch receptor neurone. f Physiol 384, 649-664. Edman, A., Gestrelius, S. & Grampp, W. 1987b. Current activation by membrane hyperpolarization in the slowly adapting lobster stretch receptor neuron:. 3 Physiol384, 671-690. Edman, A. & Grampp, W. 1989. Ion permeation through hyperpolarization-activated membrane channels (Q-channels) in the lobster stretch receptor neurone. Pjiigers Arch 413, 249-255. Edman, A. & Grampp, W. 1991. Ion (H+, Ca2+,Co”) and temperature effects on a hyperpolarizationactivated membrane current in the lobster stretch receptor neurone. Acta PhysiolScand 141,251-261. Foehring, R.C. & Waters, R.S. 1991. Contribution of low-threshold calcium current and anomalous rectifier (I,) to slow depolarizations underlying burst firing in human neocortical neurones in aitro. Neurosci Lett 124, 17-2 1. Gordon, T.R., Kocsis, J.D. & Waxman, S.G. 1991. TEA-sensitive potassium channels and inward rectification in regenerated rat sciatic nerve. Nerve & Muscle 14, 640-646.
232
.$. Edmaii et al.
Hanscn, .A.J. 1985. Efect of anoxia on ion distribution in the brain. Phj~siolKtji. 6.5, 101-148. Hille, H, 1084. Iotiil Cha ti t i ~ l s F ~ l . i / [blr i .I Ilwihr(i t i t . ; . Sinauer .Associates Inc. Publishers Sunderland, llassachusetts. Hisada, T., Ordwy. R.\f-., Kirber. tl.T., Singer, J..i. 8; \\-alsh Jr, J.\-. 1991. Hyperpolarization-actii-ated cationic channels in smooth niusclc cells are stretch sensiti! e. Pfiiig17t.s . - l r ~ . / i417, 49.
