Changes in internal ionized Ca2+ and H+ in voltage clamped squid axons

c.sf/ca/&n(lQeQ)io,

401+12

cDLongmanGroupUKLtdlQ8Q

Changes in internal ionized Ca*+ and H+ in voltage clamped squid axons L.J. MULLINS’, J. WHIlTEMBURY2 and J. REQUENA3 ‘Department of Biophysics, University of Maryland School of Medicine, Battimore, Mary/and, USA 2Department of Physiology and Biophysics, Case Western Reserve University School of Medicine, Cleveland, Ohio, USA 3Centro de Biociencias, lnstituto lnternacional de Estudios Avanzados (IDEA), Caracas, Venezuela

Abstract - Squid giant axons were injected with aequorin and tetraethylammonium and were impaled with hydrogen ion sensitive, current and voltage electrodes. A newly designed horizontal microinjector was used to introduce the aequorin. lt also served, simultaneously, as the current and voltage electrode for voltage clamping and as the reference for ion-sensitive microelectrode measurements. The axons were usually bathed in a solution containing 150 mfvl each of Nat, KC, and some inert cation, at either physiological or zero bath Ca2+ concentration ([Ca2q0), and had ionic currents pharmacologically blocked. Voltage clamp pulses were repeatedly delivered to the extent necessary to induce a change in the aequorin light emission, a measure of axoplasmic ionized Ca2+ level, ([Ca2+]i). Alternatively, membrane potential was steadily held at values that represented deviations from the resting membrane potential observed at 150 mbl [K$ (i.e. = -15 mV). In the absence of [Ca2$ a significant steady depolarization brought about by current flow increased [Ca2+]i (and acidified the axoplasm). Changes in internal hydrogen activity, [H+]i, induced by current flow from the internal Pt wire limited the extent to which valid measurements of [Ca2+]i could be made. However, there are effects on [Ca2+]i that can be ascribed to membrane potential. Thus, in the absence of [Ca2$, hyperpolarization can reduce [Ca2+]i, implying that a Ca2’ efflux mechanism Is enhanced. lt is also observed that [Ca2+]i is increased by depolarization. These results are consistent with the operation of an electrogenic mechanism that exchanges Nat for Ca2+ in squid giant axon. The use of electric currents to control membrane potential in a conventional squid axon voltage clamp set-up can cause changes in intracellular hydrogen ion activity ([H+]i) which can, in turn, lead to variations in the level of axoplasmic ionized Ca2’ However, in depolarizations level, [Ca2+]i, [ll. observed in the absence of current flow, such as during a Kt induced depolarization, there seems to be a change in [H+]i that can be ascribed to a Ca2’

entry via Nat/Ca2’ exchange [2]. A rationale for the electrogenic nature of the exchanger is that it has been shown that hyperpolarization can enhance Ca2’ efflux while depolarization can increase Ca2’ influx [3]. Therefore, it would seem that besides the effects of current flow affecting @!l+]ithere are also effects of electrical potential that can be ascribed to Nat/Ca2’ exchange. Since the unwanted effects of current flow 401

CELLCALCIUM

402

affecting [H+]lin squid WOIU GUI be minimized by carefully controlling the experimental conditions, it would be convenient to dissect from a Ca2+ signal obtained during a change in membrane voltage, the relative contributions of Nat/Ca2+ exchange and of artifacts, if any, induced by the current delivered to achieve control of the membrane voltage. Technical difficulties, such as the simultaneous measurement of [H+]i during voltage clamp, and intrinsic complications, such as sensitivity to voltage of the Nat/Ca2+ exchanger, limited somewhat the range of experimental conditions available. For instance, measurements of Ca2+entry with K+ depolarization have been traditionally made by replacing all external Nat (NatO) by extracellular K’ (K 0> a procedure that changes the axon resting potential from -60 mV to 0 mV. Under voltage clamp conditions and in squid giant axons bathed in artificial seawater (ASW, [l&, = 10 mM) it is possible to probe membrane potentials in the range of -100 to +60 mV. However, it would be more convenient to use an artificial seawater containing 150 mM Kt, 150 mM Nat and 150 mM of some inert cation, such as Lit, N-methylglucamine (NMG) or Tris, which should yield a membrane potential of about -15 mV. By using appropriate blockers of Nat and Kt channels it is possible in this solution to explore the effects of changes in the membrane potential of +120 mV (from the holding potential) without applying excessive membrane currents. Moreover, previous research into Na+/Ca2’ exchange mechanism revealed that the presence of 150 mM Na+ in artificial seawater is sufficient for the activation of the Ca2+ efflux [4]; it is an optional value for the overall operation of the exchanger 151. An additional advantage for the use of low [Nat],, stems from the observation that it has been found that high [Na+Joinhibits the ability to see changes in aequoriu light emission resulting from either depolarizing or hyperpolarizing pulses [l]. Another point worth considering is the presence of an intracellular concentration of Nat, [Na’]i, higher than 25 mM in order to observe Ca2’ movement through the axolemma [6]. Thus, axons prestimulated and exposed to low Nat seawater are a convenient ex rimental condition when aequorin is used as a CaY+ indicator.

In the present study (composed of this and a companion paper PI) it has been found that depolarizing voltage clamp pulses of the size and duration of an action potential can enhance Ca2’ entry, while similar but hyperpolarlzing pulses lead to a decrease in the aequorin luminescence even in the absence of Ca2+0, provided the pulse protocol does not significantly change [H+]i. The same proviso holds when the membrane potential is steadily changed. It is found that there is, at most, a change of 0.1 pH uniffmin for pulses of maximum 120 mV amplitude or 6 ms duration at a maximum rate of 30/s. A similar rate of change in pH1 occurs either for steady depolarization in excess of 10 mV or steady hyperpolarizalion iu excess of 20 mV. In general, steady voltage clamp current densities larger than 50 @/cm2 are liable to induce pH related artifacts. It should be stressed that these limits are for tetrodotoxin (77%) and tetraethylammoninm (TEA) treated axons injected with 40 mM of a pH buffer. Under the conditions and limitations of our experiments, we found that there is a dependence of the [Ca2+]ion membrane voltage as a consequence of the modification of the direction and rate of exchange of the Nat/Ca2’ mechanism. It is only when the current necessary to hold the membrane voltage exceeds certain limits, that artifacts induced by the formation of hydrogen ions at the current passing electrode obscure the operation of the Nat/Ca2’ exchange.

Materials and Methods Axons

The experiments were performed at the Marine Biological Laboratory, Woods Hole, MA, USA. The hindmost giant squid axon from the @late ganglion was dissected from live specimens of Lo&o pealei. After careful cleaning for connective tissue, the axon was either mounted in a chamber and cannulated at both ends [8] or, alternatively, the axon was mounted on pedestals at either end. Through slits opened in the axon the electrodes and the microinjector were inserted [9].

CHANGES IN INTERNAL IONIZED Ca AND H IN SQUID AXONS

Microinjection and experimental arrangement

A newly designed horizontal microinjector was used to introduce the aequorin. It also served, simultaneously, as the current and voltage electrode for voltage clamping while it could act as the reference for ion microelectrode measurements. A selected fused quartz capillary (60 mm long,. 0.10 mm i.d x 0.17 mm o.d., Vitro Dynamics, Rockaway, NJ, USA) was coupled to a conventional 1 pl syringe. A length of 13 mm of the outer surface of the quartz capillary was gold covered [lo]. A length of 13 mm at the extreme end of the gold coated capillary was further covered with finely deposited black platinum which then served as the current passing electrode. The internal bore of the capillary filled with 0.6 M KC1was in contact with a Ag/AgCl pellet and acted as an internal voltage sensing electrode. Trials with a deposit of silver instead of gold on the outside of the capillary followed by chloridizing it were found to be inadequate for voltage clamping. Although, this should have given a reversible electrode that minimized Ht production, in practice, it was found that it yielded toxic products. For injection purposes a tiny bubble of air was loaded at the tip of the capillary and then about 0.15 p.l of aequorin (plus TEA) solution was sucked into the capillary. Thus the aequorin solution could be injected back inside the axon once the capillary was properly placed into position. Otherwise, the experimental set-up was as described by Requena et al. [6] plus the additions listed in Mullins et al. [l] for voltage clamping. Extracellular solution changes in the experimental chamber took about 10 s to be Axons were always immersed in completed. seawater flowing at a rate of 1 ml/min. Luminescence was measured as described in Mullins and Requena [l 11. Experiments were performed at 15-16°C. Adequate arrangement of voltage clamp electrode for good spatial control of the membrane voltage over the 13 mm length of the axon exposed to the optical guide was verified by positioning an independent voltage sensing electrode at various distances away from the tip of the voltage control electrode (the latter always being placed very near an extreme of the region of axon being sampled). Comparisons of positive and negative

403

current pulses, in the presence and absence of TTX and TEA, showed that 11 mm away from the tip of the voltage control electrode, 90% of the voltage could be controlled when Nat and K” currents were pharmacologicallyblocked. pHi measurement and limitations

The hydrogen ion sensitive glass microelectrodes were constructed from a previously described batch of glass [2]. The hydrogen microelectrode was introduced into the axon from the extreme opposite to the injector-clamp-electrode and was positioned midway into the sampled axon region. The output of the H’ microelectrode was fed into an Orion 901 ionanalyzer which subtracted the electrode signal from that of the internal voltage membrane ele.ctrode. The latter was isolated and buffeti by a unity gain amplifier (model 48K, Analog Devices, Norwood, MA, USA) which was part of the voltage clamp cinzuitry. The resulting signal was referenced in the 901 to the extracellular bath electrode which, in tnm, was GmilarIy isolated and buff& High impedance leads were always properly shielded and kept to a minimum length. The signal taken from the digital output of the Orion 901 microprocessor, that converts electrode potential to concentration units, could be decoded and shown as an antiogue signal. The D/A converter used produced a signal that was linear with Ht activity over the 10 to 100 nM range @H 8-7). If the axoplasm were more acid than this pH 7 the converter would shift to produce a signal linear over the 100 to 1000 nM H range (pH 7-6). The concentration value was then passed on to a four channel chart recorder. The Ht microelectrode was calibrated before each experiment in standardized pH reference buffer solutions with elevated ionic strength (0.35 M KCl). Values reported herein correspond to the equivalent hydrogen ion activity with no further corrections. As the glass microelectrode impedance is higher than that of the reference voltage electrode there will be, with changes in membrane potential, a change in the output of the membrane voltage electrode faster than that of the H’ microelectrode. The result is that the output of the microelectrode presents an artifact with a time constant estimated to

CELL CALCIUM

404 (mV) Em_5:l

/-\

E

___-

80

photons/s x IO” 1.5

(nM)

051086

Solutions

1.0

0.5

0

balance between the electrode and its reference (such as one resulting from the alkalization due to ammonia added externally) the response time of our ion sensitive glass microelectrodes was limited to 10 to 20 s, a value comparable to that found in the literature [12]. The traces corresponding to hydrogen ion microelectrode readings in the flgtnes appear as discontinuous lines during episodes of voltage change, because, given the experimental limitations encountered, only steady state values for the equivalent hydrogen ion activity ([H+]l) are thought to be precise.

1,,1111111111111

0

5

IO 15 MINUTES

20

Ffg. 1

Calcium and hydrogen ion changes during K’ depolarization. Time coume of aequorin luminescence (bottom trace in photons/s), hydrogen ion activity (middle trace; [H+]i in nlvi) and membrane voltage (top trace; E,,, in mV), simultaneously measured in an axon, briefly exposed to 3 n&I Ca2+ seawater in which all the Na+o was replaced by K+o. Solution changes are at the time indicated by the arrows. Before the beginning of the record, the axon had been stimulated for 30 mitt at 30 imp/s in regular (Na+) artificial seawater in order to raise [Na+]i. The axon had then an initial [H+]i of 80 nh4. Note that after the depolarizing episode the [H+]i is larger than before the test as shown by the extrapolated dotted line. De@uizing test carried out 45 min after injection of aequorin. Axon diameter 500 pm

be of the order

of 2-3 min. This limitation is specially acute for a train of voltage clamp pulses, where there is not a possibility to detect the time course of the production of Ht during each episode. It should be mentioned that for a change in [H+]iin which there is not a major alteration in the charge

The regular artificial seawater used in these experiments had the following composition @I): NaCl (440); KCl (10); CaCh (10); Mgclz (50); Tris-TBS (tris(hydroxymethyl)aminomethane hydroxide, N-tris (hydroxymethyl) methyl 2 aminoetbane sulfonic acid) pH 7.8 buffer (10); EDTA (ethylene diamine tetraacetic acid) (0.1); and NaHC03 (10). Bicarbonate was added to all seawater in order to ensure proton pumping and steady pHi levels [13, 141. Most of the voltage clamp experiments were performed in a depolarizing artificial seawater containing 150 mM of the chloride salt of Nat, Kt and of an inert cation which was either Lit, N-methylglucamine (NMG) or Trls. In other circumstances Nat was sometimes replaced by either Lit, Tris or NMG, but Kt was alwa s maintained at 150 mM. To prepare low CaIt seawater Ca2’, was replaced by Mgzto. Internal injection’solutions were: aequorin (200 @I), made by dissolving 1 mg of salt-free aequorin (kindly supplied by Dr John Blinks, Mayo Clinic, Rochester, MN, USA) into 80 rJ-1of 5 pM K2EGTA (ethyleneglycol bis (~aminoethyl ether)- N,N,N’,N’ tetraacetic d&K salt) plus 3 pM IQEDTA solution at neutral pH (made from glass distilled water): 2.4 M TEA-MOPS prepared by neutralization in the cold of a 40% solution of tetraethylammonium hydroxide (TEA-OH, Alfa Products, Danvers, MA, USA) with enough MOPS (4 morpholine propanesulfonic acid, Sigma Chemicals, St Louis, MO, USA) to yield pH 8.0 f 0.1. The aequorin and the MOPS solutions were made in 0.6 M KCl. Most of the time equal parts of the aequorin and TEA-MOPS were

CHANGES IN INTERNAL IONIZED Ca AND H IN SQUID AXONS

DhOtOnS/S r

160585C

800

I I +40

mV

5min Fig. 2

Results

Changes in depolarization

400

0

Intracellular calcnun release. Time course of aequorin luminescence (in photons/s) in an axon kept for over 60 min in flowing seawater with zero Ca’+ (0.1 mM EGTA) and I50 mM each of Na+, K’ and Tris together with 60 mM Mg’+ plus 1 mM 3,4 diaminopyridine and held under voltage clamp at -10 mV. The holding potential was changed to +4O mV (50 mV steady depolarization) at the time indicated by the first arrow and returned to -10 mV at the second arrow. Note that aeqwrin luminescence increases with a delay with respect to the change in holding potential. TEA-MOPS omitted in injection. Axon diameter 525 pm

405

[Ca2+]i

and

[P]i

during

K+

Fi ure 1 shows the effect on ionized intracellular CaBt level and Ht activity of a Kto depolarization in a squid giant axon. The nerve fiber, injected with aequorin and TEA, had been previously stimulated electrically in low Ca2’ seawater in order to raise its [Na+]i by SOW 35 IIM [ll]. During this stimulation period, [H+]iincreased by some 20 nM, while [Ca2+]ihardly changed from the resting initial value, monitored as an aequorin luminescence of 150 photons/s. The experiment in Figure 1 shows, as originally demonstrated by Baker et al. [15, see also 16, 171, that [Ca2’]i rises several fold with the replacement of all of the Nat0 by Kto in the seawater. It also shows that, concomitant with this change in [Ca2+]i, the [H+]i went from about 110 nM, before depolarization, to about 130 nM Ht, as measured a few minutes after the return to the initial condition. Since no cmrent was delivered to the axon during this depolarization episode, it must be concluded that increments in [Ca2’]iand [H+]iin the axoplasm [2] are not electrically induced artifacts. It should be mentioned that the observed increment in [H+]i must be somewhat attenuated by the MOPS buffer injected (together with TEA and aequorin) which amounted to 25 mM of additional H’ buffering power. Strong depolarization

and

aequorin

signal

in

Ca2+-f/ee seawater

injected. When aequorin alone was injected, Kt If an aequorin injected axon is kept for over 60 min currents were suppressed by the addition of 3,4 in flowing seawater with zero Ca2+ (plus 0.1 mM diaminopyridine to the seawater. All seawaters EGTA) together with 150 mM each bf Nat, K’ and were freshly made and adjusted to 1010 mOsm Tris and with 60 mM Mg2+ plus 1 mM 3,4 Unless otherwise specified, all experiments were diaminopyridine, aequorin light emission declines performed in the presence of 100 nM tetrodotoxin over this period to levels that approached the (TTX; Sigma Chemicals). KC1 solution for the intrinsic resting glow of the injected aeqnorin. As reference electrode was either made from shown in Figure 2, when the voltage clamped axon, recrystallized pure salt (in 500 @viEGTA, 300 pM with its membrane potential held (zero holding EDTA solution) or made from suprapure KC1 current) at -10 mV is depolarized by +50 mV (to (Merck, Darmstadt, FRG) to avoid contamination of change its membraue potential to +40 mV) for 2 the aequorin with Ca2+carried by the salt. min there is very little change in light emission

CELL CALCIUhI

406

CpA/cm*) h

H

3 P (mV)

+g5 I, 3 -25 -IS

ionized Ca2+. The interrx&iate that releases Ca2+ during current flow delivered by an imracell~ Pt wire is a change in [H+]i. Small steady depolarizationin &‘-free, seawater

Nat-pee

Previous studies have shown that depokizing electric currents flowing from a platinized wire can acidify the axoplasm Indeed, phenol red injected into an axon becomes yellowish with a large and prolonged voltage clamp pulse [l]. Also, it is known that in the absence of Na+oand Ca2’, (in the presence of high EGTAo), Nat/Ca2+ exchange movement is inhibited [4]. To further study the effect of depolarization and to separate the operation 0515880 0.4 1 of the Nat/Ca2’ exchanger from intracellular events that might lead to the appearance of ions stored in the a~~pla~m, changes in m’]i and [Ca2+]i MINUTES secondary to steady voltage clamp depolarization Fig. 3 The effect of steady depolarization in the absence of Ca2’ were followed in axons bathed in seawater movement. Time course of aequorin luminescence (bottom tmces containing zero Ca2’ and zero Nat . In Figure 3 the in photons/s), hydrogen ion activity (middle traces; [If]] in &I) nature of the Nat0 substitute was explored in an and holding potential (HP in mV) with its integrated holding axon that had been exposed for 30 mitt to a seawater current (Imin pA/cm2)in an axon held under voltage clamp at -20 containing 150 mM Kt, 300 mM of either Lit or mV and tested with a 20 min long depolarizing change in HP of 5 NMG in zero Ca2’ (2 mM EGTA) and tested with a mV in the absence of Ca2’, (2 mh4 EGTA) and Na+o substituted steady small depolarizing change of 5 mV in by Li+ or NMG (300 mh4). Broken lines in [H’]i trace represent holding potential. The axon showed for both Nat0 the extrapolation of the last set of values for the purpose of substitutes very similar responses: a steady modest reference for each condition, namely Li+ or NMG. Solution rise in luminescence of about 20 photons/mm and a changes are at the time indicated by the armws. Note that dn was, at the most, of the order of 25 @/cm2 during depolarization. negligible reduction in [H+]i (about 4 nM), 15 min Axon diameter 550 pm after the start of the test, Thus, in the absence of Ca2’ movement through the axolemma, the effect of during the first minute to be promptly followed by a a mild steady depolarization is neither a delayed rapid rise. This enhancement in light emission rapid increase in [Ca2’]i as seen in Figure 2, nor is observed after the latent period declines promptly there an appreciable rate of change of intracellular upon repolarization. H’ activity. Since the axon was bathed in zero Ca2’ (EGTA containing) seawater there should have been no Voltageckunp pulses in a high [Nat/i axon external Ca2+ to enter upon depolarization. Thus, the delayed appearance of a substantial amount of The effect of pulsed rather than steady Ca2+ observed during a strong electrical depolarization under voltage clamp was next studied depolarization, and exemplified in Fiy+re 2, must in an axon previously loaded with Nati by come from a different source than Ca o. It must stimulation in order to enhance Nat/Ca2’ exchange have originated from the axoplasm as a result of the activity. The axon was bathed in a seawater current flow necessary to hold the membrane solution containing 150 mM Kt, Nat and NMG and potential at the large depolarized value, which 1 mM Ca2’ clamped at -7 mV. Figure 4 shows that would have released stored Ca2+ that appeared as the light emitted was enhanced during the first

407

CHANGES IN INTERNAL IONIZEDCa AND H IN SQUID AXONS

IO

20

so

so

40

60

70

MINUTES The effect of symmetrical pulses in a high Na+t axon. Time course of aequorin luminescence (bottom trace in photons/s), hydrogen ion activity (middle trace; [H+]tin x&I)and holding potential (HPin mV) with its integratedholding current(I,,, in pAkn2), in an axon held under voltage clamp at -7 mV and tested with trains of pulses applied at the time inte~al comprised by a set of armws. Eithe-rdepolarizing or hyperpolarizing(100 mV amplitude, 5 ms long at 20/s) pulses were delivered. The peak HP value is given close to the HP trace. The fust train sequence was depolarizingfollowed by a hyperpolariziq one. Then a symmetrical a based upon a double pulse protocol. a depolarixing one immediately followed by anotha hyperpolariziq one, was applied. Record ends with a repetition of the initial depolarixing train. Bmken line in IH+]ttmce nqesents the extrapolationof the last value for the putpose of mfexencc. Note that the change in huninescence talrw place synchronouslywith the pulse period mtkr than with a delay as in Figute 2. Axon bathed in 1 mM Ca2’ seawater with 150 mh4 of each K’. Na+ and NMG. Initial ma+&raised (35 mM) by pm-stimulation. Axon diameter575 pm Fig. 4

depolarizing pulse episode (amplitude 100 mV, duration of 5 ms and at 20/s), while it decreased during the hyperpolarizing train that followed The next episode, a double pulse protocol, one depolarizing pulse immediately followed by a hyperpolarizing one of the same absolute amplitude and duration, did not alter the luminescence. Finally the last depolarizing train of pulses produced an effect similar to the first one. These findings for the level of [Ca2+]iare consistent with the idea that the axon has a significant concentration of ma’]i which supports the reverse mode of operation of an electrogenic Nat/Ca2’ exchange 121. That is, in the presence of Ca2’, and during depolarization there is an enhancement of the Ca2’ entry zpromoted by the exchanger, leading to a rise in [Ca +]i. Conversely, hyperpolarization would increase Ca2’ efflux by means of a facilitation of the forward mode of operation of the Nat/Ca2+ exchanger leading to a mduction in [Ca2+]i. In relation to [H+]i,the first depolarization episode increases [H+]l by about 12 nM

while the

following hyperpolarizing train decreases it by about 8 nM. The symr~trical double pulse protocol did not induce a change in IH+]l. Voltageclamppulses in a low [Na+]i axon

Figure 5 shows a protocol similar to that of Figure 4 but in a very fresh axon, presumably with very low wa’]i. The first part shows, in an axon exposed to 3 mM Ca2’, 440 mM Na+ seawater, that a depolarizing train of pulses (75 mV amplitude, 6 ms duration and at 20/s) on the one hand increased [H+]i by about 24 nM while, on the other hand, reduced the li$rt emitted by aequorin. This decrease in [Ca ]i can be argued to be due to the presence of a low @Ja+]isince it has been shown [6] that the inwardly directed passive leakage pathway for Ca2+ entry is the most p&or&ant mode for Ca2’ influx if there is not enough Na+l to activate the NatKa2’ exchange in its reverse mode of operation, that is, exchanging Ca2'o for Na+l. It

40s

”

_

.v”

__ i

-

HP (mv)

_&!F

‘

7 - 131

-.-131

+I9 r-1

0.5EGTA +19 r-r-i-

1

+I9

MINUTES Fig. 5 The effect of [Ca2’J0 in a fresh axon. Time courses of aequorin luminescence (bottom trace in photons/s), hydrogen ion activity (middle trace; [H+]t in n?vl) and holding potential (HP in mv) with its integrated holding current (Im in pA/cm2), in an axon held under voltage clamp at -56 mV, exposed to various [Ca2’], end EaTA and tested with trains of pulses applied at the time interval comprised by a set of armws. Either depolarizing or hyperpolarizing (75 mV amplitude, 6 ms long at 24%) pulses wem delivered. The peak HP value is given close to the HP trace. Broken line in @ITi trace npresents the extrapolation of the last value for the purpose of reference. Initially, [Ca2’10in the regular @la’) seawater was 3 mM snd then a&r the second set of pulses it was changed to nominally zero Ca%. Subsequently various levels of EGTA went added at the time indicated in the Figure. Axon diameter 500 pm

should be recalled that this passive leakage pathway for Ca2” would respond to depolarization by reducing the cation flux. A few minutes later, under the same conditions of 3 mM Ca2’ and 440 mM Nat seawater, a train of hyperpolarizing pulses reduced [H+]iby about 8 nM and increased the light reduced by aequorin. 2P Again this increase in [Ca ]i can be understood if hyperpolarization enhances the Ca2’ leakate flux. Indeed, the effect on this component of Ca ’ entry must have been larger than that on the influx represented by the forward mode of operation of the exchanger, swapping Ca2+ifor Nate. The rest of Figure 5 shows the effects of pulse protocols at various levels of EGTA while the axon was bathed in nominally zero Ca2’ (Nat) seawater. First notice that the aequorin signal dropped almost by a factor of 10 with the removal of Ca2’, while the [H+]lhardly changed. This reduction in [Ca2+]i confirms that the main pathway for Ca2’ entry in this axon was leakage. At nominally zero [Ca2+], and in the absence of EGTA, a hyperpolarizin train of pulses produces no effect, in either [Cak ]i or

[H+]l levels, while a depolarizing train (same frequency, duration and absolute amplitude as those previously applied) enhances the aequorin luminescence and Ht activity. By removing most of the ca2’,, the Ca2+ leak into the axon H;as diminished significantly and a contribution of Nat/Ca2+ exchange reverse mode of operation can now be observed. Indeed, as shown by Baker and McNaughton [18], a modest rate of operation of the exchanger can be easily sustained by the Ca2’, present as a contaminant bound to the extracellular matrix given the very low activation constant for Ca2+, [41. It should be noticed, however, that a reduction in [Ca2+]lwith depolarization might be also a reflection of a drop in the ca2’ efflux mediated by the exchanger. It could represent, as well, a combination of voltage effects on the relative unidirectional rates of &ix for the exchanger. Figure 5 next shows that this rise in [Ca2’]lwith depolarization is reduced to one half within 5 min of the addition of 0.5 mM EGTA to the flowing seawater, while it is almost abolished 5 ruin after EGTA was increased to 2 mM, a concentration that

CHANGES

IN INTERNAL

409

IONIZED Ca AND H IN SQUID AXONS

l25

1

4n

L-25

(LlA/an*) 060466

magnitude and direction of the change in luminescence. Finally, notice that the integrated value of the current necessary to sustain the voltage clamp pulse protocol employed is small, as shown at the top of Figure 4. The integrated value for a pulsing episode was of the order of IO pA/cm2. Ca2’ e$lux and pulsed hypelpolarization

Hplm”I~--L- 140 photons/s I 4

41

Fig. 6 The effect of pulsed hyperpolatization cotuses

of aequorin

luminescence

in 0 [Ca2+lo. Time

(bottom

trace in photons/s),

holding potential (HP in mV; middle trace) and integrated

holding

current (Im in pA/cm2; top trace), in an axon held under voltage clamp

at -20

mV and tested,

EGTA), with hyperpolarizing

in the absence pulses (indicated

was made of pulses -120 mV amplitude, values

are

represents

given

close

during

Train

5 ms long at 20/s.

to the HP trace.

area under the baseline

of Ca2+,, (1 n&f by arrows).

HP

Shadowed

region

test episode.

Axon

bathed in 150 mh4 K+ with 300 mM Na’ plus 1 mM CN.

In Figure 5 we saw that depolarization in an axon, with a reduced or suppressed Ca2+ infhrx through the exchanger, results in an enhancement of [Ca2+]i, through a reduction of the Ca2+ efflux component of the exchanger. Therefore, it was important to explore under similar conditions the effect on [Ca2’]l of a train of hyperpolarizing pukes. In Figure 6, from a holdin potential of -20 mV, an axon immersed in a Ca2g -free (with 1 mM EGTA) seawater containing 150 mh4 Kt and 300 mM Nat, was hyperpolarized by -120 mV with pulses of 5 ms duration at 201s. A reduction in the Ca2+ signal is observed associated with hyperpolarization. Since in the absence of [Ca2’], there can not be much of a Ca2’ influx to be affected, the decrease in [Ca2’]i must be almost entirely due to a change in Ca2’ efflux. Note that the change in luminescence takes place synchronously with the pulse period rather than with a delay as in Figure 2. An alternative explanation seems unlikely namely that Ca2+ leaks out of the fiber following its new electrochemical gradient ([Caztlo = 0 mM) due to some sort of membrane damage.

Axon

diameter 525 pm

would have removed all Ca2+ from the extracellular space and, thus, considerably diminished any C!a2’ influx component of the exchanger. This observation indicates that under this condition of zero [Ca2+],, provided by high [EGTAlo, increases in [Ca2’]l produced by pulsed depolarization mainly reflect a reduction of the Ca2+ efflux component of the exchanger operation. Figure 5 also shows that during each de~~larizin+g episode in the nominal absence of Ca o, [H ]i increased about 20 nM. Notice that these changes in [H+]i are present regardless of the

The effect of steady changes in holding potential The effect of changes in holding potential on [Ca’+]i was studied in two axons examined in detail for their [H’]i response during either pulses or steady voltage clamp episodes. The results are shown in Table 1. Notice that the rate of change in pH when hyperpolarizing pulses are applied (100 mV pulses, 4 ms duration at 20/s) is O.OMnin, while steady hyperpolarization of 20 mV resulted in a rate of pH change of O.lO/min. In contrast, steady depolarization (20 mV) produced a rate of pH change of -0.3O/min, and for pulses (100 mV amplitude, 4 ms duration at 20/s) an average of -0.095 pH/min was observed.

410

CELL cAL.cTuM

Discussion The results of simultaneous measurement of intracellular Ht and Ca2’ in axoplasm during a Ca2’ load are controversial since, for instance, Mullins et al. [2] have shown a concomitant increase in H’ with a Ca2’ entry, while Baker and Honerjager [19] have claimed the opposite. More recently Baker and Umbach [20] saw an increase in [H+]ibut only when mitochondrial buffers were impaired. Mullins et al. [2] studied the effect of Kt depolarization on pHi and [Ca2+]i,Vassort et al. [21] looked at the effect of chemically induced pH changes on Ca2’ levels. However, there has not been a study of [H+]i and [Ca2+]iduring electrically induced changes in membrane voltage. Technically, the experiments am diftlcult because of the simultaneous use of high impedance electrodes and the control of bansmembrane potential. Our H’ determination set-up responded to a voltage transient with a time constant of about 2 min. Although some electronic refinement could be introduced to improve the time response, it does not seem possible at present to study effects on [H’]i

produced during rapid changes in voltage, such as those brought about by voltage clamp pulsing. In this study, the axons were frequently depolarized with 150 mM [Kt]Oin order to explore a larger range of membrane potentials. The combination of voltage changes, high &lo and zero [Ca2’lo does not seem to damage the axolemma, since axons could be repeatedly depolarized and repolarized without signs of leakiness to Ca2’. Haydon et al. [22] have reached a comparable conclusion during voltage clamp experiments in Loligoforbesi axons. Intracellularcalcium balance The present results indicate that the variation in

[Ca2+]ibrought about by membrane vo1ta.e changes is not a consequence of alterations in IH ]i since as shown in Figure 5, for instance, increments in Ht in the axoplasm do not necessarily lead to increases of [Ca2+]i, as monitored by aequorin. Changes in [Ca2+]idue to voltage excursions must reflect the operation of membrane based mechanisms, either the Nat/Ca2’ exchanger or a passive conductance

Table 1 Changes in pH induced by current flow Membrane voltage Axon

lCa2+10

number

tmW

050785

0

INa+lot hw 150

Hold hW

-5

Pulse’ (mv)

steady (mv)

PH min

if0 hhfls)

+1&I

-0.12

0.14

test 2 test 3

+100

-0.09 -0.13

0.17 0.38

test4

-100 -100 -25

0.04 0.04 0.11

-0.07 -0.10 -0.30

-35

-0.09 0.02 0.02 0.07

0.22 -0.09 -0.09 a.15

-5 -5

-0.16 -0.14

0.30 0.47

-40 0

0.02 -0.04 0.05 -0.08

-0.10 0.09 -0.12 0.13

t100

test 5 test6 050685 test 2 test 3 test 4

3

-17

t100 -100 -100

-15

test5 test6 test 7 test8 test9 test 10

3

150

-17 -20

-100 +100

‘seawater was 150 mM IC and 300 mM NMG mims fNa+J, ‘pulses4ms dudon at 20/s

CHANGES IN INTERNAL IONIZED Ca AND H IN SQUID AXONS

pathway, namely, the so-called Ca2+ leak. In our experiments, membrane potential was altered either by electric current flow or by changes in [I&. Depolarization would tend to enhance Ca2’ intlux through the exchanger provided there is enough [Na+]i (25-35 mM) to activate its reverse reaction. This would also be favored by the absence of LNa+lo. Moreover since Ca2’ influx throu h the 8 exchanger depends’upon the presence of Ca2 0 with an apparent activation constant of the order of 0.5 mM [4], the incomplete removal of Ca2+0,such as zero Ca2+in the absence of EGTA, could still leave enough Ca2’ attached to the extracellular matrix to sustain some measure of Ca2’ influx [IS]. In the absolute absence of Ca2t0, the seawater being made Ca2+-free with high [EGTAJ, the changes in membrane potential in a hyperpolarizin direction can be expected to influence mainly Ca2f efflux through the Nat/Ca2’ exchanger and, indeed, in Figure 6 as a consequence of a train of hyperpolarizing pulses, a decline in aequorin light emission was seen, a measure of a drop in [Ca2 i. Conversely, depolarization will mainly reduce Ca1’ efflux and thus result in an enhancement of [Ca2+]l, as shown in several of the experiments referred to herein. In the presence of moderately high concentrations of Ca2+0, the situation is more complex because hyperpolarizationincreases passive leak of Ca2+ into the axon as well as Ca2+ efflux while it decreases Ca2’ influx . by the Nat/Ca2’ exchanger [23] In addition Ca2’ influx could also occur through’ Ca2+ channels. However, in the present experiments, this path was minimized by the use of low levels of Ca2+o. It should be noted that the decline in aequorin glow with pulsed depolarization detected in a fresh axon, presumably with low @Ja+]i,is comparable to that observed under very similar experimental conditions but for Kt depolarizations. The explanation for the phenomenon described in Figure 4 is thus the same as that put forward by Requena et al. 161. Hydrogen and calcium

We have been able to determine the effect of voltage clamp pulses on Ir_r’]and the extent that

411

changes in this parameter alter [Ca’+]i. For ~xrlses about 100 mV in amplitude and 5 ms duration (at 20/s) sustained during 2 min. a change in pH of the order of 0.2 units at most is observe for steady depolarization, a much smaller membrane potential change can be tolerated (of the order of 5 to 10 mV) during a similar period. Within these limits, however, it has proved feasible to dissect the relative contribution of change in [H+]i to Ca2’ in axoplasm. It was shown that Ca2’ entry itself leads to a small acidification even in the absence of a voltage clamp current. There is neither change in Ca2’ signal nor a change in pHi if symmetrical voltage clamp pulses of equal amplitude and duration but of alternating polarity am applied to an axon (as shown in Fig. 4). This finding suggests that (a) both Ca2’ influx and Ca2’ efflux are affected by membrane potential changes and these changes are roughly of the same order of magnitude, and (b) the electrochemical reactions that lead to axoplasmic acidity for depolarizing pulses can be counteracted by pulses of the opposite polarity that lead to a decrease in axoplasmic acidity. It is quite possible that in the absence of Ca2’, there is a release of Ca2+from axoplasmic stores as a result of current flow mediated by an increase in Ht in axoplasm This hypothesis implies that some of the mechanisms in the axoplasm that buffer and store Ca2+ are capable of exchanging protons for stored Ca2’. In contrast, the experiment where Ca2t0 enters the axon in the absence of current flow, resulting in an increase in [H’]i, implies that Ca2’ is taken up by intracellular buffers while H+ is released, but this point has been argued before [2]. If the external bathing solution is Nat-free and Ca2+-free,where Ca2’ iuflux is nil and Ca2+efflux through the Nat/Ca2’ exchanger is minimiz a 9 +]i small depolarization results in a slow rise in [Ca (see Fig. 3). This presumably reflects a continuous leak of Ca2+from internal stores into the axoplasm. Our results, in addition to defining the limits of electrical current flow that can be used with negligible changes in intracellular pH, emphasize the interplay between [Ca2+]i and lH’]i, and compliment those found in the following py+~ [7], where it is shown that the dependence of Ca entry generated by repetitive voltage clamp pulses is a

CELL CALCIUM

412

steep function of [Na+]i. The results presented herein lend further support to the notion that Nat/Ca2+ exchange is sensitive to membrane voltage no matter whether this is applied as a constant depolarization or as an intermittent one.

Acknowledgements The authors wish to thank Dr Guillermo Whitmmbury for critical reading of the manuscript and the director and staff of the Marine Biological Laboratory, Woods Hole, MA, USA for facilities placed at their disposal. This work was partially supported by grant No. Sl-1147 from CONJCJT (Caracas, Venezuela) and ROl NS 17718 from National Institutes of Health, Bethesda, USA.

References 1. Mullins LJ. Requena J. Wbittembury J. (1985) Ca’+ entry in squid axons during voltage clamp pulses is mainly Na’/Ca*+ exchange. Pmt. Natl. Acad. Sci. USA, 82, 1847-1851. 2. Mullins LJ. Tiffert T. Vassort G. Whittembury J. (1983) Effects of internal sodium and hydrogen ions of external calcium ions and membrane potential on calcium entry in squid axons. J. Physiol., 338, 295-319. 3. Mullins LJ. Btinley Jr FJ. (1975) The sensitivity of calcium efflux from squid axons to changes in membrane potential. J. Gen. Physiol., 65, 135-152. 4. Requena J. Mullins LJ. (1979) Ca2+movement in nerve fibers. Q. Rev. Biophys., 12,371-460. 5. Requena J. (1978) Ca2+efflux from squid axons under constant Na+ electrochemical gradient. J. Gen. Physiol., 72, 443-470. 6. Requena J. Mullins LJ. Whittembury J. Brinley Jr FJ. (1986) Dependence of ionized and total Ca2’ in squid axons on Nafo-free or high K+, conditions. J. Gen. Physiol., 87, 143-159. 7. Requena J. Whittembury J. Mullins LJ. (1989) Ca2+entry in squid axons during voltage clamp pulses. Cell Calcium, 10,413-423. 8. Requena J. DiPolo R. Brinley Jr FJ. Mullins LJ. (1977) The control of ionized calcium in squid axons. J. Gen. Physiol., 70, 329-353. 9. DiPolo R. Bezanilla F. Cafuto C. Rojas H. (1985) Voltage dependence of the Na+/Ca + exchange in voltage-clamp, dialyzed squid axons. Na+-dependent Ca2’ efflux. J. Gen. Physiol., 86, 457-478.

10. VanWagoner D. Whittembury J. (1985) An improved current electrode/injection capillary for large cells voltage clamp. J. Physiol., 365, 8P. 11. Mullins LJ. Requena J. (1981) The “late” Ca” channel in squid axons. J. Gen. Physiol., 78, 683-700. 12. Boron WF. DeWeer P. (1976) Jntracellular pH transients in squid giant axons caused by CGz, NHs and metabolic inhibitors. J. Gen. Physiol., 67, 91-112. 13. Boron WF. Russell JM. (1983) Stoichiometry and ion dependencies. of the.intracelhtlar-pH-regulating mechanism in squid giant axons. J. Gen. Physiol., 81, 373-399. 14. Boron WF. (1985) Intracellular pH regulation mechanism of the squid axon. Relation between external Na+ and HCC&dependences. J. Gen. Physiol., 85, 325-346. 15. Baker PF. Hodgkin AL. Ridgway EG. (1971) Depolarization and Ca2’ entry in squid axons. J. Physiol., 218,109-155. 16. Baker PF. Meves H. Ridgway EG. (1973) Effects of manganese and other agents on the calcium uptake that fo!low depolarization of squid axons. J. Physiol., 231, 51 l-526. 17. Baker PF. Meves H. Ridgway EG. (1973) Calcium entry in response to maintained depolarization of squid axons. J. Physiol., 231,527-548. 18. Baker PF. McNaughton P. (1978) The influence of extracellular Ca2’ binding on the Ca2’ et?lux from squid axons. J. Physiol., 276, 127-150. 19. Baker PF. Hone@iger P. (1978) Infhtence of CGz on level of ionized Ca2’ in squid axons. Nature, 273,160-161. 20. Baker PF. Umbach GE. (1985) Calcium buffering in axons and axoplasm of L&go. J. Physiol., 383, 369-394. 21. Vassort G. Whittembury J. Mullins LJ. (1986) Increase in internal Ca2’ and decrease in internal H’ are induced by general anesthetics in squid axons. Biophys. J., 50, 11-19. 22. Haydon DA. Requena J. Simon AB. (1988) The potassium conductance of the resting squid axon and its blockade by clinical concentrations of general anaesthetics. J. Physiol., 402,363-374. 23. DiPolo R. Rojas H. Beauge L. (1982) Ca2+entry at rest and during prolonged depolarization in dialyzed squid axons. Cell Calcium, 3, 19-41. Please send reprint requests to : Dr J. Requena, Centro de Biociencias, Jnstituto Jntemacional de Estudios Avanzados (IDEA), Apartado 17606 Parque Central, Caracas 1015A, Venezuela Received : 5 January 1989 Accepted : 14 March 1989