Spontaneous calcium release from sarcoplasmic reticulum. Effect of local anesthetics.

Spontaneous calcium release from purified light sarcoplasmic reticulum has been previously described (Palade, P., Mitchell, R. D., and Fleischer, S. (1983) J. Biol. Chem. 258, 8098-8107) and found to be distinct from several other forms of Ca2+ release. Ca2+ release occurs after a lag period following active Ca2+ preloading and depletion of extravesicular Ca2+. In the present study, we find that local anesthetics inhibit spontaneous Ca2+ release, in a time-dependent manner, varying considerably in the preincubation time required to exert maximal effect. At pH 7.0, hydrophilic and mostly charged local anesthetics, such as procaine, procainamide, and N-(2,6-dimethylphenyl carbamoyl methyl)triethyl ammonium bromide, inhibit Ca2+ release only after long preincubations (hours), whereas more hydrophobic local anesthetics are effective after only a short incubation (minutes) with sarcoplasmic reticulum. The more hydrophobic anesthetics take somewhat longer to reach equilibrium, as studied by inhibition of unidirectional Ca2+ efflux, and there is a direct relationship between hydrophobic partition coefficient and half-time to reach equilibrium. Agents known to inhibit permeability pathways for monovalent cations i.e. K+ channel blockers (decamethonium and n-dodecane-1, 12-N,N,N,N',N',N'-hexamethyl-bis-ammonium) or the anion blocker (4,4'-diisothiocyanostilbene-2,2'-disulfonic acid), do not inhibit spontaneous Ca2+ release. Carbonyl cyanide m-fluorophenylhydrazone, a protonophore, and gramicidin D, a monovalent cation ionophore, have no effect on Ca2+ release whether local anesthetics are present or not, while the Ca2+ ionophore A23187 relieves inhibition of Ca2+ release by local anesthetics. Ruthenium red does not inhibit spontaneous Ca2+ release. These findings suggest that the binding site(s) for local anesthetics is located on the inner face of the sarcoplasmic reticulum membrane and that local anesthetics interact directly with a Ca2+ channel rather than with other permeability pathways which might indirectly influence Ca2+ channel gating.

methy1)triethyl ammonium bromide, inhibit Ca2+ release only after long preincubations (hours), whereas more hydrophobic local anesthetics are effective after only a short incubation (minutes) with sarcoplasmic reticulum. The more hydrophobic anesthetics take somewhat longer to reach equilibrium, as studied by inhibition of unidirectional Ca2+ efflux, and there is a direct relationship between hydrophobic partition coefficient and half-time to reach equilibrium. Agents known to inhibit permeability pathways for monovalent cations i.e. K+ channel blockers N,N,Nf ,N', or the anion blocker (4,4'-diisothiocyanostilbene-2,2 '-disulfonic acid), do not inhibit spontaneous Ca2+ release. Carbonyl cyanide m-fluorophenylhydrazone, a protonophore, and gramicidin D, a monovalent cation ionophore, have no effect on Ca2+ release whether local anesthetics are present or not, while the Ca2+ ionophore A23187 relieves inhibition of Ca2+ release by local anesthetics. Ruthenium red does not inhibit spontaneous Ca2+ release.
These findings suggest that the binding site(s) for local anesthetics is located on the inner face of the sarcoplasmic reticulum membrane and that local an-esthetics interact directly with a Caz+ channel rather than with other permeability pathways which might indirectly influence Ca2+ channel gating.
As reported recently (43); spontaneous Ca2+ release3 requires a critical extent of calcium preloading in the presence of high phosphate levels and takes place, after a lag period, when the free [Ca], has been reduced to the submicromolar range. The experiments presented here deal with the influence of local anesthetics on spontaneous Ca2+ release from skeletal SR. The data are consistent with the existence of a Caz+ efflux pathway (Ca2+ channel) which is selectively blocked by local anesthetics from the inner face of the SR membrane. A preliminary report has appeared (44).

Chem.
Throughout this paper, Ca2+ release is used to refer to net loss of calcium by SR vesicles and CuZ' efflux refers to unidirectional calcium efflux from the SR vesicles.

T E R T I A R Y A M I N E S
Quaternary amines, QX 572 and QX 314, are permanently in a charged, cationic form. QX 572 is much more hydrophobic than QX 314 with an estimated permeability coefficient of lo" cm/s, as compared with 2 X lo-' cm/s for QX 314 (45).
Preparation of Isolated SR and Preincubation with Local Anesthetics-Light SR, derived mostly from longitudinal sarcoplasmic reticulum of rabbit fast-twitch skeletal muscle, was isolated and purified as previously described (46). SR vesicles in 10% sucrose, 5 mM K Hepes, pH 7.2, were stored in small aliquots at -70 "C until used. The biochemical characteristics of this SR fraction were described in Mitchell et al. (46) and in the first paper of this series (43).
Preincubation of SR with local anesthetics was carried out at room temperature (23-24 "C) in the assay buffer, either Na-or K phosphate, pH 7.0, as detailed in the figure legends. Shortly before commencing assay, other ingredients were added (see below) and the reaction was started with addition of 1 mM Na, ATP.
The reaction was started by adding 20-30 pg of SR protein. The loading of 3-4 sequentially added 50-nmol CaC1, pulses was monitored by following differential absorbance changes of antipyrylazo I11 at 710-790 nm or of arsenazo I11 at 660-740 nm in a Hewlett-Packard 8450 A spectrophotometer. Antipyrylazo 111 was used in most of the experiments because it gives better signal-to-noise Characteristics and somewhat higher Ca2+ loading rates than arsenazo I11 (43). The extent and rate of spontaneous Ca2+ release were determined with respect to the absorbance change brought about by the last Ca2+ pulse. When a drug was administered after the completion of Ca2+ loading but before release, a recalibration with Ca" was required. In the experiments carried out with QX 314 and QX 572, Ca2+ fluxes were followed using only arsenazo I11 because of the interaction between these drugs and antipyrylazo 111. When ethanolic solutions of drugs were used, equal amounts of alcohol were added to the control (0.5-1%).
Caz+ Efflux Assays-Unidirectional Ca2+ efflux3 was measured in  Table I.  the presence of either quercetin or EGTA to inhibit the Ca2+ pump. The quercetin method, involving only spectrophotometric assay, was more convenient than the EGTA method. In Method A, Ca2+ efflux was measured by adding 100 p~ quercetin (5 pl of 20 mM stock solution) after the completion of CaZ+ loading, at the beginning of the lag period. The assay medium was the same as that described above. A final pulse of Ca2+ was added at the end of each experiment in order to recalibrate the absorbance changes of the Ca2+ indicator (see the shift in the base line, inset of Fig. 7). Quercetin is known to inhibit the forward and backward reaction of the Ca2+ pump (47) and not to inhibit Ca2+ release (43, 47). 100 p~ quercetin inhibited Ca2+ loading almost completely and Ca2+ reuptake after spontaneous Ca2+ release completely (see inset of Fig.  7). Quercetin did not change the length of the lag phase preceding spontaneous Caz+ release. The rate and extent of Ca2+ efflux, as well as inhibition by local anesthetics, measured using quercetin were in good agreement with those measured using EGTA and 45CaC12 (see Method B and legend to Fig. 7). Thus, under the prevailing experimental conditions, quercetin does not enhance unidirectional spontaneous eaz+ efflux from purified light SR (Fig. 9C in Ref. 43).
In Method B, ea2+ efflux was also measured by means of isotope techniques as described previously (43). The assay medium was similar to that described above except for the deletion of quercetin and the final volume was 2 ml. A small aliquot of a high specific activity 45CaC12 was added before the loading was started. The loading of 300 nmol of CaCI, by 45-60 pg of SR protein was monitored following the absorbance changes of either antipyrylazo I11 or arsenazo 111. At the completion of the loading, 1 mM EGTA was added (estimated resulting free [Ca], = 3 X lo-@ M), and 0.1-ml aliquots were withdrawn over the next few minutes. Each aliquot was rapidly filtered using a microfilter kit and 0.2 pm nitrocellulose filters (Schleicher and Schuell, Keene, NH). Filtrates were counted to monitor isotope released from the SR vesicles.
In Method C, Ca2+ efflux at fixed free [Ca], was determined by means of isotope techniques as described above (Method B). At the completion of Ca2+ loading, 40 p1 of concentrated Ca-EGTA solution (250 mM EGTA) was added to the standard assay medium (2 ml) so that free [Ca],, during Ca2+ efflux, was kept constant at either or 10" M. Aliquots (0.1 ml) were withdrawn and filtered and filtrates counted for radioactivity. Ca-EGTA solutions were prepared with the aid of a computer program (43) using the association constants and calculations reported by Fabiato and Fabiato (48).

RESULTS
The effect of three categories of local anesthetics on spontaneous calcium release from sarcoplasmic reticulum was studied. In the range of anesthetic concentrations used, Ca2+ release was selectively inhibited i.e. Ca2+ loading was affected only slightly, or not at all (Table 11).
Tertiary Amine Local Anesthetics-A typical control experiment illustrating spontaneous Ca2+ release3 is shown in Fig.   2a. After contaminating Ca2+ is sequestered by SR (arrowhead), three consecutive 50-nmol CaCI, pulses were administered (arrows). Spontaneous Ca2+ release occurred at the completion of Ca2+ loading, after a lag period of about 75 s, followed by Ca2+ reuptake.
Increasing concentrations of lidocaine lead to progressive inhibition of Ca2+ release (Fig. 2, b-d). Lidocaine was administered just prior to the first CaC12 pulse ( L in Fig. 2), 5-6 min before the onset of spontaneous release. At lidocaine concentrations of 6, 9, and 15 mM, the Ca2+ release rate decreased approximately 15, 40, and 84%, respectively.
Preincubation of SR with the anesthetics enhanced their inhibitory action on Ca2+ release (Fig. 3). The onset of the action of lidocaine was relatively rapid, but some anesthetics were much slower in taking effect. Procainamide and procaine, both less hydrophobic than lidocaine and mostly in the charged form (Table I), required long preincubations with SR in order to inhibit Ca2+ release. After 4 h of preincubation at room temperature, procainamide or procaine (10 mM of each)

TABLE I1
Effect of local anesthetics on Ca2+ loading and ea2' release rates Ca2+ loading and Ca2+ release were measured as described under "Experimental Procedures." SR vesicles were loaded with 5.4-6.8 pmol of Ca2+/mg of protein. K, is defined as the concentration of local anesthetic which inhibited the Ca2+ release rate by 50% when added just prior to initiation of Ca2+ loading, and was calculated from a linear plot of % inhibition of control activity uersu-s anesthetic concentration. Ca2+ loading rates were measured at K, concentration of anesthetic. The values are given as the mean -+ S.D. for the number of different SR preparations shown in parentheses. Some anesthetics required prolonged preincubation for inhibition. For example, Caz+ release was inhibited 50% after 2-h incubation with 10 mM procaine "In preliminary experiments, dibucaine was found to markedly inhibit CaZ+ loading rate. Whether this finding was due to inhihition of Ca2+ loading activity (29, 31, 34, 35, 71) or increase of Ca2+ permeability (11,12) was not further investigated. Nonetheless, when dibucaine (0.5-1 mM) was added after Ca2+ loading, the rate of Ca2+ release was decreased (by 35%) but the extent of release was increased (1.8 times). under "Experimental Procedures" in K medium, using antipyrylazo 111 as the calcium indicator. SR (23 pg of protein; arrowhead) was first added, then three consecutive 50-nmol CaC12 pulses (arrows) were administered. Spontaneous Ca2+ release was observed after a lag period subsequent to removal of the third pulse of CaCIZ from the medium. In b, c, and d, lidocaine ( L ) was added prior to the first CaClz pulse. a, Control trace (2.85 pmol of Caz+ released/min.mg of protein); b, c, and d. additions of 6, 9, and 15 mM lidocaine, reduced the rate of Ca2' release by 15, 40, and 84%, respectively. release were measured in Na medium. SR (28 pg of protein) was loaded with CaC12 administered in three 50-nmol pulses. In the case of procaine (0, 10 mM) and procainamide (0, 10 mM), the first time point was obtained by adding the local anesthetic prior to Ca" loading (see legend to Fig. 2). Subsequent time points were obtained after preincubation (see "Experimental Procedures"), as indicated on the abscissa. Incubation time is the effective interaction time between SR and local anesthetics before Ca2+ release takes place. In the case of lidocaine (0, 10 mM), the first four time points were obtained by adding the local anesthetic after uptake of the third Ca2+ addition, before the third Ca2+ addition, before the second Ca2+ addition, and before the first Ca2+ addition, respectively. Other time points were obtained after short preincubations up to 15 min. Rates of Ca2+ release are expressed as percentage of activity of the control which was incubated without anesthetic for the same length of time. The average control rate of Ca2+ release was 2.80 pmol of Ca*+/min.mg of protein and decreased by approximately 15% after prolonged incubations (3-4 h).
Tetracaine and SKF 525-A, both hydrophobic and mostly charged (Table I), were found to inhibit Ca2+ release when administered just prior to Ca2+ loading. Under these conditions, the rate of Ca2+ release was inhibited by 50% with either 0.49 mM tetracaine or 47 PM SKF 525-A ( Table 11). The inhibition of Ca2+ release by tetracaine (Fig. 4a) or SKF 525-A (Fig. 4 b ) was sharply time-dependent. The first time points in Fig. 4 were obtained by adding anesthetics after the completion of Ca2+ loading so that the inhibitory effect was more marked after a few minutes of preincubation.
Etidocaine, bupivacaine, prilocaine, and mepivacaine inhibited Ca2+ release when administered in the millimolar range just prior to Ca2+ loading (Table 11). Prilocaine, which is less hydrophobic but with a higher percentage of free base form than procaine (Table I), was found to be effective without a long preincubation period, whereas procaine required pro- longed preincubation to be effective. The onset of action of tertiary amines varies with both hydrophobicity (E') and the percentage of free base form (see Table I).
Neutral Local Anesthetics-Benzocaine in millimolar concentration inhibited Ca2+ release. When benzocaine was added just prior to Ca2+ loading, 50% inhibition was obtained a t 4 mM (Table 11). Benzocaine also exhibited a sharp time-dependence for its inhibition (data not shown). In the same range of concentration, ethyl-o-toluate, a steric analog of benzocaine (17), was found to be ineffective (data not shown).
Quaternary Amine Local Anesthetics-QX 572, which is membrane-permeant despite being permanently charged (45, 49) (Table I), was found to inhibit Ca2+ release when administered in the submillimolar range just prior to Ca2+ loading (Fig. 5, b-d and Table 11). QX 572 also decreased the rate of Caz+ loading more than the other anesthetics used in this study (see Table 11), and when added after the completion of Ca2+ loading (Fig. 6, first time point), it enhanced rather than inhibited Ca2+ release.
QX 314, which like procaine and procainamide is poorly membrane-permeant, was found to be a much less effective inhibitor of spontaneous calcium release unless a long preincubation was performed. Fig. 6 shows that QX 572 at 0.25 FIG. 5. Inhibition of Ca2+ release by QX 572. Ca'+ loading and Ca2+ release were measured in Na medium using arsenazo I11 as the calcium indicator. SR (23 pg of protein) was added to the assay medium, followed by three consecutive 50-nmol CaC12 pulses. a, Control trace (2.4 pmol of Ca2+ released/min . mg of protein); b, c, and d, 100, 200, and 300 PM &X 572, respectively. QX 572 was added before Ca2+ loading (arrow). In this experiment, the Ca2+ loading rate was decreased to 78% of the control with 200 pM &X 572 (c and Table II), the concentration which inhibited the Ca2+ release rate by 50%.  . 6. Time dependence of inhibition of Ca2+ release by &X 572 and &X 314. Ca2+ loading and Ca2+ release were measured in Na medium using arsenazo I11 as the calcium indicator. SR (28 pg of protein) was loaded with three 50-nmol CaClz pulses. In the case of &X 572 (E, 0.25 mM), the first four time points were obtained by adding the local anesthetic after uptake of the third Ca2+ addition, before the third Ca2+ addition, before the second Ca2+ addition, and before the first Ca2+ addition, respectively. For &X 314 (0, 0.25 mM; 0, 2.5 mM; 0 , 5 mM), the first time point was obtained by adding the local anesthetic prior to Ca2+ loading. All other time points were obtained after preincubation, as indicated on the abscissa. mM strongly decreased Ca2+ release within a few minutes of preincubation, whereas 0.25 mM &X 314 inhibited the rate of Ca2+ release only 30% after 5 h of preincubation. Higher concentrations of &X 314 were more effective (Fig. 6).
Time Dependence of Inhibition of Ca2+ Efflux by Local Anesthetics-The inhibition of Ca2+ release was found to be time-dependent (Figs. 3,4, and 6). This finding would suggest that local anesthetics must cross a hydrophobic barrier to exert their inhibitory action on spontaneous Ca2+ release from SR. Thus, an attempt was made to discriminate between equilibrium and rate effects i.e. to distinguish factors related to anesthetic binding to a site from those pertinent to reaching the site. A Ca2+ efflux assay was utilized, instead of the Ca2+ release assay, thereby measuring unidirectional Ca2+ efflux and avoiding interference by Ca2+ pumping activity. As ob-served in Fig. 4, Ca2+ release may disappear when Ca2+ pumpmediated Ca2+ influx exceeds a partially inhibited Ca2+ efflux. Under such circumstances, an equilibrium might be erroneously assumed, leading to an underestimate of the time to attain equilibrium.
Unidirectional Ca2+ efflux3 was measured by inhibiting the Ca2+pump activity with 100 PM quercetin after the completion of Ca2+ loading (see inset in Fig. 7). The time dependence of inhibition by each local anesthetic on Ca2+ efflux was examined a t several concentrations, and the ICso i.e. the concentration which a t equilibrium inhibited Ca2+ efflux by 50% ( Fig.   7 and Table 111) was determined. &X 572; which is less lipid soluble than SKF 525-A, reached equilibrium approximately three times faster. Equilibration times for benzocaine ( Fig. 7 and Table 111), prilocaine, mepivacaine, lidocaine, tetracaine, bupivacaine, and etidocaine (Table 111) were intermediate.
More hydrophobic local anesthetics took longer to reach equilibrium and the half-time to reach equilibrium could be correlated with the logarithm of the hydrophobic partition coefficient (Fig. 8a). Three anesthetics did not conform to this trend requiring very long times to approach equilibrium. Procaine and procainamide (10 mM of each drug) and &X 314 (2.5 mM), after 4 h of preincubation, decreased Ca2+ efflux by 45, 20, and 55%, respectively, without reaching equilibrium (data not shown).
At equilibrium, SKF 525-A was the most effective local anesthetic for inhibiting Ca2+ efflux. &X 572 and tetracaine also were effective in the submillimolar range. Equilibrium potency for the other local anesthetics tested exceeded the millimolar range (Table 111). The equilibrium potency (ICso) did not correlate well with the hydrophobic partition coefficient (Fig. 8b).
p H Dependence of Inhibition of Ca2+ Efflux by Local Anesthetics-Since the distribution of tertiary amines across membrane is a function of [base] (45, 50, 51), we investigated the pH dependence of inhibition of Ca2+ efflux by local anesthetics. The rate of Ca2+ efflux in the absence of local anesthetics (not shown) displayed a maximum around pH 7.0, decreasing steadily at both higher and lower pH. Procaine and tetracaine, when added to the assay medium, were least inhibitory at pH 7.0 and more effective both below and above pH 7.0 (Fig. 9a). However, &X 314 and &X 572, both permanently charged, and benzocaine, completely in the neutral form, were insensitive to pH changes above pH 7.0 (Fig. 9b). &X 572 at 0.25 mM inhibited Ca2+ efflux by 50% above neutrality, but was more potent at lower pH. At pH 6.6, there was complete inhibition. &X 314 at 5 mM did not affect Ca2+ Inhibition of Spontaneous Ca2+ Rdease by Local Anesthetics   Fig. 7. Antipyrylazo 111 was used as the calcium indicator except for QX 572 where arsenazo 111 was used. IC50 is defined as the concentration of local anesthetic which inhibits Ca2+ efflux rate by 50% at equilibrium, and was determined by trial and error. The half-time to reach equilibrium was calculated as described in Fig..   in uiuo, may be to provide an electrical shunt across the SR membrane thereby preventing charge imbalance during Ca2+ uptake and Ca'+ release (56).
Local anesthetics might conceivably exert their primary action on one or more of these permeabilities, with inhibition of Ca2+ release being a secondary effect. We, thus, investigated the effect on Ca'+ release of known inhibitors of either the anion permeability or the Na' and K+ channel.
DIDS, which has been reported by Kasai (55) and Kasai and Taguchi (57) as a potent inhibitor of anion permeability of SR (Kd = 40 p~) , was found to enhance Ca2+ release rate (Fig. lob) when administered after the completion of Ca2+ loading (177% of the control with 200 p~) .
This enhancement FIG. 10. Effect of DIDS on Ca2+ release and Ca2+ efflux from SR. Ca2+ loading was measured in K medium, in a final volume of 2 ml as described under "Experimental Procedures." SR (56 pg of protein) was loaded with 3 consecutive 100-nmol CaCl, pulses. Antipyrylazo I11 was used as the calcium indicator. a shows a control trace; the rate of Ca2+ release was 2.39 pmol of Ca2+/min .mg of protein. b shows the trace obtained with 200 p~ DIDS added during the lag phase (arrow); the rate of Ca2+ release was 4.22 pmol of Ca2+/ min.mg of protein. The arrowhead denotes the addition of 100 nmol of CaCI, for recalibration of antipyrylazo 111. Caz+ loading rate in the presence of 200 p~ DIDS was 1% of the control activity. c shows Ca2+ efflux in the absence (0) and presence (0) of 200 p~ DIDS added during the lag phase. Experimental conditions were as for a and b except that 45CaCl, (-lo6 cpm) was administered together with the SR. After completion of Caz+ loading, 1 mM EGTA was added (arrow) and 0.1-ml aliquots were withdrawn and rapidly filtered (see "Experimental Procedures"). Filtrates were counted for 45Ca. Ca" efflux rates were 4.16 and 3.94 pmol of Ca2+/min.mg of protein in the absence and presence of DIDS, respectively. The traces are displaced with respect to each other for the sake of clarity. The vertical bar represents 100 nmol of CaCI,. appears to be due to inhibition of Ca2+ loading, since Caz+ efflux was not stimulated (Fig. ~O C ) , whereas Ca2+ loading and (Ca2+, Mg")-ATPase activities were practically abolished by 200 DIDS5 (see also legend to Fig. 10).
Two specific K+ channel blockers were used i.e. bis-Ql0 and bis-Q12 which have a blocking site on the inner face of the SR membrane (58). Neither 1-5 mM bis-Q10 nor 200 PM bis-Q12 affected Ca2+ release, even when long preincubations (up to 3 h) were performed (data not shown). Effect of Ionophores on Ca2+ Release and Efflux-If local anesthetics were to inhibit Ca2+ release by blocking monovalent cation or proton permeability, addition of specific ionophores (59) should restore Ca2+ release. 5-10 PM FCCP, a protonophore, or 5-20 kg/ml of gramicidin D, a monovalent cation ionophore, did not change the characteristics of spontaneous Ca'+ release or modify Ca2+ efflux (data not shown).  Ca2+ loading was measured in Na medium, in a final volume of 2 ml, in the presence of ''CaCIZ (-lo6 cpm). All experiments were carried out on the same SR preparation, using 54 pg of protein preloaded with 300 nmol of CaC12. Arsenazo I11 was used as the calcium indicator. Local anesthetics were administered just prior to initiation of Caz+ loading. At the completion of loading, 40 p1 of concentrated Ca-EGTA solution was added so that free [Ca], was fixed at either or IO-' M. Aliquots (0.1 ml) were then withdrawn, rapidly filtered (see "Experimental Procedures"), and filtrates were counted for 45Ca. The Ca2+ efflux rate of the control was 3.35 f 0.08 (4) at IO-* M and 1.39 _t 0.19 (4 As seen in Table IV, a t pCan 8, SKF 525-A (40-60 p M ) , &X 572 (0.2-0.25 mM), and benzocaine (4-6 mM) inhibited ca2+ efflux when administered just prior to Ca2+ loading. Procaine (10 mM) and &X 314 (0.25 mM) were largely ineffective. These results are in good agreement with those obtained by other assays described earlier. In contrast, at p C a~ 4, only benzocaine (4-6 mM) appreciably inhibited Ca2+ efflux (Table IV).

DISCUSSION
This study provides the first detailed analysis of the effect of local anesthetics on spontaneous Ca'+ release, and shows that local anesthetics inhibit this form of Ca2+ release. The inhibition is not instantaneous but time-dependent and varies widely with respect to the specific local anesthetic. This study suggests the existence of a Ca2+ pathway (channel) which is blocked by anesthetics from the inner face of the SR membrane.
Model of Action of Local Anesthetics-The time dependence of inhibition of Ca2+ release (Figs. 3, 4 and 6) or Ca2+ efflux (Fig. 7) by local anesthetic might be explained either by slow binding of local anesthetic to an inhibitory site, or slow crossing of the membrane to reach the inhibitory site, or a combination of both. Previous studies (45) suggested an intrinsically rapid local anesthetic-receptor interaction. It seems unlikely that the time scale of binding would differ drastically between fairly similar compounds i.e. lidocaine and tetracaine reached equilibrium within minutes while their respective A A. Chu, personal communication.
analogs, &X 314 and procaine, did not reach equilibrium even after 4 h of preincubation. It seems more likely that the increased effectiveness of local anesthetics with incubation is due to time required for the anesthetic to permeate inside the SR vesicles. In this regard, we find that &X 314, which is poorly permeant, is effective only after prolonged preincubation, while the more permeant &X 572 (45) is rapidly effective.
Our model is that local anesthetics have to cross the SR membrane before reacting with an internal binding site(s).
The following four consecutive steps might be involved in inhibition of Ca2+ release by local anesthetics: 1) insertion into the external leaflet of the membrane, 2) distribution (flip-flop) between the two halves of the bilayer, 3) dissociation from the membrane into the aqueous internal milieu, and 4) association with the putative binding site. Different local anesthetics appear to be rate-limited in their onset of action by different steps.
The hydrophobic partition coefficient ( F ) would reflect the ability of the anesthetic to insert into the membrane (step 1).
In order for flip-flop to occur, the anesthetic must be in the uncharged form (45, 50, 51). Thus, flip-flop (step 2) is a function of both F and the percentage of free base form at a specific pH.
Step 3 would not be necessary if the anesthetic can reach the binding site while inserted into the membrane. However, the direct correlation between halftime to reach equilibrium and hydrophobicity (Fig. 8a) suggests that dissociation from the membrane is an obligatory and rate-limiting step for most of the local anesthetics studied (Table 111). More hydrophobic compounds such as SKF 525-A, etidocaine, and bupivacaine would be expected to dissociate more slowly from the membrane into the internal aqueous environment and this would be expected to retard binding to the inhibitory site when access is only via a hydrophilic pathway.
Procaine, procainamide, and QX 314 were not included in Fig. 8a because they behave differently, taking much longer time (hours) to exert an effect. For these three anesthetics, which are poorly permeant and 98 to 100% in a charged form, it would seem that step 1 and/or 2 of the model is ratelimiting. The finding that prilocaine, less hydrophobic but with a higher percentage of free base form than procaine (Table I), attains equilibrium after a few minutes (Table 111), whereas procaine does not reach equilibrium even after 4 h of preincubation, suggests that step 2 is likely to be rate-limiting for procaine. Since at pH 7.0, prilocaine contains 10-fold more of the free base form than procaine (Table I), it is possible that flip-flop is not rate-limiting for prilocaine but is for procaine. The same explanation can be advanced to interpret the increasing inhibition of Ca'+ efflux by procaine above pH 7.0 (Fig. 9a).
Comparison of the potency of different local anesthetics at equilibrium should reflect their association with a binding site (step 4) and express selectivity of interaction. The IC,, (Table 111) spans about three orders of magnitude. We find a direct relationship between M, (see Table I) and potency at equilibrium (correlation coefficient 0.71). This conforms with a suggestion by Courtney (60) that the dissociation rate is inversely related to the size of local anesthetics. There is poor correlation between lipid solubility and potency (Fig. 8b). However, when &X 572 (way out of line in Fig. 8b) is not considered, the correlation improves (correlation coefficient -0.77). This implies that hydrophobic interaction may be an integral component of the binding site, as recently reported by Wang et al. (61) for the Na+ channel of squid axon membranes.
A number of observations, e.g. the action of local anesthetics below pH 7.0 (Fig. 91, is best explained by pH-dependent changes of the SR binding site (step 4). These findings would suggest that the binding affinity for charged local anesthetics may be higher at low pH. The possibility of higher nonspecific membrane permeability at acidic pH is ruled out by the ineffectiveness of QX 314. The pH independence of benzocaine action is compatible with a different (although still internal) site of action for neutral local anesthetics (62-64). Alternatively, both neutral and charged local anesthetics may bind to the same site (45, 51, 60) with protonation increasing affinity only for charged anesthetic molecules (51). Evidence in favor of lipid versus protein as sites of action of local anesthetics on membranes has been presented (70). The mode of action described here indicates that the SR binding site for local anesthetics is protein in nature, since evidence for an inhibitory site on the inner face of the SR membrane is provided.
Selectivity of Action on a Ca2+ Channel-It appears unlikely from the studies presented, that local anesthetics inhibit spontaneous Ca" release by blocking monovalent ion permeability pathways in SR. The lack of effect of the protonophore FCCP implies that the proton gradient which transiently develops during Ca2+ uptake (65) is not relevant to spontaneous Ca'+ release or that it is dissipated before Ca2+ release takes place. This finding is in agreement with reports by Louis et al. (66) and Shoshan et al. (67) that protonophores do not induce Ca'+ release from isolated SR. Additionally, should action of local anesthetics have involved inhibition of proton or monovalent cation permeability, spontaneous Ca'+ release should have been restored by FCCP or gramicidin D, respectively. This was not the case. On the other hand, when Cat+ permeability was specifically increased with A23187, a Ca'+ ionophore, inhibition of Ca2+ release by local anesthetics was circumvented.
Experiments performed using an anion permeability blocker (DIDS) and K' channel blockers (bis-Q10 and bis-Q12) indicate that inhibition of Ca2+ release by local anesthetics cannot be attributed to a primary effect on these permeabilities. In fact, DIDS enhances Ca2+ release because the phosphate-facilitated CaZ+ re-uptake is blocked. The results obtained with bis-Q10 and bis-Ql2 are less than conclusive since we could not assess whether the bis-quaternary ammonium blockers were able to get to their internal blocking site or if a membrane potential (negative inside) favorable to such a block (58) existed. However, Miller6 found that K' conductance in a planar phospholipid bilayer containing K+ channels from SR (52) was not inhibited by 0.5 mM SKF 525-A and only partially inhibited by 2 mM QX 572 i.e. concentrations which are 10-fold higher than those required to inhibit spontaneous Ca'+ release from SR.
Taken together, our results are compatible with the view that local anesthetics interact directly (7, 16, 21-23, 26) with a Ca'+ pathway i.e. the Ca'+ channel or its gates, rather than with other ion pathways.