PP1 Inhibitors Depolarize HermissendaPhotoreceptors and Reduce K+ Currents

Haojiang Huang, Joseph Farley


Previous research indicates that activation of protein kinase C (PKC) plays a critical role in the induction and maintenance of memory-related changes in neural excitability of Type B photoreceptors in the eyes of nudibranch mollusk Hermissenda crassicornis (H.c.). The enhanced excitability of B cells is due in part to PKC-mediated reduction in somatic K+ currents. Here we examined the effects of protein phosphatase inhibitors on Type B photoreceptor excitability and K+ currents to determine the role(s) of protein phosphatases on memory formation in Hermissenda. Using electrophysiological and pharmacological methods, we found that the PP1 inhibitors calyculin A and inhibitor-2 depolarized Type B photoreceptors by 20–30 mV. A broad-spectrum kinase inhibitor, H7, blocked this effect. The depolarization induced by PP1 inhibition occluded that produced by an in vitro associative conditioning procedure. Calyculin and inhibitor-2 reduced the same B cell K+ currents (I Aand I delayed) that are reduced by in vitro and behavioral conditioning. H7 blocked the reductions. Cantharidic acid (PP2A inhibitor) and cyclosporin (PP2B inhibitor) had negligible effects on B cell resting membrane potential, K+ currents, and in vitro conditioning-produced cumulative depolarization of B cells. These results suggest that the functional activity of K+ channels in B cells is sustained by basal activity of PP1. Inhibiting PP1 appears to allow one or more constitutively active kinase(s) to reduce K+ channel activity and thus mimic the effects of conditioning. Our results suggest that PP1 may oppose and/or constrain the extent of learning-produced changes in B cell excitability.


Learning- and memory-related modulation of synaptic plasticity and neuronal excitability often involves posttranslational modifications of receptors, ion channels, and synaptic vesicle proteins. Protein phosphorylation and dephosphorylation is perhaps the most ubiquitous and best-studied posttranslational mechanism by which protein function can be altered (Cohen 1989; Krebs 1994;Levitan 1999; Nestler and Greengard 1999). Indeed, elucidation of the cellular mechanisms underlying learning and short- and intermediate-term forms of memory has profited from the characterization of how activation of second-messenger systems result in persistent modulation of membrane ion channels, receptors, and other cellular processes by phosphorylation (Bliss and Collingridge 1993;Dudai 1989). Similarly, a prominent theme in studies of long-term memory is cellular regulation of gene expression and protein synthesis, which is often controlled by phosphorylation/dephosphorylation of transcriptional activators and repressors (Silva et al. 1998). Although much early research on the roles of phosphorylation and dephosphorylation in learning and memory was focused on various protein kinases, interest in the involvement of protein phosphatases has accelerated in recent years (Asztalos et al. 1993; Blitzer et al. 1998; Endo et al. 1995; Mulkey et al. 1994).

A variety of lines of evidence indicate that persistent changes in ocular Type B and A photoreceptor excitability contribute to associative modifications of the sea snail Hermissenda crassicornis's (H.c.) phototactic behavior, produced by repeated pairings of light and rotation (Crow and Alkon 1978; Farley and Alkon 1980, 1982). These persistent excitability changes have, in turn, been linked to phosphorylation-mediated changes in K+ channel activity, with the protein kinase C (PKC) (Crow and Forrester 1993; Etcheberrigaray et al. 1992; Farley and Auerbach 1986; Farley and Schuman 1991), protein tyrosine kinase (PTK) (I. Jin and J. Farley, unpublished data), and MAP kinase (Crow et al. 1998) families receiving the greatest attention. Activators of PKC (Farley and Auerbach 1986) or PTKs (Jin and Farley, unpublished results) mimic many of the learning-produced changes in Type B cell excitability, expressed as enhanced photoresponses, reductions inI A andI K-Ca, and decreases in resting membrane conductance. Similarly, PKC activation by phorbol esters precludes the occurrence of additional excitability changes in B cells by in vitro conditioning (Jin and Farley, unpublished data; M. Smith, I. Jin, R. McEwen, H. Huang, and J. Farley, unpublished observations) or serotonin stimulation (Farley and Auerbach 1986), suggesting that these three ways of increasing B cell excitability converge biochemically. Conversely, prior enhancement of the Type B cell photoresponse and input resistance by behavioral conditioning (Smith et al., unpublished results) or serotonin stimulation (Farley and Auerbach 1986) attenuates the increases produced by PKC activation (Smith et al., unpublished observations). Broad-spectrum kinase inhibitors such as H7 [1-(5-isoquinolinesulfonyl)-2-methylpiperazine], as well as those more specific to PKC [such as sphingosine and staurosporine (Farley and Schuman 1991) and the peptide inhibitor PKC (19-31) (Jin and Farley, unpublished data)], and PTKs [genistein, lavendustin A (Jin and Farley, unpublished results)] block in vitro conditioning effects in B cells. Behavioral and in vitro conditioning have both been reported to result in translocation and apparent increases in PKC concentration/activity within the somatic membranes of Type B cells (Impey et al. 1991; McPhie et al. 1993; Muzzio et al. 1997).

In contrast to the evidence implicating phosphorylation events in memory formation in Hermissenda, the involvement of dephosphorylation has been relatively unexplored. One hint that dephosphorylation might regulate the same K+currents in Type B cells that are modulated by learning and kinase activation comes from a prior study of the ability of kinase inhibitors to prevent as well as reverse, conditioning-produced suppression of K+ currents (Farley and Schuman 1991). This study reported that exposure of Type B cells (after conditioning had already occurred) to the kinase inhibitors H7 or sphingosine resulted in a relatively rapid (∼30–60 min) reversal of the K+ current suppression that had been produced by behavioral conditioning but did not affect the K+ currents from untrained B cells. These results suggested an involvement of persistent kinase activity in learning-produced K+ current suppression in Type B cells, and also hinted at the presence of a highly active phosphatase in B cells, whose activity was unmasked when kinase activity was suppressed.

Muzzio et al. (1999) have recently suggested that phosphatase activation is responsible for massed trial learning deficits in Hermissenda. When pairings of light and rotation are administered at relatively short inter-trial intervals (ITIs), phototactic suppression of Hermissenda and facilitation of B cell excitability are reduced relative to more distributed training regimens (Farley 1987a; Farley and Alkon 1987; Muzzio et al. 1999). Increases in resting intracellular Ca2+concentration—[Ca2+]i—have been suggested to mediate the massed-trial learning deficits, in part through enhanced adaptation of photoreceptor light responses (Farley 1987b; Farley and Alkon 1987).Muzzio et al. (1999) found that basal [Ca2+]i levels increased when short ITIs separated successive light steps. This [Ca2+]i accumulation appeared to activate protein phosphatases because facilitation of Type B photoreceptor excitability could be rescued in animals trained with short ITIs if training occurred in the presence of okadaic acid (OA). Thus Muzzio et al. (1999) suggested that a PP2B-PP1/PP2A activation scheme, similar to that proposed by Mulkey et al. (1994) to explain induction of hippocampal LTD, might be responsible for massed-trial learning deficits inHermissenda.

To explore the potential contribution of dephosphorylation events to Type B cell membrane excitability, K+ channel activity, and learning in Hermissenda, Type B cells of untrained specimens were exposed to several widely used cell-permeant, as well as impermeant, inhibitors of serine/threonine protein phosphatases. We found that exposure of B cells to either calyculin A (an inhibitor of PP1 and PP2A) or inhibitor-2 (a specific peptide inhibitor of PP1) resulted in pronounced depolarization of B cells and suppression of the same K+ currents that are reduced by associative conditioning. These effects of calyculin A and inhibitor-2 were blocked by the kinase inhibitor H7, suggesting that inhibition of PP1 allowed a constitutively active protein kinase to dominate in phosphorylation/dephosphorylation cycles. Calyculin A and inhibitor-2 exposure also occluded in vitro conditioning produced changes in Type B photoreceptors, suggesting that PP1 normally serves to limit the extent of learning-produced changes in B cells. In contrast, cantharidic acid (PP2A inhibitor) and cyclosporin (PP2B inhibitor) had negligible effects on B cell resting membrane potential, input resistance, and K+ currents.

Portions of this research have previously been reported in abstract form (Huang and Farley 1996, 1997, 1999).



Adult Hermissenda were provided by Sea Life Supply (Sand City, CA). All animals were individually housed in perforated 50-ml tubes in artificial sea water (ASW) aquaria at 15°C on a 12-h light/dark cycle as previously described (Farley 1988). Animals were fed with small pieces of Mytilus edulis every other day.

Nervous-system preparation

The circumesophageal nervous system was dissected from an animal and placed on a glass microscope slide, within a ∼400-μl well of standard ASW. Each preparation was incubated in 1 mg/ml of protease (Sigma type XXVII, catalog No. 4789) for ∼10 min at room temperature (18°C) to facilitate cell impalement. After incubation, the nervous system was washed with a minimum of six volumes of 15°C ASW.

Intracellular recording and in vitro conditioning

Intracellular recording from an isolated nervous system was effected using methods described previously (Farley and Alkon 1987). Glass microelectrodes (A and M Systems, catalog No. 6020) filled with 3 M KCl (30–40 M ohms) were used to impale cells. A silver/silver chloride wire was used to connect the electrode solution to the head stage, and a similar wire was used to ground the bath. All recordings were made with an Axoclamp 2A (Axon Instruments) amplifier in current-clamp mode, and appropriate head stages. Signals were PCM-digitized at 44 MHz with the use of a Neurocorder (NeuroData Instruments, No. 284) and stored on video tapes.

The input resistances of Type B photoreceptors were measured 1–2 min after initial impalement, 2 min before the beginning of in vitro conditioning, and 2–3 min after the end of in vitro conditioning. Input resistances were measured from the voltage drops occurring across the cell's membrane in response to injections of −0.5- to +0.2-nA current steps (in 0.1-nA increments) through a balanced bridge circuit. However, in many conditions, due to the presence of a drug or peptide in the electrode, it was not possible to keep the bridge in balance during current injection because of the high resistance or plugging of the electrode. Resistance measurements were not obtained in these cases. All recordings were obtained at room temperature (∼18°C) in standard ASW with the following composition (in mM): 430 Na+, 10 K+, 10 Ca2+, 50 Mg2+, 10 Tris HCl, and 570 Cl, pH = 7.6–7.8.

Prior to in vitro conditioning (Farley 1987b;Farley and Alkon 1987), a Type B photoreceptor and an ipsilateral statocyst caudal hair cell were impaled, and the preparation was dark-adapted for 10 min. During in vitro conditioning, the nervous system was exposed to five 30-s simultaneous presentations of whole-field illumination (at an intensity of ∼300 μW/cm2) and depolarizing current stimulation of the hair cell (0.1–0.5 nA) every 2 min. The membrane potential of the B cell was continuously monitored for ≤5 min after conditioning. The 2-min postconditioning value is reported here. Peak- and steady-state light responses were also measured (to the nearest mV) during the first three light steps given following the ten min dark-adaptation period.

For those preparations and drug conditions in which the resting membrane potential of the Type B cell was stable (<2 mV change during entire dark-adaptation period, <0.50 mV change during the last 5 min of dark adaptation), cumulative depolarization of the B cell was measured by comparing the membrane potential 2 min after the last conditioning trial with the resting membrane potential value measured just before the first trial. However, cells exposed to calyculin A or inhibitor-2 often continued to depolarize throughout dark-adaptation, and a stable preconditioning membrane potential value could not be ascertained. In these cases, the membrane potential value was extrapolated to 2 min after the last conditioning trial, assuming that the B cell would have continued to depolarize at the same rate as had occurred during the 10 min prior to conditioning. Cumulative depolarization for these cells was measured by comparing the actual membrane potential 2 min after the last trial to the extrapolated membrane potential at that point (e.g., see Fig. 4 B).

The reliability of these extrapolated membrane potential values was assessed as follows. In seven separate control experiments in which preparations were exposed to calyculin A but were not exposed to in vitro conditioning, we estimated the rate of membrane depolarization during the 10-min dark-adaptation period. We then extrapolated the membrane potential value during the next 12 min, the time when five pairings of light and hair cell stimulation would have been administered had these preparations been exposed to in vitro conditioning. We then compared the actual membrane potential with the extrapolated value (extrapolated minus actual) and found that these differed from each other by an average of 0.29 ± 0.46 (SE) mV. Thus the extrapolated values provided reasonably accurate estimates of the actual membrane potential 12 min later and validated the use of extrapolated values for the purposes of assessing the extent of cumulative depolarization.


Calyculin A, inhibitor-2, and cantharidic acid were obtained from Calbiochem. H7 [1-(5-isoquinolinesulfonyl)-2-methylpiperazine], H8 {N-[2-(Methylamino)ethyl]-5-isoquinolinesulfonamide, HCl} and cyclosporin A were obtained from Sigma.

Calyculin A was dissolved in dimethylsulfoxide (DMSO) at an initial concentration of 0.1 mM. The drug was then diluted in distilled water to a 2 μM stock concentration. When either bath-application or intracellular leakage was used to deliver the drug, calyculin A was added to the ASW bath or microelectrode 3 M KCl solution to a final concentration of 20 nM (final DMSO concentration was 0.02%). When introduced directly into Type B cells, cantharidic acid was dissolved at a concentration of 1 μM in 3 M KCl solution (0.01% DMSO). Similarly, cyclosporin A was prepared at a concentration of 100 nM in 3 M KCL (0.01% DMSO). Inhibitor-2 was dissolved in 3 M KCl at a concentration of 20 nM, and the peptide was allowed to leak into Type B cells. H7 was dissolved in ASW at a concentration of either 60 or 120 μM. H8 was dissolved in ASW at a concentration of 150 μM. Stock solutions of phosphatase and kinase inhibitors were made fresh, every day or two, and care was taken to prevent photoinactivation of light-sensitive compounds.

In control experiments, preparations were bathed in normal ASW. The concentrations used for the protein phosphatase inhibitors [calyculin A (PP1 and PP2A inhibitor), inhibitor 2 (PP1 inhibitor), cantharidic acid (PP2A inhibitor), and cyclosporin A (PP2B inhibitor)], were 10–40 times higher than the IC50 values reported for inhibition of the respective phosphatases in purified enzyme assays and were very similar to those used by others in electrophysiological experiments (Furukawa et al. 1996; Holm et al. 1997;White et al. 1993). In most experiments, a drug was allowed to leak into a Type B cell through relatively low resistance electrodes (10–15 MΩ) filled with inhibitor and 3 M KCl. In several sets of experiments, calyculin A was applied in the bath (20 nM).

Voltage clamp

All experiments were performed on Type B cell somata that had been isolated by axotomy from all synaptic interactions, as well as from any impulse-generating membrane, as previously described (Alkon and Fuortes 1972; Farley and Auerbach 1986; Farley and Han 1997). The axons of the photoreceptors leave the base of the eye, where they enter the optic nerve, and pass ensheathed through the optic ganglion (Eakin et al. 1967; Stensaas et al. 1969). The photoreceptor axons within the optic nerve join the optic tract and terminate in a spray of fine endings within the cerebropleural neuropil, ∼100 μm from their somata (Auerbach et al. 1989; Eakin et al. 1967; Senft et al. 1982). The optic nerve was razor-lesioned at the point where it exited the base of the eye, proximal to its entry into the cerebropleural neuropil. Such a lesion leaves a photoreceptor cell body (∼35–50 μm in diameter) that contains the rhabdomeric phototransduction apparatus and responds to light with normal generator potentials but is without action potentials or detectable synaptic interactions (Alkon and Grossman 1978;Farley and Auerbach 1986; Farley and Han 1997).

Standard two-electrode voltage-clamp methods were used (Farley and Auerbach 1986; Farley and Han 1997). The electrodes used to monitor membrane potential had resistances of 30–40 MΩ when filled with 3 M KCl. A lower resistance electrode (10–15 MΩ) was used for current passage and drug leakage. Data acquisition and analysis were performed using Axon Instrument's pClamp 5.5 and 6.0 program suites. A Type B cell was accepted for further study if the resting membrane potential was more negative than −40 mV and the steady-state light response to a 300 μW/cm2light step was more than 12 mV. Unless otherwise indicated, holding potential was −60 mV. Four-hundred-millisecond depolarizing command steps were used to elicit voltage-dependent K+currents. The stimulation frequency used (every 7 s) resulted in no appreciable cumulative inactivation. Activation time constants (Tauon) for individualI A current traces were fit assuming a power function 〈A(1 −e −(t−K)/Tauon)n+ B〉, where n = 3, A = final current amplitude, B = vertical offset. Inactivation time constants (Tauoff) were fit with single-exponential functions 〈A*e −(t−K)/Tauoff+ B〉. The goodness-of-fit values (R 2) for all reported time constants were >0.95.

Statistical analysis

Differences in membrane potential, input resistance, cumulative depolarization, light responses, and ionic current amplitudes produced by different treatment conditions were assessed using appropriate analyses of variance (ANOVA). When a significant F value was obtained, post hoc multiple-comparisons (Tukey's HSD test) were used when comparing three or more treatment means to each other and/or a control condition. In the absence of ANOVA, a simple Student'st-test was used to compare the effects of two treatment conditions. Two-tailed significance tests were used and are reported as significant if P < 0.05, unless otherwise indicated.


PP1 inhibitors depolarize Type B cells and increase resting input resistance

To determine a role for constitutive protein phosphatase activity in regulating the excitability of Type B photoreceptors, 20 nM calyculin A, a PP1 and PP2A inhibitor, was applied to the bath, and changes in resting membrane potential (Fig.1) and input resistance of synaptically intact B cells were measured. Because of the variable duration of recording across experiments following drug application, which was never <10 min, the effects on resting membrane potential are given as rates of depolarization. Over a 10- to 30-min recording period, calyculin A depolarized B cells at a rate of 0.8 ± 0.12 (SE) mV/min (Fig. 1, A and B) and increased resting input resistance by 24 ± 8% (n = 7). The depolarization produced by calyculin was quite dramatic with cells changing by 20–30 mV during 20–30 min exposure to calyculin. Consistent with the effects of other agents that produce large depolarizations of Type B cells (e.g., high extracellular K+) (Alkon et al. 1984;Farley 1988), cells that were exposed to calyculin for 20–30 min and allowed to fully depolarize until their membrane potential stabilized (range of about −25 to −15 mV), showed markedly abnormal (slow and small) light responses and generally lost their ability to spike.

Fig. 1.

PP1 inhibitors depolarize Type B cells. A: representative intracellular recording from a Type B photoreceptor impaled with an electrode containing 20 nM calyculin. The record begins several seconds after initial impalement and stabilization of membrane potential with calyculin application (top trace, far left, ↓). During the first 10–20 s of the record, the cell hyperpolarizes slightly, probably due to continued recovery from the long-lasting depolarization response evoked by the illumination used to guide electrode placement and impalement. The B cell first begins to depolarize ∼45–50 s following the start of the record. For the next 8 min or so, the cell continues to depolarize at a rate of ∼1.6 mV/min. During the last min (9–10), the rate has slowed to ∼0.1 mV/min. This B cell depolarized ∼15 mV in 10 min. B: summary of the effects of protein phosphatase inhibitors on B cell resting membrane potential. Both PP1 inhibitors, calyculin and inhibitor-2, produced substantial depolarization, with calyculin leakage having somewhat greater effects. PP2A (cantharidic acid) and PP2B (cyclosporin A) inhibitors produced negligible depolarization. In this and all subsequent figures, each data point plotted represents the mean ± SE. The number of cells that contributed to each mean (1 cell per preparation) appears in the parenthesis of the figure legend for each condition.

To determine how much of the calyculin-produced depolarization was attributable to inhibition of phosphatase activity within the B cell itself, as opposed to processes presynaptic to B cells, the effects of bath-application of calyculin were compared for synaptically intact (calyculin bath condition, Fig. 1 B) versus isolated B cells (isolated calyculin condition, Fig. 1 B). The rate of depolarization was 0.72 ± 0.15 mV/min for synaptically-isolated cells (Fig. 1 B), which was not significantly different from that of intact cells [t(25) = 0.48].

When calyculin A was allowed to leak into synaptically intact B cells, it depolarized the B cells at a rate of 0.99 ± 0.24 mV/min (calyculin leak condition; Fig. 1 B). This rate was slightly, though not significantly [t(29) = 0.78], greater than that of bath application. These results indicate that bath-applied calyculin depolarized B cells primarily through mechanisms endogenous to the B cells, such as inhibition of PP1 and/or PP2A.

To further differentiate between PP1 and PP2A as the target(s) for calyculin, inhibitor-2, a selective peptide inhibitor of PP1, was allowed to leak into Type B photoreceptors. Inhibitor-2 depolarized B cells at a rate of 0.88 ± 0.13 mV/min (Fig. 1 B), which was not significantly different from calyculin bath-application [t(30) = 0.23] or calyculin leakage [t(17) = 0.47]. Other phosphatase inhibitors, such as cantharidic acid (a more specific PP2A inhibitor) and cyclosporin A (a specific PP2B inhibitor) produced negligible depolarization (<1.5 mV over 10 min; Fig. 1 B).

The effects of phosphatase inhibitors on the light responses of Type B photoreceptors were determined by measuring both the peak and steady-state components of the light response for different cells in the presence versus absence of inhibitors following a standard 10 min dark-adaptation period. The results for three consecutive light steps, delivered at a 2.0-min inter-stimulus interval (ISI), were averaged. In general, Type B cells exposed to the phosphatase inhibitors tested here showed smaller peak and steady-state depolarizing generator potentials than control condition cells (Fig. 2), but only some of these differences were significant.

Fig. 2.

Light response of Type B photoreceptors is reduced by protein phosphatase inhibitors. A: representative recordings of light responses (to 2nd of 3 light steps) of cells exposed to indicated protein phosphatase inhibitors. Both peak (1st arrow) and steady-state (2nd arrow) components of the light response were smaller for cells exposed to the protein phosphatase inhibitors, calyculin, inhibitor-2, and cantharidic acid. B: summary data for both peak and steady-state light responses in the presence of protein phosphatase inhibitors. Asterisk denotes significantly different (P < 0.05).

An ANOVA of the peak light-induced response data for all conditions revealed a significant effect of drug [F(5,49) = 8.91,P < 0.01], with intracellular leakage of calyculin [Tukey's HSD post hoc test: t(15) = 4.23,P < 0.01], and inhibitor-2 [t(16) = 5.87, P < 0.01] producing significant reductions of peak light response when compared with controls. The peak light responses of cells exposed to bath-applied calyculin, leakage of cyclosporin, or leakage of cantharidic acid were not significantly different from controls [t(24) = 2.24,t(9) = 2.07 and t(9) = 2.15, respectively]. Similarly, an ANOVA of the light-induced steady-state generator potential data for all conditions revealed a significant effect of drug [F(5,49) = 11.28, P < 0.01]. Bath-applied calyculin [t(24) = 4.25,P < 0.01], and intracellular leakage of calyculin [t(15) = 3.89, P < 0.01], inhibitor-2 [t(16) = 7.23, P < 0.01], and cantharidic acid [t(9) = 3.94,P < 0.01] produced significant reductions of the steady-state light response when compared with controls. The steady-state light response of cells exposed to cyclosporin was not significantly different from controls [t(9) = 1.89].

These results indicate a partial dissociation between the effects of phosphatase inhibitors on resting membrane excitability and light responses of B cells. PP1 inhibitors (calyculin and inhibitor-2) affected both. Cantharidic acid (PP2A inhibitor) reduced the steady-state light response without causing appreciable depolarization. Cyclosporin (PP2B inhibitor) failed to affect the resting membrane potential and produced a slight (though nonsignificant) reduction in the light response. The reduction of the light response by cantharidic acid is noteworthy in two respects. First, it suggests that in addition to PP1, PP2A might also regulate phototransduction processes in Type B cells. Second, the reduction indicates that the failure of cantharidic acid to depolarize Type B cells cannot be attributed to a failure of drug delivery. Despite the fact that cantharidic acid reduced both components of the light response by approximately the same amount as bath-applied calyculin and was thus clearly delivered to cells, the effects of calyculin on resting membrane potential were about seven to eight times greater than those of cantharidic acid. A similar, though weaker case, can be made for PP2B. Although cyclosporin failed to significantly reduce the light response of B cells when evaluated using relatively conservative post hoc Tukey tests, Student'st-tests indicated significantly smaller peak [t(9) = 2.49, P < 0.05], but not steady-state [t(9) = 1.65] light responses, when compared with controls. But cyclosporin, like cantharidic acid, had no effect on B cell resting membrane potential.

Collectively, our results suggest that PP1 is a primary serine/threonine protein phosphatase in B cells that is affected by calyculin A and inhibitor-2. Further, constitutive PP1 activity is involved in regulating the excitability of B cells.

H7 blocks the effects of calyculin A and inhibitor-2

Preexposure of B cells to the broad-spectrum S/T kinase inhibitor H7 blocked the effects of calyculin A and inhibitor-2 in a dose-dependent fashion (Fig. 3). In the presence of 60 μM bath-applied H7, the average rate of depolarization caused by leakage of calyculin A (∼1.0 mV/min) was reduced by 43% (0.56 ± 0.15 mV/min). When a higher concentration of H7 was used (120 μM), the effects of bath-applied calyculin and intracellular leakage of inhibitor-2 were reduced by 94% (0.05 ± 0.05 mV/min) and 89% (0.1 ± 0.1 mV/min), respectively. In contrast to H7, a high concentration of H8 (a potent and specific inhibitor of cyclic nucleotide-dependent protein kinases) produced only a partial block (18%) of calyculin's effect (n = 4).

Fig. 3.

The broad-spectrum S/T kinase inhibitor H7 blocks the effects of calyculin and inhibitor-2. H7 blocked the effects of calyculin in a dose-dependent manner (60 vs. 120 uM). H8 (150 μM) only partially blocked the depolarization produced by calyculin (∼18%).

These data imply that the effects of calyculin A and inhibitor-2 were due to a disruption of the balance between phosphorylation and dephosphorylation rather than through direct effects on membrane excitability (e.g., block of resting K+ channels, activation of channels with a reversal potential more positive than rest, etc.). When PP1 was inhibited by either calyculin A or inhibitor-2, serine/threonine kinase(s) apparently dominated the phosphorylation/dephosphorylation cycles and led to increased phosphorylation. The depolarization and increased input resistance of B cells are consistent with phosphatase inhibition (increased phosphorylation) having produced a closure of somatic K+ channels.

PP1 inhibitors occlude conditioning produced cumulative depolarization of B cells

Isolated nervous systems from untrained Hermissendawere conditioned in vitro in the presence or absence of protein phosphatase inhibitors (Fig. 4). Five pairings of light and statocyst hair-cell stimulation in standard ASW resulted in a cumulative depolarization of 6.86 ± 0.4 mV (n = 7; Fig. 4, A and D) consistent with that observed in previous studies (Farley 1987b; Farley and Alkon 1987; Farley and Schuman 1991; Grover and Farley 1987). When conditioned in the presence of either bath-applied calyculin A (n = 5; Fig. 4, B and D), intracellular leakage of calyculin A (n = 8; Fig. 4,C and D), or intracellular leakage of inhibitor-2 (n = 6; Fig. 4 D), Type B cells showed substantially and significantly reduced cumulative depolarization. In fact, in vitro conditioning following intracellular leakage of calyculin or inhibitor-2 produced nominal net hyperpolarizations: −1.13 ± 1.11 and −1.08 ± 1.38 mV, respectively (Fig.4 D). Whether these small hyperpolarizations are genuine or not is unclear. The reference membrane potentials for the majority of these experiments were extrapolated values, and an error in extrapolation could result in an artificial apparent hyperpolarization.

Fig. 4.

PP1 inhibitors occlude in vitro conditioning produced cumulative depolarization. Recordings from Type B cells prior to, during, and shortly after in vitro conditioning. B cell activity during the first (left) and fifth (right) in vitro conditioning trial are shown. A: control cell exposed to 5 pairings of light and hair cell stimulation is ∼9 mV depolarized 2 min following the fifth trial. B: cell exposed to bath-applied calyculin (20 nM) is depolarized by the PP1 inhibitor, continues to depolarize throughout conditioning, but is no more depolarized by pairings of light and hair cell stimulation than would be expected on the basis of exposure to the drug alone. - - -, the extrapolated membrane potential, assuming that the B cell would have continued to depolarize at the same rate prior to conditioning. Conditioning produced ∼0.5 mV of cumulative depolarization in this B cell, 2 min following the 5th conditioning trial. C: calyculin-exposed cell that was not depolarized by the drug (∼30 min following start of experiment), but cumulative depolarization was still blocked. D: summary of protein phosphatase inhibitors' effects on in vitro conditioning. Calyculin A and inhibitor-2 occluded cumulative depolarization. Cantharidic acid and cyclosporin A failed to affect cumulative depolarization. *, significantly different (P < 0.05).

However, in three of these experiments (2 involving calyculin-leakage and 1 involving bath-applied calyculin), the depolarization produced by calyculin was either negligible, or had ceased, by the end of the dark-adaptation period when in vitro conditioning was initiated. Thus the stable resting membrane potentials of these B cells during the 1-min period prior to in vitro conditioning were used as the reference potentials (e.g., Fig. 4 C). For these three preparations, in vitro conditioning produced only 0.33 ± 0.41 mV of depolarization. This was significantly less than that of control preparations [t(8) = 11.28, P< 0.0001] but not different [t(11) = 0.30] from the combined average hyperpolarization of −0.30 ± 1.15 mV (n = 10) observed for the majority of the calyculin-leakage and -bath-treated cells whose membrane potentials had not stabilized by the start of in vitro conditioning. Thus attenuation of cumulative depolarization by calyculin was not specific to cells whose reference potentials were extrapolated values. And it seems safe to conclude that calyculin- and inhibitor-2-treated cells were not depolarized as much by in vitro conditioning as controls.

In contrast to the effects of PP1 inhibitors, cantharidic acid (n = 4) and cyclosporin A (n = 4) failed to affect cumulative depolarization (Fig. 4 D). In the presence of these PP inhibitors, in vitro conditioning produced depolarizations similar to that of controls (cantharidic acid: 7.1 ± 1.1 mV; cyclosporin A: 7.9 ± 1.6 mV, respectively).

An ANOVA of the cumulative depolarization data for all treatment conditions revealed a significant drug effect [F(5,28) = 12.6, P < 0.01], with bath-applied calyculin [t(10) = 3.29, P < 0.05], intracellular leakage of calyculin [t(13) = 5.34,P < 0.01], and intracellular leakage of inhibitor-2 [t(11) = 4.93, P < 0.01] producing significant reductions of cumulative depolarization as compared with controls. In contrast, the depolarizations produced by in vitro conditioning for cells exposed to cyclosporin or cantharidic acid were not significantly different from controls [t(9) = 0.15 and t(9) = 0.56, respectively].

Thus prior depolarization of B cells by the PP1 inhibitors calyculin and inhibitor-2 occluded the cumulative depolarization resulting from in vitro conditioning. This suggests that PP1 phosphatase inhibition and behavioral conditioning share common biochemical mechanisms in their regulation of B cell excitability. Because conditioning-produced cumulative depolarization of B cells results from reduced K+ channel activity (Farley 1987b;Farley and Alkon 1987; Farley and Schuman 1991), the present findings suggest the hypothesis that calyculin and inhibitor-2, through inhibition of PP1, lead to phosphorylation and reduced activity of K+channels in B cells. This in turn leads to depolarization and occlusion of in vitro conditioning-produced depolarization of B cells.

PP1 inhibitors reduce B cell K+ currents

Two macroscopic K+ currents,I A (a rapidly inactivating, voltage-dependent “A type” current) andI delayed (a slow, noninactivating current) were recorded as in previous studies (Alkon et al. 1984; Farley 1987a, 1988; Farley and Auerbach 1986) (Fig.5 A). A major component ofI delayed is a slow Ca2+-activated K+ current (I K-Ca) (Alkon et al. 1984; Farley 1988; Farley and Auerbach 1986; ). In the absence of PP inhibitors,I A andI delayed showed only slight rundowns during extended recording periods, declining by ∼8% over 30 min (Fig. 5, B and C). Thirty minutes following addition of calyculin A to the bath,I A andI delayed were reduced by 45 and 67%, respectively (n = 7; Fig. 5, A–C). Similarly, when inhibitor-2 was allowed to leak into the B cell, reductions of 49 and 63% were recorded forI A andI delayed, respectively (n = 8; Fig. 5, A–C). In contrast, PP2A and PP2B inhibitors had much smaller effects on these K+ currents (Fig. 5, B andC). Thirty minutes following the establishment of voltage-clamp and initiation of drug-leakage, cantharidic acid had reduced I A andI delayed by 25 and 23%, respectively (n = 5). Similarly, cyclosporin A reducedI A andI delayed by 26 and 13%, respectively (n = 4).

Fig. 5.

PP1 inhibitors reduce B cell K+ currents. A: representative voltage-clamp current traces from B cell before (top) and 30 min after (bottom) inhibitor-2 leakage (left), or calyculin A bath-addition (right). Both I A (1st arrow) and I delayed (2nd arrow) were reduced, in a time-dependent manner. Summary plots of effects of PP-inhibitors onI A (B) andI delayed (C). Currents were elicited by command steps to 0 mV. B:I A was greatly reduced by PP1 inhibitors (calyculin and inhibitor-2). C:I delayed was also greatly reduced by calyculin and inhibitor-2. PP2A and PP2B inhibitors, cantharidic acid and cyclosporin, had much smaller effects on K+ currents.

An ANOVA of the peak K+ current amplitudes at 0 mV, 30 min following initiation of drug exposure, revealed a significant effect of drug [F(4,11) = 4.49,P < 0.05], with calyculin [Tukey's HSD post hoc tests: t(7) = 3.61], and inhibitor-2 [t(5) = 3.5] producing significant reductions of peak K+ currents when compared with controls. The peak currents of cells exposed to cantharidic acid, or cyclosporin, were not significantly different from controls [t(4) = 1.23 andt(4) = 1.34, respectively]. A similar ANOVA and post hoc comparison of the delayed K+ current data revealed a significant drug effect [F(4,13) = 6.509,P < 0.01], with calyculin [t(7) = 3.86, P < 0.05] and inhibitor-2 [t(7) = 3.6, P < 0.05] producing significant reductions of delayed K+ currents when compared with controls. The delayed K+current amplitudes from cells exposed to cantharidic acid or cyclosporin were not significantly different from controls [t(4) = 0.51 and t(4) = 0.06, respectively].

The reductions of K+ currents produced by PP1 inhibitors were blocked by 120 μM bath-applied H7 (Fig.6).

Fig. 6.

Kinase inhibitor H7 blocks the reduction in currents produced by PP1 inhibitors. Broad-spectrum kinase inhibitor H7 prevented the reduction in I A (A) andI delayed (B) induced by calyculin and inhibitor-2. Currents were measured at 0 mV, 30 min following application of the indicated PP-inhibitors.

In addition to reducing the amplitudes ofI A (Fig.7, A and C) andI delayed (Fig. 7, B andD) at potentials equal to or more positive than −25 mV, PP1 inhibitors also altered the voltage dependence ofI delayed but not ofI A. Following exposure to inhibitor-2, the I-V curve for I A showed a ∼7- to 8-mV shift toward more depolarized membrane potentials (Fig.7 A), without appreciable change in slope.1 In contrast, the slope of the I-V curve forI delayed (Fig. 7 B) was clearly reduced by inhibitor-2. In the absence of inhibitor-2,I A andI delayed exhibited e-fold increases per 14.36 ± 1.32 and 15.81 ± 0.93 mV, respectively, when measured over the range of steepest voltage-dependency. In the presence of inhibitor-2,I A andI delayed showede-fold increases per 15.39 ± 0.84 and 18.93 ± 1.23 mV.

Fig. 7.

Summary current-voltage (I-V) relations forI A and I delayedof inhibitor-2-treated type B cells. A:I-V curve for I A of inhibitor-2-treated cells (n = 8). Inhibitor-2 produced a 7- to 8-mV shift in the I-V curve toward depolarized potentials without affecting the slope of the curve.B: I-V curve forI delayed of inhibitor-2-treated cells (n = 8). Inhibitor-2 reduced the slope of theI-V curve, indicating a reduction in the voltage sensitivity of the K+ conductance. “After” currents were measured ∼30 min following exposure to the inhibitor.

Similar to inhibitor-2, the voltage dependence ofI A was not changed by calyculin A, with the I-V curve being shifted positive by ∼10 mV (Fig.8 A). However, calyculin A reduced the slope of the I-V curve forI delayed (Fig. 8 B), withe-fold changes in current-amplitude requiring 19.94 ± 0.94 mV of depolarization in the presence of calyculin versus 16.92 ± 1.17 mV in its absence.

Fig. 8.

Summary current-voltage (I-V) relations forI A and I delayedof calyculin treated type B cells. A:I-V curve for I A of calyculin treated cells (n = 6). Calyculin produced a ∼10 mV shift of the I-V curve toward depolarized potentials, without appreciable effect on the slope. B:I-V curve for I delayed of calyculin-treated cells (n = 6). Calyculin reduced the slope of the I-V curve, indicating a reduction in the voltage-sensitivity of the K+ conductance. After currents were measured ∼30 min following exposure to the inhibitor.

The activation (Tauon) and inactivation (Tauoff) kinetics ofI A (Fig.9 A) were only weakly voltage dependent, showing only slight changes with membrane potential. Inhibitor-2 produced a small, nonsignificant increase (∼15%) in Tauon (Fig. 9 B), and a modest (∼20%) but again nonsignificant decrease in Tauoff (Fig. 9 C). On average, calyculin A also produced a small, nonsignificant increase (∼12%) in Tauon (Fig. 9 D) but a slight nonsignificant increase in Tauoff (Fig.9 E). In summary, inhibitor-2 and calyculin both produced a small but consistent slowing of activation ofI A (increased Tauon), that may have contributed (slightly) to the reduction in I A amplitudes. In contrast, the effects of the PP1 inhibitors on Tauoff were inconsistent with each other and nonsignificant. Although it is possible that the modest and nonsignificant reductions in Tauoff produced by inhibitor-2 (i.e., faster inactivation ofI A) may have contributed slightly to the reduction in peak amplitudes, it seems unlikely that this mechanism was the major one accounting for reduction in peak amplitudes. The slight slowing of inactivation rate produced by calyculin obviously cannot be the mechanism that produced reductions in peak amplitudes.

Fig. 9.

Summary of Tauon and Tauoff values for cells treated with PP1 inhibitors. A: representative records and curve fits (smooth dark curves) for activation ofI A (left) and inactivation ofI A (right). Both traces are from the same cell, at 10 mV. B: summary of changes in Tauon in the presence of inhibitor-2. C: summary of changes in Tauoff in the presence of inhibitor-2. D: summary of changes in Tauonin the presence of calyculin. E: summary of changes in Tauoff in the presence of calyculin. None of these changes were statistically significant.

Functional equivalence of PP1-inhibition, PKC-activation, and conditioning

The effects of PP1 inhibitors on Type B cell K+ currents were qualitatively very similar to those of phorbol esters (PKC-activators) (Farley and Auerbach 1986; Smith et al., unpublished data), serotonin-stimulation (Farley and Auerbach 1986; Farley and Wu 1989) [which participates in the learning-induced plasticity of B cells and occludes the effects of phorbol esters (Auerbach et al. 1989; Crow and Forrester 1991;Grover et al. 1989)], in vitro conditioning of isolated nervous systems (Farley and Schuman 1991), and multi-trial behavioral training of intact animals (Alkon et al. 1985; Farley 1988). However, there were quantitative differences among these treatments in their effects on K+ current amplitudes and voltage dependency. For example, phorbol ester (Farley and Auerbach 1986) and serotonin-application (Farley and Wu 1989) both reduceI A peak amplitudes by ∼30% over most of the activation range but have little effect (∼5–10% reduction) on A-current voltage dependency. Similarly, multi-trial behavioral conditioning of intact animals reduces peakI A currents by ∼30% (Alkon et al. 1982; Farley 1988; Farley and Schuman 1991), produces a depolarizing shift of ∼10 mV in theI-V curve for I A but only slightly reduces (∼5–10%) the voltage dependency of the A-type conductance (slope of I-V curve).

In contrast, injections of exogenous PKC into B cells reduced A currents by 30–60%, depending on the activation potential, and produced a corresponding flattening of the I-V curve (Farley and Auerbach 1986). Similarly, in vitro conditioning reduced I A by 30–60%, depending on activation potential, and also produces a marked reduction in voltage dependency (Farley and Schuman 1991).

Thus manipulations that produce largeI A reductions (in vitro conditioning, PKC injections) also reduce the voltage dependency of the conductance. Manipulations that produce less dramatic reductions inI A (serotonin stimulation, phorbol esters, behavioral conditioning, and PP1 inhibition) have correspondingly smaller effects on voltage-dependency. All manipulations that reduce I A appear to produce slight decreases in activation kinetics, but these do not appear to be correlated with the magnitude ofI A suppression.

Although multi-trial behavioral conditioning reduces the peak amplitudes of composite I delayed(Alkon et al. 1985) and pharmacologically isolatedI K-Ca (Farley 1988;Farley and Schuman 1991) by ∼50% over the activation range, the effects on voltage dependency are generally much more pronounced than for the A conductance. Typically, the slopes of theI-V curves are reduced by >45%. Phorbol esters/PKC injections (Farley and Auerbach 1986), and in vitro conditioning (Farley and Schuman 1991) also reduce the voltage dependency by >45%.


PP1 inhibitors affect B cell excitability

PP1 inhibitors, calyculin A and inhibitor-2, produced large (20–30 mV in 30 min) depolarizations of Type B photoreceptors, accompanied by an increase in resting input resistance. Approximately equivalent depolarizations were observed regardless of whether the inhibitors were applied in the bath or intracellularly through the recording electrode, and whether the cells were synaptically intact or isolated. Both results imply that the principal drug targets responsible for depolarization were endogenous to Type B cells. The depolarization resulting from PP1 inhibitors mimicked and occluded that produced by in vitro conditioning, suggesting that conditioning and PP1 inhibition affected a common signal transduction cascade. In contrast, cantharidic acid and cyclosporin, potent and specific inhibitors of PP2A and PP2B, had negligible effects on B cell resting membrane potential and failed to occlude in vitro conditioning.

Voltage-clamp analysis revealed that PP1 inhibitors suppressedI A andI delayed of B cells.I A's voltage dependency was unaffected by PP1 inhibitors, with the I-V curve showing a ∼7- to 10-mV parallel displacement toward depolarized potentials. Activation kinetics were only slightly slowed (by ∼12–15%) but may have contributed somewhat to the reductions in peak current amplitudes. In the case of I delayed, the voltage dependence of the macroscopic current was clearly reduced. MacroscopicI delayed is composed of residualI A,I K,V, andI K-Ca (Alkon et al. 1984; Farley 1988; Farley and Auerbach 1986; Farley and Schuman 1991; Farley and Wu 1989), and their relative contributions toI delayed change as a function of membrane potential. At potentials between −25 to +10 mV (inclusive),I K-Ca is the major contributor accounting for ≥60% of macroscopicI delayed (Alkon et al. 1984; Farley 1988). Thus it is likely that reductions of I K-Ca account for much of the reduction in I delayed. Additional analysis of the effects of PP1 inhibitors onI delayed will require isolation of the different current components before clear conclusions can be drawn as to mechanisms responsible for the apparent reduction in voltage sensitivity. However, to a first approximation, the effects of PP1 inhibitors were very similar to those produced by behavioral conditioning, and further reinforce the conclusion that PP1-inhibition and associative training involve common biochemical mechanisms.

The broad-spectrum S/T-kinase inhibitor H7 blocked the membrane depolarization and K+ current suppression produced by PP1-inhibitors, implying that the inhibitors' effects probably resulted from perturbation of phosphorylation-dephosphorylation cycles rather than from some other process unrelated to phosphatase inhibition (e.g., a direct block of K+ channels).

Collectively, the present and previous results suggest the scheme for kinase/phosphatase involvement in learning-produced changes in B cells depicted in Fig. 10. Prior to conditioning, the functional activity of “A type” and Ca2+-activated K+ channels in somatic membranes of B cells is sustained by constitutive PP1 activity, which evidently dominates any suppressive effects of basal constitutive phosphorylation. Calyculin-A and inhibitor-2 treatments, presumably through inhibition of PP1, allow either constitutively-active kinases (and/or induce activation of kinases) to reduce K+ channel activities (Fig. 10). Because the activities of these same K+ channels are reduced by in vitro (Farley 1987b; Farley and Schuman 1991) and multi-trial behavioral conditioning of intact animals (Alkon et al. 1982, 1985; Farley 1988), through PKC- (Farley and Auerbach 1986;Farley and Schuman 1991) and PTK-mediated (Jin and Farley 2001) phosphorylation events, the result is that PP1 inhibitors occlude the effects of conditioning.

Fig. 10.

Model of the involvement of PP1 in the learning-produced changes inHermissenda Type B photoreceptor excitability. +, a stimulation/activation effect; −, an inhibition effect. See text for explanation. For the sake of figure simplicity, we have indicated the kinase responsible for reducing K+ channel activities when PP1 is inhibited as protein kinase C (PKC). But this has not yet been determined. It may be PKC or some other kinase.

Dephosphorylation and K+ channels

Protein phosphatases have been shown to regulate a wide variety of ion channels (Herzig and Neumann 2000), especially K+ channels. Because of the functional and structural diversity of K+ channels, and the different types of cell in which they have been studied, a common scheme for phosphatase-linked regulation of K+channels has not yet emerged. Nevertheless some of our results are very similar to those of others.

In voltage-clamped Aplysia sensory neurons, intracellular injections of phosphatase inhibitors, okadaic acid and microcystin-LR, altered baseline currents and occluded 5-HT and cAMP-induced reductions in K+ currents (Ichinose and Byrne 1991; Ichinose et al. 1990). These studies implicated PP1 and/or PP2A in the reversal of cAMP-induced phosphorylations. Conversely, injections of purified mammalian PP1 or PP2A catalytic subunits mimicked the effects of FMRFamide and induced outward K+ currents. Additional studies have since narrowed the focus to PP1 (Endo et al. 1995).

Dephosphorylation of Ca2+-activated K+ channels (or associated proteins) has been reported to increase their activity in a variety of nonneuronal cell types (White et al. 1991, 1993; Zhou et al. 1996). Several examples of this type of regulation have implicated a cGMP-PKG-PP2A signaling pathway. A similar mechanism may also operate in neurons (Furukawa et al. 1996;Holm et al. 1997). Similarly, Pedarzani et al. (1998) reported that the Ca2+-activated K+ channels that underlie the slow afterhyperpolarization (sI AHP) in rat CA1 pyramidal neurons are slowly suppressed by inhibitors of PP1/PP2A.

A variety of large conductance Ca2+-activated K+ channel (BKCa) subtypes, isolated from mammalian brain synaptosomes and incorporated into artificial bilayer membranes (Farley and Rudy 1988), have been found to be differentially modulated by S/T kinases and phosphatases. Type 1 channels are stimulated by PKA (Farley and Rudy 1988; Reinhart et al. 1991) and inhibited by PP2A (Reinhart et al. 1991). Type 2 channels show the reverse pattern of effects (Reinhart et al. 1991).

Collectively, these results raise the possibility that stimulation of Ca2+-activated K+ channels by dephosphorylation through PP2A/PP1 may be a common mechanism to reduce cellular excitability and promote calcium homeostasis, thereby contributing to anti-apoptotic and neuroprotective functions.

In summary, both voltage- and Ca2+-activated K+ channels are regulated in a variety of diverse ways, including phosphorylation by PKG, PKC, and PKA and dephosphorylation by PP2A and PP1. These enzymes appear to be tightly associated with channel complexes and are often activated under basal stimulation conditions (Reinhart and Levitan 1995). Thus our findings that PP1 inhibitors shut K+ channels in B cells, presumably by allowing a constitutively active kinase to dominate, has several precedents in the literature.

Physiological role(s) of PP1 in B cells?

Our results suggest that conditioning and PP1-inhibition affect a common signal transduction cascade in B cells. Our results do not, however, define the precise physiological role for PP1. One interesting possibility is that in addition to activation of PKC, concomitant inhibition of PP1 may be necessary for the induction of (excitatory) conditioning produced-changes in B cell excitability, similar to the “gating” function proposed for PP1 inhibition in the transition from the intermediate- to late-phase stages of LTP in the hippocampus (Blitzer et al. 1998). PP1 may oppose and/or constrain the extent of learning-produced changes in B cell excitability and thus function as a checkpoint for memory formation.

One problem with this hypothesis as it applies toHermissenda concerns the phosphatase inhibitor-produced large depolarizations of Type B cells, which were accompanied by the loss of light responses and spiking. Prolonged 20- to 30-mV depolarizations of B cells are likely to be lethal. It is quite likely that [Ca2+]i levels in B cells increased during these depolarizations, since the final membrane potential reached (range of −25 to −15 mV) overlaps with that for gating of voltage-dependent Ca2+channels in B cells (Alkon et al. 1984; Farley 1988). Diminished light responses in B cells can be produced by intracellular injections of Ca2+ (Alkon 1979; Sakakibara et al. 1998), consistent with the widely accepted view of Ca2+as a mediator of light adaptation in invertebrate photoreceptors (Minke and Selinger 1996; Nagy 1991).

Although B cells undergo a cumulative depolarization during conditioning, the depolarization reaches a plateau of ∼8–10 mV after 10 light-rotation pairings (Farley 1987a; Farley and Alkon 1987; Grover and Farley 1987), far less than the 20–30 mV produced by calyculin and inhibitor-2. And in vitro-conditioned B cells retain their light response and spiking ability.

Thus a problem for the preceding PP1-inhibition hypothesis is that if such a process did occur for prolonged time periods, it would be expected to trigger apoptosis if not necrosis. Therefore it may be that PP1 is further activated during conditioning to constrain and limit the depolarization that results from pairing-produced PKC- and PTK-mediated reductions in K+ channel activity. This would be consistent with the general theme that has emerged from studies of PP1's effects on Ca2+-activated K+ channels reviewed earlier: PP1-activation of K+ channels results in hyperpolarization, reduced excitability, and protection against cell death.

It is also possible that PP1 activity might be further stimulated by other training conditions. Because PP1 activity is associated with increased K+ currents in B cells, it is intriguing to consider the possibility that PP1 might be activated by conditioned inhibition training paradigms. Exposure ofHermissenda to explicitly unpaired (EU) presentations of light and rotation produces persistent decreases in the excitability of Type B cells: smaller light responses and action potential frequency (Britton and Farley 1999), enhanced K+ currents (Farley et al. 1999). These inhibitory-learning correlated changes are the opposite of those produced by light-rotation pairings. Thus PP1 is an attractive potential candidate to play a role in these increases. A related possibility is that PP1 mediates accelerated forgetting, caused by extinction training in Hermissenda (Richards et al. 1984).

One should not overlook the possibility that PP1/PP2A may directly modulate PKC in B cells. PKC is regulated in vivo by three functionally distinct phosphorylations (Keranen et al. 1995). Recombinant PKCα activity has been reported to be reversibly inhibited by PP1 and PP2A (Ricciarelli and Azzi 1998), apparently at autophosphorylation sites. Further, at low concentrations, PP1 has been found to activate PKCα (Ricciarelli and Azzi 1998), implying the existence of an inhibitory phosphorylation site. Thus the degree of PKC activity in B cells may depend on PP1 activity, and this may introduce additional complexities to the signaling networks that are activated during and after conditioning. It is possible that calyculin and inhibitor-2 activated PKC in B cells and, in combination with inhibition of PP1, stimulated phosphorylation.

PP1 inhibitor effects on light responses

Protein phosphatases are likely to affect other physiological processes in B photoreceptors that have only an incidental connection to learning, such as phototransduction.

Both membrane and soluble PP1- and PP2A-like phosphatase activity have been shown to exist in visual and nonvisual tissue inLimulus nervous system (Edwards et al. 1996). The activities of both enzymes are greater in light- than in dark-adapted lateral eyes. Furthermore, in Limulus ventral eye photoreceptors, okadaic acid (a PP1 and PP2A inhibitor) caused a delayed depolarization of resting membrane potential and slower and diminished light-activated cation currents (Edwards et al. 1996). Inhibitors of calcineurin have also been reported to increase phosphorylation of arrestin, produce smaller and sharper quantal bumps, and slow the activation kinetics of light-induced inward currents in Limulus photoreceptors (Kass et al. 1998).

The findings from Limulus photoreceptors with PP1/PP2A inhibitors are generally consistent with our results withHermissenda. Although the molecular phototransduction pathways have not been thoroughly determined forHermissenda, it seems likely that phosphatase inhibitors affect the light response of invertebrate photoreceptors, and these effects occur in addition to effects on somatic K+ channels.

Limitations of present studies

A limitation of our present studies is the fact that all of our conclusions rest on pharmacological/electrophysiological evidence. Biochemical studies and characterization of phosphatases from theHermissenda nervous system (and more specifically, Type B photoreceptors) have not yet been reported. Thus the validity of our conclusions depends heavily on the as yet untested assumptions that Type B photoreceptors express PP1 and PP2A isoforms similar to those of other eucaryotic organisms and that the Hermissenda isoforms are inhibited by the PP-inhibitors examined here with approximately the same potency and specificity as their counterparts in other organisms.

We think the likelihood that these assumptions are correct is high. The three major S/T phosphatases families (PP1, PP2A, PP2B) are found in the nervous systems of organisms at all phylogenetic levels, ranging from Drosophila and Caenorhabditis elegans to rat to humans. Mollusks that are related to Hermissenda, such asAplysia, also express these same phosphatases (Endo et al. 1992). The catalytic subunits of PP1, PP2A, and PP2B are highly conserved. Diversity appears to come about primarily through which type(s) of regulatory subunits and accessory proteins the core catalytic subunit is associated with (Cohen 1989;Shenolikar 1994; Wera and Hemmings 1996). Most inhibitors bind to the catalytic domain/subunit of the PPs (Herzig and Neumann 2000; MacKintosh and MacKintosh 1994). The relative potency and specificity of calyculin A and inhibitor-2 against PP1 versus PP2A and PP2B is well established and generally accepted. The same appears to be true for cyclosporin (PP2B-inhibitor) and cantharidic acid (PP2A).

However, it is obvious that molecular identification and biochemical characterization of PPs from the Hermissenda nervous system and photoreceptors will eventually need to be accomplished. Biochemical studies will be particularly important for answering questions like: is PP1 constitutively active in the basal state? Is learning and/or memory formation associated with a change in PP activity?


  • Address for reprint requests: J. Farley, Neural Science, 1101 E. 10th St., Rm. 370, Indiana University, Bloomington, IN 47405-7007 (E-mail:farleyj{at}indiana.edu).

  • 1 In Fig. 5, the current responses to depolarization to 0 mV are normalized to the amplitudes at time 0; in Fig.7, averages of absolute value of current amplitudes are depicted. These different ways of calculating the averages [mean of ratios (Fig. 5) vs. ratio of means (Fig. 7)] are responsible for the slight discrepancies between the two sets of figures.


View Abstract