Specific protein changes during memory acquisition and storage

Changes in several distinct types of neuronal proteins are now known to be associated with learning. In this review, we will summarize the properties of these proteins and relate these properties to prominent theories of the biochemical basis of memory.

Hippocampal lesions impair memory of short-delay conditioned eye blink in rabbits

Involvement of hippocampus in short-delay eye blink conditioning was reexamined during conditioned response (CR) consolidation. Rabbits received bilateral hippocampectomy, removal of overlying neocortex, or sham lesions and were trained with tone/puff pairings to early acquisition (consolidation) or well trained (overtraining); retention was tested. Two effects were observed: 1) Rabbits with hippocampal lesions showed less retention in the consolidation experiment than controls. Previous studies may not have found this because initial training was more complete. Overtrained hippocampal rabbits showed more retention, which agrees with this suggestion. 2) Hippocampectomized rabbits showed larger CR amplitudes in the overtraining experiment. The complementary roles of hippocampus in the consolidation process during early learning and in modulating the expression of the amplitude/time course of behavioral conditioned responses after associations are well learned are discussed.

Sequential modification of membrane currents with classical conditioning

Pavlovian conditioning of the nudibranch mollusc Hermissenda crassicornis was previously shown to produce long-lasting reduction of two K+ currents measured across the Type B photoreceptor soma membrane (Alkon et al., 1982a; Alkon et al., 1985). Pavlovian conditioning of the rabbit was also shown to be followed by persistent K+ current reduction (Disterhoft et al., 1986). Here we report the first evidence that Ca2+ currents can also be modified by conditioning. The amplitude of the currents rather than their voltage-dependence remains reduced at least 1-2 d after conditioning (but not control procedures). Conditioning-induced changes of both K+ and Ca2+ currents increased as a function of training, the Ca2+ currents only changing substantially with greater than or equal to 250 trials. The later changes of the Ca2+ current may function to limit the magnitude of excitability increases due to associative learning.

AHP reductions in rabbit hippocampal neurons during conditioning correlate with acquisition of the learned response

Young adult male albino rabbits were conditioned using a free field auditory conditioned stimulus (CS) and periorbital shock unconditioned stimulus (US) in a short delay eye blink paradigm. All rabbits received two 80-trial training sessions. Intracellular recordings were made from hippocampal CA1 pyramidal neurons within brain slices prepared 24 h following the second training session. All 46 CA1 neurons included in the analysis had stable penetration, at least 70 mV impulse amplitudes and at least 40 M omega input resistance. Recording and initial data analysis were done 'blind' regarding behavioral training performance of the rabbit from which the slices were prepared. The animals were separated into a High (86 +/- 6% CRs, n = 12), and Low (12 +/- 4% CRs, n = 10) Acquisition group based on the number of blink CRs shown on the second training day (P less than 0.001). CA1 pyramidal neurons from the High Acquisition group (n = 20) showed a significant reduction in the afterhypolarization (AHP) response following 4 impulses elicited by intracellular current injection as compared to neurons from the Low Acquisition group (n = 26). The mean maximal AHP amplitudes after 4 spikes were -2.9 +/- 0.34 mV and -4.0 +/- 0.31 mV, respectively, in the High and Low Acquisition groups (P less than 0.01). The size of the AHP examined at 100 ms intervals during the first 1.7 s after the current pulse proved to be reduced in the High group both when evaluated for all points (F = 5.88, df = 1.44, P less than 0.02) and for each of the individual time points (at least P less than 0.05).(ABSTRACT TRUNCATED AT 250 WORDS)

Prolonged RNA changes in the Hermissenda eye induced by classical conditioning

The incorporation of 32P into mRNA and the total amount of mRNA were increased 3- to 4-fold in eyes isolated from Hermissenda crassicornis trained to associate light with rotation on a turntable compared with animals trained with equal numbers of light and rotation events presented randomly and with naive animals. Incorporation of 32P into poly(A)- RNA was reduced by as much as 60%. The RNA changes were strongly correlated with the degree of learning and could not be accounted for by changes in [32P]ATP content. The RNA changes were maximal at 24 hr and were still detectable after 4 days, indicating that associative conditioning produces a period of increased DNA transcription that could be an intermediate step in memory consolidation. The RNA changes may in part account for recently observed conditioning-specific changes in the synthesis rates of specific proteins.

Regulation of Hermissenda K+ channels by cytoplasmic and membrane-associated C-kinase

Pharmacologic activation of endogenous protein kinase C (PKC) together with elevation of the intracellular Ca2+ level was previously shown to cause reduction of two voltage-dependent K+ currents (IA and ICa2+-K+) across the soma membrane of the type B photoreceptor within the eye of the mollusc Hermissenda crassicornis. Similar effects were also found to persist for days after acquisition of a classically conditioned response. Also, the state of phosphorylation of a low-molecular-weight protein was changed only within the eyes of conditioned Hermissenda. To examine the role of PKC in causing K+ current changes as well as changes of phosphorylation during conditioning (and possibly other physiologic contexts), we studied here the effects of endogenous PKC activation and exogenous PKC injection on phosphorylation and K+ channel function. Several phosphoproteins (20, 25, 56, and 165 kilodaltons) showed differences in phosphorylation in response to PKC activators applied to intact nervous systems or to isolated eyes. Specific differences were observed for membrane and cytosolic fractions in response to both the phorbol ester 12-deoxyphorbol 13-isobutyrate 20-acetate (DPBA) or exogenous PKC in the presence of Ca2+ and phosphatidylserine/diacylglycerol. Type B cells pretreated with DPBA responded to PKC injection with a persistent reduction of K+ currents. In the absence of DPBA, PKC injection also caused K+ current reduction only following Ca2+ loading conditions. However, the direct effect of PKC injection in the absence of DPBA was only to increase ICa2+-K+. According to a proposed model, the amplitude of the K+ currents would depend on the steady-state balance of effects mediated by PKC within the cytoplasm and membrane-associated PKC. The model further specifies that the effects on K+ currents of cytoplasmic PKC require an intervening proteolytic step. Such a model predicts that increasing the concentration of cytoplasmic protease, e.g., with trypsin, will increase K+ currents, whereas blocking endogenous protease, e.g., with leupeptin, will decrease K+ currents. These effects should be opposed by preexposure of the cells to DPBA. Furthermore, prior injection of leupeptin should block or reverse the effects of subsequent injection of PKC into the type B cell. All of these predictions were confirmed by results reported here. Taken together, the results of this and previous studies suggest that PKC regulation of membrane excitability critically depends on its cellular locus. The implications of such function for long-term physiologic transformations are discussed.

Classical conditioning induces long-term translocation of protein kinase C in rabbit hippocampal CA1 cells

The role of the Ca2+/phospholipid-dependent, diacylglycerol-activated enzyme protein kinase C (PKC) in rabbit eyelid conditioning was examined. PKC was partially purified from the CA1 region of hippocampal slices from naive, pseudoconditioned, and conditioned rabbits 24 hr after the rabbits were well conditioned. Crude membrane and cytosol fractions were prepared. In conditioned rabbits, significantly more PKC activity (63.3%) was associated with the membrane fraction (and significantly less with the cytosol fraction) compared to naive (42.0%) and pseudoconditioned (44.7%) animals. These differences in distribution of enzyme activity were paralleled by differences in stimulation of enzyme activity by Ca2+, phospholipid, and diacylglycerol. There were no between-group differences in basal protein kinase activity. These results suggest that there is a long-term translocation of PKC from cytosol to membrane as a result of conditioning. Autoradiographic binding of radioactive phorbol 12,13-dibutyrate to PKC demonstrated that almost all specific binding was in the stratum radiatum, a region containing the proximal apical dendrites of CA1 pyramidal neurons. Therefore, this may be the site of the conditioning-specific PKC translocation, a locus well-suited to underlie the biophysical effects of conditioning.

Enhancement of synaptic potentials in rabbit CA1 pyramidal neurons following classical conditioning

A synaptic potential elicited by high-frequency stimulation of the Schaffer collaterals was enhanced in hippocampal CA1 pyramidal cells from rabbits that were classically conditioned relative to cells from control rabbits. In addition, confirming previous reports, the after-hyperpolarization was reduced in cells from conditioned animals. We suggest that reduced after-hyperpolarization and enhanced synaptic responsiveness in cells from conditioned animals work in concert to contribute to the functioning of hippocampal CA1 pyramidal cells during classical conditioning.

Transient and persistent depolarization-induced changes of protein phosphorylation in a molluscan nervous system

Phosphoproteins in the CNS of the nudibranch mollusc, Hermissenda crassicornis, were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography. After preincubation in artificial sea-water containing 32P, nervous systems were exposed to elevation of external K+ (100 or 300 mM) for a period (e.g., 30 min) approximating a period of depolarization which occurs during classical conditioning. Elevated external K+ was found to change the state of phosphorylation of three distinct proteins (Mr 56,000, 25,000, and 20,000) in three distinct ways without consistently changing that of any other proteins. Phosphorylation of an Mr 56,000 protein was increased by high K+ about twofold only in the presence of external Ca2+ [( Ca2+]o). Phosphorylation of Mr 25,000 protein, on the other hand, was decreased up to 10-fold by high K+, irrespective of the level of [Ca2+]o. The effect of depolarization on Mr 25,000 protein phosphorylation most likely represents dephosphorylation rather than proteolysis. This interpretation is consistent with the observations that (a) reappearance of the Mr 25,000 protein occurred in the presence of the protein synthesis inhibitors cycloheximide, puromycin, or anisomycin, and (b) the Hermissenda nervous system apparently contains a NaF- and EDTA-sensitive protein phosphatase capable of dephosphorylating Mr 25,000 protein. High K+ also reduced Mr 20,000 protein phosphorylation which was dependent on [Ca2+]o even in normal low K+ (10 mM) medium. Removal of [Ca2+]o enhanced reduction of Mr 20,000 phosphorylation due to the high K+ treatment. Interestingly, reduction of the Mr 25,000 protein phosphorylation was long-lasting, i.e., its phosphorylation did not fully recover to a control level for at least 30 min after the high K+ conditions had been removed.(ABSTRACT TRUNCATED AT 250 WORDS)

A spatial-temporal model of cell activation

A spatial-temporal model of calcium messenger function is proposed to account for sustained cellular responses to sustained stimuli, as well as for the persistent enhancement of cell responsiveness after removal of a stimulus, that is, cellular memory. According to this model, spatial separation of calcium function contributes to temporal separation of distinct phases of the cellular response. At different cellular sites, within successive temporal domains, the calcium messenger is generated by different mechanisms and has distinct molecular targets. In particular, prolonged cell activation is brought about by the interaction of calcium with another spatially confined messenger, diacylglycerol, to cause the association of protein kinase C with the plasma membrane. Activity of the membrane-associated protein kinase C is controlled by the rate of calcium cycling across the plasma membrane. In some instances, a single stimulus induces both protein kinase C activation and calcium cycling and thus causes prolonged activation; but in others, a close temporal association of distinct stimuli brings about cell activation via interaction of these intracellular messengers. Persistent enhancement of cell responsiveness after removal of stimuli is suggested to be due to the continued association, or anchoring, of protein kinase C to the membrane.

Inhibition of protein synthesis prolongs Ca2+-mediated reduction of K+ currents in molluscan neurons

Elevated intracellular Ca2+ concentration within the Hermissenda type B cell has previously been shown to cause transient reduction of both the early K+ current IA and the delayed, Ca2+-dependent K+ current ICa2+-K+, a reduction that is more permanent with classical conditioning. Other earlier experiments suggested that Ca2+-mediated reduction of K+ currents initially involves the dual activation of Ca2+/calmodulin-dependent and Ca2+/lipid-dependent protein kinases. In the present study, voltage-clamp conditions that cause substantial increases in intracellular Ca2+ concentration (i.e., a Ca2+ "load") were used to produce IA and ICa2+-K+ reduction with and without the protein synthesis inhibitor anisomycin or cycloheximide or the control substance deacetylanisomycin in the bathing medium. Anisomycin (100 microM) and cycloheximide (100 microM) caused no significant change of resting membrane potential, holding current, or the non-voltage-dependent "leak" current. However, inhibition of protein synthesis prevented recovery from Ca2+-mediated K+-current reduction. This effect resembled the effect of injecting purified Ca2+-dependent kinases and was blocked by the presence of trifluoperazine in the bathing medium. Activation of protein kinase C with a water-soluble phorbol ester caused marked reduction of protein synthesis in Hermissenda neurons as monitored by two-dimensional gel electrophoresis. Synthesis of new proteins therefore may be important for reversal of initial steps during memory storage, and Ca2+-activated phosphorylation pathways may initiate long-term changes by turning off (as well as by turning on) the synthesis of particular proteins.

In vitro associative conditioning of Hermissenda: cumulative depolarization of type B photoreceptors and short-term associative behavioral changes

Cumulative depolarization of Hermissenda type B photoreceptors, a short-term neural correlate of associative learning, was produced by simulating associative training in the isolated nervous system (in vitro conditioning). This simulation entailed stimulation and recording from three classes of neurons normally affected by the associative training procedure: a type B photoreceptor, the silent/excitatory (S/E) optic ganglion cell, and a statocyst caudal hair cell. Exposure of the isolated nervous system to five simultaneous pairings of light and current-induced impulse activity of the caudal hair cell resulted in an average 10-mV depolarization of type B cells. Cumulative depolarization was found to be pairing specific, to occur with a minimal number of training trials, and was paralleled by short-term pairing-specific changes in phototactic behavior for the intact animal. Two important determinants of cumulative depolarization were found to be the magnitude and duration of the long-lasting depolarization (LLD) response of type B cells to light, and a pairing-specific synaptic facilitation of the LLD response. The synaptic facilitation arose from two distinct sources: increased excitatory postsynaptic potential (EPSP) feedback on B cells following light and caudal hair cell stimulation pairings, and disinhibition of the type B photoreceptor following pairings. The S/E optic ganglion cell was found to be a potent regulator of B cell EPSPs. Cumulative depolarization was substantially reduced when the S/E cell was hyperpolarized throughout the course of pairings. Conversely, pairings of light with depolarizing current stimulation of the S/E cell were sufficient to produce cumulative depolarization of B cells. Precluding disinhibition of the B cell from the caudal hair cell was also found to attenuate cumulative depolarization. Additional constraints, inherent to the neural organization of the visual and statocyst neural systems were found to further limit the degree of cumulative depolarization. Among the most important of these were the interpairing interval and light intensity. Exposure of intact animals of five pairings of light and rotation resulted in short-term suppression of phototactic behavior. Like the cumulative depolarization of B cells with in vitro conditioning procedures, these changes were relatively pairing specific and persisted for comparable durations of time. Cumulative depolarization of B cells appears to be an important initial step in the production of long-term associative neural and behavioral changes in Hermissenda.

Associatively reduced withdrawal from shadows in Hermissenda: a direct behavioral analog of photoreceptor responses to brief light steps

When the nudibranch Hermissenda crassicornis encounters a shadow in an otherwise uniformly illuminated field, it stops and turns back into the light within seconds. Associative conditioning, with paired light and rotation stimuli, produces learned modifications of phototaxis in illumination gradients. This same training procedure significantly reduced the ability of paired, but not random or naive control animals, to withdraw from shadows. In naive animals, after 13 min of dark adaptation, withdrawal from shadows was less apparent when animals encountered this stimulus the first time than after the second encounter. This difference in responsiveness to the first and second edge stimulus paralleled differences in type B photoreceptor impulse frequencies recorded during and after first and second steps of light. Earlier studies have shown that associative training of Hermissenda increases a long-lasting depolarization (LLD) which follows a light step. Our present findings suggest a functional relationship between the LLD of the type B photoreceptor and the behavioral response to light-dark differences. This supports the view that membrane changes which cause modifications of LLD magnitude store the learned association for later recall.

Effects of alpha 2-adrenergic agonists and antagonists on photoreceptor membrane currents

Type B photoreceptors of the nudibranch mollusc Hermissenda crassicornis receive excitatory synaptic potentials (EPSPs) whose frequency is controlled by potential changes of a neighboring cell known as the S optic ganglion cell which is thought to be electrically coupled to the presynaptic source of these EPSPs, the E optic ganglion cell. The frequency of the EPSPs increases when a conditioned stimulus (light) is paired with an unconditioned stimulus (rotation) during acquisition of a Pavlovian conditioned response. The results of the present study are consistent with an adrenergic origin for these EPSPs. Noradrenergic agonists (greater than 100 microM), norepinephrine and clonidine, only slightly depolarize the type B cell but clearly prolong its depolarizing response to light. Serotonin, by contrast, causes hyperpolarization of the type B cell's resting potential as well as after a light step. Clonidine reduces voltage-dependent outward K+ currents (IA, an early current, ICa2+-K+, a late Ca2+-dependent current) that control the type B cell's excitability (and thus its light response and membrane potential). These effects of clonidine are reduced or blocked by the alpha 2-receptor antagonist, yohimbine (0.5 microM), but not the alpha 1-blocker, prazosin. The same yohimbine concentration also blocked depolarizing synaptic excitation of the type B cell in response to depolarization of a simultaneously impaled S optic ganglion cell. Histochemical techniques (both the glyoxylic acid method of de la Torre and Surgeon and the formaldehyde-induced fluorescence or Falck-Hillarp method) demonstrated the presence of a biogenic amine(s) within a single neuron in each optic ganglion as well as three or four cells within the vicinity of previously identified visual interneurons. No serotonergic neurons were found within the optic ganglion or in proximity to visual interneurons. A clonidine-like synaptic effect on type B cells, therefore, could amplify conditioning-specific changes of membrane currents by increasing type B depolarization and possibly, as well, by elevating intracellular second messengers.

Implicating causal relations between cellular function and learning behavior

Learning in the nudibranch mollusc Hermissenda shows many features of vertebrate associative conditioning. Pairings of light and rotation produce conditioned suppression of phototaxis, which is retained for days, shows savings, extinction, contingency sensitivity, and, recently, temporal specificity. In addition, specific features of the behavior have been shown to undergo classical Pavlovian conditioning. Extensive analysis of the neural networks mediating the flow of visual and graviceptive information have demonstrated convergent pathways at specific cellular loci. These cells are critically implicated for a primary role in the conditioned modifications of behavior. A variety of experimental approaches consistently support the proposal that reductions of specific K+ currents in the Type B photoreceptor soma play a causal role for several different behavioral expressions of the conditioning. In this article, we review several of these behaviors to show how the demonstrated close temporal correspondence of cellular and behavioral functions further implicates certain causal relations. For example, studies of the shadow withdrawal behavior of Hermissenda suggest a causal relation between the long-lasting depolarization of the Type B photoreceptor and the animal's reduced ability to turn towards the light at light/dark boundaries. Whereas the shadow response corresponded to cellular events at the end of a light step, responses to the onset of light or rotation were largely unexplored. By using a different approach, we identified behavioral responses during the first few seconds of stimulation with light and rotation. These responses, for which Pavlovian conditioning was demonstrated, correspond closely in time to known cellular correlates.(ABSTRACT TRUNCATED AT 250 WORDS)

Inositol trisphosphate regulation of photoreceptor membrane currents

In previous studies elevation of intracellular Ca2+ was shown to cause prolonged reduction of two voltage-dependent K+ currents (IA and ICa2+-K+) across the membrane of the isolated Hermissenda photoreceptor, the type B cell (Alkon et al., 1982b; Alkon and Sakakibara, 1985). Here we show that iontophoretic injection of inositol trisphosphate (IP3), but not inositol monophosphate, also caused prolonged reduction of IA and ICa2+-K+. IP3 injection also caused reduction of a light-induced K+ current (also ICa2+-K+) but did not affect the voltage-dependent Ca2+ current, ICa2+, or the light-induced inward current, INa+, of the type B cell. IP3 injection caused similar effects on the K+ currents of the other type of Hermissenda photoreceptor, the type A cell. INA+ of the type A cell, unlike that of the type B cell, was, however, markedly increased following IP3 injection. The differences of IP3 effects on the two types of photoreceptors may be related to differences in regulation of ionic currents by endogenous IP3 as reflected by clear differences (before injection) in the magnitude of IA, ICa2+-K+, and INa+ between the two cell types.

Ca2+/diacylglycerol-activated, phospholipid-dependent protein kinase in the Hermissenda CNS

In mammalian systems, Ca2+/diacylglycerol-activated phospholipid-dependent protein kinase (C-kinase) appears to play an important role in regulating physiological responses that outlast the transient rise in cytosolic Ca2+. Electrophysiological experiments in neurons of the nudibranch mollusc, Hermissenda crassicornis, have suggested a role for C-kinase in the long-lasting reductions in early and late K+ currents that have been observed following associative learning. Accordingly, we have investigated the catalytic properties of C-kinase in Hermissenda CNS. Following homogenization in Ca2+-free buffer, C-kinase can be separated from Ca2+/calmodulin-dependent protein kinase by centrifugation; C-kinase activity is found in the supernatant whereas essentially all of the Ca2+/calmodulin-dependent protein kinase is found in the membrane fraction. Addition of Ca2+, phosphatidylserine, and diacylglycerol to the cytosol results in phosphorylation of at least eight endogenous proteins. The Hermissenda CNS C-kinase can also phosphorylate lysine-rich histone, a substrate for mammalian C-kinase. The molluscan enzyme exhibits phospholipid specificity in that phosphatidylserine is much more effective than phosphatidylethanolamine, phosphatidylcholine, phosphatidylinositol, and phosphatidic acid. Addition of diacylglycerol, in the presence of Ca2+ and phosphatidylserine, increases the activity of the C-kinase. The percentage of activation by diacylglycerol is larger at lower Ca2+ concentrations. Enzyme activity is inhibited by trifluoperazine and polymixin B sulfate. These studies indicate that the Hermissenda C-kinase is catalytically similar to mammalian C-kinase.

Autoradiographic measurement of tritiated agmatine as an indicator of physiologic activity in Hermissenda visual and vestibular neurons

[3H]Agmatine (amino-4-guanidobutane) has been shown to be potentially useful for identifying and assessing the ACh sensitivity of specific neurons. Small cationic amines are able to permeate ACh-activated ion channels in sympathetic neurons and vertebrate endplates. Sensory neurons of the photic pathway in the nudibranch mollusc Hermissenda crassicornis are cholinergic and the synaptic interactions between the photic and vestibular systems have been well characterized electrophysiologically. We have therefore tested the feasibility of using autoradiography with [3H]agmatine, (a) to identify known ACh-responsive postsynaptic cells and (b) to examine its ability to serve as an indicator of physiologic activity within the photic and vestibular pathways under conditions of darkness and light stimulation. Scintillation counting revealed that approximately 70% of the radioactivity was associated with the CNS while approximately 30% was found in the processing fluids, indicating that routine glutaraldehyde-osmium fixation and subsequent processing for epoxy embedding allows retention of substantial amounts of the radiolabel. The autoradiographic results consistently demonstrated that the uptake patterns for [3H]agmatine did reflect some of the known neuronal interactions under the experimental conditions of light and dark. The accuracy extended to the second order cells of the optic ganglion and to putative interneurons along the photic tract in the cerebropleural ganglion. Since all the neurons in these pathways are unipolar with their synaptic interactions occurring only at the terminal endings, the radiolabel accumulated in the somata resulted from retrograde axonal transport. In the photic-vestibular pathways, the highest silver grain densities were found over structures (cell bodies or axon tracts) with increased synaptic activity coupled with higher levels of cellular activity (i.e. increased excitatory postsynaptic potentials or increased spontaneous impulse activity). Slightly less label was found in cells which received increased numbers of inhibitory postsynaptic potentials that produced hyperpolarization and a transient cessation of impulse activity under conditions of illumination. Therefore, the uptake levels of [3H]agmatine as revealed by autoradiography appear to reflect not only changes in sensitivity or density of ACh-activated channels but also changes in cellular activity as indicated by increased amounts of retrograde transport. These results represent the first example of the effective use of this radiolabel as an indicator of synaptic activity in invertebrates and in sensory systems.

The role of neurochemical modulation in learning

Tsukahara creatively exploited the advantages of a "simple system" approach in a vertebrate context to gain cellular insights into the learning process. The molluscs Aplysia and Hermissenda have provided useful invertebrate examples of this approach. For classical conditioning of Hermissenda a temporal sequence of cellular transformations has been found to correspond to and to substantially account for a learning-specific behavioral transformation. For at least days after the conditioning a biophysical record persists: two voltage-dependent K+ currents, IA and ICa2+-K+, remain reduced in amplitude and at least IA shows an increased rate of inactivation. More recently, a similar biophysical record of associative memory has been identified in the mammalian brain (Disterhoft et al., 1986). Other experiments suggest that a synergistic interaction of C-kinase activation with Ca2+/CaM-kinase activation enhances and prolongs Ca2+-mediated K+ current reduction. The effects of alpha-receptor agonists to enhance depolarization of type B cells (a site of visual-vestibular convergence) and in turn acquisition of classical conditioning are in contrast to the effects of serotonin which can hyperpolarize and thereby reduce depolarization during the acquisition process. For both LTP and LTD, application of a neurotransmitter itself is not sufficient to produce long-lasting neural modification. In this respect, both the LTP and LTD models are more similar to the biochemical sequence implicated in Hermissenda conditioning than to the mechanism initiated by serotonin-like substances proposed for Aplysia sensitization.

Modulation of calcium-mediated inactivation of ionic currents by Ca2+/calmodulin-dependent protein kinase II

Iontophoretic injection of Ca2+ causes reduction of I0A (an early rapidly activating and inactivating K+ current) and I0C (a late Ca2+-dependent K+ current) measured across the isolated type B soma membrane (Alkon et al., 1984, 1985; Alkon and Sakakibara, 1984, 1985). Similarly, voltage-clamp conditions which cause elevation of [Ca2+]i are followed by reduction of I0A and I0C lasting 1-3 min. Iontophoretic injection of highly purified Ca2+/CaM-dependent protein kinase II (CaM kinase II) isolated from brain tissue (Goldenring et al., 1983) enhanced and prolonged this Ca2+-mediated reduction of I0A and I0C. ICa2+, a voltage-dependent Ca2+ current, also showed some persistent reduction under these conditions. Iontophoretic injection of heat-inactivated enzyme had no effect. Agents that inhibit or block Ca2+/CaM-dependent phosphorylation produced increased I0A and I0C amplitudes and prevented the effects of CaM kinase II injection. The results reported here and in other studies implicate Ca2+-stimulated phosphorylation in the regulation of type B soma ionic currents.

Classical conditioning of Hermissenda: origin of a new response

Training of the marine snail Hermissenda crassicornis with paired light and rotation was previously shown to result in acquisition and retention of a behavioral change with many features characteristic of vertebrate associative learning. Here, this behavioral change is demonstrated to be classical, Pavlovian-like conditioning. A new response to light is formed (the CR) that is pairing-specific and resembles the unconditioned response (UCR) to rotation. The conditioned and unconditioned responses are relatively rapid, occurring within seconds of the onset of light or rotation stimuli, and correspond to pairing-specific reductions in speed during the same time period. Since the CR is independent of the presentation of rotation, and it is also expressed by the same effector system (the foot) responsible for the UCR, light stimulation has assumed some of the functional character of rotation.

Conditioning-specific membrane changes of rabbit hippocampal neurons measured in vitro

Intracellular recordings were made from hippocampal CA1 pyramidal neurons within brain slices of nictitating membrane conditioned, pseudoconditioned, and naive adult male albino rabbits. All neurons included (26 conditioned, 26 pseudoconditioned, and 28 naive) had stable penetration and at least 60 mV action potential amplitudes. Mean input resistances were approximately equal to 60 mu omega for the three groups. A marked reduction in the afterhyperpolarization (AHP) following an impulse was apparent for conditioned (x = -0.98 mV) as compared to the pseudoconditioned (x = -1.7 mV) and naive (x = -2.0 mV) neurons. The AHP has been attributed previously to activation of a Ca2+-dependent outward K+ current. The distribution of AHP amplitudes for the conditioned group included a new lower range of values for which there was little overlap with the other groups. The conditioning-specific reduction of AHP may be due to reduction of ICa2+-K+ as shown previously for conditioned Hermissenda neurons. This conditioning-induced biophysical alteration of the CA1 pyramidal cell must be stored by mechanisms intrinsic to the hippocampal slice and cannot be explained as a consequence of changes of presynaptic input arising elsewhere in the brain. Our experiments demonstrate the feasibility of analyzing cellular mechanisms of associative learning in mammalian brain with the in vitro brain slice technique.

C-kinase activation prolongs Ca2+-dependent inactivation of K+ currents

Voltage-dependent K+ currents, IA and ICa2+-K+, across the soma membrane of the Hermissenda Type B photoreceptor, have been shown to remain reduced during retention of classically conditioned behavior. IA and ICa2+-K+ undergo prolonged reduction due to [Ca2+]i elevation produced by a single pairing of a light step with a command depolarization or by iontophoretic injection of Ca2+. One pathway which could contribute to the conversion of transient Ca2+-mediated reduction of K+ currents to the persistent reduction observed with conditioning is that involving C-kinase. To examine the role of C-kinase in the long-term regulation of K+ currents, isolated Type B somata were exposed to at least 25-30 minutes' incubation in artificial sea water (ASW) containing the C-kinase activators 1-oleoyl-2-acetyl-glycerol (OAG) or 12-deoxyphorbol 13-isobutyrate 20-acetate (DPBA) or control substances [e.g., distearyolglycerol (DiSG)]. After exposure to activator (but not to control solutions) and voltage-clamp conditions which caused elevation of cytosolic Ca2+, reductions of IA and ICa2+-K+ were observed which did not reverse (up to 3 hr), even after the activator was removed. Without conditions which induced elevation of cytosolic calcium prolonged incubation with the C-kinase activators had no effect on the membrane currents. Similar exposure of homogenates of the Hermissenda nervous system to OAG and Ca2+ caused enhanced phosphorylation of specific proteins, indicating the presence of C-kinase in the Hermissenda nervous system.

Implicating causal relations between cellular function and learning behavior

Learning in the nudibranch mollusc Hermissenda shows many features of vertebrate associative conditioning. Pairings of light and rotation produce conditioned suppression of phototaxis, which is retained for days, shows savings, extinction, contingency sensitivity, and, recently, temporal specificity. In addition, specific features of the behavior have been shown to undergo classical Pavlovian conditioning. Extensive analysis of the neural networks mediating the flow of visual and graviceptive information have demonstrated convergent pathways at specific cellular loci. These cells are critically implicated for a primary role in the conditioned modifications of behavior. A variety of experimental approaches consistently support the proposal that reductions of specific K+ currents in the Type B photoreceptor soma play a causal role for several different behavioral expressions of the conditioning. In this article, we review several of these behaviors to show how the demonstrated close temporal correspondence of cellular and behavioral functions further implicates certain causal relations. For example, studies of the shadow withdrawal behavior of Hermissenda suggest a causal relation between the long-lasting depolarization of the Type B photoreceptor and the animal's reduced ability to turn towards the light at light/dark boundaries. Whereas the shadow response corresponded to cellular events at the end of a light step, responses to the onset of light or rotation were largely unexplored. By using a different approach, we identified behavioral responses during the first few seconds of stimulation with light and rotation. These responses, for which Pavlovian conditioning was demonstrated, correspond closely in time to known cellular correlates.(ABSTRACT TRUNCATED AT 250 WORDS)