Associative Learning in Hermissenda

Few features of conscious experience have captured the human imagination more than the proclivity of animals to learn and to retain the consequences of experience in memory. Learning not only provides for the adaptation of organisms to changing environmental demands, but also, and more important, for the persistence of learning—that is, long-term memory—which provides us with a history of human experience. In spite of the widespread interest in learning and memory, their basic mechanisms remain among the least thoroughly understood areas of physiology.

An attractive experimental approach to this problem at a fundamental level is the analysis of learning and memory in the less-complex central nervous system of invertebrates. One animal that has contributed to the physiology of learning and memory is the nudibranch mollusk Hermissenda crassicornis, whose behavior can be modified by a Pavlovian conditioning procedure. The Hermissenda central nervous system is relatively simple, consisting of many identifiable neurons that can be studied in detail using biochemical, biophysical, and molecular techniques. An additional advantage is that the two sensory structures and their central pathways supporting the conditioned stimulus (CS) and unconditioned stimulus (US) are totally intact in the isolated central nervous system. This attractive feature facilitates the search for cellular correlates of learning that have been identified and have been the focus of biochemical and molecular analyses.

Pavlovian Conditioning

Pavlovian conditioning of Hermissenda involves changes in light-elicited locomotion and foot length (conditioned responses, CRs) produced by stimulation of the visual and vestibular systems with their adequate stimuli (Crow and Alkon, 1978; Lederhendler, Gart, and Alkon, 1986). The Pavlovian conditioning procedure consists of pairing light, the conditioned stimulus (CS), with high-speed rotation, the unconditioned stimulus (US). As shown in Figure 1, after conditioning, the CS suppresses normal light-elicited locomotion and elicits foot shortening. Retention of conditioned behavior persists for several days to weeks depending upon the number of conditioning trials used in initial acquisition (Alkon, 1989; Crow and Alkon, 1978). Pavlovian conditioning in Hermissenda exhibits CS specificity and is dependent upon the association of the two sensory stimuli involving both contiguity and contingency. Crow and Offenbach (1983) showed that conditioned animals exhibit suppressed locomotor behavior in the presence of the CS; however, their locomotor behavior in the dark was not significantly changed. Nonassociative contributions to behavior are expressed in the initial trials of the conditioning session and the decrement rapidly following the termination of multitrial conditioning.

In addition to multiple-trial conditioning of suppression of light-elicited locomotion and foot contraction, one-trial conditioning also modifies light-elicited locomotion (Crow and Forrester, 1986). Pairing the CS with the direct application of one of the transmitters of the US pathway (serotonin 5-HT, nominal US) to the exposed nervous system of otherwise intact Hermissenda produces suppression of light-elicited locomotion when the animals are tested twenty-four hours following the one conditioning trial. One-trial conditioning also produces enhanced excitability of type-B photoreceptors (see Figure 1), a component of the CS pathway that expresses cellular plasticity produced by multitrial Pavlovian conditioning (discussed further below).

Cellular and Synaptic Plasticity Associated with Pavlovian Conditioning

An essential step in the analysis of Pavlovian conditioning is the search for the loci in the animal's nervous system where memories of the associative experience are stored. Crow and Alkon (1980) identified the primary sensory neurons (photoreceptors) of the pathway mediating the CS as one site for memory storage. Intrinsic modifications of cellular and synaptic plasticity in classically conditioned animals involve both enhanced excitability and synaptic facilitation of connections between sensory neurons in the pathway mediating the CS (Alkon, 1989; Crow and Alkon, 1980; Frysztak and Crow, 1994). Enhanced excitability in identified photoreceptors of conditioned Hermissenda is expressed by a significant increase in spike activity elicited by the CS or extrinsic current, an increase in the input resistance, an alteration in the amplitude of light-elicited generator potentials, decreased spike frequency accommodation, and a reduction in the peak amplitude of voltage-dependent (I_A, I_Ca) and Ca²⁺-dependent (I_{K, Ca}) currents (Alkon et al., 1985; for reviews, see Alkon, 1989; Crow, 1988; Sahley and Crow, 1998). Enhanced excitability, expressed by an increase in both the amplitude of CS-elicited generator potentials and the number of action potentials elicited by the CS, may be a major contributor to changes in the duration and amplitude of CS-elicited complex postsynaptic potentials (PSPs) and enhanced CS-elicited spike activity observed in postsynaptic targets. However, changes in the strength of synaptic connections between identified type-B photoreceptors and other components of the CS pathway have also been detected following conditioning (Frysztak and Crow, 1994).

Facilitation of the amplitude of unitary inhibitory postsynaptic potentials (IPSPs) elicited by single spikes in identified type-B photoreceptors are detected in type-A photoreceptors and type-I_ii interneurons of conditioned animals (see Figure 1). Facilitation of type-I_e interneuron EPSPs (excitatory postsynaptic potentials) elicited by lateral type-B spikes is also observed following conditioning. Studies of the signal transduction pathways responsible for the modification of diverse K⁺ currents of type-B photoreceptors of conditioned animals have identified several second messenger systems. Both protein kinase C (PKC) (Farley and Auerbach, 1986; Crow et al., 1991) and extracellular signal-regulated protein kinase (ERK) (Crow et al., 1998) contribute to modifications of excitability and synaptic efficacy of conditioned Hermissenda. A second site of cellular plasticity in conditioned animals is the type-A photoreceptor. Lateral type-A photoreceptors of conditioned animals exhibit an increase in CS-elicited spike frequency, a decrease in generator potential amplitude, and enhanced excitability and decreased spike frequency accommodation to extrinsic current (Frysztak and Crow, 1993). Taken collectively, the evidence for localization of cellular changes in the CS pathway indicates that multiple sites of cellular and synaptic plasticity involving changes in both excitability and synaptic strength exist in the photoreceptors and interneurons of conditioned animals (see Figure 1). Anatomical studies of type-B photoreceptors indicate the existence of spatially segregated compartments (Alkon, 1989). Phototransduction occurs in the soma-rhabdomeric compartment, spike generation in the distal axon, and synaptic interactions in the axon terminal regions within the cerebropleural neuropil. Therefore, a decrease in K⁺ conductances of type-B photoreceptors could contribute both directly and indirectly to enhanced excitability by increasing the amplitude of CS-elicited generator potentials and increasing CS-elicited spike activity in the spike-generating zone by modification of conductance that influence the interspike interval.

Mechanisms of Memory Consolidation Underlying Pavlovian Conditioning

Studies of memory have identified components of memory consolidation that can be differentiated based upon the contribution of signal transduction pathways, protein synthesis, and gene induction (for review, see DeZazzo and Tully, 1995). An analysis of one-trial conditioning in Hermissenda has provided insights into the mechanisms of different stages of memory consolidation. One-trial conditioning produces long-term suppression of light-elicited locomotion (Crow and Forrester, 1986) and short-term and long-term enhancement of excitability in sensory neurons of the CS pathway (Crow and Forrester, 1991, 1993). Short-term and long-term enhanced excitability appear to be independent, parallel processes, because long-term enhanced excitability can be expressed in the absence of prior short-term enhanced excitability (Crow and Forrester, 1993). Moreover, enhanced excitability in type-B photoreceptors follows a biphasic pattern in its development following one-trial conditioning. Excitability reaches a peak three hours after one-trial conditioning, declines toward baseline control levels five to six hours after conditioning, and is followed by an increase to a stable plateau at sixteen to twenty-four hours postconditioning. In addition, one-trial conditioning produces an intermediate stage of memory that depends on translation but not transcription, whereas long-term memory for enhanced excitability depends upon both translation and transcription (Crow and Forrester, 1990; Crow, Xue-Bian, and Siddiqi, 1999).

Associated with intermediate memory is the phosphorylation of a 24 kDa protein (CSP24). A one-trial conditioning procedure that only produces short-term memory does not result in the increased phosphorylation of CSP24 (Crow and Xue-Bian, 2000). Therefore, the regulation of this phosphoprotein by one-trial conditioning is associated with experimental conditions that produce intermediate and long-term memory. The protein-synthesis inhibitor anisomycin, present during the intermediate phase of memory, blocked the increased phosphorylation of CSP24 but did not block the increased phosphorylation of other proteins associated with one-trial conditioning (Crow, Xue-Bian, and Siddiqi, 1999). Experiments examining 35 S-methionine labeling of CSP24 during the intermediate phase of memory showed similar labeling of CSP24 for the conditioned group and unpaired controls. However, both conditioned and unpaired groups were greater than unstimulated controls with respect to 35 S-methionine labeling of CSP24. Therefore the requirement for protein synthesis is necessary but not sufficient for long-term associative memory, or it may reflect an indirect involvement in phosphorylation due to changes in the synthesis of protein kinases in signal transduction pathways or phosphatase inhibitors. Experiments where CSP24 was excised from multiple two-dimensional gels and subjected to reverse phase HPLC (high pressure liquid chromatography) and automated sequence analysis showed that the sequenced peptides exhibited a homology to the β -thymosin family of actin-binding proteins (Crow and Xue-Bian, 2000). All known vertebrate and invertebrate β -thymosins bind actin monomers (Nachimias, 1993). Cytoskeletal-related proteins such as CSP24 thus may contribute to long-term structural remodeling in the CS pathway by regulating the turnover of actin filaments during the intermediate-term transition period between short-and long-term memory.

Conclusion

One-trial and multi-trial Pavlovian conditioning produce changes in both synaptic efficacy and cellular excitability in several identified neurons of the pathway supporting the CS. Activation of PKC and ERK is produced by both one-trial and multi-trial Pavlovian conditioning of Hermissenda. Enhanced cellular excitability expressed by CS-elicited spike activity or extrinsic current results from the reduction in several K⁺ conductances in type-B photoreceptors of conditioned animals. Studies of memory formation in Hermissenda have shown that intermediate and long-term memory produced by one-trial conditioning involves the synthesis and phosphorylation of cellular proteins. One protein, CSP24, a β -thymosin-like protein associated with intermediate memory, is regulated by conditioning and may contribute to reorganization of the actin cytoskeleton underlying structural remodeling supporting long-term memory.

See also:SECOND MESSENGER SYSTEMS

Bibliography

Alkon, D. L. (1989). Memory storage and neural systems. Scientific American 261 (1), 42-50.

Alkon, D. L., Sakakibara, M, Forman, R., Harrigan. J., Lederhendler, I., and Farley, J. (1985). Reduction of two voltage-dependent K⁺ currents mediates retention of a learned association. Behavioral and Neural Biology 44, 278-300.

Crow, T. (1988). Cellular and molecular analysis of associative learning and memory in Hermissenda. Trends in Neurosciences 11, 136-142.

Crow, T., and Alkon, D. L. (1978). Retention of an associative behavioral change in Hermissenda. Science 201, 1,239-1,241.

—— (1980). Associative behavioral modification in Hermissenda : Cellular correlates. Science 209, 412-414

Crow, T., and Forrester, J. (1986). Light paired with serotonin mimics the effects of conditioning on phototactic behavior in Hermissenda. Proceedings of the National Academy of Sciences of the United States of America 83, 7,975-7,978.

—— (1990). Inhibition of protein synthesis blocks long-term enhancement of generator potentials produced by one-trial in vivo conditioning in Hermissenda. Proceedings of the National Academy of Sciences of the United States of America 87, 4,490-4,494.

—— (1991). Light paired with serotonin in vivo produces both short-and long-term enhancement of generator potentials of identified B-photoreceptors in Hermissenda. Journal of Neuroscience 11, 608-617.

—— (1993). Down-regulation of protein kinase C and kinase inhibitors dissociate short-and long-term enhancement produced by one-trial conditioning of Hermissenda. Journal of Neurophysiology 69, 636-641.

Crow, T., Forrester, J., Williams, M., Waxham, N., and Neary, J. (1991). Down regulation of protein kinase C blocks 5-HT-induced enhancement in Hermissenda B photoreceptors. Neuroscience Letters 121, 107-110.

Crow, T., and Offenbach, N. (1983). Modification of the initiation of locomotion in Hermissenda : Behavioral analysis. Brain Research 271, 301-310.

Crow, T. J., and Xue-Bian, J. J. (2000). Identification of a 24 kDa phosphorylation associated with an intermediate stage of memory in Hermissenda. Journal of Neuroscience 20 (10), RC74, 1-5.

Crow, T., Xue-Bian, J. J., and Siddiqi, V. (1999). A protein synthesis-dependent and mRNA synthesis-independent intermediate phase of memory in Hermissenda. Journal of Neurophysiology 82 (1), 495-500.

Crow, T., Xue-Bian, J. J., Siddiqi, V., Kang, Y., and Neary, J. T. (1998). Phosphorylation of mitogen-activated protein kinase by one-trial and multi-trial classical conditioning. Journal of Neuroscience 18 (9), 3,480-3,487.

DeZazzo, J., and Tully, T. (1995). Dissection of memory formation from behavioral pharmacology to molecular genetics. Trends in Neuroscience 18, 212-218.

Farley, J., and Auerbach, S. (1986). Protein kinase C activation induces conductance changes in Hermissenda photoreceptors like those seen in associative learning. Nature 319, 220-223.

Frysztak, R. J., and Crow, T. (1993). Differential expression of correlates of classical conditioning in identified medial and lateral type-A photoreceptors of Hermissenda. Journal of Neuroscience 13 (7), 2,889-2,897.

—— (1994). Enhancement of type-B-and type-A photoreceptor inhibitory synaptic connections in conditioned Hermissenda. Journal of Neuroscience 14 (3), 1,245-1,250.

Lederhendler, I., Gart, S., and Alkon, D. L. (1986). Classical conditioning of Hermissenda : Origin of a new response. Journal of Neuroscience 6, 1,325-1,331.

Nachmias, V. T. (1993). Small actin-binding proteins: the β -thymosin family. Current Opinion in Cell Biology 5, 56-62.

Sahley, C., and Crow, T. (1998). Invertebrate learning: Current perspectives. In J. Martinez and R. Kesner, eds., Neurobiology of Learning and Memory, pp. 171-209. New York: Academic Press.

Terry J.Crow