Reference: Biol. Bull. 210: 271–279. (June 2006) A contribution to The Biological Bulletin Virtual Symposium © 2006 Marine Biological Laboratory on Marine Invertebrate Models of Learning and Memory. The Cellular Mechanisms of Learning in Aplysia: Of Blind Men and Elephants DAVID L. GLANZMAN* Department of Physiological Science, UCLA College, and Department of Neurobiology and the Brain Research Institute, David Geffen School of Medicine at UCLA, Los Angeles, California 90095 Abstract. Until recently, investigations of the neurobio- That each by observation logical substrates of simple forms of learning and memory Might satisfy his mind. in the marine snail Aplysia have focused mostly on plastic In the fable each of the blind men feels a single part of the changes that occur within the presynaptic sensory neurons. elephant and comes to a very different conclusion about the Here, I summarize the results of recent studies that indicate nature of the animal. While the man who holds the ele- that exclusively presynaptic processes cannot account for phant’s trunk concludes that the elephant is “very like a simple forms of learning in Aplysia. In particular, I present snake,” another, who feels only the elephant’s ear, claims evidence that postsynaptic mechanisms play a far more that the beast resembles a fan. Yet a third blind man, who important role in nonassociative learning in Aplysia than has touches one of the tusks, argues that the animal is like a been appreciated before now. Moreover, I describe recent spear. As the poem ends, data that suggests the intriguing hypothesis that the persis- tent, learning-induced changes in Aplysia sensory neurons And so these men of Indostan might depend critically on postsynaptic signals for their Disputed loud and long, induction. Finally, I discuss the potential applicability of Each in his own opinion this hypothesis to learning-related synaptic plasticity in the Exceeding stiff and strong, Though each was partly in the right, mammalian brain. And all were in the wrong! Moral: Introduction So oft in theologic wars, The disputants, I ween, A review of the literature on learning-related synaptic Rail on in utter ignorance plasticity from the last two decades brings to mind the fable Of what each other mean, of the blind men and the elephant. As re-told in John And prate about an Elephant Godfrey Saxe’s poem, Not one of them has seen! It was six men of Indostan The intensity of the scientific debates concerning one To learning much inclined, prominent form of synaptic plasticity, long-term potentia- Who went to see the Elephant tion (LTP) of synapses in the mammalian hippocampus (Though all of them were blind), (Malenka and Nicoll, 1999), has probably seemed at times reminiscent of “theologic wars” to those outside the field. Received 17 February 2006; accepted 13 April 2006. The ongoing dispute over the site, whether pre- or postsyn- * To whom correspondence should be addressed, at Gonda (Gold- aptic, of LTP expression—a term that refers to the cellular schmied) Neuroscience and Genetics Research Center, UCLA, 695 Young modifications that underlie the persistence of the synaptic Drive South, Box 951761, Los Angeles, CA 90095-1761. E-mail: enhancement—has been especially ferocious. In the past [email protected] Abbreviations: EPSP, excitatory postsynaptic potential; LLH, long-last- several years, however, it has become increasingly evident ing habituation; LTD, long-term depression; LTH, long-term habituation; that expression of LTP of synapses in the CA1 region of the LTP, long-term potentiation; STH, short-term habituation. hippocampus almost certainly involves both postsynaptic 271 272 D. L. GLANZMAN changes—particularly modulation of the trafficking of man, 2003) tested whether persistent habituation of the ␣-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid gill-withdrawal reflex depends on activation of glutamate (AMPA)-type glutamate receptors (Nicoll, 2003)—and receptors within the abdominal ganglion. (Ezzeddine and I presynaptic changes (Antonova et al., 2001; Choi et al., originally termed this form of habituation long-lasting ha- 2003; Zakharenko et al., 2003; Powell, 2006). This bituation, or LLH. Because we have recently found that the emerging consensus raises the intriguing question of persistent habituation requires gene expression [see below], whether any form of persistent (lasting for more than however, we now refer to the learning as long-term habit- Ϸ10 min), learning-induced synaptic plasticity can result uation [LTH].) In our experiments, gill withdrawal was from changes to only one side of a synapse. stimulated by weak electrical stimulation of the siphon. Many neuroscientists, if this question were posed to Electrodes were implanted into both sides of the siphon; them, would be likely to respond affirmatively; moreover, stimulation of one side of the siphon was used for habitu- they would probably cite the research on learning and ation training, while the other side of the siphon served as memory in Aplysia (Byrne and Kandel, 1996; Kandel, 2001) the control (untrained) side. The strength of the withdrawal as providing the strongest evidence for “one-sided” synaptic reflex in response to siphon stimulation was measured with plasticity. As described in most introductory textbooks of a force transducer attached to the gill. The artery leading to neuroscience, this mollusc exhibits a simple withdrawal the abdominal ganglion was cannulated so that drugs could reflex that can exhibit several simple forms of learning, be delivered directly to this ganglion, which contains those including habituation, sensitization, and classical condition- sensory and motor neurons that mediate the siphon-elicited ing. Until relatively recently, each of these forms of learning gill withdrawal. has been ascribed, predominately or exclusively, to changes We found that LTH of gill withdrawal depends on both in the sensory neurons that mediate the withdrawal reflex. protein synthesis and gene expression, because it is blocked Thus, habituation of the reflex has been thought to be due to by infusion into the abdominal ganglion of either anisomy- homosynaptic (presynaptic) depression (Castellucci and cin (Ezzeddine and Glanzman, 2003) or actinomycin-D Kandel, 1974; Kandel et al., 1975; Armitage and Siegel- (Ezzeddine et al., 2004). Furthermore, LTH requires acti- baum, 1998), sensitization to presynaptic facilitation (Kandel vation of the protein phosphatases 1 and 2A (PP1 and et al., 1975; Byrne and Kandel, 1996; Kandel, 2001), and PP2A) (Ezzeddine and Glanzman, 2003) and calcineurin classical conditioning to activity-dependent presynaptic facili- (Ezzeddine et al., 2004). Surprisingly—and in contrast to tation (Hawkins et al., 1983; Walters and Byrne, 1983; Carew the conclusions of previous cellular analyses of short- and et al., 1984). Within the last decade, however, this tidy view of long-term habituation in Aplysia (Castellucci et al., 1970; the cell biology of learning in Aplysia has undergone signifi- Castellucci and Kandel, 1974; Bailey and Chen, 1983, cant revision. It is now generally acknowledged, for example, 1988a, b)—we discovered that LTH depends on activation that classical conditioning of the withdrawal reflex requires of glutamate receptors. In particular, LTH was blocked Hebbian, N-methyl-D-aspartate (NMDA) receptor-dependent when habituation training was carried out in the presence of LTP of the sensorimotor synapse (Lin and Glanzman, either the NMDA receptor antagonist APV or the AMPA 1994a, b; Murphy and Glanzman, 1997, 1999; Antonov et receptor antagonist DNQX (Ezzeddine and Glanzman, al., 2003), as well as a rise in intracellular Ca2ϩ within the 2003). (Note that testing in the experiments was performed postsynaptic motor neuron (Murphy and Glanzman, 1996; in normal artificial seawater.) The latter result is particularly Bao et al., 1998; Antonov et al., 2003; for review, see interesting, because it distinguishes the mechanism of LTH Roberts and Glanzman, 2003). Moreover, recent evidence from that of classic homosynaptic depression. As shown by indicates that not only associative learning, but also nonas- Armitage and Siegelbaum (1998), short-term homosynaptic sociative learning, in Aplysia depends critically on postsyn- depression of the in vitro sensorimotor synapse proceeds aptic changes. Below I summarize data from my laboratory, normally with repeated activation of the sensory neuron as well as from other laboratories, that implicate postsyn- despite a blockade of (presumably) postsynaptic glutamate aptic mechanisms in habituation and sensitization. receptors by DNQX sufficient to eliminate the sensorimotor excitatory postsynaptic potential (EPSP). Therefore, induc- The Role of Postsynaptic Glutamate Receptors tion of homosynaptic depression does not require activation in Behavioral Habituation of postsynaptic AMPA-type receptors. Taken together, the data of Armitage and Siegelbaum (1998) and those of Repeated, innocuous tactile stimulation of the animal’s Ezzeddine and Glanzman (2003) demonstrate that presyn- siphon produces habituation of the defensive withdrawal aptic depression of transmitter release at sensorimotor syn- reflex in Aplysia. Both short- and long-term forms of habit- apses, although a likely cellular mechanism for short-term uation have been demonstrated (Pinsker et al., 1970; Carew habituation (STH), cannot account for LTH. et al., 1972; Carew and Kandel, 1973). Using a reduced There is good experimental evidence that depression at preparation,