| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review
Pavlovian Conditioning of Hermissenda: Current Cellular, Molecular, and Circuit Perspectives
Department of Neurobiology and Anatomy, University of Texas Medical School, Houston, Texas 77030, USA
ABSTRACT
The less-complex central nervous system of many invertebrates make them attractive for not only the molecular analysis of the associative learning and memory, but also in determining how neural circuits are modified by learning to generate changes in behavior. The nudibranch mollusk Hermissenda crassicornis is a preparation that has contributed to an understanding of cellular and molecular mechanisms of Pavlovian conditioning. Identified neurons in the conditioned stimulus (CS) pathway have been studied in detail using biophysical, biochemical, and molecular techniques. These studies have resulted in the identification and characterization of specific membrane conductances contributing to enhanced excitability and synaptic facilitation in the CS pathway of conditioned animals. Second-messenger systems activated by the CS and US have been examined, and proteins that are regulated by one-trial and multi-trial Pavlovian conditioning have been identified in the CS pathway. The recent progress that has been made in the identification of the neural circuitry supporting the unconditioned response (UR) and conditioned response (CR) now provides for the opportunity to understand how Pavlovian conditioning is expressed in behavior.
). One animal that has contributed to an understanding of the physiology of learning and memory at a cellular, synaptic, and systems level of analysis is the nudibranch mollusk Hermissenda crassicornis. Associative learning in Hermissenda has been extensively examined using a Pavlovian conditioning procedure. The Hermissenda central nervous system is relatively simple, consisting of identifiable neurons in the neural circuitry that supports conditioning. Identified neurons in the CS pathway have been studied in detail using biochemical, biophysical, and molecular techniques. The two sensory structures mediating the CS and US are central, and thus, their synaptic projections remain totally intact after surgical isolation of the nervous system. Mechanisms of CS-US contiguity have been identified and have been the focus of biophysical, biochemical, and molecular analyses. Moreover, because the neurons that contribute to the neural circuitry supporting the unconditioned responses (URs) and conditioned responses (CRs) have been recently identified, and can be studied in semi-intact nervous systems, an explanation of how conditioning is expressed in the generation of behavior is now feasible.
Pavlovian Conditioning
The same Pavlovian conditioning procedure in Hermissenda results
in the acquisition of two different CRs. Pavlovian conditioning produces both
light-elicited inhibition of normal positive phototaxis (Crow and Alkon
1978,
1980;
Crow and Offenbach 1983;
Crow 1985a) and CS-elicited
foot-shortening (Lederhendler et al.
1986). The description of the conditioning paradigm and the two
CRs are summarized in Figure 1.
The conditioning procedure consists of pairing light, the CS with high-speed
rotation, or orbital shaking, the US. Two URs are elicited by rotation, a
reduced rate of forward locomotion and foot-shortening
(Alkon 1974;
Crow and Alkon 1978;
Lederhendler et al. 1986;
Matzel et al. 1990b).
Conditioned inhibition of phototaxis is expressed by a light-dependent
inhibition in the initiation of locomotion
(Crow and Offenbach 1983) and
a reduced rate of forward locomotion in light
(Farley and Alkon 1982;
Matzel et al. 1990b). The two
CRs are proposed to develop independently
(Matzel et al. 1990b), and the
URs may involve different components of the neural circuit responsible for
foot contraction and ciliary locomotion (Crow and Tian
2003a,b).
Retention of conditioned behavior persists for several days to weeks,
depending upon the number of conditioning trials used in initial acquisition
(Crow and Alkon 1978;
Alkon 1983;
Harrigan and Alkon 1985).
Conditioning of phototactic inhibition can be extinguished with the
presentation of nonreinforced CSs
(Richards et al. 1984).
Conditioned inhibition of phototactic behavior also exhibits CS specificity,
as conditioned animals exhibit suppressed locomotor behavior in the presence
of the CS; however, their locomotor behavior in the dark is not significantly
changed (Crow and Offenbach
1983).
|
The experimental conditions that produce Pavlovian conditioning have been
investigated in some detail. Conditioning is dependent upon the temporal
association of the CS and US involving both contiguity
(Crow and Alkon 1978) and
contingency, the predictive relationship between the CS and the US (Farley
1987a,b).
Extra CS and US presentations inserted into a sequence of CS-US pairings
attenuates conditioning (Farley
1987a). However preconditioning exposure to the CS (latent
inhibition) or US pre-exposure does not impair subsequent conditioning
(Farley 1987a). Conditioned
inhibition of phototaxis can be enhanced by compound conditioning in both
overshadowing and blocking paradigms
(Farley et al. 1997).
Potentiation of phototactic suppression is produced by the addition of a
chemosensory stimulus, scallop extract, although second-order conditioning and
sensory preconditioning have not been demonstrated
(Farley et al. 1997). Studies
have also shown that a chemosensory CS, when paired with rotation, suppresses
bite-strike responses normally elicited by the chemosensory CS prior to
conditioning (Farley et al.
1990a). Rogers and Matzel
(1996) reported that an
excitatory context produced by presenting unsignaled USs (rotation) in a
context of chemosensory stimuli blocked later conditioned foot-shortening
produced by a light CS paired with rotation US within that context. More
recently, it was reported that explicitly unpaired presentations of the CS and
US produced conditioned inhibition expressed by increased phototactic behavior
(Britton and Farley 1999).
Behavioral studies of conditioning in Hermissenda have also shown
that sensitization is not an important contributor to conditioned inhibition
of positive phototactic behavior. Nonassociative contributions to phototactic
behavior are expressed in the initial trials of each conditioning session and
decrement rapidly following the termination of multitrial conditioning
sessions (Crow 1983). Because
the magnitude and temporal characteristics of associative and nonassociative
contributions to conditioning are quite different, it has been proposed that
conditioning in Hermissenda is not an elaboration or potentiation of
the mechanisms responsible for nonassociative learning
(Crow 1983).
Conditioning in the two different behavioral response systems supporting
the two CRs is sensitive to both CS-US contiguity and forward
interstimulus-interval manipulations
(Matzel et al. 1990c).
Moreover, both conditioned foot-shortening and conditioned inhibition of
phototaxis involve the development or emergence of a new response to the CS,
not the potentiation, through US presentations of an already existing response
to the CS referred to as reflex potentiation (e.g.,
Schreurs 1989;
Sahley and Crow 1998). In
both of the CRs, there is a transfer of functional aspects of the
response-evoking properties of the US to the CS
(Crow and Alkon 1978;
Lederhendler et al. 1986;
Matzel et al. 1990b). This
feature probably accounts for the increased complexity of the circuit
supporting the CS and US, and the multiple sites of CS-US pathway convergence
in the nervous system, and multiple synaptic interactions within the neural
network supporting behavior. In addition to multiple-trial conditioning of
suppression of light-elicited locomotion and foot-shortening, one-trial
conditioning also inhibits light-elicited locomotion
(Crow and Forrester 1986).
Pairing the CS with the direct application of one of the proposed transmitters
of the US pathway (5-HT, nominal US) to the exposed nervous system of
otherwise intact Hermissenda produces suppression of light-elicited
locomotion when the animals are tested 24 h following the one-conditioning
trial. In addition, procedures for in vitro conditioning of the isolated
nervous system have been developed. In vitro conditioning involves pairing the
CS (light) with stimulation of the statocyst produced by mechanical
perturbations (US). In vitro conditioning involving several conditioning
trials produces similar cellular correlates in type B photoreceptors as found
following in vivo procedures (Farley and
Alkon 1987; Matzel et al.
1996; Gandhi and Matzel
2000).
Anatomy of the CS and US Pathways
The two sensory structures that are stimulated by the CS and US have been
described in detail by Alkon and colleagues
(Alkon and Fuortes 1972; Alkon
1973a,b;
Alkon and Bak 1973;
Detwiler and Alkon 1973). In
addition, the convergence sites providing for synaptic interactions between
the CS and US pathways have been identified (Alkon
1973a,b;
Alkon et al. 1978;
Akaike and Alkon 1980; Crow and
Tian 2000,
2002a,b,
2003a,
2004).
Photoreceptors
Each eye of Hermissenda contains five photoreceptors, three
classified as type B and two as type A. The general classification of
photoreceptors can be identified further on the basis of their location within
the eye. There are medial and lateral A and B photoreceptors and one central B
photoreceptor. The synaptic connections between the type B photoreceptors and
between type B and type A photoreceptors are in the neuropil of the
cerebropleural ganglion and are mutually inhibitory
(Alkon and Fuortes 1972;
Alkon 1973a;
Crow et al. 1979;
Senft et al. 1982;
Frysztak and Crow 1993). Light
produces a depolarizing generator potential and an increase in spike activity
in both type A and B photoreceptors
(Dennis 1967;
Alkon and Fuortes 1972).
Hair Cells
The sensory structures stimulated by the US are the two central gravity
detecting statocysts (Alkon and Bak
1973; Detwiler and Alkon
1973; Detwiler and Fuortes
1973; Alkon 1975).
Each statocyst contains 13 hair cells, whose cell bodies are located around
the perimeter of the statocyst. Hair cells that are located in opposite
positions in the statocyst are mutually inhibitory
(Detwiler and Alkon 1973).
Statocyst hair cells contact calcium carbonate particles, referred to as
statoconia, by interactions with motile cilia that project into the lumen of
the statocyst from the apical region of the somas
(Alkon 1975). Rotation or
gravity causes the statoconia to press against the motile cilia of cells in
front of the centrifugal or gravitational force vector, resulting in a
depolarizing generator potential and an increase in spike activity
(Alkon 1975). Hair cells in
back of the centrifugal force vector hyperpolarize in response to
rotation.
Optic Ganglion Cells
Second-order neurons in the visual system are located in the optic ganglion
(Alkon 1973a;
Tabata and Alkon 1982) and
cerebropleural ganglion (Akaike and Alkon
1980; Crow and Tian
2000,
2002a). Type B photoreceptors,
but not type A photoreceptors, inhibit ipsilateral optic ganglion cells. The
14 optic ganglion cells have been classified into multiple types referred to
as C, D, E, and S (Alkon 1973a;
Tabata and Alkon 1982). The
type E optic ganglion cell is presumed to be electrically coupled to the S
ganglion cell (S-E complex), and produces EPSPs in all ipsilateral type B
photoreceptors, but not type A photoreceptors, and IPSPs in ipsilateral caudal
hair cells. All of the other cells within the same optic ganglion do not have
synaptic interactions, however, type C optic ganglion cells inhibit
contralateral type D optic ganglion cells. Alkon
(1973a) has proposed that the
synaptic interactions between C and D optic ganglion cells would enhance the
contrast between the cells' responses to illumination of each of the two eyes,
which would signal the approach of a moving shadow or light.
The synaptic convergence between the CS and US pathways involving the optic
ganglion is complex and a potential role in plasticity is poorly documented in
conditioned animals. Type B photoreceptors and caudal hair cells inhibit the
S-E optic ganglion cell complex. The S-E cell produces positive feedback by
exciting type B photoreceptors and cephalic hair cells, and inhibiting caudal
hair cells (Tabata and Alkon
1982). Therefore, the S-E cell complex can produce direct and
indirect excitation of ipsilateral type B photoreceptors. The indirect
excitation of B photoreceptors is the result of inhibition of caudal hair
cells that inhibit type B photoreceptors and excitation of cephalic hair cells
that inhibit caudal hair cells. Cephalic and caudal hair cells are mutually
inhibitory (Detwiler and Alkon
1973). On the basis of the interaction between hair cells,
photoreceptors, and optic ganglion cells, it has been proposed that an
increase in the frequency of S-E optic ganglion cell generated EPSPs detected
in type B photoreceptors following light and rotation could contribute to the
prolonged depolarization and increased input resistance of type B
photoreceptors observed in conditioned animals
(Tabata and Alkon 1982). The
synaptic feedback to photoreceptors that is the result of the synaptic
interactions between photoreceptors, hair cells, and optic ganglion cells
could potentially contribute to the acquisition of conditioning correlates
detected in type B photoreceptors. However, it is unlikely that the optic
ganglion contributes to either the generation of the CR or the induction of
intrinsic plasticity recently detected in other components of the CS pathway.
Optic ganglion cells have not been reported to interact synaptically with
neurons other than primary sensory neurons, for example, photoreceptors and
hair cells. Moreover, there is no evidence that any of the synaptic
connections between optic ganglion cells and sensory neurons exhibit
plasticity with conditioning. In contrast, synaptic connections between
sensory neurons and identified interneurons with synaptic projections to motor
neurons have been well documented (Akaike
and Alkon 1980; Goh and Alkon
1984; Crow and Tian
2000,
2002a,
2003a). In addition, synaptic
facilitation of monosynaptic postsynaptic potentials (PSPs) in type
Ie and Ii interneurons elicited by single type B spikes
in conditioned animals does not involve any contribution from optic ganglion
cells (Crow and Tian 2002b).
Moreover, the facilitation of complex PSPs in type Ie interneurons
of conditioned animals that contribute to the inhibition of ciliary locomotion
is most likely the result of intrinsic changes at two CS-US convergence sites,
the photoreceptors and type Ie interneurons (Crow and Tian
2002b,
2003b).
Interneurons in the UCR Pathway
Statocyst hair cells project to optic ganglion cells and three identified
types of cerebropleural interneurons
(Akaike and Alkon 1980;
Tabata and Alkon 1982;
Crow and Tian 2004). As shown
in the diagram of Figure 2,
statocyst hair cells form monosynaptic connections with type Ie and
Ii interneurons. Type Ie and Ii interneurons
project polysynaptically to type IIIi interneurons that inhibit VP1
and VP3 ciliary activating motor neurons
(Crow and Tian 2003a).
Rotation, the US depolarizes hair cells that produce excitation of type
Ie interneurons, which results in excitation of type
IIIi inhibitory interneurons and a decrease in the spike activity
of VP1 and VP3 ciliary motor neurons. An increase in the spike activity of
type IIIi interneurons results in inhibition of ciliary locomotion
mediated by inhibition of VP1 and VP3 motor neurons. The second pathway shown
in Figure 2 mediates
contraction of the foot elicited by hair-cell stimulation. This pathway
involves polysynaptic connections with interneurons that have not been
identified. Activation of the proposed circuit involves depolarization of hair
cells by rotation and excitation of type Ib interneurons through
polysynaptic pathways. Excitation of type Ib interneurons directly
excite pedal ventral contractile motor neurons (VCMNs), and posterior foot
contraction motor neurons that collectively produce foot contraction
(Crow and Tian 2004). In
summary, rotation produces a depolarizing generator potential in identified
statocyst hair cells, and by way of monosynaptic and polysynaptic connections
with identified interneurons, the elicitation of foot contraction and
inhibition of ciliary locomotion, the two UCRs.
|
Convergence of the CS and US Pathways
As summarized in Figure 3,
one site of convergence between the CS and US is at the primary sensory
neurons of the CS and US pathways. Synaptic projections from statocyst hair
cells to the photoreceptors are both monosynaptic and polysynaptic (see
Fig. 3). Hair cells and
photoreceptors form reciprocal monosynaptic inhibitory connections
(Alkon 1973b). Caudal hair
cells inhibit photoreceptors and cephalic hair cells are inhibited by type B
photoreceptors. Stimulation of statocyst hair cells elicits a monosynaptic
GABAergic IPSP in type B photoreceptors
(Alkon et al. 1993;
Sakakibara et al. 1993;
Rogers et al. 1994;
Blackwell 2002a). The evidence
for GABA as a transmitter in the US pathway is derived from biochemical,
immunohistochemical, and pharmacological studies. The statocysts contain
endogenous GABA, and immunocytochemical procedures have localized GABA in hair
cell axons and presumed terminal processes
(Alkon et al. 1993). The
GABAA antagonist bicuculline reduced the amplitude of the type B
photoreceptor IPSP elicited by hair-cell stimulation and a GABA reuptake
inhibitor increased the amplitude of the IPSP
(Alkon et al. 1993). In
addition, both microapplication of GABA and baclofen applications
hyperpolarized the type B photoreceptors.
|
The polysynaptic projection from statocyst hair cells to photoreceptors
involves the activation of serotonergic interneurons that form putative
monosynaptic connections with photoreceptors
(Land and Crow 1985;
Auerbach et al. 1989). The
evidence for 5-HT as a neurotransmitter in the US pathway is derived from
behavioral, physiological, and immunohistochemical studies.
Immunohistochemical studies have identified serotonergic neurons in the
cerebropleural ganglion (CPG triplets) with projections that form rings of
varicosities surrounding the optic nerve that contains the axons of
photoreceptors (Land and Crow
1985). In addition, type B photoreceptors project to a region of
the cerebropleural ganglion that is innervated by 5-HT immunoreaction terminal
processes (Land and Crow 1985;
Auerbach et al. 1989).
Additional evidence implicating 5-HT in the US pathway and in conditioning of
Hermissenda comes from studies showing that pharmacological agents
that affect 5-HT neurotransmission (imipramine, bufotenine, and 5,7-DHT)
attenuate in vitro conditioning correlates in type B-photoreceptors
(Grover et al. 1989). In
addition, 5-HT modulates generator potentials and membrane conductances in
type B photoreceptors, modifications that have been identified as neural
correlates of Pavlovian conditioning (Crow
and Bridge 1985; Farley and Wu
1989; Crow and Forrester
1991; Acosta-Urquidi and Crow
1993; Rogers and Matzel
1995; Yamoah and Crow
1995,
1996). A computational model
of the type B photoreceptor used to investigate the contribution of different
ionic conductances modulated by 5-HT to the enhanced excitability produced by
5-HT suggested that changes in IA, IK,Ca, or
Ih (Yamoah et al.
1998) would produce excitability changes comparable to
experimental findings (Cai et al.
2003). One-trial conditioning studies also have provided evidence
for a role for 5-HT in conditioning. Light (CS) paired with 5-HT application
to the exposed, but otherwise intact circumesophageal nervous system is
sufficient to produce long-term phototactic suppression
(Crow and Forrester 1986).
The synaptic organization of the secondary components of the visual pathway
and graviceptive pathway of Hermissenda have now been characterized
and described in considerable detail
(Alkon et al. 1978;
Crow et al. 1979;
Akaike and Alkon 1980; Crow and
Tian 2000,
2002a,
2003a,
2004). These studies have
identified type Ie and Ii cerebropleural interneurons as
an additional site of convergence between the CS and US pathways (see
Fig. 3). Photoreceptors and
statocyst hair cells form monosynaptic excitatory connections with type
Ie interneurons and monosynaptic inhibitory connections with type
Ii interneurons (Akaike and
Alkon 1980; Crow and Tian
2000). The third site of convergence between the CS and US
pathways is between statocyst hair cells and recently identified type
Ib interneurons (Crow and Tian
2004). As summarized in Figure
3, statocyst hair cells form polysynaptic excitatory connections
with type Ib interneurons, and photoreceptors exhibit variable and
weak excitatory polysynaptic connections with type Ib
interneurons.
In addition to the identification of the sites of synaptic convergence
between the CS and US pathways, most of the components of the network
generating ciliary locomotion have now been identified (Crow and Tian
2000,
2002a,
2003a,
2004). This provides for the
opportunity to investigate how modifications in a neural circuit produced by
Pavlovian conditioning are expressed in the generation of a CR. Progress
toward this goal was supported by recent studies showing that light inhibits
the activity of VP1 ciliary motor neurons after conditioning. In contrast,
light produced excitation of ciliary motor neurons in pseudorandom controls
(Crow and Tian 2003b).
The development of a semi-intact preparation has provided some insights into the physiology of the motor system mediating the foot-shortening UCR and CR (see Fig. 4). Less is known about the circuitry supporting the foot-shortening UCR and CR, although some of the motor neurons mediating anterior and posterior foot contraction have now been identified (see Fig. 2). Studies of semi-intact preparations have led to the identification of type Ib interneurons that project monosynaptically to VCMNs and posterior foot contraction motor neurons. Mechanical stimulation of statocyst hair cells evokes a depolarizing generator potential and an increase in spike activity of type Ib interneurons and VCMNs. Depolarization of type Ib interneurons with extrinsic current is sufficient to produce contraction of the anterior and posterior foot. Moreover, extrinsic current depolarization of identified hair cells elicits EPSPs and complex EPSPs in type Ib interneurons, and complex EPSPs and spikes in VCMNs. The development of a semi-intact preparation and the identification of components of the neural circuitry supporting foot contraction and ciliary locomotion provides the opportunity to study the acquisition and expression of two CRs within the same nervous system.
|
Cellular and Synaptic Plasticity at Convergence Sites of the CS and US Pathways
An essential step in the analysis of Pavlovian conditioning is the
identification of loci in the animal's nervous system, in which 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 of
site for memory storage. Studies of neural correlates of conditioning in the
primary sensory neurons of the CS pathway have identified cellular changes
involving both enhanced excitability that is intrinsic to identified type A
and type B photoreceptors and synaptic facilitation of connections between
identified photoreceptors (Crow and Alkon
1980; Alkon et al.
1982,
1985;
Farley and Alkon 1982;
West et al. 1982; Crow
1985b,
1988; Frysztak and Crow
1993,
1994,
1997;
Gandhi and Matzel 2000).
Cellular correlates of conditioning in type B photoreceptors are expressed by
enhancement of CS-elicited generator potentials and increased spike frequency,
increased excitability to extrinsic current, modification of light-dependent,
Ca2+-dependent, and voltage-activated currents, and increases in
the phosphorylation of several proteins
(Crow and Alkon 1980; Neary et
al. 1981
1986; Alkon et al.
1982,
1985,
1992;
1993;
Farley and Alkon 1982;
Alkon 1984; Crow
1985b,
1988;
Goh et al. 1985;
Alkon and Nelson 1990;
Farley et al. 1990b;
Matzel 1990a; Frysztak and
Crow 1993,
1994,
1997;
Muzzio et al. 2001). Studies
of identified type A photoreceptors have reported a decrease in the amplitude
of light-elicited generator potentials, enhanced excitability to extrinsic
current, increases in CS-elicited spike activity, and decreases in the
magnitude of two K+ currents
(Farley et al. 1990b;
Farley and Han 1997; Frysztak
and Crow 1993,
1994,
1997). In addition to changes
in primary sensory neurons, facilitation of monosynaptic and complex PSPs in
identified type Ie and Ii interneurons has been recently
shown in conditioned Hermissenda
(Crow and Tian 2002b; see
Fig. 5).
|
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 (IA, ICa) and
Ca2+-dependent (IK,Ca) currents (Alkon et al.
1982,
1985;
Collin et al. 1988; for
reviews see Crow 1988;
Alkon 1989;
Sahley and Crow 1998).
Modulation of light-induced potassium currents in type B photoreceptors has
also been proposed to contribute to correlates of conditioning
(Blackwell 2002b). 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 type I interneurons
(Crow and Tian 2002b).
However, facilitation of the monosynaptic IPSP between identified type B
photoreceptors and type A photoreceptors may be due to both pre- and
postsynaptic mechanisms (Frysztak and Crow
1994). In addition, facilitation of the amplitude of the
monosynaptic IPSP between type B photoreceptors and type Ii
interneurons and the monosynaptic EPSP between type B-photoreceptors and type
Ie interneurons of conditioned animals may also involve pre- and
postsynaptic mechanisms (see Fig.
5).
Modifications in Components of the CS Pathway Contributing to Generation of the CR
An examination of CS-elicited changes in excitability and PSPs in the
neural circuit generating ciliary locomotion has provided an explanation for
the generation of the light-elicited suppression of the locomotor CR produced
by Pavlovian conditioning. As summarized in the circuit diagram shown in
Figure 5, studies of
conditioned animals have shown that light inhibits the tonic spike activity of
VP1 ciliary activating pedal motor neurons
(Fig. 5, inset F at top) below
their prelight baseline activity (Crow and
Tian 2003b). In contrast, recordings from pseudorandom controls
exhibited a significant increase in light-elicited tonic firing of VP1 neurons
(Fig. 5, inset F at bottom). An
analysis of changes in other components of the CS pathway of conditioned
animals revealed that type Ie interneurons exhibited an intrinsic
enhanced excitability with conditioning in contrast to pseudorandom controls
(Fig. 5, inset E). Therefore, a
combination of synaptic facilitation and intrinsic enhanced cellular
excitability can account for light-elicited inhibition of locomotion.
Facilitation of the synaptic connection between type B-photoreceptors and type
Ie (Fig. 5, inset C)
interneurons in conjunction with intrinsic enhanced excitability in type
B-photoreceptors (Fig. 5,
insets A and B) and type Ie interneurons of the CS pathway would
result in an increase in spike activity of type IIIi inhibitory
interneurons and inhibition of VP1 and VP3 ciliary motor neurons of
conditioned animals (see Fig.
5, inset F). Modifications of neurons in the circuit summarized in
Figure 5 would thus account for
the light-elicited inhibition of locomotion (CR) detected in conditioned
Hermissenda.
Morphological Modifications in the CS Pathway
Ultrastructural and electrophysiological analyses have shown that synaptic
interactions between photoreceptors, other sensory neurons, and interneurons
is in the neuropil of the cerebropleural ganglion
(Crow et al. 1979). Recent
studies of labeled photoreceptors have focused on changes in the morphology of
secondary and terminal photoreceptor processes in the neuropil. Structural
changes characterized by a reduction of dendritic boundary volumes enclosing
labeled medial-type-B photoreceptor arborizations were observed in conditioned
animals as compared with unpaired controls
(Alkon et al. 1990). The
structural changes in type B photoreceptors associated with conditioning have
been examined further using an in vitro conditioning procedure. Using confocal
microscopy, it was shown that five conditioning trials produced a contraction
of terminal branches of fluorescently labeled type B photoreceptors along a
contralateral axis as compared with unpaired controls
(Kawai et al. 2002). The
changes in terminal branch morphology were detected within an hour after in
vitro conditioning. Interestingly, in double-labeling experiments of the B
photoreceptors and hair cells, terminal contraction was not observed at the
synaptic connection between the hair cell and photoreceptor
(Kawai et al. 2002). The
structural remodeling of the B photoreceptor terminal branches following in
vitro conditioning can be blocked with anisomycin pretreatment
(Kawai et al. 2003). The
further analysis of this type of structural remodeling is of interest, as
Pavlovian conditioning produces synaptic facilitation of the monosynaptic
connection between type B photoreceptors and type A photoreceptors
(Frysztak and Crow 1994;
Gandhi and Matzel 2000) and
between type B photoreceptors and type Ie and Ii
interneurons (Crow and Tian
2002b). In addition to changes in dendritic volume, changes in the
morphology of photoreceptor somas have also been reported to occur following
activation of PKC, a signaling molecule implicated in learning
(Lederhendler et al. 1990).
Phorbol-induced changes involved outgrowth from the cell surface similar to
blebs or ruffling that altered the soma volume. The functional significance of
the morphological changes in both dendritic and soma volume has not been
established.
Second-Messenger Systems
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; Neary et al.
1986; Matzel et al.
1990a; Crow et al.
1991; Farley and Schuman
1991) and extracellular signal-regulated protein kinase (ERK;
Crow et al. 1998) have been
reported to contribute to modifications of excitability and synaptic efficacy
of conditioned Hermissenda. Light and rotation have spatially
separated physiological consequences on type B photoreceptors. However, both
the CS and US increase cytosolic Ca2+ levels
(Sakakibara et al. 1993;
Blackwell 2000,
2002a;
Muzzio et al. 2001). Both GABA
and 5-HT have been proposed to mediate the effects of activation of the US
pathway during conditioning (see section on anatomy of the CS and US
pathways). Light, the CS, activates phospholipase C (PLC) to produce inositol
trisphosphate (IP3) and diacylglycerol (DAG) (Sakakibara et al.
1986,
1994). Inositol trisphosphate
opens rhabdomeric Na+ and Ca2+ channels, which result in
a depolarizing generator potential and Ca2+ influx
(Blackwell 2000). Two distinct
Ca2+ currents have been identified in the soma of photoreceptors
(Yamoah and Crow 1994).
Inositol trisphosphate can also bind to its receptor (IP3R), which
triggers Ca2+ release from the smooth endoplasmic reticulum
(Blackwell and Alkon 1999).
The Ca2+ influx from the rhabdomere and the IP3R-gated
storage compartment can cause Ca2+ release from the ryanodine
receptorgated (RyR) compartment (Blackwell
and Alkon 1999).
Rotation, the US, produces a depolarizing generator potential in identified
statocyst hair cells and elicits a monosynaptic GABAergic IPSP in the
photoreceptors (Alkon et al.
1993; Sakakibara et al.
1993; Rogers et al.
1994; Blackwell
2002a). The US is also proposed to activate a serotonergic
polysynaptic pathway that projects to type B photoreceptors
(Land and Crow 1985; Crow and
Forrester 1986,
1991). The primary focus of
5-HT release has been on the modulation of membrane conductances (e.g.,
Farley and Wu 1989;
Acosta-Urquidi and Crow 1993;
Yamoah and Crow 1996). In
addition, the induction of 5-HT-dependent enhanced excitability of type B
photoreceptors is Ca2+ dependent
(Falk-Vairant and Crow 1992).
However, the precise role of 5-HT in the activation of second-messenger
systems is poorly understood. It has been proposed that GABAergic IPSPs in
photoreceptors activate phospholipase A2 (PLA2) to
liberate arachidonic acid (AA; Muzzio et
al. 2001) and create a back-propagating wave of Ca2+
released from intracellular stores (Ito et
al. 1994; Blackwell
2002a). When the CS and US are repeatedly paired, the
Ca2+ influx, due to light IP3R stores, RyR stores, and
voltage-gated Ca2+ channels sums together
(Blackwell and Alkon 1999).
The large increase in cytosolic Ca2+ combined with DAG and AA act
to synergistically activate PKC by translocation of PKC to the membrane
(Lester et al. 1991). Each
pairing of the CS and US has been proposed to incrementally increase the
proportion of PKC translocated to the membrane
(Muzzio et al. 1997).
Both 5-HT (Rogers and Matzel
1995; Yamoah and Crow
1996) and GABA (Yamoah and
Crow 1996) are linked to a pertussis-toxin sensitive G-protein.
These proteins can activate multiple second messenger systems, several of
which are involved in one-trial and/or multitrial classical conditioning.
Activation of PKC is necessary for the induction of cellular plasticity in
Hermissenda (Crow et al.
1991; Crow and Forrester
1993a,b).
Down-regulation of PKC and pretreatment with kinase inhibitors block the
induction of short-term excitability, but not long-term excitability
(Crow and Forrester 1993b).
This indicates that short- and long-term memory in this system may involve
parallel processes. PKC may phosphorylate two K+ channels,
IK,A and IK,Ca, decreasing their maximum conductance and
producing increased input resistance and evoked spike frequency
(Farley and Auberbach 1986;
Frysztak and Crow 1997).
Conditioning also induces the activation of ERK
(Crow et al. 1998). Serotonin
activates ERK through a Ca2+-dependent PKC pathway and a
PKC-independent pathway (Crow et al.
2001).
Proteins Regulated by Pavlovian Conditioning
The regulation of several proteins has been examined following
conditioning. Calexcitin (CE) is a GTP- and Ca2+-binding protein
found in Hermissenda photoreceptors
(Neary et al. 1981;
Alkon et al. 1998;
Kuzirian et al. 2001). CE is
activated by Ca2+ influx, can decrease K+ currents, and
may bind to the RyR to increase cytosolic Ca2+ concentrations
(Nelson et al. 1996,
1999;
Ascoli et al. 1997). CE is
proposed to be phosphorylated by PKC, which produces translocation to the
membrane. Phosphorylation of CE also causes it to bind to the
Ca2+-ATPase transporter to increase the rate of Ca2+
removal from the cytosol (Alkon et al.
1998). Behavioral conditioning has been reported to increase CE in
B photoreceptors, specifically in Ca2+ sequestering organelles such
as endoplasmic reticulum and within mitochondria and photopigments
(Kuzirian et al. 2001). It has
been proposed that increased CE levels in B photoreceptors of conditioned
animals causes increased excitability via K+-channel inactivation
and internal Ca2+ release from ER due to increased CE binding to
ryanodine receptors.
One-trial conditioning regulates proteins found in the CS pathway (Crow et
al. 1996,
1997,
1999;
Crow and Siddiqi 1997). A
protein whose phosphorylation is regulated by Pavlovian conditioning is
cytoskeleton-related protein 24 (Csp24), a member of the family of
ß-thymosin repeat proteins (Crow and Xue-Bian
2000,
2002;
Crow et al. 2003). Actin
coprecipitates with Csp24, and is colocalized with Csp24 in the cytosol of
B-photoreceptor cell bodies (Crow and
Xue-Bian 2002). Csp24 is phosphorylated by procedures that produce
enhanced intermediate-term and long-term excitability, but not after
procedures that result in short-term excitability of photoreceptors
(Crow and Xue-Bian 2000).
Incubation of isolated Hermissenda nervous systems with Csp antisense
oligonucleotides decrease Csp24 expression. Treatment with antisense
oligonucleotides before one-trial conditioning blocked intermediate-term
enhanced excitability, without affecting the induction of short-term immediate
enhanced excitability (Crow et al.
2003). Because Csp24 is associated with the actin cytoskeleton,
its regulation by conditioning may influence K+ channel activity by
the spatial and temporal control of actin dynamics.
Conclusions and Discussion
The progress in determining how Pavlovian conditioning is expressed in the generation of behavior in Hermissenda is encouraging, and is supported by recent work involving the identification of the neural circuit-controlling locomotion and its modulation by light (CS) and rotation (US). The analysis of Pavlovian conditioning in the neural circuit generating ciliary locomotion showed that both enhanced cellular excitability and synaptic facilitation are expressed in identified circuit components at different loci within the network. The distributed nature of cellular and synaptic plasticity associated with this example of Pavlovian conditioning suggests that an adequate explanation of conditioned behavior requires both an analysis of neural circuits and the identification of mechanisms of CS-US contiguity at convergence sites between the CS and US pathways. Consistent with the view that learning may initially involve changes in pre-existing synaptic connections, conditioned inhibition of phototactic behavior involves modifications of existing synaptic connections between photoreceptors and identified type I interneurons. However, the formation of new connections between neurons in the neural circuit modulating locomotor behavior cannot be dismissed. The acquisition of the foot-shortening CR may involve the establishment of new synaptic connections between photoreceptors, interneurons, and type Ib interneurons involving polysynaptic pathways. De novo synaptic connections are proposed to operate at this level of the circuit, generating foot-shortening because of the weak and variable influences of light, the CS, on the synaptic activity of type Ib interneurons observed prior to conditioning. Light does not excite VCMNs or posterior foot contraction motor neurons before conditioning, which is consistent with behavioral evidence showing that the CS does not elicit foot-shortening before conditioning.
The analysis of conditioning correlates has revealed that the first site of intrinsic cellular and synaptic plasticity is at the initial site of convergence between the CS and US pathways, the primary sensory neurons of the CS pathway. The mechanisms of temporal contiguity between the CS and US involve both enhanced cellular excitability and enhanced synaptic strength. The changes in excitability produced by conditioning involve reductions in several well-characterized K+ conductances in type B photoreceptors. Moreover, recent modeling studies have indicated that reductions in IA and IK,Ca or an increase in Ih would result in enhanced excitability similar to what is detected experimentally in voltage-clamp and current-clamp studies. The second site of intrinsic enhanced excitability is the type Ie interneurons of the CS pathway. However, membrane conductances underlying enhanced excitability of type Ie interneurons have not yet been identified.
Taken collectively, the evidence shows that acquisition of Pavlovian conditioning results in the activation of several second-messenger systems. Both one-trial and multitrial Pavlovian conditioning of Hermissenda involves PKC and the ERK-signaling pathway (ERK). Conditioning is sufficient to activate PKC and ERK, and inhibition of their phosphorylation and activation can block the induction of plasticity.
As in other learning systems, protein synthesis is required to form long-term memory following one-trial and multitrial conditioning. In addition, an intermediate phase of memory has been identified that is dependent on protein synthesis, but not RNA synthesis. Interestingly, the induction of short-term memory can be blocked without blocking the expression of long-term memory, suggesting that memory may involve parallel processing. Several proteins that are regulated by Pavlovian conditioning have now been identified. CE and Csp24 are two of the proteins that have been examined in some detail. CE is a Ca2+-and GTP-binding protein proposed to enhance excitability via K+-channel inactivation and Ca2+ release from internal stores (ER). Csp24 is a cytoskeletal-related protein whose expression and phosphorylation are required for persistent enhanced excitability. The contribution of Csp24 to synaptic and cellular structural remodeling may be through regulation of the actin cytoskeleton. Excitability could be influenced by alterations in channel density or channel conductances modulated by modification of actin filament dynamics. The cellular and synaptic changes identified following conditioning are distributed at several loci within the network and, therefore, not localized to a single synaptic site or neuron. The distributed nature of learning-dependent changes may account for the complexity of Pavlovian conditioning in Hermissenda, specifically, the emergence of a new response to the CS following conditioning.
ACKNOWLEDGMENTS
We thank Diana Parker for assistance with the manuscript. This research was supported by National Institutes of Health Grants MH-40860 and MH-58698
FOOTNOTES
Article and publication are at http://www.learnmem.org/cgi/doi/10.1101/lm.70704.
E-MAIL terry.crow{at}uth.tmc.edu; FAX (713) 500-0623.
REFERENCES
Akaike, T. and Alkon, D.L. 1980. Sensory convergence
on central visual neurons in Hermissenda. J.
Neurophysiol. 44:501
-513.
Alkon, D.L. 1973a. Neural organization of a molluscan
visual system. J. Gen. Physiol.
61:444
-461.
____. 1973b. Intersensory interactions in
Hermissenda. J. Gen. Physiol.
62:185
-202.
____. 1974. Associative training of
Hermissenda. J. Gen. Physiol.
64: 70-84.
____. 1975. Responses of hair cells to statocyst
rotation. J. Gen. Physiol.
66:507
-530.
____. 1983. Learning in a marine snail. Sci. Ameri. 249:70 -84.
____. 1984. Calcium-mediated reduction of ionic currents: A biophysical memory trace. Science 30:1037 -1045.
____. 1989. Memory storage and neural systems. Sci. Amer. 261:42 -50.
Alkon, D.L. and Bak, A. 1973. Hair cell generator
potentials. J. Gen. Physiol.
61: 619.
Alkon, D.L. and Fuortes, M.G. 1972. Responses of
photoreceptors in Hermissenda. J. Gen.
Physiol. 60:631
-649.
Alkon, D.L. and Nelson, T.J. 1990. Specificity of molecular changes in neurons involved in memory storage. FASEB J. 4:1567 -1576.[Abstract]
Alkon, D.L., Akaike, T., and Harrigan, J. 1978.
Interaction of chemosensory, visual, and statocyst pathways in Hermissenda
crassicornis. J. Gen. Physiol.
71:177
-194.
Alkon, D.L., Lederhendler, I., and Shoukimas, J.J.1982
. Primary changes of membrane currents during retention of
associative learning. Science
215:693
-695.
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. Behav. Neural Biol. 44:278 -300.[CrossRef][Medline]
Alkon, D.L., Ikeno, H., Dworkin, J., McPhie, D.L., Olds, J.L.,
Lederhendler, I., Matzel, L., Schreurs, B.G., Kuzirian, A., Collin, C., et al.1990
. Contraction of neuronal branching volume: An anatomic
correlate of Pavlovian conditioning. Proc. Natl. Acad.
Sci. 87:1611
-1614.
Alkon, D.L., Sanchez-Andrés, J-V., Ito, E., Oka, K.,
Yoshioka, T., and Collin C. 1992. Long-term transformation of an
inhibitory into an excitatory GABAergic synaptic response. Proc.
Natl. Acad. Sci. 89:11862
-11866.
Alkon, D.L., Anderson, M.J., Kuzirian, A.J., Rogers, D.F., Fass, D.M., Collin, C., Nelson, T.J., Kapetanovic, I.M., and Matzel, L.D.1993 . GABA-mediated synaptic interaction between the visual and vestibular pathways of Hermissenda. J. Neurochem. 61:556 -566.[Medline]
Alkon, D.L., Nelson, T.J., Zhao, W., and Cavallaro, S.1998 . Time domains of neuronal Ca2+ signaling and associative memory: Steps through a calexcitin, ryanodine receptor K+ channel cascade. Trends Neurosci. 21:529 -537.[CrossRef][Medline]
Ascoli, G., Liu, K.X., Olds, J.L., Nelson, T.J., Gusov, P.A., Bertucci, C., Brananti, E., Raffaelli, A., Salvador, P., and Alkon, D.L.1997 . Secondary structure of Ca2+ induced conformational change of calexcitin, a learning associated protein. J. Biol. Chem. 272:29771 -29779.
Auerbach, S.B., Grover, L.M., and Farley, J. 1989. Neurochemical and immunocytochemical studies of serotonin in the Hermissenda central nervous system. Brain Res. Bull. 22:353 -361.[CrossRef][Medline]
Blackwell, K.T. 2000. Evidence for a distinct light-induced calcium-dependent potassium current in Hermissenda crassicornis. J. Comp. Neurosci. 9: 149-170.[CrossRef][Medline]
____. 2002a. Calcium waves and closure of potassium
channels in response to GABA stimulation in Hermissenda type B
photoreceptors. J. Neurophysiol.
87:776
-792.
____. 2002b. The effect of intensity and duration on
the light-induced sodium and potassium currents in Hermissenda type B
photoreceptor. J. Neurosci.
22:4217
-4228.
Blackwell, K.T. and Alkon, D.L. 1999. Ryanodine receptor modulation of in vitro associative learning in Hermissenda crassicornis. Brain. Res. 20:114 -125.
Britton, G. and Farley, J. 1999. Behavioral and neural
bases of noncoincidence learning in Hermissenda. J.
Neurosci. 19:9126
-9132.
Cai, Y., Baxter, D.A., and Crow, T. 2003. Computational study of enhanced excitability in Hermissenda: Membrane conductances modulated by 5-HT. J. Comput. Neurosci. 15:105 -121.[CrossRef][Medline]
Collin, C., Ikeno, H., Harrigan, J.F., Lederhendler, I., and Alkon,
D.L. 1988. Sequential modification of membrane currents with
classical conditioning. Biophys. J.
54:955
-960.
Crow, T. 1983. Conditioned modification of locomotion in Hermissenda crassicornis: Analysis of time dependent associative and nonassociative components. J. Neurosci. 3:2621 -2628.[Abstract]
____. 1985a. Conditioned modification of phototactic behavior in Hermissenda. I. Analysis of light intensity. J. Neurosci. 5:209 -214.[Abstract]
____. 1985b. Conditioned modification of phototactic behavior in Hermissenda. II. Differential adaptation of B-photoreceptors. J. Neurosci. 5: 215-223.[Abstract]
____. 1988. Cellular and molecular analysis of associative learning and memory in Hermissenda. Trends Neurosci. 11:136 -147.[CrossRef][Medline]
Crow, T.J. and Alkon, D.L. 1978. Retention of an
associative behavioral change in Hermissenda.
Science 201:1239
-1241.
____. 1980. Associative behavioral modification in
Hermissenda: Cellular correlates. Science
209:412
-414.
Crow, T. and Bridge, M.S. 1985. Serotonin modulates photoresponses in Hermissenda type-B photoreceptors. Neurosci. Lett. 60:83 -88.[CrossRef][Medline]
Crow, T. and Forrester, J. 1986. Light paired with
serotonin mimics the effect of conditioning on phototactic behavior of
Hermissenda. Proc. Natl. Acad. Sci.
83:7975
-7978.
____. 1991. Light paired with serotonin in vivo produces both short- and long-term enhancement of generator potentials of identified B-photoreceptors in Hermissenda. J. Neurosci. 11:608 -617.[Abstract]
____. 1993a. Protein kinase inhibitors do not block the expression of established enhancement in identified Hermissenda B-photoreceptors. Brain Res. 613: 61-66.[CrossRef][Medline]
____. 1993b. Down-regulation of protein kinase C and
kinase inhibitors dissociate short- and long-term enhancement produced by
one-trial conditioning of Hermissenda. J.
Neurophysiol. 69:636
-641.
Crow, T. and Offenbach, N. 1983. Modification of the initiation of locomotion in Hermissenda: Behavioral analysis. Brain Res. 271:301 -310.[CrossRef][Medline]
Crow, T. and Siddiqi, V. 1997. Time-dependent changes
in excitability after one-trial conditioning of Hermissenda.
J. Neurophysiol. 78:3460
-3464.
Crow, T. and Tian, L.-M. 2000. Monosynaptic
connections between identified A and B photoreceptors and interneurons in
Hermissenda: Evidence for labeled-lines. J.
Neurophysiol. 84:367
-375.
____. 2002a. Morphological characteristics and central
projections of two types of interneurons in the visual pathway of
Hermissenda. J. Neurophysiol.
87:322
-332.
____. 2002b. Facilitation of monosynaptic and complex
PSPs in type I interneurons of conditioned Hermissenda. J.
Neurosci. 22:7818
-7824.
____. 2003a. Interneuronal projections to identified
cilia-activating pedal neurons in Hermissenda. J.
Neurophysiol. 89:2420
-2429.
____. 2003b. Neural correlates of Pavlovian
conditioning in components of the neural network supporting ciliary locomotion
in Hermissenda. Learn. Mem.
10:209
-216.
____. 2004. Statocyst hair cell activation of identified interneurons and foot contraction motor neurons in Hermissenda. J. Neurophysiol. (in press).
Crow, T. and Xue-Bian, J.J. 2000. Identification of a 24 kDa phosphoprotein associated with an intermediate stage of memory in Hermissenda. J. Neurosci. 20: RC74.
____. 2002. One-trial in vitro conditioning
regulates a cytoskeletal-related protein (CSP24) in the conditioned stimulus
pathway of Hermissenda. J. Neurosci.
22:10514
-10518.
Crow, T., Heldman, E., Hacopian, V., Enos, R., and Alkon, D.L.1979 . Ultrastructure of photoreceptors in the eye of Hermissenda labelled with intracellular injections of horseradish peroxidase. J. Neurocytol. 8: 181-195.[CrossRef][Medline]
Crow, T., Forrester, J., Williams, M., Waxham, M.N., and Neary, J.T. 1991. Down-regulation of protein kinase C blocks 5-HT-induced enhancement in Hermissenda B photoreceptors. Neurosci. Lett. 121:107 -110.[CrossRef][Medline]
Crow, T., Siddiqi, V., Zhu, Q., and Neary, J.T. 1996. Time-dependent increase in protein phosphorylation following one-trial enhancement in Hermissenda. J. Neurochem. 66:1736 -1741.[Medline]
Crow, T., Siddiqi, V., and Dash, P.K. 1997. Long-term enhancement but not short-term in Hermissenda is dependent upon mRNA synthesis. Neurobiol. Learn. Mem. 68:343 -350.[CrossRef][Medline]
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. J.
Neurosci. 18:3480
-3487.
)
Inhibitory synaptic connections; (
) excitatory synaptic connections.
Solid lines represent established monosynaptic connections, and dashed lines
represent polysynaptic connections with potential interneurons not yet
identified.
