Mechanisms of Synaptic Plasticity in the Rat Olfactory Bulb

MECHANISMS OF SYNAPTIC PLASTICITY IN THE RAT OLFACTORY BULB

YUAN GAO

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Thesis Advisor: Dr. Ben W. Strowbridge

Department of Neurosciences

CASE WESTERN RESERVE UNIVERSITY

January, 2010

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Yuan Gao

candidate for the Ph.D degree *.

(signed) Stefan Herlitze (chair of the committee)

Ben W. Strowbridge

Diana L. Kunze

Christopher Wilson

(date) Oct 2 2009

* We also certify that written approval has been obtained for any proprietary material contained therein.

ii Table of Contents

Table of Contents...... iii List of Figures...... v Acknowledgements ...... vi List of Abbreviations ...... vii Abstract...... ix Chapter 1 Introduction ...... 1 Signal Transduction in Mammalian Olfactory System...... 3 Olfactory epithelium ...... 3 Olfactory bulb ...... 4 Olfactory cortex ...... 6 Plasticity in the Olfactory System...... 8 Olfactory system related behavioral plasticity...... 8 Long-term plasticity in the olfactory structures...... 11 Regulation of Odor Representation in Mitral Cells ...... 13 Spiking pattern regulated by intrinsic properties...... 13 Dendrodendritic inhibition between mitral and granule cells...... 15 Synaptic regulation by periglomerular cells ...... 20 Lateral excitation and self-excitation on mitral cells...... 22 Other mechanism ...... 25 Chapter 2 Long-term Plasticity of Excitatory Inputs to Granule Cells in the Rat Olfactory Bulb...... 40 Abstract...... 41 Results...... 42 Supplementary Methods ...... 48 Olfactory bulb brain slice and recording methods...... 48 2-Photon guided focal stimulation...... 49 Analysis of EPSPs in granule cell intracellular recordings ...... 50 Analysis of cell-attached granule cell recordings ...... 52 Analysis of IPSPs in mitral cell intracellular recordings...... 52 Chapter 3 Extra-glomerular Layer Excitation of Olfactory Bulb Mitral Cells Mediated through a Carbenoxolone and NBQX-sensitive Pathway...... 70 Abstract...... 71 Introduction...... 72 Materials and methods ...... 75 Olfactory bulb brain slice preparation and recording methods...... 75 Two-Photon imaging ...... 76 Data acquisition and analysis...... 77 Results...... 78 Discussion...... 84 Dye-coupling between mitral cells and unknown processes ...... 84 Sensitivity to carbenoxolone...... 86 The pathway of signal transduction ...... 87 Potential functional role of “off-beam” mitral cell excitation ...... 90 Chapter 4 General Discussion ...... 109

iii Overview...... 110 Mechanisms underlying granule cell spiking ...... 115 Modulation of dendrodendritic inhibition by granule cell LTP...... 117 Regulation of the granule cell distal synaptic plasticity ...... 120 Functional relevance of STDP on the granule cell ...... 122 Electrical coupling in the extraglomerular regions...... 124 Impact of proximal and distal inhibitory input onto mitral cells ...... 128 Significance and future direction...... 129 Bibliography...... 140

iv List of Figures

Figure 1-1. Zonal organization of sensory inputs from the olfactory epithelium to the olfactory bulb ...... 28 Figure 1-2. Synaptic organization of the mammalian main olfactory bulb ...... 30 Figure 1-3. Synchronization in the olfactory bulb and the olfactory cortex...... 32 Figure 1-4. Depolarization evoked clusters of spiking and subthreshold oscillation in olfactory bulb mitral cells...... 34 Figure 1-5. Reciprocal dendrodendritic synapse between mitral and granule cells ...... 36 Figure 1-6. Lateral excitation and electrical coupling between two mitral cells...... 38 Figure 2-1. Spike timing–dependent plasticity of proximal excitatory inputs to granule cells...... 54 Figure 2-2. LTP evoked by TBS...... 56 Supplementary Figure 2-1. Properties of proximal and distal EPSPs recorded in granule cells...... 58 Supplementary Figure 2-2. Pairing synaptic stimulation with intracellular depolarization induces long-term potentiation in a resting granule cell ...... 60 Supplementary Figure 2-3. Proximal EPSPs recorded in granule cells were not potentiated by trains of postsynaptic action potentials or presynaptic stimuli presented separately ...... 62 Supplementary Figure 2-4. Pairing stimulation-induced potentiation of proximal granule cell EPSPs requires NMDA receptors...... 64 Supplementary Figure 2-5. Theta-burst stimulation potentates EPSPs in a P16 rat...... 66 Supplementary Figure 2-6. Theta-burst stimulation potentates EPSPs in a P30 rat...... 68 Figure 3-1. Mapping of GCL-evoked postsynaptic potentials recorded in a mitral cell .. 91 Figure 3-2. Two types of mitral cell IPSPs evoked by GCL stimulation ...... 93 Figure 3-3. Mitral cell EPSPs evoked by GCL stimulation...... 95 Figure 3-4. The lack of the glomerular layer in dissected olfactory bulb slices...... 97 Figure 3-5. Two-photon imaging of a mitral cell that intracellulary filled with lucifer yellow in dissected olfactory bulb slice ...... 99 Figure 3-6. Blockade of the GCL-evoked mitral cell depolarization by NBQX and TTX in dissected olfactory bulb slices...... 101 Figure 3-7. The GCL-evoked mitral cell depolarization was sensitive to carbenoxolone (CBX) in dissected olfactory bulb slices...... 103 Figure 3-8. Activation of mitral cells by two independent GCL stimulation with a distance greater than 250 μm in dissected olfactory bulb slices ...... 105 Figure 3-9. Two potential models that illustrating the GCL-evoked mitral cell excitation in dissected olfactory bulb slices...... 107 Figure 4-1. Schematic representation of olfaction perception...... 134 Figure 4-2. Schematic diagram of proposed synaptic connections between the olfactory cortex and olfactory bulb...... 136 Figure 4-3. Two subpopulation of the external tufted cells...... 138

v Acknowledgements

I would first and foremost like to thank my advisor, Dr. Ben Strowbridge, for guiding me to think questions scientifically, inspiring me to design experiments, and helping me come to my own conclusions. I have benefited greatly from my interactions with him during my Ph.D training. I would also like to thank the members of my thesis committee, Drs. Stefan Herlitze, Diana Kunze and Christopher Wilson, for their encouragement, help and advice.

I would like to extend my thanks to the faculty, staff and students of Department of Neurosciences who ever helped me and companied me during these years. I would especially like to thank members of the Strowbridge lab, including Dr. Todd Pressler, Dr.

Phil Larimer, Loren Schmidt and Ross Anderson, for great discussions about research, sports, the life as a graduate student, and everything in between.

While I can never thank them enough, I must thank my parents for their love, support and encouragement; without them I could never be the person who I am today.

The last, but not the least, I would like to thank my husband Dr. Dong Zeng, for listening to those stories about the olfactory bulb at the dinner table, staying with me till midnight in the lab, and proofreading this thesis.

vi List of Abbreviations

ACSF: artificial cerebrospinal fluid

AMPA: alpha-amino-3-hydroxy-5-methylisoxazole-4proionate

AP: action potential

APC: anterior piriform cortex

APV: 2-Amino-5-phosphonovalerate

DDI: dendrodendritic inhibition

EPL: external plexiform layer

EPSC: excitatory postsynaptic current

EPSP: excitatory postsynaptic potential

GABA: gamma-aminobutyric acid

GC: granule cell

GCL: granule cell layer

GL: glomerular layer

GPCR: G-protein coupled receptor

IPSP: inhibitory postsynaptic potential

IPSC: inhibitory postsynaptic current

vii LTP: long-term potentiation

LTD: long-term depression

MC: mitral cell

MCL: mitral cell layer

NBQX: 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide

NMDA: N-methyl-D-aspartate

OB: olfactory bulb

PG: periglomerular cell

Rin: input resistance

STDP: spike timing-dependent plasticity

TBS: theta-bursting stimulation

TTX: tetrodotoxin

viii Mechanisms of Synaptic Plasticity in the Rat Olfactory Bulb

Abstract by

YUAN GAO

Understanding the general principles of information coding that are used by the brain to examine the external world is a central challenge for neuroscientists. The olfactory bulb represents an ideal model for understanding the information coding in sensory systems because of its highly organized and stereotyped anatomy. In addition, the olfactory bulb merits the study of the long-term plasticity, since its cellular roles in the olfactory learning and memory behavior remains a mystery. This thesis project used electrophysiological and novel optical imaging methods to investigate the cellular mechanisms underlying the olfactory learning and memory. Using two-photon guided focal stimulation spike timing-dependent plasticity (STDP) could be induced at the granule cell proximal synapse, which was presumably due to the feedback projection from the piriform cortex. The positive STDP protocol potentiated proximal inputs, but depressed distal inputs to granule cells. Granule cell excitatory postsynaptic potential and mitral cell inhibition were also potentiated by theta-burst stimulation in the granule cell layer. The above results provided a potential cellular basis for olfactory learning and memory.

Then I investigated the source of excitatory input on mitral cells in the main olfactory bulb. Although most previous studies in the olfactory system are based on the

ix premise that the sensory afferent in the glomeruli is the only source of excitatory input for mitral cells, this has not been rigorously tested. In my thesis work, I found that depolarization in mitral cells could be evoked by extracellular stimulation in the granule cell layer in dissected olfactory bulb slices that have no glomerular layer. This depolarization evoked in these mini slices was sensitive to AMPA receptor antagonist

NBQX, TTX, and gap junction blocker carbenoxolone. Intracellular fills of mitral cells with lucifer yellow often showed small processes apparently connected to somata or secondary dendrites. These results are consistent with the presence of gap junctions between mitral cell lateral dendrites and processes of the tufted cells. The possible tufted- mitral cell excitatory pathway may enable mitral cells to respond to sensory input into

“off-beam” glomeruli, in addition to the receptor neurons that drive EPSPs on the glomerulus innervated by their apical dendrites.

Chapter 1 Introduction

1 Elucidating the strategies that the brain use to process sensory information and make accurate responses is a major task for neuroscientists. Information-processing steps in a distinct anatomical structure may reflect a general principle shared with many other brain regions. The olfactory bulb represents an attractive model to study how a local neuronal network encodes the sense of smell into distinct spatial and temporal patterns of electrical signals that are used for the later processing by the higher brain regions.

Based on the early histological studies (Cajal, 1955) and later work on synaptic organization, the olfactory bulb has a simple laminated structure with clearly separated input and output fibers. The olfactory bulb is situated only one synapse away from the sensory periphery (Shepherd, 1972; Buck, 1996; Shepherd and Greer, 1998; Mori et al.,

1999), so the information carried by the olfactory nerve arrives here without any alteration; unlike other sensory systems, the olfactory bulb sends its output directly to olfactory cortex without modification through thalamus (Shepherd and Greer, 1998).

This simple flow of the olfactory information from sensory receptors to the cortex makes the olfactory bulb become an ideal brain area to study the strategy of sensory coding, including intrinsic neuronal properties, synaptic properties and synaptic interactions. In addition, in spite of the intensive studies of long-term behavioral plasticity in the olfactory system (Bruce, 1959; Wilson et al., 1985; Levy et al., 1990; Kendrick et al.,

1992; Brennan and Keverne, 1997), as well as the recently discovered short-term plasticity in the rat olfactory bulb (Dietz and Murthy, 2005; Balu et al., 2007), little is known about the cellular mechanisms that are responsible for olfactory learning. The olfactory bulb is the first relay station for olfactory information, and it is suggested by numerous in vivo studies that the function of the olfactory bulb is involved in olfactory

2 experience related plasticity (Baldwin and Shillito, 1974; Ferreira et al., 1992; Kendrick et al., 1992). Therefore the olfactory bulb also represents an ideal model for the study of sensory plasticity.

In this thesis work, I focused on two major questions. First, since it is suggested that the function of the olfactory bulb changes during olfactory learning and memory

(Bruce, 1959; Bruce, 1960; Bruce and Parrott, 1960; Wilson et al., 1985; Levy et al.,

1990; Kendrick et al., 1992; Brennan and Keverne, 1997; Demas et al., 1999), I wanted to understand how the bulbar network acts to mediate long-term behavioral plasticity.

Second, although gap junctional protein connexin36 is widely expressed in extraglomeurlar regions in the olfactory bulb (Kosaka et al., 2005), bulbar electrical synapses have been reported within the glomeruli (Schoppa and Westbrook, 2002;

Christie et al., 2005; Christie and Westbrook, 2006; Pimentel and Margrie, 2008; Jan et al., 2009). This raises an interesting question of what is the physiological function of extraglomerular expressed connexin36, especially in the external plexiform layer (EPL) and mitral cell layer (MCL). So the second part of my work addressed whether there were electrical synapses formed on the mitral cells outside of the glomeruli, and if there are, I also wanted to know how these gap junctions were wired into the local circuit.

Signal Transduction in Mammalian Olfactory System

Olfactory epithelium

Olfactory perception starts from olfactory sensory neurons which express G protein-coupled odorant receptors (GPCRs) (Dohlman, 1991) in the olfactory epithelium

3 in nasal cavity. Mammals can express up to 1,000 GPCR genes to respond to the diverse odor molecules; however each olfactory sensory neuron may express only one particular

GPCR gene (Buck and Axel, 1991; Ressler et al., 1993; Chess et al., 1994; Axel, 1995).

Olfactory GPCR shows a broad tuning to odorants, since each of them can interact with a subgroup of structurally related odor ligands (Buck, 1996; Malnic et al., 1999; Araneda et al., 2000). The broad tuning of odor receptors indicates that olfactory information is encoded by overlapping sets of activated olfactory sensory neurons.

The olfactory epithelium can be divided into four spatial zones in which different sets of GPCRs are expressed (Figure 1-1), although neurons expressing the same GPCR gene can be located relatively broadly in one zone (Ressler et al., 1993; Vassar et al.,

1993; Vassar et al., 1994). Moreover, the spatial patterns of these zones are identical across different individuals from a same species (Buck, 1996). The zonal pattern in the olfactory epithelium suggests that olfactory information is highly organized before it is transmitted to the brain.

Olfactory bulb

Each olfactory sensory neuron projects a single axon to the ipsilateral olfactory bulb, where it forms synapses with specific groups of olfactory principal cells, mitral and tufted cells, in spherical regions of neuropil called glomeruli. There is strong evidence suggesting that the zonal organization in the olfactory epithelium is basically preserved at the glomeruli level in the olfactory bulb (Schoenfeld et al., 1994; Yoshihara et al., 1997;

Mori et al., 2000). However, the olfactory information carried by olfactory sensory neuron axons converges dramatically at the olfactory bulb: in rodents, for example

4 rabbits, there are about 2000 glomeruli receiving inputs from 50 × 106 olfactory neurons, and each glomerulus is innervated by about 25 mitral cells and 50 tufted cells (principal neurons), therefore the information converging in mitral cells is at the ratio about 1000:1

(Buck and Axel, 1991; Buck, 1996; Shepherd and Greer, 1998; Mori et al., 1999).

Mitral/tufted cells convert the sensory input into a spatiotemporal code by firing clusters of action potentials. However, the individual mitral and tufted cells which only project into a single glomerulus, are able to receive sensory input from a large group of olfactory sensory neurons (around 3 × 104) (Shepherd and Greer, 1998; Mori et al., 1999). Since olfactory neurons have broad tuning to structurally related odorants (Buck, 1996;

Araneda et al., 2000), the activity patterns of mitral cells initially have a broad overlap in response to the sensory input. To make an accurate olfaction response, precise modulation of the bulbar output may be necessary to narrow the tuning.

It is suggested that the tuning specificity of mitral cells is enhanced by their intrinsic properties and local synaptic circuitry. The majority of inhibition on mitral cells is from periglomerular (PG) cells in the glomerular layer (Aungst et al., 2003;

Wachowiak and Shipley, 2006), and bulbar granule cells in the granule cell layer (GCL)

(Buck, 1996; Shepherd and Greer, 1998; Urban and Sakmann, 2002) (Figure 1-2). Both self- and lateral inhibition have been found in the olfactory bulb (Shepherd and Greer,

1998). The synchronized oscillatory discharge found in mitral cells is tightly controlled by the dendrodendritic inhibition from granule cells (Laurent et al., 1996; Shepherd and

Greer, 1998; Friedman and Strowbridge, 2003) as well as sniffing behaviors (Verhagen et al., 2007). This synchronization of mitral cell activity is crucial for the later information processing in the higher brain region, the olfactory cortex (Figure 1-3) (Laurent et al.,

5 1996; Mori et al., 1999; Litaudon et al., 2008). Bulbar inhibition plays an important role in refining the spatial and temporal patterns of mitral cell activity (Yokoi et al., 1995;

MacLeod and Laurent, 1996; Isaacson and Strowbridge, 1998; Laurent, 2002; Fantana et al., 2008). Perturbation of bulbar GABAergic inhibition in mice olfactory bulb (Nusser et al., 2001) or honeybee antennal lobe (Stopfer et al., 1997) disrupts odor discrimination, as well as the normal oscillation recorded in the antennal lobe.

Olfactory bulb output represented by mitral cells is also gated by self- and lateral excitation on mitral cells (Isaacson, 1999; Friedman and Strowbridge, 2000; Schoppa and

Westbrook, 2002; Christie and Westbrook, 2006), as well as other types of direct and indirect inhibition (Pinching and Powell, 1971; Pinching and Powell, 1972; Shepherd and

Greer, 1998; Halabisky and Strowbridge, 2003; Presseler and Strowbridge, 2006).

Olfactory cortex

The olfactory information is transmitted from the olfactory bulb to the olfactory cortex via the mitral/tufted cell axons (also called lateral olfactory tract). Anatomically, the olfactory cortex is divided into five regions: the anterior olfactory cortex; the piriform cortex; the olfactory tubercle; the amygdala; and the entorhinal area. The latter four regions are thought to be related to sensory perception and odor discrimination (Buck and

Axel, 1991). The traditional anatomical studies indicate that the stereotyped spatial map in the olfactory bulb is not recapitulated in the olfactory cortex (Haberly, 1998). Even using the modern genetic approaches, for example the examination of c-fos induction

(Guthrie et al., 1993) in the olfactory cortex, there are no distinct patches in the olfactory cortex that are found to convey olfactory information regarding specific odorants (Illig

6 and Haberly, 2003). The difficulty of finding a stereotyped map in the olfactory cortex may be related to the overlap in inputs and complex interconnection between pyramidal cells (Illig and Haberly, 2003). In visual cortex (Weliky et al., 1996) and the olfactory bulb (Yokoi et al., 1995) the nearby zones with different response properties perform a lateral inhibition to sharpen the sensory output, while this extraction strategy may not be used in the olfactory cortex due to the spatially distributed representation of information.

However, using a combination of the signals carried by a large group of cortical neurons to make an odor response may represent an efficient strategy for storing and discriminating the complex odor features in nature (Haberly, 2001).

Despite the lack of a sensory map in the olfactory cortex, the convergence of mitral/tufted cell outputs onto individual pyramidal neurons suggest that temporal summation of bulbar output, partially due to synchronized oscillations, contributes to triggering activity in targeted cortical pyramidal neurons (Figure 1-3). In light of the beta and gamma oscillation in the olfactory cortex (Freeman, 1960; Bressler and Freeman,

1980; Zibrowski and Vanderwolf, 1997; Neville and Haberly, 2003; Litaudon et al.,

2008), it has been found that the interaction between oscillatory global inhibition and odor-specific excitation governs the spike timing of cortical pyramidal cells and shape odor representation at the level of the olfactory cortex (Poo and Isaacson, 2009). In addition to direct processing the outputs from the olfactory bulb, pyramidal cells in the olfactory cortex also make axon collaterals (or centrifugal inputs) back to the olfactory bulb functioning as a feedback regulation (de Olmos et al., 1978; Haberly and Price, 1978;

Shipley and Adamek, 1984; Gray and Skinner, 1988; Shepherd and Greer, 1998; Balu et al., 2007). Disruption of this feedback projection to the olfactory bulb by reversible

7 cryogenic blockade changes the oscillatory activity in the olfactory bulb (Gray and

Skinner, 1988). However, more studies are required to reveal the principle strategy of sensory coding in the olfactory cortex.

Plasticity in the Olfactory System

To better cope with the environment, many brain regions have the ability to change their internal structures in response to history of previous activity patterns. Long- term change of synaptic transmission, which includes the changes in the quantity of neurotransmitters or the change of receptors, is considered to be an elementary process underlying learning and memory. Soon after the long-term potentiation (LTP) was first discovered in mammals (Bliss and Lomo, 1973), learning and memory was thought to be generated and mediated by the hippocampus. However, studies with animals strongly suggest that the olfactory system, if even not solely, is at least partially responsible for olfactory learning and memory (Bruce, 1959; Bruce, 1960; Bruce and Parrott, 1960;

Wilson et al., 1985; Levy et al., 1990; Kendrick et al., 1992; Brennan and Keverne, 1997;

Demas et al., 1999). Unlike hippocampus-related learning and memory, the cellular mechanism underlying olfactory plasticity remains a mystery.

Olfactory system related behavioral plasticity

Since 1940s, a large effort has been put on the study of pheromone learning behavior. Bruce’s group found that in mice, exposure to a strange male after the initial mating can effectively block the pregnancy with the original male, but removal of

8 olfactory bulbs can abolish the pregnancy block (Bruce, 1959; Bruce, 1960; Bruce and

Parrott, 1960). Later work on pheromone learning showed that pheromone recognition/ memory can last over 30 days after it is initially established, and more interestingly, a non-selective pheromone memory can be formed simply by blocking GABA receptors in the olfactory bulb, even without mating (Brennan et al., 1990; Brennan and Keverne,

1997). These results indicate that some changes within the olfactory bulb are likely to be responsible for this long-term sensory-related memory; therefore hippocampus is not the only brain region which is responsible for learning and memory. In Prairie Voles, the formation of their olfactory memory requires the function of olfactory cortex rather than the hippocampus (Demas et al., 1999). Although none of the above studies discovered the cellular mechanisms for pheromone learning and memory, the work from Brennan’s group strongly suggests that the change of bulbar GABAergic inhibition is related to the olfactory memory formation. In addition, norepinephrine (NE) regulation may also be involved because of its increased activity in the mice olfactory bulb during mating

(Brennan et al., 1990). In turtle, NE projection to the olfactory bulb can reduce

GABAergic inhibition at bulbar dendrodendritic synapses between mitral and granule cells (Jahr and Nicoll, 1982). If this is true for the mice, it provides a plausible explanation for why bulbar GABA receptor blockade can mimic mating behavior to form a pheromone memory.

The presence of plasticity in the olfactory system is also suggested in familiar recognition studies. Many studies show that ewe form a selective memory for their lambs.

In 2 hours of the mother-new born bond is established following parturition (Levy et al.,

1990; Kendrick et al., 1992; Levy et al., 2004). This memory is mediated by the olfactory

9 bulb, since its removal results in a loss of selective lamb recognition and abnormal maternal behaviors (Baldwin and Shillito, 1974; Levy et al., 1995). Like in pheromone learning, the change of GABA activity is also found in lamb recognition (Ferreira et al.,

1992; Kendrick et al., 1992). GABA concentration in the external plexiform layer is significantly higher when the ewe smells her own lamb than an alien one (Kendrick et al.,

1992), which suggests increased GABA inhibition, presumably at mitral cells, is involved in this memory formation. The GABA regulation of olfactory recognition is further supported by the finding that diazepam injection, which increases GABAergic activity, in

OB disrupts lamb recognition (Ferreira et al., 1992).

Olfactory conditioning in rodents is another behavior that is well studied for learning and memory. Extracellular recording in the olfactory bulb shows that the activity of both mitral cell and granule cell significantly change when neonatal rat pups approach a familiar odor compared to an unfamiliar one (Wilson et al., 1985). So a memory is able to be formed by the neonatal exposure to odors, and as a consequence, this memory can alter the bulbar response to those odors later. The decreased mitral cell excitation and increased granule cell inhibition in this olfactory conditioning suggest that there must be a long-term change of synaptic strength on either mitral cell or granule cell, or even both.

Although the above three examples have different behavioral contexts, they share a common feature that the olfactory learning and memory depends on the change within the olfactory bulb, and specifically the change of GABAergic inhibition in the bulb.

However, there is no the direct electrophysiological evidence showing that long-term plasticity occurs in mammalian olfactory bulb, and this is one of the major questions I addressed in my thesis.

10 Long-term plasticity in the olfactory structures

Many behavioral plasticity studies proposed that olfactory memory formation occurred in the olfactory bulb and presumably at the dendrodendritic mitral-to-granule cell synapses (Wilson et al., 1985; Brennan et al., 1990; Brennan and Keverne, 1997), however the underlying cellular mechanism remains unknown. Studies in carp olfactory bulb provide the first electrophysiological evidence of LTP in olfactory memory: from the granule cell layer field recording, a long-term potentiation can be induced by a tetanus given to the lateral olfactory tract and/or the middle olfactory tract (Satou et al.,

2005). The presence of LTP in the olfactory bulb provides a plausible explanation for olfactory memory formation, and it also suggest that centrifugal fibers from the higher brain region may regulate the memory and this regulation might happen in the granule cell layer. However there are some limitations in these carp olfactory LTP studies. First, although the basic structure of the carp olfactory bulb is similar to that of the mammals, the number of odorant receptor genes and glomeruli modules are an order of magnitude smaller than those of mammals, also the bulb is less laminated and less cell types are present (Satou, 1990; Korsching et al., 1997). This structural difference may result in some different strategies of olfactory information coding in mammalian olfactory bulb.

Second, the synaptic response of particular cell types cannot be discovered by extracellular field recordings. Third, very high frequency stimulation that is used to induce LTP in the carp olfactory bulb may not be close to physiological conditions, since to cope with the sniffing behavior, the most common frequency of oscillation in the olfactory system falls in the theta, beta and lower gamma band (Zibrowski and

Vanderwolf, 1997; Neville and Haberly, 2003). All these limitations may prevent the

11 mechanism of the carp olfactory LTP from becoming a general rule for the olfactory behavioral plasticity in mammals.

In neocortex, cerebellum and hippocampus, long-term plasticity is also found in the form of spike timing dependent plasticity (STDP), and unlike the classic long-term plasticity STDP is dependent on the order of pre- and postsynaptic activities (Levy and

Steward, 1983; Gustafsson et al., 1987; Bell et al., 1997; Magee and Johnston, 1997;

Markram et al., 1997). Usually the presence of postsynaptic NMDAR is required to be the coincidence detector and the primary calcium source in this type of the plasticity

(Shouval et al., 2002; Froemke et al., 2005). Recently in the insect mushroom body,

STDP is found by using intracellular recording with synapse-specific stimulation

(Cassenaer and Laurent, 2007). In insects, receptor neurons project their axons to the antennal lobe, which share some functional similarity with the olfactory bulb (Wilson et al., 2004); then the olfactory information is transmitted from the antennal lobe to the mushroom bodies with distributed spatial and temporal patterns. When the activity of the neurons in the mushroom body is paired with that of their downstream targets in a specific temporal window, long-term plasticity is induced, and this plasticity is suggested be responsible for memory retrieval (Cassenaer and Laurent, 2007). In spite of a big structural difference between insects and mammals, the STDP in the mushroom body, together with the LTP in the carp (Satou et al., 2005), suggest the long-term plasticity in the olfactory system may involve plasticity at the long range synapses between different brain regions. In addition, the importance of the temporal regulation of neuronal activity

– spike timing is reinforced by this study. The oscillation cycles in the olfactory bulb or mushroom body may not only contribute to refining odor-specific response, but also be

12 important for the memory formation and recall. STDP observed in insect mushroom body is of interest in the light of the proposed importance of the oscillation in mammalian olfactory bulb and olfactory cortex. However, whether the pre-and postsynaptic activity dependent plasticity occurs in mammalian olfactory bulb is unknown. I address this question in Chapter 2.

Regulation of Odor Representation in Mitral Cells

Mitral cells receive odor information transmitted by olfactory nerves and code it into electrical signals with distinct spatial-temporal patterns. This procedure serves as the basis for olfactory perception. For a mitral cell, its spiking pattern is odor-specific, and for a particular odor, different temporal patterns of responses can be evoked across a subpopulation of mitral cells (Laurent et al., 1996; Shepherd and Greer, 1998). Both intrinsic properties and synaptic regulation are believed to be responsible for these distributed firing patterns in mitral cells.

Spiking pattern regulated by intrinsic properties

Intrinsic properties are demonstrated to be essential for neuronal responses in central nervous system (Llinas, 1988; Connors and Gutnick, 1990). Slice recordings in the rat mitral cell shows clusters of action potentials are generated following an initial hyperpolarization in response to intracellular threshold current injection, as well as subthreshold oscillation (Figure 1-4) (Chen and Shepherd, 1997; Balu et al., 2004). This phenomenon persists in present of GABAA and ionotropic glutamatergic antagonists

13 (Desmaisons et al., 1999; Balu et al., 2004), which suggests that both spikes and subthreshold oscillation are due to mitral cell membrane properties. In addition, the membrane potential dependent subthreshold oscillation precisely controls spike timing

(Desmaisons et al., 1999), which indicates that the interaction between the intrinsic properties and synaptic inputs regulates mitral cell firing patterns.

During a depolarization step, the first spike is usually delayed and often triggers a hyperpolarization of the membrane potential (Chen and Shepherd, 1997; Balu et al.,

2004), which suggest that K+ current is present in this mitral cell. Slowly inactivating K+

(termed ID current) is known to be crucial for temporal integration of inward and outward currents in rat hippocampal neurons (Storm, 1988). In mitral cells, blocking ID – like currents with low concentration (1–10 μM) 4-AP reduces the long initial spike latency and affect intermittent discharges (Balu et al., 2004).

Mitral cell is able to generate rebound spiking following a small amplitude and short latency hyperpolarizing current injection; while during a long duration of hyperpolarization, the recovery of inactivated IA conductance prolongs the repolarization which blocks rebound spiking (Balu and Strowbridge, 2007). Rebound spiking is blocked by TTX or the subthreshold Na channel blocker riluzole (Balu and Strowbridge, 2007).

Instead of brief hyperpolarization, an inhibitory synaptic input also can evoke rebound spiking, which indicates that IA and subthreshold Na channels on mitral cell membrane allow local inhibition to gate mitral cell spiking patterns. In addition, in physiological conditions IA channels also modulate lateral inhibition through attenuating action potential back propagation along the mitral cell lateral dendrites and decrease dendritic calcium transients (Christie and Westbrook, 2003).

14 Dendrodendritic inhibition between mitral and granule cells

The structural properties of mitral cell-to-granule cell synapses have been well defined in many studies. Granule cells, the most common type of intrinsic neurons in the olfactory bulb, are the axonless GABAergic interneurons located deep to mitral cell bodies (Price and Powell, 1970; Shepherd and Greer, 1998). These interneurons have very small cell bodies (6–8 μm in the diameter) and a single apical dendrite which bifurcates and extends laterally when it arrives at the mitral cell layer, and relatively short basal dendrites in the granule cell layer (Shepherd and Greer, 1998). The branches of the granule cell apical dendrites usually extend 50–200 μm laterally in the EPL, at where they synapse with mitral cell secondary (lateral) dendrites (Price and Powell, 1970;

Shepherd and Greer, 1998). On granule cell synaptic terminals both NMDARs and non-

NMDARs are found to be distributed along the entire extent of the postsynaptic specialization, which can be activated by glutamate released from mitral cell presynaptic terminals (Wellis and Kauer, 1993; Wellis and Kauer, 1994; Sassoe-Pognetto and

Ottersen, 2000). Interestingly GABA-containing vesicles are also found at this terminal, which is believed to mediate recurrent inhibition on mitral cells (Shepherd and Greer,

1998) (Figure 1-5). Basically at this asymmetrical dendrodendritic synapse, depolarization of mitral cell synaptic terminals evokes calcium influx through voltage- sensitive calcium channels or NMDARs, which in turn triggers glutamate release; glutamate diffuses across the synaptic cleft and binds glutamatergic receptors on granule cell synaptic terminals (Jahr and Nicoll, 1982; Wellis and Kauer, 1993; Wellis and Kauer,

1994); activation of glutamatergic receptors on granule cell depolarizes its terminal triggering calcium influx which mediates a feedback GABA release to mitral cells. In

15 EPL, over 80% of all synapses belong to this class of reciprocal dendrodendritic synapses

(Shepherd and Greer, 1998).

The physiological properties of this reciprocal dendrodendritic synapse, especially the mechanism of granule cell GABA release, also have been well studied. The dendrodendritic inhibition can be recorded by mitral cell whole-cell intracellular recordings (Isaacson and Strowbridge, 1998; Schoppa et al., 1998; Chen et al., 2000;

Halabisky et al., 2000; Isaacson, 2001). Using current-clamp recording, a depolarizing current step evoke a long-lasting bicuculline–sensitive afterhyperpolarization in presence of TTX and TEA (Isaacson and Strowbridge, 1998). Since all axonal conduction and potassium channels are blocked by TTX and TEA respectively, the recorded afterhyperpolarization reflects feedback inhibition from granule cells. In voltage-clamp mode with CsCl-based internal solution, which set the reversal of GABAA receptors mediated responses to 0 mV, a brief depolarization step trigger a slow inward current when the cell is held at -70 mV, and this slow inward current is also TTX-insensitive

(Isaacson and Strowbridge, 1998). The function of the postsynaptic NMDARs and non-

NMDARs in triggering dendrodendritic inhibition is carefully examined. It is found that

GABAergic feedback inhibition on mitral cells is not affected by bath perfusion of non-

NMDAR antagonist NBQX, but abolished by NMDAR antagonist APV (Isaacson and

Strowbridge, 1998; Schoppa et al., 1998; Halabisky et al., 2000). These pharmacological data are consistent with the finding of a dramatic increase of dendrodendritic inhibition in

Mg2+ - free extracellular solution (Isaacson and Strowbridge, 1998; Schoppa et al., 1998;

Chen et al., 2000; Halabisky et al., 2000). One plausible interpretation for these results is that activation of NMDARs on granule cell postsynaptic terminal raises local calcium

16 concentration to the threshold of neurotransmitter release: NMDAR has high permeability for Ca2+ , and open NMDARs can mediate a calcium influx that is sufficient for triggering postsynaptic GABA release in the presence of voltage-gated calcium channel blockers (Chen et al., 2000). Controversially, electron microscopic analysis

(Price and Powell, 1970; Sassoe-Pognetto and Ottersen, 2000) suggests that many of

NMDARs can be distant (1 μm) from GABA releasing sites, which implies that “pure”

NMDAR activation might not be enough for GABA exocytosis in physiological conditions. This synaptic organization is confirmed by the later electrophysiological recordings. With the presence of axonal conduction, the dendrodendritic inhibition relies on N- and P/Q type calcium channels, while blockade of T-, L- and R type calcium channels has no effect (Isaacson and Strowbridge, 1998). In addition the slow deactivation kinetics of NMDAR may cause a prolonged depolarization which allows more voltage-gated calcium channels to be activated. In the presence of cyclothiazide

(CTZ), a drug that slows the gating kinetics of AMPARs and prolongs the time course of

AMPAR mediated currents, dendrodendritic inhibition can be evoked entirely by

AMPARs in TTX (Isaacson, 2001). There are two possibilities that may lead to the above controversy: first, electron microscopic analysis that is based on a single section of a structure may limit the accuracy of morphological conclusions. Serial reconstructions of granule cell spines suggest that the distance between the apposition of NMDARs and

GABA releasing sites could be a few nanometers (Woolf et al., 1991). In addition, the contractile filamentous structure in dendrites may also play a role in the dynamic control of the distance between these two types of synapses (Landis and Reese, 1983); second, as it is suggested by Chen (2000), different mechanisms may be involved in recurrent and

17 lateral inhibition (Chen et al., 2000). However, the remarkably slow kinetics of dendrodendritic inhibition cannot be solely mediated by either NMDAR or voltage-gated calcium channels (Lester et al., 1990; Isaacson and Strowbridge, 1998; Schoppa and

Westbrook, 1999). It has been found that in the presence of 2-Amino-5- phosphonovalerate (APV), blocking transient IA current can partially rescue dendrodendritic inhibition, although the recovered inhibition has a faster kinetics

(Schoppa and Westbrook, 1999). This transient IA current is likely to attenuate the fast- activation AMPAR mediated synaptic input, therefore increase the component of

NMDARs.

Taken the above results together, glutamate released from the mitral cell presynaptic terminal likely binds to the postsynaptic AMPARs to evoke a fast depolarization, which opens the postsynaptic NMDARs. The time course of AMPAR mediated depolarization is gated by the intrinsic IA current to limit its chance of activating calcium channels directly. Open NMDARs then provide a long-lasting depolarization that allows voltage-gated calcium channels to be activated. Calcium influx through NMDARs that is presumably only in recurrent inhibition and N/P/Q type calcium channels trigger the postsynaptic neurotransmitter release. The prolonged feedback inhibition regulated by both synaptic and intrinsic properties is also likely to be critical for the generation of olfactory bulb oscillations. However there are several interesting questions remaining for the dendrodendritic synapse, especially in physiological conditions. First, how does the mitral cell balance the function of calcium influx through NMDARs and calcium channels? And is this balance activity-dependent? Second, are these N/P/Q types of calcium channels evenly clustered around GABA releasing site?

18 Third, if it is as suggested by Margrie (2001), that the propagation mitral cell action potentials along lateral dendrites is decremental with distance and the extend of the propagation depends on dendrodendritic inhibition (Margrie et al., 2001), whether a distal inhibition is able to be detected in somata and the level of inhibition is distance- dependent need to be addressed.

The function of dendrodendritic inhibition also has been well described through behavioral and electrophysiological studies. Besides recurrent inhibition, the mitral cell- to-granule cell synapse also provides lateral inhibition on neighboring mitral cells because of the convergence of excitatory inputs from multiple mitral cells onto a single granule cell (Yokoi et al., 1995; Brennan and Keverne, 1997; Shepherd and Greer, 1998).

Recording from synaptically connected mitral cells demonstrates that evoked glutamate release from one mitral cell can induce inhibition on the unstimulated one (Isaacson and

Strowbridge, 1998). Odor responses in mitral cells, recorded in vivo, are altered by disruption of this dendrodendritic inhibition, presumably including both self- and lateral inhibition (Yokoi et al., 1995). That a high density of NMDARs are found at this dendrodendritic synapse (Sassoe-Pognetto and Ottersen, 2000; Balu et al., 2007), raises an intriguing probability that dendrodendritic inhibition may also be the basis of olfactory learning and memory. So far, only short-term depression on granule cells has been reported with direct activation of mitral cell inputs (Balu et al., 2007), and whether a long-term plasticity occurs here is still an open question. Rare silent AMPARs at dendrodendritic synapse which prevents the presence of classic long-term potentiation

(Balu et al., 2007) suggest the plasticity, if occurs, probably in a special form. Another major difficulty for directly testing the plasticity at mitral cell-to-granule cell

19 dendrodendritic synapse is how to manipulate a specific synaptic input, because the synaptic connections in the olfactory bulb are not laminar in organization. I address this question in Chapter 2.

Synaptic regulation by periglomerular cells

As one type of short-axon neurons, PG cells have their dendrites within a glomerulus, and their axons distributed laterally within extraglomerular regions

(Shepherd and Greer, 1998). PG cells receive excitatory inputs from both olfactory sensory neurons axonal terminal, external tufted cells and mitral/tufted cells primary dendritic tufts, as well as inhibitory inputs from other PG cell dendrites or axons

(Shepherd and Greer, 1998; Hayar et al., 2004). Their outputs target mitral/tufted cells primary dendritic tufts, sensory nerve terminals and also function as self-inhibition

(Shepherd and Greer, 1998; Murphy et al., 2005). It is believed that PG cells modulate both afferent input and the excitability of mitral cells. PG cells that receive sensory inputs are able to provide GABAergic feedforward inhibition onto mitral cells either through PG cells-to-mitral cells dendrodendritic synapses or axodendritic synapses. Since PG cell axon can extends as far as 5 glomeruli, the axodendritic synapse between PG cells and mitral cells plays an important role of lateral inhibition between PG and mitral cells

(Pinching and Powell, 1971; Shepherd and Greer, 1998; Murphy et al., 2005). The excitatory inputs from mitral cells activate PG cells which in turn generate a feedback inhibition onto mitral cells (Pinching and Powell, 1971; Shepherd and Greer, 1998).

In addition, PG cells indirectly regulate mitral cell activity. Firstly, blockade of metabotropic GABAB receptors reduces postsynaptic excitatory potentials on mitral and

20 PG cells (Aroniadou-Anderijaska et al., 2000; Murphy et al., 2004; Murphy et al., 2005), this suggests that GABA released from PG cells activates presynaptic GABAB autoreceptors on olfactory nerve terminal, therefore inhibits glutamate release. This result is consistent with a neural model of olfactory bulb processing, which shows that an increase in PG cell inhibition decreases the number of responding output neurons (Linster and Hasselmo, 1997), probably through this mechanism. Secondly, it is found that a depolarization step in a PG cell is immediately followed by a gabazine-sensitive inward current, and it persists in presence of TTX and glutamergic receptor blockers (Smith and

Jahr, 2002; Murphy et al., 2005). This inhibition is believed to be mediated through

GABAA receptor activation by GABA released from itself. Furthermore, GABA released from one PG cell may spillover to activate GABAA receptors on neighboring PG cells.

Paired-recording at two neighboring PG cells show that a depolarization step in one PG cell generates an inward current in the unstimulated PG cell (Murphy et al., 2005). The above result provides direct evidence of the presence of lateral inhibition between PG cells, which requires GABAA receptors. This lateral inhibition between PG cells can be mediated either through dendrodendritic inhibition or axodendritic inhibition (Shepherd and Greer, 1998).

Although PG cells are generally considered as GABAergic interneurons, some of them also express tyrosine hydroxylase (Kosaka et al., 1985; Gall et al., 1987; Shepherd and Greer, 1998; Maher and Westbrook, 2008). Dopamine can be spontaneously released from some PG cells, which causes a presynaptic inhibition on olfactory never terminal in presence of a transport blocker (Ennis et al., 2001; Maher and Westbrook, 2008).

However whether dopaminergic inhibition occurs within glomeruli under physiological

21 conditions remains unclear. We can speculate that PG cell is likely to provide both dopaminergic and GABAergic inhibition onto olfactory nerve terminal to reduce the glutamate release. It is important for olfactory sensory processing given the convergence of over 1,000 olfactory neurons projecting to a single glomerulus and the high release probability of their heavily bifurcated terminals (Shepherd, 1972; Shepherd and Greer,

1998; Murphy et al., 2004).

Lateral excitation and self-excitation on mitral cells

Glutamate released from mitral cells primary and secondary dendrites at the dendrodendritic synapses (Shepherd and Greer, 1998), is not only able to excite postsynaptic targets but also presynaptic autoreceptors. Because mitral cells lacks direct synaptic contact with other mitral cells (Price and Powell, 1970), activation of the presynaptic autoreceptors is proposed to be triggered by glutamate spillover. In the glomerular layer, mitral cells possess both NMDA and AMPA autoreceptors (Nicoll and

Jahr, 1982; Petralia et al., 1994; Montague and Greer, 1999; Salin et al., 2001; Schoppa and Westbrook, 2001; Schoppa and Westbrook, 2002). Paired recording from two mitral cells projecting to a same glomerulus shows that the action potential trains in one mitral cell depolarizes the unstimulated one; however recording in mitral cells projecting to different glomeruli fails to evoke a corresponding depolarization (Figure 1-6)(Schoppa and Westbrook, 2002; Christie and Westbrook, 2006; Pimentel and Margrie, 2008), and this lateral excitation is mediated by glutamatergic receptors because it is completely blocked by APV and NBQX (Christie and Westbrook, 2006). Glutamate spillover responses are typically mediated by NMDARs in light of their high affinity for glutamate and special location (Barbour and Hausser, 1997; Kullmann and Asztely, 1998).

22 Applying APV only results in a small deduction of the response, while NBQX or GYKI

(specific antagonist for AMPARs) by itself is able to block about 80% or more of the lateral excitation (Schoppa and Westbrook, 2002; Pimentel and Margrie, 2008), which suggests that AMPA autoreceptors play a dominant role in the lateral excitation within glomeruli. In addition, the correlated spiking found in a pair of neighboring mitral cells is nearly abolished by NBQX, while not affected by APV (Schoppa and Westbrook, 2002).

This result shows a functional difference between AMPA and NMDA autoreceptor in mitral cell spike synchrony within glomeruli, in part due to the lateral excitation. Unlike the NMDA autoreceptor or those AMPA autoreceptor on the external tufted cell primary dendrite (Ma and Lowe, 2007), mitral cell AMPA autoreceptor on the primary dendrite is calcium impermeable (Pimentel and Margrie, 2008). Mitral cell self-excitation in the glomerular layer is mediated by both NMDA and non-NMDA autoreceptors (Salin et al.,

2001; Schoppa and Westbrook, 2002; Christie and Westbrook, 2006; Pimentel and

Margrie, 2008). Within glomeruli, the lateral communication between mitral cells is also via gap junction (Schoppa and Westbrook, 2002; Christie et al., 2005) (Pimentel and

Margrie, 2008)(Figure 1-6), which presumably allows self-excitation in one mitral cell to be passed to the neighboring cells. In crayfish, gap junction mediated lateral excitation amplifies synaptic inputs onto the decision making neurons to control escape behavior

(Herberholz et al., 2002). In the olfactory bulb, gap junctional connexins (Cx36) is expressed in both juvenile and adult animals (Belluardo et al., 2000; Christie et al., 2005).

However, only mitral cells project to a same glomerulus shows the electrical coupling, and there is no evidence showing the present of gap junction between different glomeurli thus far (Pimentel and Margrie, 2008). For connexin36 knock out mice the lateral

23 excitation evoked by step depolarization is virtually absent in mitral cells pair projecting to the same glomerulus, which is shown to be prominent with the same experimental condition in the wild type animals (Isaacson, 1999; Christie and Westbrook, 2006).

Direct depolarization elicited NMDAR-mediated self-excitation persists, though it is slightly smaller than that in the wild type animal (Christie and Westbrook, 2006). Taking the above evidence together, glutamate spillover-generated lateral and self-excitation is relatively weak due to the regulation of glutamate transports (Christie and Westbrook,

2006). Via gap junctions, the electrical signal in one mitral cell can be more easily passed to the neighboring ones, perhaps constituting an “avalanche” effect on amplifying synaptic output; in addition a single mitral cell is able to recruit more excitatory inputs to gate dendrodendritic signaling. The cooperation between glutamate spillover and electrical coupling powerfully mediated the communication within groups of mitral cells, which provides a potential mechanism of enhancing glomerulus-specific mitral cell synchrony (Schoppa and Westbrook, 2001; Schoppa and Westbrook, 2002).

In the EPL, although mitral cell lateral dendrites do not appear to synapse with other mitral cells (Price and Powell, 1970), both NMDAR (Petralia et al., 1994) and

AMPARs (Montague and Greer, 1999; Friedman and Strowbridge, 2000) are suggested to be present on the dendritic terminals. Glutamate released from mitral cell terminal can effectively evoke an APV-sensitive inward current in the presence of TTX and bicuculine

(Isaacson, 1999; Friedman and Strowbridge, 2000; Salin et al., 2001). In addition, in

Mg2+-free extracellular solution glutamate can spillover to a neighboring mitral cell to provide lateral excitation, which is demonstrated by paired recordings from two mitral cells (Isaacson, 1999). Blocking NMDA autoreceptors abolishes the synchronous firing

24 in neighboring mitral cells (Isaacson, 1999), and also reduces clustering of spikes in most of mitral cells (Friedman and Strowbridge, 2000). This suggests that both self- and lateral excitation mediated by NMDA autoreceptors can effectively modulate mitral cell spiking patterns. In physiological conditions, back propagation of action potentials either directly from soma or antidromical propagation from axons (Hamilton and Kauer, 1988; Margrie et al., 2001) may induce glutamate release to activate extrasynaptic NMDA autoreceptors.

Backpropagation of the AP through a mitral cell may be responsible for the self- excitation, however, the function of antidromic propagation of spikes on lateral excitation in the physiological condition is not known. Furthermore, although AMPARs are not responsible for both self- and lateral excitation, exogenous AMPA application generates self-excitation (Isaacson, 1999; Friedman and Strowbridge, 2000). Therefore the failure of detecting AMPA autoreceptor mediated events might be related to a differential distribution of NMDARs and AMPARs on dendritic terminals.

In the animal studies, disruption of calcium-permeable AMPA autoreceptors results in impaired olfactory memory but enhanced odor discrimination and olfactory learning ability (Shimshek et al., 2005). Although many functional consequences of lateral and self-excitation in the behavioral model still remain to be determined, it is possible that modulation of bulbar lateral and self-excitation may impact olfactory function.

Other mechanism

25 Besides the olfactory sensory input into the glomerular layer, the olfactory bulb also receives inputs from the brain, including the olfactory cortex and brain stem

(Shepherd and Greer, 1998). All these central inputs are termed centrifugal inputs, which terminate in both the granule cell layer and the glomerular layer (Shepherd and Greer,

1998). Both electrophysiological and immunohistochemistry data show that granule cells receive glutamatergic excitatory inputs from cortical projection fibers at their proximal dendrites (Balu et al., 2007). A paired cortical tetanus and mitral cell depolarization evokes robust dendrodendritic inhibition in mitral cell (Balu et al., 2007), which suggests that centrifugal input from the brain gates mitral cell spiking through the regulation of granule cell excitability. However, whether the gating from cortical projection is plastic is still an open question, although silent NMDARs are present on the proximal synaptic spines (Balu et al., 2007). In addition, olfactory bulb granule cells are undergoing neurogenesis and wired in the bulbar circuitry throughout life (Altman, 1969; Kishi,

1987). Recent study in the newborn granule cells shows the presence of non-NMDAR mediated long-term plasticity (Nissant et al., 2009). The finding of the long-term plasticity in adult-born granule cells may represent a mechanism of their function in olfactory learning (Magavi et al., 2005). The structural plasticity, behavioral plasticity

(Baldwin and Shillito, 1974; Brennan et al., 1990; Ferreira et al., 1992; Kendrick et al.,

1992) and the evidence from Balu (2007), raise a possibility that the change of the synapses on granule cells could be the potential mechanisms for the olfactory memory.

This question is addressed in Chapter 2.

Granule cells also receive GABAergic inhibition from Blanes cells in the granule cell layer. Persistent firing in Blanes cell, triggered by glomerular stimulation, granule

26 cell layer stimulation or a simple depolarization, can induce tonic GABAA receptor- mediated inhibition on granule cells, which suggest that granule cell excitability can be effectively regulated by Blanes cells (Presseler and Strowbridge, 2006). In addition, since one Blanes cell may innervate hundreds of granule cells, it may generate synchronous activity in a group of granule cells (Presseler and Strowbridge, 2006). Its persistent firing property may represent a mechanism underlying the short-term olfactory memory.

27 Figure 1-1. Zonal organization of sensory inputs from the olfactory epithelium to the olfactory bulb

Schematic diagram illustrating the olfactory nerve projection from the nose to the main olfactory bulb. The olfactory epithelium is divided into four zones according to the expression of odor receptors. Zonal organization is preserved at the level of the glomeruli: olfactory neurons from a same zone project to glomeruli located in a corresponding zone.

From (Mori et al., 1999)

Figure 1-1

29 Figure 1-2. Synaptic organization of the mammalian main olfactory bulb

Olfactory neurons synapse with mitral (M)/tufted (T) cells at glomeruli (GL). Different glomerular modules represent innervations from sensory neurons (OSN) expressing different odor receptors. Periglomerular cells (PG) are local interneurons which modulate mitral/tufted cell output in the glomerular lay. Granule cell (Gr) forms reciprocal dendrodendritic synapses with mitral cell lateral dendrites. From (Mori et al., 1999)

Figure 1-2

31 Figure 1-3. Synchronization in the olfactory bulb and the olfactory cortex

A, B and C represent 3 individual glomerular modules. Membrane potential oscillation in the olfactory bulb (black trace) synchronizes spikes in mitral cells from A and B.

Temporal summation of bulbar outputs, partial due to the synchronized oscillation, contribute to triggering activity in targeted cortical pyramidal neurons. From (Mori et al.,

1999)

Figure 1-3

33 Figure 1-4. Depolarization evoked clusters of spiking and subthreshold oscillation in olfactory bulb mitral cells

(A) Schematic cartoon of olfactory bulb synaptic organization and the position of patch pipette. (B) Intermittent firing clusters and subthreshold oscillation in presence of APV

(25 μM) and NBQX (5 μM) suggest the intrinsic property of mitral cells. From (Balu et al., 2004)

Figure 1-4

35 Figure 1-5. Reciprocal dendrodendritic synapse between mitral and granule cells

Mitral cell lateral dendrite forms reciprocal synapses with granule cell apical dendrites.

Mitral cell dendritic terminal has glutamate-containing vesicles, which is able to evoke an excitatory response on the postsynaptic granule cell. Granule cell posses GABA- containing vesicles at the dendritic terminal, and GABA releasing result in an inhibition on the presynaptic mitral cell. From (Isaacson and Strowbridge, 1998)

Figure 1-5

37 Figure 1-6. Lateral excitation and electrical coupling between two mitral cells

(A) Schematic cartoon of the mitral cell projection intro glomeruli. Top, two mitral cells project to a same glomerulus; Bottom, two mitral cells project to different glomeruli. (B)

Two mechanisms of mitral cell-to-mitral cell communication. Left, glutamate released from one mitral cell primary dendrite (MA) activates the postsynaptic PG cell as well as the presynaptic AMPA autoreceptors. Self-excitation is passed to the other mitral cell

(MB) via gap junction. Right, glutamate released from MA diffuse and activates AMPA autoreceptors on MB presynaptic terminal. From (Schoppa and Westbrook, 2002)

Figure 1-6

Chapter 2 Long-term Plasticity of Excitatory Inputs to Granule Cells in the Rat Olfactory Bulb

40 Abstract

Using two photon–guided focal stimulation, we found spike timing–dependent plasticity of proximal excitatory inputs to olfactory bulb granule cells that originated, in part, from cortical feedback projections. The protocol that potentiated proximal inputs depressed distal, dendrodendritic inputs to granule cells. Granule cell excitatory postsynaptic potentials and mitral cell inhibition were also potentiated by theta-burst stimulation.

Plasticity of cortical feedback inputs to interneurons provides a mechanism for encoding information by modulating bulbar inhibition.

41 Results

Long-term behavioral plasticity in the olfactory system has been well established through extensive studies of pheromonal learning (Bruce, 1959; Brennan and Keverne,

1997), familial recognition in sheep (Levy et al., 1990; Kendrick et al., 1992) and olfactory conditioning in rodents (Wilson et al., 1985). However, little is known about the cellular mechanisms that are responsible for olfactory learning. No previous study has demonstrated long-term potentiation (LTP) in the main olfactory bulb using intracellular recordings, although one group recently reported LTP of field potentials evoked by tetanic stimulation in carp (Satou et al., 2005). Indirect evidence suggests that one locus of olfactory plasticity may be excitatory inputs to GABAergic granule cells, the primary interneuron in the olfactory bulb. Mitral cell discharges diminish after odor conditioning in neonatal rat pups (Wilson et al., 1985) and as sheep learn to recognize their newborn lambs at the same time as extracellular GABA increases in the olfactory bulb (Kendrick et al., 1992). We used two-photon imaging to stimulate specific excitatory inputs to granule cells and paired these inputs with postsynaptic action potentials. We found

Hebbian LTP and spike timing–dependent plasticity (STDP) of excitatory synapses onto the proximal dendrites of granule cells that arise primarily from feedback projections from piriform cortex (Haberly and Price, 1978; Balu et al., 2007).

Focal stimulation near visualized proximal granule cell dendrites evoked fast- rising excitatory postsynaptic potentials (EPSPs) that facilitated with paired stimulation

(mean paired-pulse ratio (PPR) = 2.21±0.48, n = 11; Supplementary Fig. 2-1 and

Supplementary Methods), as reported previously (Balu et al., 2007). Pairing synaptic

42 stimulation 10 ms before postsynaptic spikes (+10 ms: 50 shocks at 20 Hz; Fig. 2-1a) triggered a twofold increase in EPSP slope measured 5–15 min after pairing (slope ratio,

2.09±0.32; range, 0.88–3.95, significantly greater than 1 (P < 0.005), n = 11; Fig. 2-1b).

Granule cell LTP was robust when analyzed in individual cells. EPSP slope increased in

10 out of 11 granule cells that were tested with one +10-ms pairing protocol; this increase was statistically significant in 8 of 10 cells that were analyzed individually (P < 0.05).

EPSP slope typically remained potentiated for the duration of the recording (mean of 25.6

±1.5 min after induction); EPSP slope decayed after ~20 min in one granule cell. This form of LTP was not associated with substantial changes in input resistance (6.90±1.2% difference; Fig. 2-1b) or holding current (-0.84±1.6 pA change) being required to maintain a constant (-55 mV) membrane potential. The same pairing protocol also potentiated EPSPs that were recorded at -70 mV (Supplementary Fig. 2-2). Neither repetitive postsynaptic spiking nor extracellular stimulation alone potentiated proximal synapses (slope ratios of 0.91±0.16 and 0.81±0.16, respectively, n = 6, P > 0.05;

Supplementary Fig. 2-3). The NMDA receptor antagonist (25 μM D(-)-2-amino-5- phosphonovaleric acid) prevented proximal EPSP potentiation (slope ratio was 0.67±

0.17, which was significantly different from control (P < 0.005), but not significantly different from 1 (P > 0.05), n = 8; Supplementary Fig. 2-4), indicating that NMDA receptors are required for LTP at proximal synapses.

Reversing the pairing protocol, evoking postsynaptic spikes 10 ms before stimulating synaptic responses (-10 ms), depressed proximal EPSPs in granule cells (Fig.

2-1a). On average, the EPSP slope ratio was 0.34±0.15 with -10-ms pairing

43 (significantly less than 1, P < 0.02, n = 4; Fig. 2-1b). Neither the EPSP PPR nor distance from stimulus position along the apical dendrite was different between the +10 and -10- ms experiments. The depression of proximal EPSPs evoked by -10-ms pairing was not associated with a change in input resistance (6.93±2.0% difference, P > 0.05; Fig. 2-1b) or holding current (-2.1±2.3 pA). Although the +10-ms protocols reliably triggered LTP and the -10-ms protocols evoked long-term depression, none of the intermediate pairing intervals tested triggered statistically significant plasticity (P > 0.05; Fig. 2-1c).

The same +10-ms protocol that effectively potentiated proximal excitatory inputs to granule cells depressed distal, presumed dendrodendritic, inputs that were activated by focal stimulation near visualized distal dendritic segments (slope ratio, 0.59±0.14, significantly less than 1 (P < 0.05), n = 4; Fig. 2-1d,e). All distal EPSPs were depressed with paired-pulse stimulation (PPR = 0.73±0.08, n = 4, significantly smaller than the

PPR of proximal EPSPs, 2.66±0.25, n = 43, P < 0.02; Supplementary Fig. 2-1). These results in the olfactory bulb may be related to the recent finding of bidirectional, location- dependent long-term plasticity in neocortical pyramidal neurons (Sjostrom and Hausser,

2006).

We next asked whether tetanic stimulation alone could potentate excitatory inputs to granule cells. Theta-burst stimulation (TBS, three 50-Hz bursts of five granule cell layer (GCL) shocks, repeated at 5 Hz) increased the number of action currents that were evoked by single test shocks in 7 out of 9 granule cells tested in the cell-attached recording mode (Fig. 2-2a). In all nine cells examined, test stimuli in control conditions appeared to show paired-pulse facilitation, which is typical of proximal synapses onto

44 granule cells, and triggered short-latency spikes more reliably following the second than the first shock with paired stimulation. We found that TBS reliably facilitated responses to GCL stimulation (0.26±0.09 evoked action currents before TBS versus 0.50±0.12 from 5–10 min after TBS, P < 0.05, paired t test, n = 9; Fig. 2-2b). The ability of TBS to facilitate spiking under cell-attached conditions suggests that synaptic potentiation also can be recorded intracellularly, near the resting potential of granule cells (-67 mV in this study). We tested this in six current-clamp recordings from granule cells held at -70 mV.

In 1 of 6 granule cells tested, TBS converted subthreshold test EPSP responses to suprathreshold discharges that precluded measuring EPSP slope reliably. EPSP slope increased in all five cells whose responses remained subthreshold to 122±4.9% of control (significantly different from control, P < 0.02; example in Supplementary Fig. 2-

5). We also found that TBS effectively potentiated EPSPs in granule cells recorded at -70 mV in slices from late juvenile rats (postnatal day 30 rat; Supplementary Fig. 2-6).

Finally, we asked whether TBS also potentiated inhibition onto mitral cells. We analyzed responses from nine mitral cells that showed clear inhibitory responses to single

GCL shocks (Fig. 2-2c). IPSPs evoked by GCL stimulation appeared to be mediated predominately by GABAA receptors, as they reversed polarity near -70 mV and were blocked by gabazine (10 μM; Fig. 2-2d). TBS potentiated IPSPs in 8 out of 9 mitral cells for at least 20 min (Fig. 2-2e). IPSPs were evoked at long onset latencies (13.2 ms) relative to the group mean (4.1±1.2 ms; n = 9) in the one mitral cell in which TBS failed to potentate inhibitory responses. The increase in IPSP amplitude was statistically significant across the population of eight mitral cells with IPSP onset latencies ≤6 ms (–

1.06±0.20 control versus -1.62±0.31 mV measured from 5–15 min after TBS, P < 0.02,

45 paired t test). On average, IPSPs increased to 154±15% of control amplitudes; this increase was relatively constant over a 20-min recording period after TBS (Fig. 2-2f). In

6 of these 8 mitral cells, the increase in IPSP peak amplitude was statistically significant when tested individually (P < 0.05, unpaired t test). Membrane potential was not significantly different before and after TBS (-49.8±0.5 in control versus -49.5±0.4 mV after TBS, P > 0.05, n = 8).

These results suggest that proximal excitatory synapses may detect coincident activity in back-projecting piriform cortical cells and postsynaptic granule cell spiking.

Long-term plasticity at this synapse was readily induced by pre- and postsynaptic pairings repeated at 20 Hz, which is in the beta/gamma frequency band that is commonly recorded in piriform cortex in vivo (Zibrowski and Vanderwolf, 1997; Neville and

Haberly, 2003) and enhanced during odor stimulation (Ravel et al., 2003). Plasticity at these synapses is likely to modulate lateral inhibition onto mitral and tufted cells. We found that the same TBS protocol that reliably triggered LTP of excitatory inputs onto granule cells also evoked a long-lasting enhancement of inhibition onto mitral cells.

Although potentiation of proximal excitatory drive to granule cells provides an attractive explanation for TBS-mediated enhancement of mitral cell inhibition, other mechanisms may also be involved, including plasticity in the granule cell–to–mitral cell synapse.

Much of the lateral and self-inhibition of principal cells in the olfactory bulb is mediated through reciprocal dendrodendritic synapses that are tonically attenuated by extracellular

Mg2+, which blocks currents through the NMDA receptors that govern GABA release

(Isaacson and Strowbridge, 1998; Schoppa et al., 1998). In addition to facilitating cortically evoked disynaptic inhibition onto mitral cells, LTP of excitatory inputs to

46 granule cells may also function to enhance dendrodendritic inhibition indirectly by increasing granule cell spiking, thereby transiently relieving the Mg2+ blockade of

NMDA receptors at dendrodendritic synapses. Previous studies have demonstrated that extracellular tetanic stimulation in the GCL (Halabisky and Strowbridge, 2003) or activation of piriform cortical cells (Balu et al., 2007) can gate dendrodendritic self- inhibition of mitral cells, perhaps via this mechanism. LTP of proximal inputs to granule cells may function to facilitate spiking in specific subpopulations of granule cells, dynamically regulating lateral inhibition onto synaptically coupled mitral cells.

47 Supplementary Methods

Olfactory bulb brain slice and recording methods

Horizontal brain slices (300 μm thick) were prepared from olfactory bulbs of P14-

21 Sprague-Dawley rats (except the experiment shown in Supplementary Fig. 2-6, which was from a P30 rat) using a modified Leica (Nussloch, Germany) VT1000S vibratome, as described previously (Halabisky and Strowbridge, 2003; Presseler and Strowbridge, 2006;

Balu et al., 2007). Slice preparation and maintenance was performed in an artificial cerebrospinal fluid (ACSF) with reduced Ca2+ that contained 124 mM NaCl, 2.6 mM KCl,

1.23 mM NaH2PO4, 3 mM MgSO4, 26 mM NaHCO3, 10 mM dextrose and 1 mM CaCl2.

Patch clamp recordings were carried out at 30 oC in a submerged recording chamber using the following ACSF: 124 mM NaCl, 3 mM KCl, 1.23 mM NaH2PO4, 1.2 mM

MgSO4, 26 mM NaHCO3, 10 mM dextrose and 2.5 mM CaCl2. Both dissecting and recording ACSF solutions were continuously oxygenated with 95% CO2 / 5% O2. Whole- cell patch-clamp recordings were made onto granule and mitral cells visualized using a

60x water immersion objective, IR-DIC optics and a frametransfer CCD camera (Cohu

6412-2000, Poway, CA) attached to an upright, fixed-stage microscope (Olympus

BX51WI, Center Valley, PA). Electrophysiological data was acquired through a

Axopatch 1D amplifier (Axon Instruments, Sunnyvale, CA), low-pass filtered at 2 kHz

(FLA-01, Cygnus Technology, Delaware Water Gap, PA) and digitized at 5 kHz using

16-bit analog-to-digital converters (Instrutech ITC-18, Port Washington, NY). Data acquisition and analysis was carried out using custom programs written in Visual Basic 6

(Microsoft, Redmond, WA) and Matlab (Mathworks, Natick, MA). Patch clamp electrodes (5–8 MΩ) contained the following internal solution: 140 mM K-methylsulfate,

48 4 mM NaCl, 10 mM HEPES, 200 μM EGTA, 4 mM MgATP, 300 μM Na3GTP, 10 mM phosphocreatine. 100 μM Alexa488 was added to the internal solution in experiments using 2-photon guided stimulation. We used a sharpened tungsten microelectrode (FHC,

Bowdoin, ME) positioned in the granule cell layer in both cell-attached and whole-cell recordings testing theta-burst stimulation protocols and for the pairing experiment shown in Supplemental Fig. 2-2. All chemicals were obtained from Sigma except Alexa488 hydrazide (Invitrogen, Carlsbad, CA). Receptor antagonists (APV, NBQX and gabazine) were applied by bath perfusion. All results presented in the text, and in the summary plots represent mean ± s.e.m. unless noted, statistical significance was determined using

Student's t test. Membrane potentials reported were not corrected for the liquid junction potential.

2-Photon guided focal stimulation

A custom 2-photon laser scanning system (Presseler and Strowbridge, 2006; Balu et al.,

2007) was used for granule cell visualization and stimulating electrode placement. The rapid scanning mode (3200 lines/sec) in the system enabled the dye-filled stimulating electrode to be efficiently positioned near proximal or distal dendritic segments under visual guidance. Patch clamp electrodes used for focal stimulation contained 124 mM

NaCl, 3 mM KCl, 10 mM HEPES and 50 μM Alexa488 (pH = 7.3) and were connected to a constant-current stimulus isolation unit (WPI A360, Sarasota, Fl). The mean stimulus intensity used in 2-photon guided focal stimulation experiments was 36 μA. Responses using glass pipettes as stimulating electrodes often included a slowly decaying stimulus artifact that was insensitive to both ionotropic glutamate receptor antagonists and blockade of Ca2+ influx. The location along the apical dendrite of granule cells for +10

49 ms proximal pairing stimulating experiments (27.3 ± 3.0 μm from cell body) was not statistically different than the location used in –10 ms pairing experiments (29.7 ± 4.3 μm;

P > 0.05). All proximal excitatory responses analyzed in this study exhibited paired-pulse facilitation. To measure paired-pulse ratio, we first generated an average response from

3–5 trials (50 ms ISI) in each cell. Paired-pulse ratio was determined from the initial

EPSP slopes in this average response. Three out of 46 proximal stimulation experiments were discarded after post-hoc analysis showed paired-pulse depression in control responses; the remaining 43 proximal stimulation experiments showed paired-pulse facilitation. Distal EPSPs were evoked by stimulation near visualized dendritic segments in the external plexiform layer. Distal EPSPs had slower kinetics than proximal EPSPs and depressed with paired-pulse stimulation, as reported previously3. Since we did not record directly from the distal granule cell dendrites, we cannot determine the local amplitude of action potentials during distal pairing protocols. All experiments in Fig. 2-1 and in Supplementary Figs. 2-1, 2-3, and 2-4 were based on 2-photon guided focal stimulation near visualized proximal or distal granule cell dendritic segments.

Analysis of EPSPs in granule cell intracellular recordings

Granule cells were identified by their morphology and by the presence of an afterhyperpolarization following a burst of action potentials evoked by a 50–100 pA, 500 ms duration current step (Presseler and Strowbridge, 2006). Stimulus intensity was adjusted to generate EPSPs with rising phase slopes >0.1 mV/ms averaged over 20–30 responses in the control period. EPSP slope was calculated from the linear fit of the somatic membrane potential from 1.4 to 4.2 ms after stimulation, making this measure relatively insensitive to polysynaptic responses evoked by the stimulating electrode.

50 Granule cells were held at ~ –55 mV in most experiments by injecting a depolarizing bias current to facilitate identifying evoked IPSP responses. (Granule cells in this study rested on average at -67 mV without added bias current.) Experiments with evoked IPSPs, or mixed EPSP/IPSP responses, were discarded. Bath application of glutamate receptor antagonists NBQX (5 μM) and D-APV (25 μM) blocked evoked EPSPs (n = 4).

Calculated slopes often were slightly negative for failures and after synaptic transmission was blocked with NBQX and APV because of a slowly decaying stimulus artifact reflecting the large capacitance of glass stimulating electrodes. Synaptic potentiation also did not appear to result from changes in concomitantly activated inhibitory synapses as

EPSP slope was calculated from the initial rising phase, before the onset of any disynaptic responses, and the monosynaptic response was blocked completely by glutamate receptor antagonists. Recordings with unstable baseline responses, defined by

>10% difference between the mean EPSP slopes in two consecutive 10-episode blocks, were discarded. In theta-burst stimulation experiments, we only included intracellular granule cell recordings in which paired-pulse GCL stimulation (50 ms interval) evoked facilitating EPSPs. Postsynaptic action potentials were evoked by 3 ms duration depolarizing current pulses. Pairing intervals reported were positive when the presynaptic stimuli preceded the current pulse injected into the postsynaptic granule cell; intervals reported are between the onsets of the extracellular stimulus and the current pulse. All results presented in this report are from the first pairing epoch (50 times at 20 Hz) tested in each granule cell. The STDP timing interval to be tested was assigned before each recording was attempted. STDP intervals greater than 25 ms were not possible since paired stimuli were repeated every 50 ms. Long-term plasticity was estimated by EPSP

51 slope ratio, defined by the average ratio calculated from 60 responses, 5–15 min after pairing, divided by mean EPSP slope in the control period immediately before pairing.

Test stimuli were repeated every 10 sec. All EPSP slope analyses and average traces presented in the figures included episodes with failures. The mean normalized EPSP slopes in the STDP summary figure (Fig. 2-1c) were: 0.34 ± 0.15 at -10 ms (n = 4), 0.89

± 0.30 at -5 ms (n = 5), 1.10 ± 0.43 at 0 ms (n = 5), 1.34 ± 0.35 at +5 ms (n = 5), 2.09 ±

0.32 at +10 ms (n = 11), and 0.75 ± 0.18 at 25 ms (n = 7). Input resistance was estimated from responses to small hyperpolarizing current pulses applied at the beginning of each episode. We observed no correlation between input resistance and the magnitude of EPSP slope change following pairing (R2 = –0.16; P > 0.05).

Analysis of cell-attached granule cell recordings

Cell-attached recording used the same internal solution as the current-clamp recordings.

Action currents were recorded at a tip potential of 0 mV. We verified that neurons recorded in the cell-attached configuration were granule cells by converting to whole-cell recording conditions and injecting current pulses (4/4 neurons tested showed afterhyperpolarizations following bursts of action potentials). The number of action currents evoked was determined in a window from 2 to 50 ms following extracellular stimulation. Action currents evoked by granule cell layer stimulation were blocked by bath application of NBQX (5 μM) and D-APV (25 μM), indicating that they were synaptically driven (n = 3).

Analysis of IPSPs in mitral cell intracellular recordings

IPSPs were evoked by granule cell layer stimulation (16–30 μA), on-beam relative to the

52 mitral cell body, and approximately 100-200 μm from the mitral cell layer. Granule cell layer stimulation sometimes evoked relatively monophasic IPSPs, recorded at –50 mV, which were completely blocked by 10 μM gabazine. However, in some experiments, the same type of stimulation evoked complex, biphasic responses that included an initial

IPSP component. Both classes of responses (illustrated by the two sets of recordings in

Fig. 2-2e) were included in the analysis. Experiments with an initial EPSP response were excluded. Mitral cell IPSPs were quantified by their peak amplitude. Individual episodes in which no hyperpolarizing IPSP component occurred within 50 ms were assigned an

IPSP amplitude of 0. The mean IPSP onset latency for the 9 mitral cell recordings in this study was 4.1 ± 1.2 ms.

53 Figure 2-1. Spike timing–dependent plasticity of proximal excitatory inputs to granule cells

(a) Left, potentiation of proximal EPSPs by +10-ms pairings. The time of the pairing is indicated by the vertical line in plot. Example traces from before and 17 min after pairing are shown above the plot. Reversing the pairing protocol (-10 ms, right) triggered long- term depression in a different cell. (b) Summary plot of changes in EPSP slope following

+10 and -10-ms pairing protocols. Neither pairing protocol affected input resistance (Rin bottom). (c) Summary of changes in EPSP slope versus pairing interval (Δt). *P < 0.02,

**P < 0.005. (d) Two-photon image of a granule cell with a stimulating electrode (Stim) in the external plexiform layer. (e) Plot of normalized EPSP slope following +10-ms pairing of distal stimuli with postsynaptic action potentials in four cells. Inset, average response before and 20 min after pairing. Data are presented as mean ± s.e.m. Animal procedures were approved by the Case Western Reserve University Institutional Animal

Care and Use Committee.

Figure 2-1

55 Figure 2-2. LTP evoked by TBS

(a) Superimposed responses to ten GCL stimuli before and 5 min after TBS. The fraction of episodes with evoked action currents within 50 ms of the stimulus is indicated above each panel. GC, granule cell. (b) Summary of the effect of TBS in nine cell-attached granule cell recordings. Data points indicate the mean number of action currents per episode with latencies <50 ms after the test stimulus (n = 30 control and 30 episodes, 5–

10 min after TBS for each cell). TBS increased the effectiveness of test stimuli evoking action currents in 7 out of 9 cells. The mean number of action currents per episode evoked by test stimuli increased significantly following TBS (filled bars, *P < 0.05). (c)

Top, schematic diagram illustrating an intracellular recording from a mitral cell (MC) and extracellular stimulation in the GCL. Bottom, IPSPs evoked by GCL stimulation recorded in a mitral cell held at –51 mV. Superimposed single trials (gray traces) and the average of ten consecutive responses are shown. (d) Blockade of GCL-evoked IPSP in a mitral cell by 10 μM gabazine. (e) TBS potentiated mitral cell inhibition. The responses that are shown are averages of ten trials for each time point from two different mitral cells. Control responses (gray) are superimposed on responses 10 and 20 min after TBS

(bold traces). Dashed lines indicate peak amplitude of control responses. (f) Summary of changes in normalized IPSP amplitude after TBS (n = 8). Responses were averaged over

30 trials for each time point in each mitral cell. Data are presented as mean ± s.e.m. **P

< 0.01.

Figure 2-2

57 Supplementary Figure 2-1. Properties of proximal and distal EPSPs recorded in granule cells.

(a) Schematic diagram illustrating granule cell (GC) recording configuration for proximal stimulation. (b) 2-Photon image of dye-filled granule cell with stimulating electrode near the proximal dendrite. Stimulating electrode tip indicated by white asterisk. (c) Five superimposed sequential responses to repeated (50 ms interval) proximal stimulation at –

50 mV; average response shown below. Inset shows block of excitatory synaptic response by NBQX (5 μM) and -APV (25μM). The residual response in NBQX and APV reflects a stimulus artifact that was insensitive to removal of extracellular Ca2+ (n= 3). (d)

Schematic diagram illustrating granule cell recording configuration for distal stimulation, in the external plexiform layer. (e) Focal distal stimulation activates dendrodendritic

EPSPs that exhibit paired-pulse depression. Superimposed sequential responses to distal stimulation; average trace shown below.

Supplementary Figure 2-1

59 Supplementary Figure 2-2. Pairing synaptic stimulation with intracellular depolarization induces long-term potentiation in a resting granule cell

(a) Example traces (grey) and averages of 10 consecutive responses (red) in control conditions and at 10, 20 and 30 minutes after LTP induction (50 shocks at 20 Hz; +10 ms pairing interval). The fast excitatory response was abolished by bath application of D-

APV (25 μM) and NBQX (5μM) at the conclusion of the experiment (right traces).

Recordings and LTP induction both were at –70 mV, the resting potential of this granule cell. (b) Plot of EPSP slope versus time for the granule cell shown in a. Average EPSP slope in control conditions indicated by dashed line. We verified that test stimuli did not evoke IPSPs by recording responses at -50 mV before and at 32 min after pairing (not shown). (c) Summary of change in EPSP slope evoked by +10 ms pairing stimulation in the granule cell shown in a. Pairing significantly increased EPSP slope, measured from

5 to 15 min after pairing (***P< 0.0001; unpaired t-test). Granule cell response to the first 5 parings in the +10 ms pairing STDP protocol. Stimuli indicated by *.

Supplementary Figure 2-2

61 Supplementary Figure 2-3. Proximal EPSPs recorded in granule cells were not potentiated by trains of postsynaptic action potentials or presynaptic stimuli presented separately

Pairing both stimuli (+10 ms timing) potentiated EPSPs without altering mean input resistance.

Supplementary Figure 2-3

63 Supplementary Figure 2-4. Pairing stimulation-induced potentiation of proximal granule cell EPSPs requires NMDA receptors

(a) Left, blockade of NMDA receptors with -APV (25 μM) diminished the proximal

EPSPs recorded at –55 mV but not at –80 mV in the same granule cell. (b) Plot of normalized EPSP slope before and after +10 ms pairing in APV from 8 granule cells.

Pairing stimulation in the presence of APV induced a modest, though not statistically significant (P > 0.05), decrease in EPSP slope after ~3 min. Inset shows average responses before pairing stimulation and 15 min after pairing from one granule cell. (c)

Summary plot of change in proximal EPSP slope 5–15 min after +10 ms pairing stimulation in control conditions (n = 11) and in APV (n = 8). ** P < 0.005.

Supplementary Figure 2-4

65 Supplementary Figure 2-5. Theta-burst stimulation potentates EPSPs in a P16 rat

(a) Average EPSP responses recorded in a P16 rat granule cell at –70 mV in control conditions and following theta-burst stimulation (TBS) in the granule cell layer. Each trace shows the average of 10 consecutive traces. Dashed line indicates peak amplitude of control EPSP. (b) Response to the first theta-burst stimulation in the granule cell shown in a. (c) Summary of change in EPSP slope 5 to 15 minutes after theta-burst stimulation in the granule cell shown in a-b. TBS significantly increased EPSP slope (*** P < 0.0001; unpaired t-test) in this granule cell. Over the population of 5 granule cells analyzed,

TBS increased EPSP slope to 122 ± 4.9% of control (significantly greater than 1; P <

0.02; population means also significantly different with paired -test; P < 0.05; n = 5).

Supplementary Figure 2-5

67 Supplementary Figure 2-6. Theta-burst stimulation potentates EPSPs in a P30 rat

(a) Average EPSP responses in a P30 rat granule cell at –69 mV to granule cell layer stimulation in control conditions and after TBS. Dashed line indicates peak amplitude of control EPSP. (b) Plot of EPSP slope versus time in the granule cell shown in a.

Dashed line indicates mean control EPSP amplitude. (c) TBS significantly increased

EPSP slope, measured 5 to 15 minutes after TBS, in the granule cell shown in a. ** P <

0.005 (unpaired t-test).

Supplementary Figure 2-6

Chapter 3 Extra-glomerular Layer Excitation of Olfactory Bulb Mitral Cells

Mediated through a Carbenoxolone and NBQX-sensitive Pathway

70 Abstract

Most previous studies in the olfactory system are based on the premise that the sensory afferent in the glomeruli is the only source of excitatory input for mitral cells. All other synaptic input for mitral cells is inhibitory and functions in the modulation of the sensory afferent. In this study, I found that depolarization in mitral cells could be evoked by extracellular stimulation in the granule cell layer in olfactory bulb slices without a glomerular layer. The depolarization evoked in these mini slices was diminished in the presence of NBQX or TTX, and significantly reduced after 30 minutes bath application of the gap junction blocker carbenoxolone. In addition, dye-coupling was observed between mitral cells and small processes, which ascended through the external plexiform layer, and terminated at the cut edge of the mini slice. The electrophysiological and imaging results are consistent with the suggestion of the presence of gap junctions on mitral cell lateral dendrites, presumably with the dendrites or axons of the external tufted cells. This excitatory pathway could be simultaneously activated by two independent

GCL stimuli that were 250 μm apart. Therefore, if this excitation was from the external tufted cells, mitral cells may respond to sensory input into “off-beam” glomeruli, in addition to the receptor neurons that drive EPSPs on the glomerulus innervated by their apical dendrites. The finding of this new type of excitation, which is sensitive to NBQX,

TTX, and carbenoxolone, on mitral cell lateral dendrites may extend our understanding of the information coding in the olfactory bulb.

71 Introduction

Volatile odorants are recognized by distinct combination of olfactory receptors, each of which expresses a single type of receptor protein (Buck and Axel, 1991; Vassar et al., 1993; Axel, 1995). One olfactory receptor typically projects to only two glomeruli in each olfactory bulb to form excitatory synapses with a group of mitral cells (Buck, 1996).

It has been stated in most of olfactory systems that the sensory afferent in the glomeruli is the only source of excitatory input for mitral cells, while all other synaptic input is inhibitory acting as the modulation of excitatory input (Shepherd and Greer, 1998). Like in other brain regions (Buzsaki and Chrobak, 1995; Cobb et al., 1995; Fricker and Miles,

2001), GABAergic inhibition from the bulbar intrinsic interneurons precisely control spike timing of mitral cells and regulate mitral cell synchrony as well (Murphy et al.,

2004; Murphy et al., 2005; Schoppa, 2006).

The excitation from a local network also can regulate the output, by mutual excitation (Steriade et al., 1993) or balancing the inhibition (Shu et al., 2003). Gap junction coupling between mitral cells is found on their apical dendrites (Schoppa and

Westbrook, 2002; Christie and Westbrook, 2006), via which the excitation in one mitral cell can be passed to the coupled one. This gap junction leads to glomerulus-specific correlated spiking in mitral cells that underlies the fast synchrony (Schoppa and

Westbrook, 2002; Christie et al., 2005; Christie and Westbrook, 2006). Besides electrical coupling, lateral excitation that mediated by AMPA autoreceptors also occur between glomerulus-specific mitral cells mediating the glomeruli-specific slow synchrony

(Carlson et al., 2000; Puopolo and Belluzzi, 2001; Schoppa and Westbrook, 2001;

72 Christie et al., 2005). In the glomerular layer, mitral cells also receive excitation at their apical dendrites from the external tufted cells, that project to a same glomerulus (Jan et al., 2009). Although the external tufted cells are thought to principally drive inhibitory neurons (Hayar et al., 2004; Wachowiak and Shipley, 2006), Jan (2009) showed that the external tufted cells also drive mitral cells via both electrical coupling and unidirectional chemical synapse on the apical dendrites and the interactions only occur between the cells that project to a same glomerulus (Jan et al., 2009). However, none of the studies above excludes the possibility of excitation received by mitral cell lateral dendrites. It is known that in the external plexiform layer (EPL), mitral cells can receive NMDAR mediated self- and lateral excitation along the lateral dendrites (Isaacson, 1999; Friedman and Strowbridge, 2000; Salin et al., 2001). Both of these excitation on the lateral dendrites have functional consequence in the bulbar output: NMDAR mediated self- excitation increases the excitability of the mitral cell via prolonging periods of phasic firing (Friedman and Strowbridge, 2000), and counteracts the recurrent inhibition as well

(Salin et al., 2001); while lateral excitation increases mitral cell synchrony to enhance response specificity and contributes to downstream signal processing in the olfactory cortex. In addition, the physiological function of the widely expressed gap junctional proteins, including connexin36 (Cx36) (Christie et al., 2005; Kosaka et al., 2005), connexin43 (Cx43) (Reyher et al., 1991; Miragall et al., 1996), connexin45 (Cx45)

(Zhang and Restrepo, 2002) and connexin32 (Cx32) (Reyher et al., 1991), in extraglomerular layers especially on mitral and tufted cells is not fully understood.

Therefore, more studies are needed to reassess the excitatory input to mitral cells.

In this study, I found that extracellular stimulation in the granule cell layer (GCL)

73 evoked both IPSPs and EPSPs in mitral cells, even when the stimulation was far off- beam (>200 μm) of mitral cells somata. The GCL-evoked IPSPs represented the synaptic input from granule cells, which showed varied onset latency (Gao and Strowbridge,

2009); however these EPSPs were unexpected. To identify the source of this excitation, the intact olfactory bulb slices were dissected to remove the glomerular layer; in addition, all recorded mitral cells from these mini slices were confirmed to lack glomerular input by two-photon live imaging that showing truncated apical dendrites at the cut off edge. In the mini olfactory bulb slices, the depolarization in mitral cells was able to be evoked by the GCL stimulation. This depolarization required the function of AMPARs as well as axonal conductance. It was also shown that the independent depolarization could be linearly summated in mitral cell somata. Moreover, the EPSP-like response evoked in these mini slices was significantly reduced after 30 min bath application of gap junction blocker carbenoxolone when the mitral cells were held at a hyperpolarized membrane potential –70 mV in APV and gabazine (~ 70% blockade, P < 0.01, n = 5); the blockade was also significant when the mitral cells were held at a depolarized membrane potential around –50 mV (~ 78% blockade, P < 0.05, n = 5). Dye-coupling was also observed between mitral cells and small processes that apparently connected to somata or lateral dendrites. These neuronal processes ascended through the external plexiform layer, and some of them terminated at the cut edge of the mini slice. Several different models were proposed and discussed, basing on the anatomical and electrophysiological evidence that obtained by this study. The finding of a new type of excitation, which is sensitive to

NBQX, TTX, and carbenoxolone, on mitral cell lateral dendrites may extend our understanding of the information coding in the olfactory bulb.

74 Materials and methods

Olfactory bulb brain slice preparation and recording methods

Horizontal brain slices (300 μm thick) were prepared from olfactory bulbs of

P14–21 Sprague-Dawley rats using a modified Leica (Nussloch, Germany) VT1000S vibratome, as described previously (Halabisky and Strowbridge, 2003; Presseler and

Strowbridge, 2006; Balu et al., 2007). Slice preparation and maintenance was performed in an artificial cerebrospinal fluid (ACSF) with reduced Ca2+ that contained 124 mM

NaCl, 2.6 mM KCl, 1.23 mM NaH2PO4, 3 mM MgSO4, 26 mM NaHCO3, 10 mM dextrose and 1 mM CaCl2. For the mini slices, the entire glomerular layer was removed under a 5x objective. Patch clamp recordings were carried out at 30 oC in a submerged recording chamber using the following ACSF: 124 mM NaCl, 3 mM KCl, 1.23 mM

NaH2PO4, 1.2 mM MgSO4, 26 mM NaHCO3, 10 mM dextrose and 2.5 mM CaCl2. Both dissecting and recording ACSF solutions were continuously oxygenated with 95% CO2 /

5% O2. A stereo microscope (Zeiss) was used to take pictures of intact and dissected brain slices. Whole-cell patch-clamp recordings were made onto mitral cells visualized using a 60× water immersion objective, IR-DIC optics and a frametransfer CCD camera

(Cohu 6412-2000, Poway, CA) attached to an upright, fixed-stage microscope (Olympus

BX51WI, Center Valley, PA). Electrophysiological data was acquired through a

Axopatch 1D amplifier (Axon Instruments, Sunnyvale, CA), low-pass filtered at 2 kHz

(FLA-01, Cygnus Technology, Delaware Water Gap, PA) and digitized at 5 kHz using

16-bit analog-to-digital converters (Instrutech ITC-18, Port Washington, NY). Data acquisition and analysis was carried out using custom programs written in Visual Basic 6

(Microsoft, Redmond, WA) and Matlab (Mathworks, Natick, MA). Patch electrodes (3-6

75 MΩ) used for current clamp recording contained the following internal solution: 140 mM

K-methylsulfate, 4 mM NaCl, 10 mM HEPES, 200 μM EGTA, 4 mM MgATP, 300 μM

Na3GTP, 10 mM phosphocreatine. Patch electrodes used for voltage clamp recordings contained (in mM): Cs-methanesulfonate 115, NaCl 4, TEA-methanesulfonate 25, QX-

314 5, HEPES 10, EGTA 1, MgATP 4, Na3GTP 0.3, and phosphocreatine 10.

Fluorescence indicator 100 μM Alexa488 or 100 μM lucifer yellow was added to the internal solution in experiments using 2-photon guided stimulation. Lucifer yellow was used because of its small molecular weight that allows it to pass through gap junctions

(Stewart, 1978). We used a sharpened tungsten microelectrode (FHC, Bowdoin, ME) positioned in the granule cell layer to evoke synaptic responses in mitral cells. All chemicals were obtained from Sigma except Alexa488 hydrazide and lucifer yellow

(Invitrogen, Carlsbad, CA). Receptor antagonists (APV, NBQX, GYKI 52466, gabazine,

TTX and carbenoxolone) were applied by bath perfusion. All results presented in the text, and in the summary plots represent mean ± s.e.m. unless noted, statistical significance was determined using Student's t test. Membrane potentials reported were not corrected for the liquid junction potential.

Two-Photon imaging

Live imaging experiments utilized a custom two-photon microscope based on the

Verdi V10 pump laser, Mira 900 Ti-sapphire laser (both from Coherent, Santa Clara, CA) and a high-speed XY galvanometer mirror system (6210; Cambridge Technology).

Intracellularly loaded fluorescent dyes were excited at 830 nm through a 60× water- immersion objective (Olympus). Emitted light was detected through an epifluorescent light path that included a 700DCLPXR dichroic mirror, a BG39 emission filter (both

76 from Chroma Technology), and a cooled PMT detector module (H7422P-40;

Hamamastu). Photomultiplier output was converted into an analog voltage by a high- bandwidth current preamplifier (SR-570; Stanford Research Systems). Custom Visual

Basic software written by BWS controlled the scanning system and image analysis functions. Laser beam intensity was controlled electronically through a Pockels cell attenuator (ConOptics) and a shutter (Uniblitz). In most experiments, the output of the

Mira laser was attenuated by 90%−95%.

Data acquisition and analysis

Electrophysiological data were recorded and analyzed using custom software written in Visual Basic 6 (Microsoft) and Origin 7.5 (OriginLab). Current and voltage records were low-pass filtered at 2 kHz and then digitized at 5 kHz using a 16-bit A/D converter (ITC-18, Instrutech). Series resistance was typically <20 MΩ and was routinely compensated by >80% in voltage-clamp experiments.

Mitral cells were identified by their morphology and by the intermittent firing pattern 150−650 pA, 3000 ms duration current step. GCL evoked EPSP slope was calculated from the initial rising phase 1.2 ms after the stimulation as the linear fit of the somatic membrane potential over 4 ms period. To measure paired-pulse ratio, we first generated an average response from 3−10 trials (50 ms ISI) in each cell. Paired-pulse ratio was determined from the initial EPSP slopes in this average response. EPSP amplitude was measured as the peak value over 50 ms period of the initial rising phase.

77 Results

GABA release from granule cells mediates both recurrent and lateral inhibition on mitral cells through the reciprocal dendrodendritic synapses (Shepherd and Greer, 1998).

As previously reported (Gao and Strowbridge, 2009), extracellular stimulation in the granule cell layer (GCL) evoked mitral cell IPSPs with varied onset latency (Figure 3-1).

Interestingly, IPSPs could be evoked with the stimulation that was both “on-beam” and

“off-beam” of the mitral cell somata (Figure 3-1). The “off-beam” stimulation presumably triggered dendrodendritic inhibition (DDI) on the distal site of mitral cell lateral dendrites based on the way granule cells project their apical dendrites toward the

EPL (Shepherd and Greer, 1998); however further experiments are required to confirm the position of Cl- influx through GABA receptors (GABARs). Different dendritic positions of GABAR mediated Cl- influx may lead to functional difference between DDI on proximal and distal sites due to their distinct kinetics (Rall, 1967; Berman and Maler,

1998). The distributed onset latency of IPSPs that found during the mapping of GCL- evoked mitral cell response suggested that different synaptic pathways could be involved, for instance monosynaptic or multisynaptic transmission (Figure 3-2). Those IPSPs with relatively short onset latency (in most cases < 3.0 ms) were thought to be monosynaptic, which could be triggered by the passive depolarization of granule cell apical dendrites by

GCL stimulation (Figure 3-2 A, top). These monosynaptic IPSPs were insensitive to

NBQX (5 μM) and APV (25 μM), that AMPAR and NMDAR antagonist respectively, while completely blocked by GABAAR antagonist gabazine (10 μM) (Figure 3-2 A). The

IPSPs with long onset latency (usually > 3.0 ms) were diminished in the bath application of NBQX and APV, which suggested that these multisynaptic IPSPs required excitatory

78 glutamatergic synaptic transmission (Figure 3-2 B). It is known that GCL stimulation activates glutamatergic excitatory input on the proximal sites of granule cell apical dendrites, in part from the piriform cortical feedback projection (Balu et al., 2007). IPSPs with long onset latency were also blocked by gabazine (dat not shown).

During mapping the GCL-evoked postsynaptic potential in mitral cells, some unexpected EPSP-like responses were also recorded, even in the mitral cells that had truncated apical dendrites (Figure 3-3 A, B, n = 3). The EPSPs in the mitral cells with truncated apical dendrites usually showed all-or-none responses to the extracellular stimulation (Figure 3-3).These EPSPs were persisted in APV, while completely blocked by NBQX (Figure 3-3 C). GYKI52466 (40μM), the antagonist that is more selective to

AMPAR, also reduced most GCL-evoked EPSPs in all tested mitral cells; however those mitral cells may contain their apical dendrites (Figure 3-3 D). Extracellular stimulation in the GCL may trigger glutamate release from mitral cells lateral dendrites through back propagation of action potentials, and the glutamate spillover is able to induce self- and lateral excitation; however both of self- and lateral excitation of mitral cells are mediated by NMDARs (Isaacson, 1999; Friedman and Strowbridge, 2000; Salin et al., 2001).

Therefore, AMPAR-mediated EPSP-like response may indicate an unknown excitation, presumably on the somata or lateral dendrites.

Most previous studies in the olfactory system state that the sensory afferent in the glomeruli is the only source of excitatory input for mitral cells (Shepherd and Greer,

1998). Although two-photon live image showed the truncated apical dendrites in my tested mitral cells (n = 3), the brain slice still anatomically contained the glomerular layer.

To exclude the function of excitatory input on the apical dendrites within glomeruli, I

79 removed the glomerular layer from the intact olfactory bulb slices (Figure 3-4). The dissected slices only contained GCL, internal plexiform layer (IPL), the mitral cell layer

(MCL), and partial EPL, while the sphere structure (Allison, 1953) in the glomerular layer was not observed in mini slices (Figure 3-4 C1, C2). Two-photon live image of dye- filled mitral cells showing truncated apical dendrites at the cut-off edge (Figure 3-5 A) was used to rule out the possibility of remaining any glomerular structure in the mini slice. Intracellular fills of mitral cell with lucifer yellow (100 μM) (Stewart, 1978) leaded to dye-coupling between mitral cells and small processes (Figure 3-5). These processes only ascended through the EPL and some of them terminated at the cut off edge of the mini slice; there were no dye-coupled somata were found in my experiments, which suggested that these neurons probably located in the glomerular layer. The dye-coupling experiments were only conducted in the first neuron being attempted in one slice, avoiding dye diffusion from other dye-filled cells. Two-photon images that taken at continuous time periods showed that the foreign processes started to be observed later than the recorded neurons (Figure 3-5 B, top), which suggested that these processes might be filled with the dye diffusing through gap junctions. In addition, that the foreign processes being filled at a later time point excluded the possibility of the contamination from the dye leaked from the patch pipette. These processes were likely from a subpopulation of the external tufted cells, which has ramifying axon arbors in EPL, MCL and GCL (Macrides and Schneider, 1982; Antal et al., 2006) as well as basal dendritic arbors in EPL (Antal et al., 2006). Furthermore, z-stack of the image (Figure 3-5 B, bottom) showed some unexpected dendrites in the deeper layer, which might be from the clustered mitral cells, since a typical mitral cell has one lateral dendrite that bifurcates in

80 EPL on the each side of the somata.

I then asked whether GCL stimulation could evoke mitral cells depolarization in the dissected slice. GCL-evoked depolarization was found in both voltage clamp recordings (n = 5) and current clamp recordings (n = 8). In both recording modes, the amplitude of depolarization did not increased by hyperpolarizing the mitral cell's membrane potential (Figure 3-6 A), as “conventional” EPSP would be (Jonas et al., 1994).

Electrical coupling is a potential mechanism that underlies this unusual depolarization, although both the distal input in the lateral dendrites and the large size of mitral cell somata may also lead to this phenomenon. The depolarization was blocked by NBQX

(5μM) and TTX (1μM) (Figure 3-6 B, C, D, E). Then, I tested the effect of carbenoxolone (100 μM), an agent that blocks gap junctions, on the depolarization.

Unlike octanol and halothane (Puil and el-Beheiry, 1990; Puil et al., 1990; Pocock and

Richards, 1993; Rorig et al., 1996), carbenoxolone has been reported to have no effect on excitatory synaptic transmission and intrinsic excitability (Kohling et al., 2001; Yang and

Michelson, 2001; Rouach et al., 2003). However, other recent studies suggest that carbenoxolone can inhibit AMPAR-, NMDAR- and GABAR- mediated synaptic transmission (Rouach et al., 2003; Chepkova et al., 2008; Tovar et al., 2009). However, in the olfactory bulb carbenoxolone is demonstrated to block connexin36 mediated gap junctions on the apical dendrites between mitral cells within glomeruli at this concentration (Schoppa and Westbrook, 2002). APV and gabazine were added to unmask the GCL-evoked depolarization. EPSPs were significantly reduced after 30 minutes bath application of carbenoxolone (Figure 3-7 A, B and C, ** P < 0.01, n = 5), so was EPSC

(demonstrated in one cell). The inhibitory effect of carbenoxolone (100 μM) on AMPAR-

81 mediated current are initiated and stabilized in 8 sec, and the reduction is about 20% of the control current (Tovar et al., 2009); however, in my experiments about 70% of the depolarization was blocked after 30 min (Figure 3-7 B, n = 5). Therefore, this blockade by carbernoxolone was not likely due to the inhibition of AMPAR-mediated synaptic transmission. In absence of APV and gabazine, mitral cells were held at -50 mV, showing both EPSPs and IPSPs (Figure 3-7 D). 30 minutes bath application of carbenoxolone significantly reduced EPSPs (Figure 3-7 E, F, * P < 0.05, n = 5), while unmasked IPSPs in the time frame (demonstrated in Figure 3-7 D). The sensitivity to carbenoxolone suggested that this GCL-evoked depolarization might recruit gap junctional coupling with other neurons on mitral cells lateral dendrites or somata. As that suggested by the cable theory, some independent inputs are able to be linearly summated in somata (Rall, 1967), so I asked whether this was true for the GCL-evoked depolarization in mitral cells. Two stimulation electrodes were positioned in the GCL with the distance ranged from 250 μm to 320 μm, on the same side of the mitral cell somata (Figure 3-8 A). This distance between two stimulation as larger than the diameter of a typical glomerulus (< 200 μm)

(Allison, 1953). EPSPs from the two separated stimulation were found to be linearly

2 summated (R = 0.97, n = 13) (Figure 3-8 B1, C). Then I asked whether these two stimulations activated a same or different synaptic input? Paired-pulse stimulation (50 ms

ISI) was sequentially tested in each stimulation electrode (Figure 3-8 B2 left two panels), and also simultaneously given in both stimulation electrodes in different orders (Figure 3-

8 B2 right two panels). The different paired-pulse ratio (50 ms ISI) between the sequential and simultaneous stimulation suggested that these two stimulations did not activate a same input pathway (Figure 3-7 B2, n = 13). Simultaneous depolarization from

82 the two independent GCL stimuli raised a possibility that mitral cells may respond to sensory input into “off-beam” glomeruli, in addition to the receptor neurons that drive

EPSPs on the glomerulus innervated by their apical dendrites.

83 Discussion

Previous studies have reported that the excitation on mitral cells is within glomeruli, from olfactory neurons (Shepherd, 1972) and the external tufted cells that project to same glomeruli (Jan et al., 2009). In the regions outside of the glomerular layer, mitral cells receive inhibition from granule cells (Shepherd and Greer, 1998), as well as

NMDAR mediated self- and lateral excitation (Isaacson, 1999; Friedman and

Strowbridge, 2000; Salin et al., 2001), on their lateral dendrites. In this study, I showed an unexpected AMPAR-mediated excitation in dissected olfactory bulb slices that do not contain a glomerular layer (Figure 3-4), which suggested that this GCL-evoked excitation was on the mitral cell somata or lateral dendrites but different as either self- or lateral excitation. This AMPAR-mediated EPSP showed very slow kinetic rate (data not shown) and unusual responses to the different membrane potentials (Figure 3-6 A), which raised a possibility that it may not be only due to chemical synaptic transmission. Both optical imaging and electrophysiological approaches were applied to understand this unexpected excitation on mitral cells lateral dendrites.

Dye-coupling between mitral cells and unknown processes

Dye-coupling, the major evidence of gap junctions was examined in these mini slices. Lucifer yellow was used in my experiment as the fluorescent indicator because of its relatively small molecular weight that allows it to pass through gap junctions (Stewart,

1978). Dye-coupling was observed in small processes that were apparently connected with mitral cell lateral dendrites or somata (Figure 3-5). These neuronal processes ascended through the EPL, and some of them terminated at the cut edge of the mini slice.

84 These neuronal processes may belong to granule cells or some short-axon interneurons, the EPL interneurons, or the external tufted cells, all of which have processes expanding in the EPL and MCL. However, I never observed any processes that extend toward the

GCL, which made granule and short axon interneurons less likely to be the source. There are some multipolar interneurons in the EPL that having the soma in the superficial EPL and varicose dendrites in the deep EPL (Van Gehuchten and Martin, 1891; Schneider and

Macrides, 1978; Kosaka et al., 1994; Hamilton et al., 2005). However, these EPL interneurons appear to receive excitatory input from mitral cell apical dendrites or from the olfactory neurons, which implies that these neurons are less likely to receive excitation from the GCL stimulation in dissected slices (Hamilton et al., 2005). In addition, there is only a small subpopulation of the EPL interneurons that have processes expanding into the deep EPL or MCL (Kosaka et al., 1994). On the other hand, the external tufted cell, as one cell type near the glomerular layer, has ramifying axon arbors in EPL, MCL and GCL (Macrides and Schneider, 1982; Schoenfeld et al., 1985; Antal et al., 2006) as well as basal dendritic arbors in EPL (Antal et al., 2006). In addition, the dye-coupling between mitral-external tufted cells is reported to occur and be mediated through gap junctions on their apical dendrites (Jan et al., 2009); however the location of gap junctions has not been tested. To this extent, the external tufted cell is possible to form gap junctions with mitral cells. However, the absence of dye-coupling did not exclude the presence of gap junctions, because the coupling sites might locate in the distal dendritic sites as shown in rat neocortex (Gibson et al., 1999; Venance et al., 2000) or some junctional protein was relatively impermeable to the dye (Veenstra, 1996).

At the molecular level, several types of connexin protein may be the candidates

85 which mediate the dye-coupling that observed in my study. In the olfactory bulb the expression of Cx36 is not limited within glomeruli; it is present in EPL, mitral cell layer

(MCL) and GCL (Christie et al., 2005; Kosaka et al., 2005). The physiological function of Cx36 in the extraglomerular regions remains to be elucidated. Some other gap junctional proteins are also expressed in the mammalian olfactory bulb. For instance,

Cx34 is known to be expressed in EPL, mitral cell layer (MCL), internal plexiform layer

(IPL) and GCL to mediate electrical coupling between granule cells (Reyher et al., 1991;

Miragall et al., 1996). Cx34 is also found on mitral and tufted cell somata by immunochemical studies, while the physiological functions are unknown (Miragall et al.,

1996). Mitral cells also express Cx45 (Zhang and Restrepo, 2002) and Cx32 (Reyher et al., 1991), neither of which is understood for the physiological function.

Sensitivity to carbenoxolone

The GCL-evoked excitation on mitral cells was significantly reduced by carbenoxolone, an agent that blocks gap junction. Although carbenoxolone has less effect on synaptic transmission and intrinsic membrane properties than other gap junction blockers, like octanol and halothane (Puil and el-Beheiry, 1990; Puil et al., 1990; Pocock and Richards, 1993; Rorig et al., 1996), its inhibition on neurons is controversial and complex. Some studies report that carbenoxolone has no effect on excitatory synaptic transmission and intrinsic excitability (Kohling et al., 2001; Yang and Michelson, 2001;

Rouach et al., 2003), while other studies suggest that carbenoxolone can inhibit AMPAR-,

NMDAR- and GABAR- mediated synaptic transmission (Rouach et al., 2003; Chepkova et al., 2008; Tovar et al., 2009) and voltage-gated calcium channels (Vessey et al., 2004).

In spite of the controversy, carbenoxolone is still a popular blocker that used for testing

86 the presence of gap junctions. For example, it is demonstrated to block 91% of Cx36 mediated gap junctions on mitral cells apical dendrites within glomeruli (Schoppa and

Westbrook, 2002).

The reduction of the GCL-evoked EPSPs in dissected slices by carbenoxolone was slow, although it varied in individual mitral cells (Figure 3-7). This blockade by carbenoxolone was not likely due to its effect on intrinsic excitability, ion channels or

AMPAR-mediated synaptic transmission: first, all mitral cells that included in this study fired action potentials with normal size before and after carbenoxolone; second, carbenoxolone at 100 μM blocks 37% of Ca channels mediated response in 2 minutes

(Vessey et al., 2004), which is faster than the blockade in my study; finally, the inhibition of carbenoxolone on AMPAR-mediated synaptic transmission is also a rapid process that occurs and stabilizes in less than 1 min (Tovar et al., 2009). A full blockade by carbenoxolone was not found in most of mitral cells I examined, which could be due to a relatively short time period of the drug application. Half an hour bath application of carbenoxolone reduced the excitation to 0.30 ± 0.06 of the control in presence of APV and gabazine (n = 5), and 0.22 ± 0.16 in regular ACSF (n = 5). Both of the changes were statistically significant (in APV and gabazine: P < 0.01, n = 5; in ACSF: P < 0.05, n = 5).

The sensitivity to carbenoxolone is consistent with my anatomical and optical imaging results, suggesting that electrical coupling may be involved to triggering this excitation in mitral cells.

The pathway of signal transduction

The major question remained in this study is which signal transduction pathway is responsible for the GCL-evoked excitation in the mitral cells. There must be more than

87 one synaptic transmission that involved here to recruits the function of AMPAR, axonal conductance and gap junctions.

The GCL stimulation may drive an antidromic action potential in mitral cells, which in turn triggers glutamate release from their lateral dendrites. AMPA autoreceptors are found on the external tufted cell apical dendrites (Ma and Lowe, 2007); therefore glutamate spillover may activate AMPARs that possibly located on the external tufted cell basal dendritic terminals (Figure 3-9 A, blue processes). The slow kinetics of the

EPSP-like responses satisfies this hypothesis in the aspect of AMPA autoreceptor activation. Like that found in the cerebellar purkinje cells (Barbour et al., 1994), the slow kinetics of EPSP-like responses may indicate a slow clearance of the glutamate spillover from mitral cells, in part due to multiple releasing events (Zhou and Hablitz, 1997).

Alternatively, GCL stimulation may directly activate the external tufted cell axons, which has been shown to project to the GCL or expand in the EPL (Antal et al., 2006). In this situation, glutamate spillover from mitral cells may act on the external tufted cell axon terminals (Figure 3-9 A, red processes). As suggested by Schmitz (2001), the axonal gap junctions may provide a ultrafast synchrony compared to dendrodendritic couplings

(Schmitz et al., 2001). In addition, the processes in the two-photon images (Figure 3-5) are more likely to be axons based on their morphology. This picture, which shows the external tufted-mitral cell excitation pathway, is consistent with my anatomical, optical and electrophysiological results. In this hypothesis, the external tufted cell may drive mitral cells through their lateral dendrites, in addition to the interaction that on mitral cells apical dendrites within glomeruli (Jan et al., 2009). Moreover, the ability of responding to two GCL stimuli with the distance larger than 250 μm (Figure 3-8) raised

88 the possibility that mitral cells might be driven by sensory input into “off-beam” glomeruli, in addition to the receptor neurons that drive EPSPs on the glomerulus innervated by their apical dendrites. Glomerulus-specific and non glomerulus-specific coupling mediates different aspects of the bulbar synchrony (Mori et al., 1999; Christie et al., 2005), therefore the possible external tufted-mitral cell interaction merits a further study for understanding the mechanisms for the bulbar synchrony.

Alternatively, these neuronal processes may be from granule cells or some short axon cells, which project their dendrites into the MCL and EPL (Figure 3-9 B). The GCL stimulation activates the cortical feedback projection on granule cells or short axon cells

(Shepherd and Greer, 1998; Balu et al., 2007), and this excitatory input is mediated by

AMPARs and NMDARs. In addition, both of granule and mitral cells express Cx43

(Reyher et al., 1991; Miragall et al., 1996), which provides a molecular base of forming gap junctions between them. If either type of these interneurons is coupled with mitral cells, the GCL-evoked mitral cells depolarization represents the excitation that passed from the interneurons. In the model of Figure 3-9 B, intrinsic interneurons may modulate the bulbar output by both inhibition and excitation. However, dye-coupled processes in the IML and GCL that satisfying the possible interneuron-mitral cell coupling, were not been observed in my study. This model also seems to be independent of axonal conductance: the GCL stimulation may directly activate the excitatory input on these interneurons. Another question remained in this model is the slow rising time of mitral cells EPSPs, since the feedback projection evoked EPSPs in granule cells has the rising time less than 2 ms (Balu et al., 2007). Electrical coupling only changes the gain of the signal; therefore interneuron-mitral cell excitation should have a faster kinetic rate.

89 Although a distal dendritic location may lead to the slow kinetics, it is not likely to explain why most of the GCL-evoked EPSPs had very slow kinetics wherever the stimulation electrode was positioned. These three questions may limit the possibility of the interneuron-mitral cell excitation.

Potential functional role of “off-beam” mitral cell excitation

Usually gap junctions are thought to mediate inhibitory network oscillation, which have been described in neocortex (Galarreta and Hestrin, 1999; Gibson et al., 1999;

Deans et al., 2001), the olfactory bulb (Friedman and Strowbridge, 2003), hippocampus

(Hormuzdi et al., 2001) and thalamic reticular nucleus (Landisman et al., 2002).

Interestingly, the olfactory bulb network is also regulated by the gap junctions between excitatory neurons (Schoppa and Westbrook, 2002; Christie et al., 2005; Christie and

Westbrook, 2006; Pimentel and Margrie, 2008; Jan et al., 2009). Mitral-mitral cells coupling (Schoppa and Westbrook, 2002; Christie et al., 2005; Christie and Westbrook,

2006; Pimentel and Margrie, 2008) and mitral-external tufted cells coupling (Jan et al.,

2009) , which are glomerulus-specific and only present on their apical dendrites, mediate the intraglomerular synchrony. There is no previous study showing the presence of gap junction between bulbar principal neurons outside of the glomerular layer. If this carbenoxolone sensitive excitation is mediated by gap junctions, this study provides a physiological function of those widely expressed gap junctional protein. This excitation also suggests that mitral cells are not only driven by the sensory input through their apical dendrites, and the function of their secondary dendrites also should be reassessed. In spite of the uncertain pathway of this GCL-evoked excitation, the presence of it extends our understanding of the information coding in the olfactory bulb as well.

90 Figure 3-1. Mapping of GCL-evoked postsynaptic potentials recorded in a mitral cell

(A) Schematic diagram with 2-photon image of a dye-filled mitral cell illustrating the recording configuration for GCL stimulation. The numbers below the stimulation electrodes indicate the horizontal distance of stimulation position from mitral cell somata.

Glom Layer: glomerular layer, EPL: external plexiform layer. (B) Averaged postsynaptic responses to GCL stimulation on- and off-beam of mitral cell somata. The mean synaptic delay was 2.0 ms. The delay of onset time was shown below each trace. All traces were averaged from 5 consecutive episodes with stimulation intensity at 20 μA. Stimulation was indicated by red arrow. Lower traces represent responses evoked by the stimulation further off-beam.

Figure 3-1

92 Figure 3-2. Two types of mitral cell IPSPs evoked by GCL stimulation

(A) Upper, schematic diagram illustrating mitral cell recording configuration for direct granule cell apical dendrite activation. Lower, superimposed sequential and average (red) responses to GCL stimulation. Monosynaptic inhibitory response remained in NBQX (5

μM) and -APV (25 μM), but diminished in present of gabazine (10 μM). (B) Upper, schematic diagram illustrating mitral cell recording configuration for activating granule cell by the GCL stimulation. Lower, superimposed sequential and average (red) responses to the GCL stimulation. NBQX (5 μM) and APV (25 μM) abolished polysynaptic mitral cell IPSPs. Mean synaptic delay was indicated below each response.

Figure 3-2

94 Figure 3-3. Mitral cell EPSPs evoked by GCL stimulation

(A) GCL-evoked EPSP recorded in a mitral cell with truncated apical dendrite. Upper, 2-

Photon reconstruction of an Alexa488-filled mitral cell. Lower, evoked responses by

GCL stimulation at graded intensities with bath perfusion of APV (25 μM) and gabazine

(10 μM). The mitral cell was held at -69 mV. (B) Postsynaptic responses in a different mitral cell with truncated apical dendrite. Upper, 2-Photon reconstruction of an

Alexa488-filled mitral cell. Lower, evoked responses by GCL stimulation at graded intensities with bath perfusion of APV (25 μM) and gabazine (10 μM). Holding potential was -70 mV. (C) Blockade of mitral cell EPSPs by NBQX. Averaged response of 3-4 traces from the mitral cell shown in (A) was blocked by bath perfusion of NBQX (5 μM), and washout of NBQX, APV and gabazine recovered the GCL-evoked EPSPs. (D)

Blockade of mitral cell EPSPs by GYKI52466 (40 μM). Upper, superimposed sequential and average (black) responses to the GCL stimulation in APV and gabazine. Lower, after bath application of GYKI52466. The recordings showing in (D) was from the mitral cell with intact apical dendrites.

Figure 3-3

96 Figure 3-4. The lack of the glomerular layer in dissected olfactory bulb slices

(A) An intact olfactory bulb slice stained with methylene blue (1 mM). The acute olfactory bulb slice was stained with methylene bulb for 30 minutes before the picture was taken. GCL, granule cell layer. EPL, external plexiform layer. MCL, mitral cell layer.

GL, glomerular layer. (B1) An unstained intact olfactory bulb slice. (B2) Part of the intact slice within the blue circle in (B1). (C1) An unstained dissected olfactory bulb slice. (C2)

Part of the dissected slice within the blue circle in (C1).

Figure 3-4

98 Figure 3-5. Two-photon imaging of a mitral cell that intracellularly filled with lucifer yellow in dissected olfactory bulb slice

(A) Reconstruction of z-stack images showing a mitral cell from the dissected olfactory bulb slice. Apical dendrites were truncated at the cut-off edge that was indicated by the dash line. Two possible dye-coupled processes were indicated by asterisks. Inset, the typical morphology of the apical dendrites after they arrive in glomeruli, and this inset image was taken in a different mitral cell that from an intact slice. (B) Top, z-stack reconstructions of the mitral cell in (A) at different time points. The possible dye-coupled process started to be observed after 20 minutes. From 20 min, two more dendrite-like processes were filled. Bottom, images along z-axis at 60 min.

Figure 3-5

100 Figure 3-6. Blockade of the GCL-evoked mitral cell depolarization by NBQX and

TTX in dissected olfactory bulb slices

(A) The GCL-evoked mitral cell depolarization at different membrane potentials. (B) The

GCL-evoked mitral cell depolarization was sensitive to NBQX. Top, the average traces before/control (black) and after (blue) NBQX. Bottom, NBQX sensitive component (red) which had a similar kinetics as the control (black). Baseline was indicated by the grey dash line. (C) The GCL-evoked mitral cell depolarization was sensitive to TTX. The average traces before/control (black) and after (blue) TTX. Baseline was indicated by the grey dash line. (D) NBQX significantly blocked the GCL-evoked mitral cell depolarization. **P < 0.01, n = 5. (E) TTX completely blocked the GCL-evoked mitral cell depolarization. *P < 0.05, n = 3. In this figure, all depolarization were recorded in presence of APV and gabazine.

101

Figure 3-6

102 Figure 3-7. The GCL-evoked mitral cell depolarization was sensitive to carbenoxolone (CBX) in dissected olfactory bulb slices

(A) The GCL-evoked mitral cell EPSPs (at -70 mV, in APV and gabazine) were attenuated by bath application of CBX. The baseline was indicated by dash line. (B) The blockade by CBX in 5 different mitral cells with the presence of APV and gabazine. Two mitral cells were in CBX for 30 min, and other three were in CBX for 50 min. Dash line indicated the 40% of the control. (C) CBX significantly reduced the depolarization after

30 min. **P < 0.01, n = 5 (the same mitral cells showing in (B), in APV and gabazine).

(D) The GCL-evoked mitral cell EPSPs (at -50 mV) were significantly reduced by CBX, while IPSPs showed less effect from CBX. Baseline was indicated by dash line. (E) The blockade by CBX showing in 5 different mitral cells. One mitral cell was in CBX for 30 min, and other four were in CBX for 40 min. Dash line indicated the 75% of the control.

(F) CBX significantly reduced the depolarization after 30 min. *P < 0.05, n = 5 (the same mitral cells showing in (E)).

103

Figure 3-7

104 Figure 3-8. Activation of mitral cells by two independent GCL stimulation with a distance greater than 250 μm in dissected olfactory bulb slices

(A) Schematic diagram illustrating mitral cell depolarization evoked by two GCL stimulation with a distance > 250 μm. GC, granule cell. MC, mitral cell. (B1) EPSPs evoked by individual GCL stimulation that indicated by black and red asterisks (left and middle), and EPSPs simultaneously evoked by the two stimulation (right). The superimposed blue traces were the average response of 3-5 traces. Baseline was indicated by dash line. (B2) EPSPs in response to individual or mixed paired-pulse stimulation (50 ms ISI). The different paired-pulse ratio between pure and mixed paired-pulse stimulation suggested that these two stimulations did not activate the same synaptic input. (C) The excitation from two GCL stimulations could be linearly summated in mitral cells. R2 =

0.97, n = 13.

105

Figure 3-8

106 Figure 3-9. Two potential models that illustrating the GCL-evoked mitral cell excitation in dissected olfactory bulb slices

(A) The schematic diagram illustrating the external tufted-mitral cells excitation pathway.

One subpopulation of the external tufted cells (ETC) has ramifying basal dendrites in the

EPL and axon arbors in the EPL, MCL and GCL. Question marks indicate the possible gap junctions between ETC and mitral cells. In ETC, basal dendrites were in blue, while axons were in red. GC, granule cell. MC, mitral cell. (B) The schematic diagram illustrating the interneuron-mitral cell excitation pathway. SA, short axon cells. GC, granule cell. Question marks indicate the possible gap junctions between SA or GC and mitral cells.

107

Figure 3-9

108

Chapter 4 General Discussion

109 Overview

As the first relay station of olfaction perception, the olfactory bulb translates external odor information that captured by the olfactory receptors into an internal format for the later processing in higher brain regions (Figure 4-1). This internal format of sensory information is represented by the activity of mitral cells, which generates distinct spatial and temporal patterns in response to different odorants. Mitral cell dendrites only project to one glomerulus to receive sensory input from olfactory neurons, which express only one type of 1,000 olfactory receptors (Malnic et al., 1999). This glomerulus-specific projection results in the spatial patterns of mitral cell activation. However, the spatial patterns have a broad overlap to many different odorants owing to the broad tuning of the olfactory receptors to those structurally related odor ligands (Buck, 1996; Malnic et al.,

1999; Araneda et al., 2000). The tuning specificity of mitral cells to odorants is enhanced by the time coding system, which is composed by the synaptic interactions on their tufted and lateral dendrites. Furthermore, the temporal modulation also leads to bulbar synchrony, which is thought to be a general strategy of information coding in CNS.

Within one single glomerulus, the specific sensory input is shared by a group of mitral cells; therefore the activity of glomerulus-specific mitral cells is likely to be synchronized. This glomerulus-specific synchrony also reflects the fact that the bulbar output is organized as a mitral cell network with distinct spatial and temporal characters

(Mori et al., 1999; Bozza and Mombaerts, 2001). The intraglomerular synchrony is regulated by electrical coupling and lateral-excitation between mitral cells (Schoppa and

Westbrook, 2002; Christie et al., 2005; Christie and Westbrook, 2006; Pimentel and

110 Margrie, 2008). Gap junctional protein connexin36 expression is found in all layers in the olfactory bulb (Christie et al., 2005; Kosaka et al., 2005). Electrical coupling plays a predominant role in spreading excitation to amplify the sensitivity of mitral cell to the sensory input. Blocking gap junctions on mitral cell apical dendrites disrupts the fast synchrony (Christie et al., 2005). In contrast to the fast synchrony, glomerulus-specific slow synchrony is mediated in part by lateral excitation through glutamate spillover

(Carlson et al., 2000; Puopolo and Belluzzi, 2001; Schoppa and Westbrook, 2001;

Christie et al., 2005). The sensitivity to the sensory input, as well as mitral cell synchrony, are also enhanced by the electrical and chemical connections between external tufted and mitral cells that project into a same glomerulus (Hayar et al., 2004; Jan et al., 2009).

Besides distinct sensory afferents, interglomerular specificity is also regulated by the lateral inhibition from periglomerular cells that have neuronal processes across several glomeruli. These GABAergic periglomerular cells, which receive excitatory input from mitral/tufted cells and the external tufted cells, modulate sensory afferent and regulate mitral cell excitability (Murphy et al., 2004; Murphy et al., 2005). In general, in the glomerular layer different types of synaptic interactions on the tufted dendrites, including electrical coupling, lateral excitation and inhibition, cooperate to gate mitral cell response to specific sensory stimulation.

Signal gating on mitral cell lateral dendrites also plays an essential role in bulbar synchrony and odor discrimination. Mitral cell lateral dendrites form reciprocal dendrodendritic synapses with granule cells. Glutamate release from mitral cells activates

AMPARs and NMDARs on the granule cell distal dendrites and trigger GABA release.

Due to the connection between one granule cell and a large number of mitral cells (about

111 1:100) (Mori, 1987), GABA release from one granule cell not only mediates recurrent inhibition but also lateral inhibition. The lateral inhibition from granule cells mediates interglomerular synchrony (Schoppa, 2006), which is especially important in temporal binding of signals from different odorant receptors (Mori et al., 1999; Lledo et al., 2005).

The lateral inhibition also helps to sharpen mitral cell activity via extracting response from nearby neurons, similar to that in the visual cortex (Weliky et al., 1996). Blocking the reciprocal dendrodendritic transmission disrupts the ability of odor discrimination

(Yokoi et al., 1995). Communication with granule cells also helps to recruit indirect synaptic regulation from GCL. Piriform cortical feedback projection and Blanes cells make excitatory and inhibitory synapses onto granule cells, respectively (Presseler and

Strowbridge, 2006; Balu et al., 2007). By regulating granule cell excitability, these inputs indirectly affect dendrodendritic transmission to gate mitral cell spiking. The long lateral dendrites also provide a structural advantage for mitral cells to receive self- and lateral excitation over distance, both of which are mediated by NMDAR (Isaacson, 1999;

Friedman and Strowbridge, 2000; Salin et al., 2001). Activation of NMDA autoreceptors is one of the major functions of AP back propagation through the lateral dendrites.

NMDAR mediated self-excitation increases the excitability of the mitral cell via prolonging periods of phasic firing (Friedman and Strowbridge, 2000), and counteracts the recurrent inhibition as well (Salin et al., 2001); while lateral excitation increases mitral cell synchrony to enhance response specificity and contributes to downstream signal processing in the olfactory cortex.

Despite of all these mechanisms for gating bulbar signal and evidence from previous in vivo studies that suggest the involvement of the olfactory bulb in familiar

112 recognition, pheromone learning and odor conditioning (Bruce and Parrott, 1960;

Brennan et al., 1990; Ferreira et al., 1992; Kendrick et al., 1992; Brennan and Keverne,

1997), the cellular mechanism that is responsible for olfactory learning and memory remains a mystery. My thesis work demonstrated the presence of spike timing-dependent plasticity (STDP), which depended on coincident activation of bulbar intrinsic neurons and cortical pyramidal neurons. In the downstream targets of granule cells, dendrodendritic inhibition on the mitral cell was significantly potentiated by the same protocol that induces granule cell LTP. Previous studies suggest that the feedback excitation from anterior piriform cortex appears to tonically remove Mg2+ from

NMDARs on granule cell distal dendritic terminal (Halabisky and Strowbridge, 2003); my results show that the feedback excitation also regulates a long-term change of dendrodendritic transmission. My findings provide a potential mechanism underlying olfactory learning and memory. The interaction between two different types of short-term plasticity of GABA release, which was demonstrated in my work to be dependent on the presence of glutamatergic synaptic transmission, implicates another mechanism of signal gating through lateral dendrites, presumably that relies on the history of the neuronal activity. In addition, I found that the dendrodendritic inhibition could be detected in mitral cell soma through a long distance, up to 300 μm along the lateral dendrites, which indicates the ability in mitral cells of integration information from a large subgroup of granule cells.

My work also provides an unexpected physiological function of gap junctional protein connexin36 outside the glomerular layer. All previous studies suggest that electrical coupling only occurs between glomerulus-specific mitral cells or mitral-to-

113 external tufted cells. However, an intensive expression of connexin36 is found outside of glomeruli, including EPL, MCL and GCL (Christie et al., 2005; Kosaka et al., 2005). I demonstrated an AMPAR mediated response in mitral cell by GCL stimulation in mini olfactory bulb slices that do not contain the glomerular layer. This AMPAR component is sensitive to carbenoxolone, an agent that blocks gap junction (Schoppa and Westbrook,

2002). By two-photon imaging, dye-coupling between neighboring mitral cells were occasionally found in the mini brain slices, as well as a mitral cell and glomerular layer originated axons that were presumably from external tufted cells (Kosaka et al., 2005;

Antal et al., 2006). All the evidence above suggests the presence of extraglomerular electrical coupling that is responsible for GCL-evoked AMPAR mediated response. The finding of extraglomerular gap junctions on mitral cells provides a novel mechanism underlying the mitral cell synchrony.

Taken together, the unaltered odor information undergoes synaptic modulation in the olfactory bulb, especially via the time coding system, to be reformatted into the electrical signals with distinct spatial-temporal patterns. Both chemical and electrical interactions are used for the time coding within the olfactory bulb. The temporal factors of the bulbar output are essential for later information processing in the olfactory cortex: the co-occurrence of intraglomerular and interglomerular synchrony not only increases the likelihood of driving a cortical neuron, but also make the cortical neurons integrators of different odorants. This integration reflects the fact that odorants are recognized by distinct combination of olfactory receptors (Malnic et al., 1999); on the other hand, distinct temporal patterns among mitral cells in response to different odorants underlie the odor discrimination ability.

114 Mechanisms underlying granule cell spiking

Granule cells receive two anatomically and electrophysiologically distinct classes of excitatory input on their proximal and distal dendrites (Balu et al., 2007). The reciprocal dendrodendritic synapses with mitral cells in the EPL are the major excitatory input on the distal dendrites (Price and Powell, 1970; Price and Powell, 1970). The activation of granule cells through the distal dendrodendritic input is dependent on

NMDARs (Isaacson and Strowbridge, 1998; Chen et al., 2000). The distal dendrodendritic excitatory input is facilitated by the focal paired-pulse stimulation, which is consistent with the stimulation given to the lateral olfactory tract (Balu et al., 2007).

The axons in the granule cell layer, partially from the piriform cortex, make excitatory synapses on the granule cell proximal dendrites (de Olmos et al., 1978; Haberly and Price,

1978; Shipley and Adamek, 1984; Balu et al., 2007). However, focal stimulation close to the proximal dendritic spines or activation of anterior piriform cortex reveals the presence of a short-term facilitation at this site (Balu et al., 2007). The proximal excitatory input has the ability to gate the distal excitatory synaptic transmission. Gamma-frequency stimulation in the GCL that activates the granule cell appears to tonically remove Mg2+ from the distal NMDARs (Halabisky and Strowbridge, 2003). This interaction of the inputs from different brain regions, in part, presents a major advantage for the granule cells to be the target of mediating the bulbar long-term plasticity based on the behavior context. The STDP found on the granule cells requires coincident activation of cortical pyramidal neurons and granule cells. However granule cells are normally hyperpolarized

(-70 mV), which prevents them from firing action potentials at the resting membrane potential. Although previous studies demonstrated that GABA release occurs in the

115 presence of TTX (Jahr and Nicoll, 1980; Isaacson and Strowbridge, 1998), in vivo experiments suggest that the granule cell fires action potentials to respond to odorants

(Mori and Takagi, 1977; Wellis and Scott, 1990; Luo and Katz, 2001). In brain slice recording, granule cell spiking is also able to be elicited by the stimulation of olfactory nerves (Lagier et al., 2004). Controversially, Cang (2003) demonstrates that granule cell either does not spike or spikes at a low rate in response to odor stimulation (Cang and

Isaacson, 2003). However, the mechanism of how granule cell spiking is precisely regulated by the distal and proximal excitatory input is not fully understood.

In the EPL the convergence of excitatory inputs from multiple mitral cells onto a single granule cell provides the anatomical base for the lateral inhibition between mitral cells (Yokoi et al., 1995; Brennan and Keverne, 1997; Shepherd and Greer, 1998), also raises the possibility that the granule spiking can be evoked by the summation of glutamatergic inputs onto the distal dendrites. In addition, the action potential initially evoked in the mitral cell soma can back- propagate along the lateral dendrites with the distance dependent on the number of spikes (Margrie et al., 2001). This evidence suggests that the granule cell can receive excitatory input from very distal sites of the mitral cell lateral dendrites, which further increase the level of convergence. If the distal input gates the granule spiking, the time window between the feedforward and feedback excitation, from mitral cells and cortical neurons respectively, onto the granule cell regulates the level and polarity of the long-term plasticity.

Alternatively, the proximal excitatory input from the piriform cortical neurons controls the granule cell spiking. In the mammalian piriform cortex, fast wave elicited by olfaction falls into the beta band (15−50 Hz) (Zibrowski and Vanderwolf, 1997; Neville

116 and Haberly, 2003), as well as gamma-frequency (35-80 Hz) band presumably related to the sniffing behavior (Freeman, 1960; Bressler and Freeman, 1980; Neville and Haberly,

2003; Litaudon et al., 2008). These oscillations reflect the synchronized activity of ensembles of piriform pyramidal neurons (Freeman, 1978; Litaudon et al., 2008). Besides the contribution to the information coding at the level of the olfactory cortex (Poo and

Isaacson, 2009), the synchronized activity may also enhance the chance for the granule cell to receive a threshold excitatory feedback input. If the synchronized feedback excitation underlies the granule cell spiking, the temporal pattern of the activity in piriform cortical neurons gates the plasticity back in the olfactory bulb.

In addition, granule cell spiking may be elicited by the summation of the proximal and distal input that brings the excitation to the threshold. It is known that heterosynaptic interaction, like the interaction between homosynaptic input, is able to significantly increase the chance of driving a postsynaptic spiking in simple cells in cat visual cortex, though the time course between individual events is important (Usrey et al., 2000). The clusters of action potential found in mitral cells (Shepherd and Greer, 1998) and synchronized cortical pyramidal neuron activity (Freeman, 1978) provide an advantage of driving granule cell spiking via the interaction between proximal and distal input, although the precise control of the postsynaptic spiking may be dependent on the time course and frequency of both input (Usrey et al., 2000).

Modulation of dendrodendritic inhibition by granule cell LTP

The distinct spatial-temporal pattern of the mitral cell spiking in response to the sensory afferent is shaped, in part, by the dendrodendritic reciprocal synapses with

117 granule cells. My study demonstrated that GCL theta bursting stimulation evoked LTP in the granule cell and potentiated of the inhibition in the mitral cell as well. GCL stimulation activates the proximal excitatory inputs on granule cell, which is supported by a facilitated EPSP to the second stimulation; moreover, cell-attachment recording showed GCL theta bursting increased the likelihood of eliciting spiking in granule cell.

With GCL theta bursting stimulation, co-activation of the proximal excitatory input and granule cell induced granule cell LTP. Both action potential back propagation and a depolarization at the distal dendritic site are able to trigger GABA release onto mitral cell.

However, the correlation between granule cell LTP and dendrodendritic inhibition is not clear. It is well known that calcium influx is not only an indicator of neuronal activity, but also tightly coupled to the transmitter release. At reciprocal dendrodendritic synapse, the calcium source for triggering GABA release is complex. Activation of NMDAR may provide sufficient calcium influx for recurrent inhibition (Chen et al., 2000). In addition, the small size of granule cell spine heads provides a structure base of building up calcium concentration solely through NMDARs (Woolf et al., 1991; Chen et al., 2000). With the presence of axonal conduction dendrodendritic inhibition is demonstrated to rely on N- and P/Q type calcium channels (Isaacson and Strowbridge, 1998), which indicate GABA release for lateral inhibition may require the function of voltage-dependent calcium channels. Enhanced inhibition in mitral cell suggests that GCL theta bursting stimulation effectively increases calcium signal at the distal dendrites. However, how an increased calcium influx is triggered by the GCL theta bursting stimulation remains to be elucidated.

118 One attractive explanation is that the potentiated proximal EPSP directly increases the GABA release in EPL. In bulbar granule cells, either somatic action potential or subthreshold depolarization which is presumably > 25 mV from resting potential, evoked by the proximal synaptic event is able to elicit calcium transient to spread along the dendrites up to EPL (Egger et al., 2003). Back propagation of somatic action potentials may activate NMDARs as well as voltage-dependent calcium channels to increase calcium concentration to the threshold of GABA release. Using pharmacological blockade of NMDARs or calcium channels, it may be possible to address the calcium sources. The effect of subthreshold depolarization can be isolated by the application of

TTX. Based on the finding from Chen (2000), subthreshold activation of granule cells may result in NMDARs mediated GABA release (Chen et al., 2000), which should be blocked by APV. The dendrodendritic inhibition mediated by subthreshold activation of granule cell, may provide a mechanism underlying long-term enhancement of mitral cell inhibition induced by GCL theta bursting stimulation, since granule cell spiking in response to odor stimulation is not always observed (Cang and Isaacson, 2003). In this direct regulation model, potentiated proximal EPSPs increase the calcium signal through

NMDARs and/or voltage-dependent calcium channels, which in turn facilitates GABA release in the dendrodendritic synapse.

However, other mechanisms may also be involved in the long-term enhancement of mitral cell inhibition. When the distal dendrodendritic synapse is focally stimulated, the EPSP onto granule cells is depressed by a paired-pulse stimulation (Balu et al., 2007).

In addition, in the extracellular solution with physiological Mg2+ concentration, GCL tetanus that evokes granule cell spiking transiently facilitates GABA release, presumably

119 due to a transient removal of Mg2+ from the NMDARs (Halabisky and Strowbridge,

2003). Both of these results suggest the presence of short-term plasticity in the dendrodendritic reciprocal synapse. Furthermore, my study demonstrates that pairing the mitral cell excitatory input with granule cell activity even induces a long-term depression in the recorded granule cell, thought the underlying mechanism is not known. Therefore the distal synaptic strength has the potential to be enhanced by the granule cell proximal

LTP to mediate long-term enhancement of inhibition. GCL theta bursting stimulation, which potentiates the proximal excitatory input onto granule cells, may increase the spiking in the granule cells. The action potential propagates back to the distal dendritic terminal and temporally removes Mg2+ away from NMDARs. The open NMDARs not only mediate a directly calcium influx, but also provide a prolonged membrane depolarization that allows more voltage-dependent calcium channels to be recruited

(Isaacson and Strowbridge, 1998). In this model, the long-term enhancement of dendrodendritic inhibition is gated through an interaction between the granule cell LTP and distal synaptic short-term plasticity. However, if this is true, a threshold activation that elicits granule cell somatic spiking is required for the potentiated inhibition on mitral cells.

Regulation of the granule cell distal synaptic plasticity

It was demonstrated in my studies that the granule cell proximal LTP was mediated by postsynaptic NMDARs, since LTP was completely diminished by a short duration of bath application of APV. The same protocol used for triggering the granule cell proximal LTP induces LTD at the distal dendrodendritic synapse, which raises an

120 intriguing possibility that the synaptic plasticity on the granule cell could be pathway- specific. This possibility is supported by the structural difference between the proximal and distal dendritic sites. The synaptic machinery on the distal synapse is more complex than that on the proximal site: NMDARs are located on postsynaptic granule cell terminals (Wellis and Kauer, 1993; Petralia et al., 1994; Wellis and Kauer, 1994; Sassoe-

Pognetto and Ottersen, 2000) as well as presynaptic mitral cell terminal (Petralia et al.,

1994; Friedman and Strowbridge, 2000). In many brain regions, including entorhinal cortex (Berretta and Jones, 1996), cerebellum (Glitsch and Marty, 1999; Casado et al.,

2000), neocortex (Sjostrom et al., 2003; Bender et al., 2006; Lien et al., 2006), presynaptically located extrasynaptic NMDARs act as coincident detectors to mediate a timing-dependent LTD. Therefore, at the bulbar dendrodendritic synapse the presynaptic

NMDARs also may detect the mitral cell activity, for example, the first spiking depolarizes the presynaptic terminal to remove Mg2+ away from NMDARs while the later spiking provides a glutamate pool to activate the autoreceptors, or the reversed order.

During the induction of STDP, activation of the presynaptic NMDA receptors may regulate the postsynaptic glutamate sensitivity, in part via a transynaptic messenger. This type of mechanism has been well demonstrated in mammalian cerebellum (Casado et al.,

2000; Casado et al., 2002; Engelman and MacDermott, 2004). Activation of presynaptic

NMDA autoreceptors on the parallel fiber terminal triggers nitric oxide synthesis, which depresses the postsynaptic current in purkinje cells (Casado et al., 2000). The decreased glutamate sensitivity that directly ensues LTD is mediated by the nitric oxide triggered

AMPAR internalization (Ito, 2001; Ito, 2002). This potential mechanism is further supported by the presence of nitric oxide found around the mitral cell layer (Lowe et al.,

121 2008). In addition numerous behavioral studies already suggest that the nitric oxide level is closely related to the olfactory learning and memory (Kendrick et al., 1997; Sakura et al., 2004). All these previous work make the presynaptic NMDAR to be a potential coincidence detector for the distal timing-dependent LTD. However a specific blockade of NMDA autoreceptor is needed to clarify whether the granule cell plasticity is pathway- specific, since the interaction between the proximal and distal excitatory input may also be involved in the distal LTD.

Functional relevance of STDP on the granule cell

As one form of the long-term plasticity, the STDP recorded in the rat olfactory bulb partially explains the cellular basis of the behavior plasticity. The specific behavioral responses to familiar pheromone or newborns somehow suggest the presence of changes in sensory processing based on the established memory. The long-term plasticity I found on both granule cell and mitral cell suggests that gating of the bulbar output relies on the history of mitral cell activity, which supports the general mechanisms for the memory formation. In addition, STDP is also recorded on the synapse between

Kenyon cells, intrinsic neurons in the mushroom bodies, and downstream projection neurons in insects (Cassenaer and Laurent, 2007). Similar phenomenon in different species suggests that coincident activities induced long-term plasticity on the intrinsic neurons represents a general mechanism in the olfactory system underlying the olfactory learning and memory.

Since the STDP depends on the coincident pre- and postsynaptic activities, it may contribute to the synchronized bulbar output. Modeling of the insect mushroom body

122 suggests the STDP can actively preserve the temporal units on the projection neuron for at lease three cycles (Cassenaer and Laurent, 2007). In addition, the oscillatory activity in the mitral cell is known to dependent on the piriform cortical feedback project; blocking this projection results in changes in the bulbar EEG frequency, amplitude and correlation of unit activity (Gray and Skinner, 1988). Therefore, the STDP on the granule cells in the rat olfactory bulb also has the potential to maintain the temporal characteristic in mitral cells. If the spike in the granule cell occurs later than the feedback input, feedback input becomes potentiated to push the granule cell spike to be advanced; otherwise, the feedback input becomes depressed, which will delay granule cell spiking. However, how precisely and actively the STDP can function to the preservation of temporal pattern in mitral cells remains unknown.

Another main question remained here is how the synapse between the olfactory bulb and cortex is precisely organized during the STDP. Most of the olfactory cortical area, except the olfactory turburcle, sends projections back to the olfactory bulb (Haberly,

1998). However there is no quantitative analysis for these projections thus far. Part of the back projections from the anterior piriform cortex end at the granule cell layer and synapse with the proximal dendrites of the granule cells (Balu et al., 2007). In light of the parallel and differential signal transduction from the mitral/tufted cells to the anterior piriform cortical pyramidal neurons (Illig and Haberly, 2003) the specificity of back projection is not indicated, though this is crucial for understanding how the STDP on the granule cell functions to establish an olfactory memory. One possible organization is the cortical neuron that sends projection back to a specific subgroup of granule cells, receive afferent from specific mitral cells; then this specific subgroup of granule cells make

123 dendrodendritic synapses back with these mitral cells (Figure 4-2, top). In this situation, the STDP may function as the feedback control for the excitation level, as well as regulating the synchronization in the cortical neuron. Alternatively, the granule cells receiving the feedback excitation form synapse with a different group of mitral cells, which are not the afferent for the feedback projection neurons. Instead of self-regulation, the STDP changes the activity of the cortical neurons responding to different group of mitral cells (Figure 4-2, bottom). However, more anatomical and electrophysiological experiments are needed to define the synaptic organization involved in the signal transduction between the olfactory bulb and cortex.

Electrical coupling in the extraglomerular regions

Glomerulus is the first station for the odor processing in the brain, where the sensory input is modulated by the intrinsic neuronal network (involving juxtaglomerular cells, mitral and tufted cells) via both chemical and electrical connections. Among all three types of juxtaglomerular cells, the external tufted cells have recently received much attention for their effect on the rhythmic activity within glomeruli, presumably due to their distinct intrinsic properties and electrochemical interaction with principal neurons

(Hayar et al., 2004; Hayar et al., 2004; Hayar et al., 2005; Jan et al., 2009). All previous studies show that the electrical coupling between excitatory neurons in the olfactory bulb occurs within the glomeruli, and formed between the neurons that project into a same glomerulus (Schoppa and Westbrook, 2002; Christie et al., 2005; Pimentel and Margrie,

2008; Jan et al., 2009). By recording from the glomerular layer removed olfactory bulb slice, I discovered GCL stimulation evoked excitatory responses in mitral cells in

124 presence of APV and gabazine. This extraglomerular excitatory input onto the mitral cell appear to be from the lateral dendrites via self- and lateral excitation, both of which are demonstrated to be exclusively mediated by NMDA autoreceptors (Friedman and

Strowbridge, 2000). Though AMPAR is expressed on mitral cell lateral dendritic terminals (Montague and Greer, 1999), it is suggested to be silent by previous electrophysiological studies (Isaacson, 1999; Friedman and Strowbridge, 2000; Salin et al., 2001). Therefore, GCL-evoked excitatory events are not likely to be mediated by the self- or lateral excitation. Unlike a regular calcium impermeable AMPAR that usually has a reversal potential at 0 mV (Jonas et al., 1994), this AMPAR mediated response

(sensitive to NBQX and GYKI 52466) cannot be reversed by changing the holding potential; this suggests that the response is either from a distal dendritic synapse or an electrically coupled neuron. This excitation was also diminished by the bath application of carbenoxolone, which blocks connexin36 regulated gap junction (Schoppa and

Westbrook, 2002). The extraglomerular electrical synapses between bulbar principal cells have never been demonstrated by electrophysiological experiments; however genetic study shows that the connexin36 expression occurs in EPL, MCL and GCL in the mice olfactory bulb (Condorelli et al., 1998; Kosaka et al., 2005), which somehow suggests that gap junction on mitral cells may not be restricted within glomeruli. Moreover, lucifer yellow dye-coupling was shown between closely located up-and-down mitral cell dendrites, as well as the mitral cell and some glomerular layer originated neuronal processes, though this dye-coupling was not observed in all dye-filled imagines. The failure of dye-coupling in some case may be due to the relatively impermeable of higher molecular weight dyes used in other studies (Veenstra, 1996), or the distal location of the

125 gap junction with thin neuronal processes that is shown to happen in the neocortex

(Gibson et al., 1999; Venance et al., 2000). The glomerular layer originated neuronal processes may be the axons of the external tufted cells, which have ramifying axon arbors in EPL, MCL and GCL (Figure 4-3) (Macrides and Schneider, 1982; Antal et al., 2006).

In addition, one of the two subpopulations of the external tufted cells possesses big basal dendritic arbors in EPL (Antal et al., 2006), which also are possible to be dye-filled through gap junction. However, we never observed dye-filled somata of the external tufted cells, which are likely to have been removed together with the glomerular layer in our mini brain slice. Charpak (2009) showed that the electrical coupling between the external tufted and mitral cells occurred in the glomeruli, however in that study the possibility of gap junction outside of the glomerulus was not excluded. Moreover, the external tufted cells are known to be heterogeneous with distinct morphology, intrinsic properties and presumably different functions (Antal et al., 2006). Different type of the external tufted cell may possess different forms of electrical coupling with mitral cells, which has not been addressed in the previous work.

In olfactory bulb mini slices that do not contain glomerular layer structure, the

GCL stimulation that is close to the mitral cell axon bundles may drive antidromic action potentials in mitral cells. In addition, the evoked excitation is dominated by AMPARs.

Together with the previous study that shows the presence of AMPA autoreceptor on the external tufted cell primary dendrites (Ma and Lowe, 2007), glutamate spillover from activated mitral cells are likely to activate AMPA autoreceptors possibly located on the external tufted cell basal dendrites. This AMPA autoreceptor mediated depolarization in the external tufted cell dendritic terminals results in the excitation in the electrically

126 coupled mitral cell. Blockade of axonal conductance with TTX prevents the glutamate release triggered by antidromic action potentials, which also diminishes of EPSPs in the mitral cells. The slow kinetics of the GCL-evoked mitral cell EPSP, like those found in the cerebellar purkinje cells (Barbour et al., 1994), may indicate a slow clearance of the glutamate spillover, in part due to multiple releasing events (Zhou and Hablitz, 1997); and it also can be due to the distal electrical signal along the dendrites (Zhou and Hablitz,

1997). Another plausible hypothesis is that glutamate spillover from mitral cells may act on the external tufted cell axon terminals. Both subpopulations of the external tufted cells may have axon projection in EPL, MCL and even GCL (Pimentel and Margrie, 2008); therefore direct activation of the external tufted cell axon by the GCL stimulation is also possible.

The dendrodendritic coupling is believed to be the main type of electrical connection in different brain regions (MacVicar and Dudek, 1980; MacVicar and Dudek,

1980; Schoppa and Westbrook, 2002; Christie et al., 2005); however, our two-photon images showed the dye-coupling on or nearby the mitral cell somata. Compared to the electrical coupling on the distal primary dendrites within glomeruli, the somatic located or proximally located gap junction are able to mediate neighboring mitral cell synchrony in a more precise window (Hjorth et al., 2007). Action potentials in a neuron are elicited by the summation of synaptic inputs. To make an accurate response to stimulation, neurons usually use a narrow integration window for the detection of afferent to avoid integrating random input (Shadlen and Movshon, 1999). Therefore, the mitral cell output synchrony in a narrow integration window, which effectively promotes the summation of

EPSPs to drive an action potential in response to specific odor stimulation, is important

127 for the information processing at the level of the olfactory cortex. On the other hand, unlike mitral and tufted cells, some external tufted cells project their apical dendrites into two glomeruli (Hayar et al., 2004; Antal et al., 2006). The widely expanded basal dendrites and axon branches further increase the chance of electrical coupling with the mitral cells that project to the neighboring glomeruli. Therefore gap junctions on the external tufted cells, especially in the extraglomerular regions, is likely to mediate the synchrony among the neighboring glomeruli.

Impact of proximal and distal inhibitory input onto mitral cells

In physiological condition, a single action potential poorly propagate into the mitral cell lateral dendrites (Margrie et al., 2001; Christie and Westbrook, 2003), however clusters of spiking may make the back propagation through a long distance to induce glutamate release (Margrie et al., 2001). When a threshold glutamate release occurs at the distal dendrodendritic synapse, the postsynaptic GABA release triggers a recurrent inhibition on the presynaptic mitral cell. The mapping of mitral cell synaptic response showed that very far off-beam stimulation could evoke an inhibitory postsynaptic potential in the recorded mitral cell. The ability of detecting very distal input allows a single mitral cell to recruit synaptic signal from a large amount of granule cells.

The communication with granule cells not only provides a mitral cell an indirect regulation from cortical feedback excitation (Balu et al., 2007) but also increases the range of the lateral inhibition (Shepherd and Greer, 1998), both of which are crucial for initiating and sharpening a mitral cell firing pattern.

128 In addition, the dynamic properties of IPSP that due to the distribution of the synaptic input (Rall, 1967), such as the time to peak and decay of IPSPs, may significantly affect the odor response in a mitral cell. It has been shown that in electrical fish that proximal and distal inhibitory input on the lateral line lobe pyramidal cell act as low-pass and high-pass filter to shape its sensory response via the different IPSP dynamics (Berman and Maler, 1998). When the integrating neuron also receives

NMDAR mediated excitatory input, the functional difference of the proximal and distal inhibition can be amplified, because the activation of NMDAR requires a precise time window of membrane depolarization and the presence of a glutamate pool (Berman and

Maler, 1998). Mitral cells receive sensory input via postsynaptic AMPARs and

NMDARs within glomeruli, and NMDA autoreceptor mediated self-/lateral excitation in the EPL. Therefore, the proximal and distal inhibitory input onto the lateral dendrites may act as different sensory filters which precisely control the response of a mitral cell to the combined excitatory signal from olfactory sensory neurons and NMDA autoreceptors.

Significance and future direction

Despite of the extensive studies in the olfaction related behavioral plasticity, my work the first time demonstrated the presence of the long-term plasticity in mammalian main olfactory bulb. This long-term plasticity, to some extent, explains the function of the olfactory bulb in the olfactory memory formation and olfactory learning. The spike- timing dependent property of this plasticity suggests an important strategy of the information coding within the olfactory bulb, which integrates both feedforward projection from periphery sensory neurons and feedback input from the higher brain

129 regions. The involvement of anterior piriform cortical feedback projection in bulbar long- term plasticity also implicates one aspect of information processing in the olfactory cortex; however the understanding at the olfactory cortex level is limited thus far. Unlike the long-term plasticity identified in a variety of brain area, the bulbar long-term plasticity involves synapses onto intrinsic interneurons. Together with the finding of intrinsic neuron STDP in insect mushroom bodies, the long-term plasticity on the inhibitory neurons induced by coincident activities over two separate olfactory structures is likely to be a general principle of olfactory learning and memory. However, a recent study in the moth Manduca sexta find that the STDP in Kenyon cells alone cannot underlie olfactory learning (Ito et al., 2008). Based on many common features between vertebrate and insect olfactory system, other mechanisms other than granule cell STDP may also be responsible for mammalian olfactory learning and memory. For instance, the structural plasticity from the neurongenesis of granule cells (Altman, 1969) may make the regulation of olfactory learning and memory more complex. How the newborn granule cells are wired in the existing network and maintain the memory is not fully understood. Recent study in the newborn granule cells shows the presence of non-

NMDAR mediated long-term plasticity induced by GCL theta bursting stimulation; however the same stimulation protocol fails to induce long-term plasticity in adult granule cells (Nissant et al., 2009). Some findings in Nissant (2009) are controversial with my work: first, I found a significant long-term potentiation of granule cell EPSPs (5 of 5 tested cells) or action currents in cell-attached recording (7 of 9 tested cells) with

GCL theta bursting stimulation in my experiments. Dye-filled granule cells showed regular size of dendritic arbors reaching the EPL; second, timing-dependent long-term

130 potentiation on granule cells was demonstrated to be relied on NMDARs in my work.

The failure of finding long-term plasticity in adult granule cell may be due to the activation of non-specific synaptic pathways. Besides excitatory input from piriform cortex, granule cell also receive inhibition from local interneuron (Presseler and

Strowbridge, 2006). On the other hand, I cannot exclude the possibility that the granule cell tested in my experiments were in a critical developing period, since more than 30,000 adult-born granule cells emerge in the olfactory bulb every day; therefore genetic tools appear to be needed to facilitate my work to clarify the controversy. For the second aspect of the controversy, besides the possibility that different mechanisms may be recruited by different types of plasticity, newborn granule cells are also likely to use different mechanisms than adult ones. Calcium permeable AMPARs are found to be expressed in interneurons in neocortex (Jonas et al., 1994), dentate gyrus basket cells

(Koh et al., 1995), and cultured bulbar intrinsic neurons as well (Blakemore et al., 2005).

The calcium source for newborn granule cell long-term potentiation can be speculated to be from AMAPRs. Despite of the controversial conclusions that are discussed above,

Nissant (2009) raises a possibility that newborn granule cells are more sensitive to plasticity. This suggestion may represent a mechanism of the involvement of newborn granule cell in olfactory learning (Magavi et al., 2005). Demonstration of the presence of the long-term plasticity in the olfactory bulb is the first step of understanding olfactory learning and memory. Discovering the cellular mechanisms underlying this timing- dependent plasticity will be very important for understanding how bulbar local circuits are organized to process sensory information. However, little is known of the cellular basis of granule cell long-term plasticity besides the function of postsynaptic NMDARs

131 on the proximal dendrites. The complex synaptic machinery on the distal dendrodendritic synapse and the convergence of distinct synaptic input on the granule cells make both pathway-specific plasticity and interaction induced different plasticity possible.

Differentiation from these two possibilities will be an interesting topic for the future studies. In addition, how the cortical neurons are fitted in the big picture of feedback control of odor response is interesting. Finding out whether the cortical feedback projection is specific on a subtype of granule cells or mitral cells will be helpful for understanding of the strategy of sensory processing in the olfactory cortex.

Besides discovering a principle rule of how the olfactory behavioral plasticity is mediated by the local circuit in the olfactory bulb, I also studied the specific mechanisms underlying the signal gating on mitral cells. I provided the direct evidence showing the presence of gap junction on mitral cells outside of the glomeurli, which was thought to be the only site of electrical coupling between bulbar principal neurons in the olfactory bulb thus far. The newly identified gap junction can be located on the mitral cell soma or lateral dendrite, which provides a novel mechanism of the regulation of odor representation in mitral cells. Compared to the gap junction on the distal site on the apical dendrites, the gap junction outside of glomeruli may provide a more precise window for mitral cell synchrony. However, the functional consequence of the electrical coupling outside of the glomeruli remains to be studied. Understanding how the newly identified gap junction affects the mitral cell distinct spatial-temporal spiking pattern in response to specific odor stimulation as well as the bulbar synchrony needs further investigation. The functional difference between proximal and distal recurrent inhibition is also an open question, discovering which is important for understanding the information coding within

132 the olfactory bulb. In addition, the physiological function of AMPA autoreceptors on the mitral cell lateral dendrites remains to be investigated.

133 Figure 4-1. Schematic representation of olfaction perception

Olfaction perception begins from the olfactory receptors located in the olfactory epithelium. Chemical information is translated into electrical signals via dynamic modulation from local synaptic interactions in the olfactory bulb. Bulbar output possesses distinct spatial and temporal patterns as well as synchrony across groups of mitral cells.

Bulbar response is then transmitted to the olfactory cortex for further information coding to achieve odor recognition. From (Lledo et al., 2005)

134

Figure 4-1

135 Figure 4-2. Schematic diagram of proposed synaptic connections between the olfactory cortex and olfactory bulb

Top, the cortical neuron receives afferent from the mitral cell (black) that make dendrodendritic synapses with a specific subgroup of granule cells (blue), which receive feedback projection coming from a same cortical neuron. Bottom, the cortical neuron receives afferent from the mitral cell (black), and sends feedback to the granule cells

(blue) which does not synapse with the black mitral cell; instead, they make synapse with the other mitral cell (green).

136

Figure 4-2

137 Figure 4-3. Two subpopulation of the external tufted cells

Two-dimensional projections of 3-D reconstructed external tufted cells. Somata and dendrites are in blue, axon arbors are in red. Some external tufted cells are able to project ramifying axon arbors into EPL, MCL and GCL. (A) External tufted cells without basal dendrites. (B) External tufted cells with basal dendrites. From (Antal et al., 2006)

138

Figure 4-3

139 Bibliography

Allison, A. C. (1953). "The morphology of the olfactory system in the vertebrates." Biol Rev 28: 195-244.

Altman, J. (1969). "Autoradiographic and histological studies of postnatal neurogenesis. IV. Cell proliferation and migration in the anterior forebrain, with special reference to persisting neurogenesis in the olfactory bulb." J Comp Neurol 137: 433-457.

Antal, M., M. Eyre, et al. (2006). "External tufted cells in the main olfactory bulb form two distinct subpopulations." Eur J Neurosci 24: 1124-1136.

Araneda, R. C., A. D. Kini, et al. (2000). "The molecular receptive range of an odorant receptor." Nat Neurosci 3: 1248-1255.

Aroniadou-Anderijaska, V., F. M. Zhou, et al. (2000). "Tonic and synaptically evoked presynaptic inhibition of sensory input to the rat olfactory bulb via GABAB heteroreceptors." J Neurophysiol 84: 1194-1203.

Aungst, J. L., P. M. Heyway, et al. (2003). "Centre-surround inhibition among olfactory bulb glomeruli." Nature 426: 623-629.

Axel, R. (1995). "The molecular logic of smell." Scientific American: 154-159.

Baldwin, B. A. and E. E. Shillito (1974). "The effects of ablation of the olfactory bulbs on parturition and maternal behaviour in Soay sheep." Animal Behaviour 22: 220- 223.

Balu, R., P. Larimer, et al. (2004). "Phasic stimulation evoke precisely timed spikes in intermittently discharging mitral cells." J Neurophysiol 92: 743-753.

Balu, R., R. T. Presseler, et al. (2007). "Multiple modes of synaptic excitation of olfactory bulb granule cells." J Neurosci 27: 5621-5632.

Balu, R. and B. W. Strowbridge (2007). "Opposing inward and outward conductances regulate rebound discharges in olfactory mitral cells." J Neurophysiol 97: 1959- 1968.

Barbour, B. and M. Hausser (1997). "Intersynaptic diffusion of neurotransmitter." Trends Neurosci 20: 377-384.

Barbour, B., B. U. Keller, et al. (1994). "Prolonged presence of glutamate during excitatory synaptic transmission to cerebellar purkinje cells." Neuron 12: 1331- 1343.

140 Bell, C. C., V. Z. Han, et al. (1997). "Synaptic plasticity in a cerebellum-like structure depends on temporal order." Nature 387: 278-281.

Belluardo, N., G. Mudo, et al. (2000). "Expression of connexin36 in the adult and developing rat brain." Brain Res 865: 121-138.

Bender, V. A., K. J. Bender, et al. (2006). "Two coincidence detectors for spiking timing- dependent plasticity in somatosensory cortex." J Neruosci 26: 4166-4177.

Berman, N. J. and L. Maler (1998). "Distal versus proximal inhibitory shaping of feedback excitation in the electrosensory lateral line lobe: implications for sensory filtering." J Neurophysiol 80: 3214-3232.

Berretta, N. and R. S. Jones (1996). "Tonic facilitation of glutamate release by presynaptic N-methyl--aspartate autoreceptors in the entorhinal cortex." Neuroscience 75: 339-344.

Blakemore, L. J., M. Resasco, et al. (2005). "Evidence for Ca2+-permeable AMPA receptors in the olfactory bulb." AM J Physiol Cell Physiol 290: C925-C935.

Bliss, T. P. and T. Lomo (1973). "Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path." J Physiol. 232: 331-356.

Bozza, T. C. and P. Mombaerts (2001). "Olfactory coding: Revealing intrinsic representations of odors." Curr Biol 11: 687-690.

Brennan, P. A., H. Kaba, et al. (1990). "Olfactory recognition: a simple memory system." Science 250: 1223-1226.

Brennan, P. A. and E. B. Keverne (1997). "Neural mechanisms of mammalian olfactory learning." Prog Neurobiol 51: 457-481.

Bressler, S. L. and W. J. Freeman (1980). "Frequency analysis of olfactory system EEG in cat, rabbit, and rat " Electroencephalogr Clin Neurophysiol 50: 21-24.

Bruce, H. M. (1959). "An exteroceptive block to pregnancy in the mouse." Nature 184.

Bruce, H. M. (1960). "A block to pregnancy in the mouse caused by proximity of strange males." J Reproductive Fertility 1: 96-103.

Bruce, H. M. and D. V. Parrott (1960). "Role of olfactory sense in pregnancy-block by strange males." Science 131: 1526.

Buck, L. and R. Axel (1991). "A novel multigene family may encode odorant receptors: a molecular basis for odor recognition." Cell 65: 175-187.

141

Buck, L. B. (1996). "Information coding in the vertebrate olfactory system." Annual Reviews 19: 517-544.

Buzsaki, G. and J. J. Chrobak (1995). "Temporal structure in spatially organized neuronal ensembles: a role for interneuronal networks." Curr Opin Neurobiol 5: 504-510.

Cajal, S. R. (1955). Studies on the cerebral cortex (limbic structures) (trans. by Kraft, L.M.). London, Lloyd-Luke Medical Books.

Cang, J. and J. S. Isaacson (2003). "In Vivo whole-cell recording of odor-evoked synaptic transmission in the rat olfactory bulb." J Neurosci 23: 4108-4116.

Carlson, G. C., M. T. Shipley, et al. (2000). "Long-lasting depolarizations in mitral cells of the rat olfactory bulb." J Neurosci 20: 2011-2021.

Casado, M., S. Dieudonne, et al. (2000). "Presynaptic N-methyl-D-aspartate receptors at the parallel fiber-purkinje cell synapse." Proc Natl Acad Sci USA 97: 11593- 11597.

Casado, M., P. Isope, et al. (2002). "Involvement of presynaptic N-methyl-D-aspartate receptors in cerebellar long-term depression." Neuron 33: 123-130.

Cassenaer, S. and G. Laurent (2007). "Hebbian STDP in mushroom bodies facilitates the synchronous flow of olfactory information in locusts." Nature 448: 709-713.

Chen, W. R. and G. M. Shepherd (1997). "Membrane and synaptic properties of mitral cells in slices of rat olfactory bulb." Brain Res 745: 189-196.

Chen, W. R., W. Xiong, et al. (2000). "Analysis of relations between NMDA receptors and GABA release at olfactory bulb reciprocal synapses." Neuron 25: 625-633.

Chepkova, A. N., O. A. Sergeeva, et al. (2008). "Carbenoxolone impairs LTP and blocks NMDA receptors in murine hippocampus." Neuropharmacology 55: 139-147.

Chess, A., I. Simon, et al. (1994). "Allelic inactivation regulates olfactory receptor gene expression." Cell 78: 823-834.

Christie, J. M., C. Bark, et al. (2005). "Connexin36 mediates spike synchrony in olfactory bulb glomeruli." Neuron 46: 761-772.

Christie, J. M. and G. L. Westbrook (2003). "Regulation of backpropagating action potentials in mitral cell lateral dendrites by A-type potassium currents." J Neurophysiol 89: 2466-2472.

142 Christie, J. M. and G. L. Westbrook (2006). "Lateral excitation within the olfactory bulb." J Neurosci 26: 2269-2277.

Cobb, S. R., E. H. Buhl, et al. (1995). "Synchronization of neuronal activity in hippocampus by individual GABAergic interneurons." Nature 378: 75-78.

Condorelli, D. F., R. Parenti, et al. (1998). "Short communication cloning of a new gap junction gene (Cx36) highly expressed in mammalian brain neurons." European Neurosci 10: 1202-1208.

Connors, B. W. and M. J. Gutnick (1990). "Intrinsic firing patterns of diverse neocortical neurons." Trends Neurosci 13: 99-104. de Olmos, J., H. Hardy, et al. (1978). "The afferent connections of the main and the accessory olfactory bulb formations in the rat: an experimental HRP-study." J Comp Neurol 181: 213-244.

Deans, M. R., J. R. Gibson, et al. (2001). "Synchronous activity of inhibitory networks in neocortex requires electrical synapses containing connexin36." Neuron 31: 477- 485.

Demas, G. E., J. M. Williams, et al. (1999). "Amygdala but not Hippocampal Lesion Impair Olfactory Memory for Mate in Prairie Voles (Microtus Ochrogaster)." Am J Physiol Regulatory Integrative Comp Physiol 273: 1683-1689.

Desmaisons, D., J.-D. Vincent, et al. (1999). "Control of action potential timing by intrinsic subthreshold oscillations in olfactory bulb output neurons." J Neurosci 19: 10727-10737.

Dietz, S. B. and V. N. Murthy (2005). "Contrasting short-term plasticity at two sides of the mitral-granule reciprocal synapse in the mammalian olfactory bulb." J Physiol 569: 475-488.

Dohlman, H. G. (1991). "Model systems for the study of seven-transmembrane-segment receptors." Ann Rev Biochem 60: 653-688.

Egger, V., K. Svoboda, et al. (2003). "Mechanisms of lateral inhibition in the olfactory bulb: efficiency and modulation of spike-evoked calcium influx into granule cells." J Neurosci 23: 7551-7558.

Engelman, H. S. and A. B. MacDermott (2004). "Presynaptic ionotropic receptors and control of transmitter release." Nature Reviews 5: 135-145.

Ennis, M., F. M. Zhou, et al. (2001). "Dopamine D2 receptor-mediated presynaptic inhibition of olfactory nerve terminals." J Neurophysiol 86: 2986-2997.

143 Fantana, A. L., E. R. Soucy, et al. (2008). "Rat olfactory bulb mitral cells receive sparse glomerular Inputs." Neuron 59: 802-814.

Ferreira, A., A. Carrau, et al. (1992). "Diazepam facilitates acceptance of alien lambs by postparturient ewes." Physiol Behav 51: 1117-1121.

Freeman, W. J. (1960). "Correlation of electrical activity of prepyriform cortex and behaviour in cat." J Neurophysiol 23: 111-131.

Freeman, W. J. (1978). "Spatial properties of an EEG event in the olfactory bulb and cortex." Electroencephalogr Clin Neurophysiol 44: 586-605.

Fricker, D. and R. Miles (2001). "Interneurons, spike timing, and perception." Neuron 32: 771-774.

Friedman, D. and B. W. Strowbridge (2000). "Functional role of NMDA autoreceptors in olfactory mitral cells." J Physiol 84: 39-50.

Friedman, D. and B. W. Strowbridge (2003). "Both electrical and chemical synapses mediate fast network oscillations in the olfactory bulb " J Neurophysiol 89: 2601- 2610.

Froemke, R. C., M. M. Poo, et al. (2005). "Spike-timing-dependent synaptic plasticity depends on dendritic location." Nature 434: 221-225.

Galarreta, M. and S. Hestrin (1999). "A network of fast-spiking cells in the neocortex connected by electrical synapses." Nature 402: 72-75.

Gall, C. M., S. H. Hendry, et al. (1987). "Evidence for coexistence of GABA and dopamine in neurons of the rat olfactory bulb." J Comp Neurol 266: 307-318.

Gao, Y. and B. W. Strowbridge (2009). "Long-term plasticity of excitatory inputs to granule cells in the rat olfactory bulb." Nat Neurosci 12: 731-733.

Gibson, J. R., M. Beierlein, et al. (1999). "Two networks of electrically coupled inhibitory neurons in neocortex." Nature 402: 75-79.

Glitsch, M. and A. Marty (1999). "Presynaptic effects of NMDA in cerebellar purkinjie cells and interneurons." J Neurosci 19: 511-519.

Gray, C. M. and J. E. Skinner (1988). "Centrifugal regulation of neuronal activity in the olfactory bulb of the walking rabbit as revealed by reversible cryogenic blockade." Exp Brain Res 69: 378-86.

144 Gustafsson, B., H. Wigstrom, et al. (1987). "Long-term potentiation in the hippocampus using depolarizing current pulse as the conditioning stimulus to single volley synaptic potentials." J Neruosci 7: 774-780.

Guthrie, K. M., A. J. Anderson, et al. (1993). "Odor-induced increases in c-fos mRNA expression reveal an anatomical "unit" for odor processing in olfactory bulb." Proc Natl Acad Sci USA 90: 3329-3333.

Haberly, L. B. (1998). The olfactory cortex. The synaptic organization of the brain. G. M. Shepherd. New York, Oxford Univ. Press: 377-416.

Haberly, L. B. (2001). "Parallel-distributed processing in olfactory cortex: new insights from morphological and physiological analysis of neuronal circuitry." Chem Senses 26: 551-576.

Haberly, L. B. and J. L. Price (1978). "Association and commissural fiber systems of the olfactory cortex of the rat." J Comp Neurol 178: 711-740.

Halabisky, B., D. Friedman, et al. (2000). "Calcium influx through NMDA receptors directly evokes GABA release in olfactory bulb granule cells." J Neurosci 20: 5124-5134.

Halabisky, B. and B. W. Strowbridge (2003). "γ-Frequency oscillation excitatory input to granule cells facilitates dendrodendritic inhibition in the rat olfactory bulb." J Neurophysiol 90: 644-654.

Hamilton, K. A., T. Heinbockel, et al. (2005). "Properties of external plexiform layer interneurons in mouse olfactory bulb slices." Neuroscience 133: 819-829.

Hamilton, K. A. and J. S. Kauer (1988). "Responses of mitral/tufted cells to orthodromic and antidromic electrical stimulation in the olfactory bulb of the tiger salamander." J Neurophysiol 59: 1736-1755.

Hayar, A., S. Karnup, et al. (2004). "External tufted cells: a major excitatory element that coordinates glomerular activity." J Neurosci 24: 6676-6685.

Hayar, A., S. Karnup, et al. (2004). "Olfactory bulb glomeruli: external tufted cells intrinsically burst at theta frequency and are entrained by patterned olfactory input." J Neurosci 24: 1190-1199.

Hayar, A., M. T. Shipley, et al. (2005). "Olfactory bulb external tufted cells are synchronized by multiple intraglomerular mechanisms." J Neurosci 25: 8197- 8208.

Herberholz, J., B. L. Antonsen, et al. (2002). "A lateral excitatory network in the escape circuit of crayfish." J Neurosci 22: 9078-9085.

145

Hjorth, J., A. H. Elias, et al. (2007). "The significance of gap junction location in striatal fast spiking interneurons." Neurocomputing 70: 1887-1891.

Hormuzdi, S. G., I. Pais, et al. (2001). "Impaired electrical signaling disrupts gamma frequency oscillations in connexin 36-deficient mice." Neuron 31: 487-495.

Illig, K. R. and L. B. Haberly (2003). "Odor-evoked activity is spatially distributed in piriform cortex." J Comp Neurol 457: 361-373.

Isaacson, J. S. (1999). "Glutamate spillover mediates excitatory transmission in the rat olfactory bulb." Neuron 23: 377-384.

Isaacson, J. S. (2001). "Mechanisms governing dendritic gamma-aminobutyric acid (GABA) release in the rat olfactory bulb." Proc Natl Acad Sci USA 98: 337-342.

Isaacson, J. S. and B. W. Strowbridge (1998). "Olfactory reciprocal synapses: dendritic signaling in the CNS." Neuron 20: 749-761.

Ito, I., R. C. Ong, et al. (2008). "Sparse odor representation and olfactory learning." Nat Neurosci 11: 1177-1184.

Ito, M. (2001). "Cerebellar long-term depression: characterization, signal transduction, and functional roles." Physiol Rev 81: 1143-1195.

Ito, M. (2002). "The molecular organization of cerebellar long-term depression." Nat Rev Neurosci 3: 896-902.

Jahr, C. E. and R. A. Nicoll (1980). "Dendrodendritic inhibition: demonstration with intracellular recording." Science 207: 1473-1475.

Jahr, C. E. and R. A. Nicoll (1982). "Noradrenergic modulation of dendrodendritic inhibition in the olfactory bulb." Nature 297: 227-229.

Jan, D., D. Hirnet, et al. (2009). "External tufted cells drive the output of olfactory bulb glomeruli." J Neurosci 29: 2043-2052.

Jonas, P., C. Racca, et al. (1994). "Differences in calcium permeability of AMPA-type glutamate receptor channels in neocortical neurons caused by differential GluR-B subunit expression." Neuron 12: 1281-1289.

Kendrick, K. M., R. Guevara-Guzman, et al. (1997). "Formation of olfactory memories mediated by nitric oxide." Nature 388: 670-674.

Kendrick, K. M., F. Levy, et al. (1992). "Changes in the sensory processing of olfactory signals induced by birth in sheep." Science 256: 833-836.

146

Kishi, K. (1987). "Golgi studies on the development of granule cells of the rat olfactory bulb with reference to migration in the subependymal layer." J Comp Neurol 258: 112-124.

Koh, D. S., J. R. Geiger, et al. (1995). "Ca2+-permeable AMPA and NMDA receptor channels in basket cells of rat hippocampal dentate gyrus." J Physiol 485: 383-402.

Kohling, R., S. J. Gladwell, et al. (2001). "Prolonged epileptiform bursting induced by 0- Mg(2+) in rat hippocampal slices depends on gap junctional coupling." Neuroscience 105: 579-587.

Korsching, S. I., S. Argo, et al. (1997). "Olfaction in zebrafish: what does a tiny teleost tell us?" Semin Cell Dev Biol 8: 181-187.

Kosaka, K., C. W. Heizmann, et al. (1994). "Calcium-binding protein parvalbumin- immunoreactive neurons in the rat olfactory bulb. 1. Distribution and structural features in adult rat." Exp Brain Res 99: 191-204.

Kosaka, T., M. R. Deans, et al. (2005). "Neuronal gap junctions in the mouse main olfactory bulb: Morphological analyses on transgenic mice " Neuroscience 134: 757-769.

Kosaka, T., Y. Hataguchi, et al. (1985). "Coexistence of immunoreactivities for glutamate decarboxylase and tyrosine hydroxylase in some neurons in the periglomerular region of the rat main olfactory bulb: possible coexistence of gamma-aminobutyric acid (GABA) and dopamine." Brain Res 343: 166-171.

Kullmann, D. M. and F. Asztely (1998). "Extrasynaptic glutamate spillover in the hippocampus: evidence and implications." Trends Neurosci 21: 8-14.

Lagier, S., A. Carleton, et al. (2004). "Interplay between local GABAergic interneurons and relay neurons generates gamma oscillations in the rat olfactory bulb." J Neurosci 24: 4382-4392.

Landis, D. M. and T. S. Reese (1983). "Cytoplasmic organization in cerebellar dendritic spines." J Cell Biol 97: 1169-1178.

Landisman, C. E., M. A. Long, et al. (2002). "Electrical synapses in the thalamic reticular nucleus." J Neurosci 22: 1002-1009.

Laurent, G. (2002). "Olfactory network dynamics and the coding of multidimensional signals." Nat Rev Neurosci 3: 884-895.

Laurent, G., K. MacLeod, et al. (1996). "Dynamical representation of odors by oscillating and evolving neural assemblies." Trends Neurosci 19: 489-496.

147

Laurent, G., M. Wehr, et al. (1996). "Temporal representations of odors in an olfactory network." J Neurosci 16: 3837-3847.

Lester, R. J., J. D. Clements, et al. (1990). "Channel kinetics determine the time course of NMDA receptor-mediated synaptic currents." Nature 346: 565-567.

Levy, F., R. Gervais, et al. (1990). "Importance of beta-noradrenergic receptors in the olfactory bulb of sheep for recognition of lambs." Behav Neurosci 104: 464-469.

Levy, F., M. Keller, et al. (2004). "Olfactory regulation of maternal behavior in mammals." Hormones and Behavior 46: 284-302.

Levy, F., A. Locatelli, et al. (1995). "Involvement of the main but not the accessory olfactory system in maternal behavior of primiparous and multiparous ewes." Physiology & Behavior 57: 97-104.

Levy, W. B. and O. Steward (1983). "Temporal contiguity requirements for long-term associative potentiation / depression in the hippocampus." Neuroscience 8: 791- 797.

Lien, C. C., Y. Mu, et al. (2006). "Visual stimuli-induced LTD of GABAergic synapses mediated by presynaptic NMDA receptors." Nat Neurosci 9: 372-380.

Linster, C. and M. Hasselmo (1997). "Modulation of inhibition in a model of olfactory bulb reduces overlap in the neural representation of olfactory stimuli." Behav Brain Res 84: 117-127.

Litaudon, P., S. Garcia, et al. (2008). "Strong coupling between pyramidal cell activity and network oscillations in the olfactory cortex." Neuroscience 156: 781-787.

Lledo, P.-M., G. Gheusi, et al. (2005). "Information processing in the mammalian olfactory system." Physiol Rev 85: 281-317.

Llinas, R. R. (1988). "The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function." Science 242: 1654-1664.

Lowe, G., D. G. Buerk, et al. (2008). "Tonic and stimulus-evoked nitric oxide production in the mouse olfactory bulb." Neuroscience 153: 842-850.

Luo, M. and L. C. Katz (2001). "Response correlation maps of neurons in the mammalian olfactory bulb." Neuron 32: 1165-1179.

Ma, J. and G. Lowe (2007). "Calcium permeable AMPA receptors and autoreceptors in external tufted cells of rat olfactory bulb." Neuroscience 144: 1094-1108.

148 MacLeod, K. and G. Laurent (1996). "Distinct mechanisms for synchronization and temporal patterning of odor-encoding neural assemblies." Science 274: 976-979.

Macrides, F. and S. P. Schneider (1982). "Laminar organization of mitral and tufted cells in the main olfactory bulb of the adult hamster." J Comp Neurol 208: 419-430.

MacVicar, B. A. and F. E. Dudek (1980). "Dye-coupling between CA3 pyramidal cells in slices of rat hippocampus." Brain Res 196: 494-497.

MacVicar, B. A. and F. E. Dudek (1980). "Local synaptic circuits in rat hippocampus: interactions between pyramidal cells." Brain Res 184: 220-223.

Magavi, S. S., B. D. Mitchell, et al. (2005). "Adult-born and preexisting olfactory granule neurons undergo distinct experience-dependent modifications of their olfactory responses in vivo." J Neurosci 25: 10729-10739.

Magee, J. C. and D. Johnston (1997). "A synaptically controlled, associative signal for Hebbian plasticity in hippocampal neurons." Science 275: 209-213.

Maher, B. J. and G. L. Westbrook (2008). "Co-transmission of dopamine and GABA in periglomerular cells." J Neurophysiol 99: 1559-1564.

Malnic, B., J. Hirono, et al. (1999). "Combinatorial receptor codes for odors." Cell 96: 713-723.

Margrie, T. W., B. Sakmann, et al. (2001). "Action potential propagation in mitral cell lateral dendrites is decremental and controls recurrent and lateral inhibition in the mammalian olfactory bulb." Proc Natl Acad Sci USA 98: 319-324.

Markram, H., J. Lubke, et al. (1997). "Regulation of synaptic efficacy by coincidence of postsynaptic APs and EPSPs." Science 275: 213-215.

Miragall, F., E. Simburger, et al. (1996). "Mitral and tufted cells of the mouse olfactory bulb possess gap junctions and express connexin43 mRNA." Neurosci Lett 216: 199-202.

Montague, A. A. and C. A. Greer (1999). "Differential distribution of ionotropic glutamate receptor subunits in the rat olfactory bulb." J Comp Neurol 405: 233- 246.

Mori, K. (1987). "Membrane and synaptic properties of identified neurons in the olfactory bulb." Prog Neurobiol 29: 275-320.

Mori, K., H. V. Campenhausen, et al. (2000). "Zonal organization of the mammalian main and accessory olfactory system." Phil Trans R Soc Lond B 355: 1801-1812.

149 Mori, K., H. Nagao, et al. (1999). "The olfactory bulb: coding and processing of odor molecule information." Science 286: 711-715.

Mori, K. and S. F. Takagi (1977). Inhibition in the olfactory bulb: dendrodendritic interactions and their relation to the induced waves. Food intake and chemical senses. Food intake and chemical senses. Satoh M., Y. Katsuki, S. F. Takagi and Y. Oomura. Tokyo, University of Tokyo Press: 33-43.

Murphy, G. J., D. P. Darcy, et al. (2005). "Intraglomerular inhibition: signaling mechanisms of an olfactory microcircuit." Nat Neurosci 8: 354-364.

Murphy, G. J., L. L. Glickfeld, et al. (2004). "Sensory neuron signaling to the brain: properties of transmitter release from olfactory nerve terminals." J Neurosci 24: 3023-3030.

Neville, K. R. and L. B. Haberly (2003). "Beta and gamma oscillations in the olfactory system of the urethane-anesthetized rat." J Neurophysiol 90: 3921-3930.

Nicoll, R. A. and C. E. Jahr (1982). "Self-excitation of olfactory neurons." Nature 296: 441-444.

Nissant, A., C. Bardy, et al. (2009). "Adult neurogenesis promotes synaptic plasticity in the olfactory bulb." Nat Neurosci 12: 728-730.

Nusser, Z., L. M. Kay, et al. (2001). "Disruption of GABAA receptors on GABAergic interneurons leads to increased oscillatory power in the olfactory bulb network." J Neurophysiol 86: 2823-2833.

Petralia, R. S., N. Yokotani, et al. (1994). "Light and electron microscope distribution of the NMDA receptor subunit NMDAR1 in the rat nervous system using a selective anti-peptide antibody." J Neurosci 14: 667-696.

Pimentel, D. O. and T. W. Margrie (2008). "Glutamatergic transmission and plasticity between olfactory bulb mitral cells." J Physiol 586: 2107-2119.

Pinching, A. J. and T. P. Powell (1971). "The neuropil of the periglomerular region of the olfactory bulb." J Cell Sci 9: 379-409.

Pinching, A. J. and T. P. Powell (1972). "Experimental studies on the axon intrinsic to the glomerular layer of the olfactory bulb." J Cell Sci 10: 637-655.

Pocock, G. and C. D. Richards (1993). "Excitatory and inhibitory synaptic mechanisms in anesthesia." Br J Anesth 71: 134-147.

Poo, C. and J. S. Isaacson (2009). "Odor representations in olfactory cortex: "sparse" coding, global inhibition, and oscillations." Neuron 62: 850-861.

150

Presseler, R. T. and B. W. Strowbridge (2006). "Blanes cells mediate persistent feedforward inhibition onto granule cells in the olfactory bulb." Neuron 49: 889- 904.

Price, J. L. and T. P. Powell (1970). "The mitral and short axon cells of the olfactory bulb." J Cell Sci 7: 631-651.

Price, J. L. and T. P. Powell (1970). "The morphology of the granule cells of the olfactory bulb." J Cell Sci 7: 91-123.

Price, J. L. and T. P. Powell (1970). "The synaptology of the granule cells of the olfactory bulb." J Cell Sci 7: 125-155.

Puil, E. and H. el-Beheiry (1990). "Anesthetic suppression of transmitter actions in neocortex." Br J Anesth 101: 61-66.

Puil, E., H. el-Beheiry, et al. (1990). "Anesthetic effects on glutamate-stimulated increase in intraneuronal calcium." J Pharmacol Exp Ther 255: 955-961.

Puopolo, M. and O. Belluzzi (2001). "NMDA-dependent, network-driven oscillatory activity induced by bicuculline or removal of Mg2+ in rat olfactory bulb neurons." Eur J Neurosci 13: 92-102.

Rall, W. (1967). "Distinguishing theoretical synaptic potentials computed for different soma-dendritic distributions of synaptic input." J Neurophysiol 30: 1138-1168.

Ravel, N., P. Chabaud, et al. (2003). "Olfactory learning modifies the expression of odour-induced oscillatory responses in the gamma (60-90 Hz) and beta (15-40 Hz) bands in the rat olfactory bulb." Eur J Neurosci 17: 350-358.

Ressler, K. J., S. L. Sullivan, et al. (1993). "A zonal organization of odorant receptor gene expression in the olfactory epithelium." Cell 73: 597-609.

Reyher, C. H., J. Lubke, et al. (1991). "Olfactory bulb granule cell aggregates: morphological evidence for interperikaryal electrotonic coupling via gap junctions." J Neurosci 11: 1485-1495.

Rorig, B., G. Klausa, et al. (1996). "Intracellular acidification reduced gap junction coupling between immature rat neocortical pyramidal neurons." J Physiol 490: 31-49.

Rouach, N., M. Segal, et al. (2003). "Carbenoxolone blockade of neuronal network activity in culture is not mediated by and action on gap junctions." J Physiol 553: 729-745.

151 Sakura, M., M. Kabetani, et al. (2004). "Impairment of olfactory discrimination by blockade of nitric oxide activity in the terrestrial slug limax valentianus." Neurosci Lett 370: 257-261.

Salin, P.-A., P.-M. Lledo, et al. (2001). "Dendritic glutamate autoreceptors modulate signal processing in rat mitral cells." J Physiol 85: 1275-1282.

Sassoe-Pognetto, M. and O. P. Ottersen (2000). "Organization of ionotropic glutamate receptors at dendrodendritic synapses in the rat olfactory bulb." J Neurosci 20: 2192-2201.

Satou, M. (1990). "Synaptic organization, local neuronal circuitry, and functional segregation of the teleost olfactory bulb." Prog Neurobiol 34: 115-142.

Satou, M., S. Anzai, et al. (2005). "Long-term potentiation and olfactory memory formation in the carp (Cyprinus carpio L.) olfactory bulb." J Comp Physiol 191: 421-434.

Schmitz, D., S. Schuchmann, et al. (2001). "Axo-axonal coupling: a novel mechanism for ultrafast neuronal communication." Neuron 31: 831-840.

Schneider, S. P. and F. Macrides (1978). "Laminar distributions of interneurons in the main olfactory bulb of the adult hamster." Brain Res Bull 3: 73-82.

Schoenfeld, T. A., A. N. Clancy, et al. (1994). "The spatial organization of the peripheral olfactory system of the hamster. I. receptor neuron projections to the main olfactory bulb." Brain Res Bull 34: 183-210.

Schoenfeld, T. A., J. E. Marchand, et al. (1985). "Topographic organization of tufted cell axonal projections in the hamster main olfactory bulb: an intrabulbar associational system." J Comp Neurol 235: 503-518.

Schoppa, N. E. (2006). "Synchronization of olfactory bulb mitral cells by precisely timed inhibition inputs." Neuron 49: 271-283.

Schoppa, N. E., J. M. Kinzie, et al. (1998). "Dendrodendritic Inhibition in the olfactory bulb is driven by NMDA receptors." J Neurosci 18: 6790-6802.

Schoppa, N. E. and G. L. Westbrook (1999). "Regulation of synaptic timing in the olfactory bulb by an A-type potassium current." Nat Neurosci 2: 1106-1113.

Schoppa, N. E. and G. L. Westbrook (2001). "Glomerulus-specific synchronization of mitral cells in the olfactory bulb." Neuron 31: 639-651.

Schoppa, N. E. and G. L. Westbrook (2002). "AMPA autoreceptors drive correlated spiking in olfactory bulb glomeruli." Nat Neurosci 5: 1194-1202.

152

Shadlen, M. N. and J. A. Movshon (1999). "Synchrony unbound: a critical evaluation of the temporal binding hypothesis." Neuron 24: 111-125.

Shepherd, G. M. (1972). "Synaptic organization of the mammalian olfactory bulb." Physiological Review 52: 864-912.

Shepherd, G. M. and C. A. Greer (1998). Olfactory Bulb. The Synaptic Organization of the Brain. G. M. Shepherd. New York, Oxford University Press: 159-204.

Shimshek, D. R., T. Bus, et al. (2005). "Enhanced odor discrimination and impaired olfactory memory by spatially controlled switch of AMPA receptors." PLoS Biol 3: e354.

Shipley, M. T. and G. D. Adamek (1984). "The connections of the mouse olfactory bulb: a study using orthograde and retrograde transport of wheat germ agglutinin conjugated to horseradish peroxide." Brain Res Bull 12: 669-688.

Shouval, H. Z., M. F. Bear, et al. (2002). "A unified model of NMDA receptor-dependent bidirectional synaptic plasticity." Proc Natl Acad Sci USA 99: 10831-10836.

Shu, Y., A. Hasenstaub, et al. (2003). "Turning on and off recurrent balanced cortical activitiy." Nature 423: 288-293.

Sjostrom, P. J. and M. Hausser (2006). "A cooperative switch determines the sign of synaptic plasticity in distal dendrites of neocortical pyramidal neurons." Neuron 51: 227-238.

Sjostrom, P. J., G. G. Turrigiano, et al. (2003). "Neocortical LTD via coincident activation of presynaptic NMDA and cannabinoid receptors." Neuron 39: 641-654.

Smith, T. C. and C. E. Jahr (2002). "Self-inhibition of olfactory bulb neurons." Nat Neurosci 5: 760-766.

Steriade, M., A. Nunez, et al. (1993). "A novel slow (< 1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components." J Neurosci 13: 3252-3265.

Stewart, W. W. (1978). "Functional connections between cells as revealed by dye coupling with a highly fluorescent napthalimide tracer." Cell 14: 741.

Stopfer, M., S. Bhagavan, et al. (1997). "Impaired odour discrimination on desynchronization of odour-encoding neural assemblies." Nature 390: 70-74.

Storm, J. F. (1988). "Temporal integration by a slowly inactivating K+ current in hippocampal neurons." Nature 336: 379-381.

153

Tovar, K. R., B. J. Maher, et al. (2009). "Direct actions of carbenoxolone on synaptic transmission and neuronal membrane properties." J Neurophysiol 102: 974-978.

Urban, N. N. and B. Sakmann (2002). "Reciprocal intraglomerular excitation and intra- and interglomerular lateral inhibition between mouse olfactory bulb mitral cells." J physiol 542: 355-367.

Usrey, W. M., J.-M. Alonso, et al. (2000). "Synaptic interactions between thalamic inputs to simple cells in cat visual cortex." J Neurosci 20: 5461-5467.

Van Gehuchten, A. and I. Martin (1891). "Le bulbe olfactif chez quelques mammiferes." Cellule 7: 205-237.

Vassar, R., S. K. Chao, et al. (1994). "Topographic organization of sensory projections to the olfactory bulb." Cell 79: 981-991.

Vassar, R., J. Ngai, et al. (1993). "Spatial segregation of odorant receptor expression in the mammalian olfactory epithelium." Cell 74: 309-318.

Veenstra, R. D. (1996). "Size and selectivity of gap junction channels formed from different connexins." J Bioenerg Biomembr 28: 327-337.

Venance, L., A. Rozov, et al. (2000). "Connexin expression in electrically coupled postnatal rat brain neurons " Proc Natl Acad Sci USA 97: 10260-1-265.

Verhagen, J. V., D. W. Wesson, et al. (2007). "Sniffing controls an adaptive filter of sensory input to the olfactory bulb." Nat Neurosci 10: 631-639.

Vessey, J. P., M. R. Lalonde, et al. (2004). "Carbenoxolone inhibition of voltage-gated Ca channels and synaptic transmission in the retina." J Neurophysiol 92: 1252- 1256.

Wachowiak, M. and M. T. Shipley (2006). "Coding and synaptic processing of sensory information in the glomerular layer of the olfactory bulb." Semin Cell Dev Biol 17: 411-423.

Weliky, M., W. H. Bosking, et al. (1996). "A systematic map of direction preference in primary visual cortex." Nature 379: 725-728.

Wellis, D. P. and J. S. Kauer (1993). "GABAA and glutamate receptor involvement in dendrodendritic synaptic interactions from salamander olfactory bulb." J Physiol 469: 315-339.

Wellis, D. P. and J. S. Kauer (1994). "GABAergic and glutamatergic synaptic input to identified granule cells in salamader olfactory bulb." J Physiol 475: 419-430.

154

Wellis, D. P. and J. W. Scott (1990). "Intracellular responses of identified rat olfactory bulb interneurons to electrical and odor stimulation." J Neurophysiol 64: 932-947.

Wilson, D. A., R. M. Sullivan, et al. (1985). "Odor familiarity alters mitral cell response in the olfactory bulb of neonatal rats." Brain Res 354: 314-317.

Wilson, R. I., G. C. Turner, et al. (2004). "Transformation of olfactory representations in the Drosophila antennal lobe." Science 303: 366-370.

Woolf, T. B., G. M. Shepherd, et al. (1991). "Serial reconstructions of granule cell spines in the mammalian olfactory bulb." Synapse 7: 181-192.

Yang, Q. and H. B. Michelson (2001). "Gap junctions synchronize the firing of inhibitory interneurons in guinea pig hippocampus." Brain Res 907: 139-143.

Yokoi, M., K. Mori, et al. (1995). "Refinement of odor molecule tuning by dendrodendritic synaptic inhibition in the olfactory bulb." Proc Natl Acad Sci 92: 3371-3375.

Yoshihara, Y., M. Kawasaki, et al. (1997). "OCAM: a new member of the neural cell adhesion molecule family related to zone-to-zone projection of olfactory and vomeronasal axons." J Neruosci 17: 5830-5842.

Zhang, C. and D. Restrepo (2002). "Expression of connexin 45 in the olfactory system." Brain Res 929: 37-47.

Zhou, F. M. and J. J. Hablitz (1997). "Rapid kinetics and inward rectification of miniature EPSCs in layer I neurons of rat neocortex." J Neurophysiol 77: 2416- 2426.

Zibrowski, E. M. and C. H. Vanderwolf (1997). "Oscillatory fast wave activity in the rat pyriform cortex: relation to olfaction and behavior." Brain Res 766: 39-49.

155