Ann Marie Craig

Synaptic organizing molecules govern synapse assembly and control synaptic properties. This dissertation focuses on the synapse organizer neurexin 1, which is central to a genetic risk pathway for neuropsychiatric disorders. Using biochemistry, electrophysiology, immunofluorescence imaging, serial block-face electron microscopy, and behavior analyses, we study modes of regulating neurexin 1 function at hippocampal synapses. Neurexin’s role in synapse development was previously thought to be mediated purely by its protein domains. In Chapter 2, we show that it requires a post-translational glycan modification called heparan sulfate (HS). HS on neurexin 1 is essential for mouse survival and excitatory synaptic transmission. It serves as a distinct binding interface for postsynaptic partners neuroligins and LRRTMS, as well as novel ligands beyond canonically known partners. Because of heterogeneity in HS structure, our work provides a unique molecular basis for fine- tuning synapse function, and position glycan-binding motifs as potential therapeutic targets for diseases of the brain. Neurexins exemplify molecular diversity by way of alternative splicing and heparan sulfation. In Chapter 3, we reveal that these post-transcriptional and post-translational modes of regulation converge on neurexin 1 splice site 5 (S5). The S5 insert increases HS valency, an effect that is associated with reduced neurexin 1 protein level and lower glutamatergic transmission. Exclusion of neurexin 1 S5 in mice bolsters synaptic transmission while maintaining the AMPA/NMDA ratio and shifts behavior away from those associated with autism spectrum disorders. Our findings uncover NRXN1 S5 as a therapeutic site that can potentially be harnessed to restore function in neuropsychiatric disorders. In Chapter 4, we interrogate the consequences of a disease-associated monogenic loss of neurexin 1. In heterozygous Nrxn1 KO mice, we show a ~2-fold reduction in excitatory synaptic transmission while leaving synapse numbers unaffected. This deficit is rescued back to WT levels by the exclusion of S5 in the remaining WT Nrxn1 allele, consistent with a restoration of Nrxn1 protein level. Heterozygous deletion of Nrxn1 reduces presynaptic neurotransmitter release and AMPA/NMDA ratio, both rescued by the S5 exclusion strategy. Thus, we present a genetic platform whereby neurexin 1 S5 can be leveraged to potentially develop therapeutics to alleviate brain-based defects.

View record

Brain function is dependent on both the properties of the individual neurons of which it is made up, as well as the precise pattern of connections between them. Both aspects of brain circuitry are determined in large part by the expression of specific genes. The products of these genes, including adhesion molecules, ion channels, and transcription factors, go on to shape neurons’ properties and as well as their connectivity.Neurons direct their connectivity in part by expressing trans-synaptic cell adhesion molecules, which act as “molecular velcro” in order to form physical connections between axons and dendrites. These adhesion molecules are able to recruit the necessary components for synaptic transmission, through mechanisms which are incompletely understood. Chapter 2 focuses on one presynaptic adhesion molecule, PTPσ, and its interaction with intracellular scaffolding proteins. PTPσ is a phosphatase, but I found that its phosphatase activity is not needed for its ability to induce new synapses. However, interaction with the presynaptic scaffolding protein liprin-α is required. Because liprin-α binds to multiple presynaptic components, this finding suggests a mechanism through which PTPσ recruits liprin-α to nascent synapses, and liprin-α in turn recruits other components, eventually leading to recruitment of vesicles, calcium channels, and everything else necessary for functional presynaptic release.Phenotypes of individual neurons, including electrophysiological activity and morphology, are known to be under genetic control. However, given the number of genes expressed by one organism and the difficulty of experimentally exploring the consequences of loss of function of any one gene, our understanding of the molecular underpinnings of any particular phenotype remains incomplete. In Chapter 3, I took the approach of searching for correlations between gene expression and neuronal phenotypes in a publicly available dataset. I found that controlling for broad cell class (that is, whether cells are inhibitory or excitatory) made a substantial difference to the results, and found much better correspondence with other datasets and with previously published literature when I took this step. In this work, I generated many testable hypotheses regarding the relationship between specific genes and neuronal phenotypes, which I hope will help to guide future studies.

View record

Synaptic adhesion molecules play central roles in synapse organization and neuronal signal transduction machineries. This dissertation first focuses on calsyntenin-3, a crucial synaptic adhesion molecule identified by our lab, based on in vitro and in vivo biochemistry, immunofluorescence, electron microscopy, and electrophysiological studies. Chapter 2 shows that calsyntenin-3 interacts with α-neurexins, but not β-neurexins, at nanomolar affinity. Being calcium-dependent and regulated by the heparan sulfate modification of α-neurexins, this interaction requires the cadherin and LNS domains of calsyntenin-3, as well as the LNS5-EGFc-LNS6 domains of α-neurexins. Calsyntenin-3 full-length form triggers pre-synapse differentiation, but calsyntenin-3-shed ectodomain suppresses the ability of α-neurexin partners in mediating pre-synapse differentiation. Calsyntenin-3 is present in pyramidal neurons throughout cortex and hippocampus, and is most highly expressed in interneurons. Young adult calsyntenin-3 knockout mice present deficits in both density and transmission for both excitatory and inhibitory synapses.Chapter 3 shows that calsyntenin-3 regulates synaptic transmission in different neuronal types. Conditional calsyntenin-3 knockout in forebrain interneurons does not affect basal inhibitory transmission in CA1 pyramidal neurons, while conditional knockout in excitatory neurons enhances both basal excitatory and inhibitory transmission. Partial calsyntenin-3 knockout in primary hippocampal neuron culture and in juvenile calsyntenin-3 mice exhibit enhanced inhibitory transmission. Juvenile calsyntenin-3 knockout mice, however, show no difference in inhibitory transmission compared to wild-type littermates. Thus, calsyntenin-3 contributes in a developmentally regulated and at least partially non-cell-autonomous manner to synaptic transmission.Chapter 4 discusses our discovery of unexpected germline recombination in distinct mouse Cre driver lines and its profound implications. The Cre-loxP system (and resulting Cre mouse lines) has been widely used for cell-type specific gene manipulations and helps decipher gene functions, especially in neuroscience research. We demonstrated that unexpected germline recombination occurs in many Cre lines based on the studies of two Cre lines within our lab and collected data for fifty-seven Cre lines worldwide. Consequently, this may sway the interpretation of past experimental results as well as moving forward. This work not only provides guidelines for breeding strategies and precautions aimed at future studies, but most importantly for the first time, comprehensively raises awareness of this phenomenon among the neuroscience community.

View record

The proper functioning of the brain and central nervous system (CNS) requires the precise formation of synapses between neurons. The two main neurotransmitter systems for fast synaptic communication in the CNS are excitatory glutamate and inhibitory gamma-aminobutyric acid. A growing body of evidence has begun to uncover several shared and divergent rules for the establishment of each of these two types of synapses.At the molecular level, a number of key proteins have been shown to be involved in the initial formation and subsequent development of synaptic connection, including cell adhesion molecules (CAMs). Among the CAMs, neurexins and neuroligins are important synaptogenic proteins that act trans-synaptically to organize synapses: binding of axonal beta-neurexins by neuroligins is sufficient to cause development of a presynaptic specialization at that site, while binding of dendritic neuroligin-1 or neuroligin-2 by beta-neurexins is sufficient to cause development of postsynaptic excitatory or inhibitory specializations, respectively. In Chapter 2, we explore the role of alpha-neurexins in synapse organization. We find alpha-neurexins are able to specifically induce the formation of inhibitory synapses, presumably through clustering of postsynaptic neuroligin-2. Moreover, we find that the expression of various splice variants of alpha- and beta-neurexins is regulated both during development and by activity, suggesting a physiological role for alternative splicing in the modulation of synapse assembly.At the cellular level, it is now clear from live imaging studies that synapses and their formation are highly dynamic processes. A number of studies have established the temporal recruitment of pre- and postsynaptic components to nascent synapses and how synapse formation can influence neuron growth. However, these studies have focused on excitatory synapses. In Chapter 3, we explore the cellular mechanisms of inhibitory synapse formation and modulation. We find that entire synapses are highly mobile and can undergo dynamic structural modulation. New synapses are formed by gradual accumulation of components from diffuse cytoplasmic pools, with a significant contribution of presynaptic vesicles from previously recycling sites. These results provide new insights into the mechanisms of inhibitory synapse formation and how it is both similar and different from excitatory synapse formation.

View record

Activity through NMDA type glutamate receptors sculpts connectivity in the developing nervous system and is typically studied in the visual system in vivo where individual synapses are difficult to visualize. Here, we developed a model of NMDA-receptor dependent synaptic competition in dissociated cultured hippocampal neurons. GluN1 -/- (KO) mouse hippocampal neurons were cultured alone or in defined ratios with wild type (WT) neurons. Synapse development was assessed by immunofluorescence for PSD-95 apposed to VGlut1. Synapse density was specifically enhanced only onto minority WT neurons co-cultured with majority KO neighbour neurons and this increased synapse density was dependent on activity through NMDA receptors. This enhanced synaptic density onto NMDA receptor-competent neurons in minority co-culture represents a cell culture paradigm for studying synaptic competition. Trafficking of NMDA receptors to the cell surface is critical for proper brain function. Recent evidence suggest that surface trafficking of other ionotropic glutamate receptors requires ligand binding for exit from the endoplasmic reticulum. We show that glutamate binding is required for trafficking of NMDA receptors to the cell surface by expressing a panel of GluN2B ligand binding mutants in heterologous cells and primary rodent neurons and found that glutamate efficacy correlates with surface expression. Such a correlation was found even with inhibition of endocytosis indicating differences in forward trafficking. These results indicate that ligand binding is critical for receptor trafficking to the cell surface. NMDA receptors mediate many forms of synaptic plasticity. GluN2B is proposed to bind and recruit CaMKII to synapses to mediate multiple forms of synaptic plasticity. We find that accumulation of CFP-CaMKIIα at synapses is induced in wild-type but not in KO neurons by bath stimulation of NMDA receptors or by a chemical long-term potentiation protocol. Stimulated synaptic accumulation of CFP-CaMKIIα was rescued in KO neurons by YFP-GluN2B or chimeric GluN2A/2B tail but not by GluN2A, chimeric GluN2B/2A tail, or GluN2B with point mutations in the CaMKII binding site. Thus, activity-regulated synaptic aggregation of CaMKII is dependent on the cytoplasmic CaMKII binding site of GluN2B and not on differential permeation properties between GluN2B and GluN2A.

View record

The brain consists of billions of neurons. During development, these neuronsmust migrate to their proper position and form connections with neighboring neuronsto form networks. The specificity and maturation of these connections, or synapses,are critical for proper brain function, including learning, memory and cognition. Manycell adhesion molecules (CAMs) are involved in the formation and maturation ofsynapses, including the well-characterized neuroligin-neurexin pair. In this study, twonew synapse modifying proteins, calsyntenin and MDGA, are characterized using invitro assays and primary hippocampal neuron cultures. Calsyntenin-3 was identifiedin an un-biased screen to search for new synaptogenic proteins. It is a post-synaptictransmembrane protein that induces the formation of excitatory and inhibitorypresynaptic specializations in contacting axons via extracellular cadherin and LNSdomains. Overexpression of calsyntenin-3 in neurons increases presynaptic proteinclustering. Interestingly, calsyntenin-3 binds to α-neurexins with high affinity,suggesting presynaptic induction is mediated through trans-synaptic signaling withneurexins. MDGAs are a family of synaptic GPI-linked proteins that bind neuroligin-2with high affinity. MDGA1 blocks the presynaptic induction activity of neuroligin-2,through blocking binding to neurexins, via extracellular immunoglobulin domains.Overexpression of MDGA1 in neurons specifically decreases inhibitory synapses,while knockdown increases inhibitory synapses. Interestingly, like other synapticproteins including neurexin and neuroligin, MDGAs have recently been linked toautism spectrum disorders and schizophrenia. Thus, the characterization of thesynapse-promoting calsyntenin-3 and the synapse-reducing MDGA1 shed new light on the mechanisms by which synaptogenesis is regulated. Investigating the complexinterplay between molecular players during synaptogenesis is critical not only forunderstanding normal brain development, but also for providing insight intoneurodevelopmental disorders.

View record