Heiman, Maxwell G. et al. (2009) International Worm Meeting "Different sensory dendrites use distinct machineries to attain their lengths."

Shaham, Shai, Heiman, Maxwell G.

[International Worm Meeting, 2009]

Each embryo encodes a map, with cell shape and cell contacts following set patterns. In C. elegans, these patterns are essentially invariant, with a given cell forming the same shape and contacts in every individual. To understand how this developmental map is encoded and read, we have taken a forward genetic approach, isolating mutants in which the map is perturbed. We focused first on the amphids, a well-characterized pair of sense organs in the head of the animal. Each amphid consists of 12 neurons, which extend unbranched sensory dendrites to the tip of the nose, and two glial cells, the sheath and the socket, which also extend processes to the nose where they surround dendritic endings of the neurons. Using time-lapse imaging to observe the formation of single dendrites, we found that these cells acquire their shapes by an unusual mechanism: the neurons first form contacts with the presumptive nose, and then the cell body migrates away, dragging out the dendrite behind it. This mechanism suggests a developmental map based on the formation of specific cell contacts, for example between dendrites and the nose tip. Indeed, we isolated mutants in which a "map failure" occurs, with the amphid failing to attain its proper length, and these mutants affect a pair of secreted extracellular matrix proteins, DYF-7 and DEX-1, required to anchor dendritic tips at the nose during cell migration. DYF-7 is a zona pellucida (ZP) domain protein expressed in most sensory neurons in the developing embryo; DEX-1 is a nidogen- and zonadhesin-containing protein expressed widely in the developing head and tail. Yet, the effects of these mutants are limited to the amphids and a similar pair of tail sense organs, the phasmids. The presence of 40 predicted ZP-encoding C. elegans genes, and several zonadhesin-like genes, suggests the possibility that part of the developmental map is encoded by combinatorial matchmaking between DYF-7- and DEX-1-like proteins. To test this hypothesis, we have isolated mutants in which sensory dendrites of the URX or CEP neurons fail to attach to the nose. Although URX and dorsal CEP neurons are lineal sisters, the mutants show that these cells attach to the nose using machineries distinct from each other and from that of the amphid. While not yet cloned, these mutants might affect additional ZP or ZP-interacting proteins. Indeed, the presence of ZP proteins in sense organs is highly conserved across species and sensory modalities. The involvement of distinct ZP proteins would support the model of a developmental map partly based on cells forming contacts with specific extracellular attachment sites.

Maxwell Heiman et al. (2007) International Worm Meeting "Sense organ morphogenesis requires a conserved neuron-glia protein pair that stabilizes dendrite tips."

Shai Shaham, Maxwell Heiman

[International Worm Meeting, 2007]

The assembly of cells into organs involves the precise assignment of a vast diversity of cell shapes and the formation of numerous specific cell-cell contacts. To understand how these traits are established, we have turned to the major sense organ of C. elegans, the amphid, consisting of 14 cells - 12 neurons and two glia - with well-defined morphologies and reproducible cell-cell contacts. The 12 neurons extend dendrites to the tip of the nose, where they sense the environment. The sheath glial cell also extends a process to the nose, where it completely ensheathes all 12 dendritic endings. The socket glial cell likewise extends a process to the nose, where it forms a pore through which some dendrites directly access the environment. To identify factors required for sense organ assembly, we isolated mutants in which neuron-glia morphogenesis goes awry. We identified mutations in dyf-7 and dex-1 that cause neuronal dendrites and the sheath glial process to fail to extend, resulting in a severely shortened organ disconnected from the nose. To understand how these genes act, we developed an optical marking method using a photoconvertible fluorescent protein, allowing us, for the first time, to visualize the development of a single amphid dendrite in real time. In wild-type animals, the neuron is born near the presumptive nose and polarizes towards it; next, the tip of this projection remains at the nose while the cell body migrates posteriorly, such that a full-length dendrite forms by a stretching mechanism. In dyf-7 mutants, neurons polarize normally but the tip of the projection is dragged behind the cell body during the posterior migration. Thus, dyf-7 stabilizes dendrite tips. To understand the molecular mechanism by which dendrite tips are stabilized, we cloned the dyf-7 and dex-1 genes. DYF-7 is a transmembrane protein expressed by neurons during dendrite development and localizes to dendrite tips during cell body migration. It is homologous to zona pellucida (ZP) proteins that form the matrix around the vertebrate oocyte. DEX-1 is a transmembrane protein expressed by the socket glial cell at the same time in development and partly colocalizes with DYF-7. It contains a zonadhesin domain, similar to the protein on sperm used to bind the ZP, and two nidogen domains, implicated in neurite outgrowth. Weak loss-of-function alleles of dyf-7 and dex-1 exhibit a strong genetic interaction, suggesting these factors may be a ligand-receptor pair. DYF-7 and DEX-1 share a domain architecture with the tectorins, a pair of proteins that form a matrix in the vertebrate ear required for hearing, suggesting they represent a conserved module for sense organ development.

Maxwell G Heiman et al. (2004) East Coast Worm Meeting "Form and function of glia-neuron interactions"

Shai Shaham, Maxwell G Heiman

[East Coast Worm Meeting, 2004]

Glia are the primary source of brain tumors and are the most abundant cells in the brain, but their functions remain mysterious. To understand how glia interact with neurons and regulate neuronal behavior, we are studying the amphid sheath cell of Caenorhabditis elegans as a model glial cell. Like vertebrate glial cells, the amphid sheath cell is derived from a neuronal lineage, physically associates with neurons, and provides functions required for neuronal activity. The amphid, a sensory organ in the head of the worm, consists of 12 neurons and two glial-like cells, the sheath and the socket. The neurons project unbranched dendrites to an opening formed by the socket cell at the head of the worm. The sheath cell forms a process collateral with these dendrites, about 100 microns long in an adult animal, which terminates in a specialized structure that completely ensheathes all the dendritic endings. To identify genes that control the differentiation and morphogenesis of the glial cell and its association with neurons, we performed a genetic screen for mutants defective in glial development. We created a transgenic strain expressing fluorescent reporter genes in the amphid sheath cell and two amphid neurons. We performed a non-clonal visual screen and isolated a new class of mutants in which both the amphid sheath cell and the associated neuronal dendrites extend only a fraction of the wild-type length. Preliminary analysis suggests the glial cell may be required for dendrite extension. In a complementary approach, we are studying what signals the sheath cell supplies to the neurons. In wild-type animals, the sheath cell deposits an electron-dense matrix around the dendrites it ensheathes, while in che-12 mutants this matrix is severely lessened and neuronal functions are disrupted. To gain insight into the secretory pathway that connects the glial cell to neurons, we are fine mapping the che-12 mutation to identify the relevant gene. Additionally, we are attempting to isolate directly factors secreted by the amphid sheath cell.

Maxwell Heiman et al. (2008) C.elegans Neuronal Development Meeting "The Extracellular Proteins DEX-1 and DYF-7 Promote Sensory Dendrite Outgrowth by Anchoring Dendritic Tips During Cell Migration"

Maxwell Heiman, Shai Shaham

[C.elegans Neuronal Development Meeting, 2008]

Neuronal shape, in particular the lengths and orientations of axons and dendrites, is a major determinant of neuronal function. While the developmentally fixed nervous system of C. elegans has been used to shed light on mechanisms of axon guidance, relatively little is known about how dendrites form or how neurons and glia undergo coordinated morphogenesis. To address these issues, we turned to the amphid, the major sense organ of the worm, comprised of 12 neurons that extend dendrites to the tip of the nose to receive information about the environment, and two glia (the sheath and the socket) that extend processes to the nose and envelop these sensory dendritic endings. Using a visual genetic screen, we isolated mutations in two genes, dex-1 and dyf-7, that cause the 12 sensory dendrites and the sheath glial process to be severely shortened. To determine how these dendrites normally grow, we used optical cell marking (OCM) with the photoconvertible protein Kaede to observe the formation of single dendrites in living embryos. We found that a sensory neuron is born near the presumptive nose, polarizes towards it, and then the cell body undergoes a ~10 microm, ~30 min posterior migration while the dendritic tip remains anchored in place, forming a dendrite by a phenomenon we call retrograde extension. In dyf-7 embryos, the dendritic tip fails to anchor, and is dragged along behind the migrating cell. DEX-1 and DYF-7 are a pair of transmembrane proteins resembling tectorins, proteins found in a matrix in the inner ear that anchors sensory hair cells against the motion caused by sound waves. DEX-1 and DYF-7 are expressed at the time of dendrite formation, with DEX-1 expressed by non-neuronal cells surrounding sense organs and DYF-7 expressed by the sensory neurons themselves, localizing to dendritic tips. Like the tectorins, DEX-1 and DYF-7 undergo proteolytic release from their membrane anchors and act in the extracellular environment. dex-1 and dyf-7 mutants display a strong genetic interaction, suggesting they cooperate to anchor dendrites. Finally, DYF-7 is capable of multimerizing and this multimerization is blocked by a point mutation that disrupts dendrite anchoring, suggesting that DEX-1 and DYF-7 may work by assembling an extracellular matrix that anchors dendritic tips against the pull of cell migration.

Maxwell G Heiman et al. (2005) International Worm Meeting "A neuronal membrane protein required for dendrite and glial process extension"

Maxwell G Heiman, Shai Shaham

[International Worm Meeting, 2005]

The development of complex cell shape is crucial for nervous system assembly and function. The primary sensory organ of C. elegans, the amphid, is composed of 12 neurons that each extend an axon and a dendrite, and two glia, one of which extends a process that is tightly bundled to the sensory dendrites. Although many studies have focused on the mechanism of axon development, little is known about how dendrites or glial processes form, or how abnormal morphogenesis of these structures affects nervous system function.To understand dendrite and glia morphogenesis, we mutagenized animals expressing fluorescent markers in amphid neurons and sheath glia and isolated mutants defective in amphid morphology. We identified a novel class of mutants in which dendrites are a fraction of their wild-type length, a phenotype we call Dex (Dendrite extension defective). We characterized mutants in two genes, dex-1 and dyf-7. In addition to shortened dendrites, mutation of either gene causes shortened glial processes. In dex-1 mutants, the dendrites and glial processes are of variable lengths within a population; yet in a single amphid they are always of equal length, revealing that dendrite and glial process extensions are coupled. Remarkably, amphid neuron axons are normal in these mutants, suggesting that dendrites and axons use different molecules for outgrowth.We cloned dyf-7, mutations in which had also been isolated in previous screens for sensory neuron dysfunction, and found that it encodes a putative membrane protein with a conserved extracellular domain, and is expressed in several embryonic neurons. dyf-7 acts early in embryogenesis near the time dendrite extension begins, and promoter-swap experiments showed that dyf-7 can act in neurons or, surprisingly, in glia. A truncated DYF-7 protein, predicted to be secreted, is capable of rescuing dyf-7 mutants. Thus, DYF-7 may act in the extracellular milieu to promote dendrite and glial process extension.To understand how dyf-7 and dex-1 promote dendrite extension, we are characterizing the process in detail in wild-type embryos. We established an optical cell-marking system, using the photoconvertible fluorescent protein Kaede, which has allowed us to observe formation of individual amphid neurons or glia in real time. We are using this system to identify the steps at which dyf-7 and other Dex mutants fail to advance. Because DYF-7 contains a conserved domain implicated in neuronal morphogenesis in other systems, the mechanism underlying amphid formation may reflect a conserved means of sensory organ development.

Yip, Candice et al. (2013) International Worm Meeting "Dendrite tiling as an emergent property of self-avoidance."

Yip, Candice, Heiman, Maxwell

[International Worm Meeting, 2013]

Dendrite arbors are sculpted into a morphologically diverse array of shapes and sizes, which are critical determinants of neural circuit function and behavior. This elaborate morphogenesis is governed by fundamental organizational principles, including isoneuronal repulsion between dendrites of a single neuron, called self-avoidance, and heteroneuronal repulsion between dendrites of neurons belonging to the same class, called tiling. Here, we took advantage of a lin-22 mutation that causes the formation of extra PVD neurons to demonstrate that dendrite tiling can arise as an emergent property of self-avoidance. PVDs are mechanosensory neurons that elaborate complex, yet ordered, dendrite arbors through the contact-mediated repulsion of "self" dendrites. In wild-type animals, a single PVD extends dendrites that cover the entire body wall. In contrast, lin-22 mutants generate a total of five PVDs on each side that are evenly spaced along the length of the animal. We found that each lin-22 PVD elaborates a smaller, non-overlapping dendrite field, suggestive of tiling. We used laser ablation of the ectopic PVDs to demonstrate that these smaller arbors are not due to defective dendrite growth in the lin-22 background, but instead reflect mutual repulsion between PVDs, consistent with a tiling model. Next, we tested whether unc-6/Netrin, which has been shown to mediate PVD dendrite self-avoidance in wild-type animals, is also required for dendrite tiling between PVDs. Indeed, tiling in unc-6; lin-22 double mutants was severely diminished, suggesting that unc-6-mediated self-avoidance is sufficient for generating a tiled array of dendrite arbors in the presence of ectopic neurons of the same subtype. These results suggest that the apparently complex phenomenon of dendrite tiling can, in fact, emerge spontaneously from pre-existing mechanisms of dendrite self-avoidance.

Yip, Candice et al. (2013) International Worm Meeting "Understanding cellular mechanisms of selective fasciculation between dendrites."

Yip, Candice, Heiman, Maxwell

[International Worm Meeting, 2013]

The organization of axons and dendrites into defined bundles, known as selective fasciculation, is a conserved feature of nervous system wiring from nematodes to mammals. It shapes the development of neural circuits, including the guidance of growth cones along existing axonal bundles in grasshopper embryos, sorting Drosophila axons into longitudinal pathways, generating topographic maps of visual space in the Drosophila optic lobe, and ensuring vomeronasal axons innervate their correct targets in mice. Defining mechanisms of selective fasciculation has been hindered by two major obstacles: (1) the challenge of discerning individual processes within a bundle, which may contain hundreds of processes, and (2) the challenge of distinguishing defects in neuron outgrowth or guidance from defects in fasciculation, since these processes are intimately coupled. In order to overcome these challenges, we have turned to the major C. elegans sense organ, the amphid, which contains 12 neurons and two glial cells. Classic electron microscopic reconstructions show that each neuron extends an unbranched sensory dendrite that resides in a stereotyped position in the amphid bundle. This stereotypy in dendrite location leads to the hypothesis that dendrites establish and maintain specific adhesive connections with their neighbors. Here, we used three-color fluorescent labelling of individual amphid neurons to recapitulate the classic EM observation that amphid dendrites are located in stereotyped positions relative to one another. We are taking several approaches to elucidate cellular mechanisms underlying this selective fasciculation. First, we are imaging wild-type animals and known mutants to determine when in development and where along the dendrites selective fasciculation occurs. Second, we are using cell ablation to examine the role of each neuron in determining the positioning of its neighbors. Third, we are using cell fate transformations to test if neuron identity, rather than birthplace, specifies dendrite positioning. Understanding mechanisms of selective fasciculation will help elucidate basic developmental processes important for nervous system organization.

Mizeracka, Karolina et al. (2013) International Worm Meeting "Developing a cell contact sensor for tracking neuron-glia interactions in vivo."

Mizeracka, Karolina, Heiman, Maxwell

[International Worm Meeting, 2013]

The mechanism by which neurons selectively recognize their glial partners is an important and largely unexplored question in neurobiology. To begin to address this question, we are developing a cell contact sensor designed to report on the interaction of adjacent cells. Inspired by the GRASP system developed by Bargmann and colleagues that can be used to label synaptic contacts by the extracellular reconstitution of a GFP molecule, we aim to develop a contact sensor that will trigger a transcriptional readout in a cell of interest, providing a more permanent mark. In order to translate an external interaction to a change in intracellular gene expression, we have modeled our contact sensor on the well-characterized Delta-Notch ligand-receptor signaling pathway, a naturally occurring system in which cell-cell contact triggers transcriptional activity. In brief, interaction of an artificial ligand-receptor pair on the surfaces of adjacent cells will result in intramembrane proteolysis of the receptor, releasing a cytoplasmic transcriptional effector, and thereby genetically marking the receiver cell. First, we will test our reporter system in S2 Drosophila cells, ensuring that interaction between our artificial ligand and receptor results in cleavage of the receptor and activation of a transcriptional reporter. Next, we will test our reporter system in vivo, focusing initially on the stereotyped and highly selective interactions between the amphid sensory neuron AFD and the amphid sheath glial cell that surrounds the AFD dendrite tip. Specifically, we will express an artificial ligand on the surface of the AFD neuron and an artificial receptor on the surface of the amphid sheath glial cell. Contact between this neuron and its glial cell will result in a transcriptional output, activating a fluorescent reporter in the glial cell. This system will allow us to screen for "miswiring" mutants in which an amphid neuron improperly associates with a non-amphid glial cell, as a means to identify factors that mediate neuron-glia recognition. Once established, this contact sensor could be readily applied to report on any cell-cell interaction, in C. elegans or other animals.

McLachlan, Ian G et al. (2013) International Worm Meeting "GRDN-1/Girdin and SAX-7/L1CAM establish a glial guide for sensory dendrite extension."

Heiman, Maxwell G, McLachlan, Ian G

[International Worm Meeting, 2013]

The extension of dendrites to genetically programmed targets is a crucial step in neural circuit formation. To understand this process, we are investigating the development of the sensory anatomy in the head of C. elegans, where each neuron extends an unbranched dendrite to a distinct but stereotyped position at the nose tip. Previous work showed that amphid neurons extend dendrites by retrograde extension: the neurons are born at the nose, anchor dendrite tips there, and migrate away to stretch out dendrites behind them. The extracellular matrix component DYF-7 is required for anchoring, and its loss results in severely shortened amphid dendrites. To determine if this mechanism is shared among distinct neuron classes, we generated a panel of fluorescent markers to visualize most of the head sensory neurons, and analyzed dendrite lengths in dyf-7 animals. We found that while all glial-ensheathed sensory neurons require dyf-7 for dendrite extension, some non-ensheathed neurons-including the oxygen sensor URX-do not require dyf-7. We therefore performed a forward genetic screen for mutants that affect dendrite extension of URX, but not amphid, neurons. We identified eight alleles of two genes: grdn-1, a homolog of Girdin, a cytoskeletal adapter that localizes to cell-cell contacts in the nervous system; and sax-7, a homolog of the neuron-glia adhesion molecule L1CAM. Surprisingly, we found that GRDN-1 and SAX-7 do not act cell-autonomously in URX. Instead, grdn-1 appears to act in glia. A 450 bp grdn-1 promoter, sufficient for rescue when driving the grdn-1 cDNA, drives expression in the glial cells of the outer labial (OL) sense organs that fasciculate with URX. Expression of the grdn-1 cDNA by other glial-specific promoters is also sufficient for rescue of the URX dendrite phenotype. Embryonic imaging suggests that the URX dendrite extends from a stationary cell body, unlike retrograde extension of the amphids. We hypothesize that GRDN-1 and SAX-7 act in OL glia to form an adhesive guide, along which the URX dendrite extends. These results thus identify a novel mechanism of glial-dependent dendrite extension, distinct from the dyf-7-mediated retrograde extension of amphid and other glial-ensheathed neurons.

Mizeracka, Karolina et al. (2021) International Worm Meeting "Two of the 30 EGF domains in FBN-1/Fibrillin are required for sensory dendrite extension"

Heiman, Maxwell, Mizeracka, Karolina, Shaham, Shai

[International Worm Meeting, 2021]

Many extracellular matrix (ECM) components are large proteins composed of highly repetitive domains. For example, FBN-1/Fibrillin is a ~2700 aa protein with 30 EGF-like domains. In principle, such repetitive domains could function redundantly, for example to cross-link other matrix components, or they could be distinct modules with specialized functions. In previous studies, disruption of fbn-1 with a large internal deletion (tm290) or RNAi caused almost complete lethality and severe molting defects. In contrast, through a genetic screen for mutants affecting amphid dendrite morphogenesis, we identified two alleles of fbn-1 (ns67 and ns283) that introduce missense mutations within 32 aa in the first two N-terminal EGF-like domains (EGF1-2). These mutants cause weakly penetrant defects in amphid dendrite extension (8% and 19% defective, respectively), with minimal lethality or molting defects, suggesting that the EGF1-2 domains perform a specialized function related to dendrite extension. To further understand this, we generated a small in-frame deletion encompassing EGF1-2 (syb239) and a complete deletion of fbn-1 (syb240) using CRISPR/Cas9. Consistent with our original mutant analysis, syb239 has weakly penetrant amphid dendrite extension defects (8% defective), while syb240 exhibits severe embryonic lethality, supporting the idea of a specialized role for EGF1-2. Amphid dendrite extension also requires another ECM protein, DYF-7, which is produced by the amphid neurons and forms "caps" at the dendrite endings that prevent rupture of the developing neuron-glia sensory epithelium. FBN-1 is expressed by hypodermal cells at this stage, and we find that the EGF1-2 domains are required for proper formation of DYF-7 caps. These results show that distinct domains within FBN-1 may be specialized to support different aspects of its function, and suggest that even highly repetitive domains in ECM proteins may reflect modularity rather than redundancy.