BIOLOGICAL OVERVIEW

Integrins are cell surface receptors involved in cell adhesion to other cells and to the extracellular matrix. Cells form multiprotein complexes (adhesion plaques), for which integrins constitute a principle component. Drosophila integrins comprise a multiprotein family consisting of multiple alpha and beta subunits. The first position specific integrins to have been isolated comprise two alpha subunits (alphaPS1 and alphaPS2), and a common beta subunit (betaPS). These three proteins have been named, respectively, Multiple edematous wings (Mew), Inflated (If), and Myospheroid (Mys) (Gotwals, 1994b). The myospheroid gene, isolated by T. W. Wright in the late 1950s, received its name from the tendency of muscles in mutant embryos to detach from their attachment sites (the apodemes) after the muscles have initially made proper attachments, and subsequently round up, i.e. become spheroid.

Each integrin molecule has an alpha chain and a beta chain of amino acids. The alpha and the beta chains combine to form an alpha-beta heterodimer. AlphaPS1 betaPS and alphaPS2 betaPS are the two principle heterodimer combinations of integrins. During gastrulation they first accumulate in embryonic tissues on the basal surface of adjacent ectodermal and presumptive mesodermal layers of the blastoderm, respectively. A specific defect in gastrulation cannot be detected at this stage, but later in germband extention [Images], defects are apparent in about half the mutants. Also apparent is a defective attachment of ectodermal and mesodermal layers. Thus PS integrins appear to play an early role in the integrity of central tissues associated with the germband and in germband extention (Roote, 1994).

Later roles can be detected in the attachment of the amnioserosa [Image] to the hindgut (a region in which MYS is concentrated) and a failure in endoderm migration and midgut constriction. Endoderm forms from the anterior and posterior midgut rudiments, derived mainly from a type of cell termed the principle midgut epithelial cell. Precursors of endothelium from the anterior and posterior midgut rudiments migrate along the visceral mesoderm, changing their shape as they migrate from mesenchymal to epithelial morphology (Tepase, 1994).

PS integrins are concentrated in both the midgut tissue and the surrounding visceral mesoderm. In mys mutants, midgut and mesoderm primordia invaginate normally but remain as spherical cell masses. The cells show little migration toward the center of the embryo (Roote, 1995).

The special role of integrins in wing morphogenesis is reviewed at the apterous site. Integrins whose transcription is regulated by apterous are involved in a compartment-specific role controlling dorsal and ventral compartment integrity (Blair, 1994). Integrins are involved in eye morphogenesis, in a process anchoring rhabdomeres to the cone cell plate (Longley, 1995).

The view of integrins as mere passive hooks mediating cell adhesion events is out of date. Studies in mammalian systems reveal that signaling through integrins results in phosphorylation of a protein tyrosine kinase, known as focal adhesion kinase (FAK) (Lipfert, 1992). Interaction between FAK and Crk-associated tyrosine kinase substrate (CAS) is a consequence of this activation due to interaction with ligand. Activation of FAK results in modification of SCR by FAK and interaction with GRB2 (Polte, 1995). This developmental alphabet soup results in the building up of focal adhesion plaque, the structure by which the cell adheres to substratum. Interactions between the proteins constituting the plaque are distinctly non-linear, and quite complex.

Thus integrins are a major component of the cell's machinery for adherence to substrate (either extracellular matrix or other cells), based on a calcium dependent process that produces a focal adhesion plaque. The process involves signaling between the extracellular domain of integrins and the interior of the cell, with many intracellular responses to the adhesion event.

PS integrins and laminins: key regulators of cell migration during Drosophila embryogenesis

During embryonic development, there are numerous cases where organ or tissue formation depends upon the migration of primordial cells. In the Drosophila embryo, the visceral mesoderm (vm) acts as a substrate for the migration of several cell populations of epithelial origin, including the endoderm, the trachea and the salivary glands. These migratory processes require both integrins and laminins. The current model is that αPS1βPS (PS1, the heterodimer of Mew and Myospheroid) and/or αPS3βPS (PS3, the heterodimer of Scab and Mys) integrins are required in migrating cells, whereas αPS2βPS (PS2 the heterodimer of Inflated and Mys) integrin is required in the vm, where it performs an as yet unidentified function. This study shows that PS1 integrins are also required for the migration over the vm of cells of mesodermal origin, the caudal visceral mesoderm (CVM). These results support a model in which PS1 might have evolved to acquire the migratory function of integrins, irrespective of the origin of the tissue. This integrin function is highly specific and its specificity resides mainly in the extracellular domain. In addition, the Laminin α1,2 trimer, was identified as the key extracellular matrix (ECM) component regulating CVM migration. Furthermore, this study shows that, as it is the case in vertebrates, integrins, and specifically PS2, contributes to CVM movement by participating in the correct assembly of the ECM that serves as tracks for migration (Urbano, 2011).

Integrin cell surface adhesion receptors play an essential role in mediating cell migration during development. During Drosophila embryogenesis, integrins are implicated in the movement of three groups of epithelial cells over the vm. In all cases, migration requires the function of PS1, and in some cases PS3, in migrating cells and an unknown function for PS2 in the visceral mesoderm. This study shows that PS1 and PS2 integrins are also required for the migration over the vm of cells of mesodermal origin, the longitudinal visceral muscle progenitors or CVM. PS1 is expressed and required in migrating CVM cells and this function is highly specific, as it cannot be replaced by PS2. Furthermore, this specificity lies mainly in the extracellular domain. In addition, the results revealed that a function of PS2 in the vm is to assemble an ECM substrate necessary for migration. This function of PS2 is not specific, as it can be substituted by PS1 (Urbano, 2011).

Previous analysis of the capacities of PS1 and PS2 integrins to substitute for each other in several cellular processes have revealed specific requirements for the different functional abilities of the two integrins. In a classical view, the diversification of the function of a progenitor gene may occur through the partitioning of regulatory sequences in the original duplication event. The differential tissue expression would then lead to functional refinement and diversification. Phylogenetic analyses of α integrins support this model for the PS1 and PS2 families (Hughes, 2001). An ancestral single α subunit would have duplicated to give rise to αPS2, which retained mesodermal regulatory sequences and evolved to perform stable adhesions, and αPS1, which retained ectodermal and endodermal regulatory elements, and became specialized in mediating dynamic adhesions (Martin-Bermudo, 1997). The current results showing that αPS1 is expressed in a mesodermal derivative suggest that the scenario might not be that simple. In an alternative model, protein functional divergence may have preceded changes in expression patterns. After the duplication event, rapid changes in functionally important sequences of the protein may have gradually led to functional divergence. Then, the two duplicate genes might have undergone degeneration of some of their cis-regulatory motifs. In this context, αPS1 would have diverged to perform a migratory function. Finally, coding-sequence and expression divergence between αPS2 and αPS1 could have been coupled (Urbano, 2011).

This study shows that during cell migration, as it is the case for most integrin-dependent developmental events, PS1 and PS2 cannot substitute for each other. The results demonstrate that the specificity resides mainly in the extracellular domain. PS1 and PS2 integrins show distinct binding specificity for all ligands identified to date. Thus, one way to explain why the αPS2 extracellular domain cannot substitute the αPS1 could be that the interaction between PS2 and its own ligand might promote an adhesion that it is not appropriate for cell migration. The cytoplasmic domain of the α subunits plays important roles in specifying responses upon ligand-binding. For example, while α5 and α2 cytoplasmic domains support collagen gel contraction, α4 cytoplasmic domain promotes cell migration on collagen. The demonstration that the αPS2 cytoplasmic tail cannot fully replace the αPS1 tail suggests that during cell migration the cytoplasmic domains may also participate by transmitting distinct intracellular responses. Alternatively, the downstream effectors of PS2 may not be present in migrating cells. Finally, as PS2 has been shown to promote cell spreading on tiggrin, a third explanation could be that the ECM components that serve as PS2 ligands for migration are not present on the vm (Urbano, 2011).

In contrast, this study has shown that αPS2 can be replaced by αPS1 in the vm, as it is the case during the migration of the tracheal cells. This result suggests that PS integrin function during the assembly of an ECM substrate necessary for cell migration does not require specificity of the α subunit. This is also the case for integrin function on the regulation of gene expression in the embryo (Urbano, 2011).

During amphibian gastrulation, integrins contribute to cell movement in several ways. They are required in migrating cells to provide traction and to transmit guidance signals and in the surrounding cells to assemble an ECM substrate that serves as tracks for migration. In Drosophila, integrins have been shown to be required for the assembly of ECM components in several developmental processes, including dorsal closure, wing imaginal disc morphogenesis and maintenance of the stem-cell niche in the gonads. This study shows that in flies, as it is the case in vertebrates, integrins, and in particular PS2, contribute to cell migration by assisting in the assembly of an ECM substrate. PS2, and not PS1, mutant embryos show irregularities in the visceral mesoderm. Thus, there are at least two possible ways in which PS2 could function in the assembly of an ECM over the vm. PS2 could be required for proper morphogenesis of the vm, which in turn could be necessary for proper expression and/or localization of ECM components. Alternatively, PS2 could directly assemble an ECM essential for both cell migration and proper vm morphogenesis (Urbano, 2011).

Experiments in different model systems have demonstrated the importance of the ECM for the migration of different cell populations. It has been shown that removal of all laminins in the Drosophila embryo affects the migration of several cell populations over the vm. This study shows that laminins are also required for the migration of CVM cells. Furthermore, the results demonstrate that lamininW (Wing blister) is the main ECM molecule supporting CVM migration, and that lamininA, perlecan and collagen IV appear of lesser importance. The migration defects observed in embryos mutant for lamininW are more severe than those found in embryos lacking both the PS integrins and Dystroglycan (Dg). One explanation for this result is that other laminin receptors might be required for CVM migration. Alternatively, laminin function during CVM migration may not only be to provide an ECM substratum, but also to present binding sites for guidance cues necessary for CVM migration. Recent studies in Drosophila showing genetic interactions between laminins and the secreted guidance cue Slit during embryonic cardiac cell migration and axons across the midline support this option. This study has also found that Nidogen (Ndg), a glycoprotein that forms a non-covalent complex with laminin and collagen IV, accumulates over the vm at the time of CVM migration. Besides a structural role in the generation and maintenance of basement membranes, Ndg, free of laminins, has its own biological functions that include cell migration. Ndg is required for the migration of trophoblasts cells, neutrophils and Schwann cells. Interestingly, α3β1 integrins, which is most similar to αPS1βPS, seem to mediate this non-structural function of Ndg. A role for Drosophila Ndg in cell migration awaits the isolation of mutants in this gene. Finally, it was shown that the main source of ECM molecules in Drosophila, hemocytes and fat body, do not provide the ECM molecules required for CVM migration, suggesting that there might be an alternative source. The alternative source could be either the vm itself, as lamininW has been detected in this tissue in stage 11 embryos, or the CVM themselves. This is the case for human keratinocytes, which deposit laminin 332 to promote their linear migration (Urbano, 2011).

During embryonic development, many cells use as their migratory routes ECM components found in the interstices and/or basement membrane surrounding different tissues. The interactions between moving cells and ECM components has to be highly coordinated to guide cells towards their final destinations during the process of organogenesis. The understanding of how this is regulated is still rather fragmentary due to the complexity of the cellular and molecular interactions involved. It was recently shown that in the Drosophila embryo, ECM components, such as laminins, are required for the migration of many cell populations, including endoderm, macrophages, salivary glands, trachea and mesodermal cells. Thus, the power of Drosophila as a model system can be used to gain insights into the developmental roles of the cell-ECM interactions during cell migration. In addition, migrating tumour cells destroy and consequently rearrange the ECM that surrounds them in order to promote proliferation and tumour invasion. Thus, a detailed and comprehensive analysis of the role of the different ECM molecules and their regulation during cell migration is important to further understand not only embryonic development but also tumor metastasis (Urbano, 2011).

Tenectin recruits integrin to stabilize bouton architecture and regulate vesicle release at the Drosophila neuromuscular junction

Assembly, maintenance and function of synaptic junctions depend on extracellular matrix (ECM) proteins and their receptors. This study reports that Tenectin (Tnc), a Mucin-type protein with RGD motifs, is an ECM component required for the structural and functional integrity of synaptic specializations at the neuromuscular junction (NMJ) in Drosophila. Using genetics, biochemistry, electrophysiology, histology and electron microscopy, this study shows that Tnc is secreted from motor neurons and striated muscles and accumulates in the synaptic cleft. Tnc selectively recruits alphaPS2/betaPS integrin at synaptic terminals, but only the cis Tnc/integrin complexes appear to be biologically active. These complexes have distinct pre- and post-synaptic functions, mediated at least in part through the local engagement of the spectrin-based membrane skeleton: the presynaptic complexes control neurotransmitter release, while postsynaptic complexes ensure the size and architectural integrity of synaptic boutons. This study reveals an unprecedented role for integrin in the synaptic recruitment of spectrin-based membrane skeleton (Wang, 2018).

The extracellular matrix (ECM) and its receptors impact every aspect of neuronal development, from axon guidance and migration to formation of dendritic spines and neuromuscular junction synaptic junctions and function. The heavily glycosylated ECM proteins provide anchorage and structural support for cells, regulate the availability of extracellular signals, and mediate intercellular communications. Transmembrane ECM receptors include integrins, syndecans and the dystrophin-associated glycoprotein complex. Integrins in particular are differentially expressed and have an extensive repertoire, controlling multiple processes during neural development. In adults, integrins regulate synaptic stability and plasticity. However, integrin roles in synapse development have been obscured by their essential functions throughout development. How integrins are selectively recruited at synaptic junctions and how they engage in specific functions during synapse development and homeostasis remain unclear (Wang, 2018).

One way to confer specificity to ECM/integrin activities is to deploy specialized ECM ligands for the synaptic recruitment and stabilization of selective heterodimeric integrin complexes. For example, at the vertebrate NMJ, three laminins containing the β2 subunit (laminin 221, 421 and 521, that are heterotrimers of α2/4/5, β2 and γ1 subunits) are deposited into the synaptic cleft and basal lamina by skeletal muscle fibers and promote synaptic differentiation. However, only laminin 421 interacts directly with presynaptic integrins containing the α3 subunit and anchors a complex containing the presynaptic Cavα and cytoskeletal and active zone-associated proteins. Studies with peptides containing the RGD sequence, recognized by many integrin subtypes, have implicated integrin in the morphological changes and reassembly after induction of long-term potentiation (LTP). Several integrin subunits (α3, α5, α8, β1 and β2) with distinct roles in the consolidation of LTP have been identified, but the relevant ligands remain unknown (Wang, 2018).

Drosophila neuromuscular junction (NMJ) is a powerful genetic system to examine the synaptic functions of ECM components and their receptors. In flies, a basal membrane surrounds the synaptic terminals only in late embryos; during development, the boutons 'sink' into the striated muscle, away from the basal membrane. The synaptic cleft relies on ECM to withstand the mechanical tensions produced by the muscle contractions. The ECM proteins, including laminins, tenascins/teneurins (Ten-a and -m) and Mind-the-gap (Mtg), interact with complexes of five integrin subunits (αPS1, αPS2, αPS3, βPS, and βν). The αPS1, αPS2 and βPS subunits localize to pre- and post-synaptic compartments and have been implicated in NMJ growth. The αPS3 and βν are primarily presynaptic and control activity-dependent plasticity. The only known integrin ligand at the fly NMJ is Laminin A, which is secreted from the muscle and signals through presynaptic αPS3/βν and Focal adhesion kinase 56 (Fak56) to negatively regulate the activity-dependent NMJ growth. Teneurins have RGD motifs, but their receptor specificities remain unknown. Mtg secreted from the motor neurons influences postsynaptic βPS accumulation, but that may be indirectly due to an essential role for Mtg in the organization of the synaptic cleft and the formation of the postsynaptic fields. The large size of these proteins and the complexity of ECM-integrin interactions made it difficult to recognize relevant ligand-receptor units and genetically dissect their roles in synapse development (Wang, 2018).

This study reports the functional analysis of Tenectin (Tnc), an integrin ligand secreted from both motor neurons and muscles; Tnc accumulates at synaptic terminals and functions in cis to differentially engage presynaptic and postsynaptic integrin. tnc, which encodes a developmentally regulated RGD-containing integrin ligand, in a screen for ECM candidates that interact genetically with neto, a gene essential for NMJ assembly and function. This study found that Tnc selectively recruits the αPS2/βPS integrin at synaptic locations, without affecting integrin anchoring at muscle attachment sites. Dissection of Tnc functions revealed pre- and postsynaptic biologically active cis Tnc/integrin complexes that function to regulate neurotransmitter release and postsynaptic architecture. Finally, the remarkable features of this selective integrin ligand were explored to uncover a novel synaptic function for integrin, in engaging the spectrin-based membrane skeleton (Wang, 2018).

The ECM proteins and their receptors have been implicated in NMJ development, but their specific roles have been difficult to assess because of their early development functions and the complexity of membrane interactions they engage. This study has shown that Tnc is a selective integrin ligand that enables distinct pre- and post-synaptic integrin activities mediated at least in part through the local engagement of the spectrin-based cortical skeleton. First, Tnc depletion altered NMJ development and function and correlated with selective disruption of αPS2/βPS integrin and spectrin accumulation at synaptic terminals. Second, manipulation of Tnc and integrin in neurons demonstrated that presynaptic Tnc/integrin modulate neurotransmitter release; spectrin mutations showed similar disruptions of the presynaptic neurotransmitter release. Third, postsynaptic Tnc influenced the development of postsynaptic structures (bouton size and SSR complexity), similar to integrin and spectrin. Fourth, presynaptic Tnc/integrin limited the accumulation and function of postsynaptic Tnc/integrin complexes. Fifth, secreted Tnc bound integrin complexes at cell membranes, but only the cis complexes were biologically active; trans Tnc/integrin complexes can form but cannot function at synaptic terminals and instead exhibited dominant-negative activities. These observations support the model that Tnc is a tightly regulated component of the synaptic ECM that functions in cis to recruit αPS2/βPS integrin and the spectrin-based membrane skeleton at synaptic terminals and together modulate the NMJ development and function (Wang, 2018).

Tnc appears to fulfill unique, complementary functions with the other known synaptic ECM proteins at the Drosophila NMJ. Unlike Mtg, which organizes the active zone matrix and the postsynaptic domains, Tnc does not influence the recruitment of iGluRs and other PSD components. LanA ensures a proper adhesion between the motor neuron terminal and muscle and also acts retrogradely to suppress the crawling activity-dependent NMJ growth. The latter function requires the presynaptic βν integrin subunit and phosphorylation of Fak56 via a pathway that appears to be completely independent of Tnc. Several more classes of trans-synaptic adhesion molecules have been implicated in either the formation of normal size synapses, for example Neurexin/Neuroligin, or in bridging the pre- and post-synaptic microtubule-based cytoskeleton, such as Teneurins. However, genetic manipulation of Tnc did not perturb synapse assembly or microtubule organization, indicating that Tnc functions independently from these adhesion molecules. Instead, Tnc appears to promote expression and stabilization of αPS2/βPS complexes, which in turn engage the spectrin-based membrane skeleton (SBMS) at synaptic terminals. On the presynaptic side these complexes modulate neurotransmitter release. On the postsynaptic side, the Tnc-mediated integrin and spectrin recruitment modulates bouton morphology. A similar role for integrin and spectrin in maintaining tissue architecture has been reported during oogenesis; egg chambers with follicle cells mutant for either integrin or spectrin produce rounder eggs (Wang, 2018).

The data are consistent with a local function for the Tnc/βPS-recruited SBMS at synaptic terminals; this is distinct from the role of spectrin in endomembrane trafficking and synapse organization. Embryos mutant for spectrins have reduced neurotransmitter release, a phenotype shared by larvae lacking presynaptic Tnc or βPS integrin. However, Tnc perturbations did not induce synapse retraction and axonal transport defects as seen in larvae with paneuronal α- or β- spectrin knockdown. Spectrins interact with ankyrins and form a lattice-like structure lining neuronal membranes in axonal and interbouton regions. This study found that Tnc manipulations did not affect the distribution of Ankyrin two isoforms (Ank2-L and Ank2-XL) in axons or at the NMJ; also loss of ankyrins generally induces boutons swelling, whereas Tnc perturbations shrink the boutons and erode bouton-interbouton boundaries. Like tnc, loss of spectrins in the striated muscle shows severe defects in SSR structure. Lack of spectrins also disrupts synapse assembly and the recruitment of glutamate receptors. In contrast, manipulations of tnc had no effect on PSD size and composition. Instead, tnc perturbations in the muscle led to boutons with altered size and individualization and resembled the morphological defects seen in spectrin tetramerization mutants, spectrin^R22S. spectrin^R22S mutants have more subtle defects than tnc, probably because spectrin is properly recruited at NMJs but fails to crosslink and form a cortical network. Spectrins are also recruited to synaptic locations by Teneurins, a pair of transmembrane molecule that form trans-synaptic bridges and influence NMJ organization and function. Drosophila Ten-m has an RGD motif; this study found that βPS levels were decreased by 35% at ten-mMB mutant NMJs. Thus, Ten-m may also contribute to the recruitment of integrin and SBMS at the NMJ, a function likely obscured by the predominant role both play in cytoskeleton organization (Wang, 2018).

Previous work has shown that α-Spectrin is severely disrupted at NMJs with suboptimal levels of Neto, such as neto¹⁰⁹- a hypomorph with 50% lethality. These mutants also had sparse SSR, reduced neurotransmitter release, as well as reduced levels of synaptic βPS. In this genetic background, lowering the dose of tnc should further decrease the capacity to accumulate integrin and spectrin at synaptic terminals and enhance the lethality. This may explain the increased synthetic lethality detected in the genetic screen (Wang, 2018).

In flies or vertebrates, the ECM proteins that comprise the synaptic cleft at the NMJ are not fully present when motor neurons first arrive at target muscles. Shortly thereafter, the neurons, muscles and glia begin to synthesize, secrete and deposit ECM proteins. At the vertebrate NMJ, deposition of the ECM proteins forms a synaptic basal lamina that surrounds each skeletal myofiber and creates a ~ 50 nm synaptic cleft. In flies, basal membrane contacts the motor terminal in late embryos, but is some distance away from the synaptic boutons during larval stages. Nonetheless, the NMJ must withstand the mechanical tensions produced by muscle contractions. The current data suggest that Tnc is an ideal candidate to perform the space filling, pressure inducing functions required to engage integrin and establish a dynamic ECM-cell membrane network at synaptic terminals. First, Tnc is a large mucin with extended PTS domains that become highly O-glycosylated, bind water and form gel-like complexes that can extend and induce effects similar to hydrostatic pressure. In fact, Tnc fills the lumen of several epithelial tubes and forms a dense matrix that acts in a dose-dependent manner to drive diameter growth. Second, the RGD and RGD-like motifs of Tnc have been directly implicated in αPS2/βPS-dependent spreading of S2 cells (Fraichard, 2010). Third, secreted Tnc appears to act close to the source, presumably because of its size and multiple interactions. In addition to the RGD motifs, Tnc also contains five complete and one partial vWFC domains, that mediate protein interactions and oligomerization in several ECM proteins including mucins, collagens, and thrombospondins. The vWFC domains are also found in growth factor binding proteins and signaling modulators such as Crossveinless-2 and Kielin/Chordin suggesting that Tnc could also influence the availability of extracellular signals. Importantly, Tnc expression is hormonally regulated during development by ecdysone (Fraichard, 2010). Tnc does not influence integrin responsiveness to axon guidance cues during late embryogenesis; unlike integrins, the tnc mutant embryos have normal longitudinal axon tracks. Instead, Tnc synthesis and secretion coincide with the NMJ expansion and formation of new bouton structures during larval stages. Recent studies have reported several mucin-type O-glycosyltransferases that modulate integrin signaling and intercellular adhesion in neuronal and non-neuronal tissues, including the Drosophila NMJ. Tnc is likely a substrate for these enzymes that may further regulate Tnc activities (Wang, 2018).

In flies as in vertebrates, integrins play essential roles in almost all aspects of synaptic development. Early in development, integrins have been implicated in axonal outgrowth, pathfinding and growth cone target selection. In adult flies, loss of αPS3 integrin activity is associated with the impairment of short-term olfactory memory. In vertebrates, integrin mediates structural changes involving actin polymerization and spine enlargement to accommodate new AMPAR during LTP, and 'lock in' these morphological changes conferring longevity for LTP. Thus far, integrin functions at synapses have been derived from compound phenotypes elicited by use of integrin mutants, RGD peptides, or enzymes that modify multiple ECM molecules. Such studies have been complicated by multiple targets for modifying enzymes and RGD peptides and by the essential functions of integrin in cell adhesion and tissue development (Wang, 2018).

In contrast, manipulations of Tnc, which affects the selective recruitment of αPS2/βPS integrin at synaptic terminals, have uncovered novel functions for integrin and clarified previous proposals. This study demonstrated that βPS integrin is dispensable for the recruitment of iGluRs at synaptic sites and for PSD maintenance. An unprecedented role was reveaked for integrin in connecting the ECM of the synaptic cleft with spectrin, in particular to the spectrin-based membrane skeleton. These Tnc/integrin/spectrin complexes are crucial for the integrity and function of synaptic structures. These studies uncover the ECM component Tnc as a novel modulator for NMJ development and function; these studies also illustrate how manipulation of a selective integrin ligand could be utilized to reveal novel integrin functions and parse the many roles of integrins at synaptic junctions (Wang, 2018).

Live imaging reveals hub cell assembly and compaction dynamics during morphogenesis of the Drosophila testis niche

Adult stem cells are often found in specialized niches, where the constituent cells direct self-renewal of their stem cell pool. The niche is therefore crucial for both normal homeostasis and tissue regeneration. In many mammalian tissues, niche cells have classically been difficult to identify, which has hampered any understanding of how tissues first construct niches during development. Fortunately, the Drosophila germline stem cell (GSC) niche is well defined, allowing for unambiguous identification of both niche cells and resident stem cells. The testis niche: this study followed pro-niche cells as they assemble and assume their final form. After ex vivo culture the niche appears fully functional, as judged by enrichment of adhesion proteins, the ability to activate STAT in adjacent GSCs, and to direct GSCs to divide orthogonally to the niche, just as they would in situ. Collectively, imaging has generated several novel insights on niche morphogenesis that could not be inferred from fixed images alone. Dynamic processes were identified that constitute an assembly phase and a compaction phase during morphogenesis. The compaction phase correlates with cell neighbor exchange among the assembled pro-niche cells, as well as a burst of divisions among newly recruited stem cells. Before compaction, an assembly phase involves the movement of pro-niche cells along the outer periphery of the gonad, using the extracellular matrix (ECM) to assemble at the anterior of the gonad. Finally, live-imaging in integrin mutants allows definition of the role of pro-niche cell-ECM interaction with regard to the new assembly and compaction dynamics revealed in this study (Anllo, 2018).

This work is the first to characterize the dynamics of testis niche morphogenesis. Through live-imaging both in vivo and ex vivo, it was possible to follow pro-niche cells as they assemble and assume the final compact form of this niche. After ex vivo culture, the niche appears fully functional, as judged by enrichment of adhesion proteins, the ability to activate STAT in adjacent germline cells, and to direct the now GSCs to divide orthogonally to the niche, just as they would in situ. Collectively, imaging has generated several novel insights on niche morphogenesis that could not be inferred from fixed images alone. The compaction phase correlates with cell neighbor exchange among the assembled pro-niche cells, as well as a burst of divisions among newly recruited stem cells. Before compaction occurs, an assembly phase involves the movement of PS11 pro-niche cells along the outer periphery of the gonad, using the ECM to assemble at the anterior of the gonad. There they join PS10 pro-niche cells, forming an anterior and internally-facing cap. Finally, live-imaging in integrin mutants allows definition of the role of pro-niche cell-ECM interaction more precisely with regard to the new assembly and compaction dynamics revealed in this study (Anllo, 2018).

Previously, analysis of fixed samples nicely inferred a change from an initial cap of pro-hub cells to a compact niche. The current ex vivo imaging now visualizes compaction directly, and by comparing cell numbers at assembly and after compaction, shows that the initial assemblage of the elongate cap of cells indeed constitutes the pro-hub. Scoring the same gonad before and after compaction showed little, if any, shuffling in the number of pro-niche cells. This finding is consistent with the fact that specification of niche cells occurs much earlier in development, before the stages imaged in this study. Lastly, the current work suggests two cell biological phenomena that correlate with the process of compaction (Anllo, 2018).

Firstly, while compaction is underway, there is a burst of GSC divisions orthogonal to the hub. Interestingly, at late anaphase, the niche cells adjacent to a dividing GSC occasionally appeared to become compressed against their niche cell neighbors, subsequently occupying less space. It is known that activation of the STAT pathway is necessary to induce germline divisions in the (male) gonad. Since the ligand for pathway activation is secreted by hub cells, this suggests a possible scenario where niche cells orient GSC divisions, which subsequently generate some of the force necessary for niche compaction, pushing hub cells closer together. In principle, this scenario can now be tested by attempting to block GSC divisions during ex vivo culture (Anllo, 2018).

Secondly, hub compaction occurs concurrently with cell neighbor exchanges. These exchanges are reminiscent of convergence and extension, where directed junctional re-organization changes the overall shape of the sheet of cells. For example, at assembly, an individual cell within the elongate cap will shrink one of its boundaries, allowing cells that were previously on either side of this cell to form new contacts adhering to one another. Such events may impart an increasingly more compact overall shape to the incipient niche. In addition to lateral neighbor exchange, ingression was observed of some pro-niche cells during compaction. This results in stratification of the niche, where a few internal pro-hub cells now lie more internal to the superficial cap of peripheral hub cells. These lateral and stratifying exchanges are complex, but their comprehensive cataloging, now enabled by ex vivo imaging, may well reveal patterns of cell movements. In turn, those patterns will likely suggest specific circuits to test as regulators of compaction (Anllo, 2018).

The neighbor exchange behavior associated with compaction is reminiscent of cell sorting, where 'like' cells will maximize contacts among themselves, while minimizing those with different neighbor types. Indeed, imaging describes a significant smoothening of the boundary between the pro-niche and the tier of stem cells, an outcome expected from sorting phenomena. Increased accumulation of adhesion proteins such as ECadherin, NCadherin and Fasciclin III, has been noted at the Assembly stage, although manipulation of these did not appear to grossly affect compaction. While unidentified factors or significant redundancy among these might drive sorting, the live-imaging parameters established here now enable analysis of the dynamics of the process. Future work might reveal changes in, for example, timing of compaction and smoothening, rather than relying on endpoint analysis alone (Anllo, 2018).

The positioning of the niche at the gonad anterior could, in principle, be driven by a gonad-intrinsic mechanism. For example, SGPs across the gonad have different antero-posterior identities, as they originate from three separate parasegments (ps10, ps11, and ps12). Thus, they accumulate Hox proteins to differing levels, and AbdB mutant embryos can exhibit defects in hub position. However, it remains possible that these defects could be caused by broader patterning changes in tissues outside the gonad, rather than reflecting an intrinsic requirement for AbdB. Still, the fact that pro-hub cells are enriched for adhesion proteins from the assembly stage onward is consistent with the idea that sorting driven by homophilic adhesions could be a main driver for niche formation. Nevertheless, two novel observations from live-imaging strongly suggest the existence of an extra-gonadal influence (Anllo, 2018).

First, while the niche assembles at the anterior, it was found that it exhibits a consistent tilt towards the interior of the embryo. This biased placement of the hub towards internal organs suggests that an extra-gonadal cue is influencing niche position. Live-imaging did reveal tissues located near cells of the assembling pro-niche: a branch of the trachea, and an alary muscle. The ablations that were conducted rule out any major impact from these two tissues on niche formation. However, other tissues, such as the nearby gut and its surrounding musculature, remain to be tested (Anllo, 2018).

Second, the individual movements of PS11 cells to join the PS10 cohort suggests that homotypic adhesions and sorting alone cannot explain the position of assembly. During the stages leading up to those shown in this study, SGPs exhibit robust cytoplasmic extensions that probe the internal milieu of the newly formed gonad, tightly embracing each other, and germline neighbors. Indeed, imaging carried out in this study revealed PS11 cellular extensions around germ cells at the onset of assembly. If sorting were the main driver for assembly, these extensions should be sufficient to cause PS11 cells to move through the internal gonad milieu, seek out and coalesce with anteriorly located PS10 SGPs. Instead, PS11 cells move out to the gonad periphery, and retract their trailing extensions. They then move along the periphery towards the anterior. In addition, if PS 11 cells were sorting together one might have expected these pro-niche cells to move as one tight collective, reminiscent of how the somatic border cell cluster migrates through the internal milieu of the Drosophila egg chamber. In contrast, the PS11 cells often move individually or as a string of cells along the periphery. Taken together, PS 11 cell behavior is much more in line with individual cell migration directed by a cue (Anllo, 2018).

Previous work described an ECM visible at the gonad periphery at the end of embryogenesis. Pigment cells ensheath the gonad, and are the likely source of this ECM. Indeed, live-imaging revealed that this ECM is present while PS11 cells are moving anteriorly. Thus, it is suspected that the matrix provides traction for movement in response to the posited external guidance cue (Anllo, 2018).

In fact, previous work has suggested a role for ECM interaction, since in the absence of Integrin function, the newly formed hub was located inside the gonadal sphere (Tanentzapf, 2007). Since that work had to rely on endpoint analysis, it was reasonable to assume that the hub had fully formed at the gonad anterior and then was internalized whole due to lack of Integrin-based adhesion. Instead, the current live-imaging of Integrin mutants now reveals that Integrin-ECM interactions are necessary well before compaction, as pro-niche cells can leave the periphery of the gonad before the niche compacts. This is the case for PS10 cells, which can be observed internalizing from the anterior of the gonad before compaction, and for PS11 cells, either after they arrive at the anterior or directly internalizing from their more posterior, PS11 location. Hub compaction occurs in Integrin mutants (Tanentzapf, 2007), but the current imaging reveals this to occur mostly after the niche internalizes in the mutants. The observation that PS11 cells can directly join an internalized cluster suggests that, left to their own, pro-hub cells might sort to inappropriate locations due to preferential adhesion. This implies that Integrin-ECM interactions at the periphery antagonize the sorting tendency of pro-niche cells. If Integrin-ECM interactions during migration are compromised, the resulting testis would exhibit aberrant polarity, as the niche would not be localized at the closed end of the tubule. Although some of the components making up the ECM surrounding the gonad have been characterized, which of these engage Integrins on the pro-niche cells is not clear (Anllo, 2018).

The nature of the directional cue read by PS11 cells is not known. Imaging shows that some but not all PS11 cells successfully arrive at the anterior in Integrin mutants. This suggests that pro-niche cells have the capability to respond to the cue, at least partially, even if they cannot interact with peripheral ECM. Normally, however, since all the PS11 cells that were tracked in wild type gonads migrated to the anterior along the ECM at the periphery of the gonad, these Integrin-ECM interactions serve to ensure reproducible positioning of the niche (Anllo, 2018).

The combination of in vivo and ex vivo imaging presented in this study overcomes previous challenges to understanding how a stem cell niche is formed. Visualizing the dynamics of niche formation has proven powerful in suggesting unanticipated mechanisms for niche morphogenesis in this work. This imaging should now allow for in-depth study of both the process of assembly and compaction into a fully functioning stem cell niche (Anllo, 2018).

PROTEIN STRUCTURE

Amino Acids - 846

Structural Domains

The gene encoding the Drosophila PS2 alpha subunit is composed of 12 exons spanning 31 kb. By employing a novel method for directed cDNA cloning, over 300 independent cDNA clones were analyzed for the existence of alternate RNA products. Two forms of PS2 alpha mRNA are frequently observed: a canonical (C) form and a form lacking the 75 nucleotide exon 8 (m8). The relative ratio of these two forms varies widely during development. Although region A, derived from exon 8 and the adjacent 25 amino acids, shows weak conservation among the sequences of alpha subunits that bind to different ligands, it is highly conserved in the homologous PS2 alpha gene of the distantly related Mediterranean fruitfly. It is suggested that the variable region A may be important in determining the specificity and affinity of integrin receptors for their ligands (Brown, 1989).

Either of the two alternative spliced forms of the transcript of the mys gene is sufficient to rescue postembryonic mys phenotypes in the wing, eye and muscle but both of the two splice forms are necessary to rescue the mys embryonic defects. The location of the alternative exons suggests that the two forms of the PS beta integrin subunit may interact with alternative alpha subunits and/or ligands (Zusman, 1993).

The Myospheroid protein is an integrin beta subunit of Drosophila. myospheroid cDNA sequence predicts a cysteine-rich integral membrane protein that displays 45% sequence identity to chicken integrin and the human fibronectin receptor beta subunit.

There is a 23 amino acid N-terminal signal sequence and a 23 amino acid hydrophobic membrane-spanning domain starting at amino acid 777. Of the 57 cysteine residues in MYS protein, 56 align with chicken integrin B in four repeated units (MacKrell, 1988). A tyrosine residue at position 834 agrees with consensus sequences of for phosphorylation by protein tyrosine kinases (MacKrell, 1988).

date revised: 5 August 2023

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.