| |
Overview of current research in the Zinn group
3/23/20
Introduction:
Our group has been working on the molecular and cellular mechanisms involved in the determination of synaptic connectivity patterns in the nervous system for many years. Our primary experimental organism is the fruit fly Drosophila melanogaster. Drosophila has many advantages as an experimental system in which to study these problems. It has a complex brain and nervous system, with hundreds of thousands of neurons. Despite this complexity, neurons have individual identities that are determined by the transcription factors that they express, and synaptic connectivity patterns are primarily determined by genetic mechanisms. This means that one can uniquely identify neurons in wild-type and mutant flies and determine which perturbations alter the synapses made by those neurons.
As described below, we are also working on assessing interactions among human cell-surface and secreted proteins, with the goal of creating a complete extracellular interactome map defining in vitro interactions among all ~3000 human proteins that bind specifically to extracellular partners. This work builds on our previous analysis of global extracellular interaction patterns in Drosophila.
Receptor tyrosine phosphatases:
Our primary focus has been on cell surface proteins that mediate interactions among neurons and between neurons and non-neuronal targets such as muscle fibers. We have worked for many years on receptor tyrosine phosphatases (RPTPs), which are cell surface receptors that signal through competition with tyrosine kinases, removing phosphate groups from specific tyrosine residues. RPTPs are widely expressed in neurons, and their loss leads to alterations in axon guidance and synapse formation. We have two ongoing RPTP projects in the lab. Namrata Bali, a postdoc in the lab, has shown that interactions between the LAR RPTP and an IgSF protein called Sticks and Stones (SNS; an ortholog of mammalian Nephrin) control the organization of the mushroom body, a complex neuropil involved in learning and memory. LAR-SNS interactions also determine synapse number for specific neuromuscular junctions in the larval neuromuscular system. Peter Lee, another postdoc, has shown that Stranded at Second (Sas), a large cell surface protein expressed on epithelia, is a ligand for the Ptp10D RPTP. Sas is exported from cells on exosomes and may be a targeting receptor for these exosomes. He performed affinity purification/mass spectrometry (AP/MS) experiments to identify exosomal proteins associated with the Sas cytoplasmic domain, and showed that two of these are the Arc proteins. Arcs are related to retroviral envelope proteins and can form capsid-like structures in vitro. The capsids can carry mRNAs, and are involved in synaptic plasticity in mammals. We are trying to determine whether Sas controls the movement of Arc and its mRNA cargoes between specific cell types.
Dprs and DIPs:
Christopher Garcia’s group at Stanford and our group at Caltech conducted an in vitro screen for interactions among 200 Drosophila cell surface proteins. This led to the definition of the “Dpr-ome”, a complex network of interactions among two subfamilies of immunoglobulin superfamily (IgSF) superfamily proteins, the Dprs (21 members) and the DIPs (11 members). We found that each Dpr and DIPs is expressed on a small and unique subset of neurons in each area of the nervous system during each phase of development. Expression of Dprs and DIPs correlates with synaptic connectivity patterns. In the pupal optic lobe, neurons expressing a particular DIP tend to be postsynaptic to neurons expressing a Dpr to which that DIP binds in vitro. Because there are so many Dpr-DIP interactions, it is impractical for a small research group to study all of these genes at once. Therefore, we have focused on particular Dpr-DIP interactions that occur between synaptically connected cells, and are using genetics and molecular biology to determine how these interactions program the formation and organization of these synapses.
In one project, Kaushiki Menon, a Member of the Professional Staff, examined how interactions between DIP-γ and its partner Dpr11 control assembly of color vision circuits in the pupal optic lobe. Color vision is mediated by two types of neurons in the compound eye: R7 and R8. R7 is a UV receptor, while R8 detects visible light. R7s and R8s are each divided into two types, “yellow” (y) and “pale” (p), which have different rhodopsins and have different wavelength preferences. Dpr11 is expressed only by yR7s, while DIP-γ is expressed by the medulla amacrine neuron Dm8 and by the Tm5a projection neuron, which forms synapses in the lobula, a higher-order visual area. We found that DIP-γ is expressed by a subset of Dm8s (yDm8s), and that these Dm8s receive synapses only from Dpr11-expressing yR7s. This circuit is determined by interactions between Dpr11 and DIP-γ during early pupal development that allow yDm8s to select yR7s as their preferred synaptic partners. During normal development, Dm8s are generated in excess, and those that do not find a presynaptic partner (about 30% in wild-type) die. When Dpr11 or DIP-γ is not expressed, most yDm8s die, because they cannot select a yR7 partner. Those that do survive form abnormal synapses. To understand the structure of this circuit, we examined the electron microscopic (EM) reconstruction of the medulla, and showed that yR7s connect selectively to both yDm8s and to Tm5as, while pR7s connect to pDm8s and to Tm5bs. Of course, the EM reconstruction was only done on wild-type animals. We were able to visualize these synapses in mutant animals using expansion microscopy (ExM), which allowed us to generate views of the neurons that could be superimposed on those generated by EM. In the future, we plan to examine how the selective synaptic connections of yR7s to Tm5as are determined. We are also doing single-cell RNA sequencing of R7s, in order to define how pR7s and pDm8s selectively connect to each other. Finally, we are examining synaptic connections between visual motion-sensitive T4 and T5 neurons expressing Dpr11 and their targets in the lobula plate, a novel set of lobula plate tangential cells (LPTCs) that express DIP-γ.
Interactions between DIP-α and its partner Dpr10 control formation of synapses between motor neurons and muscle fibers in the larval neuromuscular system, and between identified neurons in the pupal optic lobe. In the neuromuscular system, work by a postdoc in my lab, Robert Carrillo, who now has his own lab at University of Chicago, showed that DIP-α is expressed by only two of the 32 motor neurons present in each hemisegment. These two neurons are “1s” neurons that each innervate a set of muscle fibers that all express Dpr10. In the absence of DIP-α or Dpr10, the axons of DIP-α expressing neurons fails to form branches onto a specific set of muscle fibers. In the pupal optic lobe, work done by Shuwa Xu, a postdoc in my lab, while she was in Larry Zipursky’s lab at UCLA showed that DIP-α is expressed on the Dm4 and Dm12 neurons, among others. These neurons are postsynaptic to neurons expressing Dpr10. When DIP-α or Dpr10 (together with its closely related paralog Dpr6) is lost, Dm12 neurons form abnormal ectopic synapses in the wrong layer of the 10-layered neuropil called the medulla. Ectopic expression of Dpr10 in a medulla layer forces Dm12 and Dm4 to form synapses in that layer, showing that DIP-α::Dpr10 interactions can define synaptic connectivity patterns. As with DIP-γ::Dpr11, these interactions are also required for cell survival; in the absence of either DIP-α or Dpr10 and Dpr6, some Dm4 neurons die. Shuwa is now examining how the affinities of interactions between DIP-α and Dpr10 program these patterns. Her preliminary results show that formation of synaptic connections requires affinities in the low micromolar range, while cells can still survive even when interactions are much weaker. Thus, different functions require different affinities.
We have other projects involving DIP-β::Dpr10 interactions in the neuromuscular system, and DIP-δ::Dpr12 interactions in chemosensory systems.
The Human Interactome Project (HIP):
We used the strategies created for the Drosophila interactome project to conduct a screen for interactions among 556 human cell surface proteins, including all IgSF proteins. Because the affinities of interactions among cell surface proteins are typically low (micromolar or weaker), they do not form stable complexes, and interactions cannot be detected by methods such as AP/MS. To detect such interactions requires taking advantage of avidity effects by clustering the proteins into multimeric complexes. For the Drosophila screen, we developed a method called the Extracellular Interactome Assay (ECIA). The extracellular domain (XCD) of each protein was fused to both human Fc, a dimerizing agent, and to a pentameric version of placental alkaline phosphatase (AP5). XCD-Fc proteins (“baits”) are bound to wells of ELISA plates and incubated with each AP5 “prey” protein, and binding is detected using a colorimetric assay for AP. For the human screen, Woj Wojtowicz, the project director within the Garcia group, and Jost Vielmetter, the director of the Caltech Protein Expression Center (PEC) adapted the ECIA to liquid-handling robots. The HIP screen identified hundreds of new interactions, which define potential new functions for the Deleted in Colon Cancer (DCC) subfamily, RPTPs, TrkA and TIE1 receptor tyrosine kinases, and Leptin, among others.
While the IgSF HIP was successful, to conduct a much larger screen of the entire human extracellular repertoire new methods need to be developed. We need higher-throughput screening methods, and also need to increase sensitivity. To address these issues, Mike Anaya, a graduate student in my lab, is developing a radically different methodology, called Modular Enzymatic Barcoding, which may allow us to convert the labor-intensive ECIA assay, in which each bait-prey interaction is assessed in a separate reaction (so that the number of assays required expands as the square of the number of proteins to be examined), to a high-throughput DNA sequencing assay that could be performed on a single reaction containing a mixture of all XCDs. Mike, together with Matt Thomson, Sisi Chen, and Jeff Park of the Caltech Single-cell Sequencing Facility, is also developing other uses for the enzymatic barcoding technology. To increase sensitivity, we are developing methods to use larger multimeric particles rather than dimers and pentamers. We have made 60-mer XCD particles and are investigating whether these allow us to detect interactions that were too weak to see using the ECIA. |
|