Our work in the past has primarily focused on the roles of cell surface proteins in determining the patterns of synaptic connectivity in the Drosophila (fruit fly) nervous system. In 2013, we published a paper with the Garcia group at Stanford on a global extracellular "interactome" screen, in which we mapped interactions among 200 fly cell surface proteins, identifying more than 80 new interactions. As described below, these interactions have become the major focus of our work in Drosophila. We have also expanded our interactome screening efforts to humans.
The global human interactome screen
We continued our collaboration with the Garcia group, publishing an interactome screen for more than 500 human IgSF cell surface proteins in 2020. The fly and human screens were performed using a modified ELISA method, the ECIA, in which pairwise interactions among proteins are assessed by binding a multimeric form of the extracellular domain (ECD) of one protein (the "bait") to a surface and incubating with a multimeric ECD for the other protein (the "prey") in solution. We wished to expand the IgSF screen to the analysis of all ~2000 human single-transmembrane cell surface proteins, but this would involve assessment of ~5 million pairwise interactions, which would be very time-consuming to do using the ECIA, even with automation. Accordingly, we developed the BPIA method, in which binding of a single prey to hundreds of bait proteins, each attached to a different fluorescent "coded" bead type, can be simultaneously examined in one tube. To increase the speed, sensitivity, and versatility of the assay, we then devised technologies to assemble prey proteins into 60-mer nanoparticles, and to use the same ECD fusion proteins as baits and preys, so that it would not be necessary to express every protein in two forms. These technologies formed the basis for a Transformative Research Grant (TRO1) proposal to the NIH, which was awarded in 2022 and provided the necessary funding to execute the global interactome screen. This is now the major project being conducted in my lab. This is a team effort involving my group, the Protein Expression Center headed by Jost Vielmetter, and the Thomson group at Caltech. The project director is Woj Wojtowicz, who previously conducted the human IgSF screen together with Jost.
The TRO1 also involves assessment of the functions of cell surface protein interactions, by binding these proteins to cells in culture and assessing the consequences of binding using single-cell RNA sequencing. This work is being done together with the Thomson group and the Single-Cell Profiling and Engineering Center. We are primarily using PBMCs, which are human immune system cells from donors.
Determination of synaptic connectivity patterns by cell surface proteins
The Drosophila interactome screen identified two major protein interaction networks. The Dpr-ome is a network of 32 interacting immunoglobulin superfamily (IgSF) proteins, in which 21 Dpr proteins in one IgSF subfamily bind selectively to one or more DIP proteins, which comprise another IgSF subfamily. The Beat-Side interaction network has a similar structure. 14 Beat IgSF proteins bind selectively to 8 Side proteins. Dprs, DIPs, Beats, and Sides are all expressed in subsets of neurons during development.
Since our discovery of the Dpr-ome, we and others have shown that DIP::Dpr interactions control synaptic connectivity and cell survival in the optic lobe and neuromuscular system. Our current work focuses on circuits in the visual brain, known as the optic lobe. We found that interactions between Dpr11 on UV-sensitive photoreceptors and its partner DIP-g on their primary synaptic target, the Dm8 amacrine neuron, control cell survival and synapse formation. In our current work, we are examining how Beat, Side, DIP, and Dpr proteins wire color vision circuits used for detection of both long- and short-wave UV.
In another optic lobe circuit, L3 lamina neurons express Dpr6 and Dpr10, while their targets in the medulla, Dm4 and Dm12, express DIP-a, which binds to both Dpr6 and Dpr10. This DIP::Dpr interaction controls survival of Dm4 and Dm12 and is required for normal Dm12 connectivity. Our work on this circuit has primarily focused on the role of affinity. In a collaboration with the Zipursky (UCLA) and Shapiro/Honig (Columbia) groups, we introduced mutations that decrease or increase the affinity of binding between DIP-a and Dpr10 into the endogenous genes. We observed that affinity reduction increases cell death and affects Dm12 targeting, while increasing affinity saves Dm4 and Dm12 cells that would normally die during development.
Dpr10 is also expressed in larval muscles and interacts with DIP partners on motor neurons. Interestingly, while high affinity is required for Dpr10's functions in the optic lobe, it binds to DIP-b with a much lower affinity, and this low affinity of interaction is necessary for normal development of specific neuromuscular junction (NMJ) synapses. We are currently developing new technologies to visualize DIP::Dpr interactions at synapses, together with synaptic markers, and we hope to be able to simultaneously detect 6 or more interactions at high resolution at optic lobe and NMJ synapses. These technologies, when fully developed, should be widely applicable to protein visualization in both vertebrate and invertebrate tissues.
Finally, we are examining RNA transport between cells in Drosophila mediated by extracellular vesicles (EVs) that contain a virus-like particle called dArc1. We found that EVs bearing dArc1 capsids are targeted to specific cells using a cell surface protein called Stranded at Second, which binds to a receptor tyrosine phosphatase (RPTP). We had previously done extensive work on the six Drosophila RPTPs, showing that they control axon guidance and synapse formation during embryonic and pupal development.