To successfully achieve the project objectives outlined in the previous section, we employed a wide range of methodologies, spanning molecular biology, gene engineering, proteomics, and advanced imaging techniques.
First, we validated that TNS1 undergoes LLPS in human osteosarcoma cancer cell line by employing (live) cell imaging and electron microscopy. We characterized the behaviour of TNS1 condensates in line with the guidelines established in the field and we confirmed that TNS1 undergoes LLPS in a concentration-dependent manner. To assess TNS1 LLPS at endogenous expression levels, we generated a cell line with endogenously tagged TNS1. Using this model, we demonstrated that TNS1 condensation in these cells is regulated by availability of integrin ligands and by the ability of actomyosin machinery to generate contractile forces within the cell.
We next focused on dissecting the molecular composition of TNS1 condensates using a proteomic approach based on enzymatic proximity labelling (BioID) in cells with either low or high expression of TNS1. As a result, we were able to identify LLPS-dependent interactome of TNS1 and subsequent validation of the screen results revealed a selective recruitment of several integrin adhesion components to TNS1 condensates. Further analysis showed that the recruited proteins were in their inactive state, suggesting that TNS1 condensates serve as reservoirs for inactive focal adhesion components.
We then turned our attention to understanding the mechanistic determinants of TNS1 LLPS and their functional implications for cancer cell behaviour. First, we employed protein mutagenesis to identify the regions contributing to formation of TNS1 condensates. There are four structurally conserved domains in TNS1, two of them located at the N-terminus and two at the C-terminus of the protein. They are interconnected by a long intrinsically disordered region (IDR) lacking any defined secondary structure. Our experimental data demonstrated that the IDR is the primary driver of TNS1 LLPS. These findings were further supported with coarse-grained molecular dynamics simulations. On a functional level, deletion of IDR led to a significant increase in integrin adhesion dynamics when compared to wild-type TNS1, leading to enhanced cell migration and invasion into 3D extracellular matrix. The IDR deletion mutant also exhibited increased formation of cell protrusions under both 2D and 3D conditions, consistent with the observed increase in invasiveness.
Finally, we investigated the role of phosphorylation in regulating TNS1 LLPS and its effects on integrin adhesion dynamics. We showed that cellular stress negatively regulates TNS1 LLPS, correlating with increased TNS1 phosphorylation and activation of key stress-responsive kinases, including p38, ERK and Akt. To identify the respective phosphorylation sites involved in TNS1 LLPS regulation, we performed phosphoproteomic analysis of TNS1. Subsequent validation of the proteomic data confirmed that TNS1 phosphorylation negatively regulates formation of TNS1 condensates. We also demonstrated that TNS1 phosphorylation has a dramatic effect on integrin adhesion assembly and disassembly rates and significantly affects integrin adhesion size and maturation.
Taken together, in this project we deciphered the molecular determinants underlying TNS1 LLPS in living cells. We have also established the role of TNS1 condensation in regulating integrin adhesion dynamics and maintaining inactive pools of integrin adhesion components, thereby introducing a novel regulatory layer in integrin adhesion signalling.
