PerspectiveCell Signaling

Liquidity in immune cell signaling

See allHide authors and affiliations

Science  29 Apr 2016:
Vol. 352, Issue 6285, pp. 516-517
DOI: 10.1126/science.aaf8179

Tlymphocytes of the immune system need to integrate myriad signals to decide on life and death matters in defending the host from pathogens and in distinguishing normal from transformed cells, all while minimizing immunopathology. On page 595 in this issue, Su et al. (1) provide insight into how phase separation of signaling proteins in the two-dimensional (2D) liquid of the plasma membrane enables these critical decisions.

Phase separation in live cells has become a useful concept in explaining the formation of µm-scale spherical, liquid-like compartments that are held together by weak, rapidly reversible interactions and that exchange components with the surrounding bulk phase (2). Phase separation is also well studied in 2D systems such as biological membranes, where coexisting liquid disordered and liquid ordered phases are important for signaling by many transmembrane receptor systems (3), including the T cell receptor (TCR) signaling process. In response to binding antigen peptides, the TCR triggers a series of biochemical events orchestrated by various enzymes and adaptor proteins. This signaling cascade impinges on the nucleus and the actin cytoskeleton to alter T cell behavior. Su et al. examined how soluble and membrane-anchored adapter proteins downstream of the TCR form a 2D phase–separated system that depends fully on protein-protein interactions on top of a liquid disordered bilayer.

Su et al. looked at the behavior of a signaling effector of the TCR called linker of activated T cells (LAT), and two of its binding partners—growth factor receptor–bound protein 2 (Grb2), which binds to phosphorylated LAT; and son of sevenless (Sos), which binds to the N- and C- termini of Grb2 (1). This protein trio is critical for activating the signaling protein Ras as part of the TCR signaling pathway. The authors found that physiological amounts of bilayer-anchored prephosphorylated LAT (pLAT), plus soluble Grb2 and Sos proteins, formed micrometer-scale pLAT clusters on a supported lipid bilayer in vitro. The pLAT clusters have characteristics of a liquid phase in that they fuse with each other while maintaining a roughly round shape and allow rapid exchange of pLAT between a cluster and surrounding unclustered pLAT. Su et al. expanded the system of purified proteins to include kinases and phosphatases that act downstream of the TCR signaling cascade and generate pLAT. The authors also overlaid an actin polymerizing system on the pLAT clusters, which consisted of globular actin (G-actin) and five proteins: Grb2-related adapted downstream of Shc (Gads), SH2 contains leukocyte protein of 76 kDa (SLP-76), noncatalytic region of tyrosine kinase adapted protein 1 (Nck), actin-related protein 2/3 (Arp2/3), and neuronal Wiskott-Aldrich syndrome protein (N-WASP). The 2D pLAT clusters could interact with the actin-nucleating factor N-WASP to trigger formation of an elongated filamentous actin (F-actin) gel (see the figure). The F-actin gel appears to act like a solid container that absorbs liquid-phase pLAT and takes the shape of the F-actin.

Fluid signals.

Signaling molecules (blue) form liquid-like phases, instead of static aggregates, in a reconstituted system using an artificial lipid bilayer. Shown is a model of clustered pLAT and associated signaling molecules that phase separate into a liquid-like droplet in response to TCR activation. The cluster induces actin polymerization, which is a solid phase gel that changes the shape of the liquid-like droplet.


The pLAT clusters are likely to be components of TCR “microclusters,” where antigenic major histocompatibility complex-peptide ligand (antigen) are engaged and signaling reactions occur on the cytoplasmic face of the T cell plasma membrane. These microclusters are highly dynamic and display fusion events to form larger clusters (4). TCR microcluster formation in response to physiological ligands depends on F-actin remodeling. It is possible that the complex interaction between the pLAT-rich liquid phase and F-actin is related to this requirement—that is, initial TCR triggering could induce pLAT phase separation and signal actin remodeling, which would feed forward to promote further TCR interaction with antigen.

Concepts of 2D phase separation can also be applied to reconstituted cell adhesion systems. The adhesion of T cells that express the receptor CD2 to an artificial lipid bilayer bearing its ligand CD58, displays characteristics of a phase-separated system with the precisely apposed membranes contributing the necessary dynamic multipoint attachments (5). Furthermore, different size receptor systems that bridge the gap between the T cell and artificial lipid bilayer further subdivide the interface into TCR and adhesion molecule–rich domains of the immunological synapse (the interface between the antigen-presenting cell and the T cell) (6), the natural setting for the reactions studied by Su et al. It may be that the actinomyosin–based transport system, which drives centripetal transport in the immunological synapse, moves the TCR and pLAT clusters by using the F-actin-based “container” that was observed by the authors in their reconstituted system.

Su et al. focused on phenomena largely confined to the 2D supported lipid bilayer surface by tethering key components to the bilayer's surface. It is not clear whether 2D phase separation of pLAT could also nucleate a 3D phase–separated structure that would be released into the cytoplasm. Such a structure appears to be formed during T cell stimulation by immobilized antibodies to TCR, where foci (containing SLP-76) are released as particles from TCR microclusters and trafficked toward the microtubule-organizing center (7). These particles seem to dissipate in the center of the immunological synapse, consistent with reversible phase separation.

The findings of Su et al. provide insights into characteristics of TCR signaling that are otherwise hard to reconcile. Recent superresolution microscopy studies suggest that substrate-adherent T cells display basal segregation of TCR and pLAT into distinct “protein islands” on a ~100- to 300-nm length scale (8). It is possible that interaction between layered dynamic liquid phases could govern signal integration. Segregation of TCR clusters from pLAT may be mediated by the tendency of the TCR to reside in liquid disordered bilayer phases, whereas LAT, which is palmitoylated, resides in liquid ordered domains. These interactions between phases may also govern important regulators like the transmembrane tyrosine phosphatase CD45. The large extracellular domain of CD45 excludes it from TCR microclusters, which is important for the phosphorylating tyrosine kinase cascades proximal to the TCR (9). Su et al. found that CD45 is also excluded from pLAT clusters, likely due to charge repulsion between the phosphatase domain and pLAT cluster components. It may be important that CD45 is excluded in this way; segregation of TCR clusters and pLAT clusters suggests that extracellular domain size does not exclude CD45 from the pLAT clusters. The tools generated by Su et al. have the potential to further address underlying principles governing the dynamic, layered circuitry of transmembrane signaling.


View Abstract

Stay Connected to Science

Navigate This Article