Rapid Rerouting of Myosin Traffic at the T Cell Immunological Synapse

Cytoskeletal motors travel in patterns set by the architecture of their tracks. Nevertheless, we have a limited understanding of how cells dynamically reorganize their traffic patterns in response to signaling events. To investigate cytoskeletal motor rerouting, we used T cells as a model system. Upon an encounter between a T cell and an antigen presenting cell, the T cell builds a specialized interface with spatially organized immunoreceptors and adhesion molecules called the immunological synapse (IS). The IS also constructs new actin networks within minutes that define the synaptic structure. Here we track the movements of single myosin motors along presynaptic and synaptic actin networks of the T cell. We find that both myosin-5 and myosin-6 reroute after IS construction. For example, most myosin-5 traffic moves inward at the IS, although most of the IS actin filaments have a barbed end out orientation. This anomalous myosin-5 traffic pattern indicates that the IS makes two types of actin networks: a structural network that controls IS shape, and a distinct trafficking network that supports myosin motility. We disrupt these trafficking networks with chemical probes against actin, which inhibits the appearance of cell surface markers of T cell activation. Our results highlight the importance of the sparse actin networks at the center of the IS in T cell function.


6
T cells form specialized contacts with other cells, which is essential for their mat-surface markers of T cell activation. Together, we find that T cells regulate and redirect 48 myosin traffic during IS construction and T cell activation. 51 To image the actin networks and myosin traffic at the mature immunological synapse, 52 we use a common approach with antibody-coated coverslips to begin synapse formation 53 in the image plane [24,28,47]. One antibody binds and activates through CD3, while a 54 second antibody binds CD28 to provide a co-stimulatory interaction (mouse anti-human 55 OKT3 and CD28.2 antibodies, respectively). We omit the anti-CD3 antibody in our non-56 activating control condition. We apply T cells to these coverslips, allow cells to adhere 57 or synapses to form, and then fix the actin in place and remove the plasma membrane.  Figure 1: The ex vivo motility assay scheme applied to the immunological synapse. We apply lymphocytes (Jurkat T cells, human PBMCs, or CD8 + T cells from PBMCs) to a coverslip coated with activating antibodies. After contact with an activating surface, an immunological synapse forms within 5 min. We then open the cell with detergent and stabilize and label the actin filaments to visualize the synapse. Finally, we apply fluorophorelabeled myosin motors (myosin-5 or myosin-6) and track their movements through the synapse. Perturbations include activating vs. non-activating surfaces (anti-CD3/CD28 vs anti-CD28 alone) and small molecule inhibitors of actin filament assembly and function (CK666 and TR100).

Myosin-5 travels outward in resting T cells but inward at the immunological synapse
shift is the nucleation of new actin between microvilli, which increases the accessible 114 area for myosin runs at the outer extent of the cell. With anti-CD28 surfaces, myosin-6 115 runs concentrate in the interior of the cell. This concentration may reflect myosin-6's 116 preference for older actin at the interior of the cell [48]. 117 To distinguish myosin motility across the SMACs, we divided the myosin runs into 118 three radial zones-central, proximal, and distal-and investigated the heading distribu-119 tions within each of the three (Fig. 4B). On the anti-CD28 surfaces, both myosins are 120 strongly polarized (5:out and 6:in) in the distal zone containing the microvilli. That po-121 larization decays to isotropic motility through the proximal and into the central zones. On The bottom row in each set shows the detected runs, colored by heading. Headings of 0°run out from the cell center, while headings of 180°run in toward the cell center. Rose plots show the distributions (PDFs) of mean headings on a cell-by-cell basis (B), and headings for all runs collectively (C). Both measures show the reversal of myosin-5 headings on the activating surfaces. All four distributions in (B, C) are nonuniform (Watson test of uniformity; U 2 = 0.73, 0.82, 13, 3.3; p < 0.01 for each). The heading distributions also differ within (B) and within (C) (Watson-Wheeler test of homogeneity; W = 27, 570; p = 9 × 10 -7 , < 2 × 10 -16 ). Scale bars, 10 μm.  myosin-6 moves out from the inner two zones and in from the outer zone. Thus, myosin-5 125 motility diverges at the proximal / distal interface, while myosin-6 converges.

126
Based upon these headings over radial zones, we predict that myosin-5 concentrates 127 in the cSMAC, while myosin-6 concentrates at the pSMAC / dSMAC boundary. We used  The T cell state governs more than just the myosin heading 135 The rerouting of myosin-5 and myosin-6 traffic is apparent by eye. use the UMAP dimension reduction algorithm to cluster cells by their myosin behavior. 143 We then associate these clusters with the cellular state.

144
Our UMAP embedding for myosin-5 appears in Fig  we also purified by negative selection CD8 + PBMCs to understand a second major T cell 158 class. These CD8 + cells cluster within the Jurkat anti-CD3/CD28 cluster, suggesting they 159 share similar motility features. Thus, the myosin traffic rerouting we see at the immune 160 synapse appears to be a general phenomenon.

161
The UMAP embedding of cells probed with myosin-6 shows a similar circular feature, To allow comparisons between cells, we normalize distances relative to the maximum radial distance observed for each cell. On activating anti-CD3/CD28 surfaces, runs of both myosin-5 and myosin-6 shift outward. We divide the radial range into thirds, defining central, proximal, and distal zones that approximate the cSMAC, pSMAC, and dSMAC. (B) Heading changes depend upon the radial zone. On the non-activating anti-CD28 surfaces, both myosin-5 and myosin-6 show strong outward and inward preferences (respectively) in the distal zone. On activating anti-CD3/CD28 surfaces, myosin-5 moves nearly isotropically through the distal zone, but moves inward in the central and proximal zones. Likewise, myosin-6 moves outward from the central and proximal zones, and inward from the distal zone. This motility pattern would tend to focus myosin-6 at the proximal-distal boundary. The three zone distributions differ within each of the four conditions (Watson-Wheeler test of homogeneity; W = 600, 740, 550, 800; p < 2 × 10 -16 ). Immunofluorescence images show that endogenous myosin-5 concentrates at the cSMAC in either a tight central cluster (C, D), or in a more scattered central pattern (E, F). Endogenous myosin-6 concentrates at the pSMAC/dSMAC (G-J). Magenta: actin, green: myosin, scale bars, 10 μm.  is that it inhibits secretion of cytotoxic granules [33,34,12]. In this picture, actin at the [29]. Although myosin-6 is not known to interact directly with LFA-1, it may have related 262 roles that require localization at the pSMAC. These roles include endocytosis [11,16,39] or the anchoring of lipid membranes [21,2].  This mean value will lie within the unit disc, and its radius is assigned to rho. Rho is 0   Figure S6: Treatment with TR100 inhibits production of the early T cell activation marker CD69, while CK666 shows no such effect. Flow cytometric analysis of permeabilized cells reveals a TR100 dose-dependent decrease of total CD69 (χ 2 (4) = 53.7, p = 6.1 × 10 -11 ), in Jurkat cells activated with anti-CD3/CD28.