Targeted chemical disruption of clathrin function in living cells

Abstract

The accurate assignment of molecular roles in membrane traffic is frequently complicated by the lack of specific inhibitors that can work on rapid time scales. Such inhibition schemes would potentially avoid the complications arising from either compensatory gene expression or the complex downstream consequences of inhibition of an important protein over long periods (>12 h). Here, we developed a novel chemical tool to disrupt clathrin function in living cells. We engineered a cross-linkable form of clathrin by using an FK506-binding protein 12 (FKBP)-clathrin fusion protein that is specifically oligomerized upon addition of the cell-permeant cross-linker FK1012-A. This approach interrupts the normal assembly-disassembly cycle of clathrin lattices and results in a specific, rapid, and reversible approximately 70% inhibition of clathrin function. This approach should be applicable to a number of proteins that must go through an assembly-disassembly cycle for normal function.

Figure 1.

Cross-linkable form of clathrin. (A) Design of the EGFP-FKBP-LC fusion construct (top) and the presumed integration of this construct into triskelia. (B) EGFP-FKBP-LC is preferentially expressed relative to endogenous clathrin light chain molecules in TRVb-EFL cells. Equal amounts of protein were loaded in each lane. (C) EGFP-FKBP-LC puncta (top) in TRVb-EFL cells colocalize with immunofluorescence derived from clathrin heavy chain puncta (bottom). Bar, 5 μm. (D) Distribution of the relative intensity of fluorescence derived from EGFP-FKBP-LC (LC) or immunofluorescence against clathrin heavy chain (HC) indicates that all clathrin pits incorporate the EGFPFKBP-LC fusion construct (n = 8; 800 puncta total).

Figure 2.

EGFP-FKBP-LC and its associated triskelion can be reversibly clustered. (A) Clathrin-associated structures are localized to the plasma membrane and the trans-Golgi network in TRVb-EFL cells. Bar, 10 μm. (B) FK1012-A (50 nM, 2 h) leads to a clearing of cytoplasmic clathrin and the formation of large intracellular aggregates of clathrin. Clathrin puncta remain associated with the plasma membrane after cross-linking. (C) FK506 (10 μM, 2 h) reverses FK1012-A–mediated cross-linking, as clathrin distributes in a similar pattern to that of the control condition (A). (D–F) After incubation with FK1012-A, EGFP-FKBP-LC fluorescence (D) and indirect immunofluorescence against clathrin heavy chain (E) overlap almost perfectly (F).

Figure 3.

FK1012-A reduces fluorescence recovery from photobleaching. (A) Plasma membrane-associated clathrin-coated pits imaged before (top), immediately after (middle) and 140 s after photobleaching (bottom). Immediately after photobleaching fluorescence was reduced to a level ∼70% of the initial fluorescence levels (middle). (B) Average time course of fluorescence recovery after photobleaching. Cross-linking clathrin triskelia with FK1012-A (50 nM, 2 h) reduces fluorescence recovery fourfold. The cholesterol sequestering agent β-methyl cyclodextrin (10 mM, 30 min) inhibits recovery by 25% (control TRVb-EFL cells, n = 12; TRVb-EFL cells + FK1012-A, n = 6; TRVb-EFL cells + β-methyl cyclodextrin, n = 6, p; TRVb-EFL cells + FK1012-A followed by FK506, n = 6).

Figure 4.

FK1012-A disrupts clathrin-coated pit structure. (A) Unroofed TRVb-EFL cells exhibit typical flat clathrin lattices (top) and curved clathrin-coated pits (bottom). (B) FK1012-A (50 nM, 55 min) alters the structure of clathrin lattices. The fundamental network within flat lattices and curved pits is extremely deformed. Additional proteinaceous material, presumably clathrin triskelia from the cytoplasm, seems to coat the cytoplasmic surface of the structures.

Figure 5.

FK1012-A reduces membrane budding of clathrin-coated pits. (A) Time-lapse images of membrane budding at the cell surface were acquired with total internal reflection fluorescence microscopy. The budding activity of each clathrin-coated pit in the field of view was documented. The arrow indicates a clathrin-coated pit leaving the field of view in successive frames. (B) A large percentage of clathrin-coated pits in untreated TRVb-EFL cells undergo a membrane budding event. FK1012-A (50 nM, 2 h) reduces the number of membrane budding events and increases the number of stationary coated pits. Subsequent application of FK506 (10 μM, 2 h) returns the number of membrane budding events to that of the control condition. (Control, n = 18; FK1012-A, n = 12; FK1012-A followed by FK506, n = 6).

Figure 6.

Cross-linking of triskelia perturbs membrane traffic in TRVb-EFL cells. (A) TRVb-EFL cells internalize transferrin at a similar rate to that of TRVb-1 cells. FK1012-A (50 nM, 2 h) reduces the rate of internalization of transferrin by ∼50% in TRVb-EFL cells but fails to inhibit internalization in TRVb-1 cells. Application of FK506 (10 μM, 2 h) to TRVb-EFL cells returns the rate of internalization of transferrin to control levels in previously cross-linked cells. The data are the averages of three separate experiments under identical conditions ± SE. (B) FK1012-A (50 nM, 2 h) reduces the LDL internalization rate constant by ∼50% compared with control TRVb-EFL cells. The data represent the average internalization values of LDL in >80 TRVb-EFL cells per data point. (C) Transferrin recycling is minimally perturbed in cross-linked TRVb-EFL cells. The graph represents the averages of three separate experiments under identical conditions ± SE. (D) FK1012-A (50 nM, 2 h) increases the steady-state surface levels of transferrin receptor. The data are the average of three separate experiments under identical conditions ± SE. (E) Internalization of FM4-64 is unaffected by FK1012-A (50 nM, 2 h) in TRVb-EFL cells. The data are the average of three separate experiments under identical conditions ± SE.

Figure 7.

Inhibition of actin in TRVb-EFL cells inhibits clathrin-dependent endocytosis. (A) Lat-A (10 μM, 30 min) weakly inhibits fluorescence recovery after photobleaching. However, when Lat-A is combined with FK1012-A (50 nM, 2 h), recovery of fluorescence is completely inhibited (control TRVb-EFL cells, n = 12; TRVb-EFL cells + Lat-A, n = 6; TRVb-EFL cells + FK1012-A, n = 6; TRVb-EFL cells + FK1012-A followed by Lat-A, n = 6). (B) Lat-A (10 μM, 30 min) reduces membrane budding events and increases the number of stationary coated pits as viewed with TIR-FM (control, n = 18; Lat-A, n = 6). (C) The combined effect of FK1012-A (50 nM, 2 h) and Lat-A (10 μM, 30 min) is a complete abolishment of membrane budding (control, n = 18; FK1012-A, n = 12; FK1012-A + Lat-A, n = 6). (D) FK1012-A (50 nM, 2 h) and Lat-A (10 μM, 30 min) each reduce the transferrin internalization rate constant to ∼50% of the rate of untreated control TRVb-EFL cells. TRVb-EFL cells incubated with FK1012-A (50 nM, 2 h) followed by Lat-A (10 μM, 30 min) reduces the transferrin internalization rate constant to 20% of the rate of control cells. The data are the average of three separate experiments under identical conditions ± SE. (E) Incubation of TRVb-EFL cells either with Lat-A, FK1012-A, or FK1012-A followed by Lat-A significantly increases the steady-state levels of transferrin receptor on the surface of TRVb-EFL cells. The data are the average of three separate experiments under identical conditions ± SE.