Chloride intracellular channel (CLIC) protein function in S1P-induced Rac1 activation requires membrane localization of the C-terminus, but not thiol-transferase nor ion channel activities

Abstract

We have established a novel and evolutionarily-conserved function for chloride intracellular channel proteins (CLICs) in regulating Rho/Rac GTPases downstream of G protein-coupled receptors (GPCRs). Endothelial CLIC1 and CLIC4 are rapidly and transiently re-localized from the cytoplasm to the plasma membrane in response to the GPCR ligand sphingosine-1-phosphate (S1P), and both CLICs are required to activate Rac1 in response to S1P, but how they perform this function remains unknown. Biochemical studies suggest that CLICs act as non-specific ion channels and/or as glutathione-S-transferases, dependent on N-terminal features, in vitro. Here we investigate CLIC functional domains and membrane localization requirements for their function in S1P-mediated Rac1 signaling. Structure-function analyses of CLIC function in endothelial cells demonstrate that CLIC1 and CLIC4-specific functions reside at their C-termini, and that the CLIC4 N-terminus encodes determinants required for S1P-induced re-localization to the plasma membrane but is dispensable for S1P-induced Rac1 activation when the C-terminus is localized to the plasma membrane via a heterologous signal. Our results demonstrate that the postulated ion channel and thiol-transferase activities of CLICs are not required for Rac1 activation and suggests that sequences in the CLIC C-termini are critical for this function. Given the importance of S1P signaling in vascular biology and disease, our work establishes a platform to further our understanding of the membrane-localized proteins required to link GPCR activity to Rho/Rac regulation.

Keywords: CLIC1; CLIC4; Rac1; RhoA; S1P Receptor; endothelium; sphingosine-1-phosphate.

Fig. 1:. CLIC1 and CLIC4 in endothelial S1P signaling, and CLIC sequence and structural features relevant to this study.

(A) description of shared and unique functions for CLIC1 and CLIC4 in endothelial S1P signaling (Mao et al., 2021). (B) Protein sequence alignment of C. elegans EXC-4, and human CLIC1 and CLIC4 indicating regions of amino-acid identity (dark grey shading) and similarity (light grey shading). Conserved beta sheets (yellow arrows) and alpha helices (pink rounded boxes) determined by crystal structure of these proteins (Littler et al., 2004; Littler et al., 2005; Littler et al., 2008) are shown. The N-terminal putative transmembrane domain (PTMD) and conserved thioredoxin fold are highlighted. Note that the N-terminal cysteine required for thiol-reductases activity (red), and the C-terminal nuclear localization signal (orange) found in human CLIC1 and 4 are not present in C. elegans EXC-4, while the C. elegans protein has a C-terminal extension with additional alpha helices (α10, α11). (C) cartoon representation of HA-tagged full-length CLIC1 and CLIC4 constructs. The C-terminal fragments swapped in the chimeric constructs analyzed in Fig. 2 are denoted in blue, and the ∆PTMD CLIC4 truncation analyzed in Figs. 3 and 4 is denoted in purple (the precise protein sequence location for these constructs are shown as blue and purple arrows in panel B).

Fig. 2:. CLIC1 and CLIC4-specific functions in S1P signaling are encoded by their C-termini.

(A) cartoon representation of the C1-C4 and C4-C1 chimeras, and western blot analysis and quantification (n=4) of their expression in HUVEC (see Methods). (B) full-length CLIC1 and the C4-C1 chimera rescue the Rac1 activation defect caused by CLIC1^KD, as assessed by G-LISA (see Methods). (C) full-length CLIC4 and the C1-C4 chimera rescue the Rac1 activation defect caused by CLIC4^KD, as assessed by G-LISA. In panels (B) and (C) each bar represents the mean of n≥3 replicates, the error bars represent standard deviation (hereafter, SD), and significance between vehicle control (Veh.) and S1P-treated cells was calculated via unpaired t-tests. HUVEC monolayers form a barrier whose strength can be measured via a trans-endothelial electrical resistance (TEER) assay, and stimulation by S1P rapidly, and reversibly, enhances HUVEC barrier function (Garcia et al., 2001). We previously showed that CLIC1 or CLIC4 knockdown reduced S1P-induced barrier strengthening (Mao et al., 2021), and here we show that (D) full-length CLIC1 and the C4-C1 chimera rescue the barrier defect caused by CLIC1^KD (see Methods), while (E) full-length CLIC4 and the C1-C4 chimera rescue the TEER defect caused by CLIC4^KD. Each trace represents the mean TEER measurement of n≥3 samples, normalized to the TEER value at time of S1P addition (see Methods). To quantify these results, the area under the curve (AUC) ± standard error of the mean (SEM) from the time of S1P addition (~30min. after beginning of resistance measurements) until time of maximal recovery to baseline in control (~60 min. after beginning of experiment, ~30 min. after S1P addition) was calculated for each condition (see Methods), and these results are shown in the accompanying bar graphs. Significance was calculated using unpaired 2-tailed t tests.

Fig. 3:. The N-terminal putative transmembrane domain (PTMD) is required for CLIC4 re-localization and Rac1 activation in response to S1P.

(A) cartoon representation of the HA-tagged (∆PTMD)CLIC4 construct, and western blot analysis and quantification (n=3) of its expression in HUVEC (see Methods), as compared to full-length HA-CLIC4. (B) Immunofluorescence of full-length HA-tagged CLIC4 (top) and (∆PTMD)CLIC4, and their localization after treatment with S1P (see Methods). Both constructs showed nuclear and cytoplasmic accumulation prior to S1P treatment. Noticeable accumulation of full-length CLIC4 was observed near the cell cortex (outlined by VE-cadherin) within 5 minutes and remained at 10 and 15 minutes after S1P addition (yellow arrows, top rows). In contrast we did not observe similar re-localization of the (∆PTMD)CLIC4 construct (bottom rows). Scale bar represents 50µm in all panels. (C) the (∆PTMD)CLIC4 construct does not rescues the Rac1 activation defect caused by CLIC4^KD, as assessed by G-LISA. Each bar represents the mean of n≥3 replicates, the error bars represent SD, and significance between vehicle control (Veh.) and S1P-treated cells was calculated via unpaired t-tests. (D) the (∆PTMD)CLIC4 construct does not rescue the TEER defect caused by CLIC4^KD. Each trace represents the mean TEER measurement of n≥3 samples, normalized to the TEER value at time of S1P addition, and quantitative AUC analysis of TEER calculation was carried out as described in Fig. 2D and 2E.

Fig. 4:. Membrane localized CLIC4 C-terminus is sufficient to promote Rac1 activation in response to S1P.

(A) cartoon representation of HA-tagged full-length myristylated (myr) CLIC4 and myr-(∆PTMD)CLIC4, lacking the N-terminal putative transmembrane domain. Western blot analysis and quantification (n=3) of construct expression in HUVEC (see Methods), shows lower expression of myr-(∆PTMD)CLIC4 as compared to full-length myr-CLIC4. (B) Immunofluorescence showing accumulation of HA-tagged full-length CLIC4 (first column), myr-CLIC4 (second column) and myr-(∆PTMD)CLIC4 (third column) at steady state (no S1P). Note that both myristylated constructs display accumulation at the membrane (yellow arrows), and decreased nuclear accumulation (cyan asterisks) when compared to wildtype full-length CLIC4. Scale bar represents 50µm in all panels. (C) both the full-length myr-CLIC4 and the myr-(∆PTMD)CLIC4 constructs rescue the Rac1 activation defect caused by CLIC4^KD, as assessed by G-LISA. Each bar represents the mean of n≥3 replicates, the error bars represent SD, and significance between vehicle control (Veh.) and S1P-treated cells was calculated via unpaired t-tests. (D) (C) both the full-length myr-CLIC4 and the myr-(∆PTMD)CLIC4 constructs rescue the TEER defect caused by CLIC4^KD. Each trace represents the mean TEER measurement of n≥3 samples, normalized to the TEER value at time of S1P addition, and quantitative AUC analysis of TEER calculation was carried out as described in Fig. 2D and 2E. Dashed lines and bars (Control, CLIC4^KD, and CLIC4^KD + CLIC4) are the same data shown in Fig. 3D