PQLC2 recruits the C9orf72 complex to lysosomes in response to cationic amino acid starvation

Abstract

The C9orf72 protein is required for normal lysosome function. In support of such functions, C9orf72 forms a heterotrimeric complex with SMCR8 and WDR41 that is recruited to lysosomes when amino acids are scarce. These properties raise questions about the identity of the lysosomal binding partner of the C9orf72 complex and the amino acid-sensing mechanism that regulates C9orf72 complex abundance on lysosomes. We now demonstrate that an interaction with the lysosomal cationic amino acid transporter PQLC2 mediates C9orf72 complex recruitment to lysosomes. This is achieved through an interaction between PQLC2 and WDR41. The interaction between PQLC2 and the C9orf72 complex is negatively regulated by arginine, lysine, and histidine, the amino acids that PQLC2 transports across the membrane of lysosomes. These results define a new role for PQLC2 in the regulated recruitment of the C9orf72 complex to lysosomes and reveal a novel mechanism that allows cells to sense and respond to changes in the availability of cationic amino acids within lysosomes.

Figure 1.

Identification of the amino acid transporter PQLC2 as a C9orf72-interacting protein. (A) Schematic diagram illustrating the Lyso–C9orf72-GFP and the Lyso-GFP proteins that were used for proteomics experiments. (B) Summary of strategy to enrich the C9orf72 complex at lysosomes to facilitate identification of lysosomal-interacting partners. (C) Summary of proteins identified by LC–MS/MS that interact with Lyso–C9orf72-GFP compared with a Lyso-GFP control. Total peptide numbers are shown. (D) HEK293FT cells that express 2xHA-tagged C9orf72 from the endogenous locus were transfected with FLAG-tagged PQLC2 or FLAG-PAT1, followed by anti-FLAG immunoprecipitation and immunoblotting for FLAG or endogenously expressed SMCR8 and C9orf72. (E) Effect of PQLC2 overexpression on C9orf72 localization in fed and starved cells. Scale bars: 10 µm. Insets: 6.0 µm wide. (F) PAT1 overexpression did not affect C9orf72 localization in fed cells. Scale bar: 10 µm. Inset: 5.6 µm wide.

Figure S1.

PQLC2-FLAG and FLAG-PAT1 localize to lysosomes. (A and B) Immunofluorescence images of FLAG and LAMP1 in HEK293FT cells transiently transfected with FLAG-tagged PQLC2 (A) or PAT1 (B). Scale bars: 10 µm. Inset is 7.6 µm wide in A and 6.7 µm wide in B.

Figure 2.

PQLC2 is required for C9orf72 recruitment to lysosomes. (A) Immunofluorescence images of C9orf72 localization (endogenously expressed 2xHA-C9orf72) in starved WT and PQLC2 KO cells. Localization of C9orf72 to lysosomes (LAMP1) is lost in PQLC2 KO cells. Scale bar: 10 µm. Insets: 8.0 µm wide. (B) For the indicated cell lines, a summary of the percentage of cells in starved conditions containing C9orf72 puncta that colocalize with LAMP1. Mean ± SEM is plotted with individual experiments displayed as open dots. Unpaired two-tailed t test; ***, P = 0.0002, n = 3, experiments with >140 cells analyzed per cell line. (C) Immunofluorescence images of C9orf72 and FLCN localization in starved WT and PQLC2 KO cells. Scale bar: 10 µm. Insets: 7.6 µm wide. (D) Immunofluorescence images of FLCN localization in starved PQLC2 KO cells. Recruitment of FLCN to lysosomes (LAMP1-positive puncta) is maintained in these cells. Scale bars: 10 µm. Inset: 6.3 µm wide. (E) Cells expressing either PQLC2-FLAG or FLAG-tagged RagB and RagD were subjected to anti-FLAG immunoprecipitations and immunoblotting for FLAG and endogenous C9orf72 and FLCN.

Figure S2.

C9orf72 recruitment to lysosomes is independent of GATOR1-associated nutrient sensing. (A) Immunoblot analysis of Nprl3, S6K, and phospho-S6K (T389) levels during starvation (2 h) and amino acid refeeding (15 min) in WT and Nprl3 KO cells. (B) Immunofluorescence images of C9orf72 localization under normal growth conditions for WT and Nprl3 KO cells. Insets are 5.5 µm wide. (C)  Immunofluorescence images of C9orf72 localization under starved conditions (2 h) for WT and Nprl3 KO cells. Localization of C9orf72 to late endosomes and lysosomes is demonstrated by colocalization with LAMP1 puncta. Scale bars: 10 µm. Insets are 4.8 µm wide.

Figure 3.

PQ motif mutations in PQLC2 disrupt C9orf72 complex interactions and recruitment of C9orf72 to lysosomes. (A) PQLC2 KO cells were transfected with WT PQLC2-FLAG or the indicated mutants. Cells were cultured under fed (F) or starved (S) conditions; PQLC2-FLAG was immunoprecipitated, followed by immunoblotting for FLAG and the endogenous WDR41, C9orf72, and SMCR8 proteins. (B) Iron-dextran–loaded lysosomes were magnetically purified from the parental HEK293FT cells, PQLC2 KO cells, and PQLC2 KO cells stably expressing WT PQLC2-FLAG or P55L/P201L PQLC2-FLAG. Immunoblots are shown of the indicated proteins in the total cell lysate (input) and in magnetically isolated lysosomes (lysosomes). (C) Quantification of immunoblots for the indicated proteins expressed as a fold increase in the lysosome fraction relative to the input (mean ± SEM, three independent experiments, data points from individual experiments shown as open circles; ****, P ≤ 0.0001; ANOVA with Tukey’s multiple comparisons test). (D) Lysosomal levels of SMCR8 or C9orf72 normalized to the respective lysosomal PQLC2-FLAG in KO cells expressing WT and mutant PQLC2. Mean ± SEM, unpaired t test; *, P ≤ 0.05, n = 3. (E–H) PQLC2 KO cells were transfected with WT PQLC2-FLAG or the indicated mutants and cultured in normal growth medium (fed) or amino acid/serum-free RPMI (starved). Cells were fixed and immunostained with FLAG and HA antibodies to examine C9orf72 localization. Transfected cells are outlined with dashed lines. Scale bars: 10 µm.

Figure S3.

PQLC2-FLAG PQ loop mutants localize to lysosomes. (A) Visualization of PQLC2 topology generated by Protter overlaid with the conservation of amino acids at each position from ConSurf (Ashkenazy et al., 2016). The positions of the prolines in the PQ loops are indicated by arrows. (B–D) Immunofluorescence images of FLAG and LAMP1 labeling in HEK293FT cells transiently transfected with FLAG-tagged PQLC2 mutants: P55L (B), P201L (C), and P55L/P201L (D). Scale bars: 10 µm. Insets are 6.4 µm wide in B, 6.8 µm wide in C, and 7.7 µm wide in D. (E and F) Immunofluorescence images of FLAG and calnexin localization in cells transfected with FLAG-tagged WT PQLC2 (E) and P55L/P201L mutant PQLC2 (F). Scale bars: 10 µm. Insets are 7.7 µm wide in E and F. (G) Quantification of WT and mutant PQLC2 colocalization with LAMP1 and calnexin (images of WT PQLC2 and LAMP1 colocalization are shown in Fig. S1). Pearson’s correlation coefficient was measured with results from individual images plotted as open circles; mean ± SEM, ANOVA with Tukey’s multiple comparisons test. ****, P ≤ 0.0001, n = 3 experiments with a total of 20 images analyzed.

Figure 4.

Amino acid–dependent regulation of the C9orf72 complex’s interaction with PQLC2. (A) WT (parental) or gene-edited cells expressing PQLC2-2xHA from the endogenous locus were incubated in medium with or without amino acids and/or dialyzed serum. PQLC2 was immunoprecipitated, followed by immunoblotting with the indicated antibodies. (B) Summary of the ratio of SMCR8 to PQLC2-2xHA in immunoprecipitations from A. All values are mean ± SEM, with data points from individual experiments (n = 4) indicated by open circles. **, P ≤ 0.01; ANOVA with Dunnett’s multiple comparisons test. (C) List of amino acids in MEM–amino acid mix. Magenta, green, and blue boxes indicate groups of amino acids used in subsequent experiments. (D) Cells were incubated in media containing the MEM–amino acid mix (+MEM, magenta box), media without amino acids (−MEM), without arginine, lysine, and histidine (−RKH, green box), or without the other amino acids in the MEM mix (−Others, blue box). PQLC2 was then immunoprecipitated to determine the effects of these conditions on the interaction. (E) Summary of the ratio of SMCR8 to PQLC2-2xHA in immunoprecipitations in D. **, P ≤ 0.01; ***, P ≤ 0.001; ANOVA with Dunnett’s multiple comparisons test (n = 3). (F) Cells were starved and then incubated in media containing the indicated amino acids, followed by immunoprecipitation of PQLC2. (G) Summary of the ratio of SMCR8 to PQLC2-2xHA in immunoprecipitations from F. **, P ≤ 0.01; *** P ≤ 0.001; ANOVA with Dunnett’s multiple comparisons test (n = 3).

Figure S4.

Epitope tagging of the endogenous PQLC2 protein. (A) Sequencing traces from an HEK293FT cell line that has a 2xHA epitope tag inserted into the endogenous PQLC2 gene. Example traces from unaffected and edited alleles are shown. This insertion results in a 2xHA tag at the C-terminus of PQLC2. (B) Summary of sequencing results from this cell line. Results are consistent with one of three copies of PQLC2 being 2xHA-tagged at the C-terminus, while the remaining two copies of the gene are unaffected. (C) The specificity of the anti-HA immunofluorescence signal in PQLC2-2xHA cells is supported by the absence of this signal in parental, non–gene-edited cells. (D) Immunofluorescence images showing the localization of PQLC2-2xHA to lysosomes (endogenous PQLC2-2xHA and LAMP1). Scale bars: 10 µm. Insets are 5.9 µm wide.

Figure 5.

C9orf72 enrichment on lysosomes is negatively regulated by the availability of cationic amino acids. Immunofluorescence images of 2xHA-C9orf72 and LAMP1 in cells expressing 2xHA-C9orf72 from the endogenous locus. Cells were starved and then incubated in media containing the indicated amino acids. Scale bar: 10 µm. Insets: 7.0 µm wide.

Figure S5.

PQLC2 localization to lysosomes is not regulated by changes in amino acid availability. Immunofluorescence images of PQLC2-2xHA and LAMP1 in cells expressing PQLC2-2xHA from the endogenous locus. Cells were starved and then incubated in media containing the indicated amino acids (as in Figs. 4 F and 5). Scale bar: 10 µm. Insets are 5.7 µm wide.

Figure 6.

Lysosomal substrates negatively regulate the interaction between PQLC2 and the C9orf72 complex. (A) Principle of the cystinosin depletion/cysteamine treatment experiment. Cystinosin (CTNS) is a lysosomal transporter of cystine, and PQLC2 is a lysosomal transporter of arginine, lysine, and histidine. Upon depletion of cystinosin, cystine accumulates within lysosomes. When cysteamine is added, it reacts with cystine in lysosomes to form cysteine and a lysine-like mixed disulfide that can be transported by PQLC2. Therefore, upon cystinosin depletion, a PQLC2 substrate can be acutely generated inside lysosomes by cysteamine addition. (B) WT or cystinosin-depleted cells expressing PQLC2-2xHA from the endogenous locus were starved and then incubated in starvation media containing cysteamine (60 min, 1 mM). PQLC2 was immunoprecipitated from cell lysates, followed by immunoblotting with the indicated antibodies. (C) Summary of the ratio of SMCR8 to PQLC2-2xHA in the immunoprecipitations from B. Values are mean ± SEM from four independent experiments with individual data points indicated by open circles; *, P < 0.05; ANOVA with Tukey’s multiple comparisons test.

Figure 7.

WDR41 mediates the interaction of the C9orf72 complex with PQLC2. (A) PQLC2-FLAG was expressed in WT and WDR41 KO cells. Following anti-FLAG immunoprecipitation, PQLC2-FLAG and endogenous WDR41, C9orf72, and SMCR8 were detected by immunoblotting. (B) Summary of results from A. Levels of the indicated proteins in anti-FLAG immunoprecipitations from WT and WDR41 KO cells (WT levels normalized to 1, n = 3, mean ± SEM plotted with data points from individual experiments [n = 3] plotted as open circles; ****, P ≤ 0.0001). (C) Immunofluorescence of PQLC2-FLAG and LAMP1 in WDR41 KO cells. Scale bar: 10 µm. Inset: 6.7 µm wide. (D) Immunofluorescence images of C9orf72 localization (endogenously expressed 2xHA-C9orf72) in WT and WDR41 KO cells transiently overexpressing PQLC2-FLAG. Consistent with immunoprecipitation data, overexpression of PQLC2 results in its colocalization with C9orf72 in WT, but not WDR41 KO, cells. Scale bar: 10 µm. Insets: 5.0 µm wide. (E) WT and C9orf72/SMCR8 DKO were transfected with FLAG-tagged PQLC2 and subjected to anti-FLAG immunoprecipitations. FLAG-tagged PAT1, an unrelated lysosomal amino acid transporter, served as a negative control. (F) Summary of WDR41 levels in anti-FLAG immunoprecipitations from E (n = 3, mean ± SEM plotted with data points from individual experiments plotted as open circles; ANOVA with Tukey’s multiple comparisons test; ***, P ≤ 0.001).

Figure 8.

PQLC2 KO cells exhibit a defect in the activation of mTORC1 by cationic amino acids. (A) Immunoblot analysis of cathepsin B, D, L, and LC3 in WT and PQLC2 KO cell lines. Ribosomal protein S6 was used as a loading control. WT cells treated with chloroquine (50 µM, 15 h) were used as a positive control of impaired lysosome function. (B) Immunoblot analysis of S6K T389 phosphorylation levels during starvation for arginine, lysine, and histidine (R/K/H) and upon subsequent R/K/H refeeding in WT, PQLC2 KO, and in KO cells stably expressing PQLC2-FLAG. (C) Quantification of phospho-S6K levels in the indicated conditions in B (normalized to total S6K levels; mean ± SEM with data points from individual experiments [n = 4] indicated by open circles; **, P ≤ 0.01; ANOVA with Tukey’s multiple comparisons test). (D) Immunoblot analysis of phospho-S6K levels during starvation for R/K/H and subsequent R/K/H refeeding in PQLC2 KO and in PQLC2 KO cells stably expressing lysosome-targeted C9orf72-GFP. (E) Quantification of phospho-S6K levels in the indicated conditions in D (*, P < 0.05; ANOVA with Tukey’s multiple comparisons test, n = 3).