Genetic analysis of genes encoding Retromer and Commander subunits using unbiased genome-wide CRISPR screens.

(A) Cartoon representation of the Retromer and Commander complexes. (B) Table listing genes encoding the subunits of the Retromer and Commander complexes. (C) Ranking of genes encoding the Retromer and Commander complexes in an unbiased CRISPR screen that was conducted to identify genes required for the surface homeostasis of GLUT-SPR [8]. The GLUT-SPR reporter was constructed by inserting a hemagglutinin (HA) epitope into an exoplasmic loop of the glucose transporter GLUT4, with a GFP tag fused to the intracellular C-terminus of GLUT4 [34, 75, 76]. Surface expression (HA staining) of the reporter was normalized to total reporter expression (GFP fluorescence) as a measure of relative surface levels of the reporter. Each dot represents a gene. The dashed line depicts the P-value cutoff at 0.05. (D) Essentiality scores of genes encoding the Retromer and Commander complexes calculated by comparing gRNA abundance in a passage cell population (without any selection) with that in the initial CRISPR library. Genes with essentiality scores below the horizontal cutoff line are predicted to be essential to cell survival or growth. Full datasets of the CRISPR screens are included in a previous report [8].

Intracellular sequestration of GLUT-SPR in COMMD3-deficient cells.

(A) Representative immunoblots showing the indicated proteins in wild-type (WT) and Commd3 KO mouse preadipocytes. (B) Normalized surface levels of GLUT-SPR measured by flow cytometry in WT and KO preadipocytes. The cells were either untreated or treated with 100 nM insulin for one hour before surface GLUT-SPR was labeled using anti-HA antibodies and APC-conjugated secondary antibodies. To inhibit insulin signaling, 100 nM wortmannin was added prior to insulin stimulation. In all figures, data normalization was performed by setting the mean value of WT data points as 100 or 1, and all data points including WT ones were normalized to that mean value. Data are presented as mean ± SD of three biological replicates. ** P < 0.01; * P < 0.05 (calculated using Student’s t-test). (C) Normalized surface levels of GLUT-SPR in WT and KO adipocytes. Data are presented as mean ± SD of three biological replicates. *** P < 0.001; n.s., P > 0.05 (calculated using Student’s t-test). (D) Representative confocal images showing the localization of GLUT-SPR in unpermeabilized WT and Commd3 KO adipocytes, which were either untreated or treated with 100 nM insulin for one hour. Surface GLUT-SPR was labeled using anti-HA antibodies and Alexa Fluor 568-conjugated secondary antibodies. Nuclei were stained with Hoechst 33342. Scale bars: 10 µm.

Abnormal endosomal morphology in COMMD3-deficient cells.

(A) Representative confocal images showing the localization of EEA1 and GLUT-SPR in permeabilized WT and Commd3 KO adipocytes stimulated with 100 nM insulin for one hour (scale bars: 10 µm). Enlarged inset images depict co-localization of EEA1 and GLUT-SPR (scale bars: 5 µm). (B) Violin plot showing quantification of EEA1-positive puncta using Fiji threshold analysis. Data of the KO adipocytes were normalized to those of WT cells. Three independent experiments are shown with ten cells analyzed in each experiment. *** P < 0.001 (calculated using Student’s t-test). (C) Violin plot depicting quantification of GFP mean fluorescence intensity (MFI) in EEA1-positive puncta using Fiji threshold analysis. Data of KO adipocytes were normalized to those of WT cells. Three independent experiments are shown with ten cells analyzed in each experiment. *** P < 0.001 (calculated using Student’s t-test). (D) Representative SIM images showing the subcellular localization of Rab5 and GLUT-SPR in WT and Commd3 KO preadipocytes (scale bars: 5 µm). (E) Quantification of Rab5 and GLUT-SPR co-localization based on SIM images, which were captured as in (D) and analyzed using ImageJ. Each dot represents data of a subcellular region of interest. Five cells were analyzed and three regions per cell were quantified. *** P < 0.001 (calculated using Student’s t-test).

COMMD3 regulates a group of cargo proteins independent of other COMMD proteins.

(A-D) Normalized surface levels of ITGA6 (A), GLUT-SPR (B), TfR (C) and LAMP1 (D) measured by flow cytometry in the indicated preadipocyte cell lines. Data of all cell samples were normalized to those of WT cells. Data of ITGA6 (n=3), GLUT-SPR (n=6), TfR (n=10) and LAMP1 (n=3) are presented as mean ± SD. ** P < 0.01; *** P < 0.001; n.s., P > 0.05 (calculated using one-way ANOVA).

Upregulation of COMMD3 in cells deficient in CCDC93 or VPS35L.

(A) Representative immunoblots showing protein expression in the indicated preadipocyte cell lines. (B-D) Normalized surface levels of TfR measured by flow cytometry in the indicated preadipocyte cell lines. Data of all cell samples were normalized to those of WT cells. Data are presented as mean ± SD of three biological replicates. ** P < 0.01; *** P < 0.001; n.s., P > 0.05 (calculated using one-way ANOVA).

The Commander-independent function of COMMD3 is mediated by its NTD.

(A) Top, diagram depicting the domain organization of COMMD3. Bottom: the structural model of COMMD3 (AlphaFold Protein Structure Database: AF-Q9UBI1-F1) showing the independently folded NTD and CTD. (B) Normalized surface levels of TfR measured by flow cytometry in the indicated preadipocyte cell lines. Data of all cell samples were normalized to those of WT cells. Data are presented as mean ± SD of three biological replicates. * P < 0.05; *** P < 0.001 (calculated using one-way ANOVA). (C) Diagrams of FL and truncated COMMD3 proteins used in proteomic experiments. The proteins were tagged with mCherry (mCh) and 3xFLAG (3xF). (D) Procedures of proteomic analysis to determine the interactomes of FL and truncated COMMD3 proteins. (E) Venn diagram showing the interactomes of COMMD3 proteins. (F) A scatter plot showing the fold-change of protein abundance over vector control in the interactomes of FL COMMD3 and the NTD of COMMD3. Selected proteins are labeled. (G) Representative immunoblots showing the interactions of ARF1 Q71L with COMMD3-NTD. HA-tagged ARF1 Q71L and 3xFLAG-tagged COMMD3-NTD were co-expressed in 293T cells. COMMD3-NTD and associated proteins were immunoprecipitated using anti-FLAG antibodies and detected using immunoblotting. The ARF Q71L mutant was used here because it adopts a GTP-bound configuration [77, 78].

The NTD of COMMD3 stabilizes ARF1.

(A) Representative SIM images showing the subcellular localization of ARF1 and COMMD3-NTD expressed in HeLa cells. HA-tagged ARF1 and 3xFLAG-tagged COMMD3-NTD were transiently expressed in HeLa cells and stained using anti-HA and anti-FLAG antibodies, respectively (scale bars: 5 µm). (B) Quantification of ARF1 and COMMD3-NTD co-localization based on SIM images, which were captured as in (A) and analyzed using ImageJ. Each dot represents data of an individual cell. In randomized samples, ARF1 images were rotated 90° clockwise, whereas COMMD3-NTD images were not rotated. *** P < 0.001 (calculated using Student’s t-test). (C) Representative immunoblots showing the ARF1-stabilizing effects of COMMD3-NTD. HA-tagged ARF1 Q71L and 3xFLAG-tagged COMMD3-NTD were transiently expressed in HeLa cells and their total expression levels were measured using immunoblotting. (D) Representative immunoblots showing protein expression in the indicated preadipocyte cell lines. (E) Normalized surface levels of TfR measured by flow cytometry in the indicated preadipocyte cell lines. Data of all cell samples were normalized to those of WT preadipocytes. Data are presented as mean ± SD of three biological replicates. *** P < 0.001; n.s., P > 0.05 (calculated using one-way ANOVA).

The COMMD3-ARF1 interaction is critical to the Commander-independent function of COMMD3.

(A) Left: the AlphaFold3-predicted structure of the COMMD3:ARF1 heterodimer visualized using ChimeraX1.8. The structural prediction was performed using COMMD3-NTD (a.a. 1-120, purple), ARF1 (pink), and GTP (green) as input. Right: key residues at the COMMD3-ARF1 binding interface. The CIF file of the structural model is included in Supplementary Dataset 1. (B) PAE heatmap of the structural model shown in (A). (C) Diagrams showing the residues mutated in a COMMD3-NTD mutant (COMMD3-NTD*, only a.a. 56-65 are shown). Mutated residues are shown in red. (D) Quantification of COMMD-NTD and NTD* stably expressed in preadipocytes based on SIM images, which were captured and analyzed as in Fig. 7A-B. Each dot represents data of an individual cell. MFI, mean fluorescence intensity. n.s., P > 0.05 (calculated using Student’s t-test). (E) Normalized surface levels of TfR measured by flow cytometry in the indicated preadipocyte cell lines. Data of all cell samples were normalized to those of WT cells. Data are presented as mean ± SD of three biological replicates. *** P < 0.001; n.s., P > 0.05 (calculated using one-way ANOVA). (F) Quantification of endogenous ARF1 in the indicated preadipocyte cell lines based on intensities of proteins on immunoblots quantified using ImageJ. Data of all samples were normalized to those of WT cells. Data are presented as mean ± SD of three biological replicates. * P < 0.05; n.s., P > 0.05; ** P < 0.01 (calculated using one-way ANOVA).

Model of the Commander-independent function of COMMD3 in endosomal trafficking.

For clarity, Retromer and other endosomal recycling regulators are not shown.