ENTREP/FAM189A2 encodes a new ITCH ubiquitin ligase activator that is downregulated in breast cancer

Abstract

The HECT-type ubiquitin E3 ligases including ITCH regulate many aspects of cellular function through ubiquitinating various substrates. These ligases are known to be allosterically autoinhibited and to require an activator protein to fully achieve the ubiquitination of their substrates. Here we demonstrate that FAM189A2, a downregulated gene in breast cancer, encodes a new type of ITCH activator. FAM189A2 is a transmembrane protein harboring PPxY motifs, and the motifs mediate its association with and ubiquitination by ITCH. FAM189A2 also associates with Epsin and accumulates in early and late endosomes along with ITCH. Intriguingly, FAM189A2 facilitates the association of a chemokine receptor CXCR4 with ITCH and enhances ITCH-mediated ubiquitination of CXCR4. FAM189A2-knockout prohibits CXCL12-induced endocytosis of CXCR4, thereby enhancing the effects of CXCL12 on the chemotaxis and mammosphere formation of breast cancer cells. In comparison to other activators or adaptors known in the previous studies, FAM189A2 is a unique activator for ITCH to desensitize CXCR4 activity, and we here propose that FAM189A2 be renamed as ENdosomal TRansmembrane binding with EPsin (ENTREP).

Keywords: CXCR4; ENTREP; FAM189A2; ITCH; breast cancer.

Figure 1. The expression of FAM189A2/ENTREP

1. A

   The mRNA expression of FAM189A2/ENTREP in the primary human mammary epithelium HMEC and human breast cancer cell lines. In qRT‐PCR analyses, the expression was normalized to the immortalized normal human mammary epithelium HMEC4tertshp16. Relative expression ratios shown as mean + SD from three biological replicates.

2. B

   Relative expression of FAM189A2/ENTREP in the normal (white) and cancer (green) tissues of breast (Curtis Breast), lung (Okayama Lung), colorectal (TCGA Colorectal), and head and neck (Estilo Head‐Neck). Data were downloaded from the Oncomine database (http://www.oncomine.org). P‐values obtained by Student’s t‐tests and the number of cases are listed on the top and bottom of figures, respectively. P < 0.05 was considered as statistically significant. Box‐whisker plot represents the interquartile range (25^th and 75^th percentiles) as a box and the median as a line. The maximum and minimum values within 1.5 × interquartile range are shown as whiskers. Outlier data are plotted as dots.

3. C

   Relative expression of FAM189A2/ENTREP in the primary (green) and the metastatic sites (blue) including lymph node, bone, liver, lung, soft tissues of prostatic cancer (Grasso Prostate) (left) and in the normal (white), primary (green), and metastatic sites (three cases of sentinel lymph node) (blue) of breast cancer (TCGA Breast) (right). Note that only three cases of the metastatic breast cancer were available in the dataset. Data were downloaded from the Oncomine database (http://www.oncomine.org). P‐values obtained by Student’s t‐tests and the number of cases are listed on the top and bottom of figures, respectively. P < 0.05 was considered as statistically significant. Box‐whisker plot represents the interquartile range (25^th and 75^th percentiles) as a box and the median as a line. The maximum and minimum values within 1.5 × interquartile range are shown as whiskers. Outlier data are plotted as dots.

4. D–F

   Relapse‐free survival (RFS) of breast cancer patients. Data were downloaded from the Kaplan–Meier plotter (https://kmplot.com/analysis/). Statistical analysis is described in the Materials and Methods.

Figure EV1. The Oncomine database analyses of FAM189A1, ENTREP/FAM189A2, and FAM189B expression in various types of cancer

Number indicates the number of datasets of gene expression analyses.

Figure 2. ENTREP associates with ITCH

1. A

   Schema of the expression vectors of ENTREP and ITCH. SS, signal sequence; CD20, CD20‐homology extracellular domain; TM, transmembrane domain; C2, C2 domain; WW, tryptophan‐tryptophan domain; HECT, the HECT domain; C830A, the catalytically inactive mutation. Red bar, cytoplasmic domain of ENTREP and a bait for the yeast two‐hybrid screening.

2. B, C

   The immunoprecipitation analysis. Co‐precipitation was impaired by either the WW domain deletion of ITCH or the mutation in the PPxY motifs of ENTREP. Data shown are representative of at least two independent experiments. IP, immunoprecipitation; IB, immunoblot.

3. D

   The immunofluorescence staining of Cos7 cells transiently transfected with indicated vectors. Merged, merged images of FLAG, Halo and DAPI staining. Scale bar, 20 mm. Data shown are representative of at least three independent experiments.

4. E

   Representative images of the proximity ligation assay (PLA) using Cos7 cells transfected with indicated vectors. PLA signals, granular dots of green color, were presented with DAPI images. Arrowheads, the plasma membrane. Scale bar, 10 mm.

5. F

   The box‐and‐whisker plot presentation of the PLA. A minimum of 10 cells per condition were counted from two independent experiments. P‐values obtained by Student’s t‐tests and P < 0.05 was considered as statistically significant. *P < 0.05. Box‐whisker plot represents the interquartile range (25^th and 75^th percentiles) as a box and the median as a line. The maximum and minimum values within 1.5 × interquartile range are shown as whiskers.

Figure 3. ITCH ubiquitinates ENTREP

1. The nickel‐pull down assay using myc/HIS‐tagged ENTREP along with HA‐tagged ubiquitin. Arrows indicate HA‐ubiquitin‐incorporating ENTREP. Arrowhead indicates ENTREP without ubiquitination. Data shown are representative of three independent experiments.

2. The ubiquitination of ENTREP. The immunoprecipitated samples using anti‐FLAG antibody under the denaturing condition were separated by the SDS‐PAGE. The gel area corresponding to band #A, #B, and #C of ENTREP‐FLAG were applied to the ubiquitin‐AQUA/PRM analysis. The absolute number of each ubiquitins of ENTREP #A, #B, and #C were calculated by subtraction of those of the corresponding area of the control vector sample (Appendix Table S2 for raw data of the ubiquitin‐AQUA/PRM analysis). Based on the ratios of each ubiquitins to the total ubiquitin of indicated gel area, which are showed as Circular graphs (ENTREP #A, #B, and #C), the number of ENTREP‐linked ubiquitin (Ub) is summarized as schema: #A is modified with one Ub, and about two thirds of #B molecules and half of #C molecules are modified with two and three of single Ubs (multi‐monoubiquitin), respectively. The result was based on three biological replicates.

Figure EV2. Ubiquitination of ENTREP

1. Sample preparation for the ubiquitin‐AQUA/PRM analysis. The immunoprecipitated samples using anti‐FLAG antibody under the denaturing condition were separated by the SDS‐PAGE and served for the immunoblot analyses and the Coomassie Brilliant Blue (CBB) gel staining. The gel area corresponding to the band #A, #B, and #C of the immunoblot were excised from the CBB‐stained gel and used for ubiquitin‐AQUA/PRM and shotgun MS analyses. The immunoblot image of FLAG IP/FLAG IB is the same with that of Fig 3B. We analyzed samples from three biological replicates (#1‐3 of the control samples and #4‐6 of ENTREP samples). The raw data of the ubiquitin‐AQUA/PRM analysis are listed on Appendix Table S2 and the result of the shotgun MS is in Appendix Fig S3.

2. The cycloheximide chase assay. Twenty‐four hours after transfection, HEK293T cells were incubated with 50 mg/ml cycloheximide for the indicated time periods and served for the immunoblot analyses. The immunoblot analyses were carried out using six independent samples, and their blot bands were semi‐quantified using ImageJ software. The relative FLAG‐ENTREP expression levels were calculated as a ratio of GAPDH‐adjusted FLAG‐ENTREP at each time points and presented as a mean + SD from six biological replicates. P‐values obtained by Student’s t‐tests and P < 0.05 was considered as statistically significant. *P < 0.05. Note that, in the immunoblot analyses, the volume of Halo‐ITCH transfected samples applied was a half of Halo‐transfected samples applied, as indicated by GAPDH.

Figure 4. ENTREP associates with EPN1

1. The immunoprecipitation analysis. The cytoplasmic ENTREP (wild or mut1+2 mutant) co‐precipitated with EPN1myc in the presence of ITCH wild or C830A. The blot bands were semi‐quantified using ImageJ software. ENTREP ratio was calculated as a ratio of ENTREP in the myc‐coprecipitaed sample to that of the total lysate. The ratios shown as mean ± SD from three biological replicates. P‐values were obtained by Student’s t‐tests and P < 0.005 was considered as statistically significant. No significant difference was observed.

2. Representative images of the proximity ligation assay (PLA) using Cos7 cells transfected with indicated vectors. PLA signals, granular dots of green color, were presented with DAPI images. Scale bar, 10 mm.

3. The box‐and‐whisker plot presentation of the PLA. A minimum of 10 cells per condition were counted from two independent experiments. P‐values obtained by Student’s t‐tests and P < 0.05 was considered as statistically significant. *P < 0.05. Box‐whisker plot represents the interquartile range (25^th and 75^th percentiles) as a box and the median as a line. The maximum and minimum values within 1.5 × interquartile range are shown as whiskers.

4. The expression of endogenous ENTREP in HMEC. ENTREP co‐localized with Clathrin, RAB5, RAB7, and RAB11. Inserts, the ring‐shaped structures of the marked areas. The images are representative of at least two independent experiments. Scale bar 10 mm.

5. Co‐localization of ENTREP with ITCH and EPN1. Cos7 cells transfected with indicated vectors were stained with specific antibodies and observed under a laser confocal microscope. Arrowheads, full‐length ENTREP (wild or mut1+2 mutant) expression at the plasma membrane. Inserts, the high‐magnification images of the marked area indicating the co‐localization of ENTREP‐FLAG, EPN1myc and Halo‐ITCH wild. The images are representative of at least two independent experiments. Scale bar 10 mm.

Figure 5. ENTREP enhances the ubiquitination of CXCR4

1. Schema of the expression vectors of ITCH deletion mutants and the carboxyl‐terminal tail of CXCR4. TM7, the seventh transmembrane domain of CXCR4; HA, hemagglutinin tag; DsRed, DsRed‐momoner (Clontech/TAKARA).

2. The immunoprecipitation analysis. Cytoplasmic ENTREP co‐precipitated with either ITCH C830A, delWW12, or delWW34 but not with delWW14. Data shown are representative of at least two independent experiments.

3. The immunoprecipitation analysis. HA‐CXCR4SS‐DsRed hardly co‐precipitated with either ITCHC830A or delWW mutants. HA‐CXCR4DD‐DsRed co‐precipitated with ITCH C830A and delWW12 but not with delWW14 and delWW34. Data shown are representative of at least two independent experiments.

4. The immunoprecipitation analysis. HA‐CXCR4DD‐DsRed co‐precipitated with ENTREP when co‐transfected with either ITCH C830A or delWW12. Data shown are representative of at least two independent experiments.

5. The immunoprecipitation analysis. The increased transfection of ENTREP vector enhanced precipitation of HA‐CXCR4DD‐DsRed with ITCH C830A. Data shown are representative of at least two independent experiments.

6. The nickel‐pull down assay using indicated vectors along with HISx6‐tagged wild‐type ubiquitin vector. HA‐tagged DsRed was used as a control for HA‐CXCR4DD‐DsRed. Arrowhead, FLAG‐tagged ITCH (wild or C830A); arrow, FLAG‐ENTREP wild cyt. Data shown are representative of at least two independent experiments.

7. Co‐localization of ENTREP and CXCR4 in the endosome. Cos7 cells were transfected with ENTREP‐FLAG, CXCR4‐DsRed, and Halo‐ITCH wild vectors and treated with CXCL12. ENTREP‐FLAG and CXCR4‐DsRed co‐localized in the RAB7‐positive endosome. Inserts, the high‐magnification images of the marked area indicating the colocalization of ENTREP‐FLAG, CXCR4‐DsRed, and RAB7. Scale bar 10 mm. The image is representative of at least two independent experiments.

Figure EV3. The ITCH WW domain responsible for the association with ENTREP and CXCR4

1. A

   Schema of the expression vectors of ITCH deletion mutants.

2. B, C

   The immunoprecipitation analysis. FLAG‐ENTREP wild cyt co‐precipitated with Halo‐ITCH delWW123 but not with delWW124, whereas HA‐CXCR4DD‐DsRed co‐precipitated equally with either Halo‐ICTH delWW123 or delWW124. Data shown are representative of at least two independent experiments.

Figure 6. ENTREP participates in CXCL12‐induced endocytosis of CXCR4

1. The immunoblot analyses of MCF‐7 ENTREP‐KO and parent MCF‐7 cells. Note that the amount of CXCR4 protein was not changed. The immunoblot analyses were carried out using six independent samples, and their blot bands were semi‐quantified using ImageJ software. CXCR4/GAPDH ratio was calculated as a ratio of GAPDH‐adjusted CXCR4 expression. CXCR4/GAPDH ratios shown as mean ± SD from at least four biological replicates. P‐values were obtained by Student’s t‐tests and P < 0.005 was considered as statistically significant. No significant difference was observed.

2. Cell proliferation analysis. Cells were cultured in DMEM supplemented with 10% FBS and counted at day1 to day3. Data are shown as mean ± SD from six biological replicates. P‐values were obtained by Student’s t‐tests and P < 0.005 was considered as statistically significant. NS, not significant.

3. The expression of endogenous CXCR4 in lentivirally transduced MCF‐7 ENTREP‐KO cells. CXCL12 (+) and (−) indicate the presence and absence of CXCL12 treatment. In CXCL12‐treated ENTREP‐EGFP expressing cells, a non‐negligible amount of CXCR4 was observed in the cytoplasm, in which CXCR4 was overlapped with ENTREP‐EGFP. CXCR4 expression on the cell surface still remained (arrows). Scale bar 10 mm. The images are representative of at least three independent experiments.

Figure 7. ENTREP fine‐tunes CXCR4 function

1. Dot plot presentation of the chemotaxis analyses using the Boyden chamber. CXCL12 (+) and (−) indicate the presence and absence of CXCL12 as an attractant in the bottom chambers. ENTREP‐knockout enhanced the chemotaxis toward CXCL12, which was blocked by pre‐treatment with AMD3100 (upper panel). The doxycyclin (DOX)‐induced ENTREP expression suppressed the chemotaxis of mouse 4T1‐Luc cells toward CXCL12 (lower panel). Ratio of migrated cells into the bottom chamber against the applied cells on the top chambers was normalized by the results of non‐treated cells (n = 7, upper panel, and n = 9, lower panel, from at least two independent experiments, respectively). P‐values were obtained by Student’s t‐tests and P < 0.005 was considered as statistically significant. Insert, the immunoblot analysis of the DOX‐induced ENTREP expression in 4T1‐Luc cells. NS, not significant.

2. Dot plot presentation of mammosphere assay (left panel) and the images of mammosphere (right panel). Each mark in dot plot presentation indicates the number of mammosphere from 3,000 cells per well in an ultra‐low attachment plate from two independent experiments. To estimate the effect of CXCL12/CXCR4 pathway, 100 ng/ml of CXCL12 and 12 μM of AMD3100 were applied as indicated. For statistical analysis, variance was assessed using two‐way ANOVA and significance was calculated using Tukey post hoc test correcting for multiple comparison. NS, not significant. Scale bar in the images of mammosphere 100 mm.

Figure EV4. Comparison of NDFIP1, N4BP1, and FAM189 family

1. The Oncomine database analysis of NDFIP1, NDFIP2, and N4BP1 expression in various types of cancer. Number indicates the number of datasets of gene expression analyses.

2. Phylogenetic analysis of FAM189A1, ENTREP/FAM189A2, FAM189B, NDFIP1, and N4BP1. The coding DNA sequences of these genes were analyzed using Phylogeny.fr software (https://www.phylogeny.fr/). Analyzed sequences were as follows: human FAM189A1, NM_015307.1; mouse Fam189a1, NM_183087.4; chicken Fam189a1, XM_025154007.1; dog Fam189a1, XM_025438066.1; human FAM189A2, NM_001127608.2; mouse Fam189a2, NM_001114174.1; chicken Fam189a2, XM_424828.6; dog Fam189a2, XM_022421428.1; human FAM189B, NM_006589.3; mouse Fam189b, NM_001014995.2; chicken Fam189b, XM_025143513.1; dog Fam189b, XM_005622739.2; human N4BP1, NM_153029.4; mouse N4bp1, NM_030563.2; chicken N4bp1, NM_001030570.1; dog N4bp1, XM_022411581.1; human NDFIP1, NM_030571.4; mouse Ndfip1, NM_001355749.1; chicken Ndfip1, XM_414658.5; dog Ndfip1, XM_022408883.1.

3. The comparison of PPxY motif. FAM189A1 and FAM189B contain the overlapped PPxY sequences which are separated with PPPY by 40 aa and 28 aa, respectively.

Figure 8. Schematic presentation of the proposed function of ENTREP

Plasma membrane‐localized ENTREP associates with ITCH by binding to its WW domain. ITCH modifies ENTREP mainly through the attachment of multi‐monoubiquitin. In addition, ENTREP enhances the ITCH association and polyubiquitination of ligand‐stimulated CXCR4, which leads to the attachment of endocytic adaptors. In collaboration with ENTREP‐associated EPN1 and CXCR4‐attached endocytic adaptors, the ENTREP‐ITCH‐CXCR4 ternary complex is efficiently endocytosed in clathrin‐coated endosomes, leading to desensitization of CXCR4.