Huntingtin interacting protein 1 modulates the transcriptional activity of nuclear hormone receptors

Abstract

Internalization of activated receptors regulates signaling, and endocytic adaptor proteins are well-characterized in clathrin-mediated uptake. One of these adaptor proteins, huntingtin interacting protein 1 (HIP1), induces cellular transformation and is overexpressed in some prostate cancers. We have discovered that HIP1 associates with the androgen receptor through a central coiled coil domain and is recruited to DNA response elements upon androgen stimulation. HIP1 is a novel androgen receptor regulator, significantly repressing transcription when knocked down using a silencing RNA approach and activating transcription when overexpressed. We have also identified a functional nuclear localization signal at the COOH terminus of HIP1, which contributes to the nuclear translocation of the protein. In conclusion, we have discovered that HIP1 is a nucleocytoplasmic protein capable of associating with membranes and DNA response elements and regulating transcription.

Figure 1.

HIP1 is found in a nuclear subcellular fraction. (A) LNCaP cells were fractionated after Mibolerone or vehicle treatment. Nuclear (N) and cytosolic (C) fractions were resolved by SDS-PAGE (50 μg per lane), transferred to nitrocellulose, and blotted for the AR, HIP1, lamin B, and clathrin, illustrated here with representative blots. (B) LNCaP cells were grown in steroid-depleted media for 48 h and then treated with 10 nM Mibolerone or ethanol. HIP1 was detected using mouse mAb and the AR was detected using a rabbit polyclonal antibody. The nuclei were stained with DAPI. Bar, 70 μm. (C) The degree of translocation of the AR and HIP1 was quantitated by densitometric analysis of the blots. Data shown represent the means of five independent experiments ± SD.

Figure 2.

Association between HIP1 and the AR. (A) LNCaP cells were transiently transfected with 2 μg pcDNA3-AR and pcDNA3-Myc-HIP1 per 90-mm dish. Cell lysates were immunoprecipitated with an anti-AR antibody and immunoblotted with an anti-AR polyclonal antibody and an anti-HIP1 mAb. (B) HIP1 acts as a transcriptional regulator in GAL4-based assays. COS-7 cells were cotransfected with a GAL4-regulated luciferase reporter construct, and chimeric constructs encompassing the GAL4 DNA-binding domain fused to HIP1 and the following domains: ANTH (encompassing aa 1–310), ΔANTH (encompassing aa 320–1037), and I/LWEQ (encompassing aa 800–1037). Luciferase activity was measured 48 h after transfection on equal amounts of total cellular lysates that expressed comparable levels of the various GAL4 fusion proteins as assessed by anti-GAL4 immunoblot (inset). Graphed data represent the means of three independent experiments with error bars for the SD. (C) COS7 cells were transfected with 2 μg pcDNA3-AR per 9-cm dish. Lysates were prepared and the AR was immunoprecipitated from 300 μg of lysate supplemented with 10 μg of recombinant proteins. The left panel is a Coomassie-stained gel illustrating the equivalent loading of GST (lane 1), GST-ANTH domain (lane 2), GST-FxDxF/coiled coil domain encompassing aa 320–800 (lane 3), and GST-I/LWEQ domain (lane 4). Immunoprecipitates were blotted for AR, HIP1, and GST as indicated.

Figure 3.

HIP1 is a transcriptional coregulator for the AR. (A) COS7 cells were transfected with 50 ng pcDNA3-AR plus increasing quantities of pcDNA3-HIP1 (0, 50, 100, 200, and 400 ng) together with 100 ng of a pPSA luciferase reporter construct (pPSALuc). Cells were treated with10 nM Mibolerone ± 1 μM bicalutamide (Casodex) for 48 h and luciferase activity was normalized for transfection efficiency as determined by β-galactosidase assays and expressed relative to vehicle-treated singly transfected AR-positive cells. Data shown represent the means of three independent experiments ± SD. (insets) Lysates were prepared and run at 50 μg per lane before blotting for HIP1 and AR. (B) COS7 cells were transfected with 50 ng pcDNA3-AR plus increasing quantities of pcDNA3-HIP1 (0, 50, 100, 200, and 400 ng) together with 100 ng of a minimal ARE luciferase reporter construct (pARE₄-Luc). After 48 h, cells were lysed and luciferase assays were performed as described. Data shown represent the means of three independent experiments ± SD.

Figure 4.

HIP1 associates with ARE. (A) Schematic diagram of the PSA promoter and ChIP assay with the length of PCR products denoted by a black bar. (B) ChIP was performed in LNCaP cells using both HIP1 and AR antibodies over a 4-h androgen time course. PCR was performed for AREs I and III. (C) ReChIP assays were performed in LNCaP cells by reprobing AR immunocomplexes with AR and HIP1 antibodies after 2-h androgen treatment. AR-HIP1 association was analyzed at ARE I and a non-ARE–containing portion of the PSA promoter (AREX).

Figure 5.

Silencing HIP1 expression reduces the transcriptional activity and protein levels of the AR. (A) LNCaP cells were transfected with a combination of HIP1 siRNAs versus a control siRNA. After 48 h, cell lysates were blotted for HIP1, HIP1R, AR, PSA, and tubulin. (B) Relative expression of AR, actin, HIP1, HIP1R, and PSA genes in LNCaP cells analyzed by real-time RT-PCR after transfection with either scrambled control or HIP1 siRNA. The data represent experimental triplicates normalized to actin levels from cells treated with a scrambled control siRNA and the error bars correspond to the SD on this data. (C) LNCaP cells were transfected with HIP1 siRNA or a scrambled control siRNA. After 40 h, cells were treated with cycloheximide (CHX). Cells were lysed during the course of the following 8 h with Western blot analysis of AR levels. Lysates were also probed by Western blotting for tubulin as a loading control.

Figure 6.

The effects of the K56E/K58E double mutation on lipid binding by HIP1. (A) A sequence alignment of the putative α1-to-α2 loop region of HIP1 with that of other ANTH domain proteins based on the crystal structure of the CALM (Altschul et al., 1997; Ford et al., 2001). Note key lipid binding residues (red); sequence identity (red); mutated residues (red arrowheads); and predicted NLS using http://cubic.bioc.columbia.edu/cgi/var/nair/resonline.pl (blue amino acids). Sequence identifications: HsHIP1 (NP_005329), HsHIP1R (NP_003950), Sp putative clathrin coat assembly protein (NP_596345), ScSla2p (NP_014156), Xl Hip1-prov protein (AAH77182), and RnAP180 (CAA48748). Hs, Homo sapiens; Sc, Saccharomyces cerevisiae; Sp, Schizosaccharomyces pombe; Xl, Xenopus laevis; Rn, Rattus norvegicus. An I/LWEQ module sequence alignment prepared using CLUSTALW incorporating the predicted fourth α helix of the module is shown (McCann and Craig, 1999). Additional sequence identifiers: HsTalin1 (AAF27330) and SpSla2p (NP_594069). (B) Coomassie-stained gel of a sedimentation assay with phosphatidylinositol (PI) and brain (Folch) liposomes. P, pellet; S, supernatant. Liposomes were incubated with 5 μM of the indicated proteins: GST, GST-HIP1N (aa 1–310), GST-HIP1N K/E (aa 1–310 K56E/K58E). (C) Isolation of an enriched clathrin-coated vesicle fraction from LNCaP cells. LNCaP cells were transfected with pcDNA3 HIP1 R1005E or HIP1 K56E/K58E or vector alone. 48 h posttransfection they were disrupted by homogenization and CCV fractions were isolated using a protocol adapted from Hirst et al. (2004). Coomassie blue–stained gel (left) and Western blots of equal protein loadings of homogenate (H), high speed supernatant (HSS), and CCV fractions from the isolation procedure.

Figure 7.

Characterization of an NLS at the COOH terminus of HIP1. COS7 cells were transfected with GFP-HIP1R (aa 992–1009; A) or GFP-HIP1NLS (aa 992–1009; B) 48 h before imaging and fixation. Nuclei were stained with DAPI. GFP-HIP1R is predominantly cytosolic, whereas GFP-HIP1 has a nucleocytoplasmic distribution. Bars, 80 μm. (C) COS7 cells were transfected with GFPHIP1, GFPHIP1_R1005E, GFPHIP1R, and GFP alone and were fractionated to give nuclear (N) and cytoplasmic (C) fractions, which were resolved by SDS-PAGE (50 μg per lane) and blotted for GFP, lamin B, and clathrin. (D) The degree of translocation of GFP, GFPHIP1, GFPHIP1_R1005E, GFPHIP1R, and clathrin was quantitated by densitometric analysis. Data shown represent the means of five independent experiments ± SD.

Figure 8.

Effects of lipid binding and NLS mutations on the coregulator functions of HIP1. (A) COS7 cells were transiently transfected with pcDNA3-HIP1, HIP1 K56E/K58E, or HIP1 R1005E constructs, pcDNA3-AR and the pPSALuc reporter. Luciferase assays were performed after treatment with 10 nM Mibolerone and 1 μM Casodex. Data shown represent the means of three independent experiments ± SD. Lysates were blotted for HIP1 and AR as well as actin as a loading control. (B) HIP1 R1005E was transfected into LNCaP cells undergoing steroid depletion. Cells were treated with 10 nM Mibolerone for 2 h, fixed, and stained for the AR (red) and with a Myc antibody for HIP1 R1005E (green). Bar, 10 μm.