Small ubiquitin-like modifier-2 modification of retinoic acid receptor-alpha regulates its subcellular localization and transcriptional activity

Abstract

The retinoic acid receptor-alpha (Rara) gene is critical for germ cell development in the testis, as demonstrated by infertile Rara knockout male mice. The encoded protein for Rara (RARA) is expressed in both Sertoli cells and germ cells, but it is not always in the nucleus. Previously, all-trans retinoic acid (ATRA) was shown to increase the nuclear localization and transcriptional activity of RARA in Sertoli cells. Here, we identified a small ubiquitin-like modifier-2 (SUMO-2) modification as a novel posttranslational regulatory mechanism controlling the ATRA-dependent RARA subcellular localization and transcription. ATRA increased the SUMO-2 modification of RARA. In the presence of ATRA, lysine 166 (K166) and K171 of RARA were modified at a physiological concentration of SUMO-2, whereas in the absence of ATRA, K399 was the only site that was modified, but at a higher SUMO-2 concentration. However, K399 was critical for ATRA-controlled nuclear trafficking of RARA. In the presence of ATRA, a K399 mutation to arginine resulted in the cytoplasmic localization of K399R mutant, indicating that K166 and K171 sumoylations were inhibitory to nuclear localization. This may be due to SUMO/sentrin-specific peptidase 6 (SENP6) not being able to bind K399R mutant to desumoylate K166 and K171 in Sertoli cells, whereas it can bind RARA with intact K399. On the other hand, functional K166 and K171 sites for sumoylation were required for a full transcriptional activity, when K399 was intact. These results together suggest that both K166 and K171 sumoylation and desumoylation are critical for optimal RARA function.

Figure 1

SUMO-2 is highly expressed in Sertoli cells, and RARA has four potential SUMO acceptors. Panel A, The translated protein sequence of Sumo-2 cDNA isolated from a LacZ-positive yeast clone, identified as containing a protein that strongly interacted with RARA. The SUMO-2 sequence was identical to mouse (mSUMO-2), rat (rSUMO-2), and human (hSUMO-2) sequences. The potential sumoylation sites are boxed, and the consensus sequence is indicated. Panel B, Relative expression of each Sumo was determined using real-time RT-PCR in rat Sertoli cells, rat germ cells, and testes from VAD rats. Expression was normalized to the internal control, the ribosomal protein S2 expression. Results are expressed as means ± sem from n = 3 independent RNA samples, with each real-time PCR conducted in triplicate. Different letters denote a significant difference from each other at P < 0.05. Panel C, Four potential sumoylation sites of RARA, K399, K171, K166, and K147, predicted by SUMOplot software, arranged according to predicted score from high to low. Panel D, Schematic of the hRARA, showing amino acid position (Pos.) K399 in the ligand-binding domain (LBD), K166 and K171 in the nuclear localization signal (NLS), and K147 in the DNA-binding domain (DBD). A–F, Modular domains in RARA.

Figure 2

RARA is covalently modified by SUMO-2 in primary Sertoli cells and MSC-1 and transfected COS-7 cells. A and B, Extracts of primary Sertoli cells (1°SC) or MSC-1 cells, treated with ATRA (1 μm), FSH (25 μm), or ATRA and FSH, as indicated, for 3 h, were loaded onto a column with antibody cross-linked to protein A agarose beads (IP). The eluate was TCA precipitated (+) or not (−), as indicated, and immunoblotted (IB) with anti-RARA antibody or anti-SUMO-2/3 antibody. C and D, COS-7 cells were transiently transfected with pFLAG-RARA cDNA and pFLAG-Sumo-2 cDNA constructs, lysed, immunoprecipitated (IP) with either anti-RARA or anti-SUMO-2/3 antibody cross-linked to True Blot antirabbit IP beads, followed by immunoblotting (IB) with anti-SUMO-2/3 or anti-RARA antibody. E, MSC-1 or COS-7 cells were treated with vehicle (−) or ATRA (1 μm) (+) for 24 h, lysed, and subjected to immunoblotting (IB) with anti-RARA antibody. IgG, IgG contaminant in the eluate from the large-scale IP; 150 kDa C and 70 kDa C, protein complexes around 150 and 70 kDa; M, 10% of input, showing RARA or SUMO-2/3. Negative controls included IgG beads incubated with the lysate and no primary antibody (labeled C in C) or BSA loaded on a column packed with anti-RARA antibody cross-linked to protein A agarose beads and immunoblotted (IB) with anti-RARA (labeled BSA in Fig. 2B). Equal loading was determined by the Coomassie-stained membrane (Com.) for E. All experiments were repeated at least three times.

Figure 3

Mutations of SUMO-2 acceptor lysines on RARA influenced the extent of sumoylation. COS-7 cells were transfected with a varying ratio, as indicated, of pHis-Sumo-2 (pSumo-2) to pFLAG-RARA WT (pRARA) or pFLAG-RARA K399R [pRARA (K399R)] mutant cDNA constructs (A–D) or a set amount of pHis-Sumo-2 (pSumo-2) and pFLAG-RARA WT or various mutants of RARA cDNA constructs, 4:4 ratio (E). Cells were treated with either vehicle or ATRA (1 μm), lysed, and immunoblotted (IB) with anti-RARA antibody (C and D) or anti-SUMO-2/3 (A) or anti-FLAG M2 antibody (B and E). Equal loading was determined by the Coomassie blue dye-stained membranes (Com.). 150 kDa C, protein complexes around 150 kDa. All the experiments were repeated at least three times.

Figure 4

K399 sumoylation site was the primary site that influenced the RARA subcellular localization. COS-7 cells were transfected with the pFLAG-RARA WT (A and I) or pFLAG-RARA mutant cDNA constructs (B–D, E–F, J–L, and M–P), treated with either vehicle (A–H) or ATRA (1 μm) (I–P) for 30 min, fixed, and analyzed by immunofluorescence with anti-FLAG M2 antibody. Bar in A, 60 μm for A–P. The images shown represent at least 65% of cells.

Figure 5

Triple mutations in K166, K171, and K399 enhanced the transcriptional activity of RARA. A and B, COS-7 cells were cotransfected with empty vector or a varying ratio, as indicated, of pHis-Sumo-2 (pSumo-2) to pFLAG-RARA (pRARA) cDNA. C and D, COS-7 cells were cotransfected with pFLAG-RARA WT or mutant RARA cDNA constructs. A–D, pcDNA-Rxra cDNA, pRARE-tk-Luc cDNA and pcDNA-β-gal cDNA constructs were also transfected. Cells were treated with either vehicle (white bars) or ATRA (1 μm) (A, C, and D) (black bars) or Am580 (0.1 μm) (B) (black bars) for 24 h and analyzed by luciferase reporter assay. Results are expressed as means ± sem from n = 3 independent experiments, with each conducted in triplicate. Different letters denote a significant difference from each other at P < 0.05. All the experiments were repeated at least three times.

Figure 6

SENP6 interacts with RARA in primary Sertoli cells. A and B, Relative expression of each Senp gene was determined using real-time RT-PCR in rat Sertoli cells (A) and rat germ cells (B). Expression was normalized to the ribosomal protein S2 expression. Results are expressed as means ± sem from n = 3 independent RNA samples, with each conducted in triplicate. Different letters denote a significant difference from each other at P < 0.05. C–H, Primary Sertoli cells were treated with vehicle (C–E) or ATRA (1 μm) (F–H) for 30 min, fixed, and immunostained with anti-RARA (C and F) and anti-SENP6 (D and G) antibodies. 4′,6-Diamidino-2-phenylindole stained the nucleus. Bar in C, 60 μm for C–H. I and J, Primary Sertoli cells (I) or COS-7 cells (J) transiently transfected with pFLAG-RARAWT cDNA or pFLAG-RARA K399R cDNA constructs were lysed, immunoprecipitated (IP) with anti-SENP6 antibody cross-linked to True Blot antigoat IP beads, followed by immunoblotting (IB) with anti-RARA (I) or anti-FLAG M2 (J) antibody. C, a negative control lane showing IgG beads incubated with the lysate and no primary antibody; IgG, IgG contaminant in the eluate from IP; M, 10% of input, showing RARA. All the experiments were repeated at least three times. K, Working model for RARA, SUMO-2, and SENP6 interaction. In the absence of ATRA, sumoylation on K399 may be very low (spiked S) at a physiological concentration of SUMO-2. Binding of ATRA (R) to RARA (rectangle) induces a conformational change to RARA (oval) that reveals K171 and K166 sumoylation sites for SUMO-2 (S) modification. SENP6 (Se6) bound to K399 in the presence of ATRA could desumoylate K171 and K166 before RARA enters the nucleus. CoA, Coactivator; tick marks represent K166, K171, and K399 sites.