Domain Structure of the NRIF3 Family of Coregulators Suggests Potential Dual Roles in Transcriptional Regulation

Abstract

The identification of a novel coregulator for nuclear hormone receptors, designated NRIF3, was recently reported (D. Li et al., Mol. Cell. Biol. 19:7191–7202, 1999). Unlike most known coactivators, NRIF3 exhibits a distinct receptor specificity in interacting with and potentiating the activity of only TRs and RXRs but not other examined nuclear receptors. However, the molecular basis underlying such specificity is unclear. In this report, we extended our study of NRIF3-receptor interactions. Our results suggest a bivalent interaction model, where a single NRIF3 molecule utilizes both the C-terminal LXXIL (receptor-interacting domain 1 [RID1]) and the N-terminal LXXLL (RID2) modules to cooperatively interact with TR or RXR (presumably a receptor dimer), with the spacing between RID1 and RID2 playing an important role in influencing the affinity of the interactions. During the course of these studies, we also uncovered an NRIF3-NRIF3 interaction domain. Deletion and mutagenesis analyses mapped the dimerization domain to a region in the middle of NRIF3 (residues 84 to 112), which is predicted to form a coiled-coil structure and contains a putative leucine zipper-like motif. By using Gal4 fusion constructs, we identified an autonomous transactivation domain (AD1) at the C terminus of NRIF3. Somewhat surprisingly, full-length NRIF3 fused to the DNA-binding domain of Gal4 was found to repress transcription of a Gal4 reporter. Further analyses mapped a novel repression domain (RepD1) to a small region at the N-terminal portion of NRIF3 (residues 20 to 50). The NRIF3 gene encodes at least two additional isoforms due to alternative splicing. These two isoforms contain the same RepD1 region as NRIF3. Consistent with this, Gal4 fusions of these two isoforms were also found to repress transcription. Cotransfection of NRIF3 or its two isoforms did not relieve the transrepression function mediated by their corresponding Gal4 fusion proteins, suggesting that the repression involves a mechanism(s) other than the recruitment of a titratable corepressor. Interestingly, a single amino acid residue change of a potential phosphorylation site in RepD1 (Ser²⁸ to Ala) abolishes its transrepression function, suggesting that the coregulatory property of NRIF3 (or its isoforms) might be subjected to regulation by cellular signaling. Taken together, our results identify NRIF3 as an interesting coregulator that possesses both transactivation and transrepression domains and/or functions. Collectively, the NRIF3 family of coregulators (which includes NRIF3 and its other isoforms) may play dual roles in mediating both positive and negative regulatory effects on gene expression.

ACKNOWLEDGMENTS

We thank Sanford Shattil for plasmids expressing GFP-EnL and GFP-EnS, Richard Heyman for providing the retinoids, Gordon Fishell for help with fluorescence microscopy, and Fred Stanley for advice on graphic preparations.

This research was supported by NIH grant DK16636 (to H.H.S.) and NRSA postdoctoral fellowship award DK09581 (to D.L.). H.H.S is a member of the NYUMC Cancer Center (CA16087).