The RNA coregulator SRA, its binding proteins and nuclear receptor signaling activity
Abstract
Nuclear receptor (NR) coregulators are key modulators of hormone signaling. Discovery of steroid receptor RNA activator (SRA), a coregulator that is active as a RNA, transformed thinking in the field of hormone action. The subsequent identification of SRA-binding coregulator proteins, including p68, SHARP and more recently SLIRP, has provided important insight into SRA's mechanism of action and potentially offers new opportunities to target NR signaling pathways for therapeutic gain. Here we outline advances in the field of NR coregulator biology, with a bias on recent progress in understanding SRA-protein interactions. © 2008 IUBMB IUBMB Life, 60(3): 159–164, 2008
THE NUCLEAR RECEPTOR SUPERFAMILY AND COREGULATORS
NRs are ligand-inducible transcription factors that regulate the expression of genes involved in metabolism, development and reproduction1, mediating the effects of sex steroids (progestins, estrogens, and androgens), glucocorticoids, mineralocorticoids, vitamin D, thyroid, and retinoid hormones, as well as a variety of metabolic ligands. The NR superfamily comprises 3 subclasses: (i) Endocrine—including estrogen (ER), androgen (AR), glucocorticoid (GR), mineralocorticoid (MR), thyroid (TR), progesterone (PR), and vitamin D (VDR) receptors; (ii) Adopted Orphan—for which ligands have been identified including peroxisome proliferator activated receptors (PPARs), liver X receptor, farnesoid X receptor; and (iii) Orphan—receptors with no known ligand, including liver receptor homolog-1 and estrogen related receptor. NRs are structurally related and are generally characterised as comprising an amino-terminal activation function (AF-1) domain; DNA-binding domain; a hinge region; and a carboxy-terminal ligand-binding domain (LBD) containing a second AF-2 domain. The pathway of ligand binding, receptor activation and nuclear translocation (for GR, MR, AR, and PR), dimerization and binding of cognate DNA response elements have been well characterized for many of the individual NRs1-3.
The discovery of NR coregulators that are recruited to receptor complexes modulating their activity has dramatically changed our understanding of hormone action. The past decade has revealed a host of such molecules that, in response to ligand, are either recruited or expelled from complexes with NRs and their cognate response elements, communicate with the general transcriptional machinery and RNA polymerase II, modulating NR directed transcription1. Specific coregulators possess enzymatic activity that can alter chromatin structure (histone acetylase) or methylation status (methyl transferase), act as kinases, ubiquitin ligases or have ATPase activity. Generally, coactivators facilitate transactivation while corepressors reduce the transcriptional rate of NR target genes. Significantly, the relative ratio of coactivator to corepressor molecules have been shown to alter the responsiveness of individual tissues to both agonist and antagonists of NR activity4. The biological significance of coregulator activities have been highlighted in knockout studies where the majority are embryonically lethal5. Viable knockouts and transgenic animal studies have characterized what might be considered subsets of NR action. For example, female mice lacking SRC-3 have reduced fertility and attenuated mammary gland development6 while SRC-3 over expressing mice are characterized by having mammary gland hyperplasia and accelerated differentiation with high incidence of mammary tumor development7. Differential expression of coregulators can have significant clinical relevance as, again using SRC-3 as an example, its over expression is associated with tamoxifen resistance in breast cancer patients8. Together these and many other studies underline the fact that changes to NR coregulator expression and activities have biological consequences on par with those of the receptors they modulate.
SRA—A REMARKABLE RNA COACTIVATOR
Until 1999, all NR coregulators identified were proteins. However, with the isolation of SRA9 a unique RNA coactivator molecule was uncovered (Fig. 1). Several laboratories have demonstrated SRA's ability to coactivate a range of NRs including ERα and β, AR, GR, RARα, PPARδ and γ, TR, and VDR9-12. Extensive studies have demonstrated that SRA-mediated NR coactivation does not require the expression of a SRA protein9.
Participation of SRA in nuclear receptor complexes and signaling. Numerous coregulators have been identified that participate in nuclear receptor (NR) signaling. Typically in the absence of ligand, corepressors, such as SHARP and SLIRP, bind NRs facilitating the recruitment of enzymes, including HDACs that act on chromatin to repress transcription of target genes. In the presence of ligand (L) corepressors are replaced by coactivators including SRC-1 and p300 in NR complexes, which direct the activities of cellular transcriptional machinery, including RNA polymerase II, up regulating target gene expression. In addition to protein participants in NR signaling is SRA. This RNA coregulator is itself the target of both corepressor and coactivator molecule binding and is thought to act as a scaffold bringing together NRs, coregulators and elements of the cell's transcriptional machinery at NR target genes. The ability of SRA to interact with other coregulators is further influenced by the activities of pseudouridine synthase molecules that differentially convert uridine (U) to pseudouridine (Ψ) at discrete sites within SRA transcripts. The presence of both NRs and SRA binding coregulators including SLIRP at the mitochondria suggests a role for these complexes in this organelle.
The SRA gene is well conserved across species and is composed of 5 exons in human, rat, and mouse genomes. Several alternative RNA products of the SRA gene have been identified13. Initially three SRA isoforms were described, with unique 5′ and 3′ ends but a shared common “core” domain9. Protein initiation and polyadenylation sequences were identified in these SRA isoforms however a reading frame of no greater than 162 amino acids (aas) was predicted. Since its initial discovery, human SRA isoforms containing larger open reading frames have been found14. For example, SRA1 codes for a 236 aa protein, referred to as SRAP, the carboxy-terminal 162 aas of which are identical to that theoretically coded for by the SRA sequence identified by Lanz and coworkers but with an additional 73 amino-terminal aas. Sequencing data base analyses predict that over 20 chordata species express a SRAP protein13 and antisera raised against human SRAP was able to detect appropriately sized proteins in cow, rabbit, chicken, pig, sheep, and avian species15. Consistent with the presence of two initiating methionines in the SRA1 transcript, two proteins of 31/32 kDa have been detected, with SRAP present in both the nucleus and cytoplasm. As discussed later, differential splicing of the SRA transcript may generate a noncoding RNA that is the target for a range of RNA-binding coregulators and a translated product whose expression may be important in human breast cancer16.
SRA—AN EMERGING ROLE IN HUMAN TUMORIGENESIS?
Multiple lines of evidence implicate SRA in human tumorigenesis. Elevated and/or aberrant SRA expression has been reported in cell lines and human breast, ovary, and uterine tumors17-19. In studies of breast cancer samples, expression of an exon 3 deletion mutant of SRA (SRA-Del) but not the wild type molecule correlated with a higher tumor grade17, 20. Notably, mutation of SRA stem-loop structure 10, encompassed by the SRA-Del deletion, results in a 50% reduction in transactivation compared with wild type SRA in transient transfection assays21 There is mounting evidence that SRAP may also play a role in breast tumor growth. Recent data suggests that patients with SRAP positive tumors have lower recurrence rates and improved outcomes22. Consistent with this observation are results showing that over expression of SRAP reduces estrogen signaling activity in the ER positive MCF-7 breast cancer cell line22 although specific down regulation of SRA in the same line resulted in increased pS2 expression and a mild increase in proliferation23.
Interestingly, transgenic mice overexpressing SRA demonstrate increased mitosis and elevated cell death of mammary epithelium but no apparent increase in tumor incidence18. When SRA transgenic mice were crossed with MMTV-ras mice, surprisingly, there was reduced frequency of mammary tumor development compared with MMTV-ras single transgenic animals. Taken together, these data suggest that SRA is an important regulator of mammary epithelial cell growth, but that the roles of SRA and SRAP in tumorigenesis are yet to be fully elucidated.
MECHANISM OF SRA ACTION AS AN RNA COACTIVATOR
Secondary structure predictions suggest the existence of multiple stem-loops within SRA. Extensive mutational analysis has shown that some of these stem-loops are critical for SRA's coactivation activity21. In fact, it appears multiple RNA substructures work together to effect SRA's overall coactivator function. Initial studies indicated that SRA coactivates a range of NR activities, including ERα and β, PR, and GR in a ligand-dependent manner. SRA could also augment ERα activity in a ligand-independent manner through its AF-1 domain involving MAPK but this was not the case for ERβ10. Over expression of SRA1 in MCF-7 cells generated contrasting results22. In stable clones constitutively expressing SRAP, ER reporter gene activity was lower than in control cells, however, when expression of the endogenous ER target PR was assessed, its levels were elevated22. Given the existence of both coding and non-coding SRA RNAs, and their opposing activities, it has been proposed that differential splicing of its transcripts may regulate the balance between coding and functional non-coding RNAs and the overall effect of SRA gene expression16.
SRA-BINDING PROTEINS INVOLVED IN NR SIGNALING
The identification of RNA-binding domains within multiple coregulators, including SHARP24, PGC-125, CoAA, CoAM26, p6827, p72, CAPERα, and β28, has generated much interest in understanding RNA-protein interactions in NR signaling (Fig. 1). It also raised the possibility that some coregulators could target specific components of the transcriptional machinery as a critical part of their function. A number of coregulators with RNA recognition motif (RRM) RNA-binding domains influence transcript splicing (e.g., CoAA, CAPERα, and β), whilst others appear to target SRA and more directly impact on NR transactivation. For example, SHARP (SMRT/HDAC1 associated repressor protein) is a NR corepressor that interacts with SRA in vitro and contains an RNA-binding domain comprised of three RRMs24. These RRMs are required by SHARP to repress SRA-augmented estrogen-induced transactivation24. Another ER coregulator which binds SRA in vitro and copurifies with SRA from cell extracts is p7229. However, to date the specific details of how each of these proteins interacts with SRA, and which stem-loop is targeted, remain to be determined.
Adding an additional layer of complexity to the regulation of NR activity, two pseudouridine synthases (PUS1p and 3p) have been reported to bind and pseudouridylate SRA transcripts11, 30. While both act on SRA they have slightly contrasting abilities to isomerase the RNA acting on different numbers of uridines. Further, the order in which SRA is processed by the individual PUS enzymes can alter patterns of pseudouridlyation potentially influencing the binding of other coactivators and repressors. Significantly U206, present within stem loop structure 521 of SRA, is a target for PUS proteins. Its mutation to adenine results in SRA becoming hyperpseudouridylated and a transcriptional repressor30. Underlining the clinical significance of these enzymes, mutation of human PUS1p causes mitochondrial myopathy and sideroblastic anemia31 leading to the speculation that abnormalities associated with these mutations may be the consequence of defective SRA-NR signaling32. The existence of this posttranscriptional mechanism of SRA regulation further highlights its importance in regulating NR activity.
SRA-BINDING PROTEINS INVOLVED IN NON-NR SIGNALING
myoD
SRA is present in conjugates with the skeletal muscle differentiation factor myoD and p68/p7233. Validating the role of SRA in these complexes, siRNA mediated reduction of SRA results in reduced myoblast differentiation and muscle specific gene expression33. Similarly, loss of p68 expression inhibits both adipocyte differentiation and lineage specific gene expression34 along with myoblast differentiation33. This is supported by chromatin immunoprecipitation (ChIP) assays where decreased p68/72 results in reduced TATA and pol II recruitment to the myosin heavy chain IIb promoter. Notably, not all genes activated by MyoD are modulated by p68/72 or SRA33. p68 is recruited to a subset of regulatory regions directing muscle gene expression in differentiated skeletal muscle cells but is absent from the same sites in myoblasts35.
p53
The activities of the p53 tumor suppressor protein have also been linked with SRA and its binding protein p6836. RNAi suppression of p68 inhibits p53 target gene expression in response to DNA damage, as well as p53-dependent apoptosis36. However, as a RNA helicase deficient mutant p68 was equally able to coactivate p53, formal validation of SRA's role in p53 target gene expression is required. It is intriguing to note, as aforementioned, that higher levels of breast cell proliferation and apoptosis were observed in SRA transgenic mice and that these animals were more resistant than wild type littermates to breast tumor formation when crossed with MMTV-ras tansgenic animals18.
SLIRP—A NOVEL SRA-BINDING PROTEIN COREPRESSOR
To identify novel SRA-binding proteins, and better understand detailed SRA-protein interactions, we screened a human breast cancer library with a SRA probe containing a key stem-loop implicated in SRA activity and identified SRA Stem-Loop Interacting RNA binding Protein [SLIRP12]. The SLIRP gene in humans codes for a 109 aa protein with a putative mitochondrial localization signal at its amino-terminal and with central, highly conserved RRM RNA-binding domain required to repress signaling of a range of NRs. It augments SHARP's corepressor activity and the effects of tamoxifen on estrogen signaling. ChIP studies confirmed SLIRP's presence in the nucleus associated with NR target genes and that its depletion by siRNA impacted on the recruitment of NRs and other coregulators, such as ER and NCoR12.
SLIRP—A PREDOMINANTLY MITOCHONDRIAL COREGULATOR
SLIRP is widely expressed in normal human tissues but elevated in high energy demand organs such as liver, skeletal, and cardiac muscle. In addition, SLIRP is expressed in a range of human cancer cell lines and readily detected in human breast cancer tissue. We were surprised however to observe that SLIRP is predominantly a mitochondrial protein. Although unexpected, the presence of SLIRP in the mitochondria adds to the mounting evidence describing NRs such as ER37, GR38, and TR39 in this organelle. For the GR, mitochondrial translocation correlates with susceptibility to glucocorticoid-induced apoptosis38. In addition, mitochondrial DNA contains putative hormone-response elements that can be bound by NRs in DNA gel-shift studies39. The presence of NRs in both the nucleus and mitochondria raises the possibility that they could coordinately regulate gene expression in both locations directing key processes in metabolism and cell growth.
To our knowledge, SLIRP is the only NR coregulator that influences transcription in the nucleus but resides predominantly in the mitochondrion. Interestingly, although not located in the mitochondrion, there are parallels between SLIRP and another NR coregulator, RIP140. The latter is also preferentially expressed in tissues similar to SLIRP where it can suppress oxidative metabolism and mitochondrial biogenesis40. Furthermore, RIP140 is a powerful negative regulator of insulin-responsive hexose uptake and oxidative metabolism in mouse adipocytes and skeletal muscle41. Depletion of RIP140 from these cells upregulates clusters of genes involved in glucose uptake, glycolysis, the TCA cycle, fatty acid oxidation, mitochondrial biogenesis and oxidative phosphorylation, together with increased mitochondrial oxygen consumption. As predicted, RIP140 null mice are resistant to diet-induced obesity. Given that SLIRP corepresses PPARδ12, is predominantly mitochondrial, and is expressed in high energy demand tissues, it will be of great interest to determine what genes are regulated by SLIRP and whether it plays a role in regulating body metabolism and mitochondrial biogenesis.
SUMMARY
The discovery of the RNA coactivator SRA represented a paradigm shift in our understanding of coregulator regulation of NR signaling. The subsequent isolation of SRA-binding proteins such as SLIRP, has opened up new avenues of investigation that will lead to a better understanding of the mechanism of SRA action and may yield novel targets for the treatment of diseases ranging from breast cancer to diabetes. In addition, the discovery of NRs and coregulators in mitochondria not only provokes challenging questions regarding the role of these molecules in controlling energy homeostasis and metabolism but how they may coordinate signaling pathways between the nucleus and other organelles.
Acknowledgements
The authors are grateful to Keith Giles and Andrew Barker for their critical reading of this manuscript and to the NHMRC, National Breast Cancer Foundation, Cancer Council of WA and Royal Perth Hospital Medical Research Foundation.
