| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Human Genetics (M.S.), Mount Sinai School of Medicine, New York, New York 10029; and National Institutes of Health (NIH) (L.N., H.L., L.J., D.F.), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), Clinical Endocrinology Branch, Bethesda, Maryland 20892-1772
Address all correspondence and requests for reprints to: Douglas Forrest, National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases (NIH/NIDDK), Clinical Endocrinology Branch, 10 Center Drive, Bethesda, Maryland 20892-1772. E-mail: forrestd{at}niddk.nih.gov
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
To understand the development of the color visual system it is necessary to identify the transcription factors that regulate the onset of S or M opsin expression. Cones are among the first cell types to be generated from the retinal progenitor cells (5). In comparison, rod photoreceptors that mediate vision in dim light are generated later. After their generation in utero, immature cones enter a poorly defined phase of development before they express opsins in temporally and spatially distinct patterns (6). S opsin appears in the neonate whereas M opsin begins to appear a week later in the mouse (7, 8). We previously found that thyroid hormone receptor ß2 (TRß2), a nuclear receptor, is essential for M opsin expression. In mice lacking TRß2, all cones instead express S opsin (9, 10, 11). Thus, TRß2 prompts a subpopulation of cones to acquire an M opsin identity. In TRß2-deficient mice, S opsin is prematurely expressed, suggesting that TRß2 also transiently suppresses S opsin. Other factors with a role in opsin regulation include CRX, a homeodomain factor that acts widely in cones and rods. Human CRX mutations are associated with cone-rod dystrophy and photoreceptor degeneration (12, 13, 14), and Crx–/– mice have reduced expression of several rod and cone genes including S and M opsins (15). Other factors that influence cone gene expression, but in an unusual manner because this occurs in the rod pathway, are Nrl (neural retina basic-leucine zipper factor) (16), and Nr2e3, an orphan nuclear receptor (17). Nrl and Nr2e3 promote rod differentiation and limit the rod precursors from expressing S opsin and other cone genes (18, 19, 20, 21).
Analysis of human (22, 23) and mouse (8) Opn1sw transgenes has shown that as little as 0.5 kb of conserved upstream sequence directs expression in cones in vivo but the factors that act on these sequences are unknown. In the course of investigating Opn1sw regulation, we have uncovered a role for RORß (retinoid-related orphan receptor ß, encoded by Rorb/Nr1f2, or RORB in humans) (24, 25, 26). RORß lacks a known physiological ligand and transactivates as a monomer through a particular type of response element (27, 28). RORß is expressed in the brain, pineal gland, and retina (29), and its deletion in mice causes loss of electroretinogram responses and retinal degeneration in adults (30). However, the specific functions and target genes of RORß are unknown. We report that RORß activates the Opn1sw promoter. Further studies using Opn1sw reporter transgenes and RORß-deficient mice indicate the physiological importance of RORß as an inducer of Opn1sw during cone development.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). Unlike typical thyroid hormone receptor response elements consisting of dimeric repeats, each element consisted of a single motif flanked on its 5'-side by an AT-rich sequence, characteristic of binding sites for the ROR/Reverb orphan receptor family (25, 31). Both elements (RORE1 and RORE2) were conserved in mammalian species. A putative CRX response element at position –154 (CRXE; consensus CTAATC) was identified previously by visual inspection (12, 14).
|
, B and C). RORß2 and CRX together stimulated a 38-fold synergistic induction. Increasing amounts of RORß2 gave a dose-dependent increase in the response. Thus, RORß2 alone modestly induced 562 S opsin-luc but cooperated with CRX to amplify the level of induction.
Rorb encodes RORß2 and RORß1 N-terminal variants (see Fig. 5). Both variants activated 562 S opsin-luc, although RORß1 gave a weaker induction than did RORß2 in 293T cells (Fig. 1C). This is consistent with the described inefficiency of RORß1 in nonneuronal cells (28, 32) and suggests that its short N terminus [nine amino acids preceding the first Cys of the DNA-binding domain (DBD)] lacks functions that are present in the 20-amino acid N terminus of RORß2. Hence, we focused on RORß2 because it is reported to be retinal specific in rat (24), and we isolated an RORß2 cDNA from mouse eye with an identical N terminus to that in rat.
|
). Deletion to –200, removing both RORE2 and RORE1, abrogated the response to RORß2 and the synergy with CRX, while permitting a reduced (4-fold) response to CRX. Deletion to –113, removing the CRXE, abrogated all response to CRX.
|
). Together with the deletion analysis, this indicates that both sites contribute to a full response to RORß2. Mutagenesis of the CRXE allowed a normal 2-fold response to RORß2 but reduced the response to CRX alone by 60% (P < 0.01) and gave no significant synergy between RORß2 and CRX (P = 0.12). The residual response to CRX alone indicates that other unknown elements partly compensate for loss of this CRXE. However, this proximal CRXE has a major role in mediating synergy with RORß2.
The function of RORE1 or RORE2 in isolation was shown by ligation of these sites onto 200 S opsin-luc, a minimal reporter that lacked ROREs but carried the CRXE (Fig. 2C). RORE1 or RORE2 conferred little response to RORß2 alone, but each restored a synergistic response to RORß2 and CRX. Together, both elements gave an augmented 3-fold response to RORß2, confirming their function and cooperativity.
RORß Binds RORE1 and RORE2
Using an RORE1 probe in EMSA, nuclear protein extracts from mouse eye at embryonic d 17.5 (E17.5) gave a band that comigrated with that formed by in vitro translated RORß2 (Fig. 3A). The band was abolished by wild-type (wt) competitor but not mutant oligonucleotides. The band was also abolished by an antibody against RORß, indicating that it contained RORß. Similar results were obtained with an RORE2 probe. RORE2 formed a fainter band than RORE1, in accord with a TG being less optimal than an AT dinucleotide preceding the AGGTCA motif on artificial ROREs (24). The lack of detection of other major factors in E17.5 retina that bind these probes suggests the selectivity of RORE1 and RORE2 for RORß.
|
). The shifted band was abolished by competitor probe but was only inefficiently competed by a mutant oligonucleotide, even at a 100-fold molar excess. An antibody against CRX supershifted the band.
Fig. 3, A and B, shows that RORß in eye extracts bound RORE1 or RORE2 as a single band but did not form more retarded complexes that would suggest binding of RORß with other factors when bound to isolated DNA elements in EMSA. To test further for binding between RORß2 and CRX when bound to these probes, increasing amounts of CRX were incubated with RORß2 using RORE1 or RORE2 probes. No additional bands were formed in EMSA (data not shown). However, EMSA does not exclude protein-protein interactions, possibly of a weaker nature or mediated by accessory factors in a host cell. To test for this possibility, a coselection assay was performed on extracts from 293T cells transfected with expression vectors for RORß2 and His-tagged CRX (Fig. 3D). His-CRX and any associated bound proteins were selected from the cell lysates using nickel beads. The selection of CRX was shown by Western blot analysis using antibody against the histidine tag (anti-poly-His). An antibody against RORß ligand-binding domain (LBD) detected RORß2 in these selected lysates, indicating copurification of RORß2 with CRX. As a positive control, Nrl (16), which is known to bind to CRX, was also found to copurify with CRX. In contrast, TRß2 did not copurify with CRX and served as a negative control. The input RORß2, Nrl, and TRß2 proteins were readily detected in transfected 293T cells by Western blot analysis of unselected cell lysates. None of these proteins were detected after nickel-chelate selection in the absence of His-CRX. The results suggest that CRX and RORß2 interact in complexes that can be copurified from host cells under these conditions.
Requirement for DNA Binding and C-Terminal Domains of RORß2
The domains of RORß2 required for induction of Opn1sw were mapped by testing truncated proteins on the 562 S opsin-luc reporter (Fig. 4). The C-terminal activation function 2 (AF2) domain of nuclear receptors promotes interactions with cofactors that mediate transactivation (33). Deletion of AF2 (del:AF2) or a larger C-terminal domain that kept only the N terminus and DBD intact, abrogated the response to RORß2 and the synergy with CRX. A deletion that retained only an intact C terminus (LBD) also abrogated the response. The entire C-terminal region of RORß is thought to be structurally important for formation of a transcriptionally active protein (33). The 
-helix in the AF2 domain adopts a folded conformation with other regions of the LBD that bind co-activators. Although a physiological ligand for RORß is unknown, residues within helices 2–12 form a ligand-binding pocket for fortuitous ligands that stabilize the structure. Residues in helices 3–7 are thought to be in close contact with such ligands. As expected, an RORß2 protein with an internal deletion of helices 3–8 (del:H3–8) was inactive in transactivation either alone or in synergy with CRX. Thus, induction of Opn1sw and synergy with CRX by RORß2 require the DBD and an intact C-terminus including the ligand-binding pocket and coactivator-binding domains.
|
). Using nuclear extracts from transfected 293T cells or in vitro translated proteins, RORß2 and the DBD and del:AF2 truncations bound to RORE1 probe, with the DBD giving a lesser shift, as expected for this smaller protein. The presence of RORß2 in the complexes was confirmed using an antibody against the RORß2 N terminus that diminished or supershifted the bands. The LBD truncation failed to bind DNA, and its expression in transfected cells was shown by Western blot analysis (Fig. 4D).
RORß and CRX Expression in Retina
Expression of RORß and CRX overlaps in the photoreceptor layer of mouse eye during the developmental period when S opsin is first induced (Fig. 5A). In situ hybridization detected RORß mRNA widely in the E17.5 retina in the outer neuroblastic layer (ONBL), which will form the photoreceptor layer, the inner neuroblastic layer (INBL), which will form the inner nuclear layer containing interneurons, and in the formative ganglion cell layer (GCL), as in rat eye (34). CRX mRNA was detected in a narrower zone in the peripheral ONL, as described previously (12, 14). At postnatal d 12 (P12) and older ages, RORß signal was weaker but still extended across all retinal layers. At P12 or adult ages, CRX was expressed over the entire photoreceptor layer (data not shown), as described previously (12, 14).
Northern blot analysis confirmed RORß mRNA expression in the immature eye and the reduction of levels in adults (Fig. 5B). A common probe for both RORß1 and RORß2 detected several mRNA sizes at E17.5. In the adult, the mRNA levels declined by 70%, suggesting that less RORß is sufficient for maintenance functions in the mature retina. An RORß2-specific probe showed that RORß2 represented a substantial proportion of the RORß mRNA present. Analysis of Rorb genomic sequences located the RORß2-specific exon 13 kb upstream of the common exon encoding the DBD of RORß, whereas the RORß1-specific exon was a further 109 kb upstream (Fig. 5C).
ROREs Are Required for S Opsin Transgene Expression in Vivo
The function of RORE1 and RORE2 in vivo was tested in transgenic mice. Expression of a wt 562-bp fragment of mouse Opn1sw fused to lacZ (562 S opsin-lacZ) was compared with that of a transgene carrying mutations in both RORE1 and RORE2 (RE1,2-mut-lacZ)(Fig. 6). These same mutations abrogated the response to RORß2, but not CRX, in luciferase assays in 293T cells (see Fig. 2A).
|
). This wt transgene therefore provided a physiologically relevant reference. The transgene was expressed in a gradient with strongest expression in the inferior retina and weaker expression in diminishing cell numbers in the superior retina (Fig. 6B), as is found for endogenous S opsin in rodents (8). The RE1,2-mut-lacZ transgene gave no detectable expression in the retina in 26 founders or stable lines of mice. In contrast, the wt transgene was expressed in nine of 21 founders or stable lines, with expression levels varying between high (one line), medium (five lines), and weak (three lines). No correlation was observed between expression levels and transgene copy number. The possibility of a rearrangement of the integrated transgene that might prevent its expression was ruled out by sequencing the RE1,2-mut-lacZ fragment after reisolation by PCR from genomic DNA of 15 founder mice. The functional potential of RE1,2-mut-lacZ was also confirmed by cotransfection of the construct into 293T cells with CRX to induce expression. Both wt and mutant constructs gave ß-galactosidase-positive cells.
Cone Defects in Rorb–/– Mice
The role for RORß in Opn1sw induction was tested further in Rorb–/– mice. These mice were originally reported to have gross retinal degeneration as adults, but no histological or other data were given on specific cell types or on the course of development before degeneration (30). We observed the formation of a photoreceptor layer during the first 2–3 wk of postnatal development that allowed a study of the initial period of induction of Opn1sw (see Fig. 8). In wt mice at P6, as expected, S opsin was detected by immunostaining in the immature cone cell bodies and pedicles (Fig. 7A). By P13, S opsin was concentrated in the outer segments of the maturing cones. In contrast, S opsin was not detected in cones in Rorb–/– mice at P6 or in outer segments at P13. At P23, weak signals were detected in scattered cells in the ONL and INL, which suggests that, at later stages, low levels of S opsin may be present in aberrant cells (data not shown). Also, signals were not detected using peanut agglutinin, which normally reacts with the inner and outer segments of all cone subtypes.
|
|
). EMSA with retinal extracts from Rorb–/– pups using the Opn1sw RORE1 probe showed a loss of the RORß-specific band detected in wt mice, confirming that RORß normally binds the site (Fig. 7C). At P28, additional, minor bands were detected in wt mice that were insensitive to abolition by antibody. These minor bands also remained when using extracts from Rorb–/– mice, suggesting that, in older mice, other factors may bind this RORE.
To rule out an absence of cones as the cause of impaired S opsin induction, other markers were studied (Fig. 7D). CRX is normally expressed in both cones and rods from early stages. In Rorb–/– mice, CRX was detected by in situ hybridization at E17.5 and at postnatal stages, although in a somewhat disorganized pattern. This result also suggests that CRX without RORß is unable to induce S opsin appropriately in neonatal cones. TRß2 is a marker of immature cones before S or M opsin is expressed. In Rorb–/– mice, TRß2-positive cells were present at E17.5, although in lower numbers than in wt mice (158 ± 16 vs. 490 ± 38 per midretinal section; n = 3 mice, mean ± SD, P < 0.001; eye diameters were not grossly different). Thus, as indicated by CRX and TRß2 expression, cones form in Rorb–/– mice, although in reduced numbers. M opsin is normally induced later than S opsin, in the second postnatal week, and it was detected by in situ hybridization in both wt and Rorb–/– mice at P14 (Fig. 7E). M opsin showed a normal gradient of distribution with most positive cells in the superior retina and fewer in the inferior, in both wt and Rorb–/– mice. Thus, cones are competent to express M opsin in Rorb–/– mice.
Histological examination yielded several findings concerning photoreceptor differentiation in Rorb–/– mice. First, Rorb–/– mice form an ONL although this was less than two thirds of normal thickness at P16 (Fig. 8, A and B) or P23 (data not shown). This photoreceptor layer endured for several weeks, suggesting that the thinner ONL reflected a shortfall in the generation of photoreceptors rather than a progressive degeneration. In contrast to the relative stability of the ONL, cell loss did occur in the inner part of the inner nuclear layer and ganglion cell layer as early as P6. Also, the outer and inner plexiform layers were severely condensed, suggesting defective formation of synaptic networks. Second, the photoreceptor cell bodies were disorganized in the ONL, and cones and rods were not readily distinguished in Rorb–/– mice. Normally, the abundant rod cell bodies have compact nuclei with dense chromatin, whereas the sparser cone cell bodies (representing only 
3% of all photoreceptors) are located in the outer zone of the ONL and have larger, less dense nuclei (35). Third, cone and rod photoreceptors in Rorb–/– mice lacked outer and inner segments (Fig. 8C). This is consistent with the lack of staining with peanut agglutinin in cones noted in Fig. 7A. These segments are prominent in wt mice, but only very rare, severely stunted structures were found in Rorb–/– mice. Thus, RORß has several functions in photoreceptor differentiation. However, it can be concluded that defective induction of S opsin in cones in Rorb–/– pups is not due to an absence of cones or degeneration. These findings support the proposed role for RORß in the developmental induction of S opsin in cones.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The results suggest that the developmental induction of both S and M opsins is controlled by nuclear receptors: S opsin by RORß and, as shown previously, M opsin by TRß2 (10). In contrast to the multiple roles of RORß in photoreceptors, TRß2 acts primarily as a switch to an M opsin identity. In TRß2-deficient mice the retina appears normal, and, although the cones lack M opsin, they express S opsin and retain responses to short wave light. In contrast to the selective roles of TRß2 and RORß on M and S opsin induction, respectively, CRX is probably a cofacilitator of expression of both cone opsins, in accord with CRX acting as a general activator in the terminal maturation of all photoreceptors (15). The decline in RORß expression levels at older postnatal ages (Fig. 5) implies that RORß has particularly important functions at earlier stages, including the determination of opsin identity in immature cones.
The timing of onset of S opsin expression in neonatal cones may be explained by the orchestrated action of these same factors. Although RORß and CRX are expressed in embryonic retina, S opsin is transiently suppressed until birth by TRß2 (10). The declining TRß2 levels after birth may allow CRX and RORß to induce S opsin in the postnatal period. It is unclear whether TRß2 acts by direct or indirect means because it does not repress an Opn1sw reporter in 293T cells (Srinivas, M., and D. Forrest, unpublished observation). An interplay between RORß and TRß2 presumably produces the opposing distribution gradients of S and M opsins across the superior-inferior axis of the mouse retina. The mechanism is unknown, and it remains an intriguing question how the balance of M or S opsin expression is set in a given cone. Other factors may be invoked because both RORß and TRß2 are expressed over the entirety of this axis. One candidate is RXR
, which is necessary for suppressing S opsin in the superior retina (36). The pattern of distribution of S and M opsins over the retina varies between mammalian species, but the same temporal sequence of induction of S before M opsin occurs in both human (6) and rodent (7, 8) retina. Cooperation between RORß and CRX may therefore provide a common stimulus for S opsin induction in mammals.
Other functions of RORß in photoreceptors, although not the main subject of this study, are evident from the phenotype of Rorb–/– mice. A major role, which parallels that of CRX (15), is in directing formation of the outer segments, the light-sensing structures of both cones and rods that contain opsins and rhodopsin, respectively. The formation of these structures may involve genes encoding structural or biosynthetic molecules that are common targets for RORß and CRX (37). At earlier stages, the commitment of progenitor cells to adopt a photoreceptor fate is directed by factors such as Otx2 (38). The expression of early cone markers in Rorb–/– mice show that RORß is not needed for the first steps in cone differentiation. However, the reduced expression of these markers (TRß2 and CRX) and the thinner photoreceptor layer in Rorb–/– mice suggest that RORß has a degree of influence over entry to the photoreceptor pathway or, subsequently, over survival. The forced expression of RORß can increase the number of cell clones in explants of embryonic rat retina, suggesting that it may have a role in early proliferating progenitors (34).
In rods, Nrl and Nr2e3 promote rod differentiation and suppress expression of cone genes, including S opsin, thus preventing rods from reverting to a cone-like phenotype (18, 19, 20, 21). It is currently unclear how Nrl and Nr2e3 suppress S opsin in rods and whether this is direct or indirect. No specific binding sites for Nrl or Nr2e3 have been mapped on the Opn1sw gene. RORß is widely expressed in the ONL in both cones and rods, and it is an interesting question whether RORß may interact with these factors in rod differentiation. In summary, the evidence indicates several roles for RORß in photoreceptors: 1) a specific role in S opsin induction during the period when cones acquire their M or S opsin identity, which is ultimately necessary for color perception; 2) a broader role in the late morphological maturation of all photoreceptors in the formation of outer segments in both rods and cones; and also, potentially 3) a specific role in rod differentiation, a possibility that is being investigated.
RORß and Activation of Opn1sw
To our knowledge, Opn1sw is the first natural target gene identified for RORß, and it is interesting to note that the presence of two ROREs in the Opn1sw promoter agrees with data on artificial reporters showing that RORß is inefficient on a single RORE (24, 27, 28). However, the modest response to RORß alone suggests that RORß activity is, by necessity, enhanced by cooperation with other factors, in this case by CRX. Such cooperation may provide cell specificity and amplify the activity of orphan receptors that lack a ligand. Candidate binding sites for homeodomain factors are relatively common in the genome, and cooperation with other factors can enhance their specificity (39, 40). In photoreceptors, therefore, we propose that, when bound to DNA, CRX confers strong activation whereas its cooperation with RORß specifies a subset of photoreceptor target genes. A specifying role is in accord with a less frequent occurrence of the 11-bp RORE consensus than the 6-bp CRXE. The low level of S opsin mRNA remaining in Crx–/– mice (15) may reflect residual induction by RORß alone. Such cooperative mechanisms are likely to be important in conferring target gene specificity on the actions of CRX, given its broad role in the terminal differentiation of all photoreceptor types (15, 41, 42).
One simple explanation for the synergy is that upon binding to Opn1sw, the combined activation domains of RORß and CRX are more potent than the sum of their parts. On Opn1sw, the synergy may be enhanced by the dual RORß-binding sites, allowing a contribution by more than one AF2 activation domain and possibly by other CRX-binding sites in addition to the proximal CRXE. The copurification of CRX and RORß2 from transfected host cells suggests that protein complexes containing RORß and CRX form and may aid the synergy, perhaps by stabilizing the protein complexes after DNA binding. Further study is needed to determine whether direct contacts occur between RORß2 and CRX or whether accessory factors are involved. Other photoreceptor genes may be regulated through combinations of distinct transcription factor binding sites. In Drosophila, several elements, including sites for the CRX ortholog Otd, regulate opsin genes (43). Also distinct elements are thought to control the UV cone opsin gene in zebra fish (44) and mouse cone arrestin analyzed in Xenopus (45).
The cooperation between RORß and CRX in opsin regulation broadens the range of known interactions between orphan receptors and homeodomain factors. In Drosophila development, binding of the Ftzf1 orphan receptor to weaker affinity DNA sites is facilitated by interaction with the Ftz homeodomain factor (39, 40). Another example, in liver metabolism, concerns the Prox1 homeodomain factor, which inhibits the LRH1 orphan receptor from binding DNA (46). Orphan receptors such as RORß display constitutive activity in transfection assays in the absence of the specific ligands that are typically required by nuclear receptors. Some orphan receptors such as RORß may be stabilized by more commonly available hydrophobic molecules of lower specificity that fit into a large ligand-binding pocket in vitro (33). However, our results suggest that, regardless of any putative ligand-like molecule in retina, RORß function is strongly enhanced by cooperation with CRX in cone differentiation.
Given the conservation of other regulators of the color visual system in mammals, RORß may play a wider role in opsin regulation in species other than the mouse. Mutations in human RORB, located on chromosome 9q22, have not been reported to date. However, mutations in human CRX are associated with photoreceptor disease (12, 13, 14). Also, mutations in THRB (encoding TRß2), cause resistance to thyroid hormone, which in a rare recessive form may be associated with color blindness (47). Orthologs of ROR (DHR3) and CRX (Otd) (43) also exist in insects, suggesting that these genes may cooperate in eye development in many species.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Plasmids
RORß1 (1.35 kb) and RORß2 (1.4 kb) cDNAs were generated from adult mouse eye cDNA by PCR using primers based on rat cDNA and mouse genomic sequences and were cloned into pSG5 (GenBank accession nos. DQ779923, DQ779924). RORß2 del:AF2, del:H3–8, DBD, and LBD truncations were made by PCR. Bovine His-tag-CRX cDNA in pCDNA3.1 was a gift of Dr. S-M. Chen. A pcDNA3.1-CRX construct was also made without the His-tag. S opsin-luc reporters were made by PCR from Opn1sw genomic DNA, cloned into HindIII and NcoI sites of pGL3 basic vector (Promega Corp., Madison, WI) with the exon 1 ATG codon fused into a luciferase cassette. The site-directed mutagenesis system (Promega) was used to change all 11 bases of RORE1 (RE1-mut) and RORE2 (RE2-mut)(see Fig. 6) or all six bases of CRXE (CRXE-mut, CTAATC to TCGGCT) in 562 S opsin-luc. RORE1 and RORE2 double-stranded oligonucleotides flanked by SacI and HindIII sites were ligated at the 5'-end of 200 S opsin-luc.
Transfection Assays
293T cells were grown in DMEM and 10% fetal bovine serum. Lipofectamine (Invitrogen, San Diego, CA) was used to transfect 1 x 106 cells in 12-well dishes with 250 ng of reporter, 100 ng of pSG5-RORß2, pSG5-RORß1, and/or pCDNA3.1-CRX and 25 ng of Renilla luciferase vector (Promega) as internal control. Bovine His6-CRX or nontagged CRX cDNAs gave similar results; data shown were derived using the former plasmid. Total DNA concentration was maintained at 1 µg using pSG5 empty vector. Experiments were performed in triplicate, and each was repeated three to five times. Cells were harvested after 48 h (unless otherwise stated in figure legends) in Passive Lysis buffer (Promega); 20 µl of lysate was assayed using a Dynex or a Centro LB 960 microplate (Berthold Technologies, Bad Wildbad, Germany) luminometer. Firefly luciferase activity was normalized to Renilla light units. Data were analyzed by the Student’s t test. Similar results were obtained in 293T or NIH3T3 cells.
Proteins, EMSA, and Copurification
Reticulocyte lysate-translated proteins were made using the TNT system (Promega), and nuclear proteins were prepared from eye tissue as described elsewhere (48). EMSA was performed as described using 2 µl of protein (49) and analysis in polyacrylamide gels in 0.5 x Tris-borate, EDTA buffer. Antibody (1 or 2 µl) was added after addition of probe. For competition, cold oligonucleotides were added at 0.5-, 1-, 10-, and 100-fold molar excess for CRXE and 1-, 10-, and 100-fold excess for ROREs, after polydeoxyinosinic deoxycytidylic acid was added, and then incubated for 10 min. For RORß2 and CRX mixing experiments, low stringency EMSA was performed in 0.25x Tris-borate EDTA to enhance detection of any weaker bands. Translation of full-length and truncated RORß2 proteins was monitored by [35S]Met-labeling.
Copurification
Combinations of expression plasmids (pcDNA3-His-tag-CRX, pSG5-TRß2, pSG5-RORß2, and pcDNA3-Nrl) were cotransfected into 293T cells in 10-cm dishes using Lipofectamine 2000 (Invitrogen). Cells were harvested after 36 h and washed once in PBS. Whole-cell lysis was performed using three freeze-thaw cycles in 180 mM KCl buffer (20 mM HEPES, pH 7.9; 180 mM KCl; 0.5 mM EGTA; 0.5 mM EDTA; 0.5% Nonidet P-40, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonylfluoride and protease inhibitor). The lysis mix was centrifuged and the supernatant was taken for analysis. To select for His-CRX and associated proteins, 20 µl nickel magnetic beads (Sigma Chemical Co., St. Louis, MO) were added and incubated at 4 C for 4 h with mixing by slow rotation. Protein was eluted by boiling in 1x SDS-PAGE-loading buffer after three washes with the 180 mM KCl buffer. The eluted samples were analyzed by 4–12% NuPAGE gel electrophoresis (Invitrogen), followed by Western blot analysis. Blots were probed in binding buffer (5% dry milk powder in 0.2% Tween 20 in 1x PBS) with the following primary antibodies, all at 1:2000 dilution: mouse anti-polyHis (Sigma); rabbit anti-RORß LBD, anti-TRß2 N terminus and anti-Nrl (see below). Secondary antibodies (antirabbit or antimouse) were used at 1:4000 dilution, with detection using enhanced chemiluminescence reagents from Amersham Biosciences (Arlington Heights, IL).
RNA Analysis
Total RNA (15 µg) or poly(A)-selected RNA (2 µg) was loaded per Northern blot lane. Hybridization was in 50% formamide buffer at 42 C with 45-mer antisense oligonucleotides common for RORß1/ß2, or specific for RORß2 end labeled with [-32P]ATP and polynucleotide kinase. Opsin and glyceraldehyde-3-phosphate dehydrogenase (G3PDH) probes were labeled by random priming (10). Signals were quantified by phosphor imaging.
In Situ Hybridization
Tissue fixation in 4% paraformaldehyde (PFA) and hybridization analysis of 10 µm cryosections using digoxigenin-labeled riboprobes at 65 C and Digoxigenin Nucleic Acid Detection Kit (Roche, Indianapolis, IN) followed standard protocols (50). RORß antisense and sense probes were made from a 648-bp cDNA including the 45-bp RORß2 N terminus and 603 bp of common DBD in pGEM3Zf(–), linearized with EcoRI and SmaI, respectively. A mouse CRX 898-bp cDNA (codons 2–299) was cloned into pGEM-TEZ, and then linearized with AvrII to make an antisense riboprobe. An M opsin 673-base probe was made from a mouse M opsin cDNA in pGEM linearized with NcoI (10).
Transgenes and Mouse Strains
The 562 S opsin-lacZ transgene was released by PstI and KpnI digestion of a 6.4-kb transgene of mouse Opn1sw fused at the exon 1 ATG to lacZ (8). RE1,2-mut-lacZ transgene was identical but carried site-directed mutations in RORE1 and RORE2. Transgenic mice were made on a B6D2/F1J background at the Mount Sinai transgenic facility. Founders were identified using Southern blots with a lacZ probe and were studied as transients or were crossed with B6D2/F1J mice. Genotyping employed a three-primer PCR giving a 450-bp transgene band and a 400-bp control band from endogenous Opn1sw. Rorb+/– mice were first created on a 129/Ola x C57BL/6 background (30) and then backcrossed for six generations onto a C3H/HeN strain, as kindly provided to us by Dr. M. Dubocovich. Rorb+/– mice were crossed and wt, +/–, and –/– littermates were studied. Genotyping employed a three-primer PCR giving a wt band of 331 bp and a mutant band of 610 bp. C3H strains have retinal degeneration; therefore, Rorb–/– mice were not studied beyond 3–4 wk of age. We also observed a sporadic hyperplasia in the retina in some mice in this colony, regardless of Rorb genotype. Animal experiments followed approved protocols at Mount Sinai School of Medicine and at NIH.
Histology, Antibodies, and Immunostaining
Eyes were fixed in 3% glutaraldehyde/2% PFA for 12 h, and then embedded in methyl methacrylate. We found that typical stains used for plastic sections did not label all retinal layers optimally. Therefore, 3-µm sections were briefly stained for 2 min in hematoxylin and eosin Y (diluted 1:4 in water). For electron microscopy, samples were postfixed in 1% osmium tetroxide, and then embedded in Spurr’s plastic resin; 600–800Å sections were cut and analyzed on a Zeiss transmission electron micoscope (JFE Enterprises, Brookville, MD). Eyes from at least three –/– and three wt mice per group were studied by histology or by electron microscopy. For immunostaining, eyes were fixed in 2% PFA and cryoprotected in 30% sucrose. ß-Galactosidase was detected on 10-µm cryosections using 5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside (X-gal) substrate for 1–2 d. Antibodies, reagents, and dilutions are as follows: ß-galactosidase, mouse monoclonal (Promega), 1:1000; S opsin, rabbit (JH455, a gift from J. Nathans and T. Reh), 1:1000, or (Chemicon AB5407) 1:3000; TRß2 N terminus, rabbit polyclonal, 1:2000 (Ng, L., and D. Forrest, in preparation); biotinylated peanut agglutinin (Vector Laboratories, Inc., Burlingame, CA), 1:500. Signals were detected using avidin, biotinylated enzyme complex (ABC) or mouse antibody-on-mouse-tissue (M.O.M.) kits (Vector Laboratories). Rabbit antiserum against RORß residues 111–459 was a gift of H. Stunnenberg (27). Anti-CRX was a gift of S.-M. Chen (42). Anti-Nrl was a gift from A. Swaroop (University of Michigan, Ann Arbor, MI). A synthetic 15-amino acid peptide of RORß2 N terminus (MCENQPKTKADGTAQ), conjugated to keyhole limpet hemocyanin (KLH) carrier, was used to raise antiserum in a NZW rabbit (Covance Laboratories, Inc., Madison, WI); titers were tested by ELISA. For Western blots, 5 µg of protein from transfected cells was loaded per lane and analyzed with a 1:1000 dilution of primary antibody, a 1:5000 dilution of horseradish peroxidase-conjugated antirabbit secondary antibody and enhanced chemiluminescence reagents (Amersham) by standard methods (50).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The authors have no conflicts of interest to disclose. M.S., L.N., H.L., L.J., and D.F. have nothing to declare.
First Published Online March 30, 2006
Abbreviations: AF2, activation function 2; CRX, cone-rod homeobox factor; CRXE, CRX response element; DBD, DNA-binding domain; GCL, ganglion cell layer; G3PDH, glyceraldehyde-3-phosphate dehydrogenase; INBL, inner neuroblastic layer; INL, inner nuclear layer; L opsin, long-wave-sensitive opsin; M opsin, medium-wave-senstive opsin; ONBL, outer neuroblastic layer; ONL, outer nuclear layer; PFA, paraformaldehyde; ROR, retinoid-related orphan receptor; RORE, ROR response element; S opsin, short-wave-sensitive opsin; TRß2, thyroid hormone receptor ß2; wt, wild type.
Received for publication December 12, 2005. Accepted for publication March 22, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
and N-terminal variant ß receptors. Development 114:39–47[Abstract](4) and ROR ß(1). Biochim Biophys Acta 1494:236–241[Medline] 1 and the Rev-erb family of orphan nuclear receptors. Mol Endocrinol 8:1253–1261[Abstract] is necessary to establish the S-opsin gradient in cone photoreceptors of the developing mouse retina. Invest Ophthalmol Vis Sci 46:2897–2904-hydroxylase gene. Mol Endocrinol 18:2424–2439NURSA Molecule Pages Link:
This article has been cited by other articles:
 |
|
 |
|
  H. Liu, P. Etter, S. Hayes, I. Jones, B. Nelson, B. Hartman, D. Forrest, and T. A. Reh NeuroD1 Regulates Expression of Thyroid Hormone Receptor 2 and Cone Opsins in the Developing Mouse Retina J. Neurosci., January 16, 2008; 28(3): 749 - 756. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Jomary and S. E. Jones Induction of Functional Photoreceptor Phenotype by Exogenous Crx Expression in Mouse Retinal Stem Cells Invest. Ophthalmol. Vis. Sci., January 1, 2008; 49(1): 429 - 437. |
, P = 0.025). C, EMSA showing expression and DNA binding ability of truncated proteins on RORE1 using protein from reticulocyte lysates or mock programmed lysates; results were similar using extracts from transfected 293T cells (data not shown). EMSA was performed ± antibody (Ab) against RORß2 N terminus. Arrowheads indicate complexes for RORß2 and del:AF2 (upper) and DBD (lower) proteins; s, supershifts. D, Western blot detection of LBD (38 kDa) and del:H3–8 (40 kDa) in lysates from transfected 293T cells using antibody against RORß LBD; mock, transfection control; RORß2 (52 kDa), full-length control.
