cone opsin gene pairs: Implications for photoreceptor subtype specification

Author's Accepted Manuscript

Transcriptional co-regulation of evolutionarily conserved microRNA/cone opsin gene pairs: Implications for photoreceptor subtype specification Yutaka Daido, Sakurako Hamanishi, Takehiro G. Kusakabe

www.elsevier.com/locate/developmentalbiology

PII: DOI: Reference:

S0012-1606(14)00239-5 http://dx.doi.org/10.1016/j.ydbio.2014.04.021 YDBIO6423

To appear in:

Developmental Biology

Received date: 6 November 2013 Revised date: 16 April 2014 Accepted date: 25 April 2014 Cite this article as: Yutaka Daido, Sakurako Hamanishi, Takehiro G. Kusakabe, Transcriptional co-regulation of evolutionarily conserved microRNA/cone opsin gene pairs: Implications for photoreceptor subtype specification, Developmental Biology, http://dx.doi.org/10.1016/j.ydbio.2014.04.021 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Transcriptional co-regulation of evolutionarily conserved microRNA/cone opsin gene pairs: implications for photoreceptor subtype specification Yutaka Daido a, Sakurako Hamanishi b, Takehiro G. Kusakabe a,b,* a

Institute for Integrative Neurobiology, Graduate School of Natural Science, Konan University, 8-9-1 Okamoto, Higashinada-ku, Kobe 658-8501, Japan

b

Department of Biology, Faculty of Science and Engineering, Konan University, 8-9-1 Okamoto, Higashinada-ku, Kobe 658-8501, Japan

*Corresponding author: Takehiro G. Kusakabe Department of Biology, Faculty of Science and Engineering, Konan University, 8-9-1 Okamoto, Higashinada-ku, Kobe 658-8501, Japan Phone: +81-78-435-2511 Fax: +81-78-435-2539 e-mail: [email protected]

1

ABSTRACT The vertebrate retina contains two types of photoreceptor cells, rods and cones, which use distinct types of opsins and phototransduction proteins. Cones can be further divided into several subtypes with differing wavelength sensitivity and morphology. Although photoreceptor development has been extensively studied in a variety of vertebrate species, the mechanism by which photoreceptor subtypes are established is still largely unknown. Here we report two microRNAs (miRNAs), miR-726 and miR-729, which are potentially involved in photoreceptor subtype specification. In the medaka Oryzias latipes, the genes encoding miR-726 and miR-729 are located upstream of the red-sensitive opsin gene LWS-A and the UV-sensitive opsin gene SWS1, respectively, and are transcribed in the opposite direction from the respective opsin genes. The miR-726/LWS pair is conserved between teleosts and tetrapods, and the miR-729/SWS1 pair is conserved among teleosts. In situ hybridization analyses and fluorescence reporter assays suggest that these miRNAs are co-expressed with the respective opsins in specific cone subtypes. Potential targets of miR-726 and miR-729 predicted in silico include several transcription factors that regulate photoreceptor development. Functional analyses of cis-regulatory sequences in vivo suggest that transcription of the paired microRNA and opsin genes is co-regulated by common cis-regulatory modules. We propose an evolutionarily conserved mechanism that controls photoreceptor subtype identity through coupling between transcriptional and post-transcriptional regulations.

Keywords: Cone photoreceptor cells; Opsin; Medaka; MicroRNA; Bidirectional gene pair

2

Introduction In the vertebrate retina, there are two major types of photoreceptor cells, rods and cones, which use distinct types of opsins (Nathans, 1987; Palczewski, 2012; Kefalov, 2012). Cones can be further divided into different subtypes based on their wavelength sensitivity and morphology. The difference in wavelength sensitivity is primarily attributable to the opsin subtypes expressed (Bowmaker, 2008; Allison et al., 2010). Transcriptional regulation through specific interactions between transcription factors and cis-regulatory DNAs is a central mechanism to control photoreceptor development (Swaroop et al., 2010; Forrest and Swaroop, 2012). Otx-class homeodomain transcription factors Otx2 and Crx are expressed in photoreceptor progenitor cells, and their expression is required for both rod and cone differentiation (Chen et al., 1997; Freund et al., 1997; Furukawa et al., 1997; Nishida et al., 2003). The Maf-family transcription factor Nrl and the orphan nuclear receptor Nr2e3 are required for rod specification and differentiation (Mears et al., 2001; Cheng et al., 2004; Chen et al., 2005; Akimoto et al., 2006; McIlvain et al., 2007; Oh et al., 2007). Development of cone subtypes is also known to be regulated by specific transcription factors: the thyroid hormone receptor TR2 and other nuclear receptor family factors, including RXR, ROR and ROR, regulate cone identity in mammals (Ng et al., 2001; Roberts et al., 2005; Roberts et al., 2006; Srinivas et al., 2006; Applebury et al., 2007; Lu et al., 2009; Fujieda et al., 2009; Onishi et al., 2010; Glaschke et al., 2011), and the T-box transcription factor Tbx2b has been suggested topromote the ultraviolet-sensitive cone fate by repressing the rod fate in zebrafish (Alvarez-Delfin et al., 2009). Recent studies suggest that the sine oculis homologue Six7 also play an important role in the regulation of photoreceptor subtype 3

development in zebrafish (Alvarez-Delfin, 2011; Saade et al., 2013). Compared to the mechanisms underlying rod-cone differences, however, the regulatory machinery that controls cone subtype identity is less understood, particularly in non-mammalian vertebrates, which have a greater diversity of cone photoreceptors than mammals. Emerging evidence suggests that microRNAs (miRNAs) post-transcriptionally control the development and function of cells in various tissues, including the retina (Ambros, 2004; Bartel, 2004; Bartel, 2009; Sundermeier and Palczewski, 2012). The importance of miRNA for development and function of the retina has been suggested by studies employing the conditional inactivation of dicer, an RNase III endonuclease required for maturation of miRNAs (Damiani et al., 2008; Georgi and Reh, 2010; Pinter and Hindges, 2010; Davis et al., 2011). In the mammalian retina, the miRNA cluster miR-183/96/182 is highly expressed in photoreceptor cells and is required for the normal differentiation and function of photoreceptor cells (Xu et al., 2007; Zhu et al., 2011; Lumayag et al., 2013). To date, however, it is not known whether miRNAs are also involved in photoreceptor subtype regulation. Here we report two miRNAs, miR-726 and miR-729, encoded near cone opsin gene loci, which are potentially involved in photoreceptor subtype specification. The genes encoding miR-726 and miR-729 are located upstream of the red-sensitive opsin gene LWS-A and the ultraviolet-sensitive opsin gene SWS1, respectively, in the genomes of various teleosts. The miR-726/LWS pair is also conserved in some tetrapods. Based on the expression patterns and cis-regulatory functions of these microRNA-cone opsin pairs, we propose a novel mechanism that controls photoreceptor subtype identity through coupling between transcriptional and 4

post-transcriptional regulations.

Materials and methods Animals and embryos Mature adults of the orange-red variety of medaka Oryzias latipes were kept in indoor tanks under artificial reproductive conditions (10 hr dark, 14 hr light; 28°C) and fed on Artemia nauplii (Miyako Kagaku Co., Ltd., Tokyo, Japan) and Otohime B2 artificial feed (Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan). Eggs and embryos were obtained and cultured as described previously (Kusakabe et al., 2012).

Comparative genomics and bioinformatics The genomic sequences of Oryzias latipes, Gasterosteus aculeatus, Takifugu rubripes, Tetraodon nigroviridis, Danio rerio, Xenopus tropicalis, and Anolis carolinensis were obtained from Ensembl (http://www.ensembl.org/index.html). VISTA tools (Frazer et al. 2004; http://genome.lbl.gov/vista/index.shtml) were used for sequence alignment and comparison with the following parameters: 50 bp windows and 50% identity threshold. The locations of the compared regions were: Chr. 5: 27,010,969-27,017,652 (O. latipes miR-726/LWS); Chr. 11: 26,409,202-26,414,002 (D. rerio miR-726/LWS); groupXVII: 10,624,091-10,629,945 (G. aculeatus miR-726/LWS); Scaffold_79: 746,356-750,681 (T. rubripes miR-726/LWS); Chr. 11: 10,120,886-10,124,588 (T. nigroviridis miR-726/LWS); Scaffold GL172911.1: 207,521-219,737 (X. tropicalis miR-726/LWS); Chr. 2: 88,641,952-88,669,311 (A. carolinensis miR-726/LWS); scaffold1021: 44,173-47,630 (O. latipes miR-729/SWS1); Chr. 4: 12,641,956-12,649,372 (D. rerio miR-729/SWS1); and 5

Scaffold_90: 456,569-463,114 (G. aculeatus miR-729/SWS1). Putative transcription factor binding sites in the conserved noncoding regions were predicted by scanning the genomic sequences with matrix models in the JASPER CORE database (http://jaspar.genereg.net/; Mathelier et al., 2014). The relative profile score threshold used for the search was 80%. The genomic sequences containing conserved noncoding regions were aligned using the ClustalW program (Thompson et al., 1994), and the alignments were used for the analysis of evolutionary conservation of the putative transcription factor binding sites. Potential targets of microRNAs in zebrafish were found using TargetScanFish (Release 6.2; http://www.targetscan.org/fish_62/) with annotated zebrafish UTRs (Lewis et al., 2005; Ulitsky et al., 2012).

Preparation of probes for in situ hybridization Total RNA was isolated from the eyes of three adult medaka using ISOGEN (Nippon Gene Co., Ltd., Tokyo, Japan). One μg of total RNA was used as the template to synthesize the first strand cDNA using an oligo(dT) primer according to the manufacturer’s protocol (SuperScript™ III First Strand Synthesis System for RT-PCR; Invitrogen Corp., Carlsbad, CA). Partial cDNA fragments of pri-ola-miR-726, pri-ola-miR-729, LWS-A, SWS1, RH2-A, RH2-B, RH2-C, and rhodopsin were amplified from the first-strand cDNA by PCR using a Taq DNA polymerase (Ex Taq; Takara Bio, Otsu, Japan) with a pair of gene-specific primers, and cloned into the EcoRV site of pBluescriptII SK(+) by the TA cloning method (Marchuk et al., 1991). The primers for O. latipes LWS-A, SWS1, RH2-A, RH2-B, RH2-C, and rhodopsin were designed based on the published nucleotide sequences (Hisatomi et al., 1997; Matsumoto et al., 2006; NCBI accession numbers AB223051, AB223053, AB223054, 6

AB223055, AB223058, and AB001606). The nucleotide sequences of the primers used were: 5’-AGTGGAAGCGATGACCCT-3’ and 5’-CGTGCATTATGATACATTCCCTCT-3’ for LWS-A; 5’-CCGTCCTCTCCTTTGAGAGATAC-3’ and 5’-TCCATCTTCTTTCCGAAAACCATC-3’ for SWS1; 5’-GCTCAACAGTAACTCTGCTTGG-3’ and 5’-CCAAGACCGTCAACAAGGAAAC-3’ for RH2-A; 5’-GGATCCTCCGCTTCAAGAGA-3’ and 5’-GAAGTATGCCGATGGAATTCCTTAG-3’ for RH2-B; 5’-GCATCAGCTGAAGCTCCA-3’ and 5’-GTTCCCACATGCATTTCTGTG-3’ for RH2-C; 5’-TTATGTCCCTATGGTGAACAC-3’ and 5’-AGCAGCCTCCTTGACA-3’ for rhodopsin; 5’-GTACCTTTGCTGTTTTGCTT-3’ and 5’-CTGACTCATGCAAACAAAAATG-3’ for pri-ola-miR-726; and 5’-AGTCTGGGCTGCAGT-3’ and 5’-ATGTAACCCGTTCATGGTT-3’ for pri-ola-miR-729. The sizes of the cDNA fragments of pri-ola-miR-726, pri-ola-miR-729, LWS-A, SWS1, RH2-A, RH2-B, RH2-C, and rhodopsin are 693 bp, 514 bp, 604 bp, 595 bp, 1267 bp, 1274 bp, 1337 bp, and 676 bp, respectively.

The plasmid

clones containing the cDNA fragments were digested with a restriction enzyme, and used as a template to synthesize digoxigenin-labeled RNA probes (pri-ola-miR-726, pri-ola-miR-729, and rhodopsin) or fluorescein-labeled RNA probes (LWS-A, SWS1, RH2-A, RH2-B, RH2-C, and rhodopsin) using a DIG or Fluorescein RNA labeling Mix (Roche Diagnostics, Tokyo, Japan).

Double fluorescence in situ hybridization Dissected adult medaka heads were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) containing 0.1% Tween-20 (PBT) at 4°C for 20 hr (overnight). Fixed tissues were washed three times with PBT, and then 7

dehydrated by ethanol/PBT and ethanol washes (25%, 50%, 75%, and 100%). The specimens were cleared in xylene and embedded in paraffin (Paraffin Wax II60; Sakura Finetek Japan Co., Ltd., Tokyo, Japan) at 65°C. Paraffin blocks containing the specimens were cut at 10-μm thickness with a microtome (RM2125; Leica Biosystems, Nussloch, Germany). The sections were mounted on adhesive glass slides (MAS-coated glass slides; Matsunami Glass Ind., Ltd., Osaka, Japan) with a flattening table (HI1220; Leica Biosystems). After deparaffinization in xylene and rehydration in a descending series of diluted alcohol, the sections were washed with PBS and partially digested with 2 μg/ml of proteinase K in PBS for 5 min at room temperature (RT). After washing with PBS, the sections were post-fixed with 4% paraformaldehyde in PBS for 10 min at RT, then washed with PBS. The sections were acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 8.0) for 20 min and washed again in PBS. The sections were then incubated with hybridization buffer for 1 hr at 63°C. A mixture of digoxigeninand fluorescein-labeled RNA probes at a concentration of 5 μg/ml in hybridization buffer was denatured for 10 min at 68°C, and then added to the sections. The sections were hybridized with for 16 hr at 63°C. The hybridization buffer contained 50% formamide, 5 x SSC (0.15 M NaCl, 0.0015 M Na3citrate, pH 7.0), 50 μg/ml yeast tRNA, 50 μg/ml heparin, and 0.1% Tween-20. After briefly washing with 4 x SSC, the hybridized sections were washed in 50% formamide/2 x SSC for 1 hr at 63°C, in 0.2 x SSC for 2 hr at 63°C, and then in PBS for 5 min at RT. The sections were blocked with 0.5% blocking reagent (Roche Diagnostics) in PBS for 30 min at RT, followed by incubation with a peroxidase-conjugated anti-digoxigenin antibody (1:1000, 11207733910; Roche Diagnostics) in the same blocking solution for 2 hr at RT. After washing with PBS (4 times, 15 min each), the 8

sections were reacted with TSA-Alexa Fluor 488 (T20912; Invitrogen) in Amplification buffer (Invitrogen) containing 0.0015% hydrogen peroxide for 16 hr at 4°C. The peroxidase was inactivated by incubation in 1.5% hydrogen peroxide for 30 min. The fluorescein-labeled probe was detected by using a peroxidase-conjugated anti-fluorescein antibody (1:200, NEF710001EA; Perkin Elmer) for 2 hr at RT and reacted with TSA-Alexa Fluor 594 (T20925; Invitrogen) in Amplification buffer/0.0015% hydrogen peroxide for 30 min (rhodopsin) or 3 hr (the other probes) at RT. Nuclei were stained with 1 μg/ml of DAPI in PBS for 5 min at RT. The sections were mounted with 2% DABCO in 80% glycerol-PBS. Fluorescent images were obtained by using a laser scanning confocal microscope (FV1200; Olympus, Tokyo, Japan).

Fluorescence-reporter constructs The dual reporter plasmid pBluescript-mCherry-EGFP was constructed as follows: an mCherry-coding sequence (provided by Dr. R. Tsien, University of California San Diego, CA) was replaced with EGFP-coding sequence of pBluescript-EGFP (Yoshida et al., 2004), and then the DNA fragment containing EGFP-coding and SV40 poly(A) sequences derived from pEGFP-1 (Clontech, Mountain View, CA) was inserted into the EcoRI/SalI sites in the opposite direction to mCherry. The O. latipes genomic DNA was extracted from one individual of the Hd-rR inbred strain as described previously (Kusakabe and Suzuki, 2000). The intergenic regions between ola-miR-726 and LWS-A (Chr. 5: 27,011,474-27,015,548) and between ola-miR-729 and SWS1 (scaffold1021: 46,113-47,306) were amplified from the genomic DNA by PCR using a thermostable DNA polymerase (PrimeStar HS DNA polymerase, Takara BIO, Japan) and a pair of gene-specific 9

oligonucleotide primers: 5’-GTCAGGATCCTCACTCAGAAGAAGGGGT-3’ and 5’-ACTGGTCGACGTGATCGCACAAGG-3’ for ola-miR-726 and LWS-A; 5’-CGATGGATCCTCTGGTTCTGGTCCTG-3’ and 5’-ACGTGTCGACGTGAAGCTGAGCTCTG-3’ for ola-miR-729 and SWS1 (underlining in the primer sequences indicates restriction enzyme recognition sites). The amplified intergenic region was inserted into the SalI/BamHI sites of pBluescript-mCherry-EGFP. The reporter constructs in which a part of the intergenic region was deleted were made as follows: the whole original plasmid except for the region that we wanted to delete was amplified by PCR using a high-fidelity DNA polymerase (PrimeSTAR HS DNA polymerase; Takara Bio) with a pair of primers (5’-CGTATCTAGATCAAAACTGTGCAAGGAATC-3’ and 5’-GCTATCTAGAGTCAGGTCGGTGTCTTG-3’ for LWS-CNR-A; 5’-CGATTCTAGAACAGGCATGTTTTGG-3’ and 5’-CGTATCTAGATGCTAATGCTACATGTCCTT-3’ for LWS-CNR-B; 5’-CGATTCTAGACGGATTTTTAGCCTGGAG-3’ and 5’-CGTATCTAGAGATTTCCACAGAGACGGG-3’ for LWS-CNR-C; 5’-CGATTCTAGACGGATTTTTAGCCTGGAG-3’ and 5’-CGTATCTAGATGCTAATGCTACATGTCCTT-3’ for LWS-CNR-B+C; 5’-CGATTCTAGACGGATTTTTAGCCTGGAG-3’ and 5’-CGTATCTAGATTCCTTGCACAGTTTTG-3’ for LWS-CNR-A; 5’-GCTATCTAGACAGTTTGGGTTCTGGTCT-3’ and 5’-GCATTCTAGAGCTCTGCTCCCAACAG-3’ for SWS1-CNR-D 5’-GCTATCTAGATTAGACAAAGCCGTAAGCT-3’ and 5’-CGTATCTAGAACCAGGACCTTCTCATAA-3’ for SWS1-CNR-E and 5’-GCTATCTAGATTAGACAAAGCCGTAAGCT-3’ and 5’-GCATTCTAGACCAGAACCCAAACTG-3’ for SWS1-CNR-D), digested with XbaI (recognition sites were underlined in the primer sequences above) and then self-ligated. 10

Microinjection of reporter DNA constructs and immunofluorescent detection of the reporter proteins Circular plasmid DNA constructs were introduced into medaka embryos by microinjection as described previously (Kusakabe et al., 1999). The injected embryos were incubated at 28C. Fluorescence was observed with a fluorescent dissection microscope (M165 FC; Leica Microsystems, Wetzlar, Germany). Embryos were dechorionated as described previously (Kusakabe et al., 2012) and fixed with 4% paraformaldehyde in PBS overnight at 4C. After washing several times with PBS containing 1% Triton X-100 (T-PBS) and once with distilled water, the embryos were permeabilized in cold acetone for 20 min at -20C, and washed again with distilled water and T-PBS. The embryos were then dehydrated in a graded series of methanol (50%, 70%, and 100%) at 4C and stored in methanol overnight at -20C. After rehydration, the embryos were washed with T-PBS containing 1% dimethyl sulfoxide (DMSO) and blocked with 10% goat serum in T-PBS containing 1% DMSO overnight at 4C. The embryos were incubated with rabbit anti-GFP polyclonal (A11122; Invitrogen) and rat anti-RFP monoclonal (5F8; ChromoTek GmbH, Martinsried, Germany) antibodies diluted 1000-fold with the blocking buffer for 9 hr at 4C, washed with T-PBS containing 1% DMSO overnight at 4°C, and then incubated with Alexa Fluor 488-conjugated anti-rabbit IgG and Alexa Fluor 594-conjugated anti-rat IgG goat antibodies (Molecular Probes, Eugene, OR) for 9 hr at 4°C. After being rinsed several times with T-PBS containing 1% DMSO, the embryos were mounted in 80% glycerol and observed under a confocal microscope (FV1000; Olympus).

11

Results MicroRNA-encoding sequences are conserved in the 5’ flanking region of cone opsin genes in teleosts, amphibians, and reptiles During the course of searches for conserved non-coding sequences in the medaka cone opsin loci, two highly conserved sequences were found to encode putative miRNAs. One of these putative miRNAs (Ensembl ID ENSORLG00000021084) is encoded at 4 kb upstream of the red-sensitive opsin gene LWS-A (Matsumoto et al., 2006; Ensembl ID ENSORLG00000014088) on the DNA strand opposite the opsin gene (Fig. 1A). The predicted mature miRNA sequence is identical to that of a zebrafish miRNA, dre-miR-726 (Kloosterman et al., 2006), so that the medaka miRNA is named ola-miR-726 (Fig. 1C). The miR-726-encoding sequence is highly conserved in the upstream region of the LWS-A orthologues not only among various teleosts, but also in amphibians and reptiles (Fig. 1A,B). Identity of this region to miR-726 and its evolutionary conservation were also pointed out by O'Quin et al. (2011). The other putative miRNA (Ensembl ID ENSORLG00000021049) is encoded at 1.2 kb upstream of the ultraviolet-sensitive opsin gene SWS1 (Matsumoto et al., 2006; Ensembl ID ENSORLG00000019293). This miRNA is also encoded on the DNA strand opposite the opsin gene (Fig. 2). The predicted mature sequence is similar to that of the zebrafish miR-729 with the “seed” sequence from position 2 to 8 being identical (Fig. 2C), and this medaka miRNA has been named ola-miR-729 by Li et al. (2010). The miR-729 is conserved in the teleost genomes (Fig. 2A,B).

Co-expression of the miRNA/opsin pairs in a specific subtype of cone photoreceptor cells 12

The spatial expression patterns of primary transcripts for ola-miR-726 and ola-miR-729 (pri-ola-miR-726 and pri-ola-miR-729) were compared with those of opsin genes by double-fluorescence in situ hybridization in the retina of the adult medaka. The hybridization signals for both pri-ola-miR-726 and pri-ola-miR-729 were specifically observed in the outer nuclear layer (ONL), where the nuclei of the photoreceptor cells reside (Fig. 3). Localization of pri-ola-miR-726 was detected in nuclei located immediately to the inner side of LWS-A-positive sites (Fig. 3A), while that of pri-ola-miR-729 was found in nuclei adjacent to SWS1-positive sites (Fig. 3B). In contrast, localization of pri-ola-miR-726 was not associated with SWS1-positive sites (Fig. 3C) or green-sensitive opsin (RH2)-positive sites (Fig. 3E). Similarly, localization of pri-ola-miR-729 was not associated with LWS-A-positive sites (Fig. 3D) or RH2-positive sites (Fig. 3F). The nuclei of rod photoreceptor cells were located in the inner region of ONL and not overlapped with localization of pri-ola-miR-726 or pri-ola-miR-729 (Fig. 3G,H). Hybridization signals for pri-ola-miR-726 and pri-ola-miR-729 were usually observed in different nuclei (Fig. 3I). These results consistently suggest that ola-miR-726 and ola-miR-729 are co-expressed with the red-sensitive opsin LWS-A and UV-sensitive opsin SWS1, respectively.

The intergenic regions between the paired miRNA and opsin genes can drive bidirectional transcription in the same photoreceptor cells Evolutionary conservation of the genomic organization and co-expression of the miRNA/opsin gene pairs suggest that their co-expression is regulated by common cis-regulatory sequences located in the intergenic region. To test this assumption and investigate the molecular mechanisms for co-expression of the 13

paired miRNA/opsin genes, we cloned the intergenic regions between ola-miR-726 and LWS-A and between ola-miR-729 and SWS1, and connected them to a red-fluorescent protein (mCherry)-coding sequence in place of the miRNA gene and a green fluorescent protein (EGFP)-coding sequence in place of the opsin gene (Figs. 4A and 5A). We introduced these dual-reporter constructs into fertilized eggs of medaka by microinjection and examined the reporter expression in the retina of injected embryos. For both constructs, expression of mCherry and EGFP was specifically observed in photoreceptor cells at 8 days post fertilization (dpf). The mCherry signals were frequently overlapped with the GFP signals in embryos injected with either of these constructs (Figs. 4B-G and 5B-G). These results suggest that both ola-miR-726 and ola-miR-729 are expressed in cones but not rods of the retina. The intergenic region is sufficient to drive bidirectional transcription of the paired miRNA and opsin genes in the same photoreceptor cells. These results also provide further evidence for co-expression of the paired miRNA and opsin genes in a specific cone subtype.

Co-expression of the miRNA/opsin pair is regulated by common cis-regulatory modules To further narrow the list of potential cis-regulatory sequences required for the expression of the miRNA and opsin genes, we looked at conserved noncoding sequences as candidates for cis-regulatory modules. In the intergenic region between miR-726 and LWS-A, there are three conserved noncoding regions (CNRs), in addition to the conserved sequences in their basal promoters, and we named them LWS-CNR-A, LWS-CNR-B, and LWS-CNR-C, respectively (Fig. 6A). Similarly, there are two CNRs between miR-729 and SWS1 and we named them SWS1-CNR- 14

and SWS1-CNR-, respectively (Fig. 7A). We made a series of dual-reporter constructs in which a portion of the intergenic region between miR-726 and LWS-A, including CNRs, was deleted from the wild type construct, as shown in Fig. 6B. Expression of the two reporters in retinal photoreceptor cells was observed in embryos developed from eggs injected with these constructs. When LWS-CNR-A was deleted, mCherry expression was completely abolished and EGFP expression was also substantially reduced, whereas the deletion of either LWS-CNR-B or LWS-CNR-C did not substantially change the reporter expression (Fig. 6B and Supplementary Fig. 1). We then tested whether LWS-CNR-A is sufficient for transcriptional activation of both miR-726 and LWS-A. When the construct in which LWS-CNR-A alone was connected to the basal promoters of miR-726 and LWS-A was injected into medaka embryos, both EGFP and mCherry expression was observed in photoreceptor cells (the “CNR-A” construct in Fig. 6B and Supplementary Fig. 1). These results suggest that LWS-CNR-A acts as an enhancer for both miR-726 and LWS-A in photoreceptor cells. The role of CNRs in the intergenic region between miR-729 and SWS1 was also examined by introducing a series of dual-reporter constructs into medaka embryos (Fig. 7B). Deletion of SWS1-CNR- substantially decreased the expression of both mCherry and EGFP, whereas deletion of SWS1-CNR-did not affect the reporter expression (Fig. 7B and Supplementary Fig. 2). Thus SWS1-CNR- in combination with the basal promoters seems to be sufficient for transcriptional activation of both miR-729 and SWS1. These results suggest that transcription of the paired miRNA and opsin genes is co-regulated by common cis-regulatory modules (Fig. 8). Transcription factors that interact with LWS-CNR-A and 15

SWS1-CNR- were predicted by searching potential binding sites for known transcription factors in silico (Figs. 6C and 7C). Medaka LWS-CNR-A contains three potential biding sites for the photoreceptor-specific transcription factor Crx (Chen et al., 1997; Freund et al., 1997; Furukawa et al., 1997). These Crx sites contain an ATTA sequence, which is a core motif of homeodomain recognition sequences (Levine and Hoey, 1988; Hayashi and Scott, 1990). Two additional ATTA sequences are also highly conserved in the LWS-CNR-A regions of teleots (Fig. 6C). Other potential sites in medaka LWS-CNR-A include binding sequences for nuclear receptor family proteins, Nr2e3 and RXR, which are known to be involved in photoreceptor development (Chen et al., 2005; Swaroop et al., 2010). These sites are less conserved among teleots. Medaka SWS1-CNR- contains two potential sites for Crx (Fig. 7C). These Crx sites are highly conserved between medaka and stickleback. Potential biding sites for various transcription factors, such as Hox, Fox, POU, bZip, and nuclear receptor family proteins, are also found, and some of these are well conserved (Fig. 7C). Another interesting sequence is a potential binding site for STAT3, which is a regulator for differentiation of rod photoreceptor cells (Rhee et al., 2004; Ozawa et al., 2004; Zhang et al., 2004). A STAT3 site is also found in medaka LWS-A (Fig. 6C).

Predicted miRNA targets include transcription factors involved in the regulation of photoreceptor development At present, information on 3' UTRs in medaka is limited compared to that in zebrafish; e.g., 358 vs. 18,233 entries in the UTR database UTRdb (http://utrdb.ba.itb.cnr.it/; Grillo et al., 2010). Therefore, we looked at putative targets of miR-726 and miR-729 in annotated zebrafish UTRs (Ulitsky et al., 2012) using the TargetScan system (Lewis et al., 2005). A total of 2,986 and 2,157 16

genes are predicted to be targets of miR-726 and miR-729, respectively. At least four transcription factors that are involved in development of photoreceptor subtypes (Nrl, thyroid hormone receptor , Six7, and ROR) are potential targets of dre-miR-726 (Supplementary Fig. 3). Potential targets of dre-miR-729 include three homeodomain proteins (Six7, Crx, and Otx2) that play an important role in photoreceptor development (Supplementary Fig. 4A,B). We compared orthologues of these potential targets between medaka and zebrafish and found that the 3' UTR of medaka Six7 contains a potential target sequence of ola-miR-729 (Supplementary Fig. 4C,D).

Discussion Cone subtype-specific expression and a putative role of miR-726 and miR-729 The present results suggest that miR-726 and miR-729 are specifically expressed in red-sensitive cones and UV-sensitive cones, respectively. In the previous study in zebrafish, expression of miR-726 and miR-729 was detected in the eye by northern blot analysis, but in situ hybridization using LNA probes did not detect any signals for miR-726 and miR-729 (Kloosterman et al., 2006). In this study we detected expression of primary transcripts using conventional RNA probes, because so far we have not succeeded to detect any signals using LNA probes for mature ola-miR-726 and ola-miR-729. As far as we know, there are no reports on expression patterns of these miRNAs in any other species. There are also no previous reports on photoreceptor subtype-specific expression of any miRNAs. Thus, the present study provides the first evidence for photoreceptor subtype-specific expression of miRNAs. 17

The photoreceptor subtype-specific expression of miR-726 and miR-729 suggests that they play a role in the regulation of photoreceptor subtypes in the retina. These miRNAs may posttranscriptionally suppress the expression of genes that are not supposed to be expressed in a particular cone subtype. For example, miR-726 expressed in long wave-sensitive cones may repress transcription factors that promote differentiation of short wave-sensitive cones, middle wave-sensitive cones, or rods. Because different subtypes of photoreceptor cells share common features as photosensory cells and develop from common progenitor cells, the genes expressed in different photoreceptor subtypes should partially overlap. For example, different subtypes of cones use the same isoforms of phototransduction proteins (Hisatomi et al., 2002; Larhammar et al., 2009). A posttranscriptional repressive mechanism by miRNAs would help ensure a clear distinction between photoreceptor subtypes. In silico analysis revealed that several transcription factors involved in the regulation of photoreceptor development are potential targets of miR-726 and miR-729. One of the potential targets of miR-726 is the Maf-family bZIP transcription factor Nrl, which promotes rod specification and development (Mears et al., 2001; McIlvain and Knox, 2007; Oh et al., 2007). Rods are transformed to functional cones in the Nrl–/– mouse retina (Mears et al., 2001; Daniele et al., 2005; Akimoto et al., 2006). Conversely, ectopic expression of NRL can transform cone precursors to functional rods (Oh et al., 2007). Another potential target of miR-726 is ROR, which acts on the same transcriptional pathway as Nrl in directing precursors to a rod state (Jia et al., 2009). Thus, one of the functions of miR-726 may be to repress rod phenotype by downregulating Nrl and ROR. The potential targets of miR-729 includes two members of 18

the Otx family, Otx2 and Crx. Function of these transcription factors are necessary for development of both rods and cones (Chen et al., 1997; Freund et al., 1997; Furukawa et al., 1997; Nishida et al., 2003). The coordinated and precisely balanced action of the six key transcription factors (Otx2, Crx, ROR, Nrl, Nr2e3, and thyroid hormone receptor 2) has been proposed to be crucial as retinal progenitor cells commit to a rod or cone photoreceptor cell fate in the mammalian retina (Swaroop et al., 2010). Because five of these factors are found to be potential targets of either miR-726 or miR-729, these microRNAs may regulate development of photoreceptor subtypes through differential downregulation of a specific subset of the key transcription factors. The sine oculis homologue Six7 is predicted to be a potential target of both miR-726 and miR-729. Conservation of the miR-729 target sequence between zebrafish and medaka supports the validity of the prediction. Because the role of Six7 in photoreceptor development is only poorly understood, however, it is currently difficult to speculate a regulatory role of Six7 suppression by these microRNAs. The predicted targets of miR-726 and miR-729 discussed above are upstream regulatory factors for photoreceptor development, and their potential biding sites are found in the conserved cis-regulatory regions (LWS-CNR-A and SWS1-CNR-) of the microRNA/cone opsin gene pairs. For example, Crx and Otx2 are potential targets of miR-729, and at the same time, SWS1-CNR- contains highly conserved binding sequences for Crx/Otx. Therefore, miR-726 and miR-729 may mediate a negative feedback loop to control a photoreceptor cell fate decision.

Evolutionary conservation of miRNA/opsin pairs 19

The miR-726/LWS pair is conserved in teleosts, amphibians, and reptiles. This conservation profile indicates that the ancestral miR-726/LWS pair was present in the last common ancestor of teleosts and tetrapods. The evolutionary conservation of this pair suggests the functional importance of miR-726 and also the importance of the genomic organization. Interestingly, miR-726 has not been found in the genomes of birds and mammals, suggesting that miR-726 has been independently lost in each of the avian and mammalian lineages. The LWS opsins constitute the most highly conserved cone opsin subclass, which is present in the agnathan lamprey as well as in all gnathostome classes (Davies et al., 2012). The miR-726/LWS pair, therefore, would provide a unique model to study the mechanisms and roles of acquisition and loss of an miRNA tightly associated with a specific cell-type during vertebrate evolution. In the medaka genome, a blue-sensitive opsin gene SWS2-A and a violet sensitive gene opsin SWS2-B are located upstream of LWS-A and another red-sensitive opsin gene LWS-B is located downstream of LWS-A, forming a tandem cluster of four opsin genes (Matsumoto et al., 2006). The ola-miR-726 gene is located between SWS2-B and LWS-A and corresponds to a recently identified highly conserved intergenic region termed Region I (Watson et al., 2010) or CNE 7a (O'Quin et al., 2011). The close linkage of SWS2 and LWS opsin genes is also evolutionarily highly conserved (Wakefield et al., 2008; Watson et al., 2010). In the present study, we could not directly compare expression of SWS2 genes and ola-miR-726, because we failed to detect SWS2 expression by in situ hybridization. Judging from the regularly arranged mosaic pattern of cone-subtypes in the medaka retina (Hisatomi et al., 1997), ola-miR-726 is probably not co-expressed with SWS2-A and SWS2-B. On the other hand, we can not exclude 20

the possibility that ola-miR-726 is co-expressed with LWS-B, because the sequences of LWS-A and LWS-B are almost identical and they seem to be expressed in the same subtype of cones (Matsumoto et al., 2006). Visual opsins of the SWS1 subclass are also well conserved across a wide range of vertebrates from agnathans to humans (Davies et al., 2012). Nevertheless, the miR-729/SWS1 pair has been found only in teleosts. On the other hand, the miR-729/SWS1 pair is highly conserved among teleosts. This situation suggests that miR-729 plays an important yet teleost-specific role related to the development or function of short-wave sensitive cones. An early teleost ancestor may have acquired miR-729 as an evolutionary novelty. Considering the current incompleteness of the genome sequencing across diverse vertebrate species, however, we cannot exclude the possibility that the miR-729/SWS1 pair is present in some non-teleost species.

Transcription factors that interact with LWS-CNR-A and SWS1-CNR- In zebrafish, the intergenic region corresponding to LWS-CNR-A has been identified as a locus control region (LAR; LWS-activating region) regulating expression of two tandemly linked LWS genes, LWS-1 and LWS-2 (Tsujimura et al., 2010). Two Otx (A/GGATTA) and one Otx-like (TGATTA) sequences are present in the zebrafish LAR. The Otx sites correspond to the conserved Crx sites identified in the present study. The region corresponding to LWS-CNR-A was also identified as Region II or CNE 7b in recent comparative studies of teleost LWS loci (Watson et al., 2010; O'Quin et al., 2011). In these studies, in addition to the conserved Crx binding sites, a highly conserved element is identified as a potential binding site for nuclear receptor family factors (HRE in Fig. 21

6C; Watson et al., 2010; O'Quin et al., 2011). Various nuclear receptor family proteins, including Nr2e3, TR2, RXR, ROR and ROR, are known to be involved in the regulation of photoreceptor development, and some of these proteins may interact with this highly conserved element.

Bidirectional cis-regulatory regions for the mRNA-miRNA pairs as a means to couple transcriptional and posttranscriptional regulations We showed that a gene encoding cone opsin and a gene encoding an miRNA precursor are closely linked in a head-to-head arrangement on opposite DNA strands. This type of bidirectional gene configuration has been implicated in the coordinated regulation of paired genes in various metazoans, including Drosophila (Logan et al., 1989; Hofmann et al., 1991; Herr and Harris, 2004), ascidians (Kusakabe et al., 1995), and mammals (Burbelo et al., 1988; Adachi and Lieber, 2002; Trinklein et al., 2004). In humans, a class of divergently transcribed gene pairs, whose intergenic regions are less than 1000 base pairs in length, represents more than 10% of the genes in the genome (Trinklein et al., 2004). Most of these human gene pairs encode housekeeping proteins and their short intergenic regions are thought to be bidirectional promoters (Adachi and Lieber, 2002; Trinklein et al., 2004). In contrast, the intergenic regions of the miR-726/LWS-A and miR-729/SWS1 pairs are relatively long, 3.9 kb and 1.1 kb, respectively, and the paired genes are highly tissue-specific. Our cis-regulatory analysis suggests that each of the paired genes has its own core promoter but that a common enhancer located in the intergenic region regulates transcription of the paired genes. Thus, bidirectional gene organization of the miRNA/opsin pairs seems to be functionally significant. They may represent another class of divergently transcribed gene pairs. 22

Another example of a divergently transcribed pair of an miRNA gene and a protein-coding gene has been reported in humans; miR-34b/c and the B-cell translocation gene 4 (BTG4) are transcribed from a bidirectional promoter (Toyota et al., 2008). Both miR-34b/c and BTG4 are thought to be tumor suppressors, so that their coordinated regulation could be adaptive. Divergently transcribed pairs of miRNA/protein genes may have evolved to efficiently add a posttranscriptional regulatory mechanism by utilizing the already existing transcriptional machinery in a particular biological pathway, such as cell-type specification and tumor suppression. At present, however, it is unclear how unique or widespread this type of gene pair is. A systematic genome-wide survey in the genomes of various organisms would reveal further examples of bidirectional miRNA/protein gene pairs.

Acknowledgments We thank Drs. Kotaro Shimai, Koki Nishitsuji, and Rie Kusakabe for their valuable input, and NBRP-Medaka for providing us with the medaka hatching enzyme. This study was supported in part by Grants-in-Aid for Scientific Research from JSPS to T.G.K. (22310120, 22657023, 25650118, 25290067). This study was also supported in part by research grants from the Hirao Taro Foundation of the Konan University Association for Academic Research.

References Adachi N., Lieber M.R., 2002. Bidirectional gene organization: a common architectural feature of the human genome. Cell 109, 807–809. 23

Akimoto, M., Cheng, H., Zhu, D., Brzezinski, J.A., Khanna, R., Filippova, E., Oh, E.C.T., Jing, Y., Linares, J.L., Brooks, M., Zareparsi, S., Mears, A.J., Hero, A., Glaser, T., Swaroop, A., 2006. Targeting of GFP to newborn rods by Nrl promoter and temporal expression profiling of flow-sorted photoreceptors. Proc. Natl. Acad. Sci. USA 103, 3890–3895. Allison, W.T., Barthel, L.K., Skebo, K.M., Takechi, M., Kawamura, S., Raymond, P.A., 2010. Ontogeny of cone photoreceptor mosaics in zebrafish. J. Comp. Neurol. 518, 4182–4195. Alvarez-Delfin, K., Morris, A.C., Snelson, C.D., Gamse, J.T., Gupta, T., Marlow, F.L., Mullins, M.C., Burgess, H.A., Granato, M., Fadool, J.M., 2009. Tbx2b is required for ultraviolet photoreceptor cell specification during zebrafish retinal development. Proc. Natl. Acad. Sci. USA 106, 2023–2028. Alvarez-Delfin, K., 2011. Transcriptional Regulation of Photoreceptor Development in the Zebrafish Retina. Electronic Theses, Treatises and Dissertations. Paper 4597. Florida State University. http://diginole.lib.fsu.edu/etd/4597 Ambros, V., 2004. The functions of animal microRNAs. Nature 431, 350–355. Applebury, M.L., Farhangfar, F., Glösmann, M., Hashimoto, K., Kage, K., Robbins, J.T., Shibusawa, N., Wondisford, F.E., Zhang, H., 2007. Transient expression of thyroid hormone nuclear receptor TR2 sets S opsin patterning during cone photoreceptor genesis. Dev. Dyn. 236, 1203–1212. Bartel, D.P., 2004. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281–297. Bartel, D.P., 2009. MicroRNAs: target recognition and regulatory functions. Cell 136, 215–233. Bowmaker, J.K., 2008. Evolution of vertebrate visual pigments. Vision Res. 48, 2022–2041. 24

Burbelo, P.D., Martin, G.R., Yamada, Y., 1988. D1(IV) and D2(IV) collagen genes are regulated by a bidirectional promoter and a shared enhancer. Proc. Natl. Acad. Sci. USA 85, 9679–9682. Chen, J., Rattner, A., Nathans. J., 2005. The rod photoreceptor-specific nuclear receptor Nr2e3 represses transcription of multiple cone-specific genes. J. Neurosci. 25, 118–129. Chen, S., Wang, Q.L., Nie, Z., Sun, H., Lennon, G., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Zack, D.J., 1997. Crx, a novel Otx-like paired-homeodomain protein, binds to and transactivates photoreceptor cell-specific genes. Neuron 19, 1017–1030. Cheng, H., Khanna, H., Oh, E.C., Hicks, D., Mitton, K.P., Swaroop, A., 2004. Photoreceptor-specific nuclear receptor NR2E3 functions as a transcriptional activator in rod photoreceptors. Hum. Mol. Genet. 13, 1563–1575. Damiani, D., Alexander, J.J., O'Rourke, J.R., McManus, M., Jadhav, A.P., Cepko, C.L., Hauswirth, W.W., Harfe, B.D., Strettoi, E., 2008. Dicer inactivation leads to progressive functional and structural degeneration of the mouse retina. J. Neurosci. 28, 4878–4887. Daniele, L.L., Lillo, C., Lyubarsky, A.L., Nikonov, S.S., Philp, N., Mears, A.J., Swaroop, A., Williams, D.S., Pugh, E.N., Jr., 2005. Cone-like morphological, molecular, and electrophysiological features of the photoreceptors of the Nrl knockout mouse. Invest. Ophthalmol, Vis. Sci. 46, 2156–2167. Davies, W.I., Collin, S.P., Hunt, D.M., 2012. Molecular ecology and adaptation of visual photopigments in craniates. Mol. Ecol. 21, 3121–3158. Davis, N., Mor, E., Ashery-Padan, R., 2011. Roles for Dicer1 in the patterning and differentiation of the optic cup neuroepithelium. 25

Development 138, 127–138. Forrest, D., Swaroop, A., 2012. The role of nuclear receptors in photoreceptor differentiation and disease. Mol. Endocrinol. 26, 905–915. Frazer, K.A., Pachter, L., Poliakov, A., Rubin, E.M., Dubchak, I., 2004. VISTA: computational tools for comparative genomics. Nucleic Acids Res. 32, W273–W279. Freund, C.L., Gregory-Evans, C.Y., Furukawa, T., Papaioannou, M., Looser, J., Ploder, L., Bellingham, J., Ng, D., Herbrick, J.A., Duncan, A., Scherer, S.W., Tsui, L.C., Loutradis-Anagnostou, A., Jacobson, S.G., Cepko, C.L., Bhattacharya, S.S., McInnes, R.R., 1997. Cone-rod dystrophy due to mutations in a novel photoreceptor specific homeobox gene (CRX) essential for maintenance of the photoreceptor. Cell 91, 543–553. Fujieda, H., Bremner, R., Mears, A.J., Sasaki, H., 2009. Retinoic acid receptorrelated orphan receptor  regulates a subset of cone genes during mouse retinal development. J. Neurochem. 108, 91–101. Furukawa, T., Morrow, E.M., Cepko, C.L., 1997. Crx, a novel otx-like homeobox gene, shows photoreceptor-specific expression and regulates photoreceptor differentiation. Cell 91, 531–541. Georgi, S.A., Reh, T.A., 2010. Dicer is required for the transition from early to late progenitor state in the developing mouse retina. J. Neurosci. 30, 4048–4061. Glaschke, A., Weiland, J., Del Turco, D., Steiner, M., Peichl, L., Glösmann, M., 2011. Thyroid hormone controls cone opsin expression in the retina of adult rodents. J. Neurosci. 31, 4844–4851. Grillo, G., Turi, A., Licciulli, F., Mignone, F., Liuni, S., Banfi, S., Gennarino, V.A., Horner, D.S., Pavesi, G., Picardi, E., Pesole, G., 2010. UTRdb and UTRsite (RELEASE 2010): a collection of 26

sequences and regulatory motifs of the untranslated regions of eukaryotic mRNAs. Nucleic Acids Res. 38, D75–D80. Hayashi, S., Scott, M.P., 1990. What determines the specificity of action of Drosophila homeodomain proteins? Cell 63, 883–894. Herr, D.R., Harris, G.L., 2004. Close head-to-head juxtaposition of genes favors their coordinate regulation in Drosophila melanogaster. FEBS Lett. 572, 147–153. Hisatomi, O., Satoh, T., Tokunaga, F., 1997. The primary structure and distribution of killifish visual pigments. Vision Res. 37, 3089–3096. Hisatomi, O., Tokunaga, F., 2002. Molecular evolution of proteins involved in vertebrate phototransduction. Comp. Biochem. Physiol. B 133, 509-522. Hofmann, A., Garfinkel, M.D., Meyerowitz, E.M., 1991. cis-acting sequences required for expression of the divergently transcribed Drosophila melanogaster Sgs-7 and Sgs-8 glue protein genes. Mol. Cell. Biol. 11, 2971–2979. Jia, L., Oh, E.C., Ng, L., Srinivas, M., Brooks, M., Swaroop, A., Forrest, D., 2009. Retinoid-related orphan nuclear receptor ROR is an early-acting factor in rod photoreceptor development. Proc. Natl. Acad. Sci. USA 106, 17534–17539. Kefalov, V.J., 2012. Rod and cone visual pigments and phototransduction through pharmacological, genetic, and physiological approaches. J. Biol. Chem. 287, 1635–1641. Kloosterman, W.P., Steiner, F.A., Berezikov, E., de Bruijn, E., van de Belt, J., Verheul, M., Cuppen, E., Plasterk, R.H., 2006. Cloning and expression of new microRNAs from zebrafish. Nucleic Acids Res. 34, 2558–2569. Kusakabe, R., Kusakabe, T., Suzuki, N., 1999. In vivo analysis of two striated muscle actin promoters reveals combinations of multiple regulatory modules required for skeletal and cardiac 27

muscle-specific gene expression. Int. J. Dev. Biol. 43, 541–554. Kusakabe, T., Hikosaka, A., Satoh, N., 1995. Coexpression and promoter function in two muscle actin gene complexes of different structural organization in the ascidian Halocynthia roretzi. Dev. Biol. 169, 461–472. Kusakabe, T., Suzuki, N., 2000. Photoreceptors and olfactory cells express the same retinal gyanylyl cyclase isofom in medaka: visualization by promoter transgenics. FEBS Lett. 483, 143–148. Kusakabe, T.G., Sakai, T., Aoyama, M., Kitajima, Y., Miyamoto, Y., Takigawa, T., Daido, Y., Fujiwara, K., Terashima, Y., Sugiuchi, Y., Matassi, G., Yagisawa, H., Park, M.K., Satake, H., Tsuda, M., 2012. A conserved non-reproductive GnRH system in chordates. PLoS One 7, e41955. Larhammar, D., Nordström, K., Larsson, T.A., 2009. Evolution of vertebrate rod and cone phototransduction genes. Phil. Trans. R. Soc. Lond. B 364, 2867–2880. Levine, M., Hoey, T., 1988. Homeobox proteins as sequence-specific transcription factors. Cell 55, 537–540. Lewis, B.P., Burge, C.B., Bartel, D.P., 2005. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120, 15–20. Li, S.C., Chan, W.C., Ho, M.R., Tsai, K.W., Hu, L.Y., Lai, C.H., Hsu, C.N., Hwang, P.P., Lin, W.C., 2010. Discovery and characterization of medaka miRNA genes by next generation sequencing platform. BMC Genomics 11, Suppl. 4, S8. Logan, S.K., Garabedian, M.J., Wensink, P.C., 1989. DNA regions that regulate the ovarian transcriptional specificity of Drosophila yolk protein genes. Genes Dev. 3, 1453–1461. Lu, A., Ng, L., Ma, M., Kefas, B., Davies, T.F., Hernandez, A., Chan, C.C., Forrest, D., 2009. Retarded developmental expression and 28

patterning of retinal cone opsins in hypothyroid mice. Endocrinology 150, 1536–1544. Lumayag, S., Haldin, C.E., Corbett, N.J., Wahlin, K.J., Cowan, C., Turturro, S., Larsen, P.E., Kovacs, B., Witmer, P.D., Valle, D., Zack, D.J., Nicholson, D.A., Xu, S., 2013. Inactivation of the microRNA-183/96/182 cluster results in syndromic retinal degeneration. Proc. Natl. Acad. Sci. USA 110, E507–E516. Marchuk, D., Drumm, M., Saulino, A., Collins, F.S., 1991. Construction of T-vectors, a rapid and general system for direct cloning of unmodified PCR products. Nucleic Acids Res. 19, 1154. Mathelier, A., Zhao, X., Zhang, A.W., Parcy, F., Worsley-Hunt, R., Arenillas, D.J., Buchman, S., Chen, C.-y., Chou, A., Ienasescu, H., Lim, J., Shyr, C., Tan, G., Zhou, M., Lenhard, B., Sandelin, A., Wasserman, W.W., 2014. JASPAR 2014: an extensively expanded and updated open-access database of transcription factor binding profiles. Nucleic Acids Res.42, D142–D147. Matsumoto, Y., Fukamachi, S., Mitani, H., Kawamura, S., 2006. Functional characterization of visual opsin repertoire in Medaka (Oryzias latipes). Gene 371, 268–278. McIlvain, V.A., Knox, B.E., 2007. Nr2e3 and Nrl can reprogram retinal precursors to the rod fate in Xenopus retina. Dev. Dyn. 236, 1970–1979. Mears, A.J., Kondo, M., Swain, P.K., Takada, Y., Bush, R.A., Saunders, T.L., Sieving, P.A., Swaroop, A., 2001. Nrl is required for rod photoreceptor development. Nat. Genet. 29, 447–452. Nathans, J., 1987. Molecular biology of visual pigments. Annu. Rev. Neurosci. 10, 163–194. Ng, L., Hurley, J.B., Dierks, B., Srinivas, M., Salto, C., Vennstrom, B., Reh, T.A., Forrest, D., 2001. A thyroid hormone receptor that is required for the development of green cone photoreceptors. 29

Nat. Genet. 27, 94–98. Nishida, A., Furukawa, A., Koike, C., Tano, Y., Aizawa, S., Matsuo, I., Furukawa, T., 2003. Otx2 homeobox gene controls retinal photoreceptor cell fate and pineal gland development. Nat. Neurosci. 6, 1255–1263. Oh, E.C.T., Khan, N., Novelli, E., Khanna, H., Strettoi, E., Swaroop, A., 2007. Transformation of cone precursors to functional rod photoreceptors by bZIP transcription factor NRL. Proc. Natl. Acad. Sci. USA 104, 1679–1684. Onishi, A., Peng, G.H., Chen, S., Blackshaw, S., 2010. Pias3-dependent SUMOylation controls mammalian cone photoreceptor differentiation. Nat. Neurosci. 13, 1059–1065. O'Quin, K.E., Smith, D., Naseer, Z., Schulte, J., Engel, S.D., Loh, Y.H., Streelman, J.T., Boore, J.L., Carleton, K.L., 2011. Divergence in cis-regulatory sequences surrounding the opsin gene arrays of African cichlid fishes. BMC Evol. Biol. 11, 120. Ozawa, Y., Nakao, K., Shimazaki, T., Takeda, J., Akira, S., Ishihara, K., Hirano, T., Oguchi, Y., Okano, H., 2004. Downregulation of STAT3 activation is required for presumptive rod photoreceptor cells to differentiate in the postnatal retina. Mol. Cell. Neurosci. 26, 258–270. Palczewski, K., 2012. Chemistry and biology of vision. J. Biol. Chem. 287, 1612–1619. Pinter, R., Hindges, R., 2010. Perturbations of microRNA function in mouse dicer mutants produce retinal defects and lead to aberrant axon pathfinding at the optic chiasm. PLoS ONE 5, e10021. Roberts, M.R., Hendrickson, A., McGuire, C.R., Reh, T.A., 2005. Retinoid X receptor  is necessary to establish the S-opsin gradient in cone photoreceptors of the developing mouse retina. Invest. Ophthalmol. Vis. Sci. 46, 2897–2904. 30

Roberts, M.R., Srinivas, M., Forrest, D., Morreale de Escobar, G., Reh, T.A., 2006. Making the gradient: thyroid hormone regulates cone opsin expression in the developing mouse retina. Proc. Natl. Acad. Sci. USA 103, 6218–6223. Rhee, K.D., Goureau, O., Chen, S., Yang, X.J., 2004. Cytokine-induced activation of signal transducer and activator of transcription in photoreceptor precursors regulates rod differentiation in the developing mouse retina. J. Neurosci. 24, 9779–9788. Saade, C.J., Alvarez-Delfin K., Fadool, J.M., 2013. Rod photoreceptors protect from cone degeneration-induced retinal remodeling and restore visual responses in zebrafish. J Neurosci. 33, 1804–1814. Srinivas, M., Ng, L., Liu, H., Jia, L., Forrest, D., 2006. Activation of the blue opsin gene in cone photoreceptor development by retinoid-related orphan receptor beta. Mol. Endocrinol. 20, 1728–1741. Steffen, P., Voss, B., Rehmsmeier, M., Reeder, J., Giegerich, R., 2006. RNAshapes: an integrated RNA analysis package based on abstract shapes. Bioinformatics 22, 500–503. Sundermeier, T.R., Palczewski, K., 2012. The physiological impact of microRNA gene regulation in the retina. Cell. Mol. Life Sci. 69, 2739–2750. Swaroop, A., Kim, D., Forrest, D., 2010. Transcriptional regulation of photoreceptor development and homeostasis in the mammalian retina. Nat. Rev. Neurosci. 11, 563–576. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680. 31

Toyota, M., Suzuki, H., Sasaki, Y., Maruyama, R., Imai, K., Shinomura, Y., Tokino, T., 2008. Epigenetic silencing of microRNA-34b/c and B-cell translocation gene 4 is associated with CpG island methylation in colorectal cancer. Cancer Res. 68, 4123–4132. Trinklein, N.D., Aldred, S.F., Hartman, S.J., Schroeder, D.I., Otillar, R.P., Myers, R.M., 2004. An abundance of bidirectional promoters in the human genome. Genome Res. 14, 62–66. Tsujimura, T., Hosoya, T., Kawamura, S., 2010. A single enhancer regulating the differential expression of duplicated red-sensitive opsin genes in zebrafish. PLoS Genet. 6, e1001245. Ulitsky, I., Shkumatava, A., Jan, C.H., Subtelny, A.O., Koppstein, D., Bell, G.W., Sive, H., Bartel, D.P., 2012. Extensive alternative polyadenylation during zebrafish development. Genome Res. 22, 2054–2066. Wakefield, M.J., Anderson, M., Chang, E., Wei, K.J., Kaul, R., Graves, J.A., Grützner, F., Deeb, S.S.,2008. Cone visual pigments of monotremes: filling the phylogenetic gap. Vis. Neurosci. 25, 257–264. Watson, C.T., Lubieniecki, K.P., Loew, E., Davidson, W.S., Breden, F., 2010. Genomic organization of duplicated short wave-sensitive and long wave-sensitive opsin genes in the green swordtail, Xiphophorus helleri. BMC Evol. Biol. 10, 87. Yoshida, R., Horie, T., Tsuda, M., Kusakabe, T.G., 2007. Comparative genomics identifies a cis-regulatory module that actives transcription in specific subsets of neurons in Ciona intestinalis larvae. Dev. Growth Differ. 49, 657–667. Xu, S., Witmer, P.D., Lumayag, S., Kovacs, B., Valle, D., 2007. MicroRNA (miRNA) transcriptome of mouse retina and identification of a sensory organ-specific miRNA cluster. J. Biol. Chem. 282, 25053–25066. 32

Zhang, S.S., Wei, J., Qin, H., Zhang, L., Xie, B., Hui, P., Deisseroth, A., Barnstable, C.J., Fu, X.Y., 2004. STAT3-mediated signaling in the determination of rod photoreceptor cell fate in mouse retina. Invest. Ophthalmol. Vis. Sci. 45, 2407–2412. Zhu, Q., Sun, W., Okano, K., Chen, Y., Zhang, N., Maeda, T., Palczewski, K., 2011. Sponge transgenic mouse model reveals important roles for the microRNA-183 (miR-183)/96/182 cluster in postmitotic photoreceptors of the retina. J. Biol. Chem. 286, 31749–31760.

33

Figure Legends

Figure 1. A highly conserved noncoding sequence in the upstream region of LWS-A encodes a microRNA (miRNA). (A) Comparison of genomic sequences of the LWS-A locus among teleosts, an amphibian, and a reptile using VISTA (Frazer et al., 2004) revealed a highly conserved noncoding region (green box). Numbers in the parentheses indicate the distance (in base pairs) between 3’ end of the putative miRNA sequence and the start codon of the opsin coding sequence in each species. (B) An alignment of the nucleotide sequences of this region. Highly conserved nucleotide positions are indicated by asterisks, and putative mature miRNA sequences are indicated in red. (C) The predicted medaka miRNA sequence is identical to that of a zebrafish miRNA, dre-miR-726 (Kloosterman et al., 2006). The “seed” sequence from position 2 to 8 is indicated in blue. (D) The stem-loop structure of pre-ola-miR-726 predicted by using RNAshapes (Steffen et al. 2006).

Figure 2. A highly conserved noncoding sequence in the upstream region of SWS1 encodes a miRNA. (A) Comparison of genomic sequences of the SWS1 locus among teleosts using VISTA revealed a highly conserved non-coding region (green box). Numbers in the parentheses indicate the distance (in base pairs) between 3’ end of the putative miRNA sequence and the start codon of the opsin coding sequence in each species. (B) An alignment of nucleotide sequences of the highly conserved region. Nucleotide positions with identical bases among three species are indicated by asterisks, and putative mature miRNA sequences are indicated in red. (C) The predicted medaka miRNA sequence is similar to that of a zebrafish microRNA, dre-miR-729 (Kloosterman et al., 2006), with the “seed” 34

region being perfectly matched. The seed sequence from position 2 to 8 is indicated in blue. (D) The stem-loop structure of pre-ola-miR-726 predicted by using RNAshapes.

Figure 3. Cone subtype-specific expression of primary transcripts for ola-miR-726 (pri-ola-miR-726) and ola-miR-729 (pri-ola-miR729). Localization of pri-ola-miR-726 (A, C, E, and G) and pri-ola-miR-729 (B, D, F, and H) was compared with that of mRNAs for red-sensitive opsin LWS-A (A and D), UV-sensitive opsin SWS1 (B and C), green-sensitive opsin RH2 (E and F) and rod-specific rhodopsin (G and H) by double-fluorescence in situ hybridization of retinal sections from a 3-month-old fish. (A and B) Signals for primary miRNAs (green) are located beneath the respective cone opsin signals (magenta). Some typical sites are indicated by arrows. (C and E) Signals for pri-ola-miR726 (green, also shown by arrows) are not located beneath the UV-sensitive and green-sensitive cone opsin signals. (D and F) Signals for pri-ola-miR729 (green, also shown by arrows) are not located beneath the red-sensitive and green-sensitive cone opsin signals. (G and H) Rod nuclei surrounded by rhodopsin mRNA localization (magenta) are on the inner side of ONL and not overlapped with primary miRNA localization (green). (I) Localizations of pri-ola-miR-726 (magenta, also shown by arrowheads) and pri-ola-miR-729 (green, also shown by arrows) were compared by double-fluorescence in situ hybridization. (J) A control section using sense probes for SWS1 (magenta) and pri-ola-miR-729 (green). Nuclei were visualized with DAPI for all sections. Note that the outer segments of cone photoreceptor cells exhibit background fluorescence (due to enhanced sensitivity for primary miRNA detection), which is also visible in sections with sense probes (J). Scale bars = 20 μm (A-I) or 30 μm (J). 35

Figure 4. Bidirectional transcription in the same photoreceptor cells from the intergenic region between ola-miR-726 and LWS-A. (A) Schematic diagram showing the dual reporter construct mCherryEGFP, in which ola-miR-726 and LWS-A were replaced with mCherry and EGFP, respectively. Numbers in parentheses indicate the nucleotide positions of the ends of the upstream regions relative to the translation start site of LWS-A. (B-D) Co-localization of mCherry (magenta in B and D) and EGFP (green in C and D) in photoreceptor cells (arrows) of the retina of 8 dpf medaka embryos developed from eggs injected with mCherryEGFP. Scale bars = 100 μm. (E-G) Higher magnification images of the region boxed with dashed lines in B-D. Scale bars = 50 μm.

Figure 5. Bidirectional transcription in the same photoreceptor cells from the intergenic regions between ola-miR-729 and SWS1. (A) Schematic diagram showing the dual reporter construct mCherryEGFP, in which ola-miR-729 and SWS1 were replaced with mCherry and EGFP, respectively. Numbers in parentheses indicate the nucleotide positions of the ends of the upstream regions relative to the translation start site of SWS1. (B-D) Co-localization of mCherry (magenta in B and D) and EGFP (green in C and D) in photoreceptor cells (arrows) of the retina of 8 dpf medaka embryos developed from eggs injected with mCherryEGFP. Scale bars = 100 μm. (E-G) Higher magnification images of the region boxed with dashed lines in B-D. Scale bars = 50 μm.

Figure 6. Function of conserved noncoding regions (CNRs) in bidirectional photoreceptor-specific transcription of the 36

ola-miR-726/LWS-A pair. (A) VISTA profile showing sequence conservation of the intergenic region among medaka, stickleback, and puffer fish. The three CNRs are designated as LWS-CNR-A, -B, and -C, respectively. In addition to these CNRs, the proximal promoter regions of ola-miR-726 and LWS-A are also conserved. Arrows indicate the positions and transcriptional direction of the ola-miR-726/LWS-A genes. (B) Structure of dual reporter constructs and their expression in photoreceptor cells of embryos at 8 dpf. The wild type or mutant forms of the intergenic region were connected to mCherry and EGFP reporter genes (left). The magenta and green boxes indicate mCherry- and EGFP-coding sequences, respectively, and the blue boxes indicate CNRs. The frequencies of embryos with reporter expression in retinal photoreceptor cells for each construct are shown as a percentage on the right. The "Co-expression" column indicates the percentage of embryos in which EGFP and mCherry were found in the same photoreceptor cells. Numbers in parentheses indicate the number of embryos scored for each construct. (C) An alignment of the nucleotide sequences of the CNR-A region of five teleost species. The location and direction of putative transcription factor binding sites in medaka CNR- predicted using the JASPAR core database (Mathelier et al., 2014) are indicated by dashed arrows. Yellow boxes indicate a putative core sequence (ATTA) for homeodomain proteins (Levine and Hoey, 1988; Hayashi and Scott, 1990). A highly conserved element identified as a hormone response element (HRE) in previous studies (Watson et al., 2010; O'Quin et al., 2011) is boxed. Nucleotide positions with identical bases among five species are indicated by asterisks.

Figure 7. Function of CNRs in bidirectional photoreceptor-specific transcription of the ola-miR-729/SWS1 pair. (A) VISTA profile 37

showing sequence conservation of the intergenic region among medaka, stickleback, and puffer fish. The two CNRs are designated as SWS1-CNR- and SWS1-CNR-, respectively. In addition to these CNRs, the proximal promoter regions of ola-miR-729 and SWS1 are also conserved. Arrows indicate the positions and transcriptional direction of the ola-miR-729/SWS1 genes. (B) Structure of dual reporter constructs (left) and their expression in photoreceptor cells of embryos at 8 dpf (right). The wild type or mutant forms of the intergenic region were connected to mCherry and EGFP reporter genes. The magenta and green boxes indicate mCherry- and EGFP-coding sequences, respectively, and the blue boxes indicate CNRs. The frequencies of embryos with reporter expression in retinal photoreceptor cells for each construct are shown as a percentage. The "Co-expression" column indicates the percentage of embryos in which EGFP and mCherry were found in the same photoreceptor cells. Numbers in parentheses indicate the number of embryos scored for each construct. (C) An alignment of the nucleotide sequences of the CNR- region of three teleost species. The location and direction of putative transcription factor binding sites in medaka CNR- predicted using the JASPAR core database (Mathelier et al., 2014) are indicated by dashed arrows. Yellow boxes indicate a putative core sequence (ATTA) for homeodomain proteins (Levine and Hoey, 1988; Hayashi and Scott, 1990). Nucleotide positions with identical bases among three species are indicated by asterisks.

Figure 8. Transcriptional regulation of the miRNA/opsin gene pairs. Co-expression of an miRNA/opsin gene pair is regulated by a common cis-regulatory module located in the shared upstream intergenic region. (A) Transcription of ola-miR-726 and LWS-A is activated in red-sensitive cones by LWS-CNR-A. (B) Transcription of 38

ola-miR-729 and SWS1 is activated in UV-sensitive cones by SWS1-CNR-.

Highlights x

Two bidirectional pairs of miRNA and cone opsin genes are conserved in vertebrates.

x

The paired miRNA and opsin genes are co-expressed in a particular cone subtype.

x

Their bidirectional transcription is regulated by a common cis-regulatory module.

x

These miRNAs may mediate a negative feedback loop for photoreceptor development.

x

These miRNAs may regulate development and/or function of photoreceptor subtypes.

39

Figure 1 (A)

LWS-A

ola-miR-726

4142 bp (2665)

Gasterosteus aculeatus

(2176)

Takifugu rubripes

(1782)

Tetraodon nigroviridis

(1609)

Danio rerio

(6741)

mg Ci-gnrh2 Xenopus

(>21572)

tropicalis

Anolis carolinensis

(B) Oryzias Gasterosteus Takifugu Tetraodon Danio Xenopus Anolis

TATTTCGAGAGTAGGTCTGGAATTCCGCTAGTTCTGAACTATTC----ATGC TGTTTCGAGGGTAGGTCTGGAATTCCGCTAGTTCTGAACTATTC----GGGA TATTTCGAGGGCAGGTCTGGAATTCCGCTAGTTCTGAACTATTC----ATGT TGTTTCGAGAGCAGGTCTGGAATTCCGCTAGTTCTGAACTATTC----ATGT TCAGTCAGTCGTATGCCTGGAATTCCGCTAGTTCTGAACTATTC----GTGA TCCTGTGGGTGCTTGTCAGGAATTCCGCTAGTTCTGAACTATTCCATGTTAG CTTTCTGGGCGCTCGTCAGGAATTCCGCTAGTTCTGAACTATTT----ACAG * * * ************************* TTGGAAAAAGTTCACTACTAGCAGAACTCGGATTTGCTCCCGACTGC-CGA TTGGTAAAAGTTCACTACTAGCAGAACTCAGATTTGCTCCCGACTAC-CGT TTGGTAAAAGTTCACTACTAGCAGAACTCAGATTTGCTCCCGACTAC-CGC TTGGTAAAAGTTCACTACTAGCAGAACTCAGATTTGCTCCCGACTAC-CGT TTGGCAAAAGTTCACTACTAGCAGAACTCGGATATACAACTGAATTCTCAA TAAGTAAAAGTTCACTACTAGCAGAACTCGGCCACGTAC-----CACACTT TTAGAAAAAGTTCACTACTAGCAGAACTCGGCCGCGCGCTCT-TCACTCTT * * ************************ * * *

(C) ola-miR-726 dre-miR-726

Ci-gnrh2::gfp 1 uucacuacuagcagaacucgg 21 ||||||||||||||||||||| 1 uucacuacuagcagaacucgg 21

(D) 5'- AUU

A

AA

C

UC

A--

U

UCG GAGUAGGUCUGG UUC GCUAGU UGAACU UUCA G ||| |||||||||||| ||| |||||| |||||| |||| ) AGC CUCGUUUAGGCU AAG CGAUCA ACUUGA AGGU C 3'- GUC C -C A UC AAA U

Figure 2

(A) SWS1

ola-miR-729

1250 bp (4212)

Gasterosteus aculeatus Ga

(2961)

Danio rerio Da

(B) Oryzias Gasterosteus Danio

G-GATCCGTACCCTCCAGAG--CAGGGGCTGTATCATAACCAGGCTGTGAGC-T TTGATCCAT---CATCAAAG--CTGGGGCTGTATCATAACCAGTCTGGTTAT-T AAGGGCCAT-----CCAGTGACCCGGGGTTGTATTGTAACCAAGCTGTAAGTGT * ** * ** * * **** ***** ****** *** * GAACGA-GCATGGGTATGATACGACCTCAGTT------TGGAGGAGCTCCTCCT TTAGAA-GCATGGGTATGATACGACCTCAGTT------TGTTGATGATTTTCTT CCATGATGCATGGGTATGATACGACCTGGGTTACAGAATGGTGATGGACATCCT * * ******************** *** ** * * ** *

(C) ola-miR-729 dre-miR-729

1 cauggguaugauacgaccucaguu 24 ||||||||||||||||||| ||| 1 cauggguaugauacgaccuggguu 24

(D) 5'- UAC

G

A

C

A

Ci-gnrh2::gfp

G

GA

CCUCCA AGC GGGG UGUAUCAUA CCA GCU-GU G |||||| ||| |||| ||||||||| ||| ||| || C GGAGGU-UUG CUCC GCAUAGUAU GGU CGA CA U 3'- CGA A A G A G AG

Figure 3

ONL

A

LWS-A antisense miR-726 antisense

C

SWS1 antisense miR-726 antisense

E

RH2 antisense miR-726 antisense

G

rhodopsin antisense miR-726 antisense

I

miR-726 antisense miR-726 antisense

B

SWS1 antisense miR-729 antisense

D

LWS-A antisense miR-729 antisense

F

RH2 antisense miR-729 antisense

H

rhodopsin antisense miR-729 antisense

J

SWS1 sense miR-729 sense

Figure 4

A

LWS-A

ola-miR-726

(+52)

(-3945)

EGFP >

< mCherry

B

C

D

E

F

G

Figure 5

A

ola-miR-729

(-1176)

SWS1

(+18)

EGFP >

< mCherry

B

C

D

E

F

G

Figure 6 (A)

ola-miR-726

LWS-A Gasterosteus aculeatus Takifugu rubripes

LWS-CNR-A

(B)

LWS-CNR-B

CNR-A

CNR-B CNR-C

< mCherry

WT

LWS-CNR-C

EGFP

comCherry expression

EGFP >

88%

63%

61% (n=41)

(mCherryEGFP)

∆A

< mCherry

EGFP >

25%

0%

0% (n=92)

∆B

< mCherry

EGFP >

90%

58%

58% (n=50)

∆C

< mCherry

EGFP >

82%

46%

42% (n=72)

∆B+C

< mCherry

EGFP >

83%

65%

63% (n=40)

CNR-A

< mCherry

EGFP >

63%

17%

17% (n=65)

(C) Nr2e3

Crx

Oryzias Gasterosteus Takifugu Tetraodon Danio

-----------> ------>-------AAGAGTCAAAG-GATTAAAATAAACCACTATAGCAGCTTCGATTAGGCCCAGCTTTCTCCTC CATTAG-CAACG-GATAAA-ATAAAC-----TAGCCGCTTTTGT------CGACCTGTGATTC -CAGAATCATCA-TATTAT-ATAAAGCAGTA-AAGGGCTGTTGTTATCCTCAGCTTTATCCCG -CAGAATCGTCA-TATTAT-TTAAAGCACTATAGGGGCTGTTGTTATCCTAAGCTCTATCCCG CAGAAACCGGGGAGATTAAAACTGACCTCATCAAGATTTTCAGATAAACCCTGCTT--TCCCT * * ** * * * * *

RXRα heterodimer

Crx ------------> ------------> AAGC-GACCCTTCAGTCCGGAGAGATTA--AAGTCTTACCGAGTGAGCATGCCAGAAGTAATT GAGCTGAACTGTCGCTCCGGAAAGATTA--ACTCGTTACGGAGTGAGCACGCCAGAAGTAATT CAAT-GAACCGTCGTTCCGGAGAGATTA--AAGTGTTACGGAGAGGCCCTGCCAGAAGTAATC CAGT-GAACCGTCATTCCGGAGAGATTA--AAGTGTTACGGAGAGGCCCTGCCAGAAGTAATT AAACTGAACCATCAATCCATTGGGATTAGGAAGGTTTGGGCTGCATCCTTGTCAATACTAATT * ** * ** *** ***** * ** * * * ** * **** STAT3

Crx HRE <-------------------> ACTCCGTGTTAATGAGTTTTGCAGAGAAACCTGAGTTGTCCGCATAATCTACTTTGGTCTGG ACTCTGTGGTAATCTGTTTTATGGAAGAACCTGACTTGTCCGTATAATCTGCTTTGGGATGA ACTCAGTGTTAATCTGTTTTGTGGAGGACCCTGAGTTGTCCCTGTAATCTTATTTGGTTTTT ACTCAGTGGTAATCTGTTTTGTGGAGGACCCTGAGTTGTCCCTGTAATCTCATTTGTTTTTT AGGATCTGGTG-TGGAGGGTACCTGGGTACCCCAGTTGTCCCAATAAGCTGATTTGGACTTG * ** * * * ** * ****** *** ** **** *

Figure 7

(A) ola-miR-729

SWS1 Gasterosteus aculeatus Danio rerio

SWS1-CNR-α

(B)

SWS1-CNR-β

CNR-α

WT

CNR-β

< mCherry

EGFP

comCherry expression

EGFP >

76%

44%

40% (n=78)

(mCherryEGFP)

∆α

< mCherry

EGFP >

16%

1%

0% (n=90)

∆β

< mCherry

EGFP >

61%

37%

37% (n=93)

CNR-α < mCherry

EGFP >

56%

22%

19% (n=67)

(C) NEF2L1::MafG Oryzias Gasterosteus Danio

Crx

STAT3

Hoxa9

<----<---------<---------<-------AGTCATTACCTC---CTAATCTGATTGG--------GTTTTCTGGATTTGCA-GGATTGAGG GGTCTTTTTGTGT--CTAATCTGATGATCTGACGATGTTTTGAGGATTTTCC-GGATTGAGC AATCCTTTTGTGACATCAATGCGGT------------TTGTAGGGATTTTTAGGGATTGA-** ** * *** * * ** * ****** *******

Nr2e3

FoxA1/POU2F2 Crx <------ ---------> Pax2 -<----------------------> AGGATGACTT-CTTAGGTAAATCTGAAACATCCGGATTAAACTTGAAGTC--ACGGA----A AGGACCACCTGCTCAGGTAAATCTGAAACATCCGGATTACATCTAAAGTCGCACGGACTGTA -----------TTTAGGTAAATCTTG--TTTCCAGATTACGGCTCAT-CCTTATGCTTGAAG * ********** *** ***** * * * * * Nr2e3

<-----AAGT-AAAGGTTC-------TTAGGGGTTTTTTATC----TGTCAGCC--ACCTGCTTTT AACCAAAAGGTTC-------TTATTGGATTTTTGTG----TGCAGACC--ACCTGCTTCT AACTTAAGAGTTTGGTTTGCCTTTCGCCTGCTTATGCTGTTGTGGACTTAACAAGCTTGT ** ** *** * * * ** * ** * ** **** *

Figure 8

(A) LWS-A

ola-miR-726

LWS-CNR-A

(B) ola-miR-729

SWS1-CNR-α

SWS1