Developmental Biology 386 (2014) 340–357

Contents lists available at ScienceDirect

Developmental Biology

journal homepage: www.elsevier.com/locate/developmentalbiology

The Prdm13 histone methyltransferase encoding is a Ptf1a–Rbpj downstream target that suppresses glutamatergic and promotes GABAergic neuronal fate in the dorsal neural tube

Julie Hanotel a, Nathalie Bessodes a, Aurore Thélie a, Marie Hedderich c, Karine Parain d, Benoit Van Driessche b, Karina De Oliveira Brandão a, Sadia Kricha a, Mette C. Jorgensen e, Anne Grapin-Botton e, Palle Serup e, Carine Van Lint b, Muriel Perron d, Tomas Pieler c, Kristine A. Henningfeld c, Eric J. Bellefroid a,n a Laboratory of Developmental Genetics, Université Libre de Bruxelles (ULB), Institute of Molecular Biology and Medicine, and ULB Neuroscience Institute, B-6041 Gosselies, Belgium b Laboratory of Molecular Virology, Université Libre de Bruxelles (ULB), Institute of Molecular Biology and Medicine, B-6041 Gosselies, Belgium c Department of Developmental Biochemistry, Center for Nanoscale Microscopy and Molecular Physiology of the Brain (CNMPB), University of Goettingen, 37077 Goettingen, Germany d UPR CNRS 3294 Neurobiology and Development, Université Paris Sud, 91405 Orsay Cedex, France e DanStem, University of Copenhagen, 3B Blegdamsvej, DK-2200 Copenhagen N, Denmark article info abstract

Article history: The basic helix–loop–helix (bHLH) transcriptional activator Ptf1a determines inhibitory GABAergic over Received 15 April 2013 excitatory glutamatergic neuronal cell fate in progenitors of the vertebrate dorsal spinal cord, cerebellum Received in revised form and retina. In an in situ hybridization expression survey of PR domain containing encoding putative 19 November 2013 chromatin-remodeling zinc finger transcription factors in Xenopus embryos, we identified Prdm13 as a Accepted 17 December 2013 histone methyltransferase belonging to the Ptf1a synexpression group. Gain and loss of Ptf1a function Available online 24 December 2013 analyses in both frog and mice indicates that Prdm13 is positively regulated by Ptf1a and likely Keywords: constitutes a direct transcriptional target. We also showed that this regulation requires the formation Prdm genes of the Ptf1a–Rbp-j complex. Prdm13 knockdown in Xenopus embryos and in Ptf1a overexpressing Ptf1a ectodermal explants lead to an upregulation of Tlx3/Hox11L2, which specifies a glutamatergic lineage and Neurog2 a reduction of the GABAergic neuronal marker Pax2. It also leads to an upregulation of Prdm13 Pax2 Tlx3 transcription, suggesting an autonegative regulation. Conversely, in animal caps, Prdm13 blocks the Histone methyltransferase ability of the bHLH factor Neurog2 to activate Tlx3. Additional gain of function experiments in the chick þ þ Spinal cord neural tube confirm that Prdm13 suppresses Tlx3 /glutamatergic and induces Pax2 /GABAergic Retina neuronal fate. Thus, Prdm13 is a novel crucial component of the Ptf1a regulatory pathway that, by GABAergic neuron modulating the transcriptional activity of bHLH factors such as Neurog2, controls the balance between Glutamatergic neuron GABAergic and glutamatergic neuronal fate in the dorsal and caudal part of the vertebrate neural tube. & 2013 Elsevier Inc. All rights reserved.

Introduction processing of somatosensory inputs from the periphery. Six dis- tinct classes of dorsal interneurons (early dI1-6 and late dILA The dorsal spinal cord is an integrative center that transmits and dILB) arise in the dorsal spinal cord from six progenitor and processes diverse somatosensory information. These neurons domains (dP1-6). Among them, early-born dI4 and dI6 and late- can be grouped into excitatory neurons and inhibitory neurons, born dILA neurons are GABA/glycinergic neurons (Helms and which use glutamate and GABA/glycine as transmitters respec- Johnson, 2003; Hori and Hoshino, 2012). These three classes of tively. These inhibitory and excitatory neurons are interconnected postmitotic interneurons express the homeodomain transcription and a correct balance between them is essential for correct factors Lbx1, Pax2 and Lhx1/5 that regulate GABA/glycinergic differentiation (Pillai et al., 2007; Batista and Lewis, 2008; Bröhl et al., 2008). n Correspondence to: Laboratoire de Génétique du Développement, Université Libre While postmitotic dorsal interneurons can be distinguished de Bruxelles, Institut de Biologie et de Médecine Moléculaires (IBMM), Rue des Profs. Jeener et Brachet 12, B-6041 Gosselies, Belgium. Fax: þ32 2 6509733. by the combinatorial expression of homeodomain transcription E-mail address: [email protected] (E.J. Bellefroid). factors, bHLH factors have a crucial role in progenitors in the

0012-1606/$ - see front matter & 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ydbio.2013.12.024 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 341 specification of excitatory versus inhibitory neurons. Ptf1a (pan- Recent reports have shown that some vertebrate Prdm family creas 1a), a bHLH transcriptional activator members play important roles as cell fate regulators in the initially identified as a cell fate determinant of the pancreas developing nervous system. For example, Prdm1/Blimp1 specifies (Krapp et al., 1998; Kawaguchi et al., 2002; Sellick et al., 2004; photoreceptor over bipolar neurons in the mouse and zebrafish Afelik et al., 2006) that is restricted to dI4 and DILA interneuron retina (Brzezinski et al., 2010; Katoh et al., 2010). It also controls progenitors in the dorsal spinal cord, plays an essential role in the cell fate decision at the neural plate border in zebrafish embryos specification of GABAergic and glycinergic inhibitory neurons (Roy and Ng, 2004; Hernandez-Lagunas et al., 2011). Prdm8 while suppressing glutamatergic excitatory neurons. Mice lacking regulates target in postmitotic cortical neurons Ptf1a show a near complete loss of dI4 and dILA GABAergic by forming a repressor complex with the transcription factor interneurons and an increase of dI5 and dILB excitatory interneur- Bhlhb5 (Ross et al., 2012). Hamlet, the Drosophila Prdm3/Evi1 and ons. Conversely, overexpression of Ptf1a in the chick neural tube Prdm16/Mel1 homolog controls olfactory neuron diversi- induces GABAergic and glycinergic neurons at the expense of fication by modulating Notch signaling (Endo et al., 2011). Tlx3-positive glutamatergic excitatory dI5 and dILB neurons Here, we identify the HMTase Prdm13 as a likely direct target of (Glasgow et al., 2005; Hori et al., 2008; Huang et al., 2008). The the Ptf1a–Rbp-j complex. We show that Prdm13 negatively reg- requirement for Ptf1a in the generation of GABAergic and glyci- ulates its own expression and that it plays an important role nergic inhibitory neurons has been also demonstrated in other downstream of Ptf1a in the suppression of glutamatergic and regions of the developing CNS where it is expressed, including the induction of GABAergic differentiation. We also provide evidence cerebellum, cochlear nucleus and the retina (Hoshino et al., 2005a, suggesting that Prdm13 represses glutamatergic neuronal fate by 2005b; Fujitani et al., 2006; Nakhai et al., 2007; Pascual et al., blocking Neurog2 ability to induce the glutamatergic-neuron 2007; Dullin et al., 2007; Fujiyama et al., 2009). Ptf1a has been specifier Tlx3 gene. shown to be required for Pax2, Lhx1 and Lhx5 expression and for the suppression of glutamatergic cell fate determinants Tlx3 (also known as Rnx and Hox11L2) and Lmx1b (Cheng et al., 2004, 2005; Materials and methods Hori et al., 2008). The molecular mechanism by which Ptf1a activates one lineage specific gene expression and represses Expression constructs expression of genes of the other alternate lineage remains unclear. Ptf1a is a component of a heterotrimeric transcription complex, The mPrdm13 expression vector used in frog microinjection called PTF1-J, that includes Ptf1a, a commonly expressed bHLH E- experiments was obtained by PCR amplifying from EST BY731300, and a member of the mammalian Suppressor of Hairless clone E130110K18, using primers 50- GAATCGCACGGAACTTCTCG- family, Rbp-j (or its paralogue, Rbp-l, to form PTF1-L in the pancreas) CACTTCG-30 and 50-AGAATGCTTTCAGAATGAGAGGCTATGCCA- that mediates Notch signaling (Beres et al., 2006; Masui et al., 2007). GAGTCT-30 a 2359 bp fragment containing an open reading This PTF1-J transcription complex is required in vivo for the specifica- frame starting from the second methionine (position 49 in tion of GABAergic neurons in a Notch signaling independent manner supplementary material Fig. S1, corresponding to the first methio- (Hori et al., 2008). The PTF1-J complex binds a bipartite DNA nine in the other species) into the EcoRI and XhoI sites of the sequence composed of an E-box for bHLH heterodimer binding and pCS2-Flag vector. The mPrdm13 expression construct used in chick a TC-box for Rbp-j binding. Only a few direct targets of the PTF1-J electroporation was generated by PCR amplifying from the same complex in the dorsal spinal cord have been identified so far. These EST using primers 50-CTCGAGCCCGAGCGTCTGCCTGGC-30 and 50- include Nephrin and Neph3 genes encoding immunoglobulin type GAATTCCTGGTCCCTGCTCAGTCCCT-30 a 2190 bp fragment contain- transmembrane , whose function in the development of ing an open reading frame starting from the second methionine Ptf1a expressing GABAergic interneurons is unknown (Nishida et al., into the XhoI and EcoRI sites of the pCIG vector. 2010) and Ptf1a itself (Meredith et al., 2009; Masui et al., 2008). The MT-Ptf1apCS2þ, MT-Ptf1aW224ApCS2þ, MT-Ptf1aW242ApCS2þ atonal related bHLH proneural gene Neurog2,whichisbroadly and MT-Ptf1aW224A/W242ApCS2þ were generated by PCR amplifi- expressed in dorsal interneurons, is another direct downstream cation using the following primers: forward 50-CAGAATTCCATG- target of PTF1-J (Henke et al., 2009). Embryos null for both Neurog2 GAAACGGTCC-30 and reverse 50-CCCTCTAGATCACATATCAAGGCAC- and another member of the bHLH family, Ascl1 (Mash1), coexpressed 30. The PCR fragments were inserted into the EcoRI and XbaI sites with Neurog2 in dP3-dP5 progenitors, have an apparent loss of dI4 of the MTpCS2þ vector. neurons. This phenotype is however not observed in the single Ptf1aW224A–GRpCS2þ,Ptf1aW242A–GRpCS2þ and Ptf1aW224A/ mutant embryos (Helms et al., 2005)suggestingthatNeurog2hasa W242A–GRpCS2þ were generated using the “QuickChange XL Site redundant function with Ascl1 in dP4. As Ascl1 is not affected in Ptf1a directed Mutagenesis” kit (Stratagene) to introduce mutations mutants, this suggests that additional targets of Ptf1a need to be that lead to an exchange of tryptophan to alanine at position 224 identified to understand the dramatic loss of the dI4 cells observed in and/or 242 using the following primers: C1_up GGACATTCTCTCT- Ptf1a mutants (Henke et al., 2009). CAGCGACTGATGAGAAGCAACTGAG, C1_down CTCAGTTGCTTCTCAT- In an in situ hybridization expression survey of the expression CAGTCGCTGAGAGAGAATGTCC, C2_up GTTGTCAGAACGGCCAAAGTG- of PR domain containing genes in the Xenopus embryo, the GCGACTCCTGAGGATCC, C2_down GGATCCTCAGGAGTCGCCAC- expression profile of the Prdm13 gene was found to overlap with TTTGGCCGTTCTGACAAC. Rbp-J–HApCS2þ was generated by PCR that of Ptf1a in the developing spinal cord and retina. Prdm13 amplification using the following primers: forward 50-CTGGATC- encodes a protein containing a N-terminal PR domain related to CATGCAACCTGGC-30 and reverse 50-CAACTCGAGGGACACTACT- the SET domain found in a large group of histone methyltrans- GCTGC-30. The PCR fragment was inserted into the BamHI and XhoI ferases, followed by four zinc finger repeats (Huang et al., 1998; sites of the HApCS2þ vector. Ptf1a/bHLH-Neurog2-GRpCS2 was Albert and Helin, 2010). Prdm13 belongs to the Prdm gene family generated by first amplifying the bHLH domain of Neurog2 (GR- consisting of 16 orthologs in rodents, birds and amphibians that Neurog2p30, Perron et al., 1999) with the following primers, which encodes important regulators in differentiation and disease. Prdm contained the surrounding sequences of the Ptf1a bHLH domain family members recognize specific DNA-sequences or act as (in italics): Ptf1a/bHLH-Neurog2_for CTGAGGTCGGACGCGGAGATG- non-DNA binding cofactors. They either function as direct histone CAGCAGCGGCGCGTTAAAGCTAACAAC and Ptf1a/bHLH-Neurog2_rev methyltransferases (HMTases) or recruit histone-modifying enzyme GCGGCAGATCGGACTGTACCATCTCGCTAAGAGCCCAGATGTAGTTGTAG. to target promoters (Fog et al., 2011; Hohenauer and Moore, 2012). The generated megaprimer were used in a second PCR reaction 342 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 with the “QuickChange XL Site directed Mutagenesis” kit (Strata- mRNA (50 pg/blastomere) to reveal the manipulated side. For gene) and Ptf1a-GRpCS2þ (Afelik et al., 2006). animal cap assays, synthesized mRNA was microinjected into the animal region of each blastomere of four-cell stage embryos. Histone methyltransferase (HMTase) assay Animal caps were dissected at blastula stage (st.9) and cultured in 1X Steinberg medium, 0.1% BSA until sibling embryos reach the The HMTase assay was conducted as reported (Eom et al., desired stage. To block protein translation, embryos and animal 2009), with some minor modifications. Briefly, Flag-Prdm13 was caps were pretreated at stage 12.5 with 10 mg/ml cycloheximide overexpressed in HEK 293T cells. 24 h post-transdection, total (CHX) for 1 h prior to Dex addition (10 mM). The expression of the protein extracts were prepared in IPH buffer (50 mM Tris/HCl pH different markers was analyzed 24 h later at stages 15–16. 8, 150 mM NaCl, 0.5% NP-40, 1 mM DTT) and precleared with protein-A sepharose beads (GE healthcare) for 1 h at 4 1C. Subse- In situ hybridization and immunohistochemistry quently, precleared cell extracts were incubated overnight with Flag antibody and for 2 h at 4 1C with protein-A sepharose beads. Whole-mount in situ hybridization analysis of Xenopus After three washes with IPH buffer containing 500 mM NaCl, the embryos was performed as described using digoxigenin- or beads were incubated in a 35 ml of reaction mixture containing fluorescein-labeled antisense probes (Sive et al., 2000) generated 3 mg of chicken core histones (Millipore) as a substrate for enzyme as indicated: pBSK(-)-xPrdm13 (EST BJ064589, NotI, T7), pCMV- activity, 1.1 mCi of S-adenosyl-[methyl-3 H]-L-methionine (Perkin SPORT6-xPrdm8 (EST BC136202, SalI, T7), pCMV-SPORT6-Dbx1 (EST Elmer) in HMTase assay buffer (50 mM Tris, pH 8.0, 1 mM EDTA, CF287983, SalI, T7), pBKS-Sox3 (EcoRI, T7). The following con- 50 mM NaCl, 1 mM MgCl2, 1 mM DTT), containing protease inhi- structs were previously described: Ptf1a, Gad1 (Dullin et al., 2007), bitors (complete EDTA free from Roche). After 2 h of incubation at Mafb (Ishibashi and Yasuda, 2001), N-tubulin (Chitnis et al., 1995), 30 1C under agitation, proteins were filtered using p81 paper Eng2 (Hemmati-Brivanlou et al., 1991), Krox-20 (Nieto et al., 1991), (Phosphocellulose Squares) (Millipore), and washed three times Neurog2 (Ma et al., 1996), Pax2 (Heller and Brändli, 1997), Tlx3/ with 10% TCA and 100% ethanol for 5 min in RT. The filters were Hox11L2 (Patterson and Krieg, 1999), TH (Parlier et al., 2008). For allowed to air dry. Five ml of OptiPhase ‘Hisafe’ 2 (PerkinElmer) sections, embryos after completion of the whole-mount procedure was then added to the filters. Quantification was done on a LS were gelatine-embedded and vibratome-sectioned at 30 μm thick- 6500 liquid scintillation counter (Beckman). ness. Double fluorescent in situ mRNA hybridization was per- formed on 12 μm cryostat section as previously described (Lea In vitro co-immunoprecipitation et al., 2012). For EdU experiments, stage 42 tadpoles were incubated 3 h in a 1 mM EdU solution, then immediately fixed in For the co-immunoprecipitation, Rbp-J-HApCS2 þ ,MT- 4% paraformaldehyde. In situ hybridization on 12 μm cryostat Ptf1apCS2þ, MT-Ptf1aW224ApCS2þ, MT-Ptf1aW242ApCS2þ and section was performed followed by EdU revelation using Click-iT MT-Ptf1aW224A/W242ApCS2þ were in vitro translated and radio- EdU Alexa Fluor 488 or 594 Imaging Kit (Molecular Probes). 35 actively labeled with S methionine using the Promega TNT For section in situ hybridization of mouse embryos, 20 mm Coupled Reticulocyte lysate systems. Immunopellets were pre- cryostat sections of 4% paraformaldehyde fixed, 30% sucrose/PBS- pared by incubating γ-bind sepharose beads (Amersham) with infused tissue frozen in gelatin (7,5% gelatin, 15% sucrose/PBS) anti-HA.11 (Covance) in NET-2 buffer (50 mM Tris–HCl, pH 7.4, were used. In situ hybridizations were performed as previously 150 mM NaCl, 0.05% NP-40) for 2 h at 4 1C and subsequent described using digoxigenin-labeled antisense Prdm13 (Kinameri washing for four times with 500 μl NET-2 buffer. The immunopel- et al., 2008), Pax2 (Dressler et al., 1990) and Neph3 (Nishida et al., lets were incubated with 2 μl of the in vitro synthesized proteins 2010) riboprobes as described (Wilkinson and Nieto, 1993). for 1 h at 4 1C and finally washed four times with 700 μl NET-2 For immunofluorescent labeling, fixed electroporated chicken buffer. After boiling the immunopellets for 5 min in 2 SDS embryos were cryoprotected with 30% (w/v) sucrose in PBS, loading buffer, the proteins were separated by 12% SDS-PAGE. embedded in gelatin (7,5% gelatin, 15% sucrose/PBS) and frozen for Images were acquired using a phosphoimager (Typhoon 9400, cryo-sectionning (16 mm thickness). Immunolabeling was performed Amersham Biosciences). as described (Francius and Clotman, 2010). The antibodies used were Pax2 (rabbit, 1:500, ab23345, Abcam), Tlx3, Lmx1b (rabbit, 1:5–10,000 Xenopus embryo culture, micro-injections and animal cap dissections or guinea pig, 1:10–20,000, gift from T.Müller) and Lbx1 (rabbit, 1:10,000,giftfromT.Muller).Fixedanimalcapswereprocessedwith Xenopus embryos were obtained from adult frogs by hormone an anti-Flag (mouse, Sigma F3165, 1:1000). In situ hybridization images induced egg-laying and in vitro fertilization using standard meth- were acquired with an Olympus SZX16 stereomicroscope and Olym- ods (Sive et al., 2000) and staged according to Nieuwkoop and pus XC50 camera, using the Imaging software Cell-Fn.Immunofluor- Faber (1967). Synthetic mRNAs were made using Sp6 mMESSAGE escence images were acquired on a Zeiss Axio Imager 2 Fluorescence mMACHINE. (Ambion). Flag-mPrdm13, Ptf1a/bHLH-Neurog2-GR microscope or a Zeiss LSM710 confocal microscope, using the Imaging and Ptf1aW224A/W242A-GR were linearized with NotI and tran- software Axio visio and Zen2010 respectively. scribed with Sp6. Template described previously include: MT- Neurog2 (Ma et al., 1996), Neurog2-GR (Perron et al., 1999), Ascl1 RT-PCR analysis (Parlier et al., 2008), NeuroD (Lee et al., 1995), Ptf1a-GR (Afelik et al., 2006) and Ptf1a-VP16 (Dullin et al., 2007), Su(H) (Wettstein Total RNA was extracted using the RNAspin Mini RNA isolation et al., 1997) and nLacZ (Chitnis et al., 1995). The Prdm13 antisense kit (GE Healthcare). cDNA was synthesized with iScript cDNA morpholino (50-ACCCCCAGTAGTTGGCAGTGACAGA-30), Prdm13 synthesis kit (Biorad), and tested by PCR amplification using histone antisense mismatch morpholino (50-ACCCCgAcTAcTTGcCAcTGA- H4 primers (27 cycles). Real time RT-PCR was performed using the CAGA-30) and control morpholino oligonucleotides were obtained Step One Plus Real Time PCR system (Applied biosystems) with from Genetools. Ptf1a and Neurog2 antisense morpholinos were as MESA GREEN qPCR kits for SYBRs Assay (Eurogentec). Xenopus previously described (Dullin et al., 2007; Klisch et al., 2006). For GAPDH was used as reference gene and all conditions were in situ analysis, embryos were injected in one cell of two to four- compared to non-injected controls. The following primers were cell stage embryos, and fixed at neurula, tailbud or early tadpole used: Prdm13 (forward: 50-CTGCCGACACATGATGAAAAAGG-30 and stage. In all experiments, embryos were coinjected with LacZ reverse: 50-AGATTTTGGGGGAGGCAGAAAAG-30); Pax2 (forward: J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 343

50-CAGTCAGCACGCCTGGGCAT-30 and reverse: 50-TGCCTCCAGTTGC- and four zinc finger motifs, one located immediately 30 to the PR TGCTGAGT-30); Neurog2 (forward 50-GGCGCGTTAAAGCTAACAAC-30 domain and the others clustered at the C-terminal end of the and reverse 50TTCGCTAAGAGCCCAGATGT-30); Gad1a (forward: 50- protein. Similarity between the identified Xenopus laevis, human, ATGGGCGTCTTACTCCAATG-30 and reverse: 50-ATGTCTACATGGC- mouse and zebrafish Prdm13 proteins is mainly limited to the PR GACCACA-30); Tlx3/Hox11L2 (forward: 50 GCCAACAAGTACAAGTGCA- and zinc finger domains (Supplementary material Fig. S1A,B). The CAG-30 and reverse: 50 CAGGAGCCAGACTCACATTGAC-30); Myt1 Xenopus laevis Prdm13 gene shows an exon/intron structure (forward: 50-TTGGGATGGTCCCATAGACT-30 and reverse 50-TCTTGCA- identical to its mammalian and lower vertebrate homologs (4 TCTCCTGCATCTC-30); Vglut1 (forward: 50-TCAGTGATTTGCTTGT- exons, with the PR domain encoded by exons 1–3 and zinc fingers TACGTGTAT-30 and reverse: 50-CATCCTCATGGATATTTTCTACACC- 1–4 by exon 4), which further supports its orthologous relation to 30); Lmx1b (forward: 50-ATCGCTGCTTCACCAGGGGC-30 and reverse: the other vertebrate Prdm13 genes. 50-GGCAGTCTGACCCGAGCAGC-30); Lbx1 (forward: 50-GCAGTGCT- As no Xenopus cDNA clone containing the full Prdm13 open CACCATCTC-30 and reverse 50-CCACATCCTCCTGATCC-30); Histone reading frame is available, we generated a plasmid encoding a H4 (50-CGGGATAACATTCAGGGTATCACT-30 and reverse 50-ATC- Flag-tagged mouse Prdm13 protein to determine the subcellular CATGGCGGTAACTGTCTTCCT-30)andGAPDH (forward 50-TAGTTG- localization of the protein. mPrdm13 mRNA was injected into the GCGTGAACCATGAG-30 and reverse 50-GCCAAAGTTGTCGTTGATGA- animal pole of fertilized Xenopus embryos at the 2–4 cell-stage. As 30). All measurements were done in duplicates or triplicates. Error expected, the overproduced protein was detected in the nucleus of bars represent standard deviation. The values in the figures are the ectodermal cells (Supplementary material Fig. S1C). To deter- representative cases of at least two independent experiments. mine whether Prdm13 has histone methyltransferase (HMT) activity, we overproduced the protein into HEK293 cells. Prdm16 that has HMT activity (Pinheiro et al., 2012) was used as a positive In ovo electroporation control. An empty-Flag vector was used to control for non-specific binding. Flag proteins were immunoprecipitated using anti-Flag Fertilized chicken eggs were incubated at 38.5 1Cinahumidified antibodies and enzymatic HMT activity was measured with the incubator and staged according to Hamburger and Hamilton immunoprecipitated proteins using core histones and [3H]-SAM as (Hamburger and Hamilton, 1992). Embryos were electroporated at the substrate and methyl donor. As shown in Fig. 1A, Prdm13 stage HH12-14. The expression constructs (1 μg/μlinendotoxinfree increases the incorporation of [3H]-SAM on core histones, as water) dyed with fast green 5X were injected into the lumen of the observed with Prdm16. Thus, as other members of the family, spinal cord and electroporated using a BTX square-wave electro- Prdm13 possess HMT activity. porator and gold electrodes (5 pulses of 20–25 V for 50 ms each). Embryos were collected 48 h (HH24-25) or 72 h (HH27-28) after electroporation and fixed in PBS/4%PFA for 30 min or 1 h, respec- Prdm13 is expressed in progenitors and post-mitotic cells of the tivelyandthenprocessedforimmunohistochemistry as above. Xenopus dorsal caudal neural tube and retina

To investigate the Prdm13 temporal expression pattern during early Xenopus embryogenesis, semi-quantitative RT-PCR was used. Results As shown in Fig. 1B, maternal Prdm13 transcripts are detected only at low levels. Zygotic Prdm13 transcripts are detected starting from Identification of Xenopus laevis HMTase Prdm13 stages 12 to 13, increase until neurula stages and persist later in tailbud and tadpole stages. The spatial expression of Prdm13 was In a previous work, we identified Prdm3 (Evi1) and Prdm16 analyzed by whole-mount in situ hybridization on embryos at (Mel1) as zinc finger genes with localized expression in the different stages. Prdm13 transcripts were first detected at stage 13 developing embryo, including the nervous system (Van in two bilateral longitudinal stripes running along the posterior Campenhout et al., 2006). was used to neural plate (Fig. 1C), which were within the sensorial layer of the identify additional Xenopus laevis orthologs of human and mouse neuroectoderm (data not shown). This pattern of expression is Prdm genes. Through BLAST searches of the National Center for maintained throughout neural fold stages of development. After Biology Information (NCBI, http://www.ncbi.nlm.nih.gov/) Xenopus neural tube closure, Prdm13 persists in the dorsal part of the laevis EST database, we identified four additional members of the developing posterior hindbrain and spinal cord (Figs. 1D,E). Prdm family, Prdm8, Prdm12, Prdm13 and Prdm14, with restricted Expression of Prdm13 was also detectable from stage 25 more expression in the nervous system. Here we focus on one of them, anteriorly in the hindbrain (Fig. 1F, arrowhead), and from stage 28 Prdm13, during Xenopus development. The detailed expression and in the developing retina (Fig. 1G). On transversal sections of stage functional analysis of some of the other Prdm orthologs identified 32 embryos at the level of the anterior and posterior hindbrain, will be reported elsewhere. The Prdm13 cDNA identified (EST Prdm13 is expressed in the ventricular zone of the dorsal spinal BJ064589) was sequenced. The resulting sequence was analyzed cord but is highest in a more restricted progenitor domain at the to localize the open reading frame and deduced peptide, and the ventricular lateral margin (Figs. 1H and I). At stage 42, strong annotated sequence was submitted to Genbank (GenBank BankIt Prdm13 expression becomes restricted to the hindbrain (Fig. 1J). submission ID: 1621537). As the identified cDNA clone did not At that stage, in the mature retina, Prdm13 is detected in the inner contain the entire open reading frame, we searched the initial part of the inner nuclear layer where GABAergic amacrine cells draft assembly of the Xenopus laevis genome and identified a reside (Supplementary material Fig. S2A,B,D,E)(Dullin et al., 2007). scaffold (32494 in genome draft 6.0) containing Prdm13. The To characterize Prdm13 relative to cell proliferation, we used EdU sequence obtained predicts a 633 amino acid protein with 59% incorporation to detect cells in S phase. In stage 42 embryos identity to human Prdm13 (accession ♯ NM_021620), 58% to exposed to EdU 3 h before analysis, a subset of Prdm13 labeling in mouse Prdm13 (accession ♯ NM_001080771), 64% to zebrafish the neural tube was found overlapping the EdU þ staining (Fig.1K). Prdm13 (accession ♯ XM_001336555) and 92% to the Xenopus In the retina, Prdm13 was also detected in some EdUþ cells in the tropicalis Prdm13 predicted protein (accession XM_002932160). ciliary marginal zone (CMZ) (Supplementary material Fig. S2C). The Xenopus laevis Prdm13 protein has a similar structure to its These data suggest that in the frog neural tube and retina, Prdm13 homologs in the other species, including a N-terminal PR domain is expressed in progenitors and post-mitotic differentiating cells. 344 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357

Fig. 1. HMTase activity and developmental expression of Prdm13. (A) mPrdm13 is a histone methyl-transferase. Core histones were used as substrates in a HMTase assay with a immunopurified Flag-Prdm13 protein produced in 293T cells. Flag-Prdm16 was used as a positive control. Empty vector transfected cells were used to control for nonspecific background. The results represent the average of triplicate samples plus SD. The amount of immunoprecipitated Prdm16 and Prdm13 proteins used in the assay is shown on the western on the right (arrows). Note that mPrdm13 is expressed at a much lower level than mPrdm16. On the autoradiograph shown, multiple bands are seen for mPrdm16 due to the high exposure time used to detect mPrdm13. Those bands likely correspond to mPrdm16 degradation products. A non-specific IgG band is indicated by an arrowhead. (B) Temporal embryonic expression of Xenopus laevis Prdm13 as determined by RT-PCR. Histone H4 is shown as a loading control. (C–K) Whole-mount in situ analysis of Prdm13 expression during embryogenesis. Panel E, transversal section of the neural tube of a stage 22 embryo. Panels H and I, transversal sections of the neural tube of a stage 32 embryo at the levels indicated in G. In panel K, transversal section of the neural tube of an embryo that has been additionally stained for proliferation by EdU incorporation. The enlargement in the inset shows that some Prdm13 are EdU positive. Nieuwkoop Faber stages are indicated. All embryos shown are with anterior to the right. (C, D) Dorsal views. (F, G, J) Lateral views. J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 345

To map the anterior and dorso-ventral limits of Prdm13 zone where it promotes inhibitory neuron differentiation. Its expression in neurula and early tailbud stage embryos, we downstream direct target Neurog2 which is broadly expressed in compared Prdm13 expression to that of other regionalized neural the spinal cord is enriched in the lateral regions of the ventricular markers using double in situ hybridization. For the anteroposterior zone (Helms et al., 2005; Henke et al., 2009). Pax2, which is axis, we used -2 (Eng2), expressed specifically at the required for Ptf1a to specify an inhibitory neuronal cell fate, is midbrain–hindbrain boundary (Hemmati-Brivanlou et al., 1991), expressed in dI4/ dILA as well as in dI6 dorsal interneurons and in Krox-20 expressed specifically within rhombomeres 3 and 5 of the ventral interneurons (Huang et al., 2008). To determine if those developing hindbrain (Bradley et al., 1993) and Mafb expressed in Ptf1a downstream targets are expressed in a similar manner in the rhombomeres 5 and 6 (Sturgeon et al., 2011). We found that Xenopus caudal neural tube and determine whether Prdm13 over- Prdm13 expression lies caudally to the En2 and Krox20 domains of laps with them, we compared by whole-mount in situ hybridiza- expression, with a gap of the size of about one rhombomere tion their expression. On transversal sections of the hindbrain in between the anterior limit of Prdm13 and Krox20 within rhombo- stage 32 embryos, as in mouse, Ptf1a is detected with Prdm13, mere 5 (Supplementary material Fig. S3A,B). Double in situ with Neurog2 and Pax2 at similar D/V levels in the dorsal part of the Mafb showed that Prdm13 expression terminates at the domain of neural tube (Fig. 2J–M). While Ptf1a appears expressed in the Mafb expression and thus has an anterior limit at rhombomere 7 lateral part of the ventricular zone and in the intermediate zone, (Supplementary material Fig. S3C). Prdm13 strongest expression, like that of Neurog2, is detected To determine the position of the stripe of Prdm13 within the slightly more laterally in the intermediate zone, where cells start neural plate, we compared its expression with that of N-tubulin, to differentiate. Pax2 is detected in the marginal zone containing Dbx1 and Prdm8 by double in situ hybridization. N-tubulin is a post-mitotic cells. These results suggest that Prdm13 may act neuronal marker expressed in the three longitudinal stripes downstream of Ptf1a, at about the same time as Neurog2, and marking the positions of the differentiated primary neurons (Ma before Pax2 during neuronal differentiation. et al., 1996). Dbx1 is expressed in p0 progenitors and marks the ventral limit of Prdm13 expression in the mouse spinal cord Ptf1a regulates Prdm13 expression in the spinal cord (Kinameri et al., 2008). In the frog, the homolog of Dbx1 is detected in between the medial and intermediate stripes of primary To address a possible role of Ptf1a in the regulation of Prdm13 neurons within the Xenopus neural plate (Gershon et al., (2000)). expression within the developing neural tube, we first used gain of Prdm8 marks ventral interneurons and motor neurons in the function experiments. mRNAs encoding glucocorticoid inducible mouse spinal cord (Komai et al., 2009). Double labeling of neurula forms of Ptf1a (Ptf1a-GR) or of Ptf1a-VP16-GR encoding a Ptf1a embryos with Prdm13 and N-tubulin revealed that Prdm13 is variant where the strong VP16 transactivation domain is fused to expressed in the intermediate stripe where differentiating primary the carboxyl-terminus of Ptf1a were injected into one blastomere interneurons are located (Supplementary material Fig. S3D). Dou- of two-cell stage embryos together with LacZ mRNA to localize the ble in situ with Prdm8 showed that Prdm13 expression is detected distribution of the injected mRNA. Dexamethasone (DEX) was more laterally within the neural plate than Prdm8 (Supplementary added at stage 12.5. Embryos were analyzed by whole mount material Fig. S3E). Double labeling with Dbx1 showed that Prdm13 in situ hybridization for Prdm13 at neurula and tadpole stages. The is expressed at the lateral boundary of the Dbx1 domain. Later, GABAergic marker Pax2, a known downstream target of Ptf1a in within the neural tube, Prdm13 is detected in the dorsal half, above the mouse (Glasgow et al., 2005) was used as a control. As shown the Dbx1 expression domain (Supplementary material Fig. S3F–H). in Fig. 3, Ptf1a-GR overexpression increases Prdm13 expression in Thus, similar to mouse and zebrafish Prdm13 (Kinameri et al., neurula and tadpole stage embryos. At neurula stage, Prdm13 2008; Sun et al., 2008), Xenopus Prdm13 is only expressed in the expression was increased within its normal expression domain. caudal neural tube where it is restricted to the dorsal region. Ectopic expression was also detected within the ectoderm, the strongest staining being observed in the anterior part of the neural plate (98%, n¼40). At tadpole stage, ectopic Prdm13 staining was Prdm13 expression is related to that of Ptf1a in the dorsal neural tube still detected in the ectoderm (95%, n¼42). Pax2 expression was as and retina expected also upregulated by Ptf1a overexpression at both neurula (90%, n¼41) and tadpole (100%, n¼28) stages (Fig. 3A–D). An even Comparing Prdm13 expression to that of the different known more dramatic increase of Prdm13 and Pax2 was observed upon bHLH transcription factors controlling neuronal specification, we injection of Ptf1a-VP16 mRNA in both neurula (Prdm13: 91%, found that the Prdm13 expression profile we identified in the early n¼45; Pax2: 100%, n¼27) and tadpole stage embryos (Prdm13: frog embryo is strikingly related to that of Ptf1a (Afelik et al., 100%, n¼48; Pax2: 100%, n¼16). In those Ptf1a-VP16 injected 2006). Indeed, as shown in Fig. 2A–F, at around stage 18, Ptf1a and tadpoles, Prdm13 upregulation was only detected in the dorsal Prdm13 are both detected along two parallel longitudinal stripes part of the neural tube. Strikingly, Prdm13 expression was also representing the caudal neural folds. At tailbud stage, both Ptf1a observed in some embryos in somitic tissue, which was not the and Prdm13 become expressed in the anterior hindbrain. At that case for Pax2 (Fig. 3E–H). Although much less dramatic, an stage, expression of Ptf1a become less detectable in the caudal upregulation of Prdm13, was also observed in embryos overex- neural tube while Prdm13 remains high. At tadpole stage, Prdm13 pressing Ascl1 (Xash1) that, like Ptf1a, induces GABAergic neurons is detected in the retina and the anterior hindbrain like Ptf1a but (Mazurier et al., unpublished)(Supplementary material Fig. S4). not in the pancreas, the other major site of Ptf1a expression (data Because Neurog2 is a direct target of Ptf1a (Henke et al., 2009)and not shown). To determine whether Prdm13 expression overlaps that it is coexpressed with Prdm13 in the dorsal spinal cord of frog with that of Ptf1a, we performed fluorescent double in situ embryos (Fig. 2K,L), we also examined Prdm13 expression in response hybridization on cross-sections of tadpole stage embryos. to increase expression of Neurog2 as above using an inducible form of Fig. 2G–I shows that in the spinal cord, the expression of Prdm13 the protein (Neurog2-GR). Prdm13 expression was mainly unaffected overlaps with that of Ptf1a. Within the retina, Ptf1a and Prdm13 are by overexpression of Neurog2 in embryos at neurula (all unaffected, also coexpressed in the central ciliary marginal zone (CMZ) n¼29) or tadpole stages (95% unaffected, n¼55). Pax2 was also (Supplementary material Fig. S2D, F–I). mainly unaffected by Neurog2 overexpression (all unaffected, n¼26 In the mouse dorsal spinal cord, the expression of Ptf1a is at neurula and 95% unaffected, n¼44 at tadpole stage) (Fig. 3I–L). In restricted to the dP4/dILA progenitor domain of the ventricular contrast, the pan-neuronal marker N-tubulin and the glutamatergic 346 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357

Fig. 2. Prdm13 embryonic expression profile is highly related to that of Ptf1a and overlaps with those of its downstream targets Neurog2 and Pax2 in Xenopus embryos. (A–F) Dorsal or lateral views of Prdm13 and Ptf1a in stage 18, 27 and 33 embryos showing that the two genes have a similar restricted expression in the neural tube. (G–I) Images of transversal sections of the neural tube at the level of the anterior spinal cord of a stage 32 embryo, following double fluorescent in situ hybridization for Prdm13 and Ptf1a. Note the overlap of the two stainings (arrows). (J–M) Transversal sections of the neural tube of stage 32 embryos at the level of the otic vesicles stained with the indicated probes. Note that Ptf1a appears coexpressed with Prdm13, Neurog2 and Pax2 in the dorsal neural tube. While Ptf1a is expressed in the lateral part of the ventricular zone and in the intermediate zone, Prdm13 and Neurog2 strongest expression is detected in the intermediate zone and Pax2 in the marginal zone. Nieuwkoop Faber stages are indicated. marker Tlx3/Hox11L2 were, as previously reported (Perron et al., 1999; more prominently induced in caps derived from embryos overexpres- Dullin et al., 2007), strongly increased in those Neurog2 mRNA- sing both Ptf1a and Neurog2 compared to those derived from embryos injected embryos (data not shown). A downregulation of Prdm13 injected with Ptf1a mRNA alone, suggesting an unexpected synergistic was even observed in embryos overexpressing a non-inducible effect (Supplementary material Fig. S5). form of Neurog2 or the downstream proneural gene NeuroD that, like To determine whether Ptf1a and Neurog2 are required for Neurog2, promotes in the Xenopus ectoderm a glutamatergic fate Prdm13 expression, we used antisense morpholinos that have (supplementary material Fig. S4). The regulation of Prdm13 by Ptf1a been shown previously to be efficient and specifictoknockdown and Neurog2 was further investigated in ectodermal explants. These their function (Klisch et al., 2006; Afelik et al., 2006; Dullin et al., explants are fated to become epidermis but can be converted to 2007). Embryos injected with the Ptf1a-MO or Neurog2-MO derivatives of all three germ layers. Embryos were injected bilaterally were analyzed at neurula and tadpole stages for Prdm13 expres- with mRNA at the two cell-stage, and animal caps dissected at blastula sion. Pax2 was also examined at tadpole stage embryos as a stages. Total RNA was analyzed at stage 16 by RT-qPCR. As observed in control for the Ptf1a-MO. A reduction of Prdm13 was detected on embryos, Ptf1a but not Neurog2 strongly induces Prdm13.Wealso the injected side within the neural plate (71% reduced, n¼57) of analyzed the ability of the GABAergic inducer Ptf1a to induce Prdm13 Ptf1a-MO injected embryos. A similar reduction was observed in when co-injected with Neurog2 that only induces in this assay the neural tube of Ptf1a-MO injected tadpole stage embryos glutamatergic markers (Dullin et al., 2007). We found that Prdm13 is (91% reduced, n¼48). Pax2 wasasexpectedstronglyreduced J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 347

Fig. 3. Prdm13 is dramatically affected by Ptf1a but not by Neurog2 gain or loss of function in Xenopus.(A–L) In situ hybridization analysis of Prdm13 expression in Ptf1a-GR, Ptf1a-VP16 or Neurog2-GR injected embryos analyzed at neurula and tadpole stages. Pax2 was used as a control. At neurula stage, embryos are shown in dorsal views, anterior to the bottom. At tadpole stage, transversal sections at the level of the posterior hindbrain or spinal cord are shown. The arrow in G indicates ectopic Prdm13 in the somites. (M–R) In situ hybridization analysis of Prdm13 expression in embryos injected with 20 ng of a Ptf1a-MO or a Neurog2-MO. Pax2 was used as a control for the Ptf1a-MO. Prdm13 expression was analyzed at the open neural plate stage and in the neural tube of tadpole stage embryos. Pax2 was only analyzed at tadpole stage. Note that Prdm13 like Pax2 is reduced in Ptf1a-MO injected embryos and is largely unaffected in Neurog2-MO injected embryos. In all cases, LacZ mRNA was used as a lineage tracer. IS, injected side. following Ptf1a-MO injection (all reduced, n¼ 46) (Fig. 3M–O). n¼49 at tadpole stage). Pax2 was also unaffected (100%, n¼22) In embryos injected with the Neurog2-MO, Prdm13 remains (Fig. 3P–R). The expression of the pan-neuronal marker N- mainly unaffected (85% unaffected and 15% slightly reduced, tubulin was as previously reported (Klisch et al., 2006) inhibited n¼55 at neurula stage and 92% unaffected and 8% reduced, (data not shown). Together, these data indicate that Ptf1a but not 348 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357

Neurog2 is a critical regulator of Prdm13 expression in the dorsal pretreated embryos, n¼45 and 51, respectively) (Fig. 4B). To confirm neural tube. this, we repeated the Ptf1a-GR induction assays in the presence or absence of CHX pretreatment in animal caps. Explants expressing Prdm13 is a likely direct Ptf1a transcriptional target Ptf1a-GR wereincubatedatstage12.5for1hinthepresenceofCHX before DEX treatment and then used for RT-qPCR analysis at the – To determine whether Prdm13 regulation by Ptf1a is likely to be equivalent of stages 15 16. Fig. 4CshowsthatPrdm13 is slightly direct or not, we first measured in animal caps the temporal induction upregulated by CHX treatment (6 fold increase), suggesting that some of Prdm13 induction by Ptf1a-GR compared to that of Pax2,anindirect negative regulators may turn over rapidly and therefore disappear in Ptf1a target. At stages 9–10, animal cap explants were either left the presence of CHX. DEX treatment in the presence of CHX also untreated or exposed to DEX for 1–6 h before harvesting. We observed resulted in a stronger induction of Prdm13, higher than that observed aslightincreaseofPrdm13 expression (4-fold induction) already after in DEX alone treated caps. As in embryos, the Ptf1a-induction of Pax2 2 h of treatment of DEX and a dramatic induction (1300-fold induc- was completely abolished by CHX treatment. These data suggest that tion) later, which is in accordance with the idea that it may be induced Ptf1a directly regulates Prdm13. In accordance with this idea, an directly by Ptf1a. By contrast, significant Pax2 induction was not analysis of a pancreatic cell line chromatin immunoprecipitation detected, even at the latest time points analyzed (Fig. 4A). We next dataset of Ptf1a genomic binding sites (Geo: GSE27295, Thompson asked if Ptf1a-GR could induce Prdm13 expression in the presence of et al., 2012) revealed two strong binding sites 48 kb upstream and cycloheximide (CHX), a protein synthesis inhibitor. Therefore, embryos 70 kb downstream of the Prdm13 start site. Although the downstream were injected unilaterally at the two-cell stage with Ptf1a-GR mRNA, site had a Rbp-J binding site (TC-box), it did not show a perfect Ptf1a – together with LacZ mRNA as a tracer. At stage 12.5, embryos were binding site (E-box) with the correct spacing of 1 3 helical turns culturedfor1hinthepresenceorabsenceofCHX,beforeexposureto (Masui et al., 2007; Nishida et al., 2010). However, the upstream peak DEX until stages 15–16 and then processed by in situ hybridization for had a Ptf1a binding element 1 and 3 helical turns apart from two Rbp- Prdm13 and Pax2 expression. This CHX pretreatment has been shown J (TC-box) binding elements (Supplementary material Fig. S6). to be efficient to reduce protein synthesis (Wettstein et al., 1997). Under the conditions used, this treatment appears to last for a Implication of the bHLH domain of Ptf1a in Prdm13 activation sufficiently long time as it did block the indirect Ptf1a target Pax2 and GABAergic specification (96% induced, n¼28 in non-CHX pretreated embryos; 97% uninduced, n¼41 in CHX pretreated embryos). In contrast, this treatment did not In Xenopus whole embryos and animal cap explants, Ptf1a over- prevent activation of Prdm13 (allinducedinbothuntreatedandCHX expression strongly induces GABAergic neurons while Neurog2

Fig. 4. Prdm13 is a likely direct Ptf1a transcriptional target. (A) Real time RT-PCR analysis of Prdm13 expression in animal cap explants derived from Ptf1a-GR overexpressing embryos either left untreated or exposed to DEX for 1–6 h before harvesting. Pax2, was used as a control. Note that Prdm13 expression is already dramatically induced after 4 h of treatment of DEX. In contrast, Pax2 is not induced at those early time points of induction. CC, control caps. (B) In situ hybridization analysis of Prdm13 and Pax2 in neurula embryos overexpressing Ptf1a-GR, treated with or without CHX for 1 h at stage 12.5 before the addition (or not) of DEX as indicated. LacZ mRNA was used as a lineage tracer. Dorsal views are shown, with anterior to the bottom. IS, injected side. (C) Real time RT-PCR analysis of the expression of Prdm13 and Pax2 in animal caps derived from Ptf1a-GR overexpressing embryos, pretreated or not with CHX before the addition (or not) of DEX as in (B). Note that both in embryos and in animal caps, this CHX treatment does not reduce Prdm13 induction while Pax2 induction is completely abolished. J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 349 induces glutamatergic neurons (Dullin et al., 2007). The importance of Neurog2, indicating that sequences outside the bHLH domain of the bHLH domain of Ptf1a for the GABAergic inducing activity of the of Ptf1a are most important for its specific functional properties. Ptf1-J complex remains unclear. To determine whether the ability of Ptf1a to induce Prdm13 expression is dependent on some specific amino acids of its bHLH domain, we generated a construct encoding a Prdm13 is activated by the Ptf1-J trimeric transcription complex DEX inducible Ptf1a/bHLH-Neurog2-GR chimeric protein where the required for the specification of GABAergic and suppression of Ptf1a bHLH has been replaced with that of Neurog2. The ability of this glutamatergic neurons construct to induce Prdm13 was evaluated in embryos. Gad1 was examined as a control. Ptf1a/bHLH-Neurog2-GR was still able to Ptf1a and Rbp-j interact in a complex through two highly induce Prdm13 (95% increased, n¼45) and Gad1 (100% increased, conserved motifs (C1/C2) located in the C-terminus of Ptf1a. This n¼13), although to a lower extent than wild type Ptf1a-GR (Fig. 5A). complex, Ptf1-J, is required for specification of GABAergic neurons A reduced but still significant level of induction of Prdm13 was also and suppression of glutamatergic expressing cells in the chick observed in animal caps overexpressing Ptf1a/bHLH-Neurog2-GR neural tube (Hori et al., 2008). In this complex, Ptf1a binds with a (Fig. 5B). Thus, the bHLH domain of Ptf1a is interchangeable with that common bHLH protein to an E-box, while Rbp-J binds to a TC box

Fig. 5. Prdm13 is activated by the Ptf1-J transcription complex. (A) In situ hybridization analysis of Prdm13 and Gad1 in embryos injected with Ptf1a-GR, Ptf1a/bHLH-Neurog2- GR or Ptf1aW224A/W242A-GR. Dorsal views, anterior to the bottom, of neurula embryos are shown for Prdm13. Whole head dorsal views and transverse sections at the level of the posterior hindbrain are shown for Gad1. (B) Real time RT-PCR analysis of the expression of Prdm13 in stage 16 animal cap explants isolated from embryos overexpressing Ptf1a-GR, Ptf1a/bHLH-Neurog2-GR or Ptf1aW224A/W242A-GR. Note that Prdm13 is dramatically activated in both assays by Ptf1a and to a lesser extent by Ptf1a/bHLH-Neurog2-GR but not by Ptf1aW224A/W242A-GR, which is also unable to activate Gad1. (C) In situ hybridization analysis of Prdm13 and Neurog2 in embryos injected with Ptf1a, Rbp-J,ora combination of Ptf1a and Rbp-J. Dorsal views, anterior to the bottom, of neurula embryos are shown. (D) Real time RT-PCR analysis of the expression of Prdm13 and Neurog2 in stage 16 animal cap explants isolated from embryos overexpressing Ptf1a, Rbp-J, or a combination of Ptf1a and Rbp-J. Note that Rbp-J does not affect Prdm13 nor Neurog2 expression and that the combination of Ptf1a and Rbp-J induces stronger Prdm13 and Neurog2 expression than that induced by Ptf1a alone. In all cases, LacZ mRNA was used as a lineage tracer to identify the injected side. IS, injected side. 350 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 positioned one or two DNA turns away (Beres et al., 2006; Masui mutant embryos, Prdm13 was strongly reduced in the spinal cord, et al., 2007). To determine whether Prdm13 is induced by the Ptf1-J as observed in Ptf1a-null embryos. The stronger staining observed complex, we first compared the capacity of Ptf1a alone to that of at the lateral border of the ventricular zone of the neural tube was Ptf1a and Rbp-j to induce Prdm13 in embryos and animal caps. in particular clearly reduced in both mutant Ptf1a-null and Neurog2 was also examined as a control. As shown in Fig. 5C,D, Ptf1aW298A mutant embryos (Fig. 6A,D,G). Pax2 expression was lost overexpression of Ptf1a in combination with Rbp-j induces higher in dI4 but maintained in dI6 differentiating neurons in both Prdm13 transcript levels than Ptf1a alone. As anticipated, Rbp-j mutants, as expected (Fig. 6B,E,H). Neph3, another Ptf1a target alone was not able to induce Prdm13. Similar results were obtained (Nishida et al., 2010) was also undetectable in both mutants in animal cap explants. Neurog2 was also more strongly induced by (Fig. 6C,F,I). These results reveal that Prdm13 is a target of the the simultaneous overexpression of Ptf1a and Rbp-j than by either Ptf1-J complex. alone. To determine whether the Rbp-j interaction domain of Ptf1a is required for its ability to induce Prdm13 expression, we next generated a form of Xenopus Ptf1a (Ptf1aW224A/W242A-GR) mutated Prdm13 is required downstream of Ptf1a for the repression of Tlx3 within the C motifs of Ptf1a that disrupts its interaction with Rbp-j, and the induction of Pax2 and negatively regulates its own expression as described previously for mouse Ptf1a (Beres et al., 2006). Immunoprecipitation experiments demonstrate that the intro- The results above suggest that Prdm13 may play a role down- duced mutations disrupt in vitro the interaction between Xenopus stream of Ptf1a. To begin to approach its function during neural Ptf1a and Rbp-j (Supplementary material Fig. S7). Fig. 5A shows development, we first used a morpholino antisense oligonucleotide that, in embryos, overexpression of Ptf1aW224A/W242A-GR does not designed to interfere with Prdm13 translation (Prdm13-MO) to induce Prdm13 (none induced, n¼30). Gad1 was, as expected, also knockdown endogenous Prdm13 in the embryo (Fig.7A–L). Upon unaffected by Ptf1aW224A/W242A-GR (77% non-induced and 23% only injection of 5 ng of Prdm13-MO in one blastomere of 2–4cellstage weakly increased, n¼13). This Ptf1aW224A/W242A-GR mutant was embryos, 90% of the embryos (n¼21) showed a reduction of Pax2 also tested in animal cap assays. Fig. 5B shows that as in embryos, expression in the dorsal part of the neural tube at stage 25–28. Gad1 Ptf1aW224A/W242A-GR is unable to induce Prdm13 in animal caps. was also reduced in Prdm13-MO injected embryos (data not shown). To further validate Prdm13 as a target of the Ptf1-J complex, we By contrast, Tlx3 was found slightly upregulated (75%, n¼28). The also examined its expression in the mouse Ptf1aW298A mutant expression of the pan-neural marker Sox3 was unperturbed (none incapable of interacting with Rbp-j (Hori et al., 2008). In Ptf1aW298A affected n¼20). Injection of a Prdm13 mismatch MO had no such

Fig. 6. Prdm13 is lost in mouse Ptf1acre/cre and Ptf1aW298A mutant embryos. (A–F) Expression of Prdm13, Pax2 and Neph3 analyzed by in situ hybridization on coronal sections of the spinal cord of E11,5 Ptf1acre/cre, Ptf1aW298A or Ptf1a þ / þ embryos. Note that the strong staining observed in the lateral domain of Prdm13 expression is lost in both mutants. J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 351

Fig. 7. Knockdown of Prdm13 reduces Pax2 and an upregulates Tlx3 in the dorsal neural tube of early tadpole embryos and in Ptf1a overexpressing animal caps. (A–H) Dorsal view and (I–L) transversal sections at the level of the posterior hindbrain or anterior spinal cord of Prdm13-MO injected embryos hybridized with the indicated probes. Note that Pax2 is markedly reduced in the dorsal neural tube (arrow) while its more ventral expression domain is unaffected. In contrast, Tlx3 and Prdm13 are upregulated. Sox3 is unperturbed by injection of the Prdm13-MO. Injection of a Prdm13-MOmis has no effect on the different markers tested. (M-R) Real-time RT-PCR analysis of animal cap explants isolated from embryos injected with Ptf1a mRNA and morpholinos as indicated and collected when sibling embryos reached stage 26. Note that the knockdown of Prdm13 inhibits the increase of Pax2 and Gad1 and upregulates Tlx3, Vglut1 and Lmx1b. Such an effect is not seen with a Prdm13-MOmis or a standard control MO (MOctrl). Prdm13 is also induced by Ptf1a. The activation of Prdm13 by Ptf1a (430 fold activation) is not seen on the graph due to the strong activation in the presence of the Prdm13- MO (19,000 fold activation). CC, control caps. effect on the expression of these markers (all unaffected for Pax2 and own expression in a feedback loop. No such effects were observed Sox3, n¼21 and 20 respectively; 92% unaffected, n¼28 for Tlx3). with injection at the same dose of a Prdm13 mismatch MO (all Surprisingly, Prdm13 was strongly upregulated in those embryos (all unaffected, n¼28). We next examined Prdm13 requirement down- increased, n¼28), suggesting that Prdm13 negatively regulates its stream of Ptf1a by simultaneously overexpressing Ptf1a while 352 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 knocking down Prdm13 in animal caps. Explants were treated with Prdm13 suppresses glutamatergic and induces GABAergic markers DEX at stage 9, cultured until stage 26 and analyzed by RT-qPCR. Fig. 7M–R shows that the level of induction of the GABAergic To further demonstrate a role for Prdm13 in the suppression of markers Pax2 and Gad1 was reduced following Prdm13 inhibition. Tlx3 and the induction of Pax2 expression, we misexpressed In contrast, the expression of the glutamatergic markers Tlx3/ Prdm13 in the Xenopus embryos. Injected embryos developed Hox11L2, VGlut1 and Lmx1b was upregulated. As observed in normally until gastrula stage but then failed to complete gastrula- embryos, the level of induction of Prdm13 by Ptf1a was increased tion (Supplementary material Fig. S9B–E). To analyze the conse- upon Prdm13 depletion. Importantly, the downregulation of Pax2 quence of Prdm13 misexpression in the dorsal neural tube and and the upregulation of Tlx3 and Lmx1b could be partially rescued by overcome its effects on early development, we turned to the chick injection of mPrdm13 mRNA, which further validate the observed model that is widely used to overexpress gene in the neural tube phenotype (Supplementary material Fig. S8). Together, these results by in ovo electroporation. Embryos were electroporated at stage indicate that Prdm13 is required downstream of Ptf1a for the HH12–14 with a vector expressing mPrdm13 plus GFP and ana- repression of Tlx3 and the induction of Pax2 and that it negatively lyzed 48 h later (stage HH24–25). Transverse sections of thoracic regulates its own expression. neural tube were examined by immunofluorescence for alterations in neuronal subtype specification relative to the non- electroporated side. We used antibodies against the GABAergic determinants Pax2 and Lbx1 and against the Tlx3 and Lmx1b Prdm13 blocks Neurog2 ability to induce Tlx3 glutamatergic markers. Ectopic expression of Prdm13 resulted in Pax2þ GABAergic neurons in the dI1-3 and dI5 domains where it To further uncover the function of Prdm13, we performed is normally absent (Fig. 9A–C). Lbx1 þ neurons that are generated mPrdm13 gain-of-function assays in the frog. We first asked from dP4–dP6 progenitors were also detected in dI1–3 and vI0–3 whether Prdm13 is sufficient to induce GABAergic markers in upon ectopic expression of mPrdm13 (Fig. 9D–F). Conversely, animal caps. Supplemental material Fig. S9A shows that over- Tlx3 þ glutamatergic neurons were reduced in the dI3 and dI5 expression of mPrdm13 is not sufficient to increase Pax2, Lbx1 and domains (Fig. 9G–I). Similarly, the number of cells expressing Gad1 in animal cap explants. Lmx1b was decreased in dI5 (Fig. 9J–L). Thus, Prdm13 phenocopies While this manuscript was under review, another study was the effects of Ptf1a in vivo, suppressing glutamatergic and inducing published identifying also Prdm13 as a critical regulator of the GABAergic markers. balance between inhibitory and excitatory neurons acting down- stream of Ptf1a in the dorsal spinal cord. In their work, Prdm13 is proposed to act through interaction with the bHLH factor Ascl1 and the repression of its ability to activate Tlx3. The authors also Discussion suggest that Prdm13 may also function by interfering with other activators of glutamatergic genes (Chang et al., 2013). We therefore The balance between excitatory versus inhibitory fate specifica- tested in animal caps whether Prdm13 affects the activity of the tion during development of the central nervous system is a tightly bHLH factor Neurog2, which is known to function in some regulated process that requires precise transcriptional control, and contexts in glutamatergic differentiation (Schuurmans et al., an imbalance leads to sensory disorders. In the dorsal spinal cord, 2004; Winpenny et al., 2011). We found that, as previously the network of transcription factors required for this process reported (Dullin et al., 2007), Neurog2 overexpression in animal includes members of the bHLH and HD families. A key transcrip- caps strongly activates the expression of the glutamatergic speci- tion factor in this process is the bHLH factor Ptf1a that functions in fier Tlx3 and pan-neuronal markers such as Myt1 but not Pax2. a complex with the Notch mediator Rbp-j to promote GABAergic Interestingly, this dramatic induction of Tlx3 by Neurog2 was and glycinergic inhibitory cell fates and to suppress a glutamater- abrogated by the coexpression of Prdm13. No such effect was gic excitatory cell fate (Beres et al., 2006; Hori et al., 2008). Ptf1a is observed on Myt1. Pax2 that is reduced upon Neurog2 overexpres- known to promote genes encoding transcription factors such as sion was slightly upregulated when Prdm13 is coexpressed (Fig. 8). Pax2 required for the formation of GABAergic neurons and to Thus, Prdm13 specifically silences Neurog2 transcriptional activity suppress Tlx3, an important determinant for glutamatergic cell fate on the Tlx3 gene. (Hori and Hoshino, 2012). However, how Ptf1a operates to fulfill

Fig. 8. Real-time RT-PCR analysis of animal cap explants isolated from embryos injected with Neurog2 and Prdm13 mRNA as indicated and collected when sibling embryos reached stage 26. Note that the coexpression of Prdm13 with Neurog2 abrogates its ability to induce Tlx3. Myt1 induction by Neurog2 is in contrast not affected. Pax2 is slightly upregulated by the coexpression of Prdm13. CC, control caps. J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 353

Fig. 9. Prdm13 overexpression induces GABAergic and suppresses glutamatergic markers in the chick neural tube. (A–L) Immunofluorescence on transverse sections of HH24–25 chick neural tube, with the right side of the neural tube electroporated with a mouse pCIG-Prdm13 expression construct with antibodies against Pax2, Lbx1, Tlx3, Lmx1b in red and GFP in green. Note the disappearance of the Pax2 gap of staining between the presumptive dI4 and dI6 neurons (arrow) and the Pax2 staining above dI4 (asterisk). Lbx1 is also detected more ventrally and dorsally (asterisks) upon Prdm13 overexpression. In contrast, Tlx3 in dI3/5 and Lmx1b in dI5 are decreased. High magnification views of the control and electroporated sides in the dorsal half of the neural tube are shown in C0–C00 and F0–F00. In contrast, in I0–I00 and L0–L00, no GFP þ Prdm13 expressing cells are Tlx3 þ or Lmx1bþ . (M) Quantification of the number of Pax2, Tlx3, Lbx1 and Lmx1b positive cells in the different progenitor regions on the Prdm13 electroporated side compared to the contralateral, control side. Po0.05n, Po0.01nn, Po0.001nnn;(nZ3). 354 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 such a function is not known because of lack of information about whose expression also overlaps with that of Prdm13 in the dorsal its targets in the nervous system. Here we report the identification spinal cord does not appear to regulate Prdm13. We however in Xenopus of a novel likely direct target of the PTF1-J complex, observed that the combined overexpression of Ptf1a and Neurog2 Prdm13, with a key role in the control of the balance of inhibitory induces a higher level of Prdm13 than that observed in Ptf1a alone versus excitatory neurons in the dorsal neural tube. or Neurog2 alone injected embryos. This observation may reflect a synergistic interaction between the two factors. It may however Prdm13 has histone methyltransferase activity also simply be a secondary consequence of Neurog2 proneural activity leading to a larger number of neurons with their fate being Prdm genes epigenetic transcriptional regulators. Some subsequently determined by Ptf1a. members of the Prdm family have been shown to possess intrinsic Prdm13 is likely directly regulated by Ptf1a as it is rapidly HMT activity while others act via interactions with enzymes induced upon Ptf1a overexpression and that this induction occurs capable of methylating histones or catalyzing other post- even in the presence of the translation blocking drug cyclohex- translational modifications of chromatin histone (Fog et al., 2011; imide. Several Ptf1a binding regions surrounding the Prdm13 gene Hohenauer and Moore, 2012). We showed that Prdm13 has HMT were recently identified on neural tube tissue, which demon- activity in 293T cells and have observed that it causes H3K9 and strates that Prdm13 is indeed a direct downstream target of Ptf1a H3K27 trimethylation (B. Van Driessche, unpublished data) which (Chang et al., 2013). induces repressive chromatin states. Chang et al. (2013) has shown Intrinsic functional properties of bHLH transcription factors is that Prdm13 has some function in vivo in neuronal specification driven by specificity of protein–protein interactions or DNA bind- without the PR domain. Whether Prdm13 has intrinsic HMT ing specificity. In structure function analysis, functional differences activity dependent on the PR domain is an important issue that between bHLH proteins have been mapped to residues within the requires further investigation. bHLH domain (Nakada et al., 2004; Skowronska-Krawczyk et al., 2005; Powell and Jarman, 2008). We observed that a Ptf1a/bHLH- Prdm13 expression and regulation in the dorsal caudal neural tube Neurog2-GR chimeric protein, where the Ptf1a bHLH has been replaced with that of Neurog2, is still able to induce Prdm13, Prdm13 is, together with the closely related Prdm8 gene, one of indicating that in the case of Ptf1a, it is the sequences outside the the members of the Prdm family that has CNS specific expression bHLH domain that are most important for its specify of function. and is not detected in invertebrates (Hohenauer and Moore, 2012). Our data indicate that the unique ability of Ptf1a to induce Prdm13 Our data shows that Prdm13 expression in Xenopus is highly is due to the C1/C2 motifs, located in its C-terminal region, which similar to that in mouse (Kinameri et al., 2008) and in zebrafish allow interaction with Rbp-j (Beres et al., 2006). This is supported (Sun et al., 2008), suggesting evolutionary conservation of function by the fact that in the frog (1) overexpression of Ptf1a in in vertebrates. As in those species, Prdm13 expression in the frog is combination with Rbp-j induces higher Prdm13 transcript levels only detected in the caudal neural tube and is highly restricted than Ptf1a alone, (2) a mutated form of Ptf1aW224A/W242A-GR along its D/V axis. At the neural plate stage, Prdm13 expression is unable to interact with Rbp-j is unable to induce Prdm13 and detected in the region of the differentiating primary interneurons, (3) in mouse Ptf1aW298A mutant embryos, Prdm13 was strongly suggesting a role in primary neurogenesis. It is detected more reduced in the spinal cord. Recent ChIP-seq data have shown that dorsally compared to Dbx1 that marks p0 progenitors and Prdm8 Rbp-j colocalizes with several Ptf1a peaks surrounding the mouse that defines the p0 to pMN progenitor regions (Kinameri et al., Prdm13 locus, further demonstrating that Prdm13 is a target of the 2008). We detected a gap of expression in between the stripes of PTF1-J complex (Chang et al., 2013). The reduced activity of the Prdm13 and Prdm8. This gap is likely to correspond to the Dbx1 Ptf1a/bHLH-Neurog2-GR chimeric protein compared to the wild domain of expression. In the mouse, low level of Prdm13 expres- type protein suggest a possible contribution of the bHLH domain sion is also detected in the ventricular zone of the neural tube of of Ptf1a to its specificity of function. This contribution, which may mouse E11.5 embryos in the region expressing Dbx1 (Kinameri be due either to some E-box binding sites preference or to some et al., 2008). Whether Prdm13 and Dbx1 crossregulate each other specific protein–protein interactions, remains to be further remains to be investigated. investigated. As in the mouse, Prdm13 in the frog is weakly expressed in the ventricular zone and stronger staining is detected at the margin of Function of Prdm13 in neuronal specification the VZ, suggesting that the gene may be the most active when cells stop to proliferate and start to differentiate. Along the A/P axis, we In the dorsal spinal cord, Ptf1a act as a selector gene, promoting found that in the frog neurula and early tailbud stage embryos, GABA/glycinergic neurons and suppressing alternate fates through Prdm13 is restricted to the caudal hindbrain and spinal cord and Tlx3 repression (Glasgow et al., 2005; Hori et al., 2008). In the frog, that its expression only later extends to the anterior hindbrain we found that knock-down of Prdm13 in embryos and in animal region. A detailed analysis of the temporal expression of Prdm13 cap explants reduces the ability of Ptf1a to induce the GABAergic during embryogenesis is not available in zebrafish and mouse, so it markers and upregulates glutamatergic determinants. Conversely, is not clear whether Prdm13 has a similar dynamic expression in in the chick neural tube, Prdm13 overexpression induces GABAergic those species. Outside the developing spinal cord, Prdm13 is also and suppresses glutamatergic markers. These results and those expressed in the frog in the retina as in the mouse and zebrafish similar obtained by Chang et al. (2013) indicate that Prdm13 is a (Kinameri et al., 2008; Sun et al., 2008). No expression was downstream target of Ptf1a that has a conserved essential function detected in the frog in olfactory placodes as reported in zebrafish in the control of the balance between GABAergic and glutamater- (Sun et al., 2008). gic neurons. Our results indicate that within both the dorsal spinal cord and In the frog mature retina, we showed that Prdm13 is coex- retina, Prdm13 expression overlaps with that of Ptf1a and that pressed with Ptf1a in the CMZ and that it is also expressed in post- Ptf1a promotes Prdm13 expression in the spinal cord. We also mitotic cells in the inner nuclear layer. This layer contains ama- observed that Prdm13 is upregulated by Ascl1. This is in accor- crine and horizontal cells, which are mainly GABAergic or glyci- dance with Ascl1 controlling GABAergic differentiation in the nergic, and ganglion cells which are considered mainly as dorsal spinal cord by regulating Ptf1a expression (Mizuguchi glutamatergic. Whether Prdm13 is expressed in all or a subset of et al., 2006; Mazurier et al., unpublished). In contrast, Neurog2 GABA and/or glycinergic neurons and is excluded from excitatory J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 355 cells remains to be investigated. As observed in the mouse (Chang of GABA/glutamatergic cell fate specification in the vertebrate et al., 2013), we found that Prdm13 is lost in the retina from Ptf1a dorsal neural tube. Future work examining its biochemical proper- depleted embryos (Hanotel, unpublished data). Based on the ties and using genetic approaches is required to get a better importance of Prdm13 in GABAergic cell fate within the caudal understanding of its mechanism of action and its function in other neural tube, it is thus tempting to speculate that it also plays a role regions of the CNS such as the retina. in the retina in the suppression of the expression of non-Ptf1a lineages. In the retina, the fates of all retinal neurons that primarily express the inhibitory neurotransmitters GABA or gly- Acknowledgments cine require the expression of Ptf1a (Fujitani et al., 2006; Dullin et al., 2007; Nakhai et al., 2007; Jusuf et al., 2011). Ptf1a alone is The authors acknowledge F. Clotman, Gregory Dressler, Jane however not sufficient to confer subtype-specificity (Jusuf et al., Johnson, Adrian Moore, Thomas Muller, Fujio Murakani, Hajime 2011). Whether Prdm13 is required downstream of Ptf1a to specify Ogino, Nicolas Pollet and Chris Wright for generously providing all inhibitory neurons or whether it biases precursors towards a reagents used in this work. We also thank Louis Delhaye and specific subtype of amacrine and horizontal destinies is an Nadège Delacourt for excellent technical assistance. This work was important question that remains to be addressed. supported by grants from the Belgian Fonds de la Recherche Scientifique, the Télévie program of the FRS-FNRS, the Belgian Mechanisms of action of Prdm13 in neuronal specification Queen Elisabeth Medical Foundation, the Van Buuren, Jean Brachet and Defay Foundation, the Fédération Wallonie-Bruxelles (Action How does Prdm13 function downstream of Ptf1a to promote de Recherche Concertée), the Interuniversity Attraction Poles Pax2þ /GABAergic over Tlx3 þ/glutamatergic neuronal fate? As Programme, Belgian State, Federal Office for Scientific, Technical many Prdm members function as repressors (Hohenauer and and Cultural Affairs and the Walloon Region First International Moore, 2012), one attractive possibility is that it may do so programme and Excellence Programme (“CIBLES”) to E.B and C.V. indirectly by repressing transcription of genes such as Tlx3 that L., and from the Agence Nationale de la Recherche and Retina induce the glutamatergic program and antagonize Lbx1 that France to M.P. Part of this work (T.P., K.A.H.) was funded through promote Pax2 and the GABAergic program (Cheng et al., 2004; support by the Cluster of Excellence and DFG Research Center Cheng et al., 2005). This is supported by our observation that Nanoscale Microscopy and Molecular Physiology of the Brain. A.T. Prdm13 overexpression in 293T cells increases the level of several is a postdoctoral fellow from the Belgian Fonds de la Recherche histone repressive marks (Van Driessche, unpublished data). This Scientifique (FNRS). N.B. is a postdoctoral fellow BRIC (Bureau des has been also very recently supported by the additional observa- Relations Internationales et de la coopération, ULB). J.H. is a tions that (1) Prdm13 occupies genomic regions near the Tlx1 and doctoral fellow from the Belgian Fonds pour la formation à la 0 0 Tlx3 genes and (2) Prdm13-bound regions have enhancer activities Recherche dans l Industrie et dans l Agriculture (FRIA). C.V.L. is that is blocked by ectopic Prdm13 (Chang et al., 2013). In their Directeur de Recherches of FRS-FNRS. B.V.D. is a postdoctoral study, Chang et al. suggest that Prdm13 acts during GABAergic fellow from CIBLES. neuron differentiation through interaction with the bHLH factor Ascl1 and the repression of its ability to activate Tlx3. Our data suggest that Prdm13 silence glutamatergic lineage genes not only Appendix A. Supplementary materials through repression of the activity of Ascl1 but also through repression of Neurog2 activity. Neurog2 is essential for the gen- Supplementary materials associated with this article can be eration of sensory neurons, spinal motor neurons, glutamatergic found in the online version at http://dx.doi.org/10.1016/j.ydbio. neurons in the cerebral cortex, dopaminergic neurons in the 2013.12.024. ventral midbrain and in the developing olfactory system (Fode et al., 1998, 2000; Ma et al., 1999; Novitch et al., 2001; Scardigli References et al., 2001; Schuurmans et al., 2004; Kele et al., 2006; Shaker et al., 2012). Its role in the formation of GABAergic interneurons in Afelik, S., Chen, Y., Pieler, T., 2006. Combined ectopic expression of Pdx1 and Ptf1a/ the spinal cord is unclear. Results of the analysis of embryos null p48 results in the stable conversion of posterior endoderm into endocrine and exocrine pancreatic tissue. Genes Dev. 20, 1441–1446. for Ascl1 or Neurog2 and of Ascl1/Neurog2 double knockouts Albert, M., Helin, K., 2010. Histone methyltransferases in cancer. Sem. Dev. Biol. 21, suggest that there may be a redundant function for Ascl1 and 209–220. Neurog2 in the generation of dI4 neurons (Helms et al., 2005). Batista, M.F., Lewis, K.E., 2008. Pax2/8 act redundantly to specify glycinergic and GABAergic fates of multiple spinal interneurons. Dev. Biol. 323, 88–97. Whether mPrdm13 forms a repressor complex with mNeurog2 Beres, T.M., Masui, T., Swift, G.H., Shi, L., Henke, R.M., MacDonald, R.J., 2006. PTF1 is and thus switches it from an activator to a repressor with respect an organ-specific and Notch-independent basic helix–loop–helix complex to Tlx3 gene expression, as seen with Ascl1, remains to be containing the mammalian Suppressor of Hairless (RBP-J) or its paralogue, RBP-L. Mol. Cell Biol. 26, 117–130. addressed (Chang et al., 2013). Bröhl, D., Strehle, M., Wende, H., Hori, K., Bormuth, I., Nave, K.-A., Müller, T., Our results also indicate that Prdm13 negatively self-regulates Birchmeier, C., 2008. A transcriptional network coordinately determines its expression. This negative autoregulation that limits Prdm13 transmitter and pettidergic fate in the dordal spinal cord. Dev. Biol. 322, 381–393. expression may be important to avoid to reach a level of Prdm13 Bradley, L.C., Snape, A., Bhatt, S., Wilkinson, D.G., 1993. The structure and that would allow unspecific binding and interference with other expression of the Xenopus Krox-20 gene: conserved and divergent pattern of factors required for GABAergic neurons. In contrast to Prdm13, expression in rhombomeres and neural crest. Mech. Dev. 40, 73–84. Ptf1a is not upregulated in Prdm13 morphants (Hanotel, unpub- Brzezinski, J.A., Lamba, D.A., Reh, T.A., 2010. Blimp1 controls photorecptor versus bipolar cell fate choice during retinal development. Development 137, 619–629. lished data), indicating that the increase of Prdm13 transcription is Chang, J.C., Meredith, D.M., Mayer, P.R., Borromeo, M.D., Lai, H.C., Ou, Y.H., Johnson, not due to a negative retrocontrol of Prdm13 on Ptf1a expression. J.E., 2013. Prdm13 mediates the balance of inhibitory and excitatory neurons in – Prdm13 negative autoregulation is thus likely to be due to other somatosensory circuits. Dev. Cell. 25, 182 195. Cheng, L., Arata, A., Mizuguchi, R., Qian, Y., Karunaratne, A., Gray, P.A., Arata, S., mechanisms. Although there is no evidence for Prdm13 directly Shirasawa, S., Bouchard, M., Luo, P., Chen, C.L., Busslinger, M., Goulding, M., binding to DNA, another possibility is a direct autorepressive Onimaru, H., Ma, Q., 2004. Tlx3 and Tlx1 are post-mitotic selector genes action of Prdm13. determining glutamatergic over GABAergic cell fates. Nat. Neurosci. 7, 510–517. Cheng, L., Samad, O.A., Xu, Y., Mizuguchi, R., Luo, P., Shirasawa, S., Goulding, M., Ma, Q., Altogether, our gain and loss of function data identify the zinc 2005. Lbx1 and Tlx3 are opposing switches in determining GABAergic versus finger transcription factor Prdm13 as a conserved critical regulator glutamatergic transmitter phenotypes. Nat. Neurosci. 8, 1510–1515. 356 J. Hanotel et al. / Developmental Biology 386 (2014) 340–357

Chitnis, A., Henrique, D., Lewis, J., Ish-Horowicz, D., Kintner, C., 1995. Primary Katoh, K., Omori, Y., Onishi, A., Sato, S., Kondo, M., Furukawa, T., 2010. Blimp1 neurogenesis in Xenopus embryos regulated by a homolog of the Drosophila suppresses Chx10 expression in differentiating retinal photoreceptor precur- neurogenic gene Delta. 375, 761–766. sors to ensure proper photoreceptor development. J. Neurosci. 30, 6515–6526. Dressler, G.R., Deutsch, U., Chowdhury, K., Nornes, H.O., Gruss, P., 1990. Pax2, a new Kawaguchi, Y., Cooper, B., Gannon, M., Ray, M., MacDonald, R.J., Wright, C.V., 2002. murine paired-box-containing gene and its expression in the developing The role of the transcriptional regulator Ptf1a in converting intestinal to excretory system. Development 109, 787–795. pancreatic progenitors. Nat. Genet. 32, 128–134. Dullin, J.P., Locker, M., Robach, M., Henningfeld, K.A., Parain, K., Afelik, S., Pieler, T., Kele, J., Simplicio, N., Ferri, A.L., Mira, H., Guillemot, F., Arenas, E., Ang, S.L., 2006. Perron, M., 2007. Ptf1a triggers GABAergic neuronal cell fates in the retina. BMC Neurogenin 2 is required for the development of ventral midbrain dopaminer- Dev. Biol. 7, 110. gic neurons. Development 133, 495–505. Endo, K., Karim, M.R., Taniguchi, H., Krejci, A., Kinameri, E., Siebert, M., Ito, K., Bray, Kinameri, E., Inoue, T., Aruga, J., Imayoshi, I., Kageyama, R., Shimogori, T., Moore, A., S.J., Moore, A.W., 2011. Chromatin modification of Notch targets in olfactory 2008. Prdm proto-oncogene transcription factor family expression and receptor neuron diversification. Nat. Neurosci. 15, 224–233. interaction with the Notch-Hes pathway in mouse neurogenesis. Plos One 3, Eom, G.H., Kim, K., Kim, S.M., Kee, H.J., Kim, J.Y., Jin, H.M., Kim, J.R., Kim, J.H., Choe, e3859. N., Kim, K.B., Lee, J., Kook, H., Kim, N., Seo, S.B., 2009. Histone methyltransferase Klisch, T.J., Souopgui, J., Juergens, K., Rust, B., Pieler, T., Henningfeld, K.A., 2006. PRDM8 regulates mouse testis steroidogenesis. Biochem. Biophys. Res. Com- Mxi1 is essential for neurogenesis in Xenopus and acts by bridging the pan- mun. 388, 131–136. neural and proneural genes. Dev. Biol. 292, 470–485. Fode, C., Gradwohl, G., Morin, X., Dierich, A., LeMeur, M., Goridis, C., Guillemot, F., Komai, T., Iwanari, H., Mochizuki, Y., Hamakubo, T., Shinkai, Y., 2009. Expression of 1998. The bHLH protein NEUROGENIN 2 is a determination factor for epibran- the mouse PR domain protein Prdm8 in the developing central nervous system. chial placode-derived sensory neurons. Neuron 20, 483–494. Gene Expr. Pattern 9, 503–514. Fode, C., Ma, Q., Casarosa, S., Ang, S.L., Anderson, D.J., Guillemot, F., 2000. A role for Hoshino, M., Nakamura, S., Mori, K., Kawauchi, T., Nishimura, Y.V., Fukuda, A., Fuse, T., neural determination genes in specifying the dorsoventral identity of tele- Matsuo, N., Sone, M., Watanabe, M., Bito, H., Terashima, T., Wright, C.V., ncephalic neurons. Genes Dev. 14, 67–80. Kawaguchi, Y., Nakao, K., Nabeshima, Y., 2005b. Ptf1a, a bHLH transcriptional Fog, C.K., Galli, G.G., Lund, A.H., 2011. PRDM proteins: important players in gene, defines GABAergic neuronal fates in cerebellum. Neuron 47, 201–213. differentiation and disease. BioEssays 34, 50–60. Krapp, A., Knöfler, M., Ledermann, B., Bürki, K., Berney, C., Zoerkler, N., Francius, C., Clotman, F., 2010. Dynamic expression of the Onecut transcription Hagenbüchle, O., Wellauer, P.K., 1998. The bHLH protein PTF1-p48 is essential factors HNF-6, OC-2 and OC-3 during spinal motor neuron development. for the formation of the exocrine and the correct spatial organization of the Neuroscience 165, 116–129. endocrine pancreas. Genes Dev. 12, 3752–3763. Fujitani, Y., Fujitani, S., Luo, H., Qiu, F., Burlison, J., Long, Q., Kawaguchi, Y., Edlund, H., Lea, R., Bonev, B., Dubaissi, E., Vize, P.D., Papalopulu, N., 2012. Multicolor fluorescent MacDonald, R.J., Furukawa, T., Fujikado, T., Magnuson, M.A., Xiang, M., Wright, C. in situ mRNA hybridization (FISH) on whole mounts and sections. Methods Mol. V., 2006. Ptf1a determines horizontal and amacrine cell fates during mouse Biol. 917, 431–444. retinal development. Development 133, 4439–4450. Lee, J.E., Hollenberg, S.M., Snider, L., Turner, D.L., Lipnick, N., Weintraub, H., 1995. Fujiyama, T., Yamada, M., Terao, M., Terashima, T., Hioki, H., Inoue, Y.U., Inoue, T., Conversion of Xenopus ectoderm into neurons by NeuroD, a basic helix–loop– Masuyama, N., Obata, K., Yanagawa, Y., Kawaguchi, Y., Nabeshima, Y., Hoshino, helix protein. Science 268, 836–844. M., 2009. Inhibitory and excitatory subtypes of cochlear nucleus neurons are Ma, Q., Kintner, C., Anderson, D.J., 1996. Identification of neurogenin, a vertebrate defined by distinct bHLH transcription factors, Ptf1a and Atoh1. Development neuronal differentiation gene. Cell 87, 43–52. 136, 2049–2058. Ma, Q., Fode, C., Guillemot, F., Anderson, D.J., 1999. Neurogenin1 and neurogenin2 Gershon, A.A., Rudnick, J., Kalam, L., Zimmerman, K., 2000. The homeodomain- control two distinct waves of neurogenesis in developing dorsal root ganglia. containing gene Xdbx inhibits neuronal différentiation in the developing Genes Dev. 13, 1717–1728. embryo. Development 127, 2945–2954. Masui, T., Long, Q., Beres, T.M., Magnuson, M.A., MacDonald, R.J., 2007. Early Glasgow, S.M., Henke, R.M., MacDonald, R.J., Wright, C.V.E., Johnson, J.E., 2005. Ptf1a pancreatic development requires the vertebrate Suppressor of Hairless (RBPJ) determines GABAergic over glutamatergic neuronal cell fate in the spinal cord in the PTF1 bHLH complex. Genes Dev. 21, 2629–2643. dorsal horn. Development 132, 5461–5469. Masui, T., Swift, G.H., Hale, M.A., Meredith, D.M., Johnson, J.E., MacDonald, R.J., Hamburger, V., Hamilton, H.L., 1992. A series of normal stages in the development 2008. Transcriptional autoregulation controls pancreatic Ptf1a expression of the chick embryo. Dev. Dyn. 195, 231–272. during development and adulthood. Mol. Cell Biol. 28, 5458–5468. Heller, N., Brändli, A.W., 1997. Xenopus Pax-2 displays multiple splice forms during Mazurier, N., Parlier, D., Parain, K., Pretto, S., Vernier, P., Locker, M., Bellefroid, E.J., embryogenesis and pronephric kidney development. Mech. Dev. 69, 83–104. Perron, M., 2014. Ascl1 as a novel player in the Ptf1a transcriptional network for Helms, A.W., Johnson, J.E., 2003. Specification of dorsal spinal cord interneurons. GABAergic cell specification in the retina. (unpublished). Curr. Opin. Neurol. 13, 42–49. Meredith, D.M., Masui, T., Swift, G.H., MacDonald, R.J., Johnson, J.E., 2009. Multiple Helms, A.W., Battiste, J., Henke, R.M., Nakada, Y., Simplicio, N., Guillemot, F., transcriptional mechanisms control Ptf1a levels during neural development Johnson, J.E., 2005. Sequential role for Mash1 and Ngn2 in the generation of including autoregulation by the PTF1-J complex. J. Neurosc 29, 11139–11148. dorsal spinal cord interneurons. Development 132, 2709–2719. Mizuguchi, R., Kriks, S., Cordes, R., Gossler, A., Ma, Q., Goulding, M., 2006. Asl1 and Hemmati-Brivanlou, A., de la Torre, J.R., Holt, C., Harland, R.M., 1991. Cephalic Gsh1/2 control inhibitory and excitatory cell fate in spinal sensory interneur- expression and molecular characterization of Xenopus En-2. Development 111, ons. Nat. Neurosci. 9, 770–778. 715–724. Nakada, Y., Hunsaker, T.L., Henke, R.M., Johnson, J.E., 2004. Distinct domains within Henke, R.M., Savage, T.K., Meredith, D.M., Glasgow, S.M., Hori, K., Dumas, J., Mash1 and Math1 are required for function in neuronal differentiation versus MacDonald, R.J., Johnson, J.E., 2009. Neurog2 is a direct dowstream target of neuronal cell-type specification. Development 131, 1319–1330. the Ptf1a–Rbpj transcription complex in dorsal spinal cord. Development 136, Nakhai, H., Sel, S., Favor, J., Mendoza-Torres, L., Paulsen, F., Duncker, G.I., Schmid, R. 2945–2954. M., 2007. Ptf1a is essential for the differentiation of GABAergic and glycinergic Hernandez-Lagunas, L., Powell, D.R., Law, J., Grant, K.A., Artinger, K.B., 2011. Prdm1a amacrine cells and horizontal cells in the mouse retina. Development 134, and olig4 act downstream of Notch signaling to regulate cell fate at the neural 1151–1160. plate border. Dev. Biol. 356, 496–505. Nieto, M.A., Bradley, L.C., Wilkinson, D.G., 1991. Conserved segmental expression of Hohenauer, T., Moore, A.W., 2012. The Prdm family: expanding roles in stem cells Krox-20 in the vertebrate hindbrain and its relationship to lineage restriction. and development. Development 139, 2267–2282. Development Suppl 2, 59–62. Hori, K., Cholewa-Waclaw, J., Nakada, Y., Glasgow, S.M., Masui, T., Henke, R.M., Nieuwkoop, P.D., Faber, J., 1967. Normal Table of Xenopus laevis (Daudin). North Wildner, H., Martarelli, B., Beres, T.M., Epstein, J.A., Magnuson, M.A., Holland, Publishing Co., Amsterdam, The Netherlands. Macdonald, R.J., Birchmeier, C., Johnson, J.E., 2008. A nonclassical bHLH Rbpj Nishida, K., Hoshino, M., Kawagushi, Y., Murakami, F., 2010. Ptf1a directly controls transcription factor complex is required for specification of GABAergic neurons expression of immunoglobulin superfamily molecules Nephrin and Neph3 in independent of Notch signaling. Genes Dev. 22, 166–178. the developing nervous system. J. Biol. Chem. 285, 373–380. Hori, K., Hoshino, M., 2012. GABAergicneuron specification in the spinal cord, the Novitch, B.G., Chen, A.I., Jessell, T.M., 2001. Coordinate regulation of motor neuron cerebellum, and the cochlear nucleus. Neural Plast., 921732. subtype identity and pan-neuronal properties by the bHLH repressor Olig2. Hoshino, M., Nakamura, S., Mori, K., Kawauchi, T., Terao, M., Nishimura, Y., Fukuda, A., Neuron 31, 773–789. Fuse, T., Matsuo, N., Sone, M., Watanabe, M., Bito, H., Terashima, T., Wright, C.V., Parlier, D., Ariza, A., Christulia, F., Genco, F., Vanhomwegen, J., Kricha, S., Souopgui, Kawaguchi,Y.,Nakao,K.,Nabeshima,Y.,2005a.Ptf1a,abHLHtranscriptionalgene, J., Bellefroid, E.J., 2008. Xenopus zinc finger transcription factor IA1 (Insm1) defines GABAergic neuronal fates in cerebellum. Neuron 47, 201–213. expression marks anteroventral noradrenergic neuron progenitors in Xenopus Huang, M., Huang, T., Xiang, Y., Xie, Z., Chen, Y., Yan, R., Xu, J., Cheng, L., 2008. Ptf1a, embryos. Dev. Dyn. 237, 2147–2157. Lbx1 and Pax2 coordinate glycinergic and peptidergic transmitter phenotypes Pascual, M., Abasolo, I., Mingorance-Le Meur, A., Martínez, A., Del Rio, J.A., Wright, in dorsal spinal inhibitory neurons. Dev. Biol. 322, 394–405. C.V., Real, F.X., Soriano, E., 2007. Cerebellar GABAergic progenitors adopt an Huang, S., Shao, G., Liu, L., 1998. The PR domain of the Rb-binding zinc finger external granule cell-like phenotype in the absence of Ptf1a transcription factor protein RIZ1 is a protein binding interface and is related to the SET domain expression. Proc. Natl. Acad. Sci. USA 104, 5193–5198. functioning in chromatin-mediated gene expression. J. Biol. Chem. 273, Perron, M., Opdecamp, K., Butler, K., Harris, W.A., Bellefroid, E.J., 1999. X-ngnr-1 and 15933–15939. Xath3 promote ectopic expression of sensory neuron markers in the neurula Ishibashi, S., Yasuda, K., 2001. Distinct roles of genes during Xenopus lens ectoderm and have distinct inducing properties in the retina. Proc. Natl. Acad. development. Mech. Dev. 101, 155–166. Sci. USA 96. (14996-5001). Jusuf, P.R., Almeida, A.D., Randlett, O., Joubin, K., Poggi, L., Harris, W.A., 2011. Origin Patterson, K.D., Krieg, P.A., 1999. Hox11-family genes XHox11 and XHox11L2 in and determination of inhibitory cell lineages in the vertebrate retina. J. Xenopus: XHox11L2 expression is restricted to a subset of the primary sensory Neurosci. 31, 2549–2562. neurons. Dev. Dyn 214, 34–43. J. Hanotel et al. / Developmental Biology 386 (2014) 340–357 357

Pillai, A., Mansouri, A., Behringer, R., Westphal, H., Goulding, M., 2007. Lhx1 and Sive, H.L., Grainger, R.M., Harland, R.M., 2000. Early development of Xenopus laevis. Lhx5 maintain the inhibitory-neurotransmitter status of interneurons in the A Laboratory Manual. CSHL Press, Cold Spring Harbor, New-York. dorsal spinal cord. Development 134, 357–366. Skowronska-Krawczyk, D., Matter-Sadzinski, L., Ballivet, M., Matter, J.M., 2005. The Pinheiro, I., Margueron, R., Shukeir, N., Eisold, M., Fritzsch, C., Richter, F.M., Mittler, basic domain of ATH5 mediates neuron-specific promoter activity during retina G., Genoud, C., Goyama, S., Kurokawa, M., Son, J., Reinberg, D., Lachner, M., development. Mol. Cell Biol 25, 10029–10039. Jenuwein, T., 2012. Prdm3 and Prdm16 are H3K9me1 methyltransferases Sturgeon, K., Kaneko, T., Biemann, M., Gauthier, A., Chawengsaksophak, K., Cordes, required for mammalian heterochromatin integrity. Cell 150, 948–960. S., 2011. Cdx1 refines positional identity of the vertebrate hindbrain by directly Powell, L.M., Jarman, A.P., 2008. Context dependence of proneural bHLH proteins. repressing Mafb expression. Development 138, 65–74. Curr Opin. Genet. Dev 18, 411–417. Sun, X.-J., Xu, P.-F., Zhou, T., Hu, M., Fu, C.-T., Zhang, Y., Jin, Y., Chen, Y., Chen, S.-J., Ross, S.E., McCord, A.E., Jung, C., Atan, D., Mok, S.I., Hemberg, M., Kim, T.K., Huang, Q.-H., Liu, T.X., Chen, Z., 2008. Genome wide Survey and developmental Salogiannis, J., Hu, L., Cohen, S., Lin, Y., Harrar, D., McInnes, R.R., Greenberg, expression mapping of zebrafish SET domain-containing genes. Plos One 3, M.E., 2012. Neuron 73, 292–303. e1499. Roy, S., Ng, T., 2004. Blimp-1 specifies neural crest and sensory neuron progenitors Thompson, N., Gésina, E., Scheinert, P., Bucher, P., Grapin-Botton, A., 2012. RNA in the zebrafish embryo. Curr. Biol. 14, 1772–1777. profiling and chromatin immunoprecipitation-sequencing reveal that PTF1a Scardigli, R., Schuurmans, C., Gradwohl, G., Guillemot, F., 2001. Crossregulation stabilizes pancreas progenitor identity via the control of MNX1/HLXB9 and a between Neurogenin2 and pathways specifying neuronal identity in the spinal network of other transcription factors. Mol. Cell. Biol 32, 1189–1199. cord. Neuron 31, 203–217. Van Campenhout, C., Nichane, M., Antoniou, A., Pendeville, H., Bronchain, O.J., Shaker, T., Dennis, D., Kurrasch, D.M., Schuurmans, C., 2012. Neurog1 and Marine, J.C., Mazabraud, A., Voz, M.L., Bellefroid, E.J., 2006. Evi1 is specifically Neurog2 coordinately regulate development of the olfactory system. Neural expressed in the distal tubule and duct of the Xenopus pronephros and plays a Dev. 20 (7), 28. role in its formation. Dev. Biol. 294, 203–219. Schuurmans, C., Armant, O., Nieto, M., Stenman, J.M., Britz, O., Klenin, N., Brown, C., Wettstein, D.A., Turner, D.L., Kintner, C, 1997. The Xenopus homolog of Drosophila Langevin, L.M., Seibt, J., Tang, H., Cunningham, J.M., Dyck, R., Walsh, C., suppressor of Hairless mediates Notch signaling during primary neurogenesis. Campbell, K., Polleux, F., Guillemot, F., 2004. Sequential phases of cortical Development 124, 693–702. specification involve Neurogenin-dependent and -independent pathways. Wilkinson, D.G., Nieto, M.A., 1993. Detection of messenger RNA by in situ EMBO J. 23, 2892–2902. hybridization to tissue sections and whole mounts. Methods Enzymol 225, Sellick, G.S., Barker, K.T., Stolte-Dijkstra, I., Fleishmann, C., Coleman, R.J., Garrett, C., 361–373. Gloyn, A., Edghill, L., Hattersley, E.L., Wellauer, A.T., Goodwin, P.K., Houlston, R. Winpenny, E., Lebel-Potter, M., Brill, M.S., Götz, M., Guillemot, F., Raineteau, O., S., G., 2004. Mutations in Ptf1a cause pancreatic and cerebellar agenesis. Nat. 2011. Sequential generation of olfactory bulb glutamatergic neurons by Genet. 36, 1301–1305. Neurog2-expressing precursor cells. Neural Dev 6, 12.