Open Collections

UBC Faculty Research and Publications

Combined serial analysis of gene expression and transcription factor binding site prediction identifies… Schmouth, Jean-François; Arenillas, David; Corso-Díaz, Ximena; Xie, Yuan-Yun; Bohacec, Slavita; Banks, Kathleen G; Bonaguro, Russell J; Wong, Siaw H; Jones, Steven J M; Marra, Marco A; Simpson, Elizabeth M; Wasserman, Wyeth W Jul 24, 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
52383-12864_2015_Article_1770.pdf [ 2.02MB ]
Metadata
JSON: 52383-1.0228466.json
JSON-LD: 52383-1.0228466-ld.json
RDF/XML (Pretty): 52383-1.0228466-rdf.xml
RDF/JSON: 52383-1.0228466-rdf.json
Turtle: 52383-1.0228466-turtle.txt
N-Triples: 52383-1.0228466-rdf-ntriples.txt
Original Record: 52383-1.0228466-source.json
Full Text
52383-1.0228466-fulltext.txt
Citation
52383-1.0228466.ris

Full Text

RESEARCH ARTICLE Open AccessCombined serial analysis of gene expressionand transcription factor binding siteprediction identifies novel-candidate-targetgenes of Nr2e1 in neocortex developmentJean-François Schmouth1,2,7, David Arenillas1, Ximena Corso-Díaz1,2, Yuan-Yun Xie1, Slavita Bohacec1,Kathleen G. Banks1, Russell J. Bonaguro1, Siaw H. Wong1, Steven J. M. Jones2,3,4,5, Marco A. Marra3,5,Elizabeth M. Simpson1,2,5,6*† and Wyeth W. Wasserman1,2,5†AbstractBackground: Nr2e1 (nuclear receptor subfamily 2, group e, member 1) encodes a transcription factor important inneocortex development. Previous work has shown that nuclear receptors can have hundreds of target genes, andbind more than 300 co-interacting proteins. However, recognition of the critical role of Nr2e1 in neural stem cellsand neocortex development is relatively recent, thus the molecular mechanisms involved for this nuclear receptorare only beginning to be understood. Serial analysis of gene expression (SAGE), has given researchers bothqualitative and quantitative information pertaining to biological processes. Thus, in this work, six LongSAGE mouselibraries were generated from laser microdissected tissue samples of dorsal VZ/SVZ (ventricular zone andsubventricular zone) from the telencephalon of wild-type (Wt) and Nr2e1-null embryos at the critical developmentages E13.5, E15.5, and E17.5. We then used a novel approach, implementing multiple computational methodsfollowed by biological validation to further our understanding of Nr2e1 in neocortex development.Results: In this work, we have generated a list of 1279 genes that are differentially expressed in response to alteredNr2e1 expression during in vivo neocortex development. We have refined this list to 64 candidate direct-targets ofNR2E1. Our data suggested distinct roles for Nr2e1 during different neocortex developmental stages. Mostimportantly, our results suggest a possible novel pathway by which Nr2e1 regulates neurogenesis, which includesLhx2 as one of the candidate direct-target genes, and SOX9 as a co-interactor.Conclusions: In conclusion, we have provided new candidate interacting partners and numerous well-developedtestable hypotheses for understanding the pathways by which Nr2e1 functions to regulate neocortex development.Keywords: SAGE, Nuclear receptor, Nr2e1, Transcriptome, Neocortex, Transcription factor* Correspondence: simpson@cmmt.ubc.ca†Equal contributors1Centre for Molecular Medicine and Therapeutics at the Child and FamilyResearch Institute, University of British Columbia, 950 West 28th Avenue,Vancouver, BC V5Z 4H4, Canada2Genetics Graduate Program, University of British Columbia, Vancouver, BCV6T 1Z2, CanadaFull list of author information is available at the end of the article© 2015 Schmouth et al. This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in anymedium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Schmouth et al. BMC Genomics  (2015) 16:545 DOI 10.1186/s12864-015-1770-3BackgroundThe proper development of the mammalian neocortexinvolves a balance between cell-intrinsic developmentalprograms and environmental factors. In this process,neurons acting as the backbone of the neuronal circuitryare generated first. These cells arise from the dorsaltelencephalon, generating cortical excitatory neurons byradial migration, and the ventral telencephalon givingrise to cortical inhibitory interneurons by tangentialmigration [1–5]. The neurogenic stage is followed by theintegration of glial cells in the circuitry during thegliogenic stage. In mice, neurons are generated from em-bryonic day 12 (E12) to E18, with astrocytes appearingat around E18 [6, 7]. Ultimately, the neocortex willcomprise six different radial layers with cell populationshaving distinct molecular identities [8].Nr2e1 (nuclear receptor subfamily 2, group e, member1, also known as Mtll, Tlx, Tll, and tailless) encodes atranscription factor important in the process of neocor-tex development [9, 10]. This complex cellular processinvolves a careful balance between proliferation of neuralstem cells (NSC), and the proper temporal differenti-ation of progenitor cells (PC) (i.e. neurons versus glia).Nr2e1 is expressed along the ventricular zone (VZ) ofthe dorsal telencephalon during neocortex developmentand is crucial for NSC self-renewal and maintenance[11–14]. Absence of Nr2e1 in mouse embryos reducesthe number of PC populating the VZ and subventricularzone (SVZ) during development, which results in re-duced thickness of the cortical plate [9]. The reductionin PC populating the VZ is more prominent in the cau-dal telencephalon whereas the reduction in the SVZ isseen at all rostrocaudal levels during development. Thiscell-reduction ultimately results in defects in structuresgenerated later, such as the upper cortical layers (layers IIand III), the dentate gyrus, and the olfactory bulb [9, 10].Absence of Nr2e1 in mouse embryos also results inpremature neurogenesis, which contributes to the defectsin the upper cortical layers [9].Previous work has shown that a nuclear receptor tran-scription factor can have hundreds of target genes [15],and the most extensively studied nuclear receptors areestimated to bind more than 300 co-interacting proteins[16, 17]. However, recognition of the critical role of nu-clear receptor Nr2e1 in NSC and neocortex developmentis relatively recent [11, 12, 18–20], thus the molecularmechanisms involved for this nuclear receptor are onlybeginning to be understood. First, in forebrain develop-ment, Nr2e1 has been shown to regulate cell cycle pro-gression via its interaction with the tumour suppressorgene Pten, and the cyclin-dependent kinase inhibitorp21 [11]. This involves a repressive mechanism mediatedvia the interaction of Nr2e1 with chromatin modifierproteins such as members of the histone deacetylasefamily (HDACs), and the demethylase protein LSD1(KDM1A) [14, 21]. Second, the balance between NSCproliferation and differentiation has been demonstrated tobe under the control of regulatory loops involving bothNr2e1, and microRNA encoding genes such as mir-9,miR-137, and let-7d [22–24]. This phenomenon includesan intricate network formed by the ability of let-7d andmir-9 to silence Nr2e1 expression by binding the 3′ UTRregions of this gene and the ability of Nr2e1 to inactivatethe expression of mir-9 in a first feedback loop [22, 24]. Asecond loop has been reported that includes the repres-sion of the co-interactor Lsd1 by miR-137 that can be re-lieved by the repression of miR-137 by Nr2e1 [23]. Finally,Nr2e1 has been shown to act as a transcriptional activatorof the deacetylase gene Sirt1, which has a role in promot-ing neuronal differentiation [25, 26]. Thus, we hypothesizethere is still much to learn about the molecular mecha-nisms underlying the role of NR2E1 in NSC and neocor-tex development. Hence, we have undertaken additionalresearch on these mechanisms, especially focused on invivo analyses, to inform our understanding of neocortexdevelopment.Large-scale transcriptome-profiling experiments, usingmethodologies such as serial analysis of gene expression(SAGE), have given researchers the advantages of bothqualitative and quantitative information pertaining tobiological processes. SAGE analysis relies on sequencingand quantification of short (14 bp) cDNA fragmentscalled tags, which are derived from messenger RNAtranscripts [27]. This approach is considered an opentranscriptome technology as no a priori knowledge ofthe transcript sequences is required [28]. For the mam-malian central nervous system, SAGE profiling experi-ments have been used to generate knowledge on avariety of topics including; fundamental studies on braindevelopment [29], connectivity, and aging [30–32], aswell as specific neuropathologies and drug responses[33–36]. Advancement in SAGE library generation suchas SAGE-lite [37], which enabled the use of extremelysmall quantities of tissues such as those from lasercapture microdissection (LCM), and LongSAGE, whichimproved tag-to-gene mapping by generating longer tagfragments (21 bp) [38], made these approaches particu-larly appropriate to reveal in vivo molecular changes inneocortex development. One of the inherent challengesin transcriptome profiling is the effective analysis oflarge-scale datasets to optimize extraction of relevantbiological meaning. By producing LongSAGE libraries atmultiple developmental times, in the presence andabsence of Nr2e1, we generated a rich dataset for com-parative analysis. Additionally, we took advantage of theintrinsic nature of transcription factors, which regulategene expression by binding to specific DNA sequences,and used it to further hone our gene list. This was partlySchmouth et al. BMC Genomics  (2015) 16:545 Page 2 of 19based on the power of transcription-factor-discovery-motif algorithms that, when coupled to cross-speciesgenome comparisons or phylogenetic footprinting, haveproven successful in making reliable binding site predic-tions [39–42]. Returning to biology to further validatethe bioinformatic studies, we of course used the litera-ture, but most importantly, we also tested our primarynew hypothesis in vitro by embryonic stem cells (ESC)differentiation and in vivo during brain development.Thus, in this work, we used a novel approach, imple-menting multiple computational methods to generatesignificant-novel-biological information regarding themolecular mechanisms underlying the role of nuclearreceptor Nr2e1 in neocortex development.Results and discussionLongSAGE libraries generated from laser capturemicrodissection tissuesTo identify novel-candidate-target and co-interactinggenes for the nuclear receptor Nr2e1, we favoured anin vivo source of RNA in order to most accuratelycapture molecular events occurring during neocortexdevelopment. Thus, LongSAGE libraries were preparedusing RNA purified from tissues obtained by LCM ofthe VZ/SVZ, of the dorsal-lateral telencephalon, fromWt and Nr2e1frc/frc embryos. This work was under-taken at three different developmental time points(E13.5, E15.5, and E17.5), which are known to expressNr2e1 in the dissected region [9, 11, 18, 43] (Fig. 1a).These libraries were sequenced to a depth ≥100,000tags (total number of tags per libraries, see Fig. 1b).To generate tags for analysis, we used a filtering pro-cedure involving the DiscoverySpace 4.0 application(http://www.bcgsc.ca/platform/bioinfo/software/ds) (fil-tering details, see Methods) [44]. On average, ~24 % ofthe total tags per library were discarded in this procedureresulting in a useful tag population averaging ~83,000 tagsper library, and corresponding to ~25,000 tag types perlibrary (Fig. 1b). Singleton tags (tags counted only once)constituted ~18 % of the useful tags population per libraryand ~68 % of the tag type population per library (Fig. 1b).These numbers were consistent with previously publishedresults, obtained using a similar filtering procedure [45].LongSAGE libraries differential statistical analyses andgene IDs recoveryThe Audic-Claverie significance test, implemented in theDiscoverySpace 4.0 application, was used to perform statis-tical analyses on the filtered tags [44, 46]. Tags differentiallyexpressed between Wt and Nr2e1frc/frclibraries at each timepoint (E13.5, E15.5, and E17.5), and falling within the confi-dence interval of 95 % (P < 0.05), according to the Audic-Claverie significance test, were retained for further analyses.The results for tags significantly increased or decreased inabundance at each time point are shown in Fig. 2a. Theproportion of differentially abundant tags (either increasedor decreased) varied between 15 to 25 % when comparedto the combined numbers of useful tags found in the Wtand Nr2e1frc/frc library at each time point (e.g. (Up at E13.5“Tags (P < 0.05)” (Fig. 2a)/(Wt +Nr2e1frc/frc at E13.5 “Totaluseful tags” (Fig. 1b))) × 100). LongSAGE tags were mappedto RefSeq (v52) and Ensembl (v66) databases [45]. On aver-age, 52 % of the differentially abundant tags mappedto genes (average of (“Tags mapped to genes”/“Tags(P < 0.05)”)×100 for each library, Fig. 2a). The numberof Refseq accession IDs corresponding to differentiallyabundant tags at the three different time points are alsoshown in Fig. 2a. These accession numbers, correspondingto Refseq genes, were retrieved and used in the subse-quent analyses.We next looked at the genes that were differentiallyregulated between the Wt and Nr2e1frc/frc libraries at theE13.5, E15.5, and E17.5 time points. This resulted in atotal of 1279 Refseq accession numbers, originatingfrom a corresponding list of 1387 tag sequences(Additional file 1: Table S1 and Additional file 2: Table S2),and distributed according to the Venn diagram in Fig. 2b.Interestingly, when performing the analyses, on average6 genes per time point corresponded to tags that werefound in both the up and down regulated populations(data not shown). This suggested that the tags mappedto these genes were corresponding to alternative tran-scripts that were expressed in opposing directions whencomparing Wt and Nr2e1frc/frc libraries. The Venndiagram results also demonstrated that on average54 % of the differentially-regulated genes (combinedup and down) were specific for each time point: E13.5,59 % ((383/650) × 100); E15.5, 59 % ((423/711) × 100);and E17.5, 44 % ((146/333) × 100). Furthermore, onaverage 17 % of the genes overlapped between at leasttwo time points: E13.5 and E15.5, 20 % (((140 + 88)/(650 + 423 + 60)) × 100); E15.5 and E17.5, 17 % (((88 +60)/(711 + 146 + 39)) × 100); and E13.5 and E17.5, 15 %(((88 + 39)/(650 + 146 + 60)) × 100). Finally, only 6.9 %of the genes overlapped between the three time points((88/1279) × 100).LongSAGE expression results suggested distinct roles forNr2e1 in different stages of neocortex developmentTo understand the role of Nr2e1 in gene expression dur-ing neocortex development, we performed hierarchicalclustering on the tag ratio values corresponding to eachof the 1279 Refseq accession numbers of differentially-regulated genes (Fig. 3, Additional file 1: Table S1 andAdditional file 2: Table S2). Tag sequences and corre-sponding tag counts of the 1279 Refseq accession num-bers were retrieved for each LongSAGE library using theDiscoverySpace 4.0 application. Fold changes from tagsSchmouth et al. BMC Genomics  (2015) 16:545 Page 3 of 19statistically differentially abundant, at least at one timepoint between the Wt and Nr2e1frc/frc libraries, were cal-culated as previously described [45], and hierarchicalclustering was performed using the Gene Cluster soft-ware as described in Methods [47]. The clustering re-sults were visualized in a heat-map display using JavaTreeView (Fig. 3a) [48]. Spearman-rank-ordering correl-ation was additionally performed on the fold changesdataset at each time point as described in Methods. Theresults demonstrated that at the E13.5 and E15.5 time-points, differential tag ratios correlated positively(Spearman R = 0.28, P < 0.001). In contrast, comparingE13.5 and E17.5, as well as E15.5 and E17.5 yielded negativecorrelation values (E13.5 vs. E17.5, Spearman R =−0.24, P< 0.001; E15.5 vs. E17.5, Spearman R = −0.23, P < 0.001)(Fig. 3b). This demonstrated that the differential-tag ratiofound between the E13.5, and E15.5 libraries were moresimilar than the one observed in the E17.5 library. This alsosuggested that Nr2e1 expression has a more similar effecton genes in early and mid-stages of neurogenesis, thanduring the switch from neurogenesis to gliogenesis occur-ring around E17.5. These results correlated with previouslypublished observations, demonstrating a progression of theNr2e1-null phenotype during neocortex development, witha greater effect between E13 and E15 [9].Bioinformatics analyses for the prediction of Nr2e1 directtargetsConsidering that Nr2e1 encodes for a transcription fac-tor, and that transcription factors regulate transcriptionby binding the promoter regions of their target genes;we hypothesized that a list of genes, differentially regu-lated between Wt and Nr2e1frc/frc, would comprise genescontaining Nr2e1 binding sites within their promoterregions. Thus, interrogation of our pooled list of 1279Refseq accession numbers was undertaken using threedifferent software tools; the ORCA toolkit (tk) to per-form the initial orthologous-sequence alignment andFig. 1 Laser microdissected LongSAGE libraries were used to reveal the transcriptome of Wt and Nr2e1frc/frc embryos. a The laser capturemicrodissection (LCM) procedure. ((a), I-IV) Embryonic day 13.5 sagittal sections stained with cresyl violet. ((a), II) The ventricular/subventricularzone (VZ/SVZ) of the dorsal lateral telencephalon cut with laser. ((a), III) The VZ/SVZ removed by LCM. ((a), IV) The VZ/SVZ of the dorsal lateraltelencephalon captured by LCM for RNA extraction. LV, Lateral Ventricle; Str, striatum. Scale bars, 100 μm. b The composition of LongSAGElibraries. Column one presents the name of the library; columns two and three, the genotype and developmental stage respectively; columnfour, the amount of RNA used as starting material; and columns five to nine, the number of tags for each library depending on the filteringcriteria usedSchmouth et al. BMC Genomics  (2015) 16:545 Page 4 of 19phylogenetic footprinting [49], a customized version ofoPOSSUM for prediction and storage of transcriptionfactor binding sites (TFBSs) (http://www.cisreg.ca/oPOSSUM/) [39, 40], and a DAVID GO term analysisto evaluate if the resulting genes were found in bio-logical processes relevant to Nr2e1 (http://david.abcc.n-cifcrf.gov/summary.jsp) [50, 51]. The “modified” versionof the oPOSSUM database used a position-weight matrix(PWM) that we designed based on the nine sequencesavailable from the literature, which were known to bebound by NR2E1 (Additional file 3: Table S3) [21, 22,52–56]. The resulting matrix and logo are depicted in Fig. 4a.The results from these sequential analyses are summa-rized in the flowchart of Fig. 4b. ORCAtk orthologous se-quence alignments between human and mouse for eachgene was initially performed; resulting in the exclusion of304 Refseq accession numbers due to poor conservationbetween human and mouse within the promoter se-quences of these genes. This resulted in 975 Refseq acces-sion numbers that were used in the modified oPOSSUMdatabase. Of these 975 accession numbers, 770 (79 %)were found to have predicted binding sites for NR2E1within their promoter regions (Fig. 4b) (Additional file 2:Table S2). [57, 58] These 770 accession numbers were fur-ther studied in a GO term enrichment analysis using theDAVID service. The 770 Refseq accession numbers werefirst converted to DAVID IDs using the DAVID knowl-edgebase, and then compared to the DAVID mouse-background list of genes [50, 59]. The enrichment resultswere visualized using the functional annotation modulebased on the relevance for each enriched gene to “bio-logical process” with an initial P value < 0.1, using themodified Fisher exact test (EASE score) [51, 60, 61]. Inthis process, 291 Refseq accession numbers werediscarded as they were not enriched in our list comparedto the mouse background (Fig. 4b). The remaining 479Refseq accession numbers were interrogated based ontheir “biological process” terms. Only terms with a Pvalue < 0.05 after multiple test correction, using the Bon-ferroni approach [59, 62], were considered interesting forfurther investigation. Table 1 shows the list of GO termspassing this criterion, Additional file 2: Table S2 list thedifferentially expressed genes within these GO terms. Asexpected, numerous terms related to cell cycle regulationwere found after performing the GO term enrichmentanalysis on the 770 Refseq list. However, terms related toFig. 2 Comparative transcriptome investigation of Wt and Nr2e1frc/frc embryos yielded hundreds of candidate-differentially-regulated genes.a Details of the number of differentially abundant tags (increased or decreased) and the corresponding number of genes between Wtand Nr2e1frc/frc embryos at each time point. In the exceptional case of a gene having both a significantly up and down tag, it wascounted in both categories. Column one presents the direction of change; column two, the embryonic day of tissue harvest; columnthree, the number of tags that had significantly different counts between Wt and Nr2e1frc/frc embryos; column four, the number of tagshaving significant different counts that mapped to genes found in the Ensembl (v66) and Refseq (v52) gene collections; and column five,the number of Refseq genes mapped by the corresponding tags. b The Venn diagram presents the number of up- and down-regulatedgenes that were exclusive or shared at each embryonic day (E13.5, E15.5, and E17.5). Bracketed numbers correspond to the total ofdifferentially-regulated genes between Wt and Nr2e1frc/frc embryos at the corresponding time pointSchmouth et al. BMC Genomics  (2015) 16:545 Page 5 of 19cell cycle regulation were also found in a similar analysis,using the initial 1279 Refseq list. This suggested that genesinvolved in cell cycle regulation were differentiallyexpressed in our LongSAGE results when comparing Wtto Nr2e1frc/frc, but were not enriched for the presence ofNR2E1 binding sites within the promoter regions. In con-trast, the term “nervous system development” (P < 0.01),with 64 differentially-regulated genes, was found to beenriched only after performing the analysis on the 770Refseq list, again, suggesting the presence of NR2E1 bind-ing sites within the promoter regions of genes enrichedfor this term. Interestingly, similar results were also ob-tained using a different GO term enrichment software;“GOstats” yielded significant results for the “nervous sys-tem development” term (P < 0.001) [63]. Thus, we subse-quently used the “nervous system development” gene listfrom the “DAVID analysis” for further investigations.Differential expression results validated by literatureWe used the tag ratio values of the 64 differentiallyexpressed genes found in the “nervous system develop-ment” GO term category to perform hierarchical cluster-ing (Fig. 5a, Additional file 2: Table S2). We used thesame hierarchical procedure as the one described for the1279 genes list. Similarly to previously obtained results,the E13.5 and E15.5 time-points, differential tag ratioscorrelated positively (Spearman R = 0.41, P < 0.001), andthe E13.5 and E17.5 yielded a negative correlation value(Spearman R = −0.34, P < 0.001). However, no signifi-cance was observed when comparing the E15.5 andE17.5 time points (Fig. 5b). This suggested that thedifferential-tag ratio found between the E13.5, and E15.5Fig. 3 Hierarchical clustering suggested a stage-specific role forNr2e1 in neocortex development. a Tag numbers for each of thecorresponding 1279 genes were retrieved from DiscoverySpaceand hierarchical-clustering was performed prior to visualization in aheat-map; displaying significant-differential-tag ratios from Wt vs.Nr2e1frc/frc libraries. The relative expression was calculated based onthe tag ratios from Wt vs. Nr2e1frc/frc libraries and corrected to accountfor library sizes; (observed tag counts/total useful tags) X 100,000. Tagshaving a count value of “0” (no expressed tags) were adjusted to avalue of “1” for fold change calculations only. Green, down; red, up;black, no difference; grey, no expressed tags. b Embryonic-stage-specificdifferences in expression profiles of the 1279 differently abundantgenes in neocortex development of Wt vs. Nr2e1frc/frc embryos wasdemonstrated using Spearman rank correlation. Spearman rankcorrelation was performed on gene lists from pairs of embryonic stagesusing STATISTICA. Results revealed a significant positive correlationbetween the lists corresponding to early and mid-neurogenic stages(E13.5 vs. E15.5); whereas significant negative correlations were obtainedwith the tag ratios corresponding to the early and mid-neurogenicstages versus the early gliogenic stages (E13.5 vs. E17.5, and E15.5 vs.E17.5). Column one presents the time-point comparisons; column two,the valid number of genes included in the analysis; columns three andfour, the corresponding R and P values obtained by Spearmanrank correlationSchmouth et al. BMC Genomics  (2015) 16:545 Page 6 of 19libraries were more similar than the one observed in theE17.5 library; highlighting again the possibility of distinctroles for Nr2e1 in the neurogenic versus early gliogenicstages of neocortex development.As expected, all the Nr2e1frc/frc libraries showed notags for Nr2e1. Interestingly, even the Wt libraries,despite being obtained by LCM for a focused region ofNr2e1 expression, showed low abundance of Nr2e1 tags(E13.5, 4; E15.5, 2; and E17.5, 0). Thus, only E13.5reached significant differential expression between Wt andNr2e1 frc/frc (−4.5 fold, P < 0.05). As expected, the numberof tags mapping to Nr2e1 in the Wt libraries showed adeclining trend (Wt E13.5 vs. Wt E17.5, P = 0.06). This isin agreement with published and publicly-availableexpression results [18] (Allen Mouse Brain Atlas,http://www.brain-map.org/); where Nr2e1 expressionhas been observed as early as E8, peaks at E13,sharply decreases until E18, and is barely detectablein new-born brains [18]. Hence, at the time point oflowest Nr2e1 expression (E17.5) the LongSAGEapproach was insufficiently sensitive to detect this lattergene transcript. Interestingly, our bioinformatics enrich-ment analysis included Nr2e1 in the list of genes with pre-dicted NR2E1 binding sites within their promoter regions,adding support to previous observations proposing a self-regulating mechanism for Nr2e1 [22, 64].When analysing large-scale transcriptome-profilingdatasets, the overall level of expression is an importantfactor influencing the outcome of statistical significance.In our LongSAGE libraries, Pten and P21 (Cdkn1a), twodirect targets of Nr2e1 [11, 14, 21], were expressed atlow levels in the VZ/SVZ (total number of tags, Pten:E13.5, Wt 1, Nr2e1frc/frc 5; E15.5, Wt 1, Nr2e1frc/frc 1; andFig. 4 Novel implementation of three computational methods togenerate a focused list of biologically-relevant Nr2e1-candidate-targetgenes. a Data mining from the literature allowed generation of aposition-weight matrix representing the binding properties of NR2E1.The matrix and resulting logo are presented. b Flow chart describingthe novel implementation of three computational methods togenerate a focused list of biologically-relevant Nr2e1-candidate-targetgenes. DiscoverySpace was used to generate a compiled list of 1279Refseq accession numbers, corresponding to genes differentiallyregulated between Wt and Nr2e1frc/frc embryos. A customizedoPOSSUM database of predicted conserved TFBS was created byfirst aligning all orthologous human-mouse genes using ORCAtk.During this alignment, 304 of the 1279 differentially regulatedgenes were excluded due to a lack of ortholog information or poorORCAtk alignment quality. Then the remaining 975 genes withconserved promoter regions were scanned with the NR2E1 matrix(as well as all the vertebrate matrices from the JASPAR CORE collectionof transcription factor binding site profiles). Of the 975 scanned genes,770 contained conserved NR2E1 binding sites. These 770 Refseqaccession numbers were submitted to a gene ontology (GO) termanalysis using DAVID. From the DAVID analysis, 479 Refseq accessionnumbers were found to be enriched in GO term categories related tobiological processesSchmouth et al. BMC Genomics  (2015) 16:545 Page 7 of 19E17.5, Wt 0, Nr2e1frc/frc 1; P21: E13.5, Wt 0, Nr2e1frc/frc0; E15.5, Wt 1, Nr2e1frc/frc 3; and E17.5, Wt 1, Nr2e1frc/frc2) and thus did not reach significance in terms of differ-ential expression between Wt vs. Nr2e1frc/frc libraries. Incontrast, Nestin, a common marker of proliferatingneural progenitors, which was expressed at mid to highlevels, was significantly down regulated in Nr2e1frc/frc atE13.5 when compared to Wt (−7.3 fold, P < 0.05) (Fig. 5a,Additional file 1: Table S1). This correlated with the pre-viously published observation of reduced numbers ofNestin-positive cells in the VZ of Nr2e1-null embryos atE14.5 [11]. In addition, our data suggests that the mech-anism involves a direct-up regulation by Nr2e1 in Wt, asNestin was found within the bioinformatics enrichmentanalysis genes with predicted NR2E1 binding sites. An-other example of our expression results being supportedby the literature is the basic helix-loop-helix (bHLH)gene Neurog2, which was significantly down regulated inNr2e1frc/frc embryos at both E13.5 and E15.5 when com-pared to Wt (E13.5, −2.8 fold, P < 0.001; E15.5, −5.5 fold,P < 0.001) (Fig. 5a, Additional file 1: Table S1). Theseresults correlated with the previously published observa-tions of reduced expression of Neurog2 in double mu-tants embryos for Pax6 and Nr2e1 in the rostralgranular zone during neocortex development [65].Disruptions in Neurog2 expression are also characteristicof alterations in the pallio-subpallial boundary observedin Nr2e1-null embryos [66]. Additionally, downstreamcandidate genes of the pathway regulated by Neurog2(i.e. Neurod2, and Tbr1) were found differentiallyexpressed in our LongSAGE comparison analysis,arguing in favour of a direct role for Nr2e1 in regulat-ing this specific pathway during neocortex develop-ment (Fig. 5a, Additional file 1: Table S1) [65].TFBS overrepresentation analysis revealednovel-candidate-NR2E1 co-interactorsSpatial-temporal gene expression is, in general, regulatedby the dual ability of transcription factors to bind specificDNA sequences and to form complexes with other regula-tory proteins. NR2E1 has previously been shown to medi-ate gene regulation with co-interacting partners; formingregulatory complexes that lead to either direct-target-generepression or activation [14, 21, 25, 26]. Interestingly, nu-clear receptors have also been shown to mediate generegulation via interaction with other transcription factorsas co regulators [64, 67, 68]. Based on our Spearman rankordering results, we hypothesized that the striking differ-ence in direction of correlation for differentially abundanttags between the E13.5-E15.5, E13.5-E17.5, and E15.5-E17.5 time points, was largely due to the presence ofdifferent Nr2e1 interacting partners at different times indevelopment. To discover novel candidate co-interactorsof Nr2e1, we designed a computational experiment toidentify TFBS within the vicinity of the predicted NR2E1binding sites for each differentially-regulated gene foundin the GO term category “nervous system development”.The identified binding sites were then scored for theirenrichment compared to a randomized list of genes,thereby generating both a Z and Fisher score. PotentialTFBSs having a Z-score value > 10 and a Fisher scorevalue < 0.01 were considered enriched and kept for fur-ther characterization as candidate-NR2E1 co-interactors(Table 2). We further ascertained the significance of ourcandidate-NR2E1 co-interactors list by performing ana-lyses on random sets of 64 genes extracted from the initiallist of 770 genes obtained through the oPOSSUM-NR2E1binding motif interrogation step. Corresponding empir-ical P values based on the Z-scores and Fisher scores ofeach of the candidate-NR2E1 co-interactors wereextracted from the random sets of 64 genes and aresummarized in Additional file 4: Table S4.Table 1 Gene ontology (GO) term analysis revealed enrichmentin relevant biological processesGO identifiers and terms No. of differentiallyexpressed genesaP value afterBonferronibGO:0006396 ~ RNA processing 55 4.39E-10GO:0016070 ~ RNA metabolic process 68 7.54E-09GO:0044267 ~ cellular proteinmetabolic process132 2.96E-04GO:0007067 ~mitosis 26 3.12E-04GO:0000280 ~ nuclear division 26 3.12E-04GO:0015031 ~ protein transport 56 3.58E-04GO:0045184 ~ establishment ofprotein localization56 4.56E-04GO:0000087 ~ M phase of mitoticcell cycle26 4.66E-04GO:0046907 ~ intracellular transport 42 6.08E-04GO:0030163 ~ protein catabolicprocess49 1.13E-03GO:0044265 ~ cellular macromoleculecatabolic process52 1.27E-03GO:0000279 ~ M phase 31 2.35E-03GO:0051246 ~ regulation of proteinmetabolic process35 3.82E-03GO:0007399 ~ nervous systemdevelopmentc64 5.13E-03GO:0022403 ~ cell cycle phase 33 6.36E-03GO:0019538 ~ protein metabolicprocess151 6.91E-03GO:0006886 ~ intracellular proteintransport28 3.18E-02GO:0034613 ~ cellular proteinlocalization29 4.95E-02a352 total non-overlapping differentially expressed genesbhttp://david.abcc.ncifcrf.gov/ease/Help/Technical%20details/Overrepresentation%20analysis.htmcterm used in subsequent analysesSchmouth et al. BMC Genomics  (2015) 16:545 Page 8 of 19The relevance of these enriched TFBS was alsoevaluated based on the expression pattern of theircorresponding transcription factors. For this, we pri-marily used data from the Allen Mouse Brain Atlas(http://www.brain-map.org/) at three different timepoints (E13.5, E15.5, and E18.5), and included datafrom other publically-available resources as required(Table 2) [69].The expression data and statistical scores obtainedmost strongly supported the biological relevance ofSOX9 (Z-score: 16.97, empirical P value: < 0.001; Fisherscore 1.58E-05, empirical P value: = 0.001); a member ofthe SRY-box family. Examination of our LongSAGE datarevealed the presence of tags corresponding to Sox9throughout the three different time points for both ge-notypes (data not shown). Additionally, Sox9 has beenreported to function in neural-stem/progenitor-cellregulation; as does Nr2e1. Together these data supportthe hypothesis that SOX9 acts as a co-interactor ofNR2E1 [70]. Thus, the differentially-regulated genesfound in the GO term category “nervous system devel-opment”, and the number of predicted TFBSs for bothNR2E1 and SOX9 in the promoter region of these genes,are presented in Table 3 (Additional file 2: Table S2).These 40 genes represent a rich resource for the bio-logical examination of Nr2e1 downstream targets. Herewe pursue the top candidate Lhx2, a LIM-homeoboxtranscription factor. Lhx2 had the highest number ofpredicted binding sites for both NR2E1 and SOX9; 35Fig. 5 Hierarchical clustering suggested a stage-specific role forNr2e1 in nervous system development. a Tag numbers for eachof the 64 genes were retrieved from DiscoverySpace andhierarchical-clustering was performed prior to visualization in aheat-map; displaying significant-differential-tag ratios from Wtvs. Nr2e1frc/frc libraries. The relative expression was calculatedbased on the tag ratios from Wt vs. Nr2e1frc/frc libraries andcorrected to account for library sizes; (observed tag counts/total useful tags) X 100,000. Tags having a count value of “0”(no expressed tags) were adjusted to a value of “1” for foldchange calculations only. Green, down; red, up; black, nodifference; grey, no expressed tags; bold, key genes in thisanalysis. b Embryonic-stage-specific differences in expressionprofiles of “nervous system development” genes in neocortexdevelopment of Wt vs. Nr2e1frc/frc embryos was demonstratedusing Spearman rank correlation. Spearman rank correlationwas performed on gene lists from pairs of embryonic stagesusing STATISTICA. Results revealed a significant positivecorrelation between the lists corresponding to early andmid-neurogenic stages (E13.5 vs. E15.5); whereas a significantnegative correlation was obtained with the tag ratioscorresponding to the early neurogenic and early gliogenicstages (E13.5 vs. E17.5). No significance was obtained betweenthe mid-neurogenic and early gliogenic stages (E15.5 vs. E17.5).Column one presents the time-point comparisons; column two,the valid number of genes included in the analysis; columnsthree and four, the corresponding R and P values obtained bySpearman rank correlationSchmouth et al. BMC Genomics  (2015) 16:545 Page 9 of 19Table 2 SOX9 revealed as a candidate co-interactor of Nr2e1 for genes of the “nervous system development”TranscriptionfactorsNo. ofbackground hitsNo. of backgroundnon-hitsNo. oftarget hitsNo. of targetnon-hitsZ score FisherscoreABA Expressionat E13.5ABA Expressionat E15.5ABA Expressionat E18.5Other expressionresourcesExpressionpattern scoreSP1 190 310 41 23 20.40 6.53E-05 Not available Not available Not available Ubiquitousa ++Nobox 248 252 47 17 17.70 2.13E-04 Weak, Ubiquitous Moderate, Ubiquitous Weak Not applicable +SOX9 172 328 40 24 16.97 1.58E-05 Moderate,VZ/SVZModerate,VZ/SVZStrong, VZ/SVZ Not applicable +++Arnt-Ahr 243 257 48 16 15.76 4.38E-05 Weak, Ubiquitous Moderate, Ubiquitous Moderate, Ubiquitous Not applicable ++Nkx2-5 309 191 58 6 15.12 7.92E-07 Not available Not available Not available Absentb -Gfi1 199 301 42 22 14.27 7.74E-05 Not available Not available Not available Strong, ubiquitousc ++Lhx3 74 426 20 44 14.00 1.58E-03 Weak Weak Weak Not applicable +TAL1-TCF3 93 407 26 38 13.60 1.22E-04 Moderate,UbiquitousModerate,UbiquitousWeak, Ubiquitous Not applicable +NHLH1 61 439 22 42 13.57 2.00E-05 Strong, Neocortex Weak, Neocortex Weak Not applicable ++Myb 211 289 46 18 13.49 5.86E-06 Weak, Neocortex Weak, Neocortex Weak Not applicable ++Roaz (Zfp423) 59 441 19 45 12.99 3.31E-04 Not availablea Not availablea Not available Strong, ubiquitousd ++FOXI1 189 311 36 28 12.27 3.72E-03 Weak Weak Weak Not applicable +Prrx2 296 204 53 11 12.26 1.13E-04 Weak Weak Weak Not applicable +Cebpa 131 369 37 27 11.92 6.32E-07 Weak Weak Weak, Neocortex Not applicable +Foxa2 177 323 35 29 11.38 2.38E-03 Weak, Ubiquitous Weak, Ubiquitous Weak, Ubiquitous Not applicable +NFYA 68 432 18 46 11.36 3.60E-03 Moderate,UbiquitousModerate,UbiquitousModerate, Ubiquitous,NeocortexNot applicable ++Sox17 237 263 48 16 11.06 2.06E-05 Weak Weak Weak Not applicable +Sox5 269 231 51 13 10.61 4.18E-05 Strong, Neocortex Strong, Neocortex Strong, Neocortex Not applicable ++SRY 267 233 48 16 10.11 6.56E-04 Weak Weak Weak Not applicable +MYC-MAX 32 468 12 52 10.07 1.87E-03 Moderate,UbiquitousModerate,UbiquitousModerate, Ubiquitous,NeocortexNot applicable ++Bold, most relevant transcription factor according to both the statistical and expression pattern scores; Not available, no expression data available in the ABA (http://www.brain-map.org/, accessed January 10th 2014);Not applicable, data available in ABA so no need to access other resourcesaGray et al. (PMID: 15618518) showed expression in the central nervous system, including the brain ventricular layers at E13.5bExpression restricted to the developing striatum at E14.5 according to Eurexpress (http://www.eurexpress.org/)cWallis et al. (PMID: 12441305) showed expression in the developing forebrain at E12.5dExpression strong and ubiquitous in the developing brain according to Eurexpress and GenePaint at E14.5 (http://www.genepaint.org/Frameset.html)Schmouthetal.BMCGenomics (2015) 16:545 Page10of19and 13 respectively. Visualization of the predicted bind-ing sites within the promoter region of Lhx2 revealed aclustered distribution that was located within highlyconserved DNA (Fig. 6). Localization within conservedDNA further suggested a function for these binding sitesthroughout evolution. Evidence from the literature high-lights a spatial-temporal dynamic role for Lhx2 in thedeveloping forebrain. Early in development (E10.5-E11.5), Lhx2 has been shown to work as a fate determin-ant of cortical identity [71]. Later in development,distinct roles have been described for Lhx2 dependingon the forebrain structures involved; including a role inregulating progenitor cell differentiation in neocortexdevelopment (E11.5-E13.5) and a role in the neurogenicto gliogenic switch in hippocampal development(E14.5-E15.5) [72, 73]. Thus, our data, combined withthe literature, support Lhx2 as a direct target of co-regulation by NR2E1 and SOX9.Differential expression of the transcription factor Lhx2validated our LongSAGE resultsTo expand our understanding of the relationship be-tween Nr2e1 and Lhx2, and simultaneously further val-idate the results obtained from the LongSAGE taglibraries, we undertook two biological assays; one eachin vitro and in vivo. First we retrieved with Discovery-Space 4.0 the LongSAGE tag sequence mapping to Lhx2,and the corrected number of tags from each library(Fig. 7a). This showed that Lhx2 levels were significantlyincreased in Nr2e1frc/frc libraries at two different timepoints (E13.5, and E15.5).For in vitro studies, we chose a method of neurogenesisfrom adherent-monoculture of ESC, which sequentiallymimics the development of cortical neurons over thecourse of 21 days of differentiation [74]. In this system:neural induction starts at day 0 (d0) of differentiation;neurogenesis starts at d6 with the generation of subplateneurons and deep layer neurons between d7 and d9,followed by upper cortical neurons around d12; finallythere is a wave of gliogenesis by d21 [75]. The formationof subplate neurons corresponds to in vivo E10.5-E13.5,deep layer neurons to E11.4-E14.5, and upper corticalneurons to E13.5-E16.5 [75]. Hence, this culture systemencompass the key time periods for the function of Nr2e1in brain development. Using this method, we first detectedNr2e1 expression at d6, which then increased and peakedat d12 (data not shown). Further investigation using qRT-PCR at this latter time point not only showed a significantdifference between Wt and Nr2e1frc/frc cells for the Nr2e1gene (Fig. 7b), but also demonstrated a significant increasein expression in Nr2e1frc/frc cells when compared to Wtcells for the Lhx2 gene (P < 0.01) (Fig. 7c) [75]. This resultwas consistent with a model of Lhx2 being a direct targetof, and repressed by, Nr2e1.Table 3 Overrepresentation analysis revealed candidate-direct-target genes of Nr2e1Gene name No. of NR2E1binding sitesNo. of SOX9binding sitesLhx2 35 13Ppp1r9a 30 2Gap43 29 3Myh10 23 2Fezf2 20 3Cux1 19 9Ppp1r9b 19 1Epha4 18 4Nr2e1 17 4Slc1a3 17 1Atrx 15 2Neurod6 14 4Rgma 14 3Mtap1b 13 5Kif5c 13 3Bzw2 13 2Cntn2 12 1Tbr1 11 1Dpysl2 10 2Msx1 9 3Efnb2 9 2Neurod1 9 2Rufy3 9 2Apc 9 1Notch1 8 3Hes6 8 2Racgap1 8 2Edg1 (S1pr1) 7 1Id3 6 3Otx1 6 3Elavl3 6 1Pfn1 5 1Neurog2 4 1Sema4g 4 1Tubb3 3 2Dner 3 1Mtpn 3 1Sept4 3 1Hes5 2 2Sox11 2 1Schmouth et al. BMC Genomics  (2015) 16:545 Page 11 of 19For in vivo studies, we examined the expression pat-tern of the Lhx2 protein by immunofluorescence in Wtand Nr2e1frc/frc E15.5 embryos. The results showed simi-lar Lhx2 protein localization when comparing Wt andNr2e1frc/frc embryos; along the VZ/SVZ of the developingforebrain. Furthermore, for both Wt and Nr2e1frc/frc, ex-pression levels varied from high in the medial pallium tolow in the dorsal pallium (Fig. 7d). However, relativequantification of Lhx2 between Wt and Nr2e1frc/frc,along the VZ/SVZ of the dorsal-lateral telencephalon, re-vealed a significant increase of Lhx2 protein in theNr2e1frc/frc embryos when compared to Wt (P < 0.01)(Fig. 7e). Thus, the significant increase at the mRNAlevel for the Lhx2 gene results in a significant increase atthe protein level along the VZ/SVZ of the dorsal-lateraltelencephalon at E15.5. These data add further supportto the hypothesis that Nr2e1 directly-negatively regulatesLhx2 expression in the dorsal-lateral telencephalon dur-ing development.ConclusionsIn this work, we have generated a list of 1279 genes thatare differentially expressed in response to altered Nr2e1expression during in vivo neocortex development; thislist was a critical part of our own studies, but is also animportant resource for others (Additional file 1: Tableand Additional file 2: Table S2). To create this list, weprofiled the transcriptomes of Wt and Nr2e1frc/frc em-bryos by generating LongSAGE libraries through LCMof the VZ/SVZ from the dorsal-lateral telencephalon. Tofurther focus the work on the role of Nr2e1 during neo-cortex development, we chose two time points thatspanned the early to mid-neurogenic stages (E13.5,E15.5), and one time point corresponding to the earlyswitch from neurogenesis to gliogenesis (E17.5). Thus,from six LongSAGE libraries we identified 1279 candi-date genes comprising both direct and indirect targets ofNr2e1 during neocortex development. This list can nowbe mined by us, and many other groups for the antici-pated numerous co-suppressors, co-activators, and dir-ect targets making up the molecular mechanisms of thenuclear-receptor transcription-factor Nr2e1.We have further refined this list of 1279 differentiallyexpressed genes, culminating in a focused list of 64 can-didate direct-targets of NR2E1 binding during nervoussystem development, for our own studies and as a re-source for others (Additional file 2: Table S2). This wasaccomplished by performing two sequential analyses; 1)a TFBSs prediction analysis, using oPOSSUM, to identifynovel direct targets of Nr2e1, and 2) a GO term overrep-resentation analysis, to extract biological meaning fromthe latter generated list. This procedure included thegeneration of a novel NR2E1 PWM based on availableinformation from the literature (Additional file 3: TableS3); the derived matrix and logo are provided (Fig. 4a).We used this matrix, in combination with the Long-SAGE results, in a bioinformatic experiment to identifynovel direct-target genes of Nr2e1. The resulting list ofGO terms coming from this analysis (Table 1) containedgenes differentially expressed, and predicted to containNR2E1 binding sites within their promoter regions(Additional file 2: Table S2). The GO term category“nervous system development” contained 64 such genes(Fig. 5a, Additional file 2: Table S2); a list that was usedin subsequent analyses.Our approach suggested distinct roles for Nr2e1 dur-ing different neocortex developmental stages. Analysesperformed on the differential-tag ratio for the 1279Refseq accession numbers of differentially-regulatedgenes retrieved from the Wt and Nr2e1frc/.frc libraries, re-vealed a positive correlation of the differential abun-dance at E13.5 and E15.5, whereas a negative correlationwas obtained when comparing the two previous timepoints to E17.5 (Fig. 3). Thus, the differential-tag ratiosFig. 6 Lhx2 contains enriched clusters of NR2E1 and SOX9 binding sites in highly conserved regions. Presented is the ~17 kb of genomicsequence 5′ to, and surrounding, the Lhx2 transcription start site. NR2E1 binding sites are indicated in black, and SOX9 binding sites are indicated inred. NR2E1 binding site coordinates were derived from the position-weight matrix (PWM) described in the current manuscript; SOX9 binding sitescoordinates were derived from the PWM stored in the JASPAR database (http://jaspar.binf.ku.dk/cgi-bin/jaspar_db.pl). Conservation, alignment of 100vertebrate species from the University of California, Santa Cruz genome browser (UCSC: http://genome.ucsc.edu/); thick horizontal line, first exon ofLhx2; arrow, transcription start site and senseSchmouth et al. BMC Genomics  (2015) 16:545 Page 12 of 19found at E13.5 and E15.5 were more similar to eachother than when compared to E17.5. From E13.5 toE17.5, the neocortex undergoes drastic changes,including the formation of the SVZ, a layer of cells be-ing seeded by the VZ, and a progressive switch fromneurogenesis to gliogenesis. Our results indicate thatFig. 7 Validation of the differential abundance of Lhx2 LongSAGE tags in Wt vs. Nr2e1frc/frc embryos. a The tag count results, mapping to Lhx2 atthe three different embryonic days, are presented. Columns one to three present the tag sequence, accession number, and gene symbolcorresponding to Lhx2; columns four to six, the corrected tag numbers recovered from DiscoverySpace in both Wt and Nr2e1frc/frc embryos ateach time point (E13.5, 15.5, and 17.5); column seven, the fold change between the tag numbers corresponding to Lhx2 found in Wt andNr2e1frc/frc embryos at each time point; and column eight, the associated P values obtained using the Audic-Claverie statistical method. Accordingto this approach, Lhx2 expression level is significantly upregulated at both E13.5 and E15.5 in Nr2e1frc/frc embryos. b Wt embryonic stem cells(ESC) submitted to a neurogenic differentiation protocol demonstrated expression of Nr2e1 at 12 days of differentiation (d12) whereas, asexpected, Nr2e1frc/frc ESC did not express Nr2e1 (*, P < 0.001). c Quantitative RT-PCR reveals that the Lhx2 mRNA level is upregulated by ~3.6 foldin Nr2e1frc/frc ESC at d12 compared to Wt ESC (*, P < 0.01). d Immunofluorescence using an anti-Lhx2 antibody (green) demonstrated a similarexpression pattern for the Lhx2 protein along the ventricular/subventricular zone (VZ/SVZ) of the lateral telencephalon in E15.5 Nr2e1frc/frc embryoscompared to Wt. White arrows, medial pallium; red arrows, dorsal pallium; scale bar, 200 μm. e Lhx2 protein level was increased by ~1.3 foldalong the VZ/SVZ of the lateral telencephalon in E15.5 Nr2e1frc/frc embryos compared to Wt (*, P < 0.05). (c), (d), (e) Sample Student’s t-test wereperformed, N = 3; error bars, standard error of the meanSchmouth et al. BMC Genomics  (2015) 16:545 Page 13 of 19Nr2e1 has a more similar effect in the early stages ofneurogenesis (E13.5 and E15.5) compared to laterstages when the switch from neurogenesis to gliogen-esis occurs. The mechanism driving these changes maydepend on the level of Nr2e1 expression, which peaksat E13.5 and gradually decreases until birth [18].The SOX9 transcription factor may be an importantco-interactor of NR2E1 in regulating numerous targetgenes during nervous system development. A co-factoranalysis revealed enrichment for binding sites predictedto be bound by SOX9 within the vicinity of the predictedNR2E1 binding sites (Table 2); results that remained sig-nificant after calculating empirical P values on our list ofcandidate co-interactors (Additional file 4: Table S4). Inaddition, the spatial, temporal, and strength of expres-sion of Sox9 strongly supported a biological relationshipwith Nr2e1 [70]. Interestingly, others have shown thatSox9 may be involved in the acquisition of gliogeniccompetence of neural stem/progenitor cells during cen-tral nervous system (CNS) development [76]. Togetherthese data suggest that co-interaction between the SOX9transcription factor and NR2E1 may regulate the expres-sion of 40 of the 64 genes involved in nervous systemdevelopment (Additional file 2: Table S2).The Sox family of transcription factors may generallybe important co-interactors of Nr2e1 in regulating targetgenes during nervous system development. This familycomprises 20 genes with several members expressed inneural stem/progenitor cells of the CNS, and peripheralnervous system [77–79]. They have been shown to actas either transcriptional activators or repressors by bind-ing to similar (A/T)(A/T)CAA(A/T)G DNA motifs [78].Recently, one of these family members, Sox2, has beenshown to form a regulatory complex with Nr2e1 in adultNSC [64]. Interestingly, our co-factor TFBSs analysis re-vealed enrichment for the presence of associated bindingsites not just for SOX9, but also three additional Soxfamily members (Sox17, Sox5, and SRY) within thevicinity of predicted NR2E1 binding sites for genes ofthe “nervous system development” GO term category(Table 2). These additional Sox family members alsoshowed expression overlap with Nr2e1, and at least oneof these members, Sox5 has been shown to bind toFezf2-conserved-enhancer sequences, resulting in a dir-ect repression of Fezf2 in neocortex development [80].LongSAGE tags mapping to Sox5 were found in our li-braries and Fezf2 was found significantly upregulated inthe Nr2e1frc/frc library at E15.5 (7.4 fold, P < 0.05, Fig. 5a).Hence our data suggests a specific testable hypothesis bywhich Nr2e1 potentially regulates Fezf2 expressionthrough its interaction via the Sox5 protein in neocortexdevelopment. In conclusion, our data supports the hy-pothesis that generally the Sox family members play animportant role as co-interactors of NR2E1.Lhx2 may be an important direct-target gene ofNr2e1, with SOX9 as a co-interactor. In Nr2e1-nullembryos, premature neurogenesis has been reportedto occur from E9.5 to E14.5 in both the dorsal andventral telencephalon [9]. Overexpression of Lhx2 hasbeen reported to prolong neurogenesis in hippocam-pal development, resulting in generation of neuronsfrom progenitors that would normally produce astro-cytes [72]. Additionally, conditional inactivation ofLhx2 in neocortical development affects the fate ofPC, resulting in a phenotype highly similar to thatobserved in Nr2e1-null embryos; with a reduction inthe number of PC populating the VZ and prematureneurogenesis in the neocortex of Lhx2-null embryos[73]. This latter phenomenon appears to involve thenotch signalling pathway, with a downregulation ofHes1 being observed along the VZ of Lhx2-null em-bryos and aberrant expression of the Notch encodinggene along the medial to lateral dorsal telencephalon[73]. Notch pathway genes such as Notch1, Hes5, andHes6 were also found differentially regulated in theNr2e1frc/frc library when compared to Wt in ourLongSAGE analysis (Notch1, E13.5, −6.8 fold, P < 0.05;Hes5, E13.5, −6.8 fold, P < 0.01; E17.5, −10.3 fold, P < 0.01;Hes6, E13.5, 9 fold, P < 0.001) (Fig. 5a, Additional file 1:Table S1 and Additional file 2: Table S2). Protein regula-tory networks in NSC have been demonstrated to behighly dosage dependent. For instance, phenotypic ana-lyses of Pax6 in gain- or loss-of-function mutant corticeshave shown similar phenotypic outcome, with both moreor less of the protein resulting in increased neurogenesisthroughout development [81]. Hence, akin to Pax6, ourcurrent results highlight a testable hypothesis in whichpremature neurogenesis observed in Nr2e1frc/frc em-bryos [9] could be due to the upregulation of Lhx2protein along the VZ/SVZ of the dorsal telenceph-alon; a phenomenon that most likely includes theconcerted effect of deregulation of other Notch path-way encoding genes. Our analysis also predicted thatthe transcription factor pathway regulated by NR2E1involves interaction with SOX9, which has beenshown to be involved in the acquisition of gliogeniccompetence of neural stem/progenitor cells duringCNS development [76]. Hence, our results highlightyet another testable hypothesis for the discovery of apossible novel pathway by which Nr2e1 regulates neuro-genesis, which includes Lhx2 as one of the direct-targetgenes, and SOX9 as a co-interactor.MethodsEthics statementAll procedures involving animals were in accordancewith the Canadian Council on Animal Care and UBCAnimal Care Committee (Protocol A11-0412).Schmouth et al. BMC Genomics  (2015) 16:545 Page 14 of 19LongSAGE libraries generationLibraries were generated from tissue samples obtainedby LCM of dorsal VZ/SVZ from the telencephalon ofwild-type (Wt) and Nr2e1frc/frc embryos at E13.5, E15.5,and E17.5 as described by us previously [31]. Briefly, oneembryo per genotype at each developmental time pointwas sectioned at 20 μm thickness to generate the tissuesamples. Sections from each embryo underwent LCM,and the isolated tissue was pooled and RNA extractedusing an RNeasy Micro Kit (Qiagen Inc., Mississauga,Ontario, Canada). The LongSAGE-lite method was usedto construct the libraries using 15 to 86 ng of highquality RNA from each embryo [31, 37, 82]. Each li-brary was sequenced to a depth of >100,000 raw tagsand the processed data is accessible on the MouseAtlas of Gene Expression project website (http://www.mouseatlas.org/) and the NIH SAGEmap data re-pository http://www.ncbi.nlm.nih.gov/projects/SAGE/) [83].LongSAGE data analysisLongSAGE libraries were analysed using the Discovery-Space 4.0 application (http://www.bcgsc.ca/platform/bioinfo/software/ds) [44]. The library data were electron-ically filtered based on procedures previously describedby us [45, 84]. Briefly, duplicated ditags (identical copiesof a ditag) and singletons (tags counted only once) wereretained for analysis. Sequence data were filtered for badtags (tags with one N-base call), and linker-derived tags(artefact tags). Only tags with a sequence quality factorgreater than 99 % were included in the analysis. Se-quence tag comparisons between Wt and Nr2e1frc/frc li-braries were performed and a P-value cutoff < 0.05 usingthe Audic-Claverie statistical method was used [46].LongSAGE tags exhibiting differential expression levelswere mapped to transcripts in the NCBI ReferenceSequence (Refseq) collection (version 52, released March8th 2012) and Ensembl gene collection (version 66, re-leased February 2012).NR2E1 binding site profile constructionNo position-weight matrix (PWM) was available in pub-lic databases to model NR2E1 transcription factor bind-ing site (TFBS) specificity. Thus, a literature review wasconducted and the sequences reported to be bound byNR2E1 were compiled. Next, the MEME motif discoverytool (http://meme-suite.org/) was applied, with defaultparameter settings, to identify a DNA sequence patternwithin the data [85].oPOSSUM promoter analysisA pooled list of RefSeq accession numbers for tran-scripts exhibiting differential expression between Wtand Nr2e1frc/frc genotypes, at least at one of the threedifferent time points, was subjected to an oPOSSUMTFBS analysis. The oPOSSUM software was runusing default settings with both the constructedNR2E1 PWM and the JASPAR core vertebrate PWMcollection (http://www.cisreg.ca/oPOSSUM/) [39, 40].Briefly, for each Refseq accession number, oPOSSUMautomatically retrieved the genomic DNA sequencesaround annotated transcription start sites (TSS) inEnsembl (plus 5000 bp of both upstream and down-stream sequence), performed an alignment of theorthologous sequences (human to mouse), and ex-tracted non-coding DNA sequences that are con-served above a predefined threshold (default: top10 % of conserved regions, minimum conservation70 %). oPOSSUM results include the positions ofpredicted TFBSs, and the scores of the sites.GO term enrichment analysisRefseq accession numbers for those genes predictedto contain NR2E1 binding sites in the oPOSSUMdatabase were submitted to the DAVID service(http://david.abcc.ncifcrf.gov/summary.jsp) for GOterm annotation enrichment analysis [50, 51]. The Refseqidentifiers were converted to DAVID identifiers (IDs),using the DAVID knowledgebase [59]. GO biological-process term enrichment was assessed relative to the en-tire set of mouse genes as provided by DAVID. Resultswere filtered to exclude those enriched GO terms associ-ated with 2 or less submitted genes. A significant P-valuethreshold was applied using a multiple testing correc-tion (Bonferroni, P value < 0.05).Clustering-correlationExpression clustering was performed on thedifferential-tag ratios of the initial list of 1279 differen-tially expressed genes, and the genes annotated withinthe enriched GO term “nervous system development”category using the Gene Cluster software [47]. Hier-archical clustering was performed on both the gene listand the embryonic stages using Spearman correlationwith complete linkage clustering. Tag counts were cor-rected to account for library sizes; (observed tagcounts/total useful tags) X 100,000, and tags having acount value of “0” (no expressed tags) were adjusted toa value of “1” for fold change calculations only. Spear-man rank correlation analyses on the differential-tag ra-tios were performed using STATISTICA 12.0 (Statsoft,Inc., Tulsa, OK, USA). For these latter analyses, geneswere considered as valid when their differential-tag ra-tios between Wt and Nr2e1frc/frc were not equal to zero.Co-factor TFBS enrichment analysis and transcriptionfactor candidate evaluationA customized bioinformatics analysis, based on the oPOS-SUM combination site analysis feature, was performed toSchmouth et al. BMC Genomics  (2015) 16:545 Page 15 of 19identify TFBS patterns that were significantly enriched inthe vicinity of predicted NR2E1 binding sites for the 64candidate genes found in the “nervous system develop-ment” GO term category. Sites within 100 bp of predictedNR2E1 binding sites were retrieved from the oPOSSUMdatabase. Sites overlapping an NR2E1 motif were ex-cluded. Both the NR2E1 sites and proximal sites were sub-ject to the default oPOSSUM parameters of conservationlevel (top 10 % of conserved regions with a minimum per-centage identity of 70 %), threshold level (default matrixscore threshold of 80 %), and search region level (5000 bpupstream and downstream of TSS). The analysis was per-formed against a background of 500 genes selected ran-domly from the oPOSSUM database. Over-representationresults were considered significant based on a Z score(>10) and a Fisher score (<0.01) according to the literature[40]. To further validate the significance of these results,additional oPOSSUM co-factor analyses were performedon 1000 sets of 64 genes selected randomly from the listof 770 genes enriched for GO terms, using the same ana-lysis parameters and the same set of 500 backgroundgenes as the one described above for the “nervous systemdevelopment” gene set. The significance of the Z andFisher score for each of the co-factors was deter-mined by empirical P value, computed as “n/N”where “n” is the number of times the Z and Fisherscore from set of the random trials for the co-factorwas more significant than the Z and Fisher scorefrom the “nervous system development” set for thatco-factor, divided by the total number “N” of randomtrials (in this case 1000). These results are shown inAdditional file 4: Table S4.Transcription factors with enriched binding site predic-tions were additionally assessed for their expression pat-tern at E13.5, 15.5, and 18.5 using images from theAllen Mouse Brain Atlas (ABA, http://www.brain-map.org/) [69]. Expression results from transcription fac-tors with enriched binding site predictions that were un-available from the ABA were evaluated using otherpublicly available resources; Eurexpress (http://www.eur-express.org/ee/), GenePaint (http://www.genepaint.org/Frameset.html), and the primary literature [86, 87]. Theexpression pattern was summarized according to the spe-cificity and strength of the expression along the VZ/SVZ,and other forebrain structures. The relevance of the ex-pression pattern for each transcription factor was scoredas absent (−), low (+), moderate (++), or high (+++);where absence of VZ/SVZ expression, or ubiquitous ex-pression in the entire embryo forebrain, was scored as +,whereas strong and restricted expression along theVZ/SVZ was scored as +++. The transcription factorhaving a high score (+++) was retained as the most in-teresting candidate. For the highest-scoring transcrip-tion factors we cross-validated the expression patternby looking for the number of corresponding tags inthe LongSAGE libraries.Embryos preparationTimed-pregnant mice were euthanized by cervical dis-location, and embryos at E15.5 were dissected andfixed in 4 % paraformaldehyde (PFA) with 0.1 M PObuffer (0.1 M Na2HPO4, pH 8.0) for 6 h at 4 °C. Theembryos were then cryoprotected as described in theliterature, and embedded in optimal cuttingtemperature (OCT) compound (Tissue-tek, Torrance,California, USA) on dry ice [11]. Embryos were sec-tioned at 20 μm using a Cryo Star HM550 cryostat(MICROM International, Kalamazoo, Michigan, USA),and mounted for immunofluorescence.Immunofluorescence and imaging analysisFor antibody staining, 20 μm sagittal cryosections fromembryos were rehydrated in sequential washes of 1xphosphate buffered saline (PBS), permeabilized in PBSwith 0.3 % triton, and blocked with 1 % BSA (bovineserum albumin) in PBS triton 0.3 % for 1 h at roomtemperature. Goat anti-Lhx2 primary antibody (1:1000)(Santa Cruz, Dallas, TX, USA, sc-19344) was incubatedovernight at 4 °C. Rabbit anti-goat Alexa 488 (1:1000)(Invitrogen, Burlington, Ontario, Canada, A21222) wasincubated for 2 h at room temperature in the dark. Tiledimages were retrieved with an Olympus BX61 motorizedfluorescence microscope at 20X magnification (OlympusAmerica Inc., Center Valley, Pennsylvania, USA). Inten-sity quantification was performed using Image-Pro(Media Cybernetics Inc., Bethesda, Maryland, USA). Therelative intensity level of Lhx2 was calculated as de-scribed in the literature [88]. Briefly, the sum of the sig-nal intensity was divided by the area selected andmultiplied by the thickness of the section and the num-ber of sections. A background correction was appliedusing the signal intensity resulting from Hoechst stain-ing for each sections quantified. A total of 28 differentsections were assessed on six embryos, three differentanimals for each genotype (Wt and Nr2e1frc/frc). Allvalues represent the mean ± standard error of the mean(SEM). Statistical analysis was performed using Student’st-test.Embryonic stem cells cultureESC from Wt and Nr2e1frc/frc blastocysts were derived,and maintained in culture as described in the litera-ture [89]. The two cell lines used were mEMS1239(B6129F1-Nr2e1frc/frc, Hprt1b-m3/Y), and mEMS1271(B6129F1-Nr2e1+/+, Hprt1b-m3/Y).The ESC differentiation procedure involved the use ofan adapted method of neurogenesis from adherent mono-culture [74, 75]. Briefly, the cells were seeded at lowSchmouth et al. BMC Genomics  (2015) 16:545 Page 16 of 19density (~10,000 cells/mm2) on gelatin coated dishes, in achemically defined medium exempt of cyclopamine, andmaintained in culture for 12 days (fresh media every twodays). RNA aliquots were prepared on day 12 and wereused for quantitative RT-PCR (qRT-PCR).Quantitative RT-PCRRNA from ESC grown in an adapted method of neurogen-esis from adherent monoculture, collected on day 12, wasextracted using Qiagen RNA Mini Plus kit (Qiagen Inc.,Mississauga, Ontario, Canada). RNA was treated withQiagen DNase kit (Qiagen Inc., Mississauga, Ontario,Canada), and cDNA was generated using Superscript IIIMaster Mix kit (Invitrogen, Burlington, Ontario, Canada).cDNA quantification was performed using ABI Taqman®assays specifically designed for Nr2e1 (Mm00455855_m1),and Lhx2 (Mm00839783_m1) (Applied Biosystems Inc.,Foster city, USA). The 7500 fast real-time PCR systemand Taqman® fast universal PCR Master Mix were used(Applied Biosystems Inc., Foster city, USA). The cyclethreshold (Ct) value was defined as the number of cy-cles required for the fluorescent signal to cross athreshold above the background signal, and is inverselyproportional to the amount of target cDNA. All valuesrepresent the mean ± SEM. Statistical analysis was per-formed using Student’s t-test.Availability of supporting dataLongSAGE processed data is accessible on theMouse Atlas of Gene Expression project website(http://www.mouseatlas.org/) and the NIH SAGEmapdata repository (http://www.ncbi.nlm.nih.gov/projects/SAGE/). All additional supporting data is included asadditional files.Additional filesAdditional file 1: Table S1. List of the 1387 differentially abundanttag sequences mapping to the 1279 genes, differentially expressed inresponse to altered Nr2e1 levels during in vivo neocortex developmentand the corresponding tag numbers, calculated fold changes and Pvalues at E13.5, E15.5, and E17.5. Columns one and two present thegene ID/symbol, associated tag sequence(s) and corresponding RefseqID for each gene. Columns three and four, seven and eight, eleven andtwelve present the number of tags found in Wt and Nr2e1frc/frc librariesat E13.5, E15.5, and E17.5 inclusively. Column five, nine, and thirteenpresent the fold change values resulting from calculations based on thetag number values found in the previous columns (details, seeMethods). Column six, ten, and fourteen present the associatedP values calculated using the Audic-Claverie statistical test for eachcorresponding tags at the threedifferent time points (E13.5, E15.5 and E17.5).Additional file 2: Table S2. List of the 1279 genes, differentiallyexpressed in response to altered Nr2e1 levels during in vivo neocortexdevelopment, and the summary of significant findings for each. Columnsone and two present the gene ID/symbol and associated transcript ID foreach gene. Columns three to five present the genes having significantlydifferentially abundant tags at E13.5, E15.5, and E17.5, respectively.Columns six to nine present the genes retained after performingsubsequent bioinformatics analyses. *, Multiple tags that were eitherup- or down-regulated but mapped to the same gene; details of the tagsequences and the direction of change are depicted. NA, not applicable.Additional file 3: Table S3. Nr2e1 binding sites containing sequencesextracted from the literature and used to generate a position weightmatrix. Columns one and two present the gene names and speciesorigin of the DNA sequence analyzed. Column three presents the DNAsequence analyzed. Column four presents the PAZAR ID matching theDNA sequence presented in column three. Column five presents thePubMed ID associated with the manuscript from which the sequenceswere derived. NA, not applicable.Additional file 4: Table S4. Empirical P values extracted from theZ-scores and Fisher scores of each candidate-NR2E1 co-interactors.Column one presents the transcription factor symbol. Column two andthree present the calculated empirical P values for both the Z-scoreand Fisher score of the corresponding transcription factor.AbbreviationsABA: Allen Mouse Brain Atlas; BSA: Bovine Serum Albumin; CNS: CentralNervous System; ESC: Embryonic Stem Cells; GO: terms Gene Ontologyterms; LCM: Laser Capture Microdissection; NSC: Neural Stem Cells;OCT: Optimal Cutting Temperature; PBS: Phosphate Buffered Saline;PC: Progenitor Cells; PFA: Paraformaldehyde; PO: Na2HPO4, pH 8.0;PWM: Position Weight Matrix; TFBS: Transcription Factor Binding Site;SAGE: Serial Analysis of Gene Expression; SELEX: Systematic Evolution ofLigands by Exponential Enrichment; TSS: Transcription Start Sites;SVZ: Subventricular Zone; UCSC: University of California, Santa Cruz;UTR: Untranslated Region; VZ: Ventricular Zone; Wt: Wild-Type.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsSJMJ, MAM, and EMS initiated the project. YYX and SB performed the lasercapture microdissections. SJMJ and MAM oversaw the LongSAGE librariesgeneration. KGB, RJB, and SHW derived the ESC from Wt and Nr2e1frc/frcblastocysts. JFS and XCD developed the ESC differentiation procedure. JFS,DA, XCD, EMS, and WWW designed the NR2E1 position weight matrix. JFS,DA, EMS, and WWW designed the bioinformatics pipeline. JFS and DAperformed the bioinformatics analyses. JFS performed the immunofluorescencesand quantitative RT-PCR experiments. JFS performed all data analyses, wrotemost of the text, and created all of the figures for this manuscript. XCD, DA, andWWW contributed to the text of this manuscript. DA, XCD, KGB, SJMJ, MAM,EMS, and WWW revised the manuscript prior to submission. All authors readand approved the final manuscript.AcknowledgementsWe thank the entire Mouse Atlas of Gene Expression Project members; withspecial thanks to Dr. Robert A. Holt and the Genome Science CentreSequencing Team for their work on the LongSAGE libraries. We also thank allthe Pleiades Promoter Project members. Lastly, we thank Dr. Charles N. deLeeuw and Katrina Bepple for aid in manuscript preparation, as well asMarina Campbell for administrative assistance.Author details1Centre for Molecular Medicine and Therapeutics at the Child and FamilyResearch Institute, University of British Columbia, 950 West 28th Avenue,Vancouver, BC V5Z 4H4, Canada. 2Genetics Graduate Program, University ofBritish Columbia, Vancouver, BC V6T 1Z2, Canada. 3Canada’s Michael SmithGenome Sciences Centre, British Columbia Cancer Agency, Vancouver, BCV5Z 4S6, Canada. 4Department of Molecular Biology and Biochemistry, SimonFraser University, Burnaby, BC V5A 1S6, Canada. 5Department of MedicalGenetics, University of British Columbia, Vancouver, BC V6T 1Z3, Canada.6Department of Psychiatry, University of British Columbia, Vancouver, BC V6T2A1, Canada. 7Current address: Montreal Neurological Institute and Hospital,McGill University, Montréal, QC H3A 2B4, Canada.Received: 10 December 2014 Accepted: 13 July 2015Schmouth et al. BMC Genomics  (2015) 16:545 Page 17 of 19References1. Anderson SA, Eisenstat DD, Shi L, Rubenstein JL. Interneuron migration frombasal forebrain to neocortex: dependence on Dlx genes. Science.1997;278:474–6.2. Angevine Jr JB, Sidman RL. Autoradiographic study of cell migration duringhistogenesis of cerebral cortex in the mouse. Nature. 1961;192:766–8.3. de Carlos JA, Lopez-Mascaraque L, Valverde F. Dynamics of cell migration fromthe lateral ganglionic eminence in the rat. J Neurosci. 1996;16:6146–56.4. Nadarajah B, Brunstrom JE, Grutzendler J, Wong RO, Pearlman AL. Twomodes of radial migration in early development of the cerebral cortex.Nat Neurosci. 2001;4:143–50.5. Tamamaki N, Fujimori KE, Takauji R. Origin and route of tangentiallymigrating neurons in the developing neocortical intermediate zone.J Neurosci. 1997;17:8313–23.6. Bayer SA, Altman J. Neocortical development. New York: Raven; 1991.7. Miller FD, Gauthier AS. Timing is everything: making neurons versus glia inthe developing cortex. Neuron. 2007;54:357–69.8. Job C, Tan SS. Constructing the mammalian neocortex: the role of intrinsicfactors. Dev Biol. 2003;257:221–32.9. Roy K, Kuznicki K, Wu Q, Sun Z, Bock D, Schutz G, et al. The Tlx generegulates the timing of neurogenesis in the cortex. J Neurosci.2004;24:8333–45.10. Land PW, Monaghan AP. Expression of the transcription factor, tailless,is required for formation of superficial cortical layers. Cereb Cortex.2003;13:921–31.11. Li W, Sun G, Yang S, Qu Q, Nakashima K, Shi Y. Nuclear receptor TLXregulates cell cycle progression in neural stem cells of the developing brain.Mol Endocrinol. 2008;22:56–64.12. Shi Y, Chichung Lie D, Taupin P, Nakashima K, Ray J, Yu RT, et al. Expressionand function of orphan nuclear receptor TLX in adult neural stem cells.Nature. 2004;427:78–83.13. Zhang CL, Zou Y, He W, Gage FH, Evans RM. A role for adult TLX-positiveneural stem cells in learning and behaviour. Nature. 2008;451:1004–7.14. Sun G, Yu RT, Evans RM, Shi Y. Orphan nuclear receptor TLX recruits histonedeacetylases to repress transcription and regulate neural stem cellproliferation. Proc Natl Acad Sci U S A. 2007;104:15282–7.15. Biddie SC, John S. Minireview: conversing with chromatin: the language ofnuclear receptors. Mol Endocrinol. 2014;28:3–15.16. Lonard DM, O’Malley BW. Nuclear receptor coregulators: judges, juries, andexecutioners of cellular regulation. Mol Cell. 2007;27:691–700.17. Lonard DM, Lanz RB, O’Malley BW. Nuclear receptor coregulators andhuman disease. Endocr Rev. 2007;28:575–87.18. Monaghan AP, Grau E, Bock D, Schutz G. The mouse homolog of theorphan nuclear receptor tailless is expressed in the developing forebrain.Development. 1995;121:839–53.19. Monaghan AP, Bock D, Gass P, Schwager A, Wolfer DP, Lipp HP, et al.Defective limbic system in mice lacking the tailless gene. Nature.1997;390:515–7.20. Young KA, Berry ML, Mahaffey CL, Saionz JR, Hawes NL, Chang B, et al.Fierce: a new mouse deletion of Nr2e1; violent behaviour and ocularabnormalities are background-dependent. Behav Brain Res. 2002;132:145–58.21. Yokoyama A, Takezawa S, Schule R, Kitagawa H, Kato S. Transrepressivefunction of TLX requires the histone demethylase LSD1. Mol Cell Biol.2008;28:3995–4003.22. Zhao C, Sun G, Li S, Shi Y. A feedback regulatory loop involving microRNA-9and nuclear receptor TLX in neural stem cell fate determination. Nat StructMol Biol. 2009;16:365–71.23. Sun G, Ye P, Murai K, Lang MF, Li S, Zhang H, et al. miR-137 forms aregulatory loop with nuclear receptor TLX and LSD1 in neural stem cells.Nat Commun. 2011;2:529.24. Zhao C, Sun G, Ye P, Li S, Shi Y. MicroRNA let-7d regulates the TLX/microRNA-9 cascade to control neural cell fate and neurogenesis. Sci Rep.2013;3:1329.25. Iwahara N, Hisahara S, Hayashi T, Horio Y. Transcriptional activation of NAD+ −dependent protein deacetylase SIRT1 by nuclear receptor TLX. BiochemBiophys Res Commun. 2009;386:671–5.26. Hisahara S, Chiba S, Matsumoto H, Tanno M, Yagi H, Shimohama S, et al. Histonedeacetylase SIRT1 modulates neuronal differentiation by its nuclear translocation.Proc Natl Acad Sci U S A. 2008;105:15599–604.27. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. Serial analysis of geneexpression. Science. 1995;270:484–7.28. Hanriot L, Keime C, Gay N, Faure C, Dossat C, Wincker P, et al. A combinationof LongSAGE with Solexa sequencing is well suited to explore the depth andthe complexity of transcriptome. BMC Genomics. 2008;9:418.29. Gunnersen JM, Augustine C, Spirkoska V, Kim M, Brown M, Tan SS.Global analysis of gene expression patterns in developing mouseneocortex using serial analysis of gene expression. Mol Cell Neurosci.2002;19:560–73.30. Koehl A, Schmidt N, Rieger A, Pilgram SM, Letunic I, Bork P, et al. Geneexpression profiling of the rat superior olivary complex using serial analysisof gene expression. Eur J Neurosci. 2004;20:3244–58.31. D’Souza CA, Chopra V, Varhol R, Xie YY, Bohacec S, Zhao Y, et al.Identification of a set of genes showing regionally enriched expression inthe mouse brain. BMC Neurosci. 2008;9:66.32. Popesco MC, Frostholm A, Rejniak K, Rotter A. Digital transcriptome analysisin the aging cerebellum. Ann N Y Acad Sci. 2004;1019:58–63.33. Ouchi Y, Kubota Y, Ito C. Serial analysis of gene expression in methamphetamine-and phencyclidine-treated rodent cerebral cortices: are there commonmechanisms? Ann N Y Acad Sci. 2004;1025:57–61.34. Guipponi M, Li QX, Hyde L, Beissbarth T, Smyth GK, Masters CL, et al. SAGEanalysis of genes differentially expressed in presymptomatic TgSOD1G93Atransgenic mice identified cellular processes involved in early stage of ALSpathology. J Mol Neurosci. 2010;41:172–82.35. Mazarei G, Neal SJ, Becanovic K, Luthi-Carter R, Simpson EM, Leavitt BR.Expression analysis of novel striatal-enriched genes in Huntington disease.Hum Mol Genet. 2010;19:609–22.36. George AJ, Gordon L, Beissbarth T, Koukoulas I, Holsinger RM, Perreau V,et al. A serial analysis of gene expression profile of the Alzheimer’s diseaseTg2576 mouse model. Neurotox Res. 2010;17:360–79.37. Peters DG, Kassam AB, Yonas H, O’Hare EH, Ferrell RE, Brufsky AM.Comprehensive transcript analysis in small quantities of mRNA by SAGE-lite.Nucleic Acids Res. 1999;27:e39.38. Wahl MB, Heinzmann U, Imai K. LongSAGE analysis significantly improvesgenome annotation: identifications of novel genes and alternativetranscripts in the mouse. Bioinformatics. 2005;21:1393–400.39. Ho Sui SJ, Fulton DL, Arenillas DJ, Kwon AT, Wasserman WW. oPOSSUM:integrated tools for analysis of regulatory motif over-representation.Nucleic Acids Res. 2007;35:W245–52.40. Ho Sui SJ, Mortimer JR, Arenillas DJ, Brumm J, Walsh CJ, Kennedy BP, et al.oPOSSUM: identification of over-represented transcription factor bindingsites in co-expressed genes. Nucleic Acids Res. 2005;33:3154–64.41. Das MK, Dai HK. A survey of DNA motif finding algorithms. BMCBioinformatics. 2007;8 Suppl 7:S21.42. Wasserman WW, Sandelin A. Applied bioinformatics for the identification ofregulatory elements. Nat Rev Genet. 2004;5:276–87.43. Stenman JM, Wang B, Campbell K. Tlx controls proliferation andpatterning of lateral telencephalic progenitor domains. J Neurosci.2003;23:10568–76.44. Robertson N, Oveisi-Fordorei M, Zuyderduyn SD, Varhol RJ, Fjell C, Marra M, et al.DiscoverySpace: an interactive data analysis application. Genome Biol. 2007;8:R6.45. Romanuik TL, Wang G, Holt RA, Jones SJ, Marra MA, Sadar MD.Identification of novel androgen-responsive genes by sequencing ofLongSAGE libraries. BMC Genomics. 2009;10:476.46. Audic S, Claverie JM. The significance of digital gene expression profiles.Genome Res. 1997;7:986–95.47. de Hoon MJ, Imoto S, Nolan J, Miyano S. Open source clustering software.Bioinformatics. 2004;20:1453–4.48. Saldanha AJ. Java Treeview–extensible visualization of microarray data.Bioinformatics. 2004;20:3246–8.49. Portales-Casamar E, Arenillas D, Lim J, Swanson MI, Jiang S, McCallumA, et al. The PAZAR database of gene regulatory information coupledto the ORCA toolkit for the study of regulatory sequences. NucleicAcids Res. 2009;37:D54–60.50. da Huang W, Sherman BT, Lempicki RA. Systematic and integrative analysisof large gene lists using DAVID bioinformatics resources. Nat Protoc.2009;4:44–57.51. Dennis Jr G, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, et al.DAVID: Database for Annotation, Visualization, and Integrated Discovery.Genome Biol. 2003;4:P3.52. Portales-Casamar E, Kirov S, Lim J, Lithwick S, Swanson MI, Ticoll A, et al.PAZAR: a framework for collection and dissemination of cis-regulatorysequence annotation. Genome Biol. 2007;8:R207.Schmouth et al. BMC Genomics  (2015) 16:545 Page 18 of 1953. Yu RT, Chiang MY, Tanabe T, Kobayashi M, Yasuda K, Evans RM, et al.The orphan nuclear receptor Tlx regulates Pax2 and is essential for vision.Proc Natl Acad Sci U S A. 2000;97:2621–5.54. Zhang CL, Zou Y, Yu RT, Gage FH, Evans RM. Nuclear receptor TLXprevents retinal dystrophy and recruits the corepressor atrophin1.Genes Dev. 2006;20:1308–20.55. Qu Q, Sun G, Li W, Yang S, Ye P, Zhao C, et al. Orphan nuclear receptor TLXactivates Wnt/beta-catenin signalling to stimulate neural stem cellproliferation and self-renewal. Nat Cell Biol. 2010;12:31–40.sup pp 31–39.56. Yu RT, McKeown M, Evans RM, Umesono K. Relationship betweenDrosophila gap gene tailless and a vertebrate nuclear receptor Tlx. Nature.1994;370:375–9.57. Hubisz MJ, Pollard KS, Siepel A. PHAST and RPHAST: phylogenetic analysiswith space/time models. Brief Bioinform. 2011;12:41–51.58. Siepel A, Bejerano G, Pedersen JS, Hinrichs AS, Hou M, Rosenbloom K, et al.Evolutionarily conserved elements in vertebrate, insect, worm, and yeastgenomes. Genome Res. 2005;15:1034–50.59. Sherman BT, Huang da W, Tan Q, Guo Y, Bour S, Liu D, et al. DAVIDKnowledgebase: a gene-centered database integrating heterogeneous geneannotation resources to facilitate high-throughput gene functional analysis.BMC Bioinformatics. 2007;8:426.60. da Huang W, Sherman BT, Tan Q, Kir J, Liu D, Bryant D, et al. DAVIDBioinformatics Resources: expanded annotation database and novelalgorithms to better extract biology from large gene lists. Nucleic Acids Res.2007;35:W169–75.61. Hosack DA, Dennis Jr G, Sherman BT, Lane HC, Lempicki RA. Identifyingbiological themes within lists of genes with EASE. Genome Biol. 2003;4:R70.62. da Huang W, Sherman BT, Lempicki RA. Bioinformatics enrichment tools:paths toward the comprehensive functional analysis of large gene lists.Nucleic Acids Res. 2009;37:1–13.63. Falcon S, Gentleman R. Using GOstats to test gene lists for GO termassociation. Bioinformatics. 2007;23:257–8.64. Shimozaki K, Zhang CL, Suh H, Denli AM, Evans RM, Gage FH. SRY-box-containing gene 2 regulation of nuclear receptor tailless (Tlx) transcriptionin adult neural stem cells. J Biol Chem. 2012;287:5969–78.65. Schuurmans C, Armant O, Nieto M, Stenman JM, Britz O, Klenin N, et al.Sequential phases of cortical specification involve Neurogenin-dependentand -independent pathways. Embo J. 2004;23:2892–902.66. Stenman J, Yu RT, Evans RM, Campbell K. Tlx and Pax6 co-operategenetically to establish the pallio-subpallial boundary in the embryonicmouse telencephalon. Development. 2003;130:1113–22.67. Peng GH, Ahmad O, Ahmad F, Liu J, Chen S. The photoreceptor-specificnuclear receptor Nr2e3 interacts with Crx and exerts opposing effects on thetranscription of rod versus cone genes. Hum Mol Genet. 2005;14:747–64.68. Aranguren XL, Beerens M, Coppiello G, Wiese C, Vandersmissen I, Lo NigroA, et al. COUP-TFII orchestrates venous and lymphatic endothelial identityby homo- or hetero-dimerisation with PROX1. J Cell Sci. 2013;126:1164–75.69. Lein ES, Hawrylycz MJ, Ao N, Ayres M, Bensinger A, Bernard A, et al.Genome-wide atlas of gene expression in the adult mouse brain. Nature.2007;445:168–76.70. Scott CE, Wynn SL, Sesay A, Cruz C, Cheung M, Gomez Gaviro MV, et al. SOX9induces and maintains neural stem cells. Nat Neurosci. 2010;13:1181–9.71. Chou SJ, Perez-Garcia CG, Kroll TT, O’Leary DD. Lhx2 specifies regional fatein Emx1 lineage of telencephalic progenitors generating cerebral cortex.Nat Neurosci. 2009;12:1381–9.72. Subramanian L, Sarkar A, Shetty AS, Muralidharan B, Padmanabhan H,Piper M, et al. Transcription factor Lhx2 is necessary and sufficient tosuppress astrogliogenesis and promote neurogenesis in the developinghippocampus. Proc Natl Acad Sci U S A. 2011;108:E265–74.73. Chou SJ, O’Leary DD. Role for Lhx2 in corticogenesis through regulation ofprogenitor differentiation. Mol Cell Neurosci. 2013;56:1–9.74. Gaspard N, Bouschet T, Herpoel A, Naeije G, van den Ameele J,Vanderhaeghen P. Generation of cortical neurons from mouse embryonicstem cells. Nat Protoc. 2009;4:1454–63.75. Gaspard N, Bouschet T, Hourez R, Dimidschstein J, Naeije G, van denAmeele J, et al. An intrinsic mechanism of corticogenesis from embryonicstem cells. Nature. 2008;455:351–7.76. Stolt CC, Lommes P, Sock E, Chaboissier MC, Schedl A, Wegner M. The Sox9transcription factor determines glial fate choice in the developing spinalcord. Genes Dev. 2003;17:1677–89.77. Castillo SD, Sanchez-Cespedes M. The SOX family of genes in cancerdevelopment: biological relevance and opportunities for therapy.Expert Opin Ther Targets. 2012;16:903–19.78. Kiefer JC. Back to basics: sox genes. Dev Dyn. 2007;236:2356–66.79. Pevny L, Placzek M. SOX genes and neural progenitor identity. Curr OpinNeurobiol. 2005;15:7–13.80. Kwan KY, Lam MM, Krsnik Z, Kawasawa YI, Lefebvre V, Sestan N. SOX5postmitotically regulates migration, postmigratory differentiation, andprojections of subplate and deep-layer neocortical neurons. Proc Natl AcadSci U S A. 2008;105:16021–6.81. Sansom SN, Griffiths DS, Faedo A, Kleinjan DJ, Ruan Y, Smith J, et al. Thelevel of the transcription factor Pax6 is essential for controlling the balancebetween neural stem cell self-renewal and neurogenesis. PLoS Genet.2009;5:e1000511.82. Khattra J, Delaney AD, Zhao Y, Siddiqui A, Asano J, McDonald H, et al.Large-scale production of SAGE libraries from microdissected tissues,flow-sorted cells, and cell lines. Genome Res. 2007;17:108–16.83. Lash AE, Tolstoshev CM, Wagner L, Schuler GD, Strausberg RL, Riggins GJ,et al. SAGEmap: a public gene expression resource. Genome Res.2000;10:1051–60.84. Siddiqui AS, Khattra J, Delaney AD, Zhao Y, Astell C, Asano J, et al. A mouseatlas of gene expression: large-scale digital gene-expression profiles fromprecisely defined developing C57BL/6 J mouse tissues and cells. Proc NatlAcad Sci U S A. 2005;102:18485–90.85. Bailey TL, Boden M, Buske FA, Frith M, Grant CE, Clementi L, et al. MEMESUITE: tools for motif discovery and searching. Nucleic Acids Res.2009;37:W202–8.86. Visel A, Thaller C, Eichele G. GenePaint.org: an atlas of gene expressionpatterns in the mouse embryo. Nucleic Acids Res. 2004;32:D552–6.87. Diez-Roux G, Banfi S, Sultan M, Geffers L, Anand S, Rozado D, et al.A high-resolution anatomical atlas of the transcriptome in the mouseembryo. PLoS Biol. 2011;9:e1000582.88. Singaraja RR, Huang K, Sanders SS, Milnerwood AJ, Hines R, Lerch JP, et al.Altered palmitoylation and neuropathological deficits in mice lacking HIP14.Hum Mol Genet. 2011;20:3899–909.89. Yang GS, Banks KG, Bonaguro RJ, Wilson G, Dreolini L, de Leeuw CN, et al.Next generation tools for high-throughput promoter and expressionanalysis employing single-copy knock-ins at the Hprt1 locus. Genomics.2009;93:196–204.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at www.biomedcentral.com/submitSchmouth et al. BMC Genomics  (2015) 16:545 Page 19 of 19

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.52383.1-0228466/manifest

Comment

Related Items