Fulltext for PDF in Flash Viewer:

Developmental Biology 292 (2006) 152 164 www.elsevier.com/locate/ydbio Evidence for motoneuron lineage-specific regulation of Olig2 in the vertebrate neural tube Tao Sun a,d,, Brian P. Hafler a , Sovann Kaing a , Masaaki Kitada a , Keith L. Ligon a , Hans R. Widlund a , Dong-in Yuk a , Charles D. Stiles b , David H. Rowitch a,c, a Department of Pediatric Oncology, Dana-Farber Cancer Institute and Harvard Medical School, 44 Binney Street, Boston, MA 02115, USA b Department of Cancer Biology, Dana-Farber Cancer Institute and Harvard Medical School, 44 Binney Street, Boston, MA 02115, USA c Divisions of Newborn Medicine and Hematology, Children�� Hospital, 300 Longwood Avenue, Boston, MA 02115, USA d Department of Cell and Developmental Biology, Cornell University Weill Medical College, 1300 York Avenue, New York, NY 10021, USA Received for publication 6 October 2005; revised 14 December 2005; accepted 21 December 2005 Available online 8 February 2006 Abstract Within the motoneuron precursor (pMN) domain of the developing spinal cord, the bHLH transcription factor, Olig2, plays critical roles in pattern formation and the generation of motor neuron and oligodendrocyte precursors. How are the multiple functions of Olig2 regulated? We have isolated a large BAC clone encompassing the human OLIG2 locus that rescues motor neuron and oligodendrocyte development but not normal pattern formation in Olig2����embryos. Within the BAC clone, we identified a conserved 3.6 kb enhancer sub-region that directs reporter expression specifically in the motor neuron lineage but not oligodendrocyte lineage in vivo. Our findings indicate complex regulation of Olig2 by stage- and lineage-specific regulatory elements. They further suggest that transcriptional regulation of Olig2 is involved in segregation of pMN neuroblasts. 2006 Elsevier Inc. All rights reserved. Keywords: Motor neuron; Lineage mapping; Oligodendrocyte; Olig2; Transcription factor; Bacterial Artificial Chromosome (BAC); cis-regulation; Enhancer; Neural tube Introduction During vertebrate central nervous system (CNS) development, multipotent precursors give rise to sequential waves of progeny in response to regionally restricted cues and temporally regulated factors (Anderson, 2001; Temple, 2001). Conceptually, this process can be divided into patterning, neurogenic and gliogenic phases. Sonic hedgehog (Shh) signaling is essential for pattern formation and instructs naive progenitors to adopt segmental expression of transcription factors in distinct Corresponding authors. T. Sun is to be contacted at Department of Cell and Developmental Biology, Cornell University Weill Medical College, 1300 York Avenue, New York, NY 10021, USA. Fax: +1 212 746 8175. D.H. Rowitch, Divisions of Newborn Medicine and Hematology, Children�� Hospital, 300 Longwood Avenue, Boston, MA 02115, USA. Fax: +1 617 632 4850. E-mail addresses: tas2009@med.cornell.edu (T. Sun), david_rowitch@dfci.harvard.edu (D.H. Rowitch). 0012-1606/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2005.12.047 domains along the dorsalventral (DV) axis of the neural tube (Jessell, 2000). Function of Olig2, encoding a Shh-regulated bHLH protein (Lu et al., 2000) is necessary to establish the identity of the ��MN��domain, which is the source of motor neuron (MN) and oligodendrocyte (OL) precursors in the ventral neural tube (Mizuguchi et al., 2001; Novitch et al., 2001; Lu et al., 2002; Park et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002; Zhou et al., 2000). In addition to early roles in neural pattern formation, Shh and retinoic acid signaling regulate the subsequent generation of MN from pMN progenitor cells (Marti et al., 1995; Roelink et al., 1995; Novitch et al., 2003). The requirement for Shh signaling extends into S-phase of the final progenitor cell division during which Shh drives differentiation into MN (Ericson et al., 1996). Additionally, ongoing Shh activity is needed in the gliogenic phase for specification of OLP (Orentas et al., 1999; Soula et al., 2001). The finding that Olig2 function is required for generation of motor neuron and oligodendrocyte T. Sun et al. / Developmental Biology 292 (2006) 152164 153 precursors indicates further mechanistic parallels in the development of these lineages (Mizuguchi et al., 2001; Novitch et al., 2001; Lu et al., 2002; Park et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002). Olig2 expression in the neural tube is bi-phasic with peak levels corresponding the major periods of MN and OLP specification at 9.5 and 12.5 days post coitum (dpc), respectively (Sun et al., 2001). During the neurogenic phase, Olig2 is expressed in proliferating pMN precursors and a subset of Olig2+ cells co-express Lhx3 and Ngn2 in neuroblasts (Mizuguchi et al., 2001; Novitch et al., 2001; Zhou et al., 2001). While the mechanisms underlying early segregation of pMN neuroblasts in the vertebrate neural tube are incompletely understood, recent work has implicated ongoing effects of Hedgehog signaling in conjunction with temporally regulated factors (Park et al., 2004). In contrast, less is known about intrinsic regulatory elements for the genes that regulate pattern formation and fate choice. We have focused transcriptional regulation of Olig2 at early stages of CNS development. We achieved rescue of motor neurons and oligodendrocytes in Olig2 null mutant animals in a dose-dependent manner using a human Bacterial Artificial Chromosome (BAC) encompassing the OLIG2 locus, and have identified a functionally conserved 3.6 kb motoneuron lineage-specific regulatory element. Our data leads to the proposal that regulation is complex with separable regulatory elements for patterning, neurogenesis and gliogenesis. Further, they implicate direct regulation of Olig2 in the process of neuroblast segregation from a pool of pMN precursors in the vertebrate embryo. Materials and methods Screening of the human BAC library Human Bacterial Artificial Chromosomes (BAC) containing human genomic sequences for OLIG2 and/or OLIG1 were screened from a human high density BAC library (Research Genetics, Invitrogen) using the mouse Olig2 cDNA as a probe. Three positive clones were identified and only one (2401C4) can be expanded for maxi-preparation of BAC DNA. BAC-2401C4 was end-sequenced using T7 and Sp6 primers. The position of BAC-2401C4 on human chromosome 21 was aligned with other BAC clones covering human OLIG2 and OLIG1. Transcription factor binding site analysis was done with MatInspector Professional. lacZ transgenic mice. Transgenic founders (n = 3) were identified by PCR to amplify a 250 bp lacZ segment (lacZ-F: 5��TTAACGCCGTGCGCTGTTCG3��lacZ-R: 5��ATCCAGCGATACAGCGCGTC-3�� The transmission of the K23-lacZ transgene was analyzed by staining the neural tube with ��alactosidase. The OLIG2/K23 enhancer-cre line was generated by the microinjection of fertilized mouse oocytes and transgenic founders were then crossed with the ROSA-26 reporter line. The transmission of the transgene was analyzed by staining the neural tube with ��galactosidase. Analysis of K23 DNA regulatory sequences Human, mouse, rat, and dog alignment was downloaded from the USCS genome browser (http://genome.ucsc.edu/) using the hg16/mm4/rn2/canFam0 assembly. Transcription factor binding sites were identified using the TRANSFAC database (http://www.genomatix.de). BAC transgene copy number and motor neuron rescue To determine transgene copy number, part of the 5��TRs of the human and mouse Olig2 genes were co-amplified by PCR using the following primers: 5��ACTCACCGCWGCATC-3��nd 5��ACCAGGCTGGCGTCCGAGTC-3��PCR products were removed during the logarithmic phase of amplification (after 20, 22, 24, 26 and 28 cycles) and electrophoresed on a 2% (w/v) agarose gel. This PCR reaction will generate two bands: a 707-bp human band and a 623-bp mouse band. The intensity of each band was measured and quantified on a SpotDenso program (Apha Innotech Corporation, model 2.1.3.). Heterozygous Olig2+/��mutant mice were crossed with hemizygous human BAC transgenic mice from lines #32 and #34 and sibling-mated to generate heterozygous Olig2+/�� BACTg/+ offspring or homozygous Olig2���� BACTg/+ offspring. Genotypes were determined, retrospectively, by PCR on tail-tip DNA. To quantify MN rescue, numbers of Hb9+ and Isl1+ MN in one side of the spinal cord were counted from more than 10 sections taken at different anterior posterior levels of three animals of the genotypes indicated in Fig. 2 (n = 9 animals in total). The average and standard deviation of MN numbers were calculated to determine level of significance. Enhancer screen and forced expression in the chick neural tube For the enhancer screen, human BAC DNA was digested with KpnI and then randomly sub-cloned into a vector containing WntI minimal promoter, WntI 110 bp minimal enhancer and lacZ. The insert of each DNA construct was sequenced to localize its position on BAC DNA. The chick neural tubes at Hamburger and Hamilton (HH) stage 1012 were electroporated unilaterally (five 50 ms pulses at 25 V) with DNA construct (3 ��/��) using an ECM830 electroporator (BTX Inc.). For co-electroporation, equal amounts of plasmid DNA (3 ��/�� each) were mixed before electroporation. Embryos were analyzed after 48 h incubation (HH stages 2123) by in situ hybridization (ISH) and immunohistochemistry (IHC) after cryosectioning or stained in whole mount with ��galactosidase for the reporter constructs. At least ten electroporated embryos were analyzed for each plasmid DNA construct and results documented in the figures are representative of at least five positive embryos. Generation of transgenic mice BAC DNA was prepared using a BAC maxi-preparation kit from the Clontech. The BAC transgenic mice were generated by microinjecting BAC DNA into fertilized oocytes (1-cell embryos), and transgenic founders were identified by PCR and Southern blot analysis of tail-tip DNA. The integrity of BAC DNA was examined by both PCR and Southern Blots (The sequences of PCR primers are available upon request). Five transgenic lines were generated and two stable lines (lines 32 and 34) transmitted through the germline. The expression of BAC transgene was examined using a human specific probe by in situ hybridization. The human specific probes for OLIG2 and OLIG1 were amplified within the 3��ntranslated regions (3��TRs) for each gene using PCR. H-OLIG1-3��TR: 5��CGAACCTTCCAGTCCAGAG-3��nd 5��TACGTATGTAGCTCTTCAACTG-3��H-OLIG2-3��TR: 5��ACCTGATCGAGCGCTGTCTGG-3��nd 5��AAGGGTGTTACACGGCAGACG-3��The DNA constructs with positive activity in the ventral chick neural tube were microinjected into fertilized mouse oocytes to generate OLIG2 enhancer- Tissue preparation, immunohistochemistry and in situ hybridization Mouse embryos were collected at different ages according to the assumption that mating occurred at midnight with plug detection at the following noon (embryonic day 0.5). Embryos were fixed in 4% paraformaldehyde (PFA) in Phosphate Buffered Saline (PBS, pH 7.0), embedded in OCT and cryosectioned at 1014 ��. For immunohistochemistry on sections, the following primary antibodies were used at the indicated dilutions: mIsl1 (1:50, DSHB), mMNR2/ Hb9 (1:50, DSHB), mNkx2.2 (1:500, DSHB), Olig2 (1:4000), Lhx3 (1:100, DSHB), Chx10 (1:5000, gift from C. Cepko, Harvard Medical School). The secondary antibodies were anti-mouse IgG FITC (1:100, Jackson) and antirabbit IgG Cy3 (1:200, Jackson). Photomicrographic images were collected on a Nikon E600 microscope and RT Monochrome SPOT digital camera. In situ hybridization was performed using a modified protocol of (Ma et al., 1997) and 154 T. Sun et al. / Developmental Biology 292 (2006) 152164 digoxygenin (DIG)-labeled antisense mRNA probe for Olig1, Olig2, PDGFR�� Sox10, PLP/DM20, MBP, hOLIG2-3��TR and hOLIG1-3��TR. Detailed instructions are available upon request. Results Regulatory sequences within a 116 kb human BAC transgene recapitulate aspects of endogenous Olig2 expression To identify cis-acting regulatory sequences for Olig2, we took advantage of a human BAC library to generate BAC transgenic mice (Heintz, 2001). Human OLIG2 lies on Chromosome 21 and a clone containing human OLIG2 coding sequence was identified using a mouse Olig2 probe under low stringency. BAC #2401C4 (Research Genetics), called BAC �� was end-sequenced and aligned with known BAC clones in Genbank, such as AP000041 and AP001716, to identify the precise locations for OLIG1 and OLIG2. As shown (Fig. 1A), BAC �� comprises approximately 70 kb of sequence upstream of OLIG2 and 43 kb of downstream sequence as well as OLIG2 and OLIG1 genes. The identity of the BAC �� was confirmed by PCR using primers covering different regions on the BAC. It was then used to generate transgenic mice by pronuclear microinjection. Of 40 pups screened, five transgenic founders were identified and two of these (lines 32 and 34) transmitted the BAC transgene to progeny. To characterize human OLIG2 expression driven by BACencoded regulatory sequences, we used a species-specific probe generated against the 3��ntranslated region of human OLIG2. BAC ��-driven expression of OLIG2 commenced at 9.5 dpc in the caudal spinal cord (Fig. 1B). At 11.5 dpc, human OLIG2 mRNA transcripts were detected at rostral (forelimb) levels, a similar pattern to endogenous Olig2. At 17.5 dpc, the human OLIG2 transgene, like the endogenous mouse Olig2, was also expressed in scattered cells evenly distributed in the spinal cord, and such expression co-localized Fig. 1. A human bacterial artificial chromosome (BAC ��) containing OLIG2 recapitulates aspects of endogenous Olig2 expression in mice. (A) Cartoon showing the structure of the OLIG2 locus 42 kb from OLIG1 on human chromosome 21 and sequenced human BAC clones. Of these, BAC-2401C4 (called ��AC ���� was endsequenced and aligned with published BAC clone sequences as shown. (B) Expression analysis of human-specific OLIG2 mRNA transcripts in situ in BAC �� transgenic mice. Expression of mouse Olig2 and human OLIG2 by in situ hybridization is shown for wild type and BAC transgenic (Tg) mice at 9.5, 11.5 and 17.5 dpc. Expression of human OLIG2 is first detected at 9.5 dpc at caudal, but not rostral, levels of the spinal cord. At 11.5 dpc, expression of mouse Olig2 in the pMN domain of rostral spinal cord is comparable to that of human OLIG2 expression in the Tg mice (black arrowheads). At 17.5 dpc, Olig2-expressing OLP are dispersed throughout the spinal cord. Expression of hOLIG2 transcripts was observed in a subset (approximately 60%) of Olig2 expressing cells (insert). T. Sun et al. / Developmental Biology 292 (2006) 152164 155 to some (��0%) but not all cells expressing Olig2 proteins (Fig. 1B, see inset). We were unable to detect human OLIG1 expression in either BAC transgenic line (data not shown). These findings indicate that BAC �� contains regulatory elements sufficient to recapitulate important aspects of endogenous Olig2 expression. Rescue of motor neuron and oligodendrocyte precursor development in Olig2����mutant mice by BAC �� Motor neurons fail to develop in the spinal cord of Olig2����embryos (Lu et al., 2002; Park et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002) and no cells expressing the MN markers Isl1, Isl2 or Hb9 are detectable in Olig2����mutants at 14.5 dpc (Lu et al., 2002). To assess whether BAC-encoded human OLIG2 transgene functions during the neurogenic phase in mice, we intercrossed BAC ��-transgenic mice (line 34) with Olig2+/��heterozygotes to obtain Olig2����BACTg34/+ animals. As shown (Fig. 2A), Olig2����BACTg34/+ animals contained surviving Isl1- and Hb9-positive cells, though in markedly reduced numbers compared to Olig2+/��BACTg34/+ heterozygous littermates at 14.5 dpc (Fig. 2C). Similar results were obtained at 11.5 dpc (not shown). Given that Olig1 is incapable of supporting MN specification (Lu et al., 2002), these data indicate that human BAC �� contains functional regulatory sequences for OLIG2 expression during neurogenesis. The fact that only partial rescue of MN is achieved in Olig2����BACTg34/+ animals might indicate different capabilities of mouse and human OLIG2 proteins, or that BAC �� lacks regulatory sequences needed for robust expression of OLIG2. To address the later point, we tested the effects of BAC �� gene dosage on the degree of MN rescue. An independent transgenic line was generated (line 32) with increased copy number. The human OLIG2 PCR product was about 1.5 times more abundant than the mouse Olig2 product, indicating that three copies of OLIG2 had integrated per diploid mouse genome in line 32, in contrast to two copies of the BAC in line 34 (Fig. 2B). The large size of BACs makes them relatively insensitive to position-dependent integration effects so that copy number correlates well to expression levels (Yang et al., 1999; Heintz, 2001). Consistent with this, we observed stronger expression of human OLIG2 mRNA transcripts in situ (Fig. 2A) and detectable levels of OLIG2 proteins (data not shown) in Olig2����BACTg32/+ relative to Olig2����BACTg34/+ animals. Augmented rescue of MN numbers was achieved in Olig2����BACTg32/+ animals (Fig. 2C). We observed spontaneous and elicited motor activity only in Olig2����BACTg32/+ fetuses at 18.5 dpc; however none survived until weaning, reflecting insufficient numbers of motor neurons generated in the hindbrain and/or spinal cord or delayed development of MN with resulting significant compromise of neurological function. In contrast, Olig2����and Olig2����BACTg34/+ animals were totally paralyzed. These data suggest that the degree MN rescue in an Olig2 null background is correlated to the dosage of BACencoded human OLIG2. We used Isl1 and HB9 because previous studies have shown they label most early motor neuron subtypes (Sharma et al., 1998). We have detected Isl1+ and HB9+ motor Fig. 2. Rescue of motor neuron development in Olig2����mice by human BAC ��. (A) BAC transgenic lines 34 and 32 were assessed for their ability to rescue motor neuron development in Olig2 null embryos. Addition of BAC 34Tg/+ was able to weakly rescue MN development at 14.5 dpc in Olig2����animals (MN markers used: Isl1 and Hb9). In contrast, BAC line 32 promotes more robust rescue of MN populations and appears to drive stronger levels of OLIG2 expression. (B) Copy number analysis of BAC transgenic lines 32 and 34. The ratios of PCR products for human and mouse OLIG2 were analyzed by the semi-quantitative PCR. The copy numbers for human BAC transgene in lines 34 and 32 are two and three copies, respectively. (C) Quantitative analysis of MN rescue in BACTg/+, Olig2 null animals. Isl1+ and Hb9+ cells in ventral neural tube were counted at ten different anteriorposterior levels from each of three animals. The increased BAC copy number in line 32 significantly enhanced MN rescue compared with line 34 (P b 0.001). Control MN numbers in Olig2+/�� BAC 34 mice were equivalent to wild type. 156 T. Sun et al. / Developmental Biology 292 (2006) 152164 neurons in the rescued spinal cord at both 11.5 dpc and 14.5 dpc, and increased Lhx3+ cells in the rescued spinal cord at these stages (Supplementary Fig. 1 and data not shown). Because Lhx3 can also be expressed in V2 interneurons, its expression does not allow us to conclude any specific effects of BAC �� on developing MN sub-populations. Olig2 function is additionally required for specification of oligodendrocyte precursors (OLP) in the neural tube (Lu et al., 2002; Zhou and Anderson, 2002). Analysis with OLP markers in Olig2����BACTg32/+ animals demonstrates delayed rescue of this early phase of development (Fig. 3). However, oligodendroglia did not achieve expression of the intermediate mature, PLP/DM20, nor the mature marker, myelin basic protein (MBP), indicating a delay or block in OLP maturation. We conclude that BAC �� contains regulatory sequences for Olig2 expression in precursors for both MN and OLP. BAC-��-driven expression of OLIG2 is insufficient to promote normal patterning of the ventral neural tube In the dorsalventral (DV) axis of the embryonic neural tube, motor neurons derive from the Olig2+ precursors of the pMN domain (see Fig. 4G). The pMN is located ventral to the V2 interneuron precursor (p2) domain and dorsal to the V3 interneuron precursor (p3) domain, demarcated by expression of Nkx2.2. We thus investigated the nature of patterning in Olig2 ����BAC Tg32/+ embryos, in which robust OLIG2 expression was only detected until 1111.5 dpc at forelimb levels of the spinal cord (Fig. 1). As shown (Fig. 4A), at 10.5 dpc Olig2+ cells were extremely difficult to detect in the pMN. Indeed, identical to findings in Olig2����mutants (Lu et al., 2002; Takebayashi et al., 2002; Zhou and Anderson, 2002), we observed dramatically increased production of Chx10+ Fig. 3. Rescue of oligodendrocyte precursor (OLP) development in Olig2����mice by human BAC ��. (A) BAC line 32 can rescue OLP development at 12.5 dpc, labeled by Sox10 (arrows). The lack of PDGFR�� cells in the Olig2����BACTg32/+ neural tube indicates a delayed rescue of OLP development. (B) While the human BAC can rescue OLP development (labeled by PDGFR�� at 16.5 and 18.5 dpc, the maturation of oligodendrocytes did not occur (labeled by PLP/DM20 and MBP). T. Sun et al. / Developmental Biology 292 (2006) 152164 157 Fig. 4. BAC �� fails to rescue normal pMN patterning in Olig2����mice. (AC) Comparison of Olig2 expression and V2 IN development in Olig2+/��BACTg32/+ controls (top panels) and Olig2����BACTg32/+ mutants (bottom panels) at 10.5 dpc. (A) Note extremely low/undetectable levels of OLIG2 proteins (green, arrow) in the pMN region, which lies just dorsal to the Nkx2.2 (red)-expressing p3 domain. (B) Consistent with previously described patterning abnormalities in Olig2����embryos (Lu et al., 2002; Zhou and Anderson, 2002), there is a marked increase in Chx10+ V2 (red) and (C) GATA3+ V2b interneuron cell production, reflecting ventral expansion of the p2 domain (white brackets). (D) Comparison of Irx3 expression in control (left) and Olig2����BACTg32/+ (right) mice. Note that Irx3 expression (green) abuts p3 progenitors that express Nkx2.2 (red). (E, F) Very rare Olig2+ and Ngn2+ cells were observed in the pMN region. (G) Despite these abnormalities, we achieved partial rescue of MN development, indicating that OLIG2 expression is sufficient to establish MN precursor identity in the absence of normal pMN patterning. V2a interneurons (IN) and GATA3+ V2b interneurons in the Olig2����BACTg32/+ neural tube, indicative of ventralization of the p2 domain (Figs. 4B and C). Irx3 is expressed in p2 precursor cells but not Olig2+ pMN cells. Ventral expansion of the p2 domain in Olig2����BACTg32/+ embryos was therefore confirmed by the observation of Irx3 cells abutting Nkx2.2+ cells of the p3 domain (Fig. 4D). However, we did observe rare Olig2+ cells in Olig2����BACTg32/+ mice, suggesting that small isolated rests of Olig2+ cells are present (Fig. 4E). Such rests could contribute to the early rescue of Hb9+ cells and low levels of Ngn2-expression in 10.5 dpc Olig2����BACTg32/+ mice. Nevertheless, it is clear that BAC��-driven expression of OLIG2 is insufficient to normally pattern the pMN domain. Screen for OLIG2 enhancer elements that are functionally conserved across species We next sought to identify regulatory elements sufficient to drive aspects of early OLIG2 expression in vivo. However, given the large size of the candidate region in BAC �� (116 kb) and the fact that enhancer activities may be conserved across species, we first employed a screening strategy of electroporation into the E2 chick neural tube (Timmer et al., 2001). Thirteen well-conserved intergenetic regions of BAC �� between human and mouse genomes were cloned into a lacZ-containing construct (Fig. 5A). The reporter gene expression was tested 48 h after unilateral electroporation by whole mount X-gal staining (Fig. 5B). As shown, lacZ 158 T. Sun et al. / Developmental Biology 292 (2006) 152164 T. Sun et al. / Developmental Biology 292 (2006) 152164 159 Fig. 6. Olig2/K23-cre fate map in the ventral neural tube. (A) (top) Scheme for Olig2-enhancer fate mapping. Olig2/K23-cre mice were crossed with the conditional ROSA26 reporter line to fate map progeny cells with ��galactosidase proteins. (B) Initial activation of reporter activity in the neural tube is observed at the forelimb level at 9.5 dpc. Expression was also observed in the hindgut consistent with the normal pattern of the Olig2 expression (Lu et al., 2000) (see Supplementary Fig. 2). (C, D) Higher power views of neural tube indicates reporter activity in ventricular zone (VZ) precursors (black arrow) and laterally located cells. (E) Pattern of conditional reporter activation at 11 dpc in whole mount. Note strong expression of ��galactosidase throughout the spinal cord and ventral hindbrain with a rostral limit at the midbrainhindbrain isthmus. Little expression is observed in the forebrain in contrast to the normal pattern of Olig2 expression. (F, G) In 50 �� vibratome sections at the forelimb level, reporter activity is detected in pMN precursors of the VZ and a mature population in the mantle zone located laterally. (H) Co-localization of ��galactosidase activity with Olig2 (black arrowheads) confirms appropriate dorsalventral restriction to the pMN. Rare ��galactosidase positive cells that did not express Olig2 are indicated (white arrowhead). This suggests that K23-cre marks an initial step of Olig2 upregulation or pMN cells that subsequently downregulate Olig2. (I) BrdU (black arrowheads) confirms co-expression of the reporter in proliferating pMN precursor cells. expression for the construct K23 was detected in the ventral neural tube and it was co-localized with motor neurons labeled with Isl1 (Fig. 5B). Six of the constructs tested showed ��galactosidase expression in the ventral neural tube and four of these were tested for cross-species activity in transgenic mice (Supplementary Table 1). Of these, only enhancer K23 was active in the developing CNS of founder mouse transgenic embryos (n = 3). The 3.6 kb K23 DNA regulatory sequences showed high conservation amongst mammals and drove reporter lacZ expression in mouse spinal cord and hindbrain from about 10 to 12.5 dpc (Fig. 5C and Supplementary Fig. 2). Coexpression of the reporter in hindbrain Olig2+ progenitors was confirmed (Fig. 5C). At more caudal levels of the neural tube, K23 driven reporter expression was relatively weak and restricted to scattered cells in a sub-population of the pMN domain (Fig. 5C). Although Olig2 expression is rapidly extinguished as MN precursors leave the cell cycle, we observed ��galactosidase activity in MN of K23-lacZ transgenic mice from 10.5 dpc until 12.5 dpc (Fig. 5C and data not shown). This might reflect relative stability of ��alactosidase proteins, or possibly that K23 regulatory sequences lack DNA-binding sites for repressors of Olig2 expression in early MN. K23-driven ��galactosidase activity was present at neurogenic, but not gliogenic, phases of pMN development and was undetectable in OLP or OL of the spinal cord, suggesting lineage-specific expression of K23 in MN precursors. Fig. 5. Conserved regulatory sequences for Olig2 drive expression in the ventral spinal cord. (A) Scheme showing the OLIG1 and OLIG2 loci and regions conserved between mouse and human syntenic sequences on chromosomes 16 and 21, respectively. The highly conserved (N50%) regions indicated with pink shade were cloned into the reporter construct shown in B, which contained insulators from chicken ��globin locus, the minimal WntI promoter and the WntI 110 bp enhancer (Rowitch et al., 1998) that was used to drive lacZ reporter gene expression. (B) Reporter constructs were electroporated into the chick E2 neural tube and harvested at E4 for analysis by X-gal whole mount staining. In addition to dorsal WntI enhancer driven expression, used as a positive control (black arrowheads), several of the constructs tested showed ventral expression (black arrows; see Supplementary Table 1) and these were selected for further analysis in transgenic mice. Construct K23 showed coexpression with motor neuron marker Isl1 in the ventral neural tube. f: floorplate. (C) Confirmation of K23 enhancer activity in transgenic mice. (left) Whole mount staining at 11.5 dpc shows expression in spinal cord and hindbrain (white arrows). Hb: hindbrain level, fl: forelimb level. (middle panels) ��galactosidase activity (turquoise, black arrowheads) was observed in Olig2+ cells of the hindbrain. At forelimb levels, scattered X-gal+ pMN cells were observed before acquisition of the MN marker Hb9 (brown). (right) Reporter activity was detectable in pMN precursors (black arrow) and post-mitotic MN at 10.5 (red arrows) until 12.5 dpc. The pMN is indicated with black bars. 160 T. Sun et al. / Developmental Biology 292 (2006) 152164 The 3.6 kb Olig2 enhancer drives lineage-specific expression in motor neurons A robust and stringent test of enhancer specificity is to fate map all progeny deriving from its expression domain. We used the K23 regulatory sequences to drive expression of bacteriophage P1 cre recombinase and analyzed the resulting pattern of activity after crosses with a conditionally active ROSA26-STOP-lacZ reporter (Soriano, 1999) (Fig. 6A). As shown (Figs. 6B to D), ��galactosidase activity was observed in the ventral spinal cord at 9.5 dpc, and little ectopic expression was noted outside of the neural tube at all stages examined (Fig. 6E and Supplementary Fig. 3). At 11 dpc, ��galactosidase activity was restricted to proliferating precursors of the pMN and laterally located post-mitotic neurons (Figs. 6F and G). We further confirmed that X-gal+ cells co-expressed Olig2 and the S-phase marker BrdU (Figs. 6H and I). These data suggest that Olig2/K23-cre allele is sufficient to drive expression in proliferating Olig2+ precursors. Fig. 7. Olig2-cre activity drives MN-lineage specific expression. (AC) Co-labeling with the MN markers Isl1, Hb9 and Lhx3 indicates that ��galactosidase expressing cells are MN in the 11.5 dpc neural tube. Lower panels are the high power views of selected areas in upper panels. (D, E) In contrast, ��galactosidase staining segregates away from expression of Nkx2.2, a proxy marker for Sim1+ V3 interneuron precursors, and Chx10, a V2 interneuron marker. (F, G) Olig2 expression in oligodendrocytes segregates away from K23-cre labeled progeny as shown by Olig2 co-staining in 17.5 dpc and postnatal day 5 (P5) mouse spinal cord. Additionally, co-labeling with Isl1 at 17.5 dpc (F) reveals that K23-cre fate maps to some MN (filled arrows) but not all (unfilled arrows) MN, consistent with K23 sequences active only at a late phase of neurogenesis. T. Sun et al. / Developmental Biology 292 (2006) 152164 161 At 11.5 dpc and later stages, more than 90% of X-gal stained cells co-localized with the MN markers Isl1, Hb9 or Lhx3 (Figs. 7A to C). In contrast, such X-gal stained cells segregated away from expression of Nkx2.2, a proxy marker for the domain of V3 IN, and Chx10, the V2 IN marker (Figs. 7D and E). Furthermore, ��galactosidase activity did not overlap with Olig2-expressing OLP at 17.5 dpc and P5, indicating that K23 enhancer has no activity in the glial lineage (Figs. 7F and G). Together, these data indicate that (1) Olig2/K23-cre expression is restricted to pMN domain precursors and progeny, (2) that the K23 enhancer drives expression in proliferating MN precursors and (3) that the kinetics of K23-cre activity corresponds to a peak period of pMN neurogenesis (9.511.5 dpc). The Olig2/K23-cre fate map at later stages remains confined to the MN lineage, confirming that it becomes downregulated in the pMN before commencement of the gliogenic phase at 12.5 dpc. Further, Isl1 co-labeling of K23cre fate mapped animals at 17.5 dpc indicates that certain surviving Isl1+ MN (presumably from an early neurogenic wave) fail to be marked by K23-cre. Together, these data suggest that BAC-encoded OLIG2 rescues several classes of MN developing at a late stage of neurogenesis and similarly that K23-cre generally labels late developing MNs. Discussion Olig genes are expressed in the Shh-responsive pMN domain of the embryonic neural tube that gives rise to motor neuron and oligodendrocyte precursors in successive waves of progeny cell production. Olig genes show ongoing expression during a Shhindependent phase of fetal oligodendrogenesis (Lu et al., 2000; Cai et al., 2005; Vallstedt et al., 2005), and in maturing oligodendrocytes as well as cycling cortical precursors in the adult rodent brain (Hack et al., 2005; KLL and DHR, unpublished observations). Using BAC transgene rescue and motoneuron lineage mapping approaches, we provide evidence that independent regulatory sequences govern baseline expression of Olig2 during the patterning phase versus neurogenesis and gliogenesis in the embryonic neural tube. Olig2 functions in neurogenesis and gliogenesis additional to pMN patterning To investigate determinants of complex Olig2 regulation, we analyzed cis-acting DNA regulatory sequences within a human BAC construct (��AC ���� that contained ��0 kb of upstream sequence and 40 kb of sequence downstream of OLIG2. In transgenic mice, we found that BAC ��-driven expression of OLIG2 was undetectable during the period of pattern formation (89.5 dpc) in the pMN of the mouse neural tube, suggesting that regulatory elements for early pMN expression are located at considerable distance from the OLIG2 ORF. Interestingly, we did observe rare, scattered Olig2+ cells in the region of the pMN of Olig2����mice. Further, in a functional test of such sequences, we achieved rescue of motor neuron and oligodendrocyte precursor development in Olig2����animals. The competence to form MN is temporally restricted (Yamada et al., 1993). Thus, these data indicate that the human BAC-driven expression of OLIG2 in neurogenesis occurs during this window in a spatially appropriate manner. Our data further demonstrate that fully normal patterning is dispensable for embryonic MN or OLP development in vivo. Indeed, in support of this view, Bai et al. (2004) have shown MN generation takes place in Gli-deficient embryos despite gross patterning abnormalities in the ventral neural tube. Motor neuron generation is thought to depend on two critical periods of Shh signaling: an early period during which naive neural plate cells are converted into ventralized progenitors (��9.5 dpc) and a late period during which Shh drives MN differentiation (Ericson et al., 1996). We cannot rule out that small rests of Olig2+ cells, below level of detection, are present in the pMN region of Olig2����BACTg/+ animals at early stages. Indeed, we observed isolated Olig2+ and Ngn2+ cells within the pMN region consistent with this possibility. Therefore, our results do not allow us to conclude that MN generation can occur independently of Olig2 activity in the early period. They do, however, rule out that non-cell autonomous functions of Olig2 are necessary for motor neuron generation in vivo. The low levels of transgene expression we observed at 9.5 d. p.c. is caused in part by the delay of activation of the human OLIG2 transgene locus compared with the endogenous mouse Olig2. The rostral-caudal gradient expression driven by the human BAC transgene matches well with the endogenous Olig2 gene, i.e., Olig2 has higher expression levels in the caudal relative to the rostral neural tube. These two facts account for the low levels of hOLIG2 especially in the rostral neural tube. The Olig2 antibody can recognize both human and mouse forms of Olig2 proteins (Ligon et al., 2004). This point is also established by analysis of human OLIG2 proteins in Olig2-null, BAC transgenic mice (this study). Analysis of human BAC transgene expression using both mRNA probes and antibody consistently showed lower levels of BAC-encoded products compared to endogenous mouse gene and protein levels. Our results also argue against auto-regulation of OLIG2, because levels of human OLIG2 mRNA transcripts are similar in an Olig2 heterozygous or null background. Evidence for dose-dependent effects of Olig2 during embryonic MN development We observed that sub-normal levels of Olig2 proteins resulted in only partial rescue of the MN deficient in Olig2 null embryos. Dose-dependent effects of other bHLH proteins have been observed outside of the CNS. For instance, E2A, encoding the bHLH protein E47, is required for B-cell lymphocyte specification, and E2A gene dosage regulates proB cell numbers (Zhuang et al., 1996; Herblot et al., 2002). Indeed, several lines of evidence suggest that increased Olig2 dosage contributes to MN development, possibly through promoting expression of the proneural bHLH protein, Ngn2. First, peak levels of Olig2 expression are observed at 9.5 dpc, corresponding to the major period of embryonic MN generation (Sun et al., 2001). Second, forced expression of Olig2 results in increased numbers of MN (Novitch et al., 2001). Third, misexpression of Olig2 in the neural tube leads 162 T. Sun et al. / Developmental Biology 292 (2006) 152164 to ectopic expression of Ngn2 (Mizuguchi et al., 2001; Novitch et al., 2001), and Ngn2 expression is lacking in the pMN domain of Olig1/2 compound null animals (Zhou and Anderson, 2002). Though our findings of dose-related MN rescue were obtained in a ��on-physiologic��Olig2����background, they are consistent with the findings generated above with wild type animals (Mizuguchi et al., 2001; Novitch et al., 2001). We observed a significant increase in the number of rescued MNs when the higher copy number BAC-�� transgene allele (line 32) was crossed into Olig2-deficient embryos. One possibility is that higher Olig2 dosage results in augmented pMN patterning and in turn a larger pool of competent precursor cells. However, despite the fact that ��0% of MN were rescued in Olig2����ACTg32/+ mice (three BAC transgene copies) compared with ��5% rescue in Olig2����ACTg34/+ mice (two copies) at 14.5 dpc, we found no evidence for normal pMN patterning in either case. We therefore favor the alternate possibility that Olig2 dosage must reach a critical threshold within pMN cells to generate MN precursors and perhaps their capabilities for self-renewal. In distinction to the studies above, Lee et al. have proposed that high levels of Olig2 inhibit a transcriptional program leading to MN differentiation (Lee et al., 2005). One way to reconcile these apparently disparate findings on Olig2 upregulation during pMN development is to consider distinct early and late stage effects of Olig2 dosage. For instance, it is possible that early Olig2 upregulation is involved in the initial segregation of pMN neuroblasts and activation of Ngn2, whereas subsequent Olig2 downregulation is necessary for MN differentiation. Further work is clearly needed to precisely quantify levels of Olig2 in individual pMN cells and determine the full consequences for cell fate acquisition during the neurogenic and gliogenic periods. cis-acting DNA regulatory sequences for regulation of Olig2 in motoneuron but not oligodendrocyte precursors A related mechanistic question is whether stage- and lineagespecific expression of Olig2 is governed by separable regulatory elements. We used cross-species comparison of intergenetic regions in BAC �� to identify candidate enhancers and functionally screened these by electroporation into the chick neural tube (Timmer et al., 2001). Another screening strategy to identify Olig2 cis-acting regulatory elements in embryonic stems cells has recently been described (Xian et al., 2005). We found evidence for differential Olig2 regulation during the neurogenic, but not patterning or gliogenic phases of development through characterization of the 3.6 kb K23 regulatory element, which lies approximately 11 kb downstream of OLIG2 on human Chromosome 21 (Fig. 5A). Because we observe purdurant ��alactosidase activity lasting until 12.5 d.p.c. in MNs, it is possible that K23 lacks DNA-binding sites for transcriptional repressors of Olig2 immediately at the time of MN lineage commitment. Nevertheless, the K23 regulatory element is sufficient to recapitulate an important aspect of Olig2 regulation in MN-dedicated precursor cells prior to cell cycle exit. The OLIG2/K23 regulatory element failed to mark OLP or oligodendrocytes at any stage examined in four independent transgenic lines, despite the fact that the BAC �� drove expression in OLP and rescued oligodendrocyte development in Olig2����mice. Results with OLIG2/K23-cre, in particular, demonstrate that OLIG2 regulatory sequences for general pMN patterning lie outside of the K23 region. Using both ��alactosidase and highly sensitive cre recombinase reporter alleles, we found that the K23 regulatory element drives expression commencing at approximately 9.259.5 dpc in scattered cells of the pMN domain. Cre fate mapping revealed, moreover, that all K23 activity is restricted to the MN lineage. Interestingly, such a specialized pMN cell type was previously suggested by mapping analysis using retroviruses in chick (Leber and Sanes, 1991). However, because retrovirus will label proliferating cells, these data together do not ambiguously distinguish between two possibilities: (1) Olig2-K23 upregulation in segregating neuroblasts of the pMN or (2) a bi-potent MN-OLP precursor that gives rise to a proliferating dedicate neuroblast (see Fig. 8). A common requirement for function of the transcription factors Gli2 and Nkx6.1/6.2 has been shown for normal development of MN and OLP (Qi et al., 2003; Cai et al., 2005; Vallstedt et al., 2005). We speculate that these or additional proteins (Supplementary Fig. 2) may function directly upstream of Olig2. Recent work shows the pMN Fig. 8. Dynamic regulation of Olig2 during pMN and MN development. During neural tube patterning, baseline levels of Olig2 expression are induced/ maintained by Shh signaling. Our results show that K23-Olig2 upregulation in scattered, proliferating cells of the pMN at 911.5 d.p.c. is specific to developing MN. Subsequently, endogenous Olig2 is downregulated as MN precursors leave the cell cycle and acquire expression of Hb9 and Isl1. Fate mapping analysis indicates that K23-cre labeled cells are dedicated to the MN lineage as no expression in oligodendrocytes was observed. However, because OLIG2-K23 activity is detected in proliferating neuroblasts (NB), our findings do not distinguish between two models of MN lineage progression, either from a bi-potent MN-OLP precursor (NG) or dedicated, independently segregating motoneuroblast (N). T. Sun et al. / Developmental Biology 292 (2006) 152164 163 domain to be heterogeneous both with respect to Olig2 expression levels and the potential of individual progenitor cells (Lee et al., 2005, Park et al., 2004). Therefore, the mosaic expression driven by K23 regulatory sequences appears indicative of the molecular mechanisms underlying pMN heterogeneity and an important clue towards identification of upstream events. Further refinement of K23 sequences by deletion analysis is underway and promises insight into mechanisms that regulate MN-specific gene expression. Acknowledgments The authors wish to thank Qiufu Ma, Paul Gray and David Anderson for comments and stimulating discussions. T.S. is grateful to the National Multiple Sclerosis Society for an Advanced Postdoctoral Fellowship and acknowledges Chris Walsh for advice and support during the course of this work. These studies were funded by grants from the NIH to D.H.R. (NS4051) and C.D.S. (NS047572), and a Collaborative Center Grant from the National Multiple Sclerosis Society. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ydbio.2005.12.047. References Anderson, D.J., 2001. Stem cells and pattern formation in the nervous system: the possible versus the actual. Neuron 30, 1935. Bai, C.B., Stephen, D., Joyner, A.L., 2004. All mouse ventral spinal cord patterning by hedgehog is Gli dependent and involves an activator function of Gli3. Dev. Cell 6, 103115. Cai, J., Qi, Y., Hu, X., Tan, M., Liu, Z., Zhang, J., Li, Q., Sander, M., Qiu, M., 2005. Generation of oligodendrocyte precursor cells from mouse dorsal spinal cord independent of Nkx6 regulation and Shh signaling. Neuron 45, 4153. Ericson, J., Morton, S., Kawakami, A., Roelink, H., Jessell, T.M., 1996. Two critical periods of Sonic Hedgehog signaling required for the specification of motor neuron identity. Cell 87, 661673. Hack, M.A., Saghatelyan, A., de Chevigny, A., Pfeifer, A., Ashery-Padan, R., Lledo, P.M., Gotz, M., 2005. Neuronal fate determinants of adult olfactory bulb neurogenesis. Nat. Neurosci. 8, 865872. Heintz, N., 2001. BAC to the future: the use of bac transgenic mice for neuroscience research. Nat. Rev., Neurosci. 2, 861870. Herblot, S., Aplan, P.D., Hoang, T., 2002. Gradient of E2A activity in B-cell development. Mol. Cell. Biol. 22, 886900. Jessell, T.M., 2000. Neuronal specification in the spinal cord: inductive signals and transcriptional codes. Nat. Rev., Genet. 1, 2029. Leber, S.M., Sanes, J.R., 1991. Lineage analysis with a recombinant retrovirus: application to chick spinal motor neurons. Adv. Neurol. 56, 2736. Lee, S.K., Lee, B., Ruiz, E.C., Pfaff, S.L., 2005. Olig2 and Ngn2 function in opposition to modulate gene expression in motor neuron progenitor cells. Genes Dev. 19, 282294. Lu, Q.R., Yuk, D., Alberta, J.A., Zhu, Z., Pawlitzky, I., Chan, J., McMahon, A.P., Stiles, C.D., Rowitch, D.H., 2000. Sonic hedgehog-regulated oligodendrocyte lineage genes encoding bHLH proteins in the mammalian central nervous system. Neuron 25, 317329. Lu, Q.R., Sun, T., Zhu, Z., Ma, N., Garcia, M., Stiles, C.D., Rowitch, D.H., 2002. Common developmental requirement for Olig function indicates a motor neuron/oligodendrocyte connection. Cell 109, 7586. Ma, Q., Sommer, L., Cserjesi, P., Anderson, D.J., 1997. Mash1 and neurogenin1 expression patterns define complementary domains of neuroepithelium in the developing CNS and are correlated with regions expressing notch ligands. J. Neurosci. 17, 36443652. Marti, E., Bumcrot, D.A., Takada, R., McMahon, A.P., 1995. Requirement of 19 K form of Sonic hedgehog for induction of distinct ventral cell types in CNS explants. Nature 375, 322325. Mizuguchi, R., Sugimori, M., Takebayashi, H., Kosako, H., Nagao, M., Yoshida, S., Nabeshima, Y., Shimamura, K., Nakafuku, M., 2001. Combinatorial roles of olig2 and neurogenin2 in the coordinated induction of pan-neuronal and subtype-specific properties of motoneurons. Neuron 31, 757771. Novitch, B.G., Chen, A.I., Jessell, T.M., 2001. Coordinate regulation of motor neuron subtype identity and pan-neuronal properties by the bHLH repressor Olig2. Neuron 31, 773789. Novitch, B.G., Wichterle, H., Jessell, T.M., Sockanathan, S., 2003. A requirement for retinoic acid-mediated transcriptional activation in ventral neural patterning and motor neuron specification. Neuron 40, 8195. Orentas, D.M., Hayes, J.E., Dyer, K.L., Miller, R.H., 1999. Sonic hedgehog signaling is required during the appearance of spinal cord oligodendrocyte precursors. Development 126, 24192429. Park, H.C., Mehta, A., Richardson, J.S., Appel, B., 2002. olig2 is required for zebrafish primary motor neuron and oligodendrocyte development. Dev. Biol. 248, 356368. Park, H.C., Shin, J., Appel, B., 2004. Spatial and temporal regulation of ventral spinal cord precursor specification by Hedgehog signaling. Development 131, 59595969. Qi, Y., Tan, M., Hui, C.C., Qiu, M., 2003. Gli2 is required for normal Shh signaling and oligodendrocyte development in the spinal cord. Mol. Cell. Neurosci. 23, 440450. Roelink, H., Porter, J.A., Chiang, C., Tanabe, Y., Chang, D.T., Beachy, P.A., Jessell, T.M., 1995. Floor plate and motor neuron induction by different concentrations of the amino-terminal cleavage product of sonic hedgehog autoproteolysis. Cell 81, 445455. Rowitch, D.H., Echelard, Y., Danielian, P.S., Gellner, K., Brenner, S., McMahon, A.P., 1998. Identification of an evolutionarily conserved 110 base-pair cis-acting regulatory sequence that governs Wnt-1 expression in the murine neural plate. Development 125, 27352746. Sharma, K., Sheng, H.Z., Lettieri, K., Li, H., Karavanov, A., Potter, S., Westphal, H., Pfaff, S.L., 1998. LIM homeodomain factors Lhx3 and Lhx4 assign subtype identities for motor neurons. Cell 95, 817828. Soriano, P., 1999. Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat. Genet. 21, 7071. Soula, C., Danesin, C., Kan, P., Grob, M., Poncet, C., Cochard, P., 2001. Distinct sites of origin of oligodendrocytes and somatic motoneurons in the chick spinal cord: oligodendrocytes arise from Nkx2.2-expressing progenitors by a Shh-dependent mechanism. Development 128, 13691379. Sun, T., Echelard, Y., Lu, R., Yuk, D.I., Kaing, S., Stiles, C.D., Rowitch, D.H., 2001. Olig bHLH proteins interact with homeodomain proteins to regulate cell fate acquisition in progenitors of the ventral neural tube. Curr. Biol. 11, 14131420. Takebayashi, H., Nabeshima, Y., Yoshida, S., Chisaka, O., Ikenaka, K., 2002. The basic helixloophelix factor olig2 is essential for the development of motoneuron and oligodendrocyte lineages. Curr. Biol. 12, 11571163. Temple, S., 2001. The development of neural stem cells. Nature 414, 112117. Timmer, J., Johnson, J., Niswander, L., 2001. The use of in ovo electroporation for the rapid analysis of neural-specific murine enhancers. Genesis 29, 123132. Vallstedt, A., Klos, J.M., Ericson, J., 2005. Multiple dorsoventral origins of oligodendrocyte generation in the spinal cord and hindbrain. Neuron 45, 5567. Xian, H.Q., Werth, K., Gottlieb, D.I., 2005. Promoter analysis in ES cell-derived neural cells. Biochem. Biophys. Res. Commun. 327, 155162. Yamada, T., Pfaff, S.L., Edlund, T., Jessell, T.M., 1993. Control of cell pattern in the neural tube: motor neuron induction by diffusible factors from notochord and floor plate. Cell 73, 673686. Yang, X.W., Wynder, C., Doughty, M.L., Heintz, N., 1999. BAC-mediated gene-dosage analysis reveals a role for Zipro1 (Ru49/Zfp38) in progenitor cell proliferation in cerebellum and skin. Nat. Genet. 22, 327335. 164 T. Sun et al. / Developmental Biology 292 (2006) 152164 Zhou, Q., Choi, G., Anderson, D.J., 2001. The bHLH transcription factor Olig2 promotes oligodendrocyte differentiation in collaboration with Nkx2.2. Neuron 31, 791807. Zhuang, Y., Cheng, P., Weintraub, H., 1996. B-lymphocyte development is regulated by the combined dosage of three basic helixloophelix genes, E2A, E2-2, and HEB. Mol. Cell. Biol. 16, 28982905. Zhou, Q., Anderson, D.J., 2002. The bHLH transcription factors OLIG2 and OLIG1 couple neuronal and glial subtype specification. Cell 109, 6173. Zhou, Q., Wang, S., Anderson, D.J., 2000. Identification of a novel family of oligodendrocyte lineage-specific basic helixloophelix transcription factors. Neuron 25, 331343.