RESEARCH ARTICLE

A zinc finger protein that regulates oligodendrocyte specification,migration and myelination in zebrafishHarwin Sidik and William S. Talbot*

ABSTRACTPrecise control of oligodendrocyte migration and development iscrucial for myelination of axons in the central nervous system (CNS),but important questions remain unanswered about the mechanismscontrolling these processes. In a zebrafish screen for myelinationmutants, we identified amutation in zinc finger protein 16-like (znf16l).znf16l mutant larvae have reduced myelin basic protein (mbp)expression and reduced CNS myelin. Marker, time-lapse andultrastructural studies indicated that oligodendrocyte specification,migration andmyelination are disrupted in znf16lmutants. Transgenicstudies indicated that znf16l acts autonomously in oligodendrocytes.Expression of Zfp488 from mouse rescued mbp expression in znf16lmutants, indicating that these homologs have overlapping functions.Our results defined the function of a new zinc finger protein withspecific function in oligodendrocyte specification, migration andmyelination in the developing CNS.

KEY WORDS: Myelin, Oligodendrocyte, Zebrafish

INTRODUCTIONPrecise temporal control of cell fate specification and migration isessential to coordinate the development of different cell types. In thenervous system, for example, newly differentiated neurons extendaxons over long distances to form connections with theirpostsynaptic targets (Araújo and Tear, 2003). Oligodendrocytes,the myelinating cells of the central nervous system (CNS), provideanother example. During development, oligodendrocyte precursorcells (OPCs) migrate throughout the CNS. After completing theirmigration, some OPCs differentiate as oligodendrocytes andmyelinate their associated axons immediately, whereas otherOPCs delay differentiation until much later stages (Young et al.,2013). Formation of new myelin from OPCs in the adult is essentialfor motor learning, underscoring the importance of the precisetemporal control of differentiation of myelinating oligodendrocytesin the healthy CNS (Scholz et al., 2009; McKenzie et al., 2014). Insome demyelinated lesions in individuals with multiple sclerosis,OPCs are present but apparently unable to form myelin (Kotteret al., 2006; Boyd et al., 2013), highlighting the need to understandthe temporal control of oligodendrocyte maturation in health anddisease.OPCs are derived from a group of Olig2-expressing progenitors

in the motoneuron progenitor (pMN) domain in the ventral CNS.Cells of the pMN domain generate motoneurons at early stages,when they express Olig2 and the neurogenic factor Ngn2(Mizuguchi et al., 2001; Novitch et al., 2001), and they form

OPCs at later stages, when they express Olig2 and the gliogenicfactor Sox10 (Stolt et al., 2002; Pozniak et al., 2010). This transitionbetween neurogenic fate and gliogenic fate is mediated by Notchreceptors and their ligands (Wang et al., 1998; Park and Appel,2003), organizing the pMN domain into temporal compartments.Diverse signals, including Pdgf and Netrin-1, direct OPC migrationaway from the midline and towards the axons that they eventuallymyelinate (de Castro and Bribián, 2005; Tsai et al., 2006; Miyamotoet al., 2008). After contacting their target axons and initiatingdifferentiation, maturing oligodendrocytes express transcriptionfactors, such as Zfp191 and Myrf, that initiate and maintainthe expression of the genes essential for myelination (Emeryet al., 2009; Howng et al., 2010). Despite much progress inidentifying essential regulators of oligodendrocyte developmentand myelination, many aspects of the developmental program arenot well understood. In particular, little is known about the factorsthat specify OPCs within the pMN domain and initiate theirmigration.

To define new genes that regulate oligodendrocyte specificationand differentiation, we have conducted a genetic screen formutations that disrupt myelin gene expression in zebrafish. In thisstudy, we report that the zinc finger protein Znf16-like (Znf16l)regulates the onset of OPC migration and differentiation. In lateembryonic stageswhenOPCs normally begin their migration, znf16lmutants had Olig2-expressing progenitors at the CNS midline butlacked OPCs in most regions of the CNS. Time-lapse imagingrevealed that although OPCs were present in the mutants at theseearly stages, their migration was delayed. At later stages, OPCmigration recovered in the mutants, but CNS myelin remainedsignificantly reduced. Through transgenic analyses, we determinedthat Znf16l acts autonomously in the oligodendrocyte lineage. Wealso discovered that the normal OPC development could be rescuedin znf16l mutants by transgenic expression of the mammalian zincfinger protein Zfp488, which was previously shown to cooperatewith Notch signaling to specify OPC fate in the chicken (Wang et al.,2006). Lastly, we showed that despite similar myelination defects inboth znf16lmutants and previously studied notch3mutants (Zauckeret al., 2013), the two genes have different roles in regulatingdevelopment of the oligodendrocyte lineage. These findingsindicate that Znf16l is an essential regulator of specification,migration and myelination in the oligodendrocyte lineage.

RESULTSZnf16-like is essential for normal myelination in the CNSIn a genetic screen for zebrafish mutants with abnormalmyelination, we identified st78 as a recessive mutation thatreduced expression of myelin basic protein (mbp) mRNAexpression in the CNS at 5 days postfertilization (dpf; Fig. 1A).Homozygous st78 mutants exhibited normal mbp expression in theperipheral nervous system (PNS) and normal gross morphology at5 dpf (Fig. 1A,B). High-resolution meiotic mapping localized theReceived 6 July 2015; Accepted 1 October 2015

Department of Developmental Biology, Stanford University, CA 94305, USA.

*Author for correspondence (wtalbot@stanford.edu)

4119

© 2015. Published by The Company of Biologists Ltd | Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

st78mutation to a 0.9 Mb region of linkage group 7 (LG7; Fig. 1C).Sequence analyses of genomic DNA from st78 mutants revealed aG-to-T transversion in the previously uncharacterized geneXM_694039, which is predicted to encode a zinc finger protein,Zinc finger 16-like (Znf16l, also known as Zinc finger 697-like/Znf697l). This mutation changed a cysteine in one of the sixpredicted Cys2His2 zinc fingers of Znf16l to a phenylalanine(C428→F; Fig. 1D,E).To construct a null mutation in znf16l, we generated a second

mutant allele using transcription activator-like effector nucleases(TALEN; Sanjana et al., 2012). We identified an allele, st97, witha deletion of coding nucleotides 133-136 of znf16l (Fig. 1E,F).This 4 bp deletion caused a frameshift in the open reading frame,resulting in a premature stop codon upstream of all six predictedzinc fingers in the Znf16-like protein. Like homozygotes for thest78 point mutation, homozygous st97 mutants lacked mbpexpression in the CNS at 5 dpf and exhibited normal grossmorphology (Fig. 1G,H). Furthermore, the st97 frameshiftmutation failed to complement the st78 point mutation: st78/st97transheterozygotes also lacked mbp expression at 5 dpf (Fig. 1G).These results confirmed that Znf16l is essential for normal mbpexpression in the larval CNS and further suggested that thedisruption of one zinc finger module in st78 mutants abolished thefunction of Znf16-like.

Specific defects in oligodendrocyte lineage in znf16lmutantsIn the developing spinal cord, most oligodendrocytes are derivedfrom a ventral region known as the pMN domain, which also givesrise to motoneurons (Rowitch, 2004). At 24 hours postfertilization(hpf), we detected no difference in the expression of pMN markers

between znf16l mutants and wild-type siblings (Fig. 2A,B),indicating that the oligodendrocyte defects did not arise from adelay in the formation of the pMN domain. Likewise, motoneurondevelopment appeared similar in the mutants and wild-typesiblings, based on the comparable expression patterns of themotoneuron marker islet1 at 52 hpf (Fig. 2C). By contrast, at thissame stage (52 hpf ), we detected differences between the mutantsand wild-type siblings in the expression of sox10 and olig2 inoligodendrocytes (Fig. 2D,E). olig2-expressing cells had migratedwidely in the brain and the spinal cord of the wild-type siblings atthis stage, whereas olig2 expression in the mutant was restricted tothe ventral midline (Fig. 2D). We also examined expression ofsox10, which is restricted to cells in the oligodendrocyte lineage inthe CNS. At 52 hpf, sox10-expressing oligodendrocyte-lineagecells were evident in the ventral spinal cord of wild-type embryosbut not in mutants (Fig. 2E′). These results support the possibilitythat znf16l functions specifically in the development of cells in theoligodendrocyte lineage.

Loss of znf16l function delays the onset of oligodendrocytemigrationTo define the role of znf16l in oligodendrocyte lineagedevelopment, we examined znf16l mutants bearing the Tg(olig2:GFP) transgene in timecourse and time-lapse studies. Thistransgene is expressed in OPCs and throughout differentiation andmaturation of myelinating oligodendrocyte-lineage cells (Shin et al.,2003; Zannino and Appel, 2009), allowing observation of differentstages of the oligodendrocyte lineage. At 60 and 84 hpf, weobserved significantly fewer olig2-expressing cells in the forebrain,midbrain and hindbrain of the mutants compared with their wild-type siblings (Fig. 3A,B,F). By 108 hpf, however, the number of

Fig. 1. Mutations in znf16-like reduce mbp mRNAexpression in the CNS. (A) At 5 dpf, expression ofmbpwas specifically reduced in the CNS but not the PNS ofst78 homozygote compared with the heterozygoussibling. (B) st78mutant had normal gross morphology at5 dpf compared with the wild-type sibling. (C) Geneticmapping linked st78 locus to XM_694039 in linkagegroup 7. (D) Sequence chromatogram shows the lesion(red box) in the coding sequence of znf16-like(XM_694039). The point mutation in st78 changes acysteine in a Cys2His2 zinc finger of Znf16l to aphenylalanine. (E) Schematic diagram depicting Znf16-like protein with all its zinc fingers, the location of thetransversion in the st78 allele, and the 4 bp deletionresulting in a frameshift in the st97 allele. (F) Nucleotidesequence of znf16l and the TALE nuclease targetedsequence (black) and diagnostic restriction enzyme site(red). Bottom line shows the st97 allele and thecorresponding amino acid sequence resulting from theframeshift. (G) At 5 dpf, expression of mbp was reducedin the CNS of st97 homozygote in comparison to theheterozygous sibling. st97 failed to complement st78,and the transheterozygous mutant lacked mbp in theCNS, similar to the homozygous mutants. Thegenotypes of all larvae shown in A,B,G,H weredetermined by PCR tests for their respective lesions.

4120

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

olig2-expressing cells in the mutants was comparable to that of thewild-type siblings (Fig. 3C,F).Time-lapse studies showed that olig2:GFP-expressing OPCs

migrated from the midline of wild-type embryos between 48 and52 hpf, but few if any OPCs migrated from the midline in znf16lmutants at these stages (Movies 1, 2; Fig. 3D). In the mutants,OPCs were observed migrating from the midline, starting at about60 hpf (Movies 3, 4; Fig. 3E). After OPCs began migrating inznf16l mutants, they appeared to move over longer distances andover longer periods of time than their wild-type counterparts. Toquantify this, we tracked migration of 30 OPCs in znf16l mutantsand wild-type siblings (ten cells per fish from three mutants andthree wild-type siblings) and measured their displacement overtime. In the 64-72 hpf time period, OPCs migrated faster inmutants than in their wild-type siblings (3.75±0.36 versus 2.20±0.18 μm/h; Fig. 3G). Mutant OPCs also spent a larger fraction ofthe observation period in movement (0.33±0.02 versus 0.20±0.01;Fig. 3H). We also noted that some mutant OPCs migrated in closeproximity to the midline, sometimes re-entering the midlineinstead of migrating away and dispersing throughout the CNS aswild-type OPCs did. Taken together, these observations indicatedthat the onset of migration of OPCs is delayed in znf16l mutantsand that some mutant OPCs displayed aberrant migratorybehavior.

Reduction of mature oligodendrocytes in znf16l mutantsTo investigate the progression of oligodendrocyte development inznf16l mutants, we examined the expression of claudin k (cldnk), amarker of mature oligodendrocyte-lineage cells in the CNS of thezebrafish larva (Münzel et al., 2012). At 3 dpf, znf16lst78 mutants

expressed little or no detectable cldnk in the CNS, in contrast to theirwild-type siblings (Fig. 4A). At 5 dpf, the mutant larvae expressedsome cldnk in the CNS but at a much lower level than theirwild-type siblings (Fig. 4B). To determine whether mutantoligodendrocytes might initiate myelination at later stages, weanalyzed the expression of mbp at 9 dpf (Fig. 4C). Despite thestrong reduction of mbp expression at 5 dpf (Fig. 1A), znf16lst78

mutants had detectable mbp expression at 9 dpf, although the levelof expression was less than in their wild-type siblings. These resultsindicated that oligodendrocyte maturation is delayed in the mutants,despite the recovery in OPC number that occurred at earlier stages.

Myelin is reduced in znf16l mutantsTo assess whether the abnormal expression of oligodendrocytemarkers reflects any defects in myelin in znf16l mutants, weexamined the ultrastructure of the myelinated axons in the ventralspinal cord. In accordance with the marker analyses, the number ofmyelinated axons in the mutants was significantly lower than intheir wild-type siblings at 5 dpf (1.7±0.8 per hemisegment of theventral spinal cord in mutants, n=6, compared with 18.7±0.9 in wildtype, n=6; Fig. 4D,E,H). At 9 dpf, the number of myelinated axonsincreased in mutants (7.7±0.7, n=3), but the number remainedsignificantly reduced compared with the wild-type siblings at thesame stage (27.3±0.3, n=3; Fig. 4F-H). These results indicated thatznf16l is required for myelination for at least several days after OPCmigration is complete. Although myelin was reduced in znf16lmutants, some myelin did form and some homozygous mutantssurvived to become fertile adults (Fig. 4I; in the progeny ofheterozygous intercrosses: 6.8% homozygous st78 mutants at90 dpf, n=177; 7.2% homozygous st97 mutants at 90 dpf, n=111).

Fig. 2. Developmental disruptions in znf16l mutants are specific to the oligodendrocyte lineage. (A,B) znf16l mutants displayed normal expression ofnkx2.2 (A) and olig2 (B) at 24 hpf. (C,C′) znf16l mutants also displayed normal expression of the motoneuron marker islet1 at 52 hpf. (D,E) Defects in znf16lmutants were specific to cells in the oligodendrocyte lineage. (D-D‴) olig2 is expressed by oligodendrocyte-lineage cells and a subset of motoneurons. Theexpression pattern of olig2 in wild-type siblings revealed groups of cells that had migrated dorsally (arrowhead), ventrally (white brackets) and laterally (box).These cell populations are missing or reduced in themutants. D′ and D″ show regions in black and red boxes, respectively, in D. (E) sox10 expression is restrictedto cells of the oligodendrocyte lineage in the CNS. Wild-type siblings displayed a group of cells (brackets in E′) immediately dorsal to the notochord that weremissing in the mutants, confirming the disruption of oligodendrocyte-lineage cells in znf16l mutants. The genotypes of all embryos shown were determined byPCR tests for the st78 lesion. Scale bars: 100 μm.

4121

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

Znf16l functions autonomously in oligodendrocytes forproper CNS myelinationTo determine which cell types in the CNS require Znf16l functionfor myelination, we performed transgenic rescue experiments usingfull-length Znf16l expressed under the control of different tissue-specific regulatory elements. Control uninjected mutants wereraised and stained alongside to ensure that rescued mbp expressionwas not a result of the weak expression of mbp seen in older mutantembryos. In transient transgenic assays, we analyzed constructs thatexpressed Znf16l in neurons (huC, mnx1), oligodendrocytes (cldnk,sox10) and macrophages or microglia (mpeg1; Fig. 5A,B). Thecldnk:znf16l and sox10:znf16l transgenes rescued mbp expressionin the CNS of mutants at 5 dpf (cldnk: 45% of the mutants, n=77;sox10: 57% of the mutants, n=47; Fig. 5B,C). No rescue wasdetected with the mnx1:znf16l construct (Fig. 5B). The huc:znf16lconstruct rescued at a low frequency (17% of the mutants, n=59;Fig. 5B,C), perhaps because of expression of the huC regulatory

sequences in some cells of the oligodendrocyte lineage (Fig. S1).The mpeg1:znf16l construct did not rescue mbp expression in themutants at 5 dpf (Fig. 5B). These results provide evidence thatznf16l acts autonomously in OPCs and early oligodendrocytes topromote CNS myelination.

Mouse Zfp488, but not Zfp191, can rescue znf16l mutantsSequence analysis did not identify a clear ortholog of Znf16l inmammals. Previous studies have identified several zinc fingerproteins that have specific functions in the oligodendrocyte lineagein mammals, including Zfp191, which is required for matureoligodendrocyte-lineage cells to myelinate (Howng et al., 2010),and Zfp488, which can promote oligodendrocyte lineagedifferentiation and maturation in combination with Notchsignaling (Wang et al., 2006). The zebrafish genome contains anortholog of Zfp191, but no gene clearly orthologous to Zfp488.Thus, we analyzed the previously characterized mouse genes in

Fig. 3. Disruption of OPC migration in znf16l mutants. (A-C) znf16lst78 mutants and wild-type siblings in Tg(olig2:GFP) background. The number of OPCs isgreatly reduced in the mutants at 60 hpf (A,A′) and 84 hpf (B,B′), but recovers quickly by 108 hpf (C,C′). (D,E) Dorsal view of time-lapse imaging from 48-52 and64-72 hpf, with trajectories of 10 individual OPCs as they migrated laterally out of the midline, shown using the MTrackJ cell-tracking tool (n=3 fish per genotype,per time point). Cells were tracked every 15 min in real time (see Movies 1-4 for time-lapse images). Genotypes of all embryos were determined by PCRafter imaging. (F) Quantification of the number of GFP-expressing cells in the hindbrain of Tg(olig2:GFP) znf16lst78mutants and wild-type siblings (n=15 wild type,n=9mutant). The numbers were significantly different at 60 and 84 hpf, but not at 108 hpf (two-tailed Student’s t-test, ****P<0.0001). (G) Quantification of averageOPC displacement (μm) per hour of 30 tracked cells from each timepoint. Mutant OPCs traveled farther than wild-type siblings. (H) Quantification of the fraction oftime for which each OPC was actively moving. Mutant OPCs spent more time moving compared with wild-type siblings at either stage. Error bars show s.e.m.significance with one-way ANOVA and post hoc comparisons. ***P=0.0003 in G; ****P<0.0001 in H.

4122

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

these functional studies. To test whether either of these zinc fingerproteins shares any functional overlap with Znf16l, we performed atransgenic rescue experiment with full-length mZfp191 or mZfp488under the control of sox10 promoter (Fig. 5A). Zfp488 rescuedmbpexpression in the CNS of znf16l mutants (47% of the mutants,n=30), whereas expression of Zfp191 had little or no effect (2.9% ofthe mutants, n=35; Fig. 5B,C). These data suggest that mammalianZfp488 and zebrafish Znf16l have a shared function in promotingoligodendrocyte-lineage cell migration and differentiation despitethe divergent sequences of these zinc finger proteins.

Different requirements for znf16l and notch3 inoligodendrocyte developmentPrevious analysis showed that notch3 mutants share phenotypicsimilarities with znf16l mutants (Zaucker et al., 2013). In notch3mutants, an early reduction in olig2-expressing OPCs underlies areduction in CNS mbp expression (Zaucker et al., 2013). In notch3mutants, mbp expression begins to recover by 7 dpf, and somehomozygous notch3 mutants survive to adulthood (Zaucker et al.,2013), similar to znf16l mutants. In addition, previous work in thechicken indicates that Zfp488, which can compensate for loss ofZnf16l in our in vivo rescue assays, cooperates with Notch signaling,because overexpression of Zfp488 promotes ectopic oligodendrocyteprecursor formation only when Notch signaling is activated (Wanget al., 2006).To investigate the possible relationship of Znf16l and Notch3 in

oligodendrocyte development, we examined the expression of early

markers in both mutants in parallel. As previously described(Zaucker et al., 2013), notch3st51 and notch3ZM homozygousmutants show marked reductions in expression of deltaD andnotch3, whereas heterozygous siblings had increased expression ofboth genes compared with wild-type siblings (Fig. 6B,C,E,F). Bycontrast, znf16lst78 mutants expressed both deltaD and notch3 at thesame level as their wild-type siblings (Fig. 6A,D), suggesting thatZnf16l and Notch3 have distinct functions, despite the similartimecourse of mbp reduction and recovery in the mutants. Inaddition, we observed that znf16l mutants lacked a ventralpopulation of olig2-expressing progenitors in the hindbrain(Fig. 2D′; Fig. 6G″), whereas notch3 mutants lack a more dorsalpopulation (Fig. 6H″,I″). Furthermore, mutants in these two genesdisplayed differences in the lateral migration of olig2-expressingOPCs (Fig. 2D‴; Fig. 6G′-I′). These results indicate that Znf16l andNotch3 have different effects on OPCs in the developing embryo.

Severe oligodendrocyte defects in znf16l;notch3 doublemutantsTo investigate the relationship of znf16l and notch3 inoligodendrocyte development further, we analyzed znf16l;notch3double mutants. The double mutants had a greater reduction ofolig2-expressing OPCs at 52 hpf compared with either singlemutant (Fig. 7A-H). The double mutants also lacked laterallymigrating OPCs, similar to znf16l mutants (Fig. 7C,D,G,H). By4 dpf, znf16l and notch3 single mutants significantly recovered thenumber and distribution of their OPCs (Fig. 7I-K), but znf16l;

Fig. 4. CNS myelin is reduced in znf16lmutants. (A-C) Expression of mature oligodendrocyte markers claudin k (A,B) and mbp (C) was reduced in znf16lst78

mutants between 3 and 9 dpf. Expression of cldnk was reduced but detectable at 5 dpf (B), and expression of mbp was reduced but detectable at 9 dpf (C).(D,E) Transverse transmission electron microscopy images of ventral spinal cord showed that many wild-type axons were already myelinated in wild type (D) butnot mutants (E). (F,G) At 9 dpf, more axons were myelinated in the mutants (G) than at 5 dpf, but the number was still much less than in wild-type siblings at9 dpf (F). (H) Quantification of the number of myelinated axons in ventral spinal cord at 5 and 9 dpf. Error bars show s.d.; significance was determined withtwo-tailed Student’s t-test. ***P<0.001 in H. (I) Adult znf16l homozygous mutants are viable and fertile, with no gross morphological defects compared with theirwild-type siblings for both of the alleles. Genotypes of all fish analyzed were determined by PCR assay.

4123

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

notch3 double mutants were mostly devoid of OPCs (Fig. 7L). At5 dpf, notch3 single mutants had reduced mbp expression in thehindbrain and in spinal cord (Fig. 7N), whereas znf16l mutantslackedmbp in both the hindbrain and the spinal cord (Fig. 7O). Likeznf16l single mutants, the znf16l;notch3 double mutants also lackedmbp in the CNS at 5 dpf (Fig. 7P). At 9 dpf, the double mutants stillhad a strong reduction ofmbpmRNA in the CNS (Fig. 7T), whereasmbp expression had recovered in both znf16l and notch3 singlemutants (Fig. 7R,S).

DISCUSSIONMany studies indicate that Hedgehog signaling activity specifiesthe progenitors in the pMN domain in the ventral neural tube(Richardson et al., 2006) and that these pMN progenitorssequentially form motoneurons at early stages of embryogenesisand oligodendrocyte precursor cells at later stages (Wu et al., 2006).Notch signaling activity is important for this transition (Wang et al.,1998; Genoud et al., 2002; Park and Appel, 2003; Kim et al., 2008;Zaucker et al., 2013), but the other factors that direct pMNprogenitors to become migratory OPCs are not well understood. Ourstudy demonstrates that Znf16l is essential for neural progenitors tobecome migratory OPCs. In znf16l mutants, early patterning of thepMN domain is normal, but oligodendrocyte precursors arespecifically delayed in initiating their migration from the midlineof the CNS. Moreover, CNS myelin is greatly diminished in znf16lmutants. Despite the severe disruption of OPC migration in mutant

embryos, migratory OPCs are evident in the mutant CNS at earlylarval stages. CNS myelin also partly recovers in larval mutants, butmyelin remained significantly reduced in mutants at stages afterOPCs have dispersed throughout the CNS. This reduction in myelinindicates that znf16l also regulates the onset of myelination, inaddition to early events in OPC specification and migration.Expression of Znf16l in the oligodendrocyte lineage rescues mbpexpression in the mutants, indicating that Znf16l acts autonomouslyin these cells. Our analysis identifies Znf16l as a key autonomousregulator of OPC specification, migration and myelination.

In light of the severe defects in oligodendrocyte development inznf16l mutant embryos and early larvae, it is interesting that someCNS myelin is present at later stages. Likewise, previous work inzebrafish indicates that the notch3 gene is required for OPCspecification and myelination in the early larva and that myelinpartly recovers at later stages (Zaucker et al., 2013). These studiespoint to redundancy in the control of oligodendrocyte developmentin zebrafish. There are at least two levels at which this redundancymight operate. First, it is possible that there are genetically distinctpopulations of OPCs that are able to compensate for loss of eachother. Second, it possible that different members of the Zinc fingerand Notch receptor families have overlapping functions, such thathomologous genes are able to substitute for each other at somestages.

Previous studies discovered heterogeneity in oligodendrocytepopulations in mammals (Tomassy and Fossati, 2014). Spatially,oligodendrocytes arise in distinct regions of the neocortex (Kessariset al., 2006), cerebellum (Buffo and Rossi, 2013), dorsal spinal cordand ventral spinal cord (Cai et al., 2005). Temporally, multiplewaves of oligodendrogenesis occur in the healthy, developing CNS(Kessaris et al., 2006). Ablation studies show that these OPCpopulations can compensate for the loss of each other, highlightingthe redundancy in the lineage despite the apparent heterogeneity inorigins of OPCs (Kessaris et al., 2006). Our study is consistent with,but does not demonstrate conclusively, the possibility that suchredundancy in the oligodendroglial lineage also exists in thezebrafish. Different domains of olig2 expression are disrupted inznf16l and notch3 single mutants, and the combined loss of bothNotch3 and Znf16l in the double mutants resulted in a more severedisruption of oligodendrocyte development and myelination. Thus,it is possible that znf16l and notch3 are required in distinctpopulations of OPCs that can compensate for loss of each other. Aconclusive test of this possibility will require identification ofmolecular markers that distinguish different populations of OPCs.

There is also evidence for redundancy of individual genefunctions in the control of oligodendrocyte development. Forexample, Notch receptors other than Notch3, notably Notch1,regulate neural progenitor differentiation to promote gliogenesis(Genoud et al., 2002). It is likely that Notch1 activity can partlycompensate for the loss of Notch3, contributing to the recovery ofCNS myelination in notch3 mutants. Likewise, our in vivo rescueexperiments indicate that a mammalian homolog of znf16l, Zfp488,can substitute for znf16l despite their sequence divergence.Although we have not detected an ortholog of Zfp488 in thezebrafish genome, its characteristic structure of two zinc fingermotifs flanking a nuclear localization signal is shared with anotherprotein, PR domain containing 8 (Prdm8), that is conserved amongzebrafish, mice and humans (Wang et al., 2006; Ross et al., 2012).Prdm8 has been shown to interact with Bhlhb5, a basic helix-loop-helix protein related to Olig2, to form a repressor complex in thecontext of neurogenesis (Ross et al., 2012). It will be of interest toexamine the role of prdm8 and the possibility that Prdm8 and Znf16l

Fig. 5. Znf16l functions cell autonomously in the oligodendrocytelineage. (A) Diagram of the transgenic constructs driving expression of znf16lby tissue-specific gene promoters in neurons (huC), oligodendrocytes (cldnk,sox10) or macrophages (mpeg). Bottom panel shows another transgenicconstruct strategy to express mammalian zinc finger proteins, mZfp191 andmZfp488, in oligodendrocytes under the control of the sox10 promoter. (B) Plotquantifies the percentage of mbp rescue using different tissue promoters anddifferent zinc finger proteins. (C) Representative images of mbp rescue incldnk:znf16l-, sox10:znf16l-, huc:znf16l- and sox10:mZfp488-injected znf16lmutants. There is no visible difference in the level of rescues with the differentconstructs. Genotypes of all fish analyzed were determined by PCR assay.

4124

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

have overlapping functions in the control of oligodendrocytespecification, migration and myelination.Future work is required to identify Znf16l target genes, but genes

controlling OPC migration are among the likely candidates. Netrin-1 in the ventral CNS repels migrating OPCs, and in Netrin-1mutants, OPCs remain near the CNS midline rather than dispersingthroughout the CNS as they do in wild type (Tsai et al., 2006). Thephenotypic similarity between Netrin-1 and znf16l mutantssuggested that receptors for Netrin-1 or other midline repellantsmight be among the genes activated by Znf16l. In preliminaryanalyses, however, we observed normal expression of the Netrinreceptor genes dcc and unc5a in znf16l mutants (data not shown),suggesting that that Znf16l promotes OPC migration by othermechanisms.

Concluding remarksPrecise control of different steps of OPC development is criticalfor the development, function and repair of the CNS, but theregulatory mechanisms controlling these processes are not wellunderstood. We have identified Znf16-like as a novel regulatorof OPC specification, migration and myelination. Our in vivorescue assays established that Znf16-like shows functionaloverlap with its mammalian homolog, Zfp488. Comparison ofdefects in znf16l and notch3 mutants indicates a degree ofredundancy in oligodendrocyte development. Many znf16lmutants survive to adulthood, providing an opportunity to test

the role of this gene in response to CNS injury and myelinrepair.

MATERIALS AND METHODSZebrafish lines and embryosExperiments involving zebrafish were approved by the Stanford UniversityInstitutional Animal Care and Use Committee. Embryos from wild-type(TL, AB/TU and WIK), Tg(olig2:GFP) (Park et al., 2002), znf16lst78,znf16lst97, notch3st51 and notch3ZM (Zaucker et al., 2013) strains were raisedat 28.5°C and staged as described by Kimmel et al. (1995). Embryos weretreated with 0.003% 1-phenyl-2-thiourea (PTU) in Methylene Blue embryowater to inhibit pigmentation.

N-ethyl-N-nitrosourea mutagenesis, genetic screen and geneticmappingFounder P0 wild-type males were mutagenized with N-ethyl-N-nitrosoureaand subsequently crossed to raise F1 and F2 families for screening asdescribed by Pogoda et al. (2006). An F3 genetic screen was conducted toidentify putative mutants with defects in PNS or CNS myelination bywhole-mount in situ hybridization at 5 dpf using a riboprobe for myelinbasic protein (mbp; Pogoda et al., 2006). st78 mutant embryos werephenotypically sorted fromwild-type siblings by a lack ofmbp expression inthe CNS. Bulk segregant analysis with 480 simple sequence lengthpolymorphisms (SSLPs; Talbot and Schier, 1999) identified markers onLG7 flanking the st78 mutation. High-resolution mapping was conductedusing additional SSLPs and single nucleotide polymorphisms linked to themutation, which were found by sequencing PCR fragments amplified fromgenomic DNA of mutants and wild-type siblings.

Fig. 6. znf16l and notch3mutants lack distinct populations of OPCs. In situ hybridization of znf16lst78, notch3st51 and notch3ZM embryos at 54 hpf for deltaD(A-C), notch3 (D-F), and olig2 (G-I) mRNA expression. (A-F) znf16l homozygous mutants do not exhibit decreased deltaD (A) or notch3 (D) expression seen innotch3 homozygous mutants (asterisks in B,C,E). (G′-I′) znf16l mutants exhibit a delay in lateral migration of OPCs (red brackets in G′) that is not seen in eithernotch3 mutant (H′,I′). (G-I,G″-I″) Lateral view of the hindbrain shows that znf16l mutants are mostly lacking the ventral OPC population (white arrowheads),whereas notch3 mutants are mostly lacking the dorsal OPC population (black arrowheads). Genotypes of all fish analyzed were determined by PCR assay.Scale bars: 100 μm.

4125

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

TALEN-mediated genomic deletionTranscription activator-like effector nucleases (TALEN; Sanjana et al.,2012) were employed to generate the st97 allele. Transcription activator-likeeffector binding sites were identified in the znf16l genomic sequence usingthe TALE-NT 2.0 tool (Cermak et al., 2011; Doyle et al., 2013). pCS2+

vectors expressing TALENs targeting the following sequences were cloned:5′-TCCGAGCTCGAGCCCGAC-3′ and 5′-TCCGTGATCTCGGTCACC-GA-3′. Expression vectors were linearized with SmaI and transcribed invitro using the mMessage mMachine T7 Ultra Kit (Ambion). A mixturecontaining equal amounts of each mRNA (∼400 pg each) was injected intoone-cell stage zebrafish embryos. On the next day, some surviving injectedembryos were lysed and genotyped following the protocols described belowto measure the efficiency of inducing genomic deletions. The remaining fishwere raised to adulthood, and mosaic carriers were identified by assaying fortransmission of genomic deletions to their progeny. Sequencing identifiedst97 as a 4 bp deletion that induces a frameshift in the open reading frame ofznf16l.

GenotypingTo genotype the st78 lesion, PCR was conducted (primers: 5′-CAGAG-TCCATTCCCTTGTCCACAAT-3′ and 5′-CAGTTTGCACCTTGTCTT-CTTG-3′), and the amplification products were cleaved with therestriction enzyme MluCI; the st78 mutation disrupts one of the tworestriction sites in the PCR product, so that the mutant allele producesfragments of 170 and 130 bp, whereas the wild-type allele produces

fragments of 170, 110 and 20 bp. To genotype the st97 allele, PCR wasconducted (primers: 5′-TGGTGCTCTATGGTGTCTGTCT-3′ and 5′-GTG-ATGTTCCTGCCCAGATG-3′), and the amplification products werecleaved with PvuII; the 4 bp deletion in st97 removes the restriction sitein the PCR product, so that the mutant allele results in an uncleaved, full-length PCR product.

Whole-mount RNA in situ hybridizationIn situ hybridization on whole zebrafish embryos and larvae from 24 hpf to9 dpf was performed using standard methods (Thisse and Thisse, 2008).Antisense riboprobes were transcribed from the following cDNAs cloned inthe pCRII-Topo vector: mbp (Pogoda et al., 2006), claudin k (Münzel et al.,2012), nkx2.2a (Barth and Wilson, 1995), olig2 (Park et al., 2002), islet1(Inoue et al., 1994), sox10 (Dutton et al., 2001), deltaD and notch3 (Zauckeret al., 2013). Imaging was performed using a Zeiss Stemi SV 11 Apostereomicroscope using the 1.6× and 10× objectives and images werecaptured using the AxioCam Hrc and AxioVision imaging software. Fordifferential interference contrast microscopy, a Zeiss Axio Imager M2microscope was used using the 20× objective, and images were captured onan AxioCam Mrc with AxioVision imaging software.

Transmission electron microscopyTransmission electron microscopy was performed as described by Lyonset al. (2009). Stained copper grids were imaged on a JEOL JEM-1400transmission electron microscope.

Fig. 7. Severe defects in oligodendrocyte development in mutants lacking function of Znf16l and Notch3. In situ hybridization analyses of oligodendrocytedevelopment and myelination in mutants lacking Notch3, Znf16l, or both. (A,E,I,M,Q) Wild-type oligodendrocyte precursors migrating laterally can be detected at52 hpf (A), and strong expression ofmbp can be detected at 5 dpf (M) continuing to 9 dpf (Q). (B,F,J,N,R) notch3mutants (st51) exhibit reduced lateral migration ofOPCs at 52 hpf and gaps in olig2 expression (red arrows in B,F and bracket in B). The number and distribution of OPCs recovers by 4 dpf (J), although mbpexpression is still reduced strongly in hindbrain of notch3mutants (N). (R) By 9 dpfmbp expression recovers, but remains distinguishably reduced compared withwild type. (C,G,K,O,S) znf16l single mutants (st97) show reduction of OPCs at an early stage (bracket in C,G) and a lack ofmbp at 5 dpf that recovers by 9 dpf asshown before (O and S). (D,H,L,P,T) Mutants lacking both Znf16l and Notch3 have a more severe phenotype than either single mutant. Numbers of OPCsare strongly reduced and no OPCs migrate laterally similar to the znf16lmutant at 52 hpf (red arrows and bracket in D,H). Numbers of OPCs are still reduced by4 dpf (L).mbp expression in double mutants is strongly reduced at 5 dpf (P) and very slight expression is observed by 9 dpf (T) compared with either single mutant(R and S). Genotypes of all fish analyzed were determined by PCR assay. Scale bars: 50 μm.

4126

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

Expression constructsFull-length znf16l coding sequence (XM_694039) of 1587 bp with a Kozaksequence at the 5′ end was amplified from a 3 dpf embryonic cDNA pool(using the primers: 5′-GCCGCCACCATGAGCCGAAAAAGGAA-3′ and5′-TTACCCAGTTTGCACCTTGTCT-3′) and directionally inserted intothe pCR8-GW-TOPO vector (Invitrogen). Tissue-specific expressionplasmids were made by LR recombination between this plasmid, pTol2,p3′E-polyA plasmid and the following tissue-specific promoter-containingp5′E plasmids: huC promoter (neurons; Shiau et al., 2013), mpeg1(macrophages and microglia; Shiau et al., 2013) and claudin k(oligodendrocytes; Münzel et al., 2012). sox10 promoter flanked by Tol2sequences was previously described by Glenn and Talbot (2013). Briefly,znf16lwith adapter sequences were amplified with the following primers: 5′-GTGGCCGCAGAACGAGTGGACCGGCCGCCACCatgagccgaaaaaggaa-3′and 5′-CATGTCTGGATCATCATCGATTttacccagtttgcaccttgtc-3′ (uppercase indicates homology to the sox10 promoter vector and lower caseindicates gene-specific sequence) and cloned into the sox10-promotervector with CloneEZ PCR-Cloning kit (GenScript). mZfp191 and mZfp488were amplified from mouse cDNA (a gift from Natasha O’Brown, KingsleyLab, Stanford) using the following primers: mZfp191: 5′-GTGGCCGCA-GAACGAGTGGACCGGCCGCCACCatgtctgcacagtcagtggaa-3′ and 5′-CA-TGTCTGGATCATCATCGATTtcctaatttcttaaaccttcacaacat-3′; and mZfp488:5′-GTGGCCGCAGAACGAGTGGACCGGCCGCCACCatggctgctggaa-catc-3′ and 5′-CATGTCTGGATCATCATCGATTctctgtagccacctgctaactat-gt-3′. PCR fragments were cloned into sox10-promoter vector describedabovewithCloneEZ PCR-Cloning kit. Another neuronal expression constructwas made by inserting a 3.1 kb fragment containing previously definedregulatory sequences from mnx1/hb9 into the p5′E donor vector (Arkhipovaet al., 2012). Transgeneswere transiently expressed by co-injecting∼12-25 pgof Tol2 constructs and∼50-100 pg of Tol2 transposase mRNA at the one-cellstage. Uninjected embryos were raised in parallel as negative controls.

Time-lapse and fluorescent imagingLive imaging was performed by anesthetizing embryos in 0.016% (w/v)tricaine and mounting them with the dorsal side up in 1.5% low-melting-point agarose. Static fluorescent images were captured using a Zeiss LSM5 Pascal confocal microscope with the Axioplan 2 imaging system underthe Plan-Neofluar 10× objective (numerical aperture 0.30) and the 488 nmlaser line. Time-lapse imaging was performed on a Zeiss Axio Observermicroscope coupled to a PerkinElmer UltraView vox spinning-diskconfocal microscope. Images were captured using a Hamamatsu camera,using the Volocity Acquisition suite set to record one z-stack every 3 min.Image analyses were performed on the Volocity software and tracking wasperformed on ImageJ software using the MtrackJ plugin (Meijering et al.,2012). Tracking was done manually by marking the center of each cell everyfive frames (15 min in real time). As embryos continued to grow througoutthe imaging period, tracked cells drifted over time as embryos grew larger.Drift was measured individually for every embryo, and cells were activelymoving when the displacement was greater than the drift.

Statistical analysesStatistical analyses were performed using the Prism6 software (GraphPad).Student’s t-test was used for all comparisons between mutants andheterozygous siblings. Oligodendrocyte displacement and activityappeared to be normally distributed, and analysis was performed withone-way ANOVA with post hoc comparisons between individual means.

AcknowledgementsWe are grateful to M. Barna and A. Villeneuve for sharing equipment. We thankC. E. Shiau for help with the genetic screen and expert advice; A. M. Meireles,K. Shen and D. Lysko for helpful discussions; T. Reyes and C. Hill for maintaining thefish facility; and T. D. Glenn and J. Perrino for help with electron microscopy.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsH.S. identified and mapped the st78 mutation, generated st97 and conducted allexperiments on znf16l and notch3 mutants. H.S. and W.S.T. analyzed the data andwrote the manuscript.

FundingH.S. was supported by an A*STAR fellowship (Singapore). W.S.T. is a CatherineR. Kennedy and Daniel L. Grossman Fellow in Human Biology. This work wassupported by the National Institutes of Health [R01NS050223]; and the NationalMultiple Sclerosis Society [RG-4756-A-3 to W.S.T]. Deposited in PMC for releaseafter 12 months.

Supplementary informationSupplementary information available online athttp://dev.biologists.org/lookup/suppl/doi:10.1242/dev.128215/-/DC1

ReferencesAraujo, S. J. and Tear, G. (2003). Axon guidance mechanisms and molecules:

lessons from invertebrates. Nat. Rev. Neurosci. 4, 910-922.Arkhipova, V., Wendik, B., Devos, N., Ek, O., Peers, B. and Meyer, D. (2012).

Characterization and regulation of the hb9/mnx1 beta-cell progenitor specificenhancer in zebrafish. Dev. Biol. 365, 290-302.

Barth, K. A. and Wilson, S. W. (1995). Expression of zebrafish nk2.2 is influencedby sonic hedgehog/vertebrate hedgehog-1 and demarcates a zone of neuronaldifferentiation in the embryonic forebrain. Development 121, 1755-1768.

Boyd, A., Zhang, H. and Williams, A. (2013). Insufficient OPC migration intodemyelinated lesions is a cause of poor remyelination in MS and mouse models.Acta Neuropathol. 125, 841-859.

Buffo, A. and Rossi, F. (2013). Origin, lineage and function of cerebellar glia. Prog.Neurobiol. 109, 42-63.

Cai, J., Qi, Y., Hu, X., Tan, M., Liu, Z., Zhang, J., Li, Q., Sander, M. and Qiu, M.(2005). Generation of oligodendrocyte precursor cells from mouse dorsal spinalcord independent of Nkx6 regulation and Shh signaling. Neuron 45, 41-53.

Cermak, T., Doyle, E. L., Christian, M., Wang, L., Zhang, Y., Schmidt, C., Baller,J. A., Somia, N. V., Bogdanove, A. J. and Voytas, D. F. (2011). Efficient designand assembly of custom TALEN and other TAL effector-based constructs for DNAtargeting. Nucleic Acids Res. 39, e82.

de Castro, F. and Bribian, A. (2005). The molecular orchestra of the migration ofoligodendrocyte precursors during development. Brain Res. Brain Res. Rev. 49,227-241.

Doyle, E. L., Stoddard, B. L., Voytas, D. F. and Bogdanove, A. J. (2013). TALeffectors: highly adaptable phytobacterial virulence factors and readilyengineered DNA-targeting proteins. Trends Cell Biol. 23, 390-398.

Dutton, K. A., Pauliny, A., Lopes, S. S., Elworthy, S., Carney, T. J., Rauch, J.,Geisler, R., Haffter, P. and Kelsh, R. N. (2001). Zebrafish colourless encodessox10 and specifies non-ectomesenchymal neural crest fates. Development 128,4113-4125.

Emery, B., Agalliu, D., Cahoy, J. D., Watkins, T. A., Dugas, J. C., Mulinyawe,S. B., Ibrahim, A., Ligon, K. L., Rowitch, D. H. and Barres, B. A. (2009). Myelingene regulatory factor is a critical transcriptional regulator required for CNSmyelination. Cell 138, 172-185.

Genoud, S., Lappe-Siefke, C., Goebbels, S., Radtke, F., Aguet, M., Scherer,S. S., Suter, U., Nave, K.-A. and Mantei, N. (2002). Notch1 control ofoligodendrocyte differentiation in the spinal cord. J. Cell Biol. 158, 709-718.

Glenn, T. D. and Talbot, W. S. (2013). Analysis of Gpr126 function defines distinctmechanisms controlling the initiation and maturation of myelin. Development 140,3167-3175.

Howng, S. Y. B., Avila, R. L., Emery, B., Traka, M., Lin, W., Watkins, T., Cook, S.,Bronson, R., Davisson, M., Barres, B. A. et al. (2010). ZFP191 is required byoligodendrocytes for CNS myelination. Genes Dev. 24, 301-311.

Inoue, A., Takahashi, M., Hatta, K., Hotta, Y. and Okamoto, H. (1994).Developmental regulation of islet-1 mRNA expression during neuronaldifferentiation in embryonic zebrafish. Dev. Dyn. 199, 1-11.

Kessaris, N., Fogarty, M., Iannarelli, P., Grist, M., Wegner, M. and Richardson,W. D. (2006). Competing waves of oligodendrocytes in the forebrain and postnatalelimination of an embryonic lineage. Nat. Neurosci. 9, 173-179.

Kim, H., Shin, J., Kim, S., Poling, J., Park, H.-C. and Appel, B. (2008). Notch-regulated oligodendrocyte specification from radial glia in the spinal cord ofzebrafish embryos. Dev. Dyn. 237, 2081-2089.

Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. and Schilling, T. F.(1995). Stages of embryonic development of the zebrafish. Dev. Dyn. 203,253-310.

Kotter, M. R., Li, W.-W., Zhao, C. and Franklin, R. J. M. (2006). Myelin impairs CNSremyelination by inhibiting oligodendrocyte precursor cell differentiation.J. Neurosci. 26, 328-332.

Lyons, D. A., Naylor, S. G., Scholze, A. and Talbot, W. S. (2009). Kif1b is essentialfor mRNA localization in oligodendrocytes and development of myelinated axons.Nat. Genet. 41, 854-858.

McKenzie, I. A., Ohayon, D., Li, H., Paes de Faria, J., Emery, B., Tohyama, K.and Richardson, W. D. (2014). Motor skill learning requires active centralmyelination. Science 346, 318-322.

Meijering, E., Dzyubachyk, O. and Smal, I. (2012). Methods for cell and particletracking. Methods Enzymol. 504, 183-200.

4127

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT

Miyamoto, Y., Yamauchi, J. and Tanoue, A. (2008). Cdk5 phosphorylation ofWAVE2 regulates oligodendrocyte precursor cell migration through nonreceptortyrosine kinase Fyn. J. Neurosci. 28, 8326-8337.

Mizuguchi, R., Sugimori, M., Takebayashi, H., Kosako, H., Nagao, M., Yoshida,S., Nabeshima, Y.-i., Shimamura, K. and Nakafuku, M. (2001). Combinatorialroles of olig2 and neurogenin2 in the coordinated induction of pan-neuronal andsubtype-specific properties of motoneurons. Neuron 31, 757-771.

Munzel, E. J., Schaefer, K., Obirei, B., Kremmer, E., Burton, E. A., Kuscha, V.,Becker, C. G., Brosamle, C., Williams, A. and Becker, T. (2012). Claudin k isspecifically expressed in cells that formmyelin during development of the nervoussystem and regeneration of the optic nerve in adult zebrafish. Glia 60, 253-270.

Novitch, B. G., Chen, A. I. and Jessell, T. M. (2001). Coordinate regulation of motorneuron subtype identity and pan-neuronal properties by the bHLH repressorOlig2. Neuron 31, 773-789.

Park, H. C. and Appel, B. (2003). Delta-Notch signaling regulates oligodendrocytespecification. Development 130, 3747-3755.

Park, H.-C., Mehta, A., Richardson, J. S. and Appel, B. (2002). olig2 is required forzebrafish primarymotor neuron and oligodendrocyte development.Dev. Biol. 248,356-368.

Pogoda, H.-M., Sternheim, N., Lyons, D. A., Diamond, B., Hawkins, T. A.,Woods, I. G., Bhatt, D. H., Franzini-Armstrong, C., Dominguez, C., Arana, N.et al. (2006). A genetic screen identifies genes essential for development ofmyelinated axons in zebrafish. Dev. Biol. 298, 118-131.

Pozniak, C. D., Langseth, A. J., Dijkgraaf, G. J. P., Choe, Y., Werb, Z. andPleasure, S. J. (2010). Sox10 directs neural stem cells toward theoligodendrocyte lineage by decreasing Suppressor of Fused expression. Proc.Natl. Acad. Sci. USA 107, 21795-21800.

Richardson, W. D., Kessaris, N. and Pringle, N. (2006). Oligodendrocyte wars.Nat. Rev. Neurosci. 7, 11-18.

Ross, S. E., McCord, A. E., Jung, C., Atan, D., Mok, S. I., Hemberg, M., Kim,T.-K., Salogiannis, J., Hu, L., Cohen, S. et al. (2012). Bhlhb5 and Prdm8 form arepressor complex involved in neuronal circuit assembly. Neuron 73, 292-303.

Rowitch, D. H. (2004). Glial specification in the vertebrate neural tube. Nat. Rev.Neurosci. 5, 409-419.

Sanjana, N. E., Cong, L., Zhou, Y., Cunniff, M. M., Feng, G. and Zhang, F. (2012).A transcription activator-like effector toolbox for genome engineering.Nat. Protoc.7, 171-192.

Scholz, J., Klein, M. C., Behrens, T. E. J. and Johansen-Berg, H. (2009). Traininginduces changes in white-matter architecture. Nat. Neurosci. 12, 1370-1371.

Shiau, C. E., Monk, K. R., Joo, W. and Talbot, W. S. (2013). An anti-inflammatoryNOD-like receptor is required for microglia development. Cell Rep. 5, 1342-1352.

Shin, J., Park, H.-C., Topczewska, J. M., Mawdsley, D. J. and Appel, B. (2003).Neural cell fate analysis in zebrafish using olig2 BAC transgenics. Methods CellSci. 25, 7-14.

Stolt, C. C., Rehberg, S., Ader, M., Lommes, P., Riethmacher, D., Schachner,M., Bartsch, U. andWegner, M. (2002). Terminal differentiation of myelin-formingoligodendrocytes depends on the transcription factor Sox10. Genes Dev. 16,165-170.

Talbot, W. S. and Schier, A. F. (1999). Positional cloning of mutated zebrafishgenes. Methods Cell Biol. 60, 259-286.

Thisse, C. and Thisse, B. (2008). High-resolution in situ hybridization to whole-mount zebrafish embryos. Nat. Protoc. 3, 59-69.

Tomassy, G. S. and Fossati, V. (2014). How big is the myelinating orchestra?Cellular diversity within the oligodendrocyte lineage: facts and hypotheses. Front.Cell Neurosci. 8, 201.

Tsai, H.-H., Macklin, W. B. and Miller, R. H. (2006). Netrin-1 is required for thenormal development of spinal cord oligodendrocytes. J. Neurosci. 26, 1913-1922.

Wang, S., Sdrulla, A. D., diSibio, G., Bush, G., Nofziger, D., Hicks, C.,Weinmaster, G. and Barres, B. A. (1998). Notch receptor activation inhibitsoligodendrocyte differentiation. Neuron 21, 63-75.

Wang, S.-Z., Dulin, J., Wu, H., Hurlock, E., Lee, S.-E., Jansson, K. and Lu, Q. R.(2006). An oligodendrocyte-specific zinc-finger transcription regulator cooperateswith Olig2 to promote oligodendrocyte differentiation. Development 133,3389-3398.

Wu, S., Wu, Y. andCapecchi, M. R. (2006). Motoneurons and oligodendrocytes aresequentially generated from neural stem cells but do not appear to share commonlineage-restricted progenitors in vivo. Development 133, 581-590.

Young, K. M., Psachoulia, K., Tripathi, R. B., Dunn, S.-J., Cossell, L., Attwell, D.,Tohyama, K. and Richardson, W. D. (2013). Oligodendrocyte dynamics in thehealthy adult CNS: evidence for myelin remodeling. Neuron 77, 873-885.

Zannino, D. A. and Appel, B. (2009). Olig2+ precursors produce abducens motorneurons and oligodendrocytes in the zebrafish hindbrain. J. Neurosci. 29,2322-2333.

Zaucker, A., Mercurio, S., Sternheim, N., Talbot, W. S. and Marlow, F. L. (2013).notch3 is essential for oligodendrocyte development and vascular integrity inzebrafish. Dis. Model. Mech. 6, 1246-1259.

4128

RESEARCH ARTICLE Development (2015) 142, 4119-4128 doi:10.1242/dev.128215

DEVELO

PM

ENT