PATTERNS & PHENOTYPES

whitesnake/sfpq Is Required for Cell Survivaland Neuronal Development in the ZebrafishLaura Anne Lowery, Jamie Rubin, and Hazel Sive*

Organogenesis involves both the development of specific cell types and their organization into a functionalthree-dimensional structure. We are using the zebrafish to assess the genetic basis for brain organogenesis.We show that the whitesnake mutant corresponds to the sfpq (splicing factor, proline/glutamine rich) gene,encoding the PSF protein (polypyrimidine tract-binding protein-associated splicing factor). In vitro studieshave shown that PSF is important for RNA splicing and transcription and is a candidate brain-specificsplicing factor, however, the in vivo function of this gene is unclear. sfpq is expressed throughoutdevelopment and in the adult zebrafish, with strong expression in the developing brain, particularly inregions enriched for neuronal progenitors. In the whitesnake mutant, a brain phenotype is visible by 28 hrafter fertilization, when it becomes apparent that the midbrain and hindbrain are abnormally shaped.Neural crest, heart, and muscle development or function is also abnormal. sfpq function appears to berequired in two distinct phases during development. First, loss of sfpq gene function leads to increased celldeath throughout the early embryo, suggesting that cell survival requires functional PSF protein. Second,sfpq function is required for differentiation, but not for determination, of specific classes of brain neurons.These data indicate that, in vertebrates, sfpq plays a key role in neuronal development and is essential fornormal brain development. Developmental Dynamics 236:1347–1357, 2007. © 2007 Wiley-Liss, Inc.

Key words: whitesnake; splicing factor proline glutamine rich; PSF; zebrafish; brain development; neuronaldetermination; neuronal differentiation; cell death

Accepted 26 February 2007

INTRODUCTION

The development of organs is complex,involving both generation of appropri-ate cell types and tissues and organi-zation of these in the correct three-dimensional structure. We havebegun to examine brain organogene-sis, using the zebrafish as a model andby studying mutants suggested to bedefective in brain development (Jianget al., 1996; Schier et al., 1996; Am-sterdam et al., 2004).

The whitesnake (wis) mutant wasisolated from a chemical mutagenesis

screen, and shows a distinct brainphenotype (Jiang et al., 1996; Schieret al., 1996). Here, we show that thezebrafish whitesnake (wis) mutantcorresponds to disruption of the sfpqgene. sfpq (splicing factor, proline/glu-tamine rich), which encodes the Poly-pyrimidine tract binding protein asso-ciated Splicing Factor (PSF), isenriched in differentiating neurons inthe mouse brain and has been sug-gested to play a role in neural-specificsplicing (Chanas-Sacre et al., 1999).PSF was first identified as an essen-

tial pre-mRNA splicing factor (Pattonet al., 1993) and has since been shownto exhibit multiple functions in nucleicacid synthesis and processing in vitroand in tissue culture, including tran-scriptional corepression, DNA un-winding, and linking RNA transcriptswith RNA polymerase II (Emili et al.,2002; Shav-Tal and Zipori, 2002). Inaddition, it has been suggested to playa role in tumorigenesis as well as ap-optosis, as sfpq translocation occurs inpapillary renal cell carcinoma (Clarket al., 1997), and nuclear relocaliza-

The Supplementary Material referred to in this article can be found at http://www.interscience.wiley.com/jpages/1058-8388/suppmatWhitehead Institute for Biomedical Research, and Massachusetts Institute of Technology, Cambridge, MassachusettsGrant sponsor: NIH; Grant Number: MH70926; Grant sponsor: NIH-NCRR; Grant number: RR12546; Grant sponsor: NIH NRSA.*Correspondence to: Hazel Sive, Whitehead Institute for Biomedical Research, and Massachusetts Institute of Technology,Nine Cambridge Center, Cambridge, MA 02142. E-mail: sive@wi.mit.edu

DOI 10.1002/dvdy.21132Published online 28 March 2007 in Wiley InterScience (www.interscience.wiley.com).

DEVELOPMENTAL DYNAMICS 236:1347–1357, 2007

© 2007 Wiley-Liss, Inc.

tion and hyperphosphorylation of PSFoccur during apoptosis (Shav-Tal etal., 2001).

In this report, we present the firstwhole animal study of sfpq loss offunction. Our data suggest that sfpqfunction is required for cell survivaland that it is also required for thedifferentiation, but not determination,of specific neuronal classes.

RESULTS AND DISCUSSION

The whitesnake MutationDisrupts Normal Brain andBody Development

The whitesnake (wis) mutant was iso-lated from a chemical mutagenesisscreen, and shows a distinct brainphenotype (Jiang et al., 1996; Schieret al., 1996). Phenotypic abnormali-ties are first visible during mid-somi-togenesis. By 22 hr postfertilization(hpf), there is a subtle change in thecurvature of the entire body, with thetail remaining more curved than wild-type siblings and the hindbrain moreflat (not shown). In addition, wis mu-tants lack myotomal contractions, andthe somites are abnormally organized(not shown). The wis phenotype ismore pronounced by 24 hpf, with ab-sence of eye pigmentation, andslightly reduced brain ventricle width(Fig. 1B,E). By 28 hpf, mutants showan absence or severe reduction in bodypigmentation, flattening of the brain,a curved and thick yolk extension, re-duced heartbeat, and a lack of touchresponse (Fig. 1H, and not shown). In-jecting the brain cavity with a fluores-cent dye (Lowery and Sive, 2005)highlights structural abnormalities bythis stage—all the embryonic brainventricles are reduced, with the mid-brain ventricle most severely affected;however, there is variability in the ex-tent of ventricle reduction (Fig. 1H andSupplementary Figure S1, which can beviewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat).At 2 days postfertilization (dpf), the wisbrain is smaller than in wild-type sib-lings, and the size of all brain ventriclecavities are severely reduced, with themidbrain ventricle almost completelyabsent (Fig. 1J–Q). The tail does be-come straight, although all other 24hours postfertilization (hpf) defects per-sist. wis mutant embryos die by 4 dpf,

when cells appear to become necroticand the embryo disintegrates. The twoalleles analyzed, tr241 and m427, haveindistinguishable phenotypes.

These results show that loss of wisfunction leads to profound morpholog-ical defects affecting several organsystems, with specific defects in thebrain. The brain morphology defectappears relatively late, after the ini-tial shaping of the brain.

The whitesnake MutantCorresponds to the sfpqGene, Encoding the PSFProtein

We observed that the wis phenotypeappears very similar to that of thesfpq hi1779 retroviral insertion mu-tant (compare Fig. 1B and C, E and F,H and I), although only preliminaryanalysis of the sfpq phenotype hasbeen reported (Amsterdam et al.,2004). The sfpq gene (Splicing Factor,Proline/glutamine (Q) rich) encodesthe PSF (Polypyrimidine tract-bind-ing protein-associated Splicing Fac-tor) protein, a 619-amino-acid nuclearfactor. In vitro assays using cell ex-tracts indicate that PSF can partici-pate in a variety of functions, includ-ing RNA splicing and transcriptionalregulation (Patton et al., 1993; Shav-Tal and Zipori, 2002).

Crosses between the wistr241 andsfpq mutants showed that they fail tocomplement, and are, therefore, likelyto be different alleles of the same lo-cus. In a cross of wis and sfpq het-erozygotes, 144 (73%) showed a wild-type phenotype and 54 (27%) amutant phenotype, as expected fornoncomplementing loci (198 embryostotal).

sfpqhi1779 has a 6-kb retroviral in-sertion at base 553 in the coding se-quence, which leads to a truncation ofthe protein at amino acid 197 (Fig.1R). This truncation removes the RNArecognition motifs and the two nuclearlocalization sequences needed for pro-tein function (Dye and Patton, 2001),suggesting that the hi1779 PSF pro-tein is not functional. We askedwhether the wis phenotype is due to amutation in the sfpq gene by compar-ing cDNA sequences from wild-type,wistr241 and wism427 mutants. The

tr241 allele of wis contains a C to Tmutation at position 491 in the sfpqcoding sequence, which results in apremature stop codon at amino acidposition 167 (Fig. 1R) and severe trun-cation of the protein. The wism427 allelehas incorrect splicing of the last exon,which truncates the last 39 amino ac-ids, removing the last nuclear localiza-tion sequence previously shown to benecessary for protein function (Dye andPatton, 2001). There are no obvious mu-tations in the wism427 genomic DNAcoding sequence that could account forthe splicing error, and thus it is likelythat the mutation exists elsewhere, per-haps in intronic sequence. We used re-verse transcriptase-polymerase chainreaction (RT-PCR) to analyze sfpq RNAlevels in wism247 and did not see grossdifferences in levels between mutantand wild-type (data not shown).

We further confirmed gene assign-ment by rescuing the wis phenotypewith sfpq mRNA. Injection of 100 –200 pg of sfpq mRNA partially res-cues the wis brain ventricle, yolk ex-tension, pigment, and movementdefects at 28 hpf (n � 17, 94% res-cue, Fig. 2B,E, and Movie in Supple-mentary Materials). The incompleterescue is likely to be due to degrada-tion of the injected mRNA by laterstages of development. Consistently,we were able to partially phenocopythe wis phenotype by injecting 8 ngof antisense morpholino oligonucleo-tides targeted to the sfpq start site(Fig. 2G–L). These morphant em-bryos showed the wis brain morphol-ogy defect, mild tail curvature, andreduced pigmentation in 100% of in-jected embryos (n � 89; Fig. 2H,K),although myotomal contractions andtouch response were not eliminated,suggesting we were not able to com-pletely deplete sfpq function withthe morpholino.

sfpq Is ExpressedThroughout the Embryo andAdult Zebrafish and IsEnriched in the BrainDuring NeuronalDevelopment

PSF protein is localized to differenti-ating neurons in the embryonic mousebrain (Chanas-Sacre et al., 1999), but

1348 LOWERY ET AL.

the expression pattern of this genehas not been thoroughly examinedduring development. We, therefore,examined the expression patterns andlevel of sfpq RNA during zebrafish de-velopment, using RT-PCR and wholemount in situ hybridization. RT-PCRshows that sfpq is expressed both ma-ternally and zygotically, with maxi-mal expression at 18 hpf and continu-ing through 3 dpf. Expression is alsostrong in the adult brain, but thereare varying levels of weaker expres-sion in the eyes, gut, heart, and bodymuscle (Fig. 3A). In situ hybridizationdemonstrates that sfpq expression isubiquitous until mid-somitogenesis(Fig. 3B–G). By 18 hpf, sfpq expres-sion is slightly stronger in presump-tive rhombomere 4 (Fig. 3H, arrow) as

Fig. 1. Phenotype of whitesnake/sfpq mutants. A–Q: Brain ventricleswere visualized by microinjecting a fluorescent dye, Rhodamine–dext-ran, into the hindbrain ventricle of living anesthetized embryos at 24hours postfertilization (hpf, A–F), 28 hpf (G–I), and 2 dpf (J–Q). A–F: At 24hpf, the wistr241 mutant has reduced brain ventricles and abnormalcurvature of the tail (B,E), as compared with wild-type (A,D), and thesfpqhi1779 mutant phenotype is similar (C,F). G–I: At 28 hpf, both wistr241

and sfpqhi1779 show variable reduction in brain ventricle size and re-duced pigmentation. J–Q: By 48 hpf, the wis brain ventricle reductionbecomes more severe compared with wild-type, especially in the mid-brain. A–C,G–I,N–Q: Dorsal views. D–F,J–M: Side views. F, forebrain; M,midbrain; H, hindbrain. Scale bar � 100 �m. R: sfpq gene/PSF proteinand corresponding mutations. wistr241 has a C to T mutation at position491 of coding sequence, which results in an early stop codon at aminoacid 167. sfpqhi1779 has a 6-kb retroviral insertion (which has a stopcodon early in the insertion sequence). wism427 mRNA has aberrantsplicing resulting in 200 base pairs of intronic sequence inserted beforethe last exon. Both sfpqhi1779 and wistr241 are truncated near the end ofP/Q-rich region, and wism427 lacks the last NLS. P/Q, proline/glutamine-rich region; RRMs, RNA recognition motifs; NLS, nuclear localizationsequence.

Fig. 2. sfpq mRNA partially rescues the whitesnake phenotype andsfpq morpholino phenocopies mutant phenotype at 28 hours postfertil-ization (hpf). A–F: Wild-type (WT) sibling of wistr241 (A,D), wistr241 mutantinjected with �100 pg sfpq mRNA (B,E), and wistr241 mutant (C,F). G–L:WT injected with 8 ng of control morpholino (G,J), WT injected with 8ngsfpq morpholino (H,K), and sfpqhi1779 mutant (I,L). A–C,G–I: Dorsal views.D–F,J–L: Lateral views. Scale bar � 100 �m.

ROLE OF WIS/SFPQ IN BRAIN DEVELOPMENT 1349

well as in the developing telencepha-lon (Fig. 3H, bracket). By 24 hpf, ex-pression appears much stronger in thebrain than in non-neural tissue (Fig.3I), with strongest expression in thetelencephalon, midbrain, hindbrain,and retina (Fig. 3J,K). At 2–3 dpf, sfpqexpression is apparent only in distinctregions of the brain (Fig. 3L–Q). At 2dpf, expression is strong in the ventralhindbrain and in dorsoventral stripesthroughout the hindbrain (Fig. 3M,N).At 3 dpf, expression is strongly de-tected in a region ventral to the brain,but also faintly along many axontracts (Fig. 3O–Q). By the more sensi-tive RT-PCR assay, sfpq is also ex-pressed in the tail at 3 dpf (Fig. 3A,lane 9).

The strong expression of sfpq in dis-tinct areas of the brain prompted us toask whether this gene is expressed inregions of active neurogenesis. We ad-dressed this question by double-label-ing 24 hpf embryos for sfpq RNA by insitu hybridization followed by anti-body labeling for HuC/D, an earlymarker for postmitotic neurons (Ma-rusich et al., 1994). This confirmedthat multiple regions in the brain withthe strongest sfpq expression overlapwith regions labeled with HuC/D (Fig.4). This overlap is particularly appar-ent in the telencephalon (Fig. 4A,Cbracket) and in the hindbrain (Fig.4B,D bracket). A higher magnificationview of the hindbrain shows twodarker stripes along its anteroposte-rior extent, with less expression medi-ally (Fig. 4B, dark stripe in brackets).HuC/D labeling in the hindbrain oc-curs only within the region of strongersfpq expression (Fig. 4D).

These data show that differentiat-ing neurons express high levels of sfpqin the zebrafish, and together with areport of high PSF levels in differen-tiating mouse brain neurons (Chanas-Sacre et al., 1999), suggest that PSFmay have brain-specific activity, func-tioning during neuronal differentia-tion. However, sfpq is also expressedbroadly during zebrafish embryonicdevelopment and in the adult, consis-tent with PSF expression in non-neu-ral human and rat tissues (Shav-Talet al., 2000; Dong et al., 2005). Remi-niscent of the case for sfpq, the RNAbinding proteins, Nova1 and Nova2,have been proposed to be brain-spe-cific splicing factors in mammals, and

aberrant splicing of neural genes isseen in Nova knockout mutants (Yanget al., 1998; Jensen et al., 2000; Ule etal., 2005). However, both Nova1 andNova2 are also expressed in one ormore non-neural tissues, includingliver and lung (Buckanovich et al.,1993; Yang et al., 1998). Thus, splic-ing factors with tissue-specific activitymay also have more general activity.This is the case for the essential splic-ing factor ASF/SF2, which has beenimplicated in cardiac-specific splicing(Xu et al., 2005b), while the ubiqui-tously expressed PTB, the polypyrimi-dine-tract–binding protein, specificallyrepresses neuron-specific splicing of the�-aminobutyric acid (A) receptor in non-neuronal cells (Ashiya and Grabowski,1997).

Loss of sfpq Function DoesNot Affect Cell Proliferationbut Leads to WidespreadCell Death

While extensive in vitro analysis hasshown that PSF protein can partici-pate in a variety of functions (Shav-Tal and Zipori, 2002), loss of PSF func-tion has not been examined in anywhole animal. To characterize thePSF loss of function phenotype, we an-alyzed levels of cell proliferation inwis embryos at 24 hpf, by labeling mi-totic cells with an antibody to phos-

phorylated histone H3 (PH3; Hendzelet al., 1997; Saka and Smith, 2001).There are no obvious differences incell proliferation levels between wismutants and their wild-type siblings(Fig. 5A,B), and quantitation of PH3-positive cells in the hindbrain andtrunk region show no statistical differ-ence between mutant and wild-typeembryos (n � 8; P � 0.8046; Fig. 5C,and data not shown).

In contrast, analysis of cell death at24 hpf by terminal deoxynucleotidyltransferase–mediated deoxyuridinetri-phosphate nick end-labeling (TUNEL)labeling shows that wis mutants dis-play twice the normal amount of celldeath throughout the embryo (Fig.5D–F; n � 14; P � 0.0001, and datanot shown). Increased cell death con-tinues through 2 dpf (data not shown),occurring in many tissues, even thosethat express sfpq at low levels. Thisfinding could be due to loss of low levelsfpq expression, or it could be a resultof loss of sfpq at earlier time points.Regardless, these results suggest thatsfpq is not required for regulation ofcell proliferation, but is involved inregulating cell death. Several lines ofevidence suggest that PSF responds toand/or can modulate apoptosis. In ap-optotic cells, PSF dissociates from apartner, PTB, and binds new part-ners, including the splicing factorsU1-70K and SR proteins (Shav-Tal et

Fig. 3. sfpq expression patterns. A: Reverse transcriptase-polymerase chain reaction (RT-PCR) forsfpq in embryonic tissue shows that sfpq is expressed beginning at 1 hpf (lane 1) and peaks at 18hours postfertilization (hpf, lane 5). Adult tissue also has sfpq expression, with the brain showingthe highest level (lane 11) and heart and body muscle showing lower levels (lanes 14, 15). ActinRT-PCR was used as control for RNA levels. B–Q: sfpq in situ hybridization time course showsthat sfpq is expressed throughout embryogenesis. B: At 1 hpf (four-cell stage), side view. C: At 4hpf (blastula stage), side view. D: At 8 hpf (75% epiboly stage), side view, dorsal right. E: At 12 hpf(six-somite stage), side view, dorsal right, anterior top. F–H: At 18 hpf, strongest expression in theforebrain (H, bracket) and in presumptive rhombomere 4 (H, arrow). I–K: At 24 hpf. L–N: At 2 dpf,expression appears restricted to strong longitudinal strips in the ventral brain (M, arrow) and inweaker transverse stripes in the hindbrain. O–Q: At 3 dpf, expression ventral to brain (P, arrow)and along axon tracts. F–G,I–J,L–M,O–P: Side views, anterior left. H,K,N,Q: Dorsal views, anteriorleft. Note: in I–Q, regions without staining may not be in focus, as to allow high magnificationimaging of the stained regions. In areas of staining, the fuzziness sometimes observed is due tolow levels of diffuse staining, not poor imaging. Scale bar � 100 �m.

Fig. 4. sfpq is strongly expressed in regions of neurogenic activity. A–D: At 24 hours postfertil-ization (hpf) wild-type embryos labeled for sfpq expression by in situ hybridization only (A,B) andwild-type sibling embryos double labeled for sfpq mRNA expression and HuC protein by immu-nohistochemistry (C,D) show that regions of strongest sfpq expression in the brain (brackets)overlap with HuC, a marker for postmitotic neuronal precursors. B,D: Higher magnification ofhindbrain; brackets mark strong sfpq expression overlapping with HuC labeling. Midline stainingis an artifact of the staining process, because it is not observed in embryos cut open beforestaining. A–D: Dorsal views. Anterior left. Scale bar � 100 �m.

1350 LOWERY ET AL.

al., 2001). Additionally, during apo-ptosis, PSF relocalizes into globularnuclear structures, rather than in thenuclear “speckles” (interchromatingranules) normally observed (Shav-Tal et al., 2001). Of interest, PSF in-duces apoptosis when overexpressedin mammalian cell culture (Xu et al.,2005a), and it is not clear how thisobservation relates to the increase inapoptosis we observe in wis/sfpq mu-tants.

Fig. 3.

Fig. 4.

ROLE OF WIS/SFPQ IN BRAIN DEVELOPMENT 1351

sfpq Is Required for theDifferentiation, but NotDetermination, of CertainNeuronal Cell Types

Because sfpq is strongly expressed inthe brain, and particularly in deter-mined, postmitotic neuronal precur-sors, we asked whether and when sfpqalters neuronal development. We firstanalyzed neural patterning in wis mu-tants at 24 hpf. Expression of the an-teroposterior neural markers krox20,pax2.1, and the dorsoventral markersopl, and shh (Krauss et al., 1991,1993; Oxtoby and Jowett, 1993; Grin-blat et al., 1998) is identical in wismutants and wild-type siblings (Fig.6A,B, and data not shown), suggestingthat sfpq is not required for early neu-ral patterning. Second, we analyzedearly neurogenesis in wis mutants.Expression of zash1B, a proneuralgene expressed exclusively in cyclingneural progenitors (Allende and Wein-berg, 1994; Mueller and Wullimann,2003), is similar between wis mutantsand wild-type siblings (data notshown). Similarly, expression of

Fig. 5.

Fig. 6.

Fig. 5. Cell proliferation and cell death analysisin whitesnake mutants. A–C: Cell proliferationanalysis, using PH3 antibody labeling. A,B:Fixed and labeled wild-type and wis brain at 24hours postfertilization (hpf). C: Quantificationcomparing labeling in hindbrain shows no dif-ference between wild-type and mutant, n � 8;P � 0.8046. D–F: Cell death analysis, usingterminal deoxynucleotidyl transferase–medi-ated deoxyuridinetriphosphate nick end-label-ing (TUNEL) staining. D,E: Fixed and labeledwild-type and wis brain at 24 hpf. F: Quantifi-cation comparing labeling in hindbrain showsapproximately twice the amount of cell death inthe mutant than in wild-type, n � 14; P �0.0001. Error bars denote standard error.A–B,D–E: Dorsal views. Boxes mark regionsused for quantitation. F, forebrain; M, midbrain;H, hindbrain. Scale bar � 100 �m.

Fig. 6. Neuronal determination is normal inwhitesnake mutants. A,B: In situ hybridizationfor pax2a (labeling nasal placodes, midbrain–hindbrain boundary, and otic vesicles, arrow-heads) and krox20 (labeling rhombomeres 3and 5, arrows) show similar staining patterns inwild-type and mutant. C,D: Immunohistochem-istry for HuC, a marker for postmitotic neurons,shows identical patterns between wild-type andmutant. E–G: This finding has been quantified ingraph E, which depicts the number of HuC-labeled cells per unit area in the spinal cords ofwild-type and wis (F,G). F, forebrain; M, mid-brain; H, hindbrain. Scale bar � 100 �m.

1352 LOWERY ET AL.

HuC/D, one of the earliest markers forpostmitotic neuronal precursors (Ma-rusich et al., 1994), shows no differ-ence between wis and wild-type sib-lings (Fig. 6C–G). Quantitation ofHuC/D cell labeling in the spinal cordconfirms that the number of Hu-posi-tive cells is the same between wis andwild-type embryos (Fig. 6E–G; n � 5;P � 0.2672). These results indicatethat sfpq is not required for early neu-ral patterning and neuronal determi-nation and show that the number ofpostmitotic neuronal precursors isnormal in the wis mutant.

In contrast, markers for later neu-ronal differentiation show obvious ex-pression defects in wis mutants. La-beling the 24 hpf embryonic axonscaffolds with anti-acetylated tubulinor anti–HNK-1 (Metcalfe et al., 1990)demonstrates a large reduction in thenumber of axon tracts throughout thebrain (Fig. 7A–J). Immunostainingwith a monoclonal antibody directedagainst Neurofilament M, which la-bels both cell bodies and axons ofhindbrain reticulospinal neuronal(Pleasure et al., 1989), shows a com-plete absence of all reticulospinal neu-rons in the hindbrain except theMauthner neurons (Fig. 7K,L). By 2dpf, Mauthner neurons appear some-what abnormal, with a rounded cellbody and occasional axonal pathfind-ing abnormalities (Fig. 7M,N, and notshown). No other hindbrain reticu-lospinal neurons are present. Impor-tantly, not all neuronal cell types areaffected in wis, as labeling with ananti-islet antibody, which marks dif-ferentiated motoneuron cell bodies(Ericson et al., 1992) shows no differ-ences between wis and wild-type sib-lings in the brain (Fig. 7O,P) or in thespinal cord (Fig. 7Q,R). Quantitationin the brain demonstrates identicalnumbers of labeled cells are present inthis region (n � 5; P � 0.7517; Sup-plementary Figure S2). One caveat tothese conclusions is that we have an-alyzed survival of motoneuron cellbodies but not axons, raising the in-teresting possibility that cell bodiessurvive while axons degenerate. How-ever, this possibility is clearly not gen-erally true, as both cell bodies andaxons were absent from wis reticu-lospinal neurons.

These data suggest that a secondphase of sfpq function is neural-spe-

cific. In particular, only specificclasses of differentiated neurons areabsent in the sfpq/wis mutants, whileother classes are present in normalnumbers. Furthermore, the number ofpostmitotic neuronal precursors iden-tified by HuC/D labeling is indistin-guishable between wild-type and mu-tant at 24 hpf. Thus, even withincreased cell death throughout theembryo, normal numbers of deter-mined and postmitotic neurons are re-cruited. Together, the data indicatethat neuronal differentiation, but notneuronal determination, requires sfpqfunction.

CONCLUSION: SEPARABLEFUNCTIONS FOR SFPQ?

Is the increase in cell death we ob-serve in sfpq/wis mutants simply theresult of removing a housekeepingfunction from the embryo? The notionof housekeeping functions has becomecomplex, where some genes have bothubiquitously required functions andcell type-specific functions, such as theessential splicing factor ASF/SF2, im-plicated in cardiac-specific splicing(Xu et al., 2005b), and the ubiqui-tously expressed PTB, the polypyrimi-dine-tract–binding protein, which re-presses neuron-specific splicing innon-neuronal cells (Ashiya andGrabowski, 1997). PSF may also haveboth general and tissue-specific func-tions.

Clearly, there is an embryo-wide re-quirement for PSF protein. Our datasuggest that PSF normally suppressesapoptosis, although the mechanismsby which it interfaces with the apopto-tic machinery are not clear. A reportthat overexpressed PSF promotes ap-optosis in mammalian tissue culturecontradicts our finding in whole ani-mals (Xu et al., 2005a) and suggestseither that cells respond to either toomuch or too little PSF, or indicates aspecies-specific difference in PSFfunction.

Later in development, a second ef-fect of loss of PSF function is appar-ent, as specific neuronal classes fail todifferentiate. This finding is not a gen-eral phenotype, as formation of someneuronal classes is unaffected, whilenumbers of neurons in other classesare profoundly decreased or are ab-sent. Consistent with two phases of

sfpq function is the replacement of lowlevel, ubiquitous early expression ofsfpq with later and very strong ex-pression that is enriched in develop-ing brain relative to the rest of theembryo. This finding suggests separa-ble early and late functions for sfpq.Our data are consistent with other re-ports that demonstrate PSF has nu-merous cellular functions (Shav-Taland Zipori, 2002).

Is specific neuronal loss linked tothe effects of PSF on apoptosis? Wecannot yet distinguish whether partic-ular neuronal classes are absent be-cause they do not survive or becausethey cannot differentiate. This ques-tion will be best answered by definingwhether neuronal progenitors die inspfq/wis mutants. Definition of PSFtarget genes will also help answer thisquestion. An alternate explanation isthat another function of PSF is neces-sary for differentiation of specific neu-ronal classes. In particular, PSF couldplay a role in brain-specific splicing ortranscriptional regulation. white-snake/sfpq mutants will be indispens-able tools to investigate whether PSFfunction depends on its splicing activ-ity or on another function of this pro-tein, and to characterize potential tar-gets of PSF.

EXPERIMENTALPROCEDURES

Fish Lines and Maintenance

Danio rerio were raised and bred ac-cording to standard methods (Wester-field, 1995). Embryos were kept at28.5°C and staged according to Kimmelet al. (1995). Times of development areexpressed as hours postfertilization(hpf) or days postfertilization (dpf).

Lines used were AB, Tubingen LongFin, sfpqhi1779 (Amsterdam et al., 2004),wistr241 (Jiang et al., 1996), and wism427

(Schier et al., 1996). For PCR genotyp-ing, tails or embryos were digested withproteinase K (1 mg/ml) in lysis buffer(10 mmol/L Tris pH 8, 1 mmol/L eth-ylenediaminetetraacetic acid, 0.3%Tween-20, 0.3% NP40). Because sf-pqhi1779 has a retroviral insertion inthe sfpq gene, mutant individualscould be identified by PCR. Primersused are as follows: 1779c, 5�-cagca-gactcccaccgtcg-3�; MSL4, 5�-gctagctt-gccaaacctacaggt-3� (MWG Biotech).

ROLE OF WIS/SFPQ IN BRAIN DEVELOPMENT 1353

MSL4 detects the retroviral insertion,which is located 553 base pairs down-stream from the sfpq start site in thehi1779 mutant. The C to T mutationpresent in wistr241 introduces an AccIrestriction site and, thus, can be de-tected by PCR followed by digestion.All procedures on live animals and

embryos were approved by the Massa-chusetts Institute of Technology Com-mittee on Animal Care.

Brain Ventricle Imaging

Methods for brain ventricle imaginghave been described previously (Low-

ery and Sive, 2005). Briefly, embryoswere anesthetized in 0.1 mg/ml Tri-caine (Sigma) dissolved in embryo me-dium before hindbrain ventricle mi-croinjection with 2–10 nl of dextranconjugated to Rhodamine (5% in 0.2mol/L KCl, Sigma), and then embryoswere imaged by light and fluorescence

Fig. 7. Specific neurons are absent in whitesnake mutants. A–J: Acetylated tubulin antibody labeling (A–F) and HNK-1 antibody labeling (G–J), both of whichmark most differentiated neurons and their axons, show reduced number of axons in wis mutant in the hindbrain (B,D,H), midbrain (F), and eyes/forebrain(J) compared with wild-type (A,C,E,G,I). K–N: The RM044 Ab, which labels reticulospinal neurons, shows an absence of all reticulospinal neurons in the wishindbrain except Mauthner neurons (labeled with M) at both 24 hours postfertilization (hpf, L) and 2 days postfertilization (dpf, N), compared with wild-type(K,M). O–R: The 4D5 Islet antibody, labeling motoneurons, shows no loss of motoneurons in whitesnake in the dorsal diencephalon (P) or spinal cord (R)compared with wild-type (O,Q). A, anterior; P, posterior. A–D,G–P: Dorsal views. E–F,Q–R: Side views. Scale bar � 50 �m.

1354 LOWERY ET AL.

microscopy with a Leica dissecting mi-croscope, using a KT Spot DigitalCamera (RT KE Diagnostic Instru-ments). Images were superimposed inPhotoshop 6 (Adobe).

Detection of whitesnakeMutation

Total RNA was extracted from mutantembryos and wild-type siblings usingTrizol reagent (Invitrogen), followed bychloroform extraction and isopropanolprecipitation. cDNA synthesis was per-formed with Super Script II ReverseTranscriptase (Invitrogen) and randomhexamers. PCR was then performed us-ing five sets of primers, which amplifythe coding region of sfpq. Primers usedare as follows: sfpq1F, 5�-tgagggtgctcc-tctctttg -3�; sfpq1R, 5�- gagagcgttgccttc-aattc-3�; sfpq2F, 5�-tccaccgaagatccagtc-tc-3�; sfpq2R, 5�-ggcagctggcttagaagaaa-3�; sfpq3F, 5�-gaggttgcgaaagcagagtt-3�;sfpq3R, 5�-tcatctcctcctctctcctacg-3�;sfpq4F, 5�-aggcagcaagtggagaaaaa-3�;sfpq4R, 5�-gccacaaatgggatgagttt-3�;sfpq5F, 5�-gcaaacgcgaggaatcttac-3�;sfpq5R, 5�-tttttgggagaaccaactgc-3�. RT-PCR products were used for sequencinganalysis, performed by NorthwoodsDNA, Inc. (Solway MN). Sequencingdata was analyzed using the BLASTprogram (http://www.ncbi.nlm.nih.gov/BLAST/), and the cDNA sequence ofsfpq was obtained from the GenBankdatabase (NM_213278).

RNA Injections

Full-length sfpq cDNAs were gener-ated by RT-PCR with primers, in-cluding ClaI/Xba1 sites, ClaI-sfpq1F5�-ccat/cgattgagggtgctcctctctttg-3�;XbaI-sfpq5R 5�-gct/ctagatttttgg-gagaaccaactgc-3�. The PCR frag-ments were subcloned into pCS2�,and the pCS2� constructs were lin-earized by NotI. Capped mRNA wastranscribed in vitro using the SP6mMessage mMachine kit (Ambion).Embryos were injected at the one-cell stage with 100 –300 pg of mRNA.The embryos phenotypically rescuedby mRNA injection were identifiedas mutants by genotyping.

Morpholino Oligo Injections

A morpholino antisense oligonucleo-tide (MO) targeted to the translation

start site of zebrafish sfpq (5�-ccatgc-caccgcgcatccccattcc-3�) was injectedinto one- to two-cell embryos. The fi-nal amounts used were 8 ng of sfpq orcontrol MO (provided by Gene Tools,Inc.).

RT-PCR Time Course

Total RNA was extracted from stagedwild-type embryos or adult dissectedtissue using Trizol reagent (Invitro-gen), followed by chloroform extrac-tion and isopropanol precipitation.cDNA synthesis was performed withSuper Script II Reverse Transcriptase(Invitrogen) and random hexamers.PCR was then performed using sfpq4Fand sfpq4R (listed above), and actinF5�-tatccacgagaccaccttcaactcc-3�, act-inR 5�-ctgcttgctgatccacatctgctgg-3�.

In Situ Hybridization

RNA probes containing digoxigenin(DIG) -11-UTP were synthesized fromlinearized plasmid DNA for sfpq,pax2.1 (Krauss et al., 1991), krox20(Oxtoby and Jowett, 1993), opl (Grin-blat et al., 1998), shh (Krauss et al.,1993), and zash1B (Allende and Wein-berg, 1994) as described (Harland,1991). Standard methods for hybrid-ization and for single color labelingwere used as described elsewhere(Sagerstrom et al., 1996). After stain-ing, embryos were fixed in 4% parafor-maldehyde overnight at 4°C, washedin phosphate buffered saline (PBS),dehydrated in methanol, and thencleared in a 3:1 benzyl benzoate/ben-zyl alcohol (BB/BA) solution beforemounting and imaging with a Nikoncompound microscope.

Immunohistochemistry

Whole-mount immunostaining wascarried out using rabbit anti-phospho-histone H3 (Upstate Biotechnology,1:800), mouse anti-acetylated tubulin(Sigma, 1:1,000), mouse anti-neuro-filament RM044 (Zymed #13-0500,1:50), mouse HNK-1/zn-12 (ZebrafishInternational Resource Center, 1:500),mouse 39.4D5 anti-islet (Developmen-tal Studies Hybridoma Bank, 1:100),mouse anti-HuC (Molecular Probes,1:500), and mouse MF20 IgG2b (Devel-opmental Studies Hybridoma Bank,1:10). Goat anti-rabbit IgG Alexa

Fluor 488 (Molecular Probes, 1:500),goat anti-mouse Alexa Fluor 488 (Mo-lecular Probes, 1:500), goat anti-mouse IgG horseradish peroxidase(HRP; Sigma, 1:500), goat anti-rabbitIgG HRP (Sigma, 1:500), and goat an-ti-mouse IgG2b Alexa Fluor 568 (Mo-lecular Probes, 1:500) were used assecondary antibodies.

For labeling with anti-phosphohis-tone H3, anti-acetylated tubulin,HNK-1 antibody, 39.4D5 antibody,anti-HuC, and MF20 antibody, decho-rionated embryos were fixed in 4%paraformaldehyde for 2 hr at roomtemperature, then rinsed in PBS. Foranti-phosphohistone H3 and anti-HuC, endogenous peroxidase activitywas quenched in 10% hydrogen perox-ide in PBT for 1 hr. For all antibodies,blocking was done for 4 hr at roomtemperature in 0.5% Triton X, 4% nor-mal goat serum, in phosphate buffer.Detection of HRP was performed withthe diaminobenzidine substrate per-oxidase kit (Vector Laboratories). Flu-orescent secondary antibodies were vi-sualized by confocal microscopy(LSM510). Brains were flat-mountedin glycerol and imaged. Images arecomposites of several scans.

For labeling with RM044 antibody,dechorionated embryos were fixed in2% tricloroacetic acid 3 hr at roomtemperature, washed in PBS, blockedin 10% normal goat serum in PBT for3 hr before incubation in antibody.The brains were partially dissectedand mounted for visualization by con-focal microscopy. To block pigmenta-tion, embryos were incubated in 0.2mM 1-phenyl-2-thiorurea in embryomedium beginning at 22 hpf.

Cell Death Labeling

DNA fragmentation during apoptosiswas detected by the TUNEL method,using ‘ApopTag’ kit (Chemicon). Em-bryos were fixed in 4% paraformalde-hyde in PBS for 2 hr, then rinsed inPBS and dechorionated. Embryoswere dehydrated to 100% ethanol,stored at �20°C overnight, then rehy-drated in PBS. Embryos were furtherpermeabilized by incubation in pro-teinase K (5 �g/ml) for 5 min, thenrinsed in PBS. TdT labeling was fol-lowed per manufacturer’s instruc-tions. Anti–DIG-AP (Gibco, 1:100) wasused to detect the DIG-labeled ends.

ROLE OF WIS/SFPQ IN BRAIN DEVELOPMENT 1355

Brains were flat-mounted in glyceroland imaged.

Statistical Analysis

To quantify amount of cell prolifera-tion, cell death, or antibody staining,labeled cells in a defined area of thebrain and/or tail were counted andthen compared statistically using anunpaired t-test (InStat v.3, GraphPadsoftware).

ACKNOWLEDGMENTSWe thank members of the Sive Lab forhelpful comments and Olivier Paugoisfor fish husbandry. Many thanks tothe Nusslein-Volhard lab for provid-ing us with the wis tr241 mutant,the Zebrafish International ResourceCenter for the wis m427 mutant andzn-12 antibody, Adam Amsterdamand Nancy Hopkins for sfpq hi1779mutant. The 39.4D5 antibody devel-oped by Thomas Jessell, and theMF20 antibody developed by DonaldFischman, were obtained from the De-velopmental Studies Hybridoma Bankdeveloped under the auspices of theNICHD and maintained by the Uni-versity of Iowa, Department of Biolog-ical Sciences (Iowa City, IA). Thiswork was conducted using the W.M.Keck Foundation Biological ImagingFacility at the Whitehead Institute.The Zebrafish International ResourceCenter is supported by the NIH-NCRR. H.S. is supported by NIH.L.A.L. is supported by NIH NRSA pre-doctoral fellowship and an AbrahamJ. Siegel Fellowship at the WhiteheadInstitute.

REFERENCES

Allende ML, Weinberg ES. 1994. The ex-pression pattern of two zebrafishachaete-scute homolog (ash) genes is al-tered in the embryonic brain of the cy-clops mutant. Dev Biol 166:509–530.

Amsterdam A, Nissen RM, Sun Z, SwindellEC, Farrington S, Hopkins N. 2004.Identification of 315 genes essential forearly zebrafish development. Proc NatlAcad Sci U S A 101:12792–12797.

Ashiya M, Grabowski PJ. 1997. A neuron-specific splicing switch mediated by anarray of pre-mRNA repressor sites: evi-dence of a regulatory role for the poly-pyrimidine tract binding protein and abrain-specific PTB counterpart. RNA 3:996–1015.

Buckanovich RJ, Posner JB, Darnell RB.1993. Nova, the paraneoplastic Ri anti-gen, is homologous to an RNA-bindingprotein and is specifically expressed inthe developing motor system. Neuron 11:657–672.

Chanas-Sacre G, Mazy-Servais C, WattiezR, Pirard S, Rogister B, Patton JG, Bela-chew S, Malgrange B, Moonen G, Le-prince P. 1999. Identification of PSF, thepolypyrimidine tract-binding protein-as-sociated splicing factor, as a develop-mentally regulated neuronal protein.J Neurosci Res 57:62–73.

Clark J, Lu YJ, Sidhar SK, Parker C, GillS, Smedley D, Hamoudi R, Linehan WM,Shipley J, Cooper CS. 1997. Fusion ofsplicing factor genes PSF and NonO(p54nrb) to the TFE3 gene in papillaryrenal cell carcinoma. Oncogene 15:2233–2239.

Dong X, Shylnova O, Challis JR, Lye SJ.2005. Identification and characterizationof the protein-associated splicing factoras a negative co-regulator of the proges-terone receptor. J Biol Chem 280:13329–13340.

Dye BT, Patton JG. 2001. An RNA recog-nition motif (RRM) is required for thelocalization of PTB-associated splicingfactor (PSF) to subnuclear speckles. ExpCell Res 263:131–144.

Emili A, Shales M, McCracken S, Xie W,Tucker PW, Kobayashi R, Blencowe BJ,Ingles CJ. 2002. Splicing and transcrip-tion-associated proteins PSF andp54nrb/nonO bind to the RNA polymer-ase II CTD. RNA 8:1102–1111.

Ericson J, Thor S, Edlund T, Jessell TM,Yamada T. 1992. Early stages of motorneuron differentiation revealed by ex-pression of homeobox gene Islet-1. Sci-ence 256:1555–1560.

Grinblat Y, Gamse J, Patel M, Sive H.1998. Determination of the zebrafishforebrain: induction and patterning. De-velopment 125:4403–4416.

Harland RM. 1991. In situ hybridization:an improved whole-mount method forXenopus embryos. Methods Cell Biol 36:685–695.

Hendzel MJ, Wei Y, Mancini MA, Van Hoo-ser A, Ranalli T, Brinkley BR, Bazett-Jones DP, Allis CD. 1997. Mitosis-spe-cific phosphorylation of histone H3initiates primarily within pericentro-meric heterochromatin during G2 andspreads in an ordered fashion coincidentwith mitotic chromosome condensation.Chromosoma 106:348–360.

Jensen KB, Dredge BK, Stefani G, ZhongR, Buckanovich RJ, Okano HJ, Yang YY,Darnell RB. 2000. Nova-1 regulates neu-ron-specific alternative splicing and isessential for neuronal viability. Neuron25:359–371.

Jiang YJ, Brand M, Heisenberg CP,Beuchle D, Furutani-Seiki M, Kelsh RN,Warga RM, Granato M, Haffter P, Ham-merschmidt M, Kane DA, Mullins MC,Odenthal J, van Eeden FJ, Nusslein-Vol-hard C. 1996. Mutations affecting neuro-genesis and brain morphology in the ze-

brafish, Danio rerio. Development 123:205–216.

Kimmel CB, Ballard WW, Kimmel SR,Ullmann B, Schilling TF. 1995. Stages ofembryonic development of the zebrafish.Dev Dyn 203:253–310.

Krauss S, Johansen T, Korzh V, Fjose A.1991. Expression of the zebrafish pairedbox gene pax[zf-b] during early neuro-genesis. Development 113:1193–1206.

Krauss S, Concordet JP, Ingham PW.1993. A functionally conserved homologof the Drosophila segment polarity genehh is expressed in tissues with polarizingactivity in zebrafish embryos. Cell 75:1431–1444.

Lowery LA, Sive H. 2005. Initial forma-tion of zebrafish brain ventricles occursindependently of circulation and re-quires the nagie oko and snakehead/atp1a1a.1 gene products. Development132:2057–2067.

Marusich MF, Furneaux HM, Henion PD,Weston JA. 1994. Hu neuronal proteinsare expressed in proliferating neuro-genic cells. J Neurobiol 25:143–155.

Metcalfe WK, Myers PZ, Trevarrow B,Bass MB, Kimmel CB. 1990. Primaryneurons that express the L2/HNK-1 car-bohydrate during early development inthe zebrafish. Development 110:491–504.

Mueller T, Wullimann MF. 2003. Anatomyof neurogenesis in the early zebrafishbrain. Brain Res Dev Brain Res 140:137–155.

Oxtoby E, Jowett T. 1993. Cloning of thezebrafish krox-20 gene (krx-20) and itsexpression during hindbrain develop-ment. Nucleic Acids Res 21:1087–1095.

Patton JG, Porro EB, Galceran J, TempstP, Nadal-Ginard B. 1993. Cloning andcharacterization of PSF, a novel pre-mRNA splicing factor. Genes Dev 7:393–406.

Pleasure SJ, Selzer ME, Lee VM. 1989.Lamprey neurofilaments combine in onesubunit the features of each mammalianNF triplet protein but are highly phos-phorylated only in large axons. J Neuro-sci 9:698–709.

Sagerstrom CG, Grinbalt Y, Sive H. 1996.Anteroposterior patterning in the ze-brafish, Danio rerio: an explant assayreveals inductive and suppressive cellinteractions. Development 122:1873–1883.

Saka Y, Smith JC. 2001. Spatial and tem-poral patterns of cell division duringearly Xenopus embryogenesis. Dev Biol229:307–318.

Schier AF, Neuhauss SC, Harvey M, Mal-icki J, Solnica-Krezel L, Stainier DY,Zwartkruis F, Abdelilah S, Stemple DL,Rangini Z, Yang H, Driever W. 1996.Mutations affecting the development ofthe embryonic zebrafish brain. Develop-ment 123:165–178.

Shav-Tal Y, Zipori D. 2002. PSF andp54(nrb)/NonO—multi-functional nu-clear proteins. FEBS Lett 531:109–114.

Shav-Tal Y, Lee B, Bar-Haim S, Vandeker-ckhove J, Zipori D. 2000. Enhanced pro-teolysis of pre-mRNA splicing factors in

1356 LOWERY ET AL.

myeloid cells. Exp Hematol 28:1029–1038.

Shav-Tal Y, Cohen M, Lapter S, Dye B,Patton JG, Vandekerckhove J, Zipori D.2001. Nuclear relocalization of the pre-mRNA splicing factor PSF during apo-ptosis involves hyperphosphorylation,masking of antigenic epitopes, andchanges in protein interactions. Mol BiolCell 12:2328–2340.

Ule J, Ule A, Spencer J, Williams A, Hu JS,Cline M, Wang H, Clark T, Fraser C,Ruggiu M, Zeeberg BR, Kane D, Wein-stein JN, Blume J, Darnell RB. 2005.

Nova regulates brain-specific splicing toshape the synapse. Nat Genet 37:844–852.

Westerfield M. 1995. The Zebrafish book:a guide for the laboratory use of ze-brafish. Eugene: University of OregonPress.

Xu J, Zhong N, Wang H, Elias JE, Kim CY,Woldman I, Pifl C, Gygi SP, Geula C,Yankner BA. 2005a. The Parkinson’sdisease-associated DJ-1 protein is atranscriptional co-activator that protectsagainst neuronal apoptosis. Hum MolGenet 14:1231–1241.

Xu X, Yang D, Ding JH, Wang W, ChuPH, Dalton ND, Wang HY, Berming-ham JR Jr, Ye Z, Liu F, Rosenfeld MG,Manley JL, Ross J, Jr., Chen J, XiaoRP, Cheng H, Fu XD. 2005b. ASF/SF2-regulated CaMKIIdelta alternativesplicing temporally reprograms excita-tion-contraction coupling in cardiacmuscle. Cell 120:59 –72.

Yang YY, Yin GL, Darnell RB. 1998. Theneuronal RNA-binding protein Nova-2 isimplicated as the autoantigen targetedin POMA patients with dementia. ProcNatl Acad Sci U S A 95:13254–13259.

ROLE OF WIS/SFPQ IN BRAIN DEVELOPMENT 1357