Developmental Biology 341 (2010) 167–175

Contents lists available at ScienceDirect

Developmental Biology

j ourna l homepage: www.e lsev ie r.com/deve lopmenta lb io logy

The PAF1 complex component Leo1 is essential for cardiac and neural crestdevelopment in zebrafish

Catherine T. Nguyen a,1, Adam Langenbacher a,1, Michael Hsieh a, Jau-Nian Chen a,b,c,d,⁎a Department of Molecular, Cell and Developmental Biology, University of California, Los Angeles, CA 90095, USAb Molecular Biology Institute, University of California, Los Angeles, CA 90095, USAc Jonsson Cancer Center, University of California, Los Angeles, CA 90095, USAd Cardiovascular Research Laboratory, University of California, Los Angeles, CA 90095, USA

⁎ Corresponding author. Department of Molecular, CeUniversity of California, Los Angeles, CA 90095, USA.

E-mail address: chenjn@mcdb.ucla.edu (J.-N. Chen).1 These authors contributed equally.

0012-1606/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.ydbio.2010.02.020

a b s t r a c t

a r t i c l e i n f o

Article history:Received for publication 7 September 2009Revised 8 February 2010Accepted 15 February 2010Available online 21 February 2010

Keywords:PAF1 complexHeart developmentNeural crestZebrafish

Leo1 is a component of the Polymerase-Associated Factor 1 (PAF1) complex, an evolutionarily conservedprotein complex involved in gene transcription regulation and chromatin remodeling. The role of leo1 invertebrate embryogenesis has not previously been examined. Here, we report that zebrafish leo1 encodes anuclear protein that has a similar molecular structure to Leo1 proteins from other species. From a geneticscreen, we identified a zebrafish mutant defective in the leo1 gene. The truncated Leo1LA1186 protein lacks anuclear localization signal and is distributed mostly in the cytoplasm. Phenotypic analysis showed that whilethe initial patterning of the primitive heart tube is not affected in leo1LA1186 mutant embryos, the differ-entiation of cardiomyocytes at the atrioventricular boundary is aberrant, suggesting a requirement for Leo1in cardiac differentiation. In addition, the expression levels of markers for neural crest-derived cells such ascrestin, gch2, dct and mitfa are greatly reduced in leo1LA1186 mutants, indicating a requirement for Leo1 inmaintaining the neural crest population. Consistent with this finding, melanocyte and xanthophore popu-lations are severely reduced, craniofacial cartilage is barely detectable, and mbp-positive glial cells are absentin leo1LA1186 mutants after three days of development. Taken together, these results provide the first geneticevidence of the requirement for Leo1 in the development of the heart and neural crest cell populations.

ll and Developmental Biology,

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Embryonic developmental processes are controlled spatiallyand temporally by precise genetic programs. In addition to tissue-specific transcription factors, tightly regulated transcriptional controlsinvolving the recruitment of RNA polymerase II-associated initiationand elongation complexes are required for the initiation of mRNAsynthesis and the stability of nascent transcripts. The subsequentrecruitment of other RNA polymerase II-associated complexes such asthe Srb-Mediator complex or Polymerase-Associated Factor (PAF1)complex promotes chromatin remodeling which further affectstranscription efficiency.

While the roles of tissue-specific transcription factors requiredfor regulating cardiac development have been studied extensively,mechanisms by which global transcription regulatory complexes andchromatin remodeling proteins influence cardiac differentiation arejust being discovered. The important roles of general transcriptioncomplex components in cardiac development have been exemplified

by studies of the zebrafish mutants pandora (pan/spt6) and foggy(fog/spt5). Spt6 and Spt5 are elongation factors associated with RNAPolymerase II following the initiation of transcriptional elongation.Both pan and fog mutant embryos display normal expression of theearly cardiac markers gata4 and nkx2.5, but terminal differentiation ofmyocardial precursors is severely affected (Yelon et al., 1999; Keeganet al., 2002), demonstrating a critical regulatory role for generaltranscription factors in cardiac development.

The PAF1 complex consists of five evolutionarily conservedproteins: Paf1, Ctr9, Leo1, Rtf1, and Cdc73 (Mueller and Jaehning,2002; Rozenblatt-Rosen et al., 2005; Adelman et al., 2006). The PAF1complex has been shown to associate with RNA polymerase II,transcription elongation factors and chromatin remodeling proteins(Shi et al., 1996, 1997; Mueller and Jaehning, 2002; Krogan et al.,2003; Simic et al., 2003), but the precise requirements of individualPAF1 complex components in development are just being evaluated.In yeast, mutants deficient in members of the PAF1 complex exhibitpleiotropic phenotypes and changes in a subset of transcripts,indicative of general transcription defects (Betz et al., 2002; Muellerand Jaehning, 2002). Loss of function of Arabidopsis homologs of Paf1and Ctr9 results in an early flowering phenotype due to a reduction inhistone H3methylation (He et al., 2004; Oh et al., 2004). InDrosophila,RNAi-mediated knockdown of Rtf1 enhances the wing phenotype in

168 C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

a Notch hypomorphic background (Tenney et al., 2006) and aberrantexpression levels of Cdc73 homologues modulate Wnt and Hhsignaling (Mosimann et al., 2006, 2009). Similarly, loss of functionof Rtf1 and Ctr9 in zebrafish results in a host of developmental defectsincluding cardiogenesis abnormalities, a reduction of the neural crestpopulation, and reduced Notch signaling in the developing somites(Akanuma et al., 2007). Furthermore, patients carrying mutations inthe human CDC73 homologue (the Hyperparathyroidism-jaw tumorsyndrome tumor suppressor gene HRPT2) suffer from parathyroidtumors and kidney cysts among other defects (Carpten et al., 2002).Whether mutations in other components of the PAF1 complex arecausative to other congenital defects or diseases is not yet known.

Unlike other PAF1 complex proteins, loss of Leo1 function inyeast causes no obvious phenotypes (Magdolen et al., 1994; Muellerand Jaehning, 2002), and the requirements for Leo1 in vertebratedevelopment have not previously been investigated. Here we reportthe isolation of a zebrafish leo1mutant from a genetic screen designedto identify genes critical for heart development. The leo1LA1186 mutantembryos have dysmorphic hearts and severely reduced blood cir-culation due to a differentiation defect in cardiomyocytes particularlyat the atrioventricular boundary. In addition, neural crest-derivedtissues such as pigment cells, glial cells and craniofacial cartilage areseverely reduced in leo1LA1186 mutant embryos. Overexpression ofwild-type leo1 mRNA in leo1LA1186 mutant embryos suppresses thedefects in both cardiac and neural crest-derived tissues, providing thefirst genetic evidence that Leo1 is essential for the development ofthe heart and neural crest cells in vertebrates.

Materials and methods

Zebrafish husbandry and ENU mutagenesis

Male fish of the Tg(kdr:GFP)LA116 line were mutagenized with ENUas previously described (Mullins et al., 1994; Solnica-Krezel et al.,1994; Choi et al., 2007). LA1186 mutants were identified based ontheir cardiac defects from a screen that surveyed 900 mutagenizedgenomes. Zebrafish colonies were cared for and bred under standardconditions and developmental stages of zebrafish embryos weredetermined using standard morphological features of fish raised at28.5 °C (Westerfield, 2000).

Positional cloning

LA1186 heterozygotes were crossed to the polymorphic WIK strainto generate a hybrid line for mapping. Embryos used for mapping werelysed in embryo lysis buffer (10 mMTris pH8.0, 2 mMEDTA, 0.2% TritonX-100, 100 µg/ml Proteinase K) at 55 °C overnight to obtain genomicDNA.GenomicDNA from24embryos (wild-typeormutant)waspooledand used for bulk segregant analysis with a panel of 200 microsatellitemarkers designed at the Cardiovascular Research Center of theMassachusetts General Hospital (Michelmore et al., 1991; Mably et al.,2003). Primer sequences of the custommarkers 18-839-3 and 18-189-5are: 18-839-3-F, 5′-TACAAACACTGGCACGCCATTAC; 18-839-3-R, 5′-ACTTGCTGTGGGGATTGCAGT; 18-189-5-F, 5′-CCAGATCATTTGTGTGT-CACTATG; and 18-189-5-R, 5′-CTTGGAGCCAATAAATCATTTGTA.

Total RNA was isolated from 1 day post fertilization (dpf) LA1186mutants and their wild-type siblings using RNA Wiz (Ambion) andcDNA was synthesized using the Superscript II Kit (Invitrogen). cDNAfragments were amplified with Phusion polymerase (Finnzymes) andcloned into pCR-Blunt II-TOPO (Invitrogen) for sequencing.

Constructs and injections

The leo1 cDNA constructs were amplified from 1 dpf wild-typeembryo cDNA using Phusion polymerase (Finnzymes) and clonedinto pCS2+3XFLAG for tagging with the FLAG epitope. The plasmids

were cut with NotI and SP6 RNA polymerase was used to generatemRNA for injection. For rescue experiments, embryos from a crossof LA1186/+ fish were injected with 75 pg leo1 mRNA or leo1LA1186

mRNA at the one-cell stage.

Histology

Fixed embryos were dehydrated, embedded in plastic blocks (JB-4,Polysciences), sectioned at 10 μm and stained with 0.1% methyleneblue as previously described (Chen and Fishman, 1996).

Whole mount in situ hybridization

Embryos for in situ hybridization were raised in embryo mediumsupplemented with 0.2 mM 1-phenyl-2-thiourea to maintain opticaltransparency (Westerfield, 2000). Whole mount in situ hybridizationwas performed as described previously (Chen and Fishman, 1996). Theantisense RNA probes used in this study include leo1, nkx2.5, cmlc2(myl7), vmhc, amhc (myh6), bmp4, notch1b, wnt2bb, versican (vcana),crestin, mbp, snai1b, sox9b, foxd3, sox10, dct, gch2, mitfa, tbx1,endothelin-1, fkd2, pax2, and hand2. Wnt2bb was amplified from 2dpf wild-type embryo cDNA by PCR with Phusion polymerase(Finnzymes) and cloned into pCR-Blunt II-TOPO (Invitrogen). Theplasmidwas linearizedwithNotI and SP6 RNApolymerasewas used togenerate antisense riboprobe.

Antibody staining

Embryos injected with RNA encoding FLAG-tagged zebrafish leo1(100 pg) were fixed in 4% PFA in PBS at 75% epiboly. The fixedembryos were incubated in primary antibody (1:50 mouse anti-FLAGM2 (Sigma)) in blocking solution (10% goat serum in PBDT) for 2 h atroom temperature followed by detection with fluorescent secondaryantibody (1:200 anti-mouse IgG1-R (Santa Cruz Biotechnology)).Stained embryos were embedded in 1% low-melt agarose and imagedon a Zeiss LSM510 confocal microscope equipped with a 63× waterobjective.

Alcian blue staining

Embryos were fixed in 4% PFA in PBST after three days of devel-opment. Staining was carried out as previously described (Gollinget al., 2002). In brief, embryos were stained with 0.1% Alcian bluesolution dissolved in 50% ethanol/0.1 M HCL for 2 h. After staining,embryos were rinsed in ethanol and rehydrated gradually into PBST,before being transferred into 0.05% trypsin solution dissolved insaturated sodium tetraborate for 3 h. Embryos were then bleached in1% KOH/3% H2O2 for 1 h.

Results

To identify genes critical for cardiac development, we conductedan ethylnitrosourea mutagenesis screen. From this screen we found amutant, designated LA1186, that exhibits a dysmorphic and dysfunc-tional heart. The primitive heart tube forms normally in LA1186mutants and the expression of multiple early cardiac markers, in-cluding nkx2.5 and cmlc2, are indistinguishable between LA1186mutant embryos and their wild-type siblings (Figs. 1D–G). However,by 30 h post fertilization (hpf), LA1186 hearts exhibit reduced con-tractility. As a result, a cardiac edema occurs and the outflow tract,sinus venosus, and ventricle collapse by 2 days post fertilization (dpf)(Fig. 1I). These findings suggest that LA1186 is required formaintaining the proper function and morphology of the embryonicheart, but is not involved in the initial steps of cardiac specification orheart tube elongation.

Fig. 1. Loss of Leo1 function results in cardiac abnormalities and absence of pigmentation (A, B, C) Lateral view of a two-day-old wild-type embryo (A), a leo1LA1186 mutant (B) and aleo1LA1186 mutant injected with wild-type leo1 mRNA (C). (D, E, F, G) Dorsal view. Expression patterns of cardiac markers, nkx2.5 (D, E) and cmlc2 (F, G), are indistinguishablebetween wild-type siblings (D, F) and leo1LA1186 mutants (E, G) at the 10 somite stage and 24 h post fertilization, respectively. (H, I) Histological sections. In contrast to wild-typehearts (H), the cardiac chambers of leo1LA1186 mutants are collapsed by 2 dpf (I). o, outflow tract; v, ventricle; a, atrium; sv, sinus venosus.

169C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

In addition to cardiac defects, LA1186 mutant embryos exhibita severe reduction in both black (melanophores) and yellow(xanthophores) pigment cells after two days of development(Fig. 1B) and a severe defect in the formation of the jaw after threedays of development (Fig. 6), indicating that LA1186 is required forthe formation of pigment cells and craniofacial structures as well.To assess whether the phenotypes observed in LA1186 mutants aremerely the result of developmental delay, we analyzed the develop-ment of several other organ systems at 2 dpf. We found that themorphogenesis of the brain, otic vesicles, pronephric ducts, thyroid,liver, pancreas, and trunk vasculature are normal in LA1186 mutantsat 2 dpf (Supplementary Fig. 1), suggesting that they are notdevelopmentally delayed compared to wild-type embryos.

LA1186 encodes Leo1, a PAF1 complex protein

As the first step toward identifying the molecular lesion of LA1186,we mapped LA1186 to zebrafish linkage group 18 between markerz8343 and the custom marker 18-189-5. We then generated addi-tional custom markers in this region and detected no recombinationbetween LA1186 and marker 18-839-3 in 1600 meiosis, indicating agenetic distance of less than 0.06 cM between LA1186 and 18-839-3(Fig. 2A). The leo1 gene is in close proximity to 18-839-3, making it agood candidate gene for LA1186. We amplified the coding regionof leo1 from LA1186 mutant embryos and their wild-type siblings.Sequencing analysis revealed a C to T transition at nucleotide 1744that produces a premature stop codon in LA1186 mutant embryos(Fig. 2B), suggesting that the nonsense mutation in leo1 may beresponsible for the cardiac, pigmentation and craniofacial defectsobserved in LA1186.

Transcripts of leo1 can be detected in wild-type embryos by wholemount in situ hybridization analysis as early as the 6-somite (S) stage.During the segmentation stage, leo1 transcripts are present in a widerange of tissues (Figs. 3B,C,E,F). The expression of leo1 becomes re-stricted to the eyes, brain, pharyngeal arches, and anterior somites by

24 hpf, and the transcripts of leo1 are no longer detectable by in situhybridization in zebrafish embryos by 48 hpf (Figs. 3G,H and data notshown). No significant reduction of leo1 signal was noted in LA1186mutant embryos (data not shown), suggesting that leo1LA1186 tran-scripts are not subject to nonsense-mediated decay.

Leo1 is a component of the PAF1 complex, which is associated withRNA polymerase II in the nucleus and has an important role intranscription regulation (Betz et al., 2002; Mueller and Jaehning,2002). The premature stop codon in LA1186 truncates Leo1 prior to itsputative nuclear localization sequence (Fig. 2C) and may therebydisturb nuclear localization of Leo1LA1186. To examine the subcellularlocalization of Leo1, we tagged Leo1wt and Leo1LA1186 with FLAGepitopes and injected these constructs into wild-type embryos at the1-cell stage. As expected, FLAG-Leo1wt is localized to the nucleus,while FLAG-Leo1LA1186 is present predominantly in the cytoplasm(Figs. 2D–G). Given the previously described role of the PAF1 complexduring transcription, the inability of Leo1LA1186 to localize to thenucleus would likely impair the function of the Leo1 mutant protein.

To further evaluate the causative relationship between loss ofLeo1 activity and the LA1186 phenotypes, we injected leo1mRNA intoone-cell stage embryos collected from a cross of LA1186 hetero-zygotes. All embryos from the LA1186 heterozygous cross that wereinjected with 75 pg of leo1 mRNA had normal jaw and cardiac mor-phology and normal levels of pigmentation (n=180) (Fig. 1C), eventhough genotyping these embryos with marker 18-939-3 indicatedthat one quarter were homozygous for LA1186. On the contrary,injection of leo1LA1186 mRNA to embryos collected from a LA1186heterozygous cross was unable to rescue LA1186 mutant phenotypes(data not shown). These findings indicate that the Leo1LA1186 proteinis not functional and demonstrate that wild-type leo1 mRNA is suf-ficient to rescue both cardiac and pigmentation phenotypes in LA1186mutant embryos. Furthermore, overexpression of leo1LA1186 transcriptin wild-type or LA1186 heterozygotes did not result in any noticeablemorphological defects (data not shown), suggesting that Leo1LA1186 isnot acting as a dominant-negative protein in mutant embryos.

Fig. 2. Positional cloning of Leo1LA1186. (A) LA1186wasmapped to linkage group 18 between markers z8343 and 18-189-5. No recombinants were detected between LA1186 and thecustommarker 18-839-3 from 1600meioses. Asteriskmarks the position of the nonsensemutation. (B) Sequencing of the leo1 locus inwildtype and LA1186 embryos identified a C toT transition, resulting in a premature stop codon (*). (C) Schematic diagram of the molecular structure of leo1. Zebrafish Leo1WT contains a Leo1 domain (green), nuclear localizationsequence (red), and an aspartic acid-rich tail (dark blue). The Leo1LA1186 protein is truncated prior to the nuclear localization sequence. (D, E, F, G) Cellular localization of Leo1WT andLeo1LA1186. FLAG-tagged Leo1WT (D) colocalized with DAPI (E) in the nucleus. FLAG-tagged Leo1LA1186 was detected mostly in the cytoplasm (F, G). FLAG, red; DAPI, blue.

170 C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

Zebrafish Leo1 is a 696 amino acid protein that shares a high levelof similarity to yeast Leo1 (55%) and other vertebrate Leo1 proteins(N70%) (Supplementary Fig. 2). As with all the Leo1 proteinsanalyzed, the zebrafish Leo1 protein contains a unique Leo1 domain

Fig. 3. In situ analysis of leo1mRNA expression. (A–F) Leo1 transcripts are detected inmany tiin the brain, pharyngeal arches and eyes and tapers posteriorly in the somites. (A, B, C, E, G

and a putative nuclear localization sequence. In addition, zebrafishLeo1 contains an aspartic acid-rich C-terminal region that is notpresent in the yeast, mouse and human Leo1 proteins (Fig. 2C andSupplementary Fig. 2). To determine whether this C-terminal tail was

ssues beginning from the 6S to the 20S stage. (G–H) At 24 hpf, leo1 is strongly expressed) Lateral view. (D, E, F, H) Anterior dorsal view.

171C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

essential for Leo1 function, we generated a truncated form of thezebrafish Leo1 protein lacking sequences after amino acid 647. Weinjected the transcripts generated from this truncated Leo1 constructinto embryos collected from an LA1186 heterozygous cross and foundthat the truncated protein was able to rescue the LA1186 mutantphenotype (n=69), suggesting that the aspartic acid-rich tail is notnecessary for Leo1 activity.

Leo1 is required for the differentiation of the atrioventricular boundaryof the heart

LA1186 mutant hearts are unable to establish proper circulation,which is often associated with abnormal morphology and cardiacgene expression in zebrafish. Indeed, the sinus venosus, outflow tract,and ventricle are collapsed in 2 dpf leo1LA1108 mutants (Fig. 1I). Weutilized several cardiac markers to assess whether leo1LA1108 mutanthearts differentiate properly. We found that while the atrial marker,amhc, is properly expressed in the atrium, the expression domain ofthe ventricular marker, vmhc, is expanded to the atrium (Figs. 4A–D),suggesting a defect in the regionalization of cardiac chamber-specificgenes. The differentiation defects of the cardiomyocytes are mostprominent at the atrioventricular (AV) boundary. In wild-type hearts,expression of bmp4, versican and wnt2bb is restricted to myocardialcells at the AV boundary after two days of development. In leo1

Fig. 4. Cardiac differentiation phenotypes of leo1LA1186. Normally, vmhc expression is restrictemutants (B). Expression of amhc, however, is restricted properly to the atrial chamber in botthe endocardium at the AV boundary of wild-type embryos (E), but are absent in leo1LA1186

myocardium of the AV boundary in wild-type embryos. Both bmp4 (H) and versican (J) tramutant embryos (L). Dashed lines (in A–D) and arrows (in E, G, I, K) mark the boundary o

mutant hearts, the expression levels of bmp4, versican andwnt2bb aregreatly reduced and the expression domains of bmp4 and versican areno longer restricted to the AV boundary (Figs. 4G–L). Furthermore,expression of notch1b in endocardial cells at the AV boundary isabsent in leo1 mutant hearts (Fig. 4F). Taken together, these dataindicate an important role for Leo1 in cardiac differentiation, espe-cially for those cells located at the AV boundary.

Leo1 is required for the development of neural crest cells

Leo1mutant embryos exhibit a severe reduction in both the mela-nophore and xanthophore pigment cell populations after two days ofdevelopment (Fig. 1B). To further investigate the pigmentationdefects in leo1 embryos, we examined the expression of gch2, dct,and mitfa (markers for xanthophores (Knight et al., 2004), melano-phores (Lamason et al., 2005) and early melanoblasts (Hong et al.,2005), respectively), at 1 dpf. Consistent with the loss of melano-phores and xanthophores in 2-day-old leo1LA1186 mutant embryos,we found that the expression levels of these markers were greatlyreduced in leo1LA1186 mutant embryos after one day of development,suggesting that Leo1 may regulate the differentiation of pigment cells(Fig. 5).

In addition, after three days of development defects in craniofacialformation become apparent in leo1LA1186 mutant embryos. Staining

d in the ventricle in wild-type embryos (A), but is expanded into the atrium in leo1LA1186

h wild-type (C) and leo1LA1186 mutant embryos (D). Notch1b transcripts are detected inmutants (F). Expression of bmp4 (G), versican (I), and wnt2bb (K) are detected in the

nscripts are no longer restricted in leo1LA1186 mutants, while wnt2bb is not detected inf atrium and ventricle.

Fig. 5. Pigmentation phenotypes of leo1LA1186. (A, B)Gch is expressed in xanthophores of the head and trunk (A). Expression levels of gch are reduced in leo1LA1186mutants (B). (C, D)Dct, amarker for differentiatedmelanophores, is severely reduced in leo1LA1186mutants (D) compared towild-type siblings (C). (E, F)Mitfa transcripts are detected in earlymelanoblasts alongthe head and trunk (E). Leo1LA1186 mutants display a reduction inmitfa expression (F).

172 C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

with Alcian blue showed that the ethmoid plate and all pharyngealarches are completely absent and only remnants of trabeculae craniiremain in 3-day-old leo1LA1186 mutant embryos (Figs. 6A–D). By 5 dpf,a very reduced craniofacial skeleton has formed in leo1LA1186 mutants(Supplementary Fig. 3). However, the expression of tbx1 andendothelin-1, markers for the pharyngeal arch epithelium and meso-dermal core (Piotrowski et al., 2003), are unaffected in leo1 mutantembryos (Figs. 6E–H). These findings indicate that Leo1 activity is notrequired for the formation of the epithelium or the mesodermal coreof the pharyngeal arches, but is required for the differentiation of anappropriate number of craniofacial chondrocytes, or the proliferationor survival of these cells.

Melanophores, xanthophores and the cartilaginous componentof the craniofacial skeleton are derived from neural crest cells (Kelshet al., 1996; Kimmel et al., 2001; Barrallo-Gimeno et al., 2004;Montero-Balaguer et al., 2006). Another cell type that neural crestcells are known to contribute to in zebrafish are glia of the peripheralnervous system (PNS) (Kelsh et al., 2000). To determine if Leo1 is alsorequired for PNS glial cell development, we examined the expressionof the glial marker myelin basic protein (mbp) in wild-type andleo1LA1186 mutant embryos. Wild-type embryos display mbp expres-sion in cranial glia of the CNS and the Schwann cells of the trunk.Interestingly, no mbp positive cells were detected in leo1LA1186

mutants after 3 and 5 days of development, indicating a completeloss of both cranial and trunk glia (Supplementary Fig. 4). The loss ofperipheral glia (Schwann cells) is consistent with the notion that thedevelopment of neural crest-derived tissues is defective in leo1LA1186

mutants and the loss of CNS glia suggests that Leo1 may play a rolein the development of glia in addition to its function in neural crestdevelopment.

The loss of multiple neural crest-derived cell types in leo1LA1186

mutant embryos suggests that Leo1 may be required for neural crestspecification and/or differentiation. To determine if Leo1 is requiredfor the initial specification of the neural crest cell population, weevaluated the expression pattern of early neural crest markers. Wefound that the expression patterns of pre-migratory neural crestmarkers (sox9b, sox10, foxd3, and snai1b; (Barrallo-Gimeno et al.,2004; Hong et al., 2005; Yan et al., 2005)) are indistinguishablebetweenwild-type and leo1LA1186 embryos at the 10S stage, indicatingthat neural crest cells are properly specified in leo1LA1186 mutantembryos (Supplementary Fig. 5). Furthermore, we detected nosignificant changes in the expression of the pharyngeal arch neuralcrest markers dlx2 (Figs. 7A,B) and hand2 (data not shown) inleo1LA1186 mutant embryos at the 18S stage, suggesting that thereduction of the craniofacial skeleton is not due to an absence ofneural crest progenitors, but rather the inability of a sufficient numberof these cells to develop into craniofacial chondrocytes (Hong et al.,2005). Interestingly, at both the 12S stage and 24 hpf, the expressionlevel of the pan-neural crest marker crestin is reduced in leo1LA1186

mutant embryos (Figs. 7C–J). We also noted that the loss of crestinexpression is most severe in the cranial neural crest region, thepopulation of neural crest cells anterior to the otic vesicle. Further-more, we could barely detect any crestin-positive cells in the cranialneural crest region at the 12S stage and only residual crestin-positive

Fig. 6. Craniofacial phenotypes of leo1LA1186. (A, B, C, D) Alcian blue staining of wild-type (A, C) and leo1LA1186 embryos (B, D) at 3 days post fertilization. Leo1LA1186 exhibit severecraniofacial defects including absence of the ethmoid plate from the anterior neurocranium and complete loss of pharyngeal arches. (A, B) Lateral view. (C, D) Ventral view. ANC,anterior neurocranium; t, trabecular cranii; m, mandibular arch; h, hyoid arch; P3–7, pharyngeal arches 3–7. (E, F, G, H) Tbx1 and endothelin-1 (edn-1) expression patterns in thepharyngeal arch epithelium and mesenchymal core are indistinguishable between wild-type embryos (E, G) and leo1LA1186 mutants (F, H) at 27 hpf.

Fig. 7.Development of neural crest cells is dependent upon Leo1 function. (A, B) Dorsal anterior viewof dlx2 expression in pharyngeal arch neural crest at the 18S stage. The expressionpattern of dlx2 is normal inwild-type (A) and leo1LA1186 embryos (B). (C–J) Pan-neural crestmarker, crestin, at the 12 somite stage (C–F) and 24hpf (G–J). Crestin expression is detectedin anterior and trunk neural crest of wild-type embryos (C, E, G, I), but is severely reduced in the region anterior to the otic vesicle of leo1LA1186mutants (D, F, H, J). (C, D, G, H) Lateralview. (E, F, I, J) Dorsal anterior view.

173C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

174 C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

cells are detected in this region at 24 hpf (Fig. 7). Taken together,these data suggest that Leo1 is not required for the initial specificationof neural crest cells, but is required for their subsequent developmentinto derivatives such as pigment cells and craniofacial cartilage.

Discussion

From a forward genetic screen, we isolated the zebrafish LA1186mutant exhibiting abnormalities in cardiac development as well ascraniofacial defects and a reduction in pigmentation. Subsequentmolecular cloning showed that the LA1186 locus encodes Leo1, acomponent of the PAF1 complex. Consistentwith previous findings thatthe PAF1 complex is associatedwith RNA polymerase II and plays a rolein chromatin remodeling (Shi et al., 1996, 1997; Mueller and Jaehning,2002; Krogan et al., 2003), the Leo1 protein is localized to the nucleus,and this nuclear localization is required for its biological function. Therequirements for Leo1 in the development of multi-cellular organisms,however, have not been previously examined. Our study shows that theleo1LA1186 mutation is embryonic lethal, indicating that Leo1 is anessential gene for vertebrates. Our phenotypic analysis of leo1LA1186

mutant embryos further reveals defects in the differentiation of theheart, especially at the AV boundary, and the development of neuralcrest-derived tissues, providing the first genetic evidence for therequirements of Leo1 in vertebrate development.

The PAF1 complex is an evolutionarily conserved protein com-plex. A large number of biochemical studies in yeast and culturedmammalian cells have shown that the PAF1 complex consists ofPaf1, Ctr9, Rtf1, Cdc73 and Leo1(Mueller and Jaehning, 2002;Rozenblatt-Rosen et al., 2005; Adelman et al., 2006). In zebrafish,embryos lacking the activity of Ctr9 or Rtf1 have defects in a widerange of tissues including the heart and neural crest cells (Akanumaet al., 2007). The similarity of the defects in heart development andneural crest-derived tissues among leo1, rtf1 and ctr9 mutants andmorphants is in agreement with the fact that these proteins aremembers of the same protein complex. However, while the cardiacand neural crest phenotypes of ctr9 and rtf1 morphants are highlysimilar (Akanuma et al., 2007), the effect of loss of leo1 functionis significantly milder, suggesting that each component of thePAF1 complex may be differentially required for various biologicalprocesses. This notion is further supported by the findings thatembryos deficient in Ctr9 and Rtf1 are small and have pleiotropicdefects. It has been carefully documented that both ctr9 morphantsand rtf1 mutants are defective in somite and otic vesicle formationin addition to the aforementioned cardiac and neural crest abnor-malities (Akanuma et al., 2007). The leo1LA1186 mutant embryos, onthe contrary, exhibit normal body size and have defects that arerestricted to the heart and some neural crest-derived tissues. Futurestudies on the precise requirements of each component of the PAF1complex would provide insights into mechanisms by which thePAF1 complex regulates vertebrate development.

Leo1 mutant embryos have a severe reduction of craniofacial car-tilage, melanophores, and xanthophores indicating a requirement forLeo1 activity in neural crest-derived tissues. The expression patternsof all early neural crest markers analyzed are normal in leo1LA1186

mutant embryos suggesting that the specification of neural crest cellsis not affected. However, the severe decrease in the expression of dct,gch and mitfa in leo1LA1186 mutants suggests that Leo1 may be impor-tant for the survival or differentiation of neural crest cells. The loss offunction of Leo1 affects a subset of anterior neural crest cells mostseverely. When analyzing the expression of crestin, a member of afamily of retroelements that is expressed in neural crest cells inzebrafish (Rubinstein et al., 2000), we found that while the overallcrestin expression level is reduced in leo1LA1186 mutant embryos,crestin expression is barely detectable in the region anterior to the oticvesicle. This observation is consistent with the severe loss of cra-niofacial cartilage observed in leo1LA1186 mutant embryos. Interest-

ingly, the expression of other neural crest markers, such as dlx2 andhand2, is indistinguishable between leo1LA1186 mutant embryos andtheir wild-type siblings suggesting that the neural crest populationdoes exist in the anterior region of leo1mutant embryos. It is possiblethat Leo1 directly regulates the expression of crestin and other, as ofyet unknown, genes that are critical for the development of neuralcrest-derived cell types, especially in the population of neural crestcells located anterior to the otic vesicle. Alternatively, the anteriorneural crest might consist of crestin-positive and crestin-negativepopulations, with the formation of the crestin-positive neural crestcell population requiring Leo1 activity. Furthermore, if two distinctneural crest cell populations exist, our finding that the craniofacialcartilage formation is severely defective even though a substantialamount of dlx2 and hand2-positive, crestin-negative neural crest cellsremain in leo1LA1186 mutant embryos would suggest that crestin-positive neural crest cells are major contributors to the cartilagecomponent of the craniofacial skeleton.

Leo1 is also required for proper cell differentiation in the heart. Theinitial patterning of the primitive heart tube appears normal inleo1LA1186 mutant embryos as evidenced by their normal morphologyand the proper expression pattern of multiple cardiac markers ana-lyzed. However, the heart begins to show signs of dysmorphism anddysfunction after 30 hpf. Genes that are normally expressed in car-diomyocytes at the AV boundary, such as bmp4 and versican fail to beproperly regionalized and the expression of wnt2bb is absent inleo1LA1186 mutant hearts, indicating an essential role for Leo1 in thedifferentiation of cardiomyocytes. The mechanisms underlying thecardiac differentiation defect observed in leo1LA1186 embryos are notyet clear. It is conceivable that the failure of the AV boundary to beproperly regionalized in leo1LA1186 mutants is secondary to defects incardiac contractility, since proper cardiac function is required fornormal gene expression and cellular morphology of the AV boundaryendocardium (Bartman et al., 2004; Beis et al., 2005). Alternatively,Leo1 may function as an intrinsic factor to drive cardiomyocyte dif-ferentiation. One caveat of this possibility is that we have not beenable to detect significant leo1 expression in the lateral plate meso-derm and leo1 expression appeared to be excluded from the heart inembryos 24 hpf and older. It is formally possible that the expression ofleo1 in cardiac progenitors or cardiomyocytes is below the level thatcan be detected by whole mount in situ hybridization. However, amore exciting possibility is that Leo1 influences heart developmentvia regulation of the differentiation of neural crest cells. In mouseand chick, neural crest cells have a pivotal role in the developmentof the cardiac outflow tract septum, aorticopulmonary septum, andsemilunar and atrioventricular valves (for review see (Hutson andKirby, 2003; Nakamura et al., 2006; Hutson and Kirby, 2007)). Thecontribution of neural crest cells to the development of the two-chambered zebrafish heart is less well understood. Lineage studieshave demonstrated that a subset of neural crest cells (located in theregion from the midbrain–hindbrain boundary to the first somiteat the 10-somite stage) migrate to the heart and differentiate intocardiomyocytes in the outflow tract, ventricle, and AV boundary andto a lesser extent, the atrium (Li et al., 2003; Sato and Yost, 2003).Removal of these cells by laser ablation results in cardiac ventriculardysfunction and failure of cardiac looping (Li et al., 2003). Intriguingly,the differentiation of the anterior neural crest population, whichwould include a substantial portion of cardiac neural crest, is mostaffected in the leo1LA1186 mutant and the cardiac defects of theleo1LA1186 mutants are similar to those observed in cardiac neuralcrest-ablated embryos. Like cardiac neural crest-ablated embryos,leo1LA1186 mutants exhibit reduced cardiac contractility shortly afterthe heart begins to beat and subsequently develop pericardial edemaand abnormal cardiac looping. Further studies on the causative rela-tionship between the neural crest and cardiac defects in leo1LA1186

mutant embryos using transgenic approaches to drive expressionof leo1 specifically in the neural crest may help to expand our

175C.T. Nguyen et al. / Developmental Biology 341 (2010) 167–175

understanding of the role of neural crest contribution in zebrafishheart development.

Acknowledgements

We thank Drs. A. Sagasti, A. T. Look, T. Schilling, I. Dawid, S. Hong,T. Pitrowski, K. Cheng and J. Talbot for sharing plasmids for in situhybridization analysis. We are grateful to members of the Chen lab forexperimental suggestions and critiques and to Yuan Dong for par-ticipating in the genetic screen. This work was supported by NIH R01HL081700 to JNC, a predoctoral fellowship from the Training Programof Genetic Mechanisms to CTN and an NSF Graduate Research fel-lowship to ADL.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.ydbio.2010.02.020.

References

Adelman, K., Wei, W., Ardehali, M.B., Werner, J., Zhu, B., Reinberg, D., Lis, J.T., 2006.Drosophila Paf1 modulates chromatin structure at actively transcribed genes. Mol.Cell. Biol. 26, 250–260.

Akanuma, T., Koshida, S., Kawamura, A., Kishimoto, Y., Takada, S., 2007. Paf1 complexhomologues are required for Notch-regulated transcription during somitesegmentation. EMBO Rep. 8, 858–863.

Barrallo-Gimeno, A., Holzschuh, J., Driever, W., Knapik, E.W., 2004. Neural crest survivaland differentiation in zebrafish depends on mont blanc/tfap2a gene function.Development 131, 1463–1477.

Bartman, T., Walsh, E.C., Wen, K.K., McKane, M., Ren, J., Alexander, J., Rubenstein, P.A.,Stainier, D.Y., 2004. Early myocardial function affects endocardial cushiondevelopment in zebrafish. PLoS Biol. 2, E129.

Beis, D., Bartman, T., Jin, S.W., Scott, I.C., D'Amico, L.A., Ober, E.A., Verkade, H., Frantsve, J.,Field, H.A., Wehman, A., Baier, H., Tallafuss, A., Bally-Cuif, L., Chen, J.N., Stainier, D.Y.,Jungblut, B., 2005. Genetic and cellular analyses of zebrafish atrioventricularcushion and valve development. Development 132, 4193–4204.

Betz, J.L., Chang, M., Washburn, T.M., Porter, S.E., Mueller, C.L., Jaehning, J.A., 2002.Phenotypic analysis of Paf1/RNA polymerase II complex mutations revealsconnections to cell cycle regulation, protein synthesis, and lipid and nucleic acidmetabolism. Mol. Genet. Genomics 268, 272–285.

Carpten, J.D., Robbins, C.M., Villablanca, A., Forsberg, L., Presciuttini, S., Bailey-Wilson, J.,Simonds, W.F., Gillanders, E.M., Kennedy, A.M., Chen, J.D., Agarwal, S.K., Sood, R.,Jones, M.P., Moses, T.Y., Haven, C., Petillo, D., Leotlela, P.D., Harding, B., Cameron, D.,Pannett, A.A., Hoog, A., Heath III, H., James-Newton, L.A., Robinson, B., Zarbo, R.J.,Cavaco, B.M., Wassif, W., Perrier, N.D., Rosen, I.B., Kristoffersson, U., Turnpenny, P.D., Farnebo, L.O., Besser, G.M., Jackson, C.E., Morreau, H., Trent, J.M., Thakker, R.V.,Marx, S.J., Teh, B.T., Larsson, C., Hobbs, M.R., 2002. HRPT2, encoding parafibromin, ismutated in hyperparathyroidism-jaw tumor syndrome. Nat. Genet. 32, 676–680.

Chen, J.N., Fishman, M.C., 1996. Zebrafish tinman homolog demarcates the heart fieldand initiates myocardial differentiation. Development 122, 3809–3816.

Choi, J., Dong, L., Ahn, J., Dao, D., Hammerschmidt, M., Chen, J.N., 2007. FoxH1 negativelymodulates flk1 gene expression and vascular formation in zebrafish. Dev. Biol. 304,735–744.

Golling, G., Amsterdam, A., Sun, Z., Antonelli, M., Maldonado, E., Chen, W., Burgess, S.,Haldi, M., Artzt, K., Farrington, S., Lin, S.Y., Nissen, R.M., Hopkins, N., 2002.Insertional mutagenesis in zebrafish rapidly identifies genes essential for earlyvertebrate development. Nat. Genet. 31, 135–140.

He, Y., Doyle, M.R., Amasino, R.M., 2004. PAF1-complex-mediated histone methylationof FLOWERING LOCUS C chromatin is required for the vernalization-responsive,winter-annual habit in Arabidopsis. Genes Dev. 18, 2774–2784.

Hong, S.K., Haldin, C.E., Lawson, N.D., Weinstein, B.M., Dawid, I.B., Hukriede, N.A., 2005.The zebrafish kohtalo/trap230 gene is required for the development of the brain,neural crest, and pronephric kidney. Proc. Natl. Acad. Sci. U. S. A. 102, 18473–18478.

Hutson, M.R., Kirby, M.L., 2003. Neural crest and cardiovascular development: a 20-yearperspective. Birth Defects Res. C Embryo Today 69, 2–13.

Hutson, M.R., Kirby, M.L., 2007. Model systems for the study of heart development anddisease. Cardiac neural crest and conotruncal malformations. Semin. Cell Dev. Biol.18, 101–110.

Keegan, B.R., Feldman, J.L., Lee, D.H., Koos, D.S., Ho, R.K., Stainier, D.Y., Yelon, D., 2002.The elongation factors Pandora/Spt6 and Foggy/Spt5 promote transcription in thezebrafish embryo. Development 129, 1623–1632.

Kelsh,R.N., Brand,M., Jiang, Y.J., Heisenberg, C.P., Lin, S., Haffter, P.,Odenthal, J.,Mullins,M.C.,van Eeden, F.J., Furutani-Seiki, M., Granato, M., Hammerschmidt, M., Kane, D.A.,Warga, R.M., Beuchle, D., Vogelsang, L., Nusslein-Volhard, C., 1996. Zebrafishpigmentation mutations and the processes of neural crest development. Devel-opment 123, 369–389.

Kelsh, R.N., Dutton, K., Medlin, J., Eisen, J.S., 2000. Expression of zebrafish fkd6 in neuralcrest-derived glia. Mech. Dev. 93, 161–164.

Kimmel, C.B., Miller, C.T., Keynes, R.J., 2001. Neural crest patterning and the evolution ofthe jaw. J. Anat. 199, 105–120.

Knight, R.D., Javidan, Y., Nelson, S., Zhang, T., Schilling, T., 2004. Skeletal and pigmentcell defects in the lockjaw mutant reveal multiple roles for zebrafish tfap2a inneural crest development. Dev. Dyn. 229, 87–98.

Krogan, N.J., Dover, J., Wood, A., Schneider, J., Heidt, J., Boateng, M.A., Dean, K., Ryan, O.W.,Golshani, A., Johnston, M., Greenblatt, J.F., Shilatifard, A., 2003. The Paf1 complex isrequired for histone H3 methylation by COMPASS and Dot1p: linking transcriptionalelongation to histone methylation. Mol. Cell 11, 721–729.

Lamason, R.L., Mohideen, M.A., Mest, J.R., Wong, A.C., Norton, H.L., Aros, M.C., Jurynec,M.J.,Mao, X., Humphreville, V.R., Humbert, J.E., Sinha, S., Moore, J.L., Jagadeeswaran, P.,Zhao, W., Ning, G., Makalowska, I., McKeigue, P.M., O'Donnell, D., Kittles, R., Parra, E.J.,Mangini, N.J., Grunwald, D.J., Shriver,M.D., Canfield, V.A., Cheng, K.C., 2005. SLC24A5, aputative cation exchanger, affects pigmentation in zebrafish andhumans. Science 310,1782–1786.

Li, Y.X., Zdanowicz, M., Young, L., Kumiski, D., Leatherbury, L., Kirby, M.L., 2003. Cardiacneural crest in zebrafish embryos contributes to myocardial cell lineage and earlyheart function. Dev. Dyn. 226, 540–550.

Mably, J.D., Mohideen, M.A., Burns, C.G., Chen, J.N., Fishman, M.C., 2003. heart of glassregulates the concentric growth of the heart in zebrafish. Curr. Biol. 13, 2138–2147.

Magdolen, V., Lang, P., Mages, G., Hermann, H., Bandlow, W., 1994. The gene LEO1 onyeast chromosome XV encodes a non-essential, extremely hydrophilic protein.Biochim. Biophys. Acta 1218, 205–209.

Michelmore, R.W., Paran, I., Kesseli, R.V., 1991. Identification of markers linked todisease-resistance genes by bulked segregant analysis: a rapid method to detectmarkers in specific genomic regions by using segregating populations. Proc. Natl.Acad. Sci. U. S. A. 88, 9828–9832.

Montero-Balaguer, M., Lang, M.R., Sachdev, S.W., Knappmeyer, C., Stewart, R.A., De LaGuardia, A., Hatzopoulos, A.K., Knapik, E.W., 2006. The mother superior mutationablates foxd3 activity in neural crest progenitor cells and depletes neural crestderivatives in zebrafish. Dev. Dyn. 235, 3199–3212.

Mosimann, C., Hausmann, G., Basler, K., 2006. Parafibromin/Hyrax activates Wnt/Wgtarget gene transcription by direct association with beta-catenin/Armadillo. Cell125, 327–341.

Mosimann, C., Hausmann, G., Basler, K., 2009. The role of Parafibromin/Hyrax as anuclear Gli/Ci-interacting protein in Hedgehog target gene control. Mech. Dev. 126,394–405.

Mueller, C.L., Jaehning, J.A., 2002. Ctr9, Rtf1, and Leo1 are components of the Paf1/RNApolymerase II complex. Mol. Cell. Biol. 22, 1971–1980.

Mullins, M.C., Hammerschmidt, M., Haffter, P., Nusslein-Volhard, C., 1994. Large-scalemutagenesis in the zebrafish: in search of genes controlling development in avertebrate. Curr. Biol. 4, 189–202.

Nakamura, T., Colbert, M.C., Robbins, J., 2006. Neural crest cells retain multipotentialcharacteristics in the developing valves and label the cardiac conduction system.Circ. Res. 98, 1547–1554.

Oh, S., Zhang, H., Ludwig, P., van Nocker, S., 2004. A mechanism related to the yeasttranscriptional regulator Paf1c is required for expression of the Arabidopsis FLC/MAF MADS box gene family. Plant Cell 16, 2940–2953.

Piotrowski, T., Ahn, D.G., Schilling, T.F., Nair, S., Ruvinsky, I., Geisler, R., Rauch, G.J.,Haffter, P., Zon, L.I., Zhou, Y., Foott, H., Dawid, I.B., Ho, R.K., 2003. The zebrafish vangogh mutation disrupts tbx1, which is involved in the DiGeorge deletion syndromein humans. Development 130, 5043–5052.

Rozenblatt-Rosen, O., Hughes, C.M., Nannepaga, S.J., Shanmugam, K.S., Copeland, T.D.,Guszczynski, T., Resau, J.H., Meyerson, M., 2005. The parafibromin tumorsuppressor protein is part of a human Paf1 complex. Mol. Cell. Biol. 25, 612–620.

Rubinstein, A.L., Lee, D., Luo, R., Henion, P.D., Halpern, M.E., 2000. Genes dependent onzebrafish cyclops function identified by AFLP differential gene expression screen.Genesis 26, 86–97.

Sato, M., Yost, H.J., 2003. Cardiac neural crest contributes to cardiomyogenesis inzebrafish. Dev. Biol. 257, 127–139.

Shi, X., Finkelstein, A., Wolf, A.J., Wade, P.A., Burton, Z.F., Jaehning, J.A., 1996. Paf1p, anRNA polymerase II-associated factor in Saccharomyces cerevisiae, may have bothpositive and negative roles in transcription. Mol. Cell. Biol. 16, 669–676.

Shi, X., Chang, M., Wolf, A.J., Chang, C.H., Frazer-Abel, A.A., Wade, P.A., Burton, Z.F.,Jaehning, J.A., 1997. Cdc73p and Paf1p are found in a novel RNA polymerase II-containing complex distinct from the Srbp-containing holoenzyme. Mol. Cell. Biol.17, 1160–1169.

Simic, R., Lindstrom, D.L., Tran, H.G., Roinick, K.L., Costa, P.J., Johnson, A.D., Hartzog, G.A.,Arndt, K.M., 2003. Chromatin remodeling protein Chd1 interacts with transcriptionelongation factors and localizes to transcribed genes. EMBO J. 22, 1846–1856.

Solnica-Krezel, L., Schier, A.F., Driever, W., 1994. Efficient recovery of ENU-inducedmutations from the zebrafish germline. Genetics 136, 1401–1420.

Tenney, K., Gerber, M., Ilvarsonn, A., Schneider, J., Gause, M., Dorsett, D., Eissenberg, J.C.,Shilatifard, A., 2006. Drosophila Rtf1 functions in histone methylation, geneexpression, and Notch signaling. Proc. Natl. Acad. Sci. U. S. A. 103, 11970–11974.

Westerfield, M., 2000. The Zebrafish Book. The University of Oregon Press.Yan, Y.L., Willoughby, J., Liu, D., Crump, J.G., Wilson, C., Miller, C.T., Singer, A., Kimmel, C.,

Westerfield, M., Postlethwait, J.H., 2005. A pair of Sox: distinct and overlappingfunctions of zebrafish sox9 co-orthologs in craniofacial and pectoral fin develop-ment. Development 132, 1069–1083.

Yelon, D., Horne, S.A., Stainier, D.Y., 1999. Restricted expression of cardiac myosin genesreveals regulated aspects of heart tube assembly in zebrafish. Dev. Biol. 214, 23–37.