Mutation of pescadillo Disrupts Oligodendrocyte Formation in Zebrafish

Background In vertebrates, the myelin sheath is essential for efficient propagation of action potentials along the axon shaft. Oligodendrocytes are the cells of the central nervous system that create myelin sheaths. During embryogenesis, ventral neural tube precursors give rise to oligodendrocyte progenitor cells, which divide and migrate throughout the central nervous system. This study aimed to investigate mechanisms that regulate oligodendrocyte progenitor cell formation. Methodology/Principal Findings By conducting a mutagenesis screen in transgenic zebrafish, we identified a mutation, designated vu166, by an apparent reduction in the number of oligodendrocyte progenitor cells in the dorsal spinal cord. We subsequently determined that vu166 is an allele of pescadillo, a gene known to play a role in ribosome biogenesis and cell proliferation. We found that pescadillo function is required for both the proper number of oligodendrocyte progenitors to form, by regulating cell cycle progression, and for normal levels of myelin gene expression. Conclusions/Significance Our data provide evidence that neural precursors require pes function to progress through the cell cycle and produce oligodendrocyte progenitor cells and for oligodendrocyte differentiation.


Introduction
In the vertebrate nervous system, the rapid and efficient transmission of electrical impulses along many axons requires the presence of an insulating myelin sheath. In the central nervous system (CNS), the myelin sheath is formed by a population of glial cells known as oligodendrocytes [1]. Within the spinal cord, most oligodendrocytes arise from a ventral region known as the pMN domain, which also gives rise to motor neurons and interneurons [2]. Once specified, oligodendrocyte progenitor cells (OPCs) migrate out of the pMN domain toward their target axons in the lateral spinal cord. These OPCs divide and extend multiple fine membrane processes as they migrate [3]. Upon reaching their target axons, OPCs stop dividing and extend processes that contact and wrap multiple axons in tube-like structures that are then compacted to form the myelin sheath [4].
The mechanisms that pattern the dorsoventral axis of the neural tube to establish pMN precursors are thoroughly described [5]. By contrast, the mechanisms that specify OPCs from pMN precursors and regulate their subsequent division, migration and differentiation to myelinating cells are poorly understood. In an effort to fill this gap we conducted an ENU based mutagenesis screen in zebrafish carrying the Tg(olig2:EGFP) reporter, which labels pMN precursors and oligodendrocyte lineage cells [6]. Here we describe one mutation, vu166, which causes a reduction of oligodendro-cytes. We determined that the vu166 allele is a nonsense mutation of pescadillo (pes), which has been implicated in ribosome biogenesis. Our analysis indicates that the oligodendrocyte deficit of pes mutant larvae results from disruption of cell cycle progression among neural precursors. In addition, we found that OPCs fail to migrate normally in mutant embryos as a consequence of an altered cellular environment and that oligodendrocytes express abnormally low levels of myelin genes in mutant larvae, raising the possibility that Pes function is necessary for oligodendrocyte differentiation.

Results
The vu166 mutation truncates the protein coding sequence of the pescadillo gene We recently conducted an ethyl nitrosourea (ENU) mutagenesis screen using a Tg(olig2:EGFP) transgenic line of zebrafish, which expresses enhanced green fluorescent protein (EGFP) under the control of olig2 regulatory DNA [6]. This transgenic reporter reveals dorsally migrating OPCs in living embryos ( Figure 1C), permitting identification of mutations that disrupt the number and distribution of OPCs. We identified the vu166 allele because homozygous mutant larvae had a deficit of dorsal spinal cord OPCs and oligodendrocytes at 4 days post fertilization (dpf) ( Figure 1D). vu166 mutant larvae also had defects in forebrain, eye, jaw, otolith, pectoral fin, and body length ( Figure 1B). We used bulked segregant analysis [7] to map the vu166 locus to chromosome 5 (data not shown). Subsequent mapping experiments revealed that the vu166 mutation was between microsatellite markers z6727 and z7351 ( Figure 1E). We noted that the pescadillo (pes) gene maps to the same genomic location [8]. A previously described mutant allele of pes, pes hi2 , had a retroviral insertion in the first exon upstream of the translational start site, which abolished transcript accumulation [8,9]. pes hi2 and vu166 mutant larvae had indistinguishable morphological defects, therefore, we performed a complementation test using pes hi2+/2 and vu166 +/2 carriers. Approximately one quarter of the progeny had morphological defects similar to pes and vu166 mutant larvae ( Figure 1F,G) indicating that the vu166 allele disrupts pes function. Sequencing of pescadillo genomic DNA obtained from vu166 mutant larvae revealed a single base pair change predicted to introduce a premature stop codon in the Pescadillo homology domain of the protein ( Figure 1H). Therefore, we hereafter refer to the vu166 allele as pes vu166 . Mutations that disrupt the BRCA1 Cterminal (BRCT) domain of Yph1p, the yeast Pes homolog, prevent growth indicating that the BRCT domain is critical to Pes function [10]. The pes vu166 mutation truncates the Pes protein before the BRCT domain, likely resulting in a complete loss of pes function.

Loss of pes function disrupts oligodendrocyte development
The Tg(olig2:EGFP) reporter revealed an apparent deficit of dorsally migrated OPCs in pes mutant larvae. To validate this observation and more carefully investigate the distribution of OPCs, we performed immunohistochemistry to detect expression of Sox10, a specific marker of oligodendrocyte lineage cells [11], and b-Acetylated Tubulin, which labels axon tracts that mark the white matter of the spinal cord after myelination occurs (Figure 2A,B). At 3 dpf, the number of Sox10 + cells that occupied the medial, axon-poor gray matter region of the spinal cord and the ventral spinal cord was indistinguishable between pes mutant larvae and their wild-type siblings ( Figure 2C). However, fewer Sox10 + cells occupied the dorsal spinal cord and the axonrich white matter region in mutant larvae than in wild type ( Figure 2C). At 4 dpf, the deficit of dorsal and white matter Sox10 + cells persisted in pes mutant larvae ( Figure 2B,D). However, 4 dpf mutant larvae had excess ventral Sox10 + cells, which mostly occupied gray matter, resulting in equivalent total number of Sox10 + cells in wild-type and mutant larvae ( Figure 2B,D). These data indicate that pes loss of function results in a transient deficit of oligodendrocyte lineage cells and disrupts their normal distribution, perhaps by disrupting OPC migration.
The deficit of white matter oligodendrocyte lineage cells raised the possibility that pes mutant larvae have a deficit of myelination. To test this, we performed in situ RNA hybridization to detect expression of genes characteristic of differentiating oligodendrocytes. Whereas mutant and wild-type larvae had similar number of ventral mbp + cells, mutant larvae had significantly fewer dorsal mbp + cells and fewer ventral and dorsal plp1a + and cldnk + cells than wild type ( Figure 3A-G). Additionally, the amount of reaction product detected in individual cells of mutant larvae was substantially less that in wild-type larvae for all three probes, including cells that occupied their normal positions near the pial surface ( Figure 3A-F). Although the apparent reduction of reaction product could result from decreased probe penetration in mutant larvae, a sox2 RNA probe detected higher levels of gene expression in mutants than in wild-type (see below). To further investigate oligodendrocyte differentiation, we performed quantitative real-time PCR to assay the relative expression levels of the CNS myelin genes myelin protein zero (mpz) and 36K at 4 dpf. Both genes were expressed at levels fourfold lower in mutant larvae relative to wild type ( Figure 3H). Therefore, pes function is necessary both for the proper number and distribution of oligodendrocyte lineage cells and for oligodendrocyte differentiation.

Loss of pes function disrupts progression of spinal cord cells through the cell cycle
The reduction in dorsal OPCs and the previously proposed role of pes in cell proliferation [12] raised the possibility that pes function is necessary for the division of OPCs after they migrate out of the pMN domain. To assay the proliferation of OPCs in mutant larvae we exposed the larvae to the thymidine analogue BrdU, an S-phase marker, at various time-points during early development. In wild-type larvae, most OPCs were labeled with BrdU when the pulses were conducted between 52 and 96 hours post fertilization (hpf), and these labeled OPCs represented the vast majority of the BrdU + cells in the spinal cord ( Figure 4A,C and data not shown). Consistent with a requirement for pes in cell division, there were no BrdU positive OPCs in the spinal cords of mutant larvae. However, this experiment revealed that mutant spinal cords had a five-fold increase in the total number of BrdU + cells, which were mostly adjacent to the central canal and medial septum ( Figure 4B,D,I). Additionally, the number of cells that were labeled by the M-phase marker phosphohistone H-3 (PH-3) ( Figure 4E,F) was dramatically increased in mutants ( Figure 4I). Nevertheless, there was no significant difference in the total number of spinal cord cells between mutant and wild type ( Figure 4J).
One possible explanation for an increase of cell cycle activity without an accompanying increase in cell number is that cell division is balanced by cell death. To investigate this possibility we performed a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay to label dying cells ( Figure 4G,H). Mutant larvae had less than two-fold more TUNEL + cells than wild type ( Figure 4J), less than the five-fold increase of dividing cells. Therefore, elevated levels of cell death may not fully account for the lack of increase in total cell number despite the substantial increase in apparent dividing cells.
An alternative explanation is that pes function is necessary for cell cycle progression. To test this idea, we investigated the cell cycle kinetics of pes null larvae. To determine the length of time that spinal cord cells spend in S/G2 phase, we performed a BrdU pulse-chase experiment in which larvae were labeled with BrdU and then fixed at various time-points followed by co-labeling with anti-BrdU and anti-PH3 antibodies. By finding the period of time that it took for BrdU + cells to become labeled with PH-3, we calculated the time it took them to progress from S-phase to mitosis. Our results indicate that whereas spinal cord cells of wild-type larvae progressed from labeling to M phase within roughly six hours, mutant larvae had cells that were still entering mitosis as long as 10 hours after the pulse, the longest interval we collected ( Figure 5A).
In light of this information and the known role of pes in regulating S-phase progression in yeast [10], we predicted that the delay in cell cycle progression in mutant zebrafish is due to a failure to progress through S-phase. To test this possibility, we conducted a second BrdU pulse-chase experiment in which we first pulsed cells with BrdU and then, at varying time-points, pulsed them with the thymidine analog EdU and fixed them immediately. Cells that co-labeled with BrdU and EdU were in Sphase during the initial pulse and had not yet exited it by the start of the second pulse. By three hours after the initial pulse, the majority of the cells in wild-type spinal cords had progressed through S-phase ( Figure 5B). In contrast, only about four percent of the cells in mutant spinal cords had exited S-phase. Therefore, loss of pes function impairs the ability of spinal cord precursors to progress through S phase.
These data indicate that spinal cord precursors in pes mutant larvae failed to progress through the cell cycle at their normal rates, therefore, one possible explanation for the transient deficit of OPCs is that spinal cord cells are maintained as precursors at the expense of specified cell types. To test this, we examined the expression of the precursor marker sox2. At 4 dpf in wild-type larvae, sox2 expression was limited to a few cells bordering the spinal cord central canal ( Figure 5C). By contrast, many spinal cord cells bordering the central canal and medial septum of pes mutant larvae expressed sox2 ( Figure 5D), consistent with the idea that pes function in necessary for neural precursors to progress through the cell cycle and differentiate.

Loss of pes function affects OPC and axon morphology cell-nonautonomous
The abnormal distribution of OPCs in pes mutant larvae raised the possibility that pes also functions in OPC migration. To test this we performed in vivo time-lapse imaging of OPCs in wild-type and mutant larvae expressing Tg(nxk2.2a:mEGFP), which encodes a membrane tethered green florescent protein under the control of nkx2.2a regulatory DNA. This reporter reveals the migratory behavior and process activity in the myelinating subpopulation of OPCs, as well as some axons that descend from the hindbrain through the spinal cord [3]. In wild-type larvae, OPCs extended numerous fine processes that eventually contacted and wrapped axons in tubular sheaths ( Figure 6A). By contrast the OPCs in pes vu1662/2 larvae rarely contacted axons, and instead exhibit thickened, disorganized processes ( Figure 6B). These movies also revealed that, compared to wild type ( Figure 6E), axons extended abnormally in mutant larvae, typified by axon wandering from the normal straight trajectories, frequent branching and repeated growth cone collapse ( Figure 6F).
Are these defects caused by a disruption in the environment through which the cells migrate, or do they instead reflect a direct, cell autonomous requirement for pes in cell migration, process extension and axogenesis? To address this question we created genetic mosaic embryos containing both mutant and wild-type cells. We found that wild-type OPCs transplanted into pes vu1662/2 larvae had thickened, disorganized membrane processes ( Figure 6C), similar to those of mutant OPCs in nonmosaic larvae. Although mutant cells transplanted into wild-type larvae rarely developed as OPCs (data not shown), those that did form extended processes that contacted and wrapped axons in a wildtype fashion ( Figure 6D). These results indicate that the cellular morphology defects of mutant larvae are cell nonautonomous. Consistent with this interpretation, pes vu1662/2 neurons transplanted in wild-type larvae extended axons through the spinal cord in a manner that was indistinguishable from that of wild-type neurons ( Figure 6G).
To gain greater insight into the cellular environment within the spinal cords of pes vu1662/2 larvae, we performed immunohistochemistry using an antibody against Zrf-1, which labels the processes of radial glia in the spinal cord at 4 dpf. When compared to wild-type radial glia, we found that mutant larvae had a disorganized labeling pattern ( Figure 6H,I), as well as regions of the spinal cord that appear to be devoid of radial glia processes. These data indicate that the spinal cords of pes vu1662/2 larvae have an abnormal cellular organization, which likely impairs cell migration and axon extension.

Discussion
Pescadillo is required for cell cycle progression pes was first identified in a zebrafish insertional mutagenesis screen on the basis of its loss-of-function phenotype, which included abnormally small eyes, brain, fins, liver and gut [8]. Subsequently, human cancer cell lines were found to express the homologous gene, PES1, at elevated levels [13] and PES1 overexpression induced transformation of non-tumorogenic fibroblast cell lines [14], thus raising the possibility that Pes proteins promote cell proliferation. Consistent with this, yeast cells lacking function of Yph1p, the yeast pes homolog, underwent cell cycle arrest [10], Pes1 mutant mouse embryos arrested as early as the eight-cell stage [12] and knockdown of pes1 function by RNAi in two breast cancer cell lines resulted in a significant inhibition of cell growth [15].
In yeast, Yph1p, also known as Nop7p, is required for ribosome biogenesis [16], suggesting that loss of Yph1p function resulted in cell cycle arrest because of ribosome depletion. However, Yph1p can interact with many proteins that regulate cell cycle progression, particularly those that are active during S-phase [10]. In the absence of Yph1p function, yeast cells released from hydroxyurea-induced arrest proceeded through S-phase at a significantly slower rate than normal. Consistent with this result, we observed a dramatic elongation of the cell cycle, and S-phase in particular, in the spinal cords of pes mutant zebrafish. However, the mechanism by which Pes regulates the cell cycle has yet to be elucidated. One possibility is that Pes directly regulates cell cycle progression through its interaction with MEC1, the yeast ATM/ ATR homologue, which regulates cell cycle checkpoint control [10]. Pes also interacts with CDC28, the yeast cyclin dependant kinase and a known regulator of cell cycle progression. Therefore, the failure of neural precursors to progress normally through the cell cycle in pes mutant zebrafish may reveal a specific role for Pes in cell cycle regulation.

The requirement for Pes function in OPC migration and axogenesis is cell non-autonomous
Our time-lapse imaging experiments showed that OPCs had abnormal morphologies and dorsalward migration movements and that axons followed abnormal trajectories within spinal cords of pes null larvae. In frogs, reduction of pes function with antisense morpholino oligonucleotides caused craniofacial defects by interfering with the migration of neural crest cells into the brachial arches [17]. Pes interacts with insulin receptor substrate 1, which regulates insulin and insulin-like growth factor signaling [14], leading the authors to propose that Pes regulates cell intrinsic mechanisms necessary for neural crest migration. However, whether pes function is required in neural crest cells or the cells with which neural crest cells interact as they migrate was not tested. Our genetic mosaic experiments showed that, for OPC migration and axon extension through the spinal cord, the requirement for pes function is cell nonautonomous, indicating that these behaviors do not rely on cell intrinsic functions of pes. Both OPCs and axons require multiple extrinsic guidance cues [18,19] expressed by specific cell populations along their trajectories. Disruption of neural cell specification and differentiation from loss of pes function could interfere with production of these cues, leading to defects in OPC migration and axogenesis.

Pes function and oligodendrocyte differentiation
Knockdown of pes function in frogs blocked expression of retinal marker genes, suggesting that pes promotes cell differentiation [17]. Consistent with this, Pes can bind DNA and act as a transcription factor in reporter gene assays [20]. Our studies indicate that Pes might also promote oligodendrocyte gene expression and differentiation. Although pes mutant larvae had a transient deficit of oligodendrocyte lineage cells, perhaps because of an extended cell cycle, OPCs eventually formed in normal number. However, these cells expressed myelin genes at barely detectable levels by in situ RNA hybridization, a result we validated using quantitative PCR. One possible explanation is that, as with OPC migration, the failure to express myelin genes at wild-type levels is a secondary consequence of an abnormal cellular environment. In particular, because OPCs do not migrate and wrap axons normally in pes mutant larvae they might fail to receive signals from axons that promote oligodendrocyte differentiation. However, oligodendrocytes express myelin genes robustly in cell culture, even in the absence of axons [21]. Notably, lack of pes function does not cause a general block of cell differentiation because neurons appeared to express the differentiation markers acetylated Tubulin ( Figure 2) and Hu proteins (unpublished data) normally. Taken together, these observations raise the possibility Pes functions within the oligodendrocyte lineage to promote myelin gene expression and differentiation.

Ethics Statement
The Institutional Animal Care and Use Committee at University of Colorado Anschutz Medical Campus approved all experiments in which animals were used. (IACUC protocol # B-85408(09)).

Genetic Mapping
The vu166 mutation was mapped to chromosome 5 using bulked segregant analysis of microsatellite markers using standard methods [7]. Whole RNA was isolated from 4 dpf mutant and wild-type larvae and converted into cDNA using a Superscript III reverse transcriptase kit (Invitrogen #18080-051). Candidate genes were sequenced using ABI 3730 DNA sequencing instruments using gene specific primers. No new sequence information was generated for submission to GenBank.

BrdU and EdU labeling
Dechorionated embryos were labeled with 5-bromo-29-deoxyuadine (BrdU) (Roche) by incubating them in 20 mM BrdU in EM with 10% DMSO at room temperature for the first 30 minutes. For longer incubations, the initial BrdU solution was replaced with one containing 20 mM BrdU in EM with 1% DMSO for the indicated amounts of time at 28.5uC. For co-labeling with 5ethynyl-29-deoxyuridine (EdU), following BrdU pulse embryos were allowed to develop at room temperature for the indicated amount of time, then incubated in 2 mM EdU (Click-iT EdU Alexafluor 555 detection kit, Invitrogen #c10338) in EM with 10% DMSO for 30 minutes at room temperature. The fish were then fixed in 4% paraformaldehyde (PFA) in PBS with 116 mM sucrose and 150 mM CaCl 2 at 4uC overnight.

Immunohistochemistry and EdU detection
Fixed embryos were imbedded in 1.5% Agar with 5% sucrose and transferred to a 30% sucrose solution in scintillation vials and incubated at 4uC overnight. The blocks were then frozen and cut into 10 mm transverse section on a cryostat microtome. The sections were incubated with rabbit anti-Sox10 (1:1,000) [9], mouse anti-b-acetylated tubulin (1:1000, Sigma), rabbit antiphosphohistone H-3 (1:1,000, Upstate Biotechnology) and anti-BrdU (1:100, Developmental Studies Hybridoma Bank) antibodies. For fluorescent detection of antibody labeling we used Alexafluor 568 and Alexafluor 647 goat anti-mouse and goat anti-rabbit secondary antibodies (1:200, Invitrogen). To detect EdU incorporation, we incubated the slides in 250 mL of the EdU Detection Reaction mix (prepared according to the manufacturer's direction) for 40 minutes at room temperature. Images were collected on a Zeiss Axio Observer 200 microscope with a PerkinElmer Ultraview Vox spinning disk confocal unit equipped with Volocity imaging software (PerkinElmer). Images were contrast enhanced using either Volocity or Photoshop (Adobe CS4).

RNA in situ hybridization
Embryos were fixed in 4% PFA in DEPC treated PBS overnight at 4uC and stored in methanol at 220uC. Anti-sense RNA probes were synthesized from a linearized plasmid DNA template using Roche digoxygenin labeling reagents and either T3 or T7 RNA polymerases (New England Biolabs). The in situ hybridization was performed using the standard protocol [23]. Images were obtained on a Zeiss Axio Observer 200 microscope with Volocity imaging software. Images were contrast enhanced using Photoshop (Adobe CS4).

Quantitative PCR analysis
Total RNA was isolated from 4 dpf mutant and wild-type larvae and converted to cDNA as above. Relative gene expression was determined using TaqMan reagents (Applied Biosystems #4304437) according to the manufacture's instructions. Probes were used against zebrafish mpz (Assay ID Dr03131917_m1) and 36K (Assay ID Dr03438575_m1) using Rpl13a (Assay ID Dr03101115_g1) as an internal control.

Chimeric analysis of zebrafish embryos
We used the Tg(nkx2.2a:mEGFP) line for chimeric analysis by cell transplantation. 30-50 cells from donors at the 'high' stage of blastula development obtained by crossing pes vu166+/2 ; Tg(nkx2.2a: mEGFP) or Tg(nkx2.2a:mEGFP) wild-type adults were transplanted into same-stage wild-type hosts or hosts derived from pes vu166+/2 adults respectively. Host embryos were raised at 28.5uC and imaged at 3 dpf. Donor embryos were immediately genotyped by PCR. Larvae used for imaging were anesthetized using 3aminobenzoic acid ethyl ester, and mounted in a 2% agarose lateral imaging mold [24]. Images were captured using a 206 dry (NA = 0.75) objective mounted on a motorized Zeiss Axiovert 200 microscope equipped with a PerkinElmer Ultraview Vox spinning disk confocal unit and heated stage to maintain embryos at 28.5uC. Z image stacks were collected and three-dimensional data sets were created using Volocity imaging software. Images were deconvolved followed by contrast enhancement using Volocity.