RESEARCH ARTICLE
Molecular Reproduction & Development 78:22–32 (2011)
cables1 Is Required for Embryonic Neural Development:
Molecular, Cellular, and Behavioral Evidence From the
Zebra？sh
1,21,22,32,3JOLIJN W. GROENEWEG,YVONNE A.R. WHITE,DAVID KOKEL,RANDALL T. PETERSON, 2,41,21,21,2LAWRENCE R. ZUKERBERG,INNA BERIN,BO R. RUEDA,AND ANTONY W. WOOD*
1 Vincent Center for Reproductive Biology, MGH Vincent Department of Obstetrics and Gynecology,
Massachusetts General Hospital, Boston, Massachusetts
2 Harvard Medical School, Boston, Massachusetts
3 Cardiovascular Research Center, Massachusetts General Hospital, Charlestown, Massachusetts 4 Department of Pathology, Massachusetts General Hospital, Boston, Massachusetts
SUMMARY
In vitro studies have suggested that the Cables1 gene regulates epithelial cell proliferation, whereas other studies suggest a role in promoting neural differentiation. In efforts to clarify the functions of Cables1 in vivo, we conducted gain- and loss-of- function studies targeting its ortholog (cables1) in the zebra？sh embryo. Similar to
rodents, zebra？sh cables1 mRNA expression is detected most robustly in embryonic neural tissues. Antisense knockdown of cables1 leads to increased numbers of apoptotic cells, particularly in brain tissue, in addition to a distinct behavioral pheno- type, characterized by hyperactivity in response to stimulation. Apoptosis and the behavioral abnormality could be rescued by co-expression of a morpholino-resistant cables1 construct. Suppression of p53 expression in cables1 morphants partially rescued both apoptosis and the behavioral phenotype, suggesting that the phenotype * Corresponding author: of cables1 morphants is due in part to p53-dependent apoptosis. Alterations in the 55 Fruit Street, THR 933 Boston, MA 02114. expression patterns of several neural transcription factors were observed in cables1 E-mail: awwood@partners.org morphants during early neurulation, suggesting that cables1 is required for early
Supported by Vincent Memorial Research neural differentiation. Ectopic overexpression of cables1 strongly disrupted embry- Funds; Advanced Medical onic morphogenesis, while overexpression of a cables1 mutant lacking the C-terminal Research Foundation and US cyclin box had little effect, suggesting functional importance of the cyclin box. Lastly, National Institutes of Health; Grant marked reductions in p35, but not Cdk5, were observed in cables1 morphants. numbers: RO1CA098333; Collectively, these data suggest that cables1 is important for neural differentiation R01MH086867; R21MH085205
during embryogenesis, in a mechanism that likely involves interactions with the Cdk5/ The authors declare no conflicts of p35 kinase pathway. interest. Published online 10 December 2010 in Mol. Reprod. Dev. 78: 22–32, 2011. ß 2010 Wiley-Liss, Inc. Wiley Online Library (wileyonlinelibrary.com). Received 1 June 2010; Accepted 11 November 2010 DOI 10.1002/mrd.21263
INTRODUCTION
Additional Supporting Information may be found in the online version of this The CDK5 and ABL1 Enzyme Substrate 1 (Cables1/ik3-1) article.
gene, a member of the cyclin superfamily, was identi？ed in Abbreviations: Cdk, cyclin-dependent kinase; hpf, hours post-fertilization; two independent screens for genes that interact with the MO, morpholino oligonucleotide; PMR, photomotor response; TUNEL, terminal cyclin-dependent kinase (Cdk) family of serine/threonine dUTP nick-end labeling.
ß 2010 WILEY-LISS, INC.
CABLES1 AND ZEBRAFISH NEURAL DEVELOPMENT protein kinases (Matsuoka et al., 2000; Zukerberg et al., 2000). In a yeast-two hybrid screening assay, Cables1 pro- tein interacted strongly with Cdk5 (a non-cell cycle-associat- ed kinase required for neural differentiation), while weaker interactions were also observed with cell cycle-associated kinases Cdk2 and Cdk3. Subsequent in vitro studies utilizing COS7 cells, E15 mouse brain lysates and cultured primary rat cortical neurons suggested that Cables1 can function in
two distinct cellular contexts: Firstly, Cables1 promoted
protein–protein interactions among c-ABL kinase, Cdk5 and
p35 that enhanced tyrosine phosphorylation of Cdk5
(Zukerberg et al., 2000); secondly, Cables1 could negative-
ly regulate cell proliferation, by augmenting inhibitory phos-
phorylation cascades that suppress cyclin-dependent
kinase activity (Wu et al., 2001). While informative, these in vitro studies have yet to provide de？nitive insights into the functions of the Cables1 gene in vivo, particularly during early development. Although Cables1 knockout mice are viable, early (non- congenic) strains were notably subfertile, with evidence of fetal resorption in pregnant dams (Zukerberg et al., 2004). These observations suggest that embryonic development is compromised in Cables1 knockouts. Interpretation of these data remains dif？cult, however, due to the potential redundancy of a related gene (Cables2) in mice. In light of multiple studies reporting a loss of Cables1 expression in selected tumor types (Wu et al., 2001; Dong et al., 2003; Tan et al., 2003; Zukerberg et al., 2004; DeBernardo et al., 2005; Zhang et al., 2005; Kirley et al., 2005a; Park do et al., 2007) there remains a strong rationale to clarify the in vivo functions of the Cables1 gene. To broaden our understanding of Cables1 functions in vivo, we have performed gain- and loss-of-function studies of the zebra？sh Cables1 ortholog (cables1; NM_ 001105665.1). Importantly, the lack of a functional paralog in this species permitted a direct investigation of cables1 functions in vivo. The data from these studies support the notion that cables1 is critically important for neural develop- ment in the zebra？sh embryo. The developmental impact of cables1 suppression appears substantially more dramatic than that observed in murine or avian models (Zukerberg et al., 2000; Rhee et al., 2007). In the zebra？sh, knockdown of cables1 results in disruptions in the expression of multiple neural transcription factors, increased apoptosis in embryonic neural tissues, and subsequent behavioral defects charac- Figure 1. A: RT-PCR showing expression of cables1 mRNA through- terized by hyperactivity in response to stimulation. In addi- out embryogenesis in zebra？sh. Numbers indicate hours post-fertili- tion, we present evidence suggesting that the embryonic zation (hpf). B: In situ hybridization showing neural expression functions of cables1 likely involve interactions with the Cdk5/ of cables1 mRNA at 24 hpf; (C) dorsal view of 24 hpf embryo; p35 kinase pathway. Collectively, these data from the zebra- (D) cryosectioned embryo (24 hpf) showing neural expression of ？sh model provide in vivo evidence supporting an essential cables1 mRNA; (E) control (sense probe) showing no signal at 24 hpf. functional role for cables1 during embryonic neural differentiation. (0 hr post-fertilization, hpf), indicating that cables1 mRNA is
provided to embryos as a maternal transcript. By in situ RESULTS hybridization, cables1 mRNA was detected most robustly in Embryonic Cables1 mRNA Expression anterior neural tissues (Fig. 1B,C). Expression in neural
cables1 mRNA was detected by RT-PCR throughout tissues was con？rmed by cryosectioning embryos after in embryogenesis (Fig. 1A), including one cell-stage embryos situ hybridization (Fig. 1D).
23 Mol Reprod Dev 78:22–32 (2011)
Molecular Reproduction & Development GROENEWEG ET AL.
Figure 2. Phenotypes of MO-injected embryos: (A) Control-injected embryo at 24 hpf; (B) cabMO-injected embryo (cables1 morphant) showing
evidence of necrotic degeneration in brain tissues (arrows); (C) cables1 morphant embryo co-injected with MO-resistant cables1 mRNA
(cabRESC), showing rescue of brain necrosis; (D) relative proportions of each phenotype in wild-type (WT), control-injected (Control), cables1
morphant (cabMO) and cables1 rescued (cabRESC) groups. KDP, knockdown phenotype (brain necrosis); OEP, over-expression phenotype (see Fig. 6).
tive experiment, this phenotype was observed in 81% Brain Necrosis in cables1 Morphants
(58/72) of morphants, in a pattern that was highly consistent Embryos microinjected at the one-cell stage with a
among injected embryos. control morpholino oligonucleotide (MO) were indis-
tinguishable from wild-type (non-injected) embryos To ensure target speci？city of cabMO, rescue experi- (Fig. 2A). By contrast, microinjecting a morpholino oligonu- ments were performed using a MO-resistant mRNA con- cleotide targeting zebra？sh cables 1 (cabMO; 2 ng/embryo) struct (cabRESC) encoding the complete open reading
frame of cables1. Co-injection of cabRESC (15 pg/embryo) led to visible developmental defects, most notably affecting
anterior (neural) tissues. By 24 hpf, the most readily observ- into cabMO-injected embryos markedly attenuated the able phenotype was the appearance of cloudy (necrotic) brain phenotype; substantially fewer embryos (25/78, 32%) tissue in midbrain and hindbrain (Fig. 2B); in a representa- showed evidence of brain necrosis (Fig. 2C), while the
Figure 3. A–C: TUNEL staining to identify apoptotic cells in zebra？sh embryos at 18 hpf; (A) control-injected embryo; (B) cables1 morphant (cabMO-injected) embryo; (C) cables1 morphant embryo co-injected with cabRESCD) cables1 morphant co-injected with p53MO; ; ((E) quantitative estimation of TUNEL-positive cells in each embryo group; y-axis represents mean TUNEL index, as described in the Materials
and Methods Section. Different letters denote statistically signi？cant differences (P < 0.05); results of individual pair-wise comparisons are provided in the Results Section. Arrows in (B) indicate high density of TUNEL-positive cells.
24 Mol Reprod Dev 78:22–32 (2011)
CABLES1 AND ZEBRAFISH NEURAL DEVELOPMENT
mulation. Activity measurements were made over a 3-sec proportion of embryos resembling control-injected embryos
more than doubled (29/78; 37%, vs. 17% for cabMO- interval within the excitation phase, as previously described injected embryos). The remaining embryos showed mild (Kokel et al., 2010).
overexpression phenotypes (described below). Control-injected embryos exhibited a PMR (Fig. 4A;
Supplementary Movie 1) that is typical of equivalent
stage wild-type embryos at all stages examined (Kokel Increased Apoptosis in cables1 Morphants et al., 2010). In marked contrast, cables1 morphants
(Fig. 4B,D; Supplementary Movie 2) displayed a signi？cantly It was previously reported that cables1 can modulate
apoptosis, via both p53-dependent and p53-independent increased PMR relative to control-injected embryos pathways (Tsuji et al., 2002). To determine whether or not (P < 0.001).
the necrotic appearance of brain tissues in cables1 mor- To determine whether the aberrant PMR could be phants resulted from increased apoptosis, embryos were attributed to a loss of cables1 expression or not, PMR assayed for DNA fragmentation at 18 hpf using the terminal measurements were made in cables1 morphants co-injected dUTP nick-end labeling (TUNEL) assay. Relative to control- with mRNA encoding cabRESC. As shown (Fig. 4C,D), the
injected embryos (Fig. 3A), a signi？cant increase was mean PMR of cabMO-injected embryos co-injected observed in the mean number of TUNEL-positive cells with cabRESC was signi？cantly reduced relative to embryos (TUNEL index, see Materials and Methods Section) in injected with cabMO alone (P < 0.05), and did not differ
signi？cantly from the PMR of control-injected embryos cables1 morphants (Fig. 3B; P < 0.001), with the highest intensity of TUNEL-positive cells observed in regions (P > 0.05). Collectively, these data (Fig. 4D), strongly corresponding to areas of necrotic degeneration (Fig. 2B). suggest that the aberrant behavioral phenotype is We next examined whether or not the increased apopto- attributable to a loss of cables1 expression. sis could be rescued by co-expressing cabRESC. As shown We next sought to determine whether or not the aberrant (Fig. 3C,E), the number of TUNEL-positive cells in cabMO- behavioral phenotype in cables1 morphants was associat- injected embryos co-injected with cabRESC was signi？cantly ed with the observed increased in p53-dependent cell reduced relative to embryos injected with cabMO alone death, by measuring the PMR in cables1 morphants co-
injected with p53MO. Similar to the above experiment, (P < 0.001). The TUNEL index of cabRESC-injected embry-
os also did not differ signi？cantly from the mean TUNEL injection of cabMO (Fig. 4F) resulted in a signi？cant in-
index in control-injected embryos (P > 0.05), indicating a crease (P < 0.001) in the mean PMR relative to control- robust rescue of the phenotype. injected embryos (Fig. 4E). Co-injection with p53MO (Fig. To determine if the increased number of TUNEL-positive 4G) signi？cantly reduced the mean PMR (P < 0.05) relative cells in cabMO-injected embryos was mediated by a p53- to cabMO-injected embryos. However, the mean PMR dependent apoptotic pathway, endogenous p53 expression of p53MO co-injected embryos was still signi？cantly
was suppressed in cabMO-injected embryos by co-injec- higher (P < 0.01) than the mean PMR of control-injected tion with a MO targeting p53 (p53MO). Co-injection of embryos, suggesting that the behavioral defects in cables1
morphants are only partially attributable to an increase in p53MO (Fig. 3D) markedly rescued apoptosis in cabMO-
injected embryos, with signi？cantly fewer TUNEL-positive p53-dependent apoptosis. These data are summarized in cells detected relative to embryos injected with cabMO Figure 4H.
alone (P < 0.01). The mean TUNEL index of p53MO co-
injected embryos was similar to the mean TUNEL index of Neural Tube Patterning Defects in cables1 cables1 morphants co-injected with cabRESCP > 0.05), but (Morphants was still signi？cantly greater (P < 0.05) than the mean In efforts to narrow down the origins of the neural TUNEL index of control-injected embryos. These data
indicate a signi？cant, but incomplete, rescue of apoptosis defects in cables1 morphants, we examined the expression by suppressing p53 expression. of a selection of marker genes by in situ hybridization. To
ensure that p53-dependent cell death did not contribute to
any observed changes in marker gene expression, control Behavioral Abnormalities in cables1 Morphants and morphant embryos were co-injected with p53MO. A A distinct behavioral phenotype was subsequently minimum of 20 embryos (10 treatments and 10 controls) observed in cables1 morphants allowed to develop were examined for each marker gene. As shown (Fig. 5), to the hatching stage ( 48 hpf). At this stage of develop- distinct alterations in the expression of selected neural ment, wild-type embryos exhibit a rapid and short-lived transcription factors were observed in cables1 morphants. motor response (burst-swimming) following stimulation We consistently observed reductions in the expression of (Granato et al., 1996). In contrast to control-injected and emx3 (presumptive forebrain), eng2a (midbrain/hindbrain wild-type embryos, cables1 morphants exhibited a notably boundary), and gsc (prechordal plate, diencephalon), sustained motor response following stimulation (not but increased expression of otx2 (anterior neural plate). shown). To investigate this behavioral phenotype in greater In contrast, the expression patterns of pax2a (midbrain/ detail, we employed an automated behavioral assay hindbrain boundary) and ntl (axial mesoderm) were (Photomotor Response, PMR) that provides a quantitative unchanged among control-injected embryos and cables1 readout of behavior (movement) in response to photosti- morphants.
25 Mol Reprod Dev 78:22–32 (2011)
Molecular Reproduction & Development GROENEWEG ET AL.
Figure 4. Photomotor response (PMR) data demonstrating behavioral abnormalities in cables1 morphant embryos. A–C: Representative PMR
plots from (A) control-injected embryos, (B) cabMO-injected embryos, and (C) cabMO-injected embryos co-injected with cabRESC mRNA.
D: Quantitative summary of PMR data in (A–C). Values are means SEM of the Motion Index, corresponding to a 3 sec measurement interval following the light pulse (see Materials and Methods Section). Different letters denote statistically signi？cant differences (P < 0.05).
E,F: Representative PMR plots from (E) control-injected embryos, (F) cabMO-injected embryos, and (G) cabMO-injected embryos co-injected
with p53MO; (H) quantitative summary data in (E–G). Different letters denote statistically signi？cant differences (P < 0.05); results of individual pair-wise comparisons are provided in the Results Section. Arrows in PMR plots indicate timing of 1 sec light pulse; horizontal bar above each PMR plot indicates time interval used for quanti？cation.
Cyclin Box Confers Biological Activity to Cables1 p35 Expression Is Reduced in cables1 Morphants To further investigate the in vivo biological activity of The potent biological activity of the cables1 cyclin box cables1, we performed ectopic overexpression experi- suggests that interactions with cyclin-related proteins are ments in zebra？sh embryos. Injection of synthetic mRNA an important aspect of cables1 function. The neural defects (100 pg/embryo) encoding full-length Cables1 protein in cables1 morphants are further suggestive of an interac- potently disrupted embryonic morphogenesis (Fig. 6). In tion with Cdk5, with which cables1 was ？rst shown to
contrast to embryos injected with control mRNA (GFP, interact in a yeast two-hybrid screen (Zukerberg et al., 100 pg/embryo; Fig. 6B), of which only 4/48 (8%) showed 2000). To determine if a loss of cables1 affects the Cdk5/ mild phenotypic defects, embryos injected with full-length p35 kinase pathway in vivo, we used Western immunoblot cables1 mRNA had a very high frequency (39/45, 87%) of analysis to examine protein levels of both Cdk5 and p35 in morphological defects. These ranged from mild/moderate whole embryo lysates after cables1 knockdown. No signi？-
axis defects (Fig. 6C,D) to complete dysmorphogenesis cant differences were detected in Cdk5 protein levels be- (Fig. 6E). tween control-injected embryos and cables1 morphants The region of the Cables1 protein most highly conserved (Fig. 7); by contrast, levels of p35 in cables1 morphants among species is the C-terminal cyclin box (CB; amino acid showed a signi？cant ( 65%) reduction relative to control- residues 413–501 of zebra？sh Cables1), which is pre- injected embryos (P < 0.01).
sumed to mediate its interactions with cyclin-related pro-
teins. To determine whether or not this region confers Cell Proliferation Is Unaffected in cables1 biological activity to zebra？sh Cables1, we microinjected
Morphants synthetic mRNA (100 pg/embryo) encoding a truncation
Previous studies have suggested an inhibitory function mutant of Cables1 protein (DCB) that lacks the cyclin box.
for cables1 in epithelial cell proliferation (Wu et al., 2001; Overexpression of DCB in zebra？sh embryos resulted in a Kirley et al., 2005b). In our knockdown studies, cables1 distribution of phenotypes that differed markedly from em-
morphants progressed through early stages of embryonic bryos-injected with full-length cables1 mRNA (Fig. 6F); 73%
(33/45) of embryos developed completely normally, 20% development (blastula, epiboly, gastrulation) at a similar (9/45) had very mild phenotypic defects, and only 4% (2/45) rate as control-injected embryos, as evidenced by their showed severe effects. simultaneous onset of somitogenesis (Fig. 8A,B). While
26 Mol Reprod Dev 78:22–32 (2011)
CABLES1 AND ZEBRAFISH NEURAL DEVELOPMENT
this observation suggests that cell cycle progression is
not affected in cables1 morphants, we opted to examine cell proliferation dynamics in more detail during early development. The number of mitotic cells was quanti？ed in control-injected and cables1 morphants at the shield stage (6 hpf), using immunohistochemical staining for phosphorylated histone H3 (PH3) as a marker of prolifer- ating cells (Fig. 8C,D). No signi？cant differences were detected in the number of PH3-positive cells between control-injected and cables1 morphants (Fig. 8E; P > 0.05).
DISCUSSION
Using the zebra？sh as an experimental model, this study provides molecular, cellular, biochemical, and behavioral evidence that suggest an important functional role for cables1 in embryonic neural development. Knockdown of cables1 in zebra？sh embryos led to increased apoptosis in embryonic neural tissues, enhanced neural excitability in response to stimulation, signi？cant reductions in p35 protein levels, and alterations in the expression of neural transcription factors during early neurulation. It is important to note that not all transcription factors examined were affected equally. We observed consistent decreases in emx3, eng2a, and gsc expression in cables1 morphants, whereas otx2 expression was consistently increased. However, no changes were detected in the expression patterns of pax2a, or the axial mesoderm mark- er ntl. These observations suggest an underlying deregu-
lation of neural differentiation pathways, rather than simply
global shifts in gene expression. The down-regulated
expression of some transcription factors also cannot be
explained by increased p53-dependent cell death during
differentiation, as these observations were made in
embryos in which p53-dependent apoptosis had been sup- pressed. Taken together, these data suggest that cables1 is functionally important during neural tube differentiation, such that a lack of cables1 expression results in neural tube patterning defects. These ？ndings are in agreement with in vitro studies in the chick model, where Cables1 was shown to link the Slit/Robo complex to the N-cadherin pathway, leading to b-catenin phosphorylation, nuclear translocation and increased transcription of genes involved in CNS patterning (Rhee et al., 2007). The behavioral response to a loss of cables1 is interest- ing, although the underlying mechanism for this is not entirely clear. The rescue experiments suggest that this phenotype is linked partly to an increase in p53-dependent cell death, but the precise mechanism underlying the Figure 5. In situ hybridization analyses showing marker gene mRNA behavioral phenotype remains unclear. Notably, hyper- expression in representative control-injected embryos (left column) excitability is not observed in all cases of excess cell death and cables1 morphants (right column). Reduced expression of emx3 in the embryonic CNS. For example, knockdown of an (A, B), eng2a (C, D), and gsc (E, F), and increased expression of otx2 insulin-like growth factor receptor (Igf1rb) led to widespread (G, H) was observed in cables1 morphants. No changes were detected in the expression of pax2a (I, J) or the axial mesoderm neural apoptosis in the zebra？sh embryo, but a distinct marker ntl (K, L). Results shown are representative images from a loss of motor activity (Schlueter et al., 2006). It is noteworthy minimum of 10 embryos examined per treatment group. that the aberrant behavior was only partially rescued by suppressing p53 expression, whereas a complete rescue was achieved by co-expressing full-length cables1. In
27 Mol Reprod Dev 78:22–32 (2011)
Molecular Reproduction & Development GROENEWEG ET AL.
Figure 6. Phenotypes of (A) wild-type, (B) control-injected (GFP, 100 pg/embryo), and (C–E) treatment
embryos, injected with synthetic mRNA encoding full-length zebra？sh Cables1 protein. cables1mRNA- injected embryos exhibited a range of overexpression phenotypes, ranging from mild (C) or moderate (D) axis defects, to severe dysmorphogenesis (E); (F) quantitative summary of embryo phenotype propor- tions, pooled from two independent experiments. WT, wild-type (uninjected) embryos; GFP, embryos ORFinjected with synthetic mRNA encoding GFP; cab, embryos injected with synthetic mRNA encoding
the complete open reading frame (ORF) of cables1; DCB, embryos injected with synthetic mRNA
encoding a truncation mutant of cables1 that lacks the C-terminal cyclin box. OEP; overexpression phenotype.
mammals, Cables1 has been reported to influence both et al., 2000). Unlike related cyclin-dependent kinases, p53-dependent and p73-dependent cell death pathways, Cdk5 is not activated by cyclins and does not function in functions that were attributed to distinct functional domains cell cycle regulation. Rather, its kinase activity is con？ned to
within the Cables1 protein (Tsuji et al., 2002). Further postmitotic (differentiating) neurons, via interactions with studies are required to determine whether or not the p35 and p39, that exhibit neural-speci？c expression
zebra？sh ortholog of Cables1 also contains discrete (Ohshima et al., 1996; Ko et al., 2001). p35 is a highly labile domains with distinct functional properties. protein that is rapidly degraded by the ubiquitin-proteasome The signi？cant reduction in p35 protein levels in cables1 pathway (Patrick et al., 1998; Kerokoski et al., 2002; morphant embryos is consistent with previous in vitro data Saito et al., 2003), and this degradation is dependent upon suggesting that cables1 regulates neural differentiation via phosphorylation by Cdk5 (Kerokoski et al., 2002; Saito interactions with the Cdk5/p35 kinase pathway (Zukerberg et al., 2003). The marked reduction of p35 protein levels
28 Mol Reprod Dev 78:22–32 (2011)
CABLES1 AND ZEBRAFISH NEURAL DEVELOPMENT Figure 7. Western immunoblot analyses of pan-Actin, Cdk5 and p35 protein levels in de-yolked embryo lysates from cables1 morphant embryos at 24 hpf. Inset shows representative Western immunoblot; plot shows protein levels, expressed as a ratio of intensities (measured by densitometry) between control-injected and cables1 morphant groups. Values shown are the means ( SEM) of three experimental replicates. p35 protein levels were signi？cantly lower (**) in cables1 morphants relative to control-injected embryos (P < 0.01).
in cables1 morphants is therefore suggestive that cables1 may function to stabilize p35, similar to its reported function in stabilizing p63 (Wang et al., 2010). The results of overexpressing a C-terminal truncation mutant lacking the cyclin box suggest that the cyclin box confers potent biological activity to cables1. This observation is consistent with the model involving interac- tions between cables1 and Cdk5/p35, as the cyclin box is the domain predicted to interact with the Cdk5/p35 complex. While deletion of the cyclin box substantially reduced Figure 8. A: Control-injected, and (B) cables1 morphant embryos at the activity of cables1 overexpressed in vivo, this could 12 hpf; images demonstrate simultaneous onset of somitogenesis potentially be explained by precocious degradation (e.g., (arrows) among embryos, indicating no gross disruptions to early nonsense-mediated decay) of the DCB mRNA (Wen and development, or notable differences in developmental rate; (C) control Brogna, 2008). However, our data are consistent with in -injected, and (D) cables1 morphant embryos (shield stage, 6 hpf) vitro studies showing that the cyclin box is essential to immunohistochemically stained for phosphorylated histone H3 (PH3) to identify proliferative cells (brown dots); (E) quantitative mediate important functions of cables1, including p53- representation of the number of PH3-positive cells in shield stage dependent cell death (Tsuji et al., 2002). embryos; values represent the mean density ( SEM) of PH3-positive Lastly, the lack of a signi？cant difference in the number cells in 6–8 random ？elds per embryo. No signi？cant differences were of phosphorylated histone H3-positive cells among detected in the mean densities of PH3-positive cells between control- injected and cables1 morphant embryos at this stage of development cables1 morphants and controls suggests that during early
embryogenesis, cables1 does not directly regulate cell (P > 0.05).
proliferation. This differs from what has been reported in
other biological contexts. For example, CABLES1 over-
may be due in part to its interactions with the Cdk5/p35 expression in human endometrial cells markedly slowed
the rate of cell proliferation (Zukerberg et al., 2004), while kinase pathway during neural differentiation. Future studies /embryonic ？broblast from Cables1mice proliferated using conditional gene ablation approaches are required to more rapidly than wild-type cells (Kirley et al., 2005b). It determine if cables1 adopts other biological functions after is unclear whether these differences reflect species- the embryonic period. The mounting evidence of a role for speci？c variations in gene function, or reflect the different this gene in early development and possibly tumorigenesis experimental or biological contexts in which the genes were (Dong et al., 2003; Zukerberg et al., 2004; Zhang et al., studied. 2005; Kirley et al., 2005a,b; Lee et al., 2007; Park do et al., In summary, this study provides in vivo evidence that 2007; Sakamoto et al., 2008) suggests that much remains cables1 is critically important for neural development in the to be learned about its functions in humans and other zebra？sh embryo, and suggests that the observed defects vertebrates.
29 Mol Reprod Dev 78:22–32 (2011)
Molecular Reproduction & Development ROENEWEG ET AL. G
morpholino microinjection. A commercially available MO MATERIALS AND METHODS
targeting p53 (p53MO) has been previously described Experimental Animals (Robu et al., 2007). Wild-type zebra？sh (T?ubingen AB strain) were main- tained in standard conditions, and staged as described
Synthetic mRNA Overexpression (Kimmel et al., 1995). Embryos were generated by natural
crosses, and reared at 28 1 C in embryo medium. Three expression constructs were made to overexpress ORF variants of cables1 in zebra？sh embryos. pCS2 þ cab
corresponded to the complete open reading frame (525 Chemicals and Reagents
amino acids) of zebra？sh cables1. pCS2 þ cabRESC also Reagents for RT-PCR (Superscript II reverse transcrip- encoded the complete open reading frame, but excluded tase, Taq DNA polymerase, dNTPs) were purchased from the morpholino (cabMO) target region. pCS2 þ DCB en- Invitrogen (Carlsbad, CA). FailSafe PCR buffers were coded a truncated variant of cables1; this was generated by purchased from Epicenter Biotechnologies, and antibodies RT-PCR using primers that introduced a stop codon (TAA) were purchased from Santa Cruz Biotechnology, Inc. at amino acid residue 407, thereby encoding a 406 amino (Santa Cruz, CA; rabbit polyclonal anti-Cdk5 [C-8]; mouse acid protein (DCB) lacking the C-terminal cyclin box. monoclonal anti-p35), ThermoFisher Scienti？c (Waltham, Each fragment was ampli？ed by RT-PCR using total MA; mouse monoclonal anti-Actin, pan Ab-5) and Millipore RNA puri？ed from adult brain tissue as template. The PCR (Billerica, MA; rabbit anti-phosphorylated histone H3, products were subcloned into pCRII-TOPO (Invitrogen), #06.570). Morpholino oligonucleotides were purchased sequenced to con？rm accuracy, digested with EcoRI and from Gene Tools LLC (Philomath, OR). TUNEL analyses ligated into pCS2þ. After identifying clones containing the were performed using the In situ Cell Death Detection Kit appropriate inserts, they were sequenced to ensure accu- (Roche Applied Science, Indianapolis, IN). All other racy and used for synthetic mRNA synthesis, as described reagents were from Sigma, St. Louis, MO or ThermoFisher (Turner and Weintraub, 1994). unless otherwise stated. Synthetic mRNA was generated using the SP6 mES- SAGE mACHINE (Sang et al., 2008), precipitated in LiCl, Reverse-Transcriptase PCR (RT-PCR) washed extensively in nuclease-free 70% ethanol and Total RNA was puri？ed from whole embryos at desired resuspended at the desired concentration in sterile 1 stages using Tri Reagent (20 embryos/ml), resuspended in Danieau microinjection buffer. nuclease-free water and quanti？ed by absorbance at 260 nm. First-stand synthesis was performed as previously Embryo Microinjection described (Sang et al., 2008). Briefly, 500 ng of total RNA Approximately 1 nl of each solution (morpholino, syn- was reverse-transcribed using Superscript III reverse- thetic mRNA) was microinjected into 1-cell stage zebra？sh transcriptase, according to the manufacturer’s protocol. embryos, as previously described (Sang et al., 2008). Aliquots of ？rst-strand cDNA were then PCR-ampli？ed, Microinjection volumes were calibrated using the oil drop using the following gene-speci？c primers: cables1-F2: method (Yuan and Sun, 2009). For rescue experiments, 50-TGCTCTCTCATTCCTCAGCA-30cables1-R4: 50 ;-synthetic mRNA (cabRESC) was simultaneously co-injected GGCAGGATTCCCTGAGTGTA-30. This primer pair ampli- with cabMO (2 ng/embryo). Synthetic mRNA encoding GFP ？ed a 345 base-pair fragment corresponding to nucleotides (100 pg/embryo) was injected into a subset of embryos to 81–425 (ATG start codon designated as position 1). PCR con？rm the quality of in vitro-derived synthetic mRNA. conditions were as follows: 94 C for 5 min; 30– 35 cycles of 94 C 1 min, 58 C 30 sec, and 72 C TUNEL Analyses 30 sec; 72 C 5 min. Each reaction (25 ml) contained 2 ml
of ？rst-strand cDNA, 0.8 mM of each primer and 100 U of TUNEL analyses for DNA fragmentation were performed Platinum Taq polymerase, diluted in 1 FailSafe PCR on paraformaldehyde-？xed whole embryos, as previously Buffer D. PCR products were electrophoretically separated described (Schlueter et al., 2007). For quanti？cation pur-
in 1.5% agarose gels, and products were visualized under poses, a mean TUNEL index was generated by counting the UV light. Ampli？cation of ornithine decarboxylase 1 (odc1) number TUNEL-positive cells in ？ve random-selected, non- served as an internal control for RNA and cDNA quality, overlapping ？elds of view per embryo. This method of using the following primer pair: odc-F1: 50-CAGAA- quanti？cation was replicated in each of ？ve embryos per
treatment group. The observer was blinded with respect to GACGCTCAACCAAACC-30; odc-R1: 50-ATCCCATCTC-
the identity of each embryo during quanti？cation. TTTCACGTCC-30.
In Situ Hybridization Morpholino Oligo-Mediated Gene Knockdown
The MO targeted against zebra？sh cables1 (cabMO: Embryos were ？xed in buffered paraformaldehyde (4%)
and stored in methanol for a minimum of 24 hr prior 50-AAGTCAGCGGAGACTAAAGGTGTTC-30) is comple-
to analysis. The methods used for in situ hybridization mentary to a 25-mer region in the 50-UTR ( 52 to 28 relative analyses have been previously described (White et al., to ATG codon); a standard control MO designed by the
manufacturer was used to control for the effects of 2009). Plasmid templates for target genes were generously
30 Mol Reprod Dev 78:22–32 (2011)
CABLES1 AND ZEBRAFISH NEURAL DEVELOPMENT
Data Analyses provided by Iain Drummond (Department of Nephrology,
Massachusetts General Hospital, Charlestown, MA), and Data are presented as mean values standard error of Igor Dawid (NIH, Bethesda, MD). Cryosectioning was the mean (SEM). Mean values for TUNEL, PMR, and performed as previously described (Wood et al., 2005). densitometry data were statistically compared by analysis
of variance (ANOVA), followed by the Student–
Newman–Keuls multiple comparison test. An unpaired Immunohistochemistry Student’s t-test was used to compare mean values for Immunohistochemical staining was performed on para- phosphorylated histone H3 staining. Differences between formaldehyde-？xed embryos after dehydration and storage mean values were considered statistically signi？cant if ( 30 C) in methanol. Embryos were rehydrated by serial P < 0.05. washes in phosphate-buffered saline containing 0.1% Tween (PBST), incubated in blocking buffer (PBST contain-
ing 2% serum, 1% DMSO, 1% BSA), and incubated in ACKNOWLEDGMENTS primary antibody solution overnight at 4 C. Embryos were This study was supported by Vincent Memorial then were washed extensively in PBST, and incubated
Research Funds, the Advanced Medical Research Foun- overnight (4 C) in biotinylated secondary antibody diluted
dation, and US National Institutes of Health grants (1:500) in blocking buffer. After washing in PBST, embryos
RO1CA098333 (B.R.R.), R01MH086867 (R.T.P.) and were incubated in Vectastain ABC reagents for 30 min at
R21MH085205 (R.T.P.). RT, washed in PBST and incubated in DAB solution
(DAKO) to complete the color reaction.
REFERENCES Western Immunoblot Analyses A total of 20 embryos from each sample group were DeBernardo RL, Littell RD, Luo H, Duska LR, Oliva E, Kirley SD, 2suspended in Caþ2þ-free Hank’s balanced salt solu- /MgLynch MP, Zukerberg LR, Rueda BR. 2005. De？ning the extent tion (HBSS), deyolked by passage through a narrow-bore of cables loss in endometrial cancer subtypes and its effective- pipette, and washed brie，y in fresh HBSS. The carcasses ness as an inhibitor of cell proliferation in malignant endometrial were then reconstituted in SDS sample buffer (5 ml/embryo) cells in vitro and in vivo. Cancer Biol Ther 4:103–107. and homogenized for 1 min in a Mini-Beadbeater-8 Dong Q, Kirley S, Rueda B, Zhao C, Zukerberg L, Oliva E. 2003. (Biospec Products, Inc., Bartlesville, OK), using 1.0 mm Loss of CABLES, a novel gene on chromosome 18q, in ovarian glass beads. The homogenates were then heated for 15 min cancer. Mod Pathol 16:863–868. at 75 C, centrifuged for 1 min at 14,000g, and aliquots of the soluble fraction (20 ml homogenate/lane) were electropho- Granato M, van Eeden FJ, Schach U, Trowe T, Brand M, Furutani- retically separated in 4–12% Bis–Tris polyacrylamide gels, Seiki M, Haffter P, Hammerschmidt M, Heisenberg CP, Jiang with MOPS running buffer. Proteins were transferred to YJ, Kane DA, Kelsh RN, Mullins MC, Odenthal J, Nusslein- PVDF membrane, washed brie，y in Tris-buffered saline Volhard C. 1996. Genes controlling and mediating locomotion containing 0.1% Tween-20 (TBST), and blocked for a behavior of the zebra？sh embryo and larva. Development minimum of 2 hr at room temperature (RT) in blocking buffer 123:399–413. (10% non-fat dry milk in TBST). Primary antibody was Kerokoski P, Suuronen T, Salminen A, Soininen H, Pirttila T. 2002. diluted (1:1,000) in blocking buffer and applied overnight Influence of phosphorylation of p35, an activator of cyclin- (4 C) with gentle agitation. After extensive washes in TBST, dependent kinase 5 (cdk5), on the proteolysis of p35. Brain Res membranes were incubated in HRP-conjugated secondary Mol Brain Res 106:50–56. antibody solution (diluted 1:5,000 in blocking buffer) for 1 hr at RT, and target proteins were detected using the Kimmel CB, Ballard WH, Kimmel SR, Ullmann B, Schilling TF. Enhanced Chemiluminescence system (ECL, GE Health- 1995. Stages of embryonic development of the zebra？sh. Dev care, Milwaukee, WI). Proteins were visualized by exposure Dyn 203:253–310. to X-ray ？lm and quanti？ed by densitometry on a FluorChem Kirley SD, D’Apuzzo M, Lauwers GY, Graeme-Cook F, Chung DC, ChemiImager 5500. Zukerberg LR. 2005a. The CABLES gene on chromosome 18Q regulates colon cancer progression in vivo. Cancer Biol Ther Photomotor Response (PMR) 4:861–863.
PMR measurements were conducted in a 96-well plate Kirley SD, Rueda BR, Chung DC, Zukerberg LR. 2005b. format, using 8–10 embryos per well, similar to methods Increased growth rate, delayed senescence and decreased previously described (Kokel et al., 2010). PMR measure- serum dependence characterize cables-de？cient cells. Cancer ments were performed on embryos at 40–42 hpf. Reported Biol Ther 4:654–658. PMR values correspond to the 3rd quartile of the Motion Ko J, Humbert S, Bronson RT, Takahashi S, Kulkarni AB, Li E, Tsai Index, for a measurement interval lasting 3 sec after the light
LH. 2001. p35 and p39 are essential for cyclin-dependent kinase pulse. Mean PMR data represent the average of three
replicate measurements, with each replicate derived from 5 function during neurodevelopment. J Neurosci 21:6758–
6771. a separate group of 8–10 embryos.
31 Mol Reprod Dev 78:22–32 (2011)