Author Correction: The Reprimo gene family member, reprimo-like (rprml), is required for blood development in embryonic zebrafish

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

www.nature.com/scientificreports www.nature.com/scientificreports/ somite-derived endothelial cells 6 , arise in the VDA in a region also known as the aorta-gonad-mesonephros (AGM) and characteristically express cmyb and/or runx1 after 26 hpf 7,8 . The HSPCs emerge from the hemogenic endothelium in the floor of the AGM or VDA, through a process termed endothelial hematopoietic transition (EHT). Previous studies using CD41:GFP low transgenic animals have defined the migration of HSPC from the AGM to their niche in the caudal hematopoietic tissue (CHT) and then to thymus and pronephros 3,9 . Recent evidence by our group has reported the expression pattern of the RPRM gene family members during blood vessel development 10 .
The Reprimo (RPRM) gene family is a group of single-exon genes exclusive to the vertebrate group, that diversified as a product of two rounds of whole genome duplications occurred in the last common ancestor of vertebrates between 676 and 615 million years ago [10][11][12] . The repertoire of Reprimo genes includes Reprimo (RPRM), Reprimo-like (RPRML) and Reprimo 3 (RPRM3). To date, only the human RPRM gene has been studied 10,12 , with no literature describing the role of RPRML or RPRM3 in either physiology or pathophysiology. In humans, RPRM is a highly glycosylated cytoplasmic protein, which has been characterized as a potential tumor-suppressor gene involved in the regulation of the p53-mediated cell cycle arrest at G2 through the regulation of the Cdc2-cyclin B1 complex activity by a yet undiscovered mechanism 13 . Zebrafish (Danio rerio) have two co-orthologs of the mammalian RPRM gene (designated rprma, rprmb) and a single co-ortholog of RPRML (rprml) and RPRM3 (rprm3), all of which are expressed during normal embryogenesis 10,14 . Gene expression profiling of the rprm genes suggested that during their evolutionary history, they have undergone a process of subfunctionalization with some degree of redundancy, as they exhibit uniquealthough partially overlapping -expression profiles during embryonic and larval blood vessel development. Interestingly, epigenetic silencing of RPRM has been linked to human blood cancers such as pediatric myeloid leukemia 15 . These studies suggest that RPRM genes may be involved in the formation and differentiation of the blood lineages in a process known as hematopoiesis.
Herein, we characterize the functions of rprml during hematovascular development in zebrafish. By loss-of-function assays either using generated knockdown CRISPR mutants (G0) or morphant embryos, we demonstrate that rprml is required during both waves of definitive hematopoiesis for the specification of both EMP and HSPC. Furthermore, we report that rprml regulates the normal formation of the HSPC perivascular PCVP/CHT niche. To our knowledge, this is the first report describing a physiological role for rprml.

Results
rprml loss-of-function using MO and CRISPR-Cas9 systems. We previously reported distinctive RNA expression pattern profiles for RPRM gene family members 10,14 . We also showed that rprml embryonic expression domains include vascular and mesodermal-derived tissues. Here, to assess a possible function of rprml during blood development, we used a combination of antisense oligonucleotide morpholinos (MOs) 16 and CRISPR-Cas9 17 genetic approaches in a transgenic Tg(fli1:GFP) background, which enabled us to visualize endothelial cells (EC) by confocal microscopy. After comparing uninjected WT embryos with MO-injected and mutant specimens, where we observed no evidence of developmental delay generated by the injection (Fig. S1), we compared the phenotypes between the embryos injected with MOs targeting rprml and the CRISPR-Cas9 G0 rprml mutants 17 . The observed phenotypes were comparable in both experiments ( Fig. S2A-C), characterized by reduced CHT territory and mild alterations in vascular morphology at the PCVP, which did not affect normal blood flow. Additionally, macroscopic observations of the different specimens showed no apparent decrease in the length of the trunk (Fig. S1). Abnormal blood accumulation in yolk, trunk edema and or hemorrhage were absent in CRISPR-Cas9 and morphant specimens. Cranial vasculature was also normal in controls compared with morphant and mutant embryos (Fig. S1).
In accordance with the recent guidelines and standard use of MO in zebrafish embryos 18 , we assessed the expression of p53 19,20 by qRT-PCR in MO-injected and Tab5 WT uninjected embryos. This control was performed to rule out activation of p53 and associated cell death in MO-injected embryos. We observed a moderate decrease in p53 expression levels in a pool of rprml-MO injected specimens compared with the control group (Fig. S2G), supporting that this MO dosage was not toxic to the embryos. To assess the effectiveness of the rprml MO knockdown approach, we performed immunohistochemical analysis of Rprml protein. We observed a significant reduction of Rprml in rprml-MO embryos when compared with controls ( Fig. S2D,E). Additionally, we performed phenotypic rescue experiments to assess the specificity of the MO. Co-injection of rprml mRNA and rprml-MO renders significantly rescued rprml-MO phenotype (Fig. S5), indicating that the rprml-MO effects over hematopoiesis are specific. Taken together, these results and the comparable phenotypes obtained by antisense MO and CRISPR-Cas9 genetic approaches, validate the use of MOs for the assessment of the physiological role of rprml during zebrafish embryogenesis by gene loss-of-function. rprml is dispensable for primitive hematopoiesis but required for HspC development. At early developmental stages, hemangioblasts contain the common precursor cells for hematopoietic and vascular lineages at the anterior and posterior lateral plate mesoderm (ALM and PLM) 1,2 . We used whole mount in situ hybridization (WISH) to determine the expression of well-known hemangioblast molecular markers in control and morphant embryos at 5-10 somite-stage, including lmo2, fli1a and tal1/scl1. We found comparable expression levels of these markers, suggesting that rprml does not play a relevant role during cell specification in primitive hematopoiesis (Fig. 1A-F). By 22 hpf, expression of tal1/scl1 (a primitive erythroid marker) remained mostly unaffected at the intermediate cell mass (ICM) (Fig. 1G,H). At around 24 hpf, we did not detect significant changes in the expression of lmo2 in rprml morphants compared with control embryos (Fig. S3A,B). However, we did observe a moderate reduction of fli1a in the PCVP at the CHT (Fig. S3C,D). In contrast, after 26 hpf, abrogation of rprml function led to a drastic reduction in the cmyb + HSPCs population at the ICM and www.nature.com/scientificreports www.nature.com/scientificreports/ the posterior blood island (PBI) (Fig. 1I,J,I′,J′ in detail). At the same developmental stage, the expression of cmyb remained unaltered in neuronal territories such as the retina in rprml morphants (Fig. 1I,J, arrowheads). Therefore, loss-of-function of rprml does not alter hemangioblast in the ALM/PLM or primitive hematopoiesis but is likely required for HSPCs formation. www.nature.com/scientificreports www.nature.com/scientificreports/ rprml is required for transient-definitive wave of hematopoiesis. As stated previously, definitive hematopoiesis occurs in two successive waves during zebrafish development 1 . The first transient-definitive wave involves the formation of erythroid-myeloid progenitors (EMP) at the PBI ( Fig. 2A,B). The second definitive wave produces HSPCs from the hemogenic endothelium located in the AGM. Compared with HSPCs, EMP in zebrafish and mouse lack capacity for self-renewal 5 . To assess the possible functional role of rprml during transient-definitive hematopoiesis, we analyzed the expression pattern of the erythroid progenitor marker tal1/ scl1 by using WISH. We noted a significant reduction of gene expression at the PBI in rprml kd (Fig. 2C,D), while normal expression was retained in the brain (compared Fig. 2C,D, arrowheads). Furthermore, expression of the myeloid progenitor marker pu.1 was also reduced at the PBI of rprml kd compared with control embryos (Fig. 2E,F). Of note, pu.1 expression was unaffected in the primitive myeloid precursors of the rostral ICM (compared Fig. 2E,F, arrowheads). We also investigated the development of granulocytic cells in rprml kd . To this end, we analyzed the expression of the neutrophil marker mpx and noted reduced expression in rprml kd when compared to control embryos (Fig. 2G,H). Also, the number of neutrophils was drastically reduced in the CHT of transgenic Tg(mpx:GFP) embryos between 32 and 48 hpf ( Fig. 2I-J′), consistent with the decreased mpx expression visualized by WISH. Altogether, these results indicate that rprml is required for EMP formation during the transient wave of definitive hematopoiesis.
Loss of rprml hinders HSPC specification. As mentioned above, zebrafish rprml kd animals exhibited normal primitive hematopoiesis including specification of erythroid and myeloid cells ( Fig. 1A-F). During definitive hematopoiesis, at 28 hpf, the cmyb + HSPC population was greatly reduced in rprml kd when compared with control embryos (Fig. 1I,J). Emergence of HSPCs from the ventral DA (VDA) has been characterized by the expression of CD41:GFP low in the AGM (Fig. 3A). To visualize HSPC dynamics, we used time-lapse confocal imaging in double transgenic Tg(flt0.8:mcherry;CD41:GFP) and single Tg(CD41:GFP) embryos (Fig. 3). Consistent with our previous results, zebrafish rprml kd showed a drastic reduction of CD41:GFP low+ HSPCs budding derived from the VDA (yellow arrows, Fig. 3B,C,N), a result that was also confirmed in CRISPR-Cas9-rprml mutants (Fig. 3D,E). Thus, the emergence of HSPCs from the AGM/VDA seems to be strongly dependent on rprml activity. Consequently, initial HSPCs migration to the CHT was substantially reduced at 54 hpf (Figs 3F,G,N and S4). This reduction in HSPCs number at the CHT persisted later in development, as the rprml morphants showed greatly reduced CD41:GFP low+ cells compared with control embryos at 84 hpf (Fig. 3H,I,N). In parallel, the dorsal-ventral width of the CHT was drastically reduced after 48 hpf. Additionally, in rprml-MO injected Tg(CD41:GFP) embryos, CD41:GFP low+ lymphocyte progenitors failed to colonize the thymus (Fig. 3J,K,N). Furthermore, rag2 + lymphocytes, which depend on upstream HSPCs 21 , were almost completely absent in rprml morphants (Fig. 3L,M). Finally, the number of CD41 + -GFP and mpx + -GFP expressing cells was significantly restored by co-injection of zebrafish rprml mRNA in rprml-morphants, indicating that the effects over hematopoiesis by rprml-MO injections are specific (Fig. S5).
There is a close relationship between HSPC formation and the aortic endothelium 7,8 . To characterize arterial development, we performed WISH using several early arterial-specific markers. We examined the expression of two typical zebrafish artery markers, kdr1 (also known as vegfr-2) and dll4 (Notch ligand). Our results show that the expression of both markers did not drastically change in morphants when compared with control embryos (Fig. S6A-D). Consistently, the overall expression for transgenic arterial markers was normal in rprml-MO injected embryos with Tg(dll4:GFP) and/or Tg(flt0.8:RFP) background ( Fig. S6E-I). Given that Notch activation is crucial for HSC specification 22 , we examined the spatiotemporal pattern of Notch signaling in the VDA using the Tg(Tp1:GFP) transgenic zebrafish line 23 . This reporter line is characterized by a fluorescent red signal in cells which respond to Notch signal activation. In these assays, we detected fewer Notch + cells in the VDA in both rprml-morphant and mutant embryos ( Fig. S6J-L). This indicates that VDA-cells are not actively responding to Notch signaling, thus affecting HSC specification from the VDA. Collectively, these results suggest that rprml is required for definitive HSPCs specification at the VDA in the hemogenic endothelium. rprml is required for the formation of the CHt. To explore if rprml has a function during HSPC niche formation, we studied CHT morphogenesis in rprml kd double transgenic Tg(flk:mcherry;fli1:eGFP) zebrafish embryos. rprml kd showed reduced thickness of the HSPC niche territory compared with wild type siblings ( Fig. 4 and S7). Specifically, rprml kd embryos displayed increased intervascular spaces at the posterior caudal vein plexus (PCVP; the network of vessels adjacent to the CHT; Fig. 4B,C,J-K and S7). To determine whether the cellular defects of the CHT in rprml kd were due to increased apoptosis, we monitored cell death by using activated caspase-3 (Casp-3) immunostaining. Surprisingly, PCVP/CHT apoptosis in rprml kd animals was increased (Fig. 4E,F,H,I). The significant difference in apoptosis between knockdown embryos and wild type/control siblings by 48 hpf, indicates that rprml is required for cell survival during the formation of the CHT. This finding partly explains the reduction in the HSPC niche by rprml loss-of-function. Alternatively, lack of rprml function may inhibit proper sprouting angiogenesis in the PCVP, thus hindering the establishment of proper sprouting connectivity by ECs at the PCVP/CHT. These mechanisms need to be further explored in the future.

Discussion
Our previous studies found that RPRM family genes are expressed in the vasculature in zebrafish and humans, specifically in vascular smooth muscle cells and ECs 10 . Although a potential role of RPRML in human hematopoiesis has not been yet assessed, the recent association between the loss of RPRM and highly aggressive acute myeloid leukemia (AML) 15 raises the question of whether products from the RPRM gene family play a relevant role during hematopoiesis. Here we provide the first report of a functional role of rprml in physiology, specifically during hematovascular development. Our results suggest that rprml is dispensable for primitive hematopoiesis but essential for the definitive wave of hematopoiesis. In absence of rprml activity, key hemangioblast genes such www.nature.com/scientificreports www.nature.com/scientificreports/ www.nature.com/scientificreports www.nature.com/scientificreports/ as lmo2, fli1a and tal1/scl1 are normally expressed, indicating that rprml does not participate at early stages of hemangioblast specification in the ALM and PLM. In contrast, cmyb + HSPC expression is strongly reduced in the VDA at the PBI (Fig. 1G-J) suggesting that rprml is essential for the activation of HSPC production. Together, these findings strongly suggest that rprml is not required for the initiation of hematopoiesis such as the primitive wave in the ICM but essential for definitive hematopoiesis along the VDA and PBI.
Previous studies have reported that MOs might present off-target RNA interactions at high concentrations. Accordingly, we followed the Stainier et al. guidelines 18 for the use of MO in zebrafish: (i) we used both MOs and CRISPR-Cas9 gene-knockdown methods, (ii) we performed p53 qRT-PCR in rprml MO-injected embryos to discard potential toxicity 19,20 , and (iii) we performed rescue experiments to assess the specificity of the MOs. Notably, recent evidence has shown that MOs can reveal gene functions concealed by mutants 20 . Indeed, stable mutations induced by CRISPR-Cas9 might lead to gene compensation by up-regulation of other genes in the same signaling pathway. In contrast, MOs do not trigger as much compensation, therefore producing more extreme phenotypes than their mutant counterparts 20 . This allows for probing of gene function that is not possible using mutants alone. Our results indicate that our rprml MOs do not produce considerable off-target effects and the observed phenotype between morphants and CRISPR-Cas9 G0 genetic approaches are comparable. Therefore, gene knockdown by MOs is suitable to assess rprml gene function. www.nature.com/scientificreports www.nature.com/scientificreports/ We also found that rprml deficiency drastically compromised HSPC specification from the VDA and the formation of the PCVP at 30-36 hpf, when budding of the HSPC from the artery floor occurs. The PCVP or HSPC niche appear incompletely branched at a critical period to receive the first HSPCs derived from the AGM/VDA. www.nature.com/scientificreports www.nature.com/scientificreports/ Initial formation of HSPCs from the VDA is significantly reduced, as characterized by a drastic reduction of CD41:GFP low+ HSPCs. Hence, we favor the hypothesis that reduction in HSPCs is due to a defect in the specification and formation of the HSPC perivascular niche at the CHT. We tracked the defect in CHT niche formation to increased apoptosis in rprml kd embryos, as concluded by increased Casp-3 activity. This is consistent with results previously published in a conference paper 24 , which reported increased apoptosis secondary to knockdown of RPRML in human cultured cells. Theoretically, rprml knockdown could cause instability of HSPCs through increased apoptosis. Interestingly, overexpression of RPRM causes increased apoptotic activity 25,26 . Therefore, RPRML might have an opposing role when compared with RPRM 24 . All these possibilities need to be thoroughly examined in future research efforts.
During transient-definitive hematopoiesis, EMPs give rise to the first granulocytes produced in the PBI of the zebrafish embryos 5 . Our data show that generation of EMPs is impaired by rprml loss-of-function. tal1/ scl1 + -erythroid-, pu.1 + -myeloid precursor-and early mpx + -granulocyte (including neutrophils)-cells fail to form in the PBI of rprml-deficient zebrafish embryos. Thus, rprml seems required for both waves of definitive hematopoiesis: formation of EMPs and HSPCs in the PBI and VDA, respectively. Whether this finding represents a conserved feature of vertebrates will have to be evaluated by future studies using cell culture and murine models. However, the retention of rprml in all main groups of vertebrates (e.g. tetrapods, coelacanths, bony fish, cartilaginous fish) 11 , could be an indication of the importance of its role in hematopoiesis in all vertebrate groups. In agreement with the phyletic distribution of rprml, other genes that are fundamental for hematopoiesis (e.g. cmyb) are also present in all main groups of vertebrates 27 .
In conclusion, the data presented here demonstrate for the first time a physiological role for the rprml gene. Specifically, we showed that rprml plays an essential role during both waves of definitive hematopoiesis by regulating the formation of EMPs and HSPCs. Additionally, rprml is required for proper HSPC niche formation in the PCVP/CHT. To our knowledge, this is the first study that provides an understanding of rprml function during hematopoiesis. Thus, the identification and validation of rprml as a regulator of definitive hematopoiesis may lead to a better understanding of both human physiology and pathology. Future studies of the physiological roles and functional interplay between the other members of the Reprimo gene family will shed light not only about the biological process in which the different RPRM genes are involved, but also of the functional evolutionary specification of genes that were originated as a product of whole genome duplications.

Materials and Methods
Zebrafish lines. Wild type TAB5 and transgenic strains were bred according to standard methods 28 . Embryos were raised in system water at 28 °C and staged according to either hpf or morphological criteria 29 . Transgenic lines used in this study correspond to Tg(fli1:GFP;flk1:mCherry), Tg(flt0.8:mcherry;CD41GFP), Tg(flt0.8:RFP), Tg(fli1:GFP) 28 , Tg(kdrl:eGFP), Tg(Cd41:GFP) 9 , Tg(dll4:GFP), Tg(Tp1:nRFP) 23 and Tg(mpx:GFP) 29 . All zebrafish studies and experimental protocols were conducted under the guidance and approval of the Institutional Animal Care and Use Committee and Bioethical Advisory Board at Pontificia Universidad Católica de Chile; and followed the AALAC reference resource on guidance on the housing and care of zebrafish (Danio rerio) 30 .
Knockdown of rprml by use of CRISPR-Cas9 genetic system. We used the CRISPR-Cas9 system to disrupt rprml. CRISPR short guide RNAs (sgRNAs) targeting the rprml exon 1 (the only exon present in rprm gene family) were designed using open access internet-based software CRISPRscan (http:// www.crisprscan.org/) 31 . This software was used to determine an on-target score to predict Cas9 cleavage efficiency using previously described algorithms 31 . sgRNAs were tested in vitro for efficiency (New England Biolabs: https://doi.org/10.17504/protocols.io.ch2t8d). Using the sgRNA forward primer for T7 template 5′-TAATACGACTCACTATAGGTAGGACTGAGTTAGCCGCGTTTTAGAGCTAGAA-3′ and the invariance reverse primer 5′-AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATA ACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC-3′. Next, fertilized zebrafish eggs from the transgenic line Tg(fli1:GFP) were injected with a mixture of Cas9 protein (400 ng/µL) and gene-specific sgR-NAs (300 ng) at one-cell developmental stage, as previously described 31 . To identify indel mutations in injected embryos with vascular defects, we used the T7 endonuclease assay (Fig. S2F) 32,33 . This assay relies on the detection of single-base mismatches in the target DNA sequence. A 20 μL reaction volume containing 300 ng of each target DNA amplified from genomic DNA of microinjected and wild type embryos and 1 μL of T7 endonuclease (New England Biolabs) were incubated at 37 °C for an hour. Results were observed on a 1.5% agarose gel. The rprml primer sequences used for T7 endonuclease were; forward primer (5′-3′): CTCCCTCCATCCATCCATGC, and reverse primer (5′-3′): GAACCTCTCGTCCACATCGG.
Morpholino and RNA injections. Embryos were microinjected at 1-2 cell-stage with 3 nL of morpholino (MO) solution (along with phenol red) at a concentration of 0.5 mM for rprml and 1 mM for the control MOs. MOs were obtained from Gene Tools (Gene Tools, LLC) and designed to block the translation start site. MO sequences were: rprml MO (5′-3′) = ACGTTCCGTTCATCCTGAGCAGAGA, and control MO (5′-3′) = CCTCTTACCTCAGTTACAATTTATA. Since rprml is a single-exon gene, it is not possible to design MOs targeting the splicing donor sites. Efficiency of knockdown was assessed by reduction in protein synthesis, as confirmed by immunohistochemistry (IHC) staining in rprml morphants compared to control MO-injected embryos (Fig. S2D,E). This antibody was validated in our previous work 14 . Messenger RNA (mRNA) was co-injected with MOs in phenotypic rescue experiments (Fig. S5).

RpRMs molecular cloning and rescue experiments.
Rprml gene was amplified by PCR by using the following primer sequences Fwd (5′-3′) = AGTTGCTCGACTCAACAGGC and Rv (5′-3′) = TGTTAACAGGTGTGACCCGC. PCR products were gel purified by using AxyPrep ™ DNA Gel Extraction www.nature.com/scientificreports www.nature.com/scientificreports/ Kit (Axygen) and cloned into pCRII-TOPO cloning vector (Invitrogen) according to manufacturer´s instructions. cDNA fragment identity was confirmed by sequencing. In order to generate mRNA, rprml was subcloned into pCS2+ plasmid vector by using BamHI and XhoI restriction enzyme sites. In turn, the plasmid was linearized by using NotI and messenger RNA was obtained by using mMessage-mMachine kit (Ambion) according to manufacturer's instructions.
RNA extraction and cDNA synthesis from zebrafish embryos. For cDNA synthesis, total RNA was collected from pools of 50 rprml MO-injected and 50 uninjected control embryos at 48 hpf using E.Z.N.A total RNA kit I (OMEGA), according to manufacturer's instructions. cDNA was synthesized from 1 µg of total RNA with oligo (dT) using ImProm-II Reverse transcriptase II (PROMEGA) following manufacturer's guidelines.
Quantitative polymerase chain reaction (qpCR). Quantitative (q) PCR was carried out with a Stratagene  Table 1).
Whole mount in situ hybridization. Templates for probe synthesis were amplified from embryonic zebrafish cDNA using primers including T7 RNA polymerase promoter sequence. The 5′ untranslated regions (5′-UTR) of rprml gene was used for primer design. Expression pattern analysis of hematopoietic (tal1/scl1, pu.1, mpx, cmyb), vascular (fli1, lmo2) and T-lymphocyte (rag-2) markers was carried out as described previously 21,34,35 . Purified PCR products were transcribed in vitro and labeled using digoxigenin (DIG) RNA labeling Kit (Roche) according to manufacturer's instructions. Riboprobes were purified by mini Quick Spin RNA Columns (Roche) and stored at −80 °C with deionized formamide for posterior use. Whole mount in situ hybridization (WISH) was carried out as previously described 36 . Immunohistochemistry (IHC) staining. 24 to 48 hpf embryos were fixed overnight in 4% paraformaldehyde at 4 °C and dehydrated in methanol for subsequent use. Rehydrated embryos were treated with acetone for 20 minutes at −20 °C, washed and treated with Proteinase K 10 ng/mL for 15 min as antigen exposure steps. Blocking was carried out with 10% FBS + 1% DMSO in PBT twice for 1 hour each. Embryos were incubated with the primary antibody overnight at 4 °C in blocking solution at corresponding dilution. Rabbit polyclonal anti-activated Caspase-3 antibody (Abcam, ab13847) was used at a 1:500 dilution, mouse monoclonal anti-GFP at 1:1000 dilution (Thermo Fisher Scientific, A11120), rabbit polyclonal anti-GFP at 1:500 dilution (Thermo Fisher Scientific, A11122) and rabbit polyclonal anti-RPRM at 1:1000 dilution (USbiological life science, 364937). This polyclonal antibody recognizes a common sequence for RPRM and RPRML, justifying its use for assessment of Rprm/Rprml protein expression in rprml-morphant embryos. For secondary antibody labeling, goat anti-rabbit Alexa-555, goat anti-mouse Alexa 488, goat anti-mouse Alexa 488, and goat anti-rabbit Alexa 546 (Invitrogen) were used at a 1:200 dilution in blocking buffer for 2 hours at room temperature. Image acquisition, processing and quantification. WISH stained zebrafish embryos were embedded in 75% glycerol/PBS and imaged using a NIKON eclipse 80i microscope equipped with a DS-Vi1 (NIKON) camera. For whole-mount fluorescent IHC images, embryos were mounted in acrylic rings with glass bottoms containing 1% low melting point agarose and images were acquired using a Zeiss LSM780 laser-scanning spinning disc microscope. VE-DIC and fluorescent microscopy were performed as described previously 37 through the 60X/1.00 water-immersion objective of a Nikon eclipse E200 microscope with a 519CU 5.0 M CMOS Camera using the Micrometrics SE Premium 4 program. To quantify the effect of the morpholinos on the development of the vascular system, CHT thickness was measured using ImageJ64. Pooled data were calculated as the mean plus or minus SD with minimum number of three independent experiments. Pairwise comparisons were subject to statistical analysis using unpaired two-tailed Student's t-test. P-value < 0.05 was considered statistically significant.