The predominant protein arginine methyltransferase PRMT1 is critical for zebrafish convergence and extension during gastrulation Yun-Jung Tsai 1 , Huichin Pan 1,2 , Chuan-Mao Hung 1 , Po-Tsun Hou 1 , Yi-Chen Li 1 , Yu-Jen Lee 3 , Yi-Ting Shen 1,4 , Trang-Tiau Wu 4,5 and Chuan Li 1,2 1 Department of Biomedical Sciences, Chung Shan Medical University, Taichung, Taiwan 2 Department of Medical Research, Chung Shan Medical University Hospital, Taichung, Taiwan 3 Institute of Biochemistry and Biotechnology, Chung Shan Medical University, Taichung, Taiwan 4 Department of Pediatric Surgery, Chung Shan Medical University Hospital, Taichung, Taiwan 5 School of Medicine, Chung Shan Medical University, Taichung, Taiwan Introduction Protein arginine methylation is a post-translational modification involved in various cellular functions, such as signal transduction, protein subcellular locali- zation, transcriptional regulation, protein–protein interactions and DNA repair [1]. At least 11 protein arginine methyltransferase (PRMT) genes have been identified in the mammalian system that catalyze the transfer of methyl groups from S-adenosylmethionine (AdoMet) to the side-chain x-guanido nitrogens of arginine residues in protein substrates. The activity can be further divided into types I and II, depending on the catalyses of formation of asymmetric di-x-N,N- methylarginines or symmetric di-x-N,N¢-methylargi- nine residues respectively [2,3]. Keywords convergence and extension; gastrulation; PRMT1; protein arginine methylation; zebrafish Correspondence C. Li, Department of Biomedical Sciences, Chung Shan Medical University, Taichung, Taiwan Fax: +886 4 23248187 Tel: +886 4 24730022 11807 E-mail: cli@csmu.edu.tw (Received 27 August 2010, revised 19 December 2010, accepted 5 January 2011) doi:10.1111/j.1742-4658.2011.08006.x Protein arginine methyltransferase (PRMT)1 is the predominant type I methyltransferase in mammals. In the present study, we used zebrafish (Danio rerio) as the model system to elucidate PRMT1 expression and function during embryogenesis. Zebrafish prmt1 transcripts were detected from the zygote period to the early larva stage. Knockdown of prmt1 by antisense morpholino oligo (AMO) resulted in delayed growth, shortened body-length, curled tails and cardiac edema. PRMT1 protein level, type I protein arginine methyltransferase activity, specific asymmetric protein argi- nine methylation and histone H4 R3 methylation all decreased in the AMO-injected morphants. The morphants showed defective convergence and extension and the abnormalities were more severe at the posterior than the anterior parts. Cell migration defects suggested by the phenotypes were not only observed in the morphant embryos, but also in a cellular prmt1 small-interfering RNA knockdown model. Rescue of the phenotypes by co-injection of wild-type but not catalytic defective prmt1 mRNA con- firmed the specificity of the AMO and the requirement of methyltransferase activity in early development. The results obtained in the present study demonstrate a direct link of early development with protein arginine methylation catalyzed by PRMT1. Abbreviations AdoMet, S-adenosylmethionine; AMO, antisense morpholino oligo; C ⁄ E, convergence ⁄ extension; hpf, hours post-fertilization; NR, nuclear receptor; PRMT, protein arginine methyltransferase; r, rhombomere; Sam68, Src-associated substrate during mitosis with a molecular mass of 68 kDa; siRNA, small-interfering RNA; STAT1, signal transducer and activator of transcription 1; WISH, whole-mount in situ hybridization; xPRMT1b, Xenopus protein arginine methyltransferase type I b. FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS 905 PRMT1 is the predominant and most abundant type I methyltransferase in mammals [2,3]. RNA bind- ing proteins such as fibrillarin, Sam68 (Src-associated substrate during mitosis with a molecular mass of 68 kDa) and many hnRNPs with arginine and glycine rich RGG motifs [2,4,5] or a RXR sequence [6] are typical substrates of PRMT1. Methylation of proteins such as hnRNPA2, Sam68 and hnRNPQ that shuttle between the cytoplasm and nucleus can affect their subcellular localization [7–9]. Arginine methylation has been reported to affect the protein–RNA or protein– protein interaction of some RNA binding proteins. For example, the interaction of hnRNPK with c-Src is reduced with arginine methylation [10]. PRMT1 also plays multiple roles in various signal- ing pathways and transcriptional regulation. For example, interaction of PRMT1 with the cytoplasmic domain of interferon-a receptor [11], and the putative methylation of signal transducer and activator of tran- scription 1 (STAT1) [12–14] and protein inhibitor of activated STAT1 [15] by PRMT1, indicate its role in interferon signaling. Furthermore, methylation of the transcriptional factor FOXO1 by PRMT1 can inhibit its phosphorylation by AKT and promote nuclear localization and transactivation of FOXO1 [16]. In addition, PRMT1 is a transcriptional coactivator of various nuclear receptors (NRs) [17] as another PRMT family member PRMT4 ⁄ CARM1 (coactivator-associ- ated arginine methyltransferase) [18]. Methylation of R3 of histone H4 by PRMT1 is part of the epigenetic histone code critical for chromatin structure and tran- scriptional activation [19]. Increased H4R3 methylation by the recruitment of PRMT1 has been reported with transcription factors other than NRs, including p53 [20] and YY1 [21]. Furthermore, PRMT1 can directly methylate some transcription factors, coactivators or transcriptional elongation factor to modulate tran- scription. For example, methylation of an orphan NR HNF4 by PRMT1 can increase its DNA binding affin- ity [22]. Methylation of the transcriptional elongation factor SPT5 by PRMT1 also regulates its promoter association and RNA polymerase II interaction [23]. Mouse embryos homozygous for PRMT1knockout failed to develop beyond the onset of gastrulation (embryonic day 6.5), indicating that PRMT1 is critical in early embryogenesis [3]. Xenopus protein arginine methyltransferase type I b (xPRMT1b) is maternally expressed and subsequently transcribed zygotically throughout the developing stages. Overexpression of xPRMT1 was found to induce the expression of a spectrum of neural markers, and antisense morpholino oligonucleotides (AMOs) against xPRMT1b impaired neural development, indicating that xPRMT1b plays a role in the early steps of neural determination [24]. However, the correlation of the phenotypes with pro- tein arginine methylation catalyzed by the methyltrans- ferase was not studied. The PRMT genes are highly conserved from zebrafish to humans, and the identity of the PRMT1 proteins is close to 90% [25]. Because zebrafish is amenable to genetic manipulation and the transparent embryos can be directly observed under microscope, we used zebrafish (Danio rerio) as a model system to monitor the relationship between protein arginine methylation and early developmental changes in fish embryos. Results Ubiquitous expression of prmt1 RNA and protein in zebrafish embryonic development Alternative splicing of prmt1 results in various mRNA and protein isoforms in mammals [3,26,27]. However, no support for alternative splicing of zebrafish prmt1 could be obtained from a database search. Ensembl (ENSDARG00000010246) illustrates that zebrafish prmt1 contains 10 exons and the prmt1 mRNA appears to be analogous to the v1 form of mammalian prmt1 mRNA (connecting the first exon and the con- stitutive 102 nucleotide exon with no alternative exons in between; Fig. 1A). A primer set to amplify the puta- tive alternatively spliced region (Fig. 1A) detected a single RT-PCR product of 138 bp for RNA prepared from embryos from one cell to 72 h post-fertilization (hpf) (Fig. S1A). The results opposed alternative splic- ing at the 5¢ end of the zebrafish prmt1 gene. The RT-PCR product further confirmed the presence of an upstream in-frame ATG within a Kozak sequence located 21 nucleotides upstream of the start site sug- gested in NCBI (NM_200650) (Fig. 1B). The predicted N-terminal amino acid sequence is indicated (Fig. 1C). Ubiquitous expression of prmt1 in various adult tis- sues, such as the brain, heart, spleen, swim bladder, gill, testis, ovary and muscle, was also demonstrated by RT-PCR (Fig. S1B). Western blot analyses further detected a 42 kDa PRMT1 protein signal expressed at different zebrafish embryonic stages (Fig. S1C). There- fore, PRMT1 protein is expressed both maternally and zygotically, comparable to the mRNA. Spatial and temporal expression pattern of prmt1 mRNA by whole-mount in situ hybridization (WISH) Zebrafish prmt1 mRNA was strongly and ubiquitously expressed in embryos through the one- to four-cell PRMT1 expression and function during embryogenesis Y J. Tsai et al. 906 FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS stages (Fig. 2A–C), demonstrating the maternal origin and homogeneous distribution of prmt1 mRNA during the very early cleavages. Continuing homogenous expression at 6 and 12 hpf indicated zygotic transcrip- tion from gastrulation to the early segmentation period (Fig. 2D–F). At 24 hpf, prmt1 was strongly expressed in the head regions, including the eyes, otic vesicle, forebrain, midbrain and hind-brain (Fig. 2G). Expres- sion in somites was also detected. As development pro- ceeded, the expression of prmt1 decreased in most parts of the brain but continued at somites at 48 and 72 hpf (Fig. 2H, I). The signals are specific to prmt1 because the sense riboprobe did not detect any signifi- cant signals (Fig. 2J). Immunofluorescent analyses of the PRMT1 protein also revealed similar expression patterns (Fig. 2K). Knockdown of prmt1 with specific morpholino oligonucleotides affects zebrafish development AMOs designed to hybridize the 5¢ region of a target mRNA can selectively block translation and knock- down gene activity [28]. Because two in-frame ATGs are present at the 5¢ region of prmt1, we synthesized two non-overlapping AMOs to target the upstream and downstream ATG (MO1 and MO2 respectively) (Fig. 1C). Injection of high-dosed MO1 (8 ng) resulted in the lysis of some embryos and a severely truncated phenotype in most survived embryos (data not shown). Similar phenotypes with different degrees of defects were observed when MO1 was injected at 4 ng or MO2 at 8 ng (Fig. 3B). The abnormalities were classi- fied as mild, moderate and severe at 48 hpf, with dif- ferent degrees of body curvature being associated with curved or shortened tails (Fig. 3B–D). Other abnor- malities, such as cardiac edema, enlarged yolk and shortened yolk stalk, smaller eyes and seriously trun- cated or bended tails, were also observed in some mor- phants. At 72–120 hpf, the phenotype of edema and swollen yolk became even worse (Fig. S2F–H, J–L, N–P), indicating poor circulation and metabolism. The ratio of morphants with abnormal phenotypes increased as the dose of the injected AMO increased (Fig. 4F; 58–98% for 2–4 ng of MO1; 75–94% for Fig. 1. Genomic structure and partial nucleotide and amino acid sequences of the zebrafish prmt1 gene. (A) Genomic structure of human and zebrafish prmt1. Three major human splicing variants [27] and the only identified zebrafish splicing form are shown. Exons are repre- sented as boxes and introns by the connecting lines. Numbers in the boxes represent the exon length in base pairs. Arrows indicate the position of the start and stop codons. Filled boxes are coding and open boxes are noncoding regions. The start ATG in human prmt1 was in accordance with that suggested in a previous study [27]. According to the zebrafish prmt1 mRNA sequence, an ATG (arrow) 21 nucleotides upstream of the previously identified ATG (NM_200650, arrowhead) is mostly likely to be the translational start site. The positions of primers used in the present study are indicated. HsPRMT1v3 (NP_938075.2), HsPRMT1v2 (NP_001527.3), HsPRMT1v1 (NP_938074.2), DrPRMT1 (NP_956944.1). Hs, Homo sapiens; Dr, Danio rerio. (B) The DNA sequence around the ATG translational start site of prmt1 (from the 38 nucleotides of NM_200650) is shown. The two in-frame ATGs are boxed. AMO binding sites complementary to the antisense morpholino oligonucleotide MO1 and MO2 are underlined. MO2 begins 20 bp downstream of the first ATG of zebrafish prmt1. (C) Comparison of the N-terminal sequences of human and zebrafish PRMT1. Y J. Tsai et al. PRMT1 expression and function during embryogenesis FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS 907 4–8 ng of MO2). The percentage of moderate or severe phenotypes also increased significantly with the raised AMO dose. This dose-dependent phenotypic severity indicates the specificity of prmt1 knockdown. With the aim of observing phenotypes beyond gastrulation, we studied the morphants by the injection of 4 ng of MO1 or 8 ng of MO2 in subsequent experiments. Reduced level of PRMT1 protein, type I PRMT activity and protein arginine methylation in prmt1 morphants A reduced PRMT1 protein level appeared to correlate with the phenotypic severity in the MO1-injected embryos (Fig. 4A). A decrease of PRMT1 expression was found at 24, 48 and 72 hpf in MO1 or MO2- injected morphants (data not shown). Thus, injection of MO1 and MO2 indeed blocked the expression of PRMT1 protein in zebrafish embryos, both effectively and persistently. Because PRMT1 is the predominant type I protein arginine methyltransferase, the type I activity in the morphants should be reduced correspondingly. In vitro methylation reaction with a typical type I PRMT sub- strate fibrillarin showed that fibrillarin methylation cat- alyzed by the morphant extract was reduced compared to that by the wild-type extract (Fig. 4B). The type I activity remained low from 24–72 hpf. We further examined the level of protein arginine methylation in the embryos with an antibody ASYM24 that recognizes asymmetrically dimethylated arginines in alternate RG sequences [8]. Dozens of zebrafish embryonic proteins of different molecular masses were detected and most of the methylarginine-specific sig- nals were reduced in the prmt1 morphants (Fig. 4C). We then examined protein arginine methylation of specific PRMT1 substrates. Histone H4 arginine 3 methylation catalyzed by PRMT1 was abolished in PRMT ) ⁄ ) mouse embryonic stem cells [17]. We thus determined H4 R3 methylation in the embryos. As shown in Fig. 4D, asymmetric arginine dimethylation at this residue detected by a modification-specific anti- body was reduced in the morphants. Detection with another H4-specific antibody confirmed an equal load- ing of H4 protein. These results confirm that the reduction of H4 R3 methylation was not a result of decreased H4 protein but instead was caused by the reduced expression of PRMT1 in the morphants. Reduced medial–lateral convergence and a shortened anterior–posterior axis in the morphants at early segmentation stage Because defective phenotypes observed in the prmt1 morphants are probably a consequence of earlier defects, we evaluated zebrafish development at the Fig. 2. Spatial and temporal expression of prmt1 by WISH and immunofluorescent analysis. Zebrafish embryos at the one-cell stage (A), two-cell stage (B), four-cell stage (C), 6 hpf (D), 12 hpf (E, F), 24 hpf (G), 48 hpf (H) and 72 hpf (I) were analyzed by WISH. A dorsal view of the 12 hpf is shown in (F). WISH with sense riboprobe is shown in (J). Immunofluorescent analysis with anti- PRMT1 of 24-hpf embryos is shown in (K). a, adaxial cells; e, eye; f, forebrain; h, hind- brain; m, midbrain; mhb, mid-hindbrain boundary; ov, otic vesicle; som, somites. PRMT1 expression and function during embryogenesis Y J. Tsai et al. 908 FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS segmentation stage with different markers to pinpoint the defects. Expression of krox20 is restricted to rhom- bomeres (r)3 and 5 (r5) in the hindbrain region. At the 10-somite stage, krox20 expression in the morphants at r3 and r5 was laterally extended (by 1.2–2.5-fold) and the posterior r5 is more extended than r3 (Fig. 5A). Generally, the anterior–posterior distance between r3 and r5 was reduced, and the extent of reduction was also correlated with the degree of lateral extension. We grouped the morphants according to the degree of abnormal krox20 expression. Abnormal somite development in the prmt1 mor- phants at 10-somite stage was clearly revealed by a muscle and somite-specific marker myoD. As shown in Fig. 5B, myoD expression in the two rows of adaxial cells flanking notochord was irregularly bent at the posterior end in type 1 morphants. In type 2 mor- phants, the width between the two rows increased, with lateral myoD expression being diminished at the end of one side, and extended and compressed at the other. The width was greatly broadened and the lateral myoD expression was greatly expanded in type 3 mor- phants. Even though the same number of segments was present in the morphants, the distances between the segments were extremely compressed. Generally, the markers showed shortened anterior– posterior axes in the morphants and the abnormalities were more severe at the posterior than the anterior part of the embryos. The percentages of the three types of abnormal phenotypes observed for each marker gene are shown in Fig. 5C. Developmental defects of prmt1 morphants at gastrulation The shortened anterior–posterior axes in the prmt1 morphants at the segmentation stage indicate defects in convergence and extensions (C ⁄ E) at gastrulation. At gastrulation, the three germ layers and the body plan are established by directed and coordinated cell movements, including epiboly to cover the yolk cells by spreading the blastomeres, involution to internalize the marginal cells to form the precursors of the meso- derm and endoderm, and C ⁄ E, in which cells accumu- late on the dorsal side and lead to axis formation. Gastrulation begins at 50% of epiboly (6 hpf) and ends at 100% (10 hpf). Defective epiboly can be observed in most mor- phants at 10 hpf. Although wild-type embryos showed complete blastopore closure, the MO2-injected embryos cannot close the yolk plug and demonstrated varying degrees of open blastopores (Fig. 6A). Staining with notail (ntl, expressed in the ring mesoderm and endodermal precursors around the margin as a pan- mesendodermal marker) showed shortened but wid- ened notochords in the prmt1 morphants. The axial mesendoderm failed to migrate to the anterior. The morphants are grouped according to the degree of Fig. 3. Defective phenotypes in prmt1 knockdown zebrafish. Phenotypes of embryos injected with zprmt1 MO at 48 hpf. Uninjected wild-type embryos are shown (A). The injected embryos are classi- fied into mild, moderate and severe accord- ing to the phenotypes at 48 hpf. The three types of MO injected embryos at 48 hpf are shown in (B–D). The injected embryos with normal body axes as the wild-type are classified as ‘normal’. (E) Frequencies of three phenotypes caused by injection of prmt1 MO (2 or 4 ng of MO1 and 4, or 8 ng of MO2). The injected embryos with normal body axes as the wild-type are classified as ‘normal’. Y J. Tsai et al. PRMT1 expression and function during embryogenesis FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS 909 Fig. 4. Reduced PRMT1 protein expression, type I protein arginine methyltransferase activity and specific protein arginine methylation in prmt1 morphants. (A) Proteins were prepared from embryos either not injected, or injected with prmt1 MO1. Western blot analysis of PRMT1 protein in embryos injected with 4 ng of MO1 with phenotypes classified as mild or moderate at 48 hpf are shown. Detection by anti-b-actin was used as a loading control. WT, wild-type; M, morphants. (B) In vitro methylation was conducted with extracts from MO2 (8 ng) injected embryos at 24, 48 and 72 hpf as the source of protein arginine methyltransferase and recombinant mouse fibrillarin as the methyl-accepting protein. The samples were separated by SDS ⁄ PAGE and the methylated proteins were detected by fluorography. (C) Argi- nine-methylated proteins in 48 hpf embryos were detected by western blotting with an asymmetric dimethylarginine-specific antibody ASYM24. Detection by anti-b-actin was used as a loading control. (D) Western blot analysis of H4R3me2 levels in 48 hpf morphant embryos. Analysis of histone H4 served to normalize levels of H4R3me2 in morphants and wild-type embryos. Fig. 5. Defective phenotypes at segmentation stage for prmt1 morphants. (A) Dorsal view of embryos (10-somite stage) for krox20 staining. The positions of r3 and r5 are indicated. The widths of r3 and r5 and the vertical distance between r3 and r5 are indicated. In type 1 mor- phants, r3 was almost normal but r5 was slightly extended laterally. The width of r3 and r5 were extended to  1.5-fold in type 2 and even to 2–2.5-fold in type 3. (B) Expression of myoD at paraxial ⁄ adaxial mesoderm at the 10-somite stage. Dorsal views, anterior at top. The lengths of myoD expressed paraxial ⁄ adaxial mesoderm are indicated. In type 1, 2 and 3 morphants, the length was  0.8–0.9, 0.6–0.7 and 0.5 compared to normal. (C) The phenotypes were classified according the degree of abnormality as type 1, 2, and 3. Percentages of wild- type and morphants embryos within each phenotypic category are shown in the bar graphs. n, total embryos counted in the experiments. PRMT1 expression and function during embryogenesis Y J. Tsai et al. 910 FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS shortening and widening of the notochord (Fig. 6B). By contrast, the expression of goosecoid (gsc), a mes- endoderm marker expressed mainly in the prechordal plate, did not reveal clear differences between wild- type and morphants (Fig. S3A). Expression of a ventral mesodermal marker tbx6, a member of the Brachyury-related T-box family, revealed a slight epibolic delay and a thickened germ ring in morphant embryos at 6 hpf. The tbx6 expres- sion at 10 hpf showed a margin with a larger unen- closed blastopore in the morphants (Fig. S3B). Even though the expression of an endodermal marker sox17 was not eliminated from the endoderm progenitors, the strong single dot stained by sox17 at dorsal fore- runner cells (i.e. that will become Kupffer’s vesicle) split into two (or a few) spots in some morphants (Fig. S3C). Rescue of the C ⁄ E during gastrulation of the zebrafish prmt1 morphants by injection of prmt1 cRNA To further demonstrate the direct relationship between AMO-mediated knockdown of PRMT1 and the phe- notypes described, rescue experiments with prmt1 cRNA were conducted. No significant phenotypic changes were observed when 50 ng of wild-type or 5¢ mutated (AMO-mismatched nucleotide sequences without changing amino acid sequences) prmt1 RNA were injected alone. We then co-injected the AMO with prmt1 cRNA. As observed at the early gastrula- tion stage, the cRNA (50 ng) can partially rescue the abnormal phenotypes induced by MO-2 (4 ng). The defective C ⁄ E phenotypes revealed by ntl staining at 10 hpf were classified as shown in Fig. 6B. The per- centage of the severe phenotypes decreased greatly in co-injected embryos (Table 1). To examine whether the phenotypes of prmt1 knockdown and the rescue of the morphants were a result of the methyltransferase activ- ity of PRTM1 or the PRMT1 protein per se, we pre- pared cRNA of catalytically inactive PRMT1. Three conserved amino acids SGT at the AdoMet-binding site were mutated to AAA, as previously reported by Balint et al. [29]. We showed that the abnormal pheno- types of the morphants cannot be rescued by the catalytic-defective cRNA (Table 1). Increased methyl- transferase activity assayed by in vitro methylation Fig. 6. Knockdown of prmt1 induces gastrulation defects. Wild-type and prmt1 MO2 (8 ng) injected embryos at 10 hpf were examined. (A) The morphants showed abnormal morphology at the end of epiboly (10 hpf) as reflected by different degrees of open blastopores. Dashed arrows and semicircles depict embryo lengths and angles between anterior–posterior ends. Lateral views, dorsal to the right. (B) ntl (staining the forerunner cell group, axial chorda mesoderm) staining of the embryos at 10 hpf. Dorsal views, anterior at top. The morphants are grouped according to the degree of shortening and widening of the notochord. Table 1. Rescue of gastrulation defects by catalytic active but not catalytic inactive zebrafish prmt1 cRNA. Normal (%) Type1 (%) Type2 (%) Type3 (%) Total (n) WT 100 120 MO2 5 18 51 26 176 MO2 + WT cRNA 29 39 25 7 100 MO2 + MT cRNA 8 24 50 18 161 MO2 (4 ng) were co-injected with zebrafish prmt1 cRNA (50 pg). The WT cRNA contains mismatches at the MO target site without changing the encoded amino acids. The MT cRNA contains the same mismatches and mutations at the AdoMet-binding site. The embryos were analyzed at 10 hpf by staining with ntl. The pheno- typic categories are classified as shown in Fig. 6. Y J. Tsai et al. PRMT1 expression and function during embryogenesis FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS 911 could be detected in extracts from the rescued embryos compared to that from morphants or morphants res- cued by catalytically inactive RNA (data not shown). The results obtained in the present study thus confirm that the phenotypes in the morphants were specifically a result of the reduced PRMT1 methyltransferase activity caused by the knockdown. Reduced PRMT1 level and defective cell movements in human Huh7 cells From the above analyses, prmt1 knockdown did not affect cell speciation, although cells in the morphant embryos appeared to migrate slower and resulted in the observed shortened anterior–posterior axes and lateral-expanded defects. The developmental program was not blocked but progressed with a slight delay in the prmt1 morphants from gastrulation to segmenta- tion. The phenotypes are thus likely to be the result of defective cell movements. Genes involved in cell movements during embryo- genesis are usually also involved in cellular migration. We thus studied whether reduced PRMT1 can affect cell movement in a cellular model. Huh7 is a human hepatocarcinoma cell line in which cell migration can be detected under normal growth conditions without any induction. PRMT1 small-interfering RNA (siRNA) knockdown reduced the PRMT1 protein level to  60% of that of control siRNA-treated Huh7 cells (Fig. 7A). Reduced cell movement can be observed in PRMT1 siRNA-treated cells compared to control cells, as shown in Fig. 7B. The capacity of cell movement in PRMT1 knockdown cells is reduced to  75% compared to that of control cells. The decreased cell movement in PRMT1 knockdown cells is statistically significant (Fig. 7C). The results obtained indicate that PRMT1 functions in the regula- tion of cell movement. Discussion In the present study, we demonstrate that the prmt1 gene is actively and ubiquitously expressed at both RNA and protein levels at the early developmental stages of zebrafish. The mRNA and protein are pres- ent before mid-blastula transition and thus are mater- Fig. 7. Reduced cell migration in a prmt1-deficient cell model. Huh7 cells were treated with control or prmt1 siRNA. (A) Cell extracts from the siRNA-treated cells were immunoblotted with anti-PRMT1. Detection by anti-b-tubulin was used as a loading control. Reduced PRMT1 protein expression by siRNA was normalized with the b-tubulin signal. (B) Images of the pre-migration and post-migration cells stained with crystal violet are shown. The white circles indicate areas covered by the stoppers before cell migration. (C) Quantification of cell movement is represented as the percentage of the area covered by migrated cells in prmt1 siRNA-treated cells compared to that in control cells. Data are shown as the mean ± SD of two independent experiments performed in quadruplicate. A statistically significant difference between the two siRNA-treated cells is indicated (**P < 0.01; Student’s t-test). PRMT1 expression and function during embryogenesis Y J. Tsai et al. 912 FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS nally derived. The continuous prmt1 expression indi- cates zygotic expression. The results obtained are con- sistent with previous reports of prmt1 expression in mice or Xenopus early development [3,24]. We also detected wide expression of prmt1 in various adult zebrafish tissues. Ubiquitous expression of PRMT1 was reported in different human, rodent or fish (Japa- nese flounder Paralichthys olivaceus) adult tissues [2,3,26,30]. The results obtained in the present study show that the ubiquitous expression of PRMT1 in adult tissues starts at early embryogenesis. In mice, the prmt1 homozygous mutants die at approximately embryonic day 6.5 when gastrulation begins [3]. On the other hand, prmt1 knockout embry- onic stem cells are viable, indicating the specific require- ment of prmt1 in early embryogenesis. Knockdown and overexpression of xPRMT1b in Xenopus were pre- viously found to provide valuable information about the gene in early neural development [24], although the focus of that study was on the roles of PRMT1 involv- ing Ca 2+ neural induction, and its effects on other developmental aspects were less discussed. In the pres- ent study, we successfully knocked down the expres- sion of prmt1 in zebrafish by injection of AMO into one-cell embryos. Observation of the defective pheno- types of the zebrafish prmt1 morphants provides the possibility of evaluating the effects of PRMT1 viably beyond gastrulation. We observed a shortened body length and curved tails in the majority of prmt1 mor- phants. Body axis shortening and lateral expansion in the morphants were even obvious in the posterior part of the embryos, as revealed by marker gene staining. Generally, prmt1 knockdown did not affect cell specia- tion, although cells in the morphants appeared to migrate slower, resulting in the observed anterior–pos- terior shortening and lateral expansion. Even though prmt1 has been implicated in many cellular processes, its involvement in cell movement or migration has not been described. In the present study, we also used a cellular model to demonstrate that PRMT1 knock- down cells migrated more slowly in a simple cell move- ment experiment. Besides PRMT1, PRMT6 knockdown affects genes involved in cellular move- ments and inhibits cell migration [31]. We confirmed the reduced expression of PRMT1 protein in the morphants. Consistently, the level of arginine methyltransferase activity and arginine-methy- lated proteins was reduced upon the injection of AMO. Rescue of the prmt1 morphants with prmt1 cRNA can partially reverse the early defective pheno- types, confirming the specificity of the AMO. Most importantly, catalytic defective mutant prmt1 cRNA lost the ability to rescue the morphants, further supporting the importance of active PRMT1 methyl- transferase in early embryogenesis. Reduced methyl- transferase activity and protein arginine methylation should thus be responsible for the abnormalities in zebrafish prmt1 morphants. The defective phenotypes in epiboly and C ⁄ E indicate wide effects of prmt1 in early embryogenesis. Considering the substrate spec- trum and the coactivator function of PRMT1, it is likely that no single target can explain the wide range of phenotypes. There would be numerous proteins and target genes that might be affected. First, PRMT1 might affect transcriptional regulation through its coactivator activity or by direct modifica- tion of histones or various transcriptional factors. PRMT1 has been shown to be the coactivator of a few NRs [17] and can also serve as a coactivator of p53 [20]. Epigenetic controls play critical roles in develop- ment. The importance of methyltransferases involved in epigenetic regulation, such as DNA methyltransfer- ase Dnmt1 and histone lysine methyltransferase Suv39h1 (specific for H3K9), have been reported in zebrafish development [32]. Methylation of histone H4 R3 is responsible for active chromatin and transcrip- tional activation [17,19]. We showed that overall asym- metric arginine dimethylation of H4R3 was decreased in the prmt1morphants. A low level of methylated H4 R3 bound to certain promoters at critical developmen- tal stages should be responsible for part of the abnor- mal phenotypes of the prmt1 morphants. Second, many typical PRMT1 substrates containing preference RGG or GAR sequences comprise RNA binding proteins that are abundant in the early embryos. Abnormal protein arginine methylation of these substrate proteins might affect their subcellular localization, as well as interactions with RNA or pro- teins, and thus lead to the developmental defects. For example, methylation of a typical RGG box-containing PRTM1 substrate Sam68 is important for its RNA binding activity and nuclear localization [8]. Sam68 has also been reported to be associated with RhoA [33], the downstream key regulator of the noncanonical Wnt pathway controlling C ⁄ E [34]. In addition, Sam68 is required for growth factor-induced migration [35]. We observed a decreased asymmetric arginine dimethyla- tion of Sam68 in both zebrafish morphants and PRMT1 siRNA knockdown cells (data not shown). Whether reduced arginine methylation of Sam68 might be related to defective cell migration requires further investigation. Furthermore, even though embryonic stem cells from mouse with a PRMT1 hypomorphic allele with residual PRMT1 activity are viable, PRMT1-deficient mouse embryonic fibroblasts showed spontaneous DNA damage, G2 ⁄ M accumulation, cell cycle delay Y J. Tsai et al. PRMT1 expression and function during embryogenesis FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS 913 and genome instability [36]. The defects indicate that PRMT1 is involved in the DNA damage response pathway. Knockdown of PRMT1 might thus affect early development as a result of defective cell prolifera- tion or apoptosis. Increased apoptotic cells were detected in the prmt1 morphants (data not shown), which may be correlated with the phenotypes. In summary, in the present study, we demonstrate the importance of the enzyme activity of PRMT1 with zebrafish embryogenesis. We show the relationships between prmt1 knockdown, reduced protein arginine methylation and H4 R3 methylation with respect to early developmental defects at gastrulation in zebra- fish. The present study describes the first thorough investigation of a protein arginine methyltransferase family member in zebrafish. The investigation also establishes zebrafish as a good study platform for pro- tein arginine methylation. Experimental procedures Zebrafish rearing Adult zebrafish (Danio rerio) were maintained under a 14 : 10 h light ⁄ dark cycle at 28 °C. All embryos were collected by natural spawning and staged according to Kimmel et al. [37]. mRNA expression analyses by RT-PCR Total RNA was isolated from embryos at different stages of embryogenesis and different adult tissues by TRIzol reagent (Molecular Research Center, Inc., Cincinnati, OH, USA). First-strand cDNA was synthesized from 5 lgof total RNA by M-MLV Reverse Transcriptase (Promega, Madison, WI, USA). RT-PCR was performed with the pri- mer set ZF1-F and ZF1-R to amplify the conserved regions in zebrafish prmt1 gene (GenBank NM_200650.1) or ASF and ASR for putative alternative splicing at the 5¢ end of prmt1 (Fig. 1A and Table S1). Amplification of the elonga- tion factor 1a (primer set Ef1 and Ef2) was used as an internal control. Zebrafish embryonic extract preparation, western blot analyses and in vitro methylation Zebrafish embryos were manually deyolked [38], resus- pended in extraction buffer (150 mm NaCl, 100 mm Tris ⁄ HCl, pH 7.5, 5% glycerol, 1 mm dithithreitol, 1% Tri- ton X-100, 1 mm phenylmethanesulfonyl fluoride and com- plete protease inhibitor cocktail; Roche Diagnostics, Basel, Switzerland) and then homogenized (400 lL per 100 embryos) by a homogenizer (IKA T10; IKA Ò Works Staufen, Germany). The homogenate was centrifuged at 17 530 g at 4 °C for 20 min and the supernatant was stored at )20 °C as the embryonic extract. Aliquots of the embryonic extract (30 lg of protein) were resolved by SDS ⁄ PAGE followed by western blot analyses with antibodies specific to PRMT1 (Upstate Biotechnology, Lake Placid, NY, USA) and methylarginines (ASYM24; Upstate Biotechnology). In vitro methylation was conducted as described previously [39]. Essentially, embryonic extracts (35 lg of protein), recombinant mouse fibrillarin protein and 1.5 lCi of [methyl- 3 H]-AdoMet (60 Ci ⁄ mmol; Amer- sham Biotech, Little Chalfont, UK) were incubated at 37 °C for 60 min in methylation buffer (50 mm sodium phosphate, pH 7.5) with a total volume of 15 lL. The sam- ples were subjected to SDS ⁄ PAGE. The gels were then stained, treated with EN3HANCE (Perkin Elmer, Wal- tham, MA, USA) and dried for fluorography. Isolation of zebrafish histones and assay for histone methylation Histones were prepared essentially in accordance with the protocol previously described by Gurvich et al. [40]. Zebrafish embryos harvested at 48 hpf were manually deyolked and dissolved in extraction buffer. Nuclei were collected by centrifugation at 17 530 g at 4 °C for 20 min, and histones were extracted by shaking in 0.2 m sulfuric acid for 1 h at 4 ° C. After centrifugation, histones were precipitated with ethanol at )20 °C overnight, washed once with ethanol, and resuspended in distilled water. Aliquots of the zebrafish embryonic extract (10 lg) were resolved by SDS ⁄ PAGE followed by western blot analyses with anti-H4 (Upstate Biotechnology) and anti-H4Me R3 (Upstate Biotechnology). WISH and immunofluorescent analysis Zebrafish prmt1 cDNA obtained from imaGenes (Berlin, Germany) was amplified with the primers set ZF1-F and ZF1R. The fragment was cloned into a modified pGEM vector with partial deletion in the multiple cloning sites and the resulting pGEM-zprmt1 was used for riboprobe preparation. In situ hybridization was performed according to Wester- field [41]. Essentially, after rehydration, proteinase treat- ment and prehybridization, hybridization was performed with 100–200 ng of digoxigenin-UTP labeled riboprobes. The pGEM-zprmt1 plasmid was linearized by EcoRI or SalI restriction enzyme and the RNA was transcribed with SP6 or T7 RNA polymerase to prepare the antisense or sense RNA probe respectively. The embryos were washed and incubated with anti-DIG antiserum and stained. Embryos were then mounted in 100% glycerol for observa- tion using a dissecting microscope (Zeiss AXioskop2; Carl PRMT1 expression and function during embryogenesis Y J. Tsai et al. 914 FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation ª 2011 FEBS [...]... A, Black MH, Talieri M & Diamandis EP (2000) Genomic organization, physical mapping, and expression analysis of the human protein arginine methyltransferase 1 gene Biochem Biophys Res Commun 278, 349–359 Goulet I, Gauvin G, Boisvenue S & Cote J (2007) Alternative splicing yields protein arginine methyltransferase 1 isoforms with distinct activity, substrate specificity, and subcellular localization... Chuang for fish rearing, cDNA preparation and WISH probe preparation References 1 Bedford MT & Clarke SG (2009) Protein arginine methylation in mammals: who, what, and why Mol Cell 33, 1–13 2 Lin WJ, Gary JD, Yang MC, Clarke S & Herschman HR (1996) The mammalian immediate-early TIS21 protein and the leukemia-associated BTG1 protein interact with a protein- arginine N -methyltransferase J Biol Chem 271, 15034–15044... USA) The full prmt1 coding region was prepared from RT-PCR with primers ASF and ZF1-R and cloned into the vector pCS2+ Rescue experiments were performed with prmt1 cRNA synthesized in vitro using the mMESSAGEmMACHINE kit in accordance with the manufacturer’s protocol (Ambion Europe Ltd, Huntingdon, UK) The 5¢ region recognized by the AMOs was further mutated using QuikChangeÒ II sitedirected mutagenesis... of the Oris Cell Migration Assembly Kit-FLEX (Platypus Technology, Madison, WI, USA) and cell migration assays were carried in accordance with the manufacturer’s instructions After the cells were allowed to attach for 10 h, well inserts were removed, and the cells were allowed to migrate into the clear field for 16 h Cells were fixed with formaldehyde, stained with crystal violet and photographed The. .. and in vitro methylation by protein arginine methyltransferases PRMT1 and PRMT3 J Biol Chem 274, 13229–13234 7 Nichols RC, Wang XW, Tang J, Hamilton BJ, High FA, Herschman HR & Rigby WF (2000) The RGG domain in hnRNP A2 affects subcellular localization Exp Cell Res 256, 522–532 8 Cote J, Boisvert FM, Boulanger MC, Bedford MT & Richard S (2003) Sam68 RNA binding protein is an in vivo substrate for protein. .. introduce mismatches aiming to avoid quenching in the rescue experiments by MO1 or MO2 with the primer sets (MO1 forward and MO1 reverse; MO2 forward and MO2 reverse; Table S1) Catalytic inactive PRMT1 with S69A, G70A and T71A mutations at the conserved AdoMet-binding site was designed as described by Balint et al [29] The point mutations were also introduced by site-directed mutagenesis to produce the catalytic... material is available: Fig S1 Expression of zebrafish prmt1 mRNA and protein during early development Fig S2 Defective phenotypes in prmt1 knock-down zebrafish at 72, 96 or 120 hpf Fig S3 Knockdown of prmt1 induces gastrulation defects Table S1 Primers used in the present study This supplementary material can be found in the online version of this article Please note: As a service to our authors and readers,... catalytic mutant by the primer set (SGT forward and SGT reverse; Table S1) Synthesized capped cRNA was co-injected with the AMO into the onecell stage embryos Cell culture and migration assay Huh-7 (human hepatocarcinoma cell line) cells were cultured at 37 °C in DMEM medium (Gibco, Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum PRMT1 expression and function during embryogenesis (Gibco), 1%... Science Council and CSMU 94-OM-A-023, 97-OM-A-136 and 98OMA-060 from Chung Shan Medical University The authors would like to express thanks to Drs Bon-chu Chung, Yi-Chuan Cheng and Shye-Jye Lee for valuable discussions and Dr Wen-Wei Chang for his valuable suggestions on the cell migration assay We also thank Yuling Lin, Pei-Hsin Chang, Hsiao-Yun Cheng, Li-Chun Tu and Han-Ni Chuang for fish rearing,... to our authors and readers, this journal provides supporting information supplied by the authors Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset Technical support issues arising from supporting information (other than missing files) should be addressed to the authors FEBS Journal 278 (2011) 905–917 ª 2011 The Authors Journal compilation . The predominant protein arginine methyltransferase PRMT1 is critical for zebrafish convergence and extension during gastrulation Yun-Jung Tsai 1 , Huichin. MO1 and MO2 indeed blocked the expression of PRMT1 protein in zebrafish embryos, both effectively and persistently. Because PRMT1 is the predominant type I protein arginine methyltransferase, the. 48 and 72 hpf as the source of protein arginine methyltransferase and recombinant mouse fibrillarin as the methyl-accepting protein. The samples were separated by SDS ⁄ PAGE and the methylated proteins