Proper Balance of Small GTPase rab10 Is Critical for PGC Migration in Zebrafish

MicroRNAs (miRNAs) play important roles in post-transcriptional repression in nearly every biological process including germ cell development. Previously, we have identified a zebrafish germ plasm-specific miRNA miR-202-5p, which regulates PGC migration through targeting cdc42se1 to protect cdc42 expression. However, knockdown of cdc42se1 could not significantly rescue PGC migration in maternal miR-202 mutant (MmiR-202) embryos, indicating that there are other target genes of miR-202-5p required for the regulation of PGC migration. Herein, we revealed the transcriptional profiles of wild type and MmiR-202 PGCs and obtained 129 differentially expressed genes (DEGs), of which 42 DEGs were enriched cell migration-related signaling pathways. From these DEGs, we identified two novel miR-202-5p target genes prdm12b and rab10. Furthermore, we found that disruption of rab10 expression led to significantly migratory defects of PGC by overexpression of rab10 siRNA, or WT, inactive as well as active forms of rab10 mRNA, and WT rab10 overexpression mediated migratory defects could be partially but significantly rescued by overexpression of miR-202-5p, demonstrating that rab10 is an important factor involved miR-202-5p mediated regulation of PGC migration. Taken together, the present results provide significant information for understanding the molecular mechanism by which miR-202-5p regulates PGC migration in zebrafish.

Like Dictyostelium discoideum or Entamoeba histolytica, zebrafish PGC migration utilizes the term of bleb, which is driven by hydrostatic pressure and cytoplasmic flow to emerge cells detaching from the actin cortex and forming a round protrusion in the cell front [21]. PGC migration requires high contractility, flexible protrusion formation and reduced adhesion levels. Firstly, the onset of PGC migration coincides with a decreased level of E-cadherin controlled by the regulator of G-protein signaling 14a protein [22]. Then, the morphology and motility of PGCs is maintained by three molecular modules including myosin light chain kinase (Mlck), Zeb1 and A5b, through elevating the cell contractility, The expression level was calculated using FPKM (fragments per kilobase million mapped reads). To yield reliable results, only genes with FPKM ≥ 1 in at least three samples were subjected to analyze differentially expressed genes (DEGs). We obtained 129 DEGs including 46 up-regulated and 83 down-regulated DEGs in MmiR-202 PGCs compared to WT PGCs ( Figure 1A,B). These DEGs were distributed in 39 GO terms and were mostly enriched in cell part, cellular process, biological regulation and binding ( Figure 1C). In summary, we established the transcriptional files of WT and MmiR-202 PGCs.

Analysis of the DEGs Associated with PGC Migration
To explore genes involved in miR-202-5p mediated PGC migration, we performed bioinformatic analysis to predict functions of these DEGs and found that 42 DEGs were enriched in cell migration associated biological processes including extracellular matrix, cell polarity, cytoskeleton, cell adhesion, cell motility and small GTPase activity (Table 2, Figure 2A). Among these DEGs, eight DEGs were involved in the regulation of GTPase activity, a process essential for the regulation of PGC polarity [27]; seven DEGs were associated with the regulation of cadherin binding, and proper regulation of cadherin was required for the onset of PGC migration [42]. In addition, there were seven DEGs participating in assembly of myosin such as myl7, tpm4b, rab36 and cmlc1, bleb formation panx1a as well as two positive regulator associated cell motility cavin1a and rab36, and the transcriptome data indicated that their expression levels were down-regulated in MmiR-202 PGCs compared to WT PGCs.  The results were representative of more than three independent experiments in triplicate. Three reference genes ef1a, β-actin and rpl13a were used in every PCR examination to normalize gene expression levels with 2 −∆∆Ct method. The statistics were calculated and analyzed by Student's t-test, * p < 0.05; ** p < 0.01; *** p < 0.001.
To validate the reliability of transcriptome data, we performed qPCR analysis of 12 representative DEGs including two PGC marker genes dnd and gra, four DEGs with over 10-fold change such ashmgxb4a, prdm12b and myl7, an ECM-related DEG megf6a, two cytoskeleton-related DEGs anxa1c and synpo2lb, three GTPase-related DEGs rab10, mcf2a and rab36, as well as cell adhesion-related DEG cldne. qPCR results exhibited a high consistency with the transcriptome data ( Figure 2B-M): the transcriptional levels of dnd, gra, myl7 and rab36 were significantly downregulated, whereas hmgxb4a, prdm12b, megf6a, rab10, mcf2a, anxa1c and synpo2lb were significantly up-regulated in MmiR-202 PGCs compared to WT PGCs. Hence, the transcriptome data demonstrated that maternal loss of miR-202-5p led to dysregulation of multiple PGC migration-related signaling pathways.

rab10 and prdm12b Are Novel miR-202-5p Target Genes
To identify novel miR-202-5p target genes associated with PGC migration, we screened whether the 3 untranslated region (UTR) of any DEG contained miR-202-5p binding sites by bioinformatic analysis. We found that the 3 UTR of rab10 and prdm12b contained the canonical seed sequences of miR-202-5p ( Figure 3A,B). Next, we employed luciferase and RNA reporter systems to clarify whether rab10 and prdm12b are subjected to miR-202-5p regulation. Firstly, we constructed recombinant psiCheck2 plasmids by linking WT or mutant 3 UTR of rab10 or prdm12b to the open reading frame (ORF) of luciferase and transiently transfected them to HEK293T cells. Dual luciferases assay indicated that the relative activity of luciferase fusing WT 3 UTR of prdm12b or rab10 was significantly decreased in miR-202-5p mimics transfected HEK293T cells, whereas mutation of miR-202-5p binding sites significantly attenuated the repression ( Figure 3C). Secondly, we employed an RNA reporter by linking the ORF of red fluorescent protein (RFP) with WT or mutant 3 UTR of rab10 or prdm12b and co-injected it with miR-202-5p mimics or control mimics into 1-cell embryos. Quantitative pixel-intensity analysis revealed a significant decrease of RFP in the presence of miR-202-5p mimics, whereas the decrease was abolished by the mutation of miR-202-5p binding sites ( Figure 3E,F). Thus, as previously shown for cdc42se1 and trim25 [15,43], the 3 UTRs of rab10 or prdm12b were responsible for conferring negative regulation on the translation of linked ORFs through the action of miR-202-5p. Consistently, qPCR analysis of the RNA extracted from miR-202-5p or control mimic-injected embryos revealed a significant decrease of endogenous rab10 or prdm12b in miR-202-5p mimic-injected embryos ( Figure 3D), meanwhile there was a significant elevation of rab10 or prdm12b in MmiR-202 PGCs ( Figure 2G,K). These data indicated that rab10 or prdm12b are miR-202-5p target genes.

The Balance of rab10 Is Critical of PGC Migration
To investigate the possible roles of rab10 and prdm12b in zebrafish PGC development, we employed loss-and gain-of-function by injection of corresponding siRNA and mRNA, respectively. Firstly, qPCR analysis on the extracted RNA from control or rab10 siRNA injected embryos revealed a significant reduction in levels of rab10 in rab10 siRNA injected embryos, and the reduction relied on the dosages of rab10 siRNA ( Figure 4A). Secondly, we employed the fluorescence reporter of rab10-rfp mRNA and co-injected it with control or rab10 siRNA into 1-cell embryos. Quantitative pixel-intensity analysis revealed a significant decrease in the levels of RFP in rab10 siRNA injected embryos compared to that in control siRNA injected embryos ( Figure 4B,C). These results demonstrated that rab10 siRNA is efficient to decrease endogenous rab10 expression in vivo.
Next, we investigated whether knockdown of rab10 affected PGC migration in the PGC fluorescence labeling transgenic line (kop: egfp-UTR-nos3). We observed a significant mislocalization of PGCs in rab10 siRNA injected embryos at 24 hpf ( Figure 4D-G). We observed that the proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs was 29.27% and 5.1, 34.88% and 6.4, 45.72% and 7.3 in 50, 150 and 300 pg rab10 siRNA injected embryos, respectively ( Figure 4N,O). Because the expression of rab10 was elevated in MmiR-202 PGC, we speculated that overexpression of rab10 also affected PGC migration. To verify it, ectopic overexpression of rab10 was performed by injection of WT rab10 mRNA. We observed that ectopic overexpression of rab10 led to a significantly migratory defect of PGCs. The proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs in 50, 100 and 200 pg rab10 mRNA was 37.0% and 6.5, 46.4% and 7.5, 59.5% and 8.6, respectively ( Figure 4H-K,N,O). Therefore, the migratory defect of PGC exhibited a significant dosage effect associated with the injected dosage of either rab10 siRNA or mRNA. Rab proteins alternate between GTP-and GDP-bound states, which functions as the molecular switch in regulation of cell trafficking and movement [44]. It has been reported that canine rab10 mutants rab10T23N and rab10Q68L prefer to maintain inactive GDP-and active GTP-bound state, respectively [45]. Homological comparison indicated that the two functional sites were conserved in zebrafish rab10 proteins. We constructed inactive and active mRNA mutants of rab10T23N and rab10Q68L and then injected them in 1-cell embryos, respectively. We observed that injection of either rab10T23N or rab10Q68L mRNA resulted in migratory defects in PGCs ( Figure 4L,M). The canonical miR-202-5p binding sites and its mutant were shown in red and blue, respectively. (C) Relative luciferase activity in HEK293T cells transfected with psiCHECK2−3 utr wt (prdm12b or rab10) or psiCHECK2-3 utr mut (prdm12b or rab10) in the presence of control mimics (con mimics) or miR-202-5p mimics. (D) qRT-PCR analysis of prdm12b and rab10 mRNA in 24hpf embryos injected with control or miR-202-5p mimics. (E) The representative images of embryos co-injected with mRNAs of rfp-prdm12b 3 utr wt, rfp-prdm12b 3 utr mut, rfp-rab10 3 utr wt or rfp-rab10 3 utr mut in the presence of control or miR-202-5p mimics. All embryos were co-injected with pEGFP-N3 vector as a control. (F) The relative signal ratio of RFP/GFP in the experiment presented in (E). (C,D,F) The results were representative of more than three independent experiments in triplicate. β-actin was used as an internal control to normalize gene expression levels with 2 −∆∆Ct method * p < 0.05; ** p < 0.01; *** p < 0.001. β-actin was used as an internal control to normalize gene expression levels with 2 −∆∆Ct method. ** p < 0.01; *** p < 0.001, **** p < 0.0001. Scale bar, 100 µm.
Furthermore, PGC-specific overexpression of rab10 by injection of rab10-rfp-utr-nos3 mRNA into 8-cell embryos led to a significant PGC mislocalization ( Figure 5A-C). In summary, the proper balance of rab10 ensures PGC migration in zebrafish.

miR-202-5p Negatively Regulates rab10 in PGC Migration
To clarify the regulatory relationship between miR-202-5p and rab10 during PGC migration, we injected rab10 siRNA or miR-202-5p mimics to rescue the migratory defect of PGCs in embryos injected WT rab10 mRNA. Compared to WT rab10 mRNA injected embryos, co-injection of rab10 mRNA with either rab10 siRNA or miR-202-5p mimics could significantly reduce both the proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs (Figure 6), although some embryos still had a modest number of mislocalized PGCs. Therefore, overexpression of miR-202-5p could partially rescue PGC migratory defects caused by overexpression of rab10.

Knockdown of rab10 Fails to Rescue PGC Migration in MmiR-202 Embryos
On the other hand, we investigated whether knockdown of rab10 could rescue the migratory defect in MmiR-202 PGC. Different dosages (50, 150 and 300 pg) of rab10 siRNA were injected into 1-cell MmiR-202 embryos. In 300 pg control siRNA injected WT embryos, there were 6.21% embryos with an average number (0.91) of mislocalized PGC ( Figure 7A). On the other hand, the proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs was 30.56% and 4.76 in control siRNA injected MmiR-202 embryos ( Figure 7B). However, the proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs increased to 36.36% and 6.22, 37.50% and 6.50, 43.59% and 7.29 in 50, 150 and 300 pg rab10 siRNA injected MmiR-202 embryos, respectively ( Figure 7C-E,G,H). Therefore, knockdown of rab10 in MmiR-202 embryos resulted in a severer defect in PGC migration compared to MmiR-202 embryos. Previously, we have found that miR-202-5p protects PGC migration by targeting and repressing the expression of cdc42se1 [15]. Therefore, we attempted to rescue PGC migration in MmiR-202 embryos by double knockdown of rab10 and cdc42se1. However, we observed that the proportion of embryos with mislocalized PGCs and average number of mislocalized PGCs further increased to 45.71% and 7.34 in 300 pg rab10 siRNA and 4000 pg cdc42se1 MO co−injected MmiR-202 embryos, demonstrating that double knockdown of rab10 and cdc42se1 led to a stronger migratory defect in the PGC migration of MmiR-202 embryos ( Figure 7F-H). Therefore, the migratory defect caused by maternal loss of miR-202-5p was complex.

Discussion
miR-202-5p is a conserved in animals, which plays an important role in reproduction [41], gametogenesis [46,47], immunology [43], tumorigenesis [48,49], and cell migration [15,50,51]. In mammals, miR-202 is specific to Sertoli cell and might play an important role in the interaction between germ cell and Sertoli cell during spermatogenesis [46,52]. In frog, miR-202-5p is a germline−enriched miRNA in early stage of oogenesis [53]. Chicken miR-202-5p is highly expressed in spermatogonia rather than PGCs and might regulate germ cell development by targeting LIMK2 [47]. The expression patterns of miR-202-5p are divergent among fish species. In Japanese medaka (Oryzias latipes), miR-202-5p is co-localized with sox9 in primordial follicles. Loss of miR-202 impairs early folliculogenesis, decreases the number of mature follicles and leads to a dramatic reduction in female fecundity [41]. In zebrafish, miR-202-5p is a component of germ plasm. Loss of miR-202 seems to not affect zebrafish fertility but results in a significant defect in PGC migration [16]. Previously, we have revealed a mechanism by which miR-202-5p uses to regulate PGC migration. miR-202-5p negatively regulates cdc42se1 to ensure the proper expression of cdc42 in PGCs. However, it remains vague whether there are other mechanisms for miR-202-5p regulating PGC migration by other target genes. In this study, we performed transcriptome sequencing of WT and MmiR-202 PGCs and obtained 146 DEGs. Furthermore, we found that rab10 acts as a downstream target gene of miR-202-5p and plays an important role in PGC migration. Therefore, we revealed a novel mechanism underlying miR-202-5p regulation of PGC migration.
Zebrafish PGC forms at about 4 hpf, initiates migration at 4.5 hpf and reaches the genital ridge at 24 hpf [54,55]. It is better to collect PGCs during the whole migratory process (4-24 hpf) to analyze the dynamic gene expression change in WT and MmiR-202 PGCs. However, because of lacking the PGC-specific transgenic miR-202 −/− line, it is difficult to collect MmiR-202 PGCs from the onset of PGC migration. The 3 UTRs of nos1, dnd and tdrd7 regulate differential localization between soma and PGCs [8,9,56]. However, in gfp-nanos 3 utr mRNA injected embryos, the PGC-specific fluorescence could be found at the late segmentation period, usually at 24 hpf. Therefore, we collected PGCs in gfp-nanos-3 utr mRNA injected WT and MmiR-202 embryos at 24 hpf to ensure the purity of isolated PGCs. The transcriptome results revealed that 42 DEGs in WT and MmiR-202 PGCs were significantly enriched in cell migration-associated biological processes ( Table 2). The motility of PGC requires a decrease of cadherin−mediated tight junction and a proper regulation of myosin-based contractility [3,23,26]. However, we found a remarkable elevation in the levels of tight junction-associated genes including cldne, itga9 and styk1b, but a reduction in the levels of myosin, polarity and bleb associated genes including myl7, tpm4b, rab36, cmlc1, panx1a, cavin1a, fgf13b and rab36 in MmiR-202 PGCs compared to WT PGCs. Chemokine-guided signaling pathway is essential for the direction of PGC migrating towards genital ridges [24,38]. We found two chemokineassociated genes chia.4 and ccl20a.3 were differentially expressed in WT and MmiR-202 PGCs, suggesting that maternal loss of miR-202 might disrupt chemokine signaling to affect PGC migration. In addition, many DEGs were enriched in the regulation of cell adhesion, cytoskeleton, ECM and GTPase activity. Together, the transcriptome files of WT and MmiR-202 PGCs provided global and useful data for understanding the potential signaling pathways involved in miR-202-5p regulation of PGC migration.
Because miRNAs rarely encode functional protein products, they usually initiate hierarchical biological events by binding to the mRNAs of target genes to direct post-transcriptional repression. Therefore, identifying target genes will be helpful to understand how miR-202-5p regulates PGC migration. From these DEGs, we identified two novel miR-202-5p target genes prdm12 and rab10. The members of PR domain-containing (PRDM) family are critical in regulation of DNA methylation and stem cell population maintenance. PRDM1 and PRDM14 are critical determinants of mammalian germ cell lineages from embryonic cells [57,58]. However, zebrafish PRDM family mainly regulate craniofacial development [59]. We found that overexpression or knockdown of prdm12b had no significant effect on PGC development in zebrafish (data not shown), suggesting PRDM family exert divergent roles in mammals and fish.
The members of small GTPase superfamily participate in almost every aspect of cell biology, which is comprised of five families conserved across eukaryotes: Ras, Rho, Rab, Arf, and Ran [60]. They alternate between GTP-and GDP-bound states, which act as molecular switches in regulation of polarity, cytoskeleton change, membrane invaginations in all eukaryotic cells [44,60]. The activity of RhoA/ROCK is essential for correct localization of germ plasm mRNAs at cleavage stage [28]. The Rac1 activation is essential for establishing the cell front and polarity of PGCs [61]. Cdc42 and ROCK regulate membrane invaginations, which is essential for motility of PGCs [62]. Inhibition of Rac/ROCK/Cdc42 activities by expression domain negative proteins significantly impairs PGC motility and leads to ectopic localization of PGCs. Rab proteins are the largest branch of Ras superfamily, which is critical in the regulation of vesicle formation and trafficking [44]. The Clathrinand Rab5-mediated endocytosis is required for activation of Rac, and the Rab5-to-Rac circuitry controls the morphology of motile PGCs in zebrafish [63]. Other members of Rab family participate in regulation of gametogenesis. Rab7 plays an important role in actin dynamics and mitochondria function in oocyte meiosis [64]. rab10 binds with the male germ cell-specific Rab GTPase-activating protein (MGCRABGAP) and the MGCRABGAP-RAB complex localizes in the manchette structure of mature sperm, indicating a possible role in spermatogenesis [65]. However, whether rab10 participates in PGC development remains unclear. Herein, we found that disruption of rab10 balance through siRNA mediated reduction, overexpression of WT rab10, inactive or active rab10 mutants led to PGC mislocalization, demonstrating that proper expression of rab10, especially for the balance of rab10 GDP-and GTP-bound states, is critical for PGC migration (Figure 4). Because rab10 is negatively regulated by miR-202-5p, we used miR-202-5p mimics to rescue PGC migratory defects in rab10 overexpressed embryos. However, overexpression of miR-202-5p partially rescued the migratory defects caused by rab10 overexpression (Figure 6), which might be because although miR-202-5p might rescue the expression level of rab10, the balance between rab10 GDP-and GTP-bound states remains in disorder. Taken together, miR-202-5p mediated balance of rab10 is important for PGC migration.
Finally, we observed that knockdown of rab10 could not significantly rescue PGC migration and even led to a severer migratory defect of PGCs in MmiR-202 embryos, no matter knockdown of cdc42se1 or not. There might be two possible explanations for why we failed to rescue PGC migration by knockdown of rab10 and cdc42se1 in MmiR-202 embryos. Firstly, injection of rab10 siRNA could rescue the expression level of rab10 but might cause a more significant disorder in the balance between rab10 GDP-and GTP-bound states in MmiR-202 embryos. Secondly, loss of miR-202-5p led to a complex dysregulation of downstream signaling networks besides rab10 and cdc42se1, such as the chemokine signaling. Chemokine-guided cell migration is essential for directing PGC migration towards the genital ridge [38]. Herein, we found that the ccl20a.3 chemokine was differentially expressed between WT and MmiR-202 PGCs, providing possible evidence suggesting that miR-202-5p might regulate PGC migratory behavior via chemokine signaling. However, which type of cell expressing ccl20a.3 chemokine receptor-ccr6 during early embryogenesis; whether and how ccr6 positive cells contribute to PGC migration towards the genital ridge needs to be further investigated. In summary, the present work provided transcriptional profiles of MmiR-202 PGCs, revealed the differentially expressed signaling pathways between WT and MmiR-202 PGCs, and identified two novel miR-202-5p target genes rab10 and prdm12b. Loss-and gain-of-function demonstrated that rab10 is required for PGC migration. The present work provides important data for understanding the molecular mechanism underlying miR-202-5p regulation of zebrafish PGC.

Ethics Statement
All procedures with zebrafish were approved by the Ethics Committee of Sun Yat−sen University (31771587, 20 February 2017), and the methods were carried out in accordance with the approved guidelines.

Zebrafish Strains and Cell Lines
The AB line and transgenic line Tg kop: egfp-UTR-nos3 zebrafish were purchased from the China Zebrafish Resource Center. The maternal miR-202 mutant (MmiR-202) zebrafish has been generated in our previous study [15]. Fish were raised at 28 • C with 10 h (h) darkness and 14 h light in the zebrafish circulatory breeding system. HEK 293T cells were cultured at 37 • C in Dulbecco's modified Eagle's medium (DMEM) supplied with 10% FBS (Invitrogen, Carlsbad, CA, USA) under a humidified atmosphere of air containing 5% CO 2 .

Isolation of PGC and Transcriptome Analysis
To label the PGCs with green fluorescence protein (GFP), 400pg gfp−nos 3 utr mRNA was injected into 1-cell WT and MmiR-202 embryos. At 24 hpf, the embryos were digested to single−cell suspension by trypsin according to the zebrafish book (http://zfin.org/zf_ info/zfbook/zfbk.html, 4th version, accessed date: 7 October 2021). A total of 900 WT and 900 MmiR-202 PGCs were collected by FACS based on the GFP signal. The isolated PGCs were equally divided into three technical replicates of WT PGCs and MmiR-202 PGCs groups for Illumina sequencing on the MGIseq2000 platform (Anoroad genome, Beijing, China).

Transcriptome Sequencing and Bioinformatic Analysis
After sequencing, clean reads were obtained with Perl scripts to clear raw read containing more than 5 adapter-polluted bases, the number of N bases accounting for more than 5%, as well as the phred quality value less than 19 accounting for more than 15%. Clean reads were mapped to the reference genome and annotation file downloaded from Ensembl (Danio rerio GRCz10.89.gtf) with HISAT2 [66]. Read Count for each gene in each sample is counted by HTSeq and gene expression level was calculated using FPKM (Fragments Per Kilobase Millon Mapped Reads). To yield reliable results, genes with FPKM ≥ 1 in at least three samples were taken for differential expression analysis between WT and MmiR-202 PGCs with DEGseq with BH method (false discovery rate correction with Benjamini-Hochberg), and genes with q ≤ 0.05 and |log2_ratio| ≥ 1 were identified as differentially expressed genes. The biological signaling pathway of DEGs was analyzed by functional enrichment in GO terms (Gene Ontology, http://geneontology.org/, 1 January 2021) and KEGG pathways (Kyoto Encyclopedia of Genes and Genomes, http://www.kegg.jp/, Release 97.0, 1 January 2021). GO terms with q < 0.05 were considered significantly enriched.

Embryo Collection and RNA Isolation
For RNA isolation, WT embryos, embryos injected control or miR-202-5p mimics, control siRNA or different dosages (50/150/300 pg each embryo) of rab10 siRNAs, as well as embryos injected with gfp mRNA or different dosages (50/100/200 pg each embryo) of rab10 mRNA were collected at 24 hpf, frozen in liquid nitrogen and stored in −80 • C. The collected embryos were homogenized in 1 mL of trizol reagent (Invitrogen, Carlsbad, CA, USA), and total RNA was extracted according to the manufacturer's instructions.

Reverse Transcription and qPCR Analysis
To examine the expression levels of predicted DEGs between WT and MmiR-202 PGCs, the RNA of collected PGCs was reverse-transcribed with single-cell sequence-specific amplification kit to establish first the cDNA library (Vazyme, Nanjing, China). The extracted RNA of other embryos was reverse−transcribed into single-strand cDNA with primescript™ first-strand cDNA synthesis kit (Takara, Kyoto, Japan). To avoid the error caused by single reference genes in single-cell PCR validation, we used three common reference genes including ef1a, β-actin and rpl13a to normalize gene expression levels using 2 −∆∆Ct method in every PCR examination. The results were representative of more than three independent experiments in triplicate. Each independent experiment was performed in triplicate.
qPCR analysis was performed on the LightCycler480 II (Roche, Basel, Switzerland) with the chamQ universal SYBR qPCR master mix (Vazyme, Nanjing, China) as previously described [67]. Primer sequences were listed in Table A1.

Plasmid Construction and mRNA Transcription
For luciferase or RFP reporters, the 3 UTRs of rab10 (ENSDARG00000046106) or prdm12b (ENSDARG00000007430) were cloned and fused to the ORF of luciferase and RFP in the psiCHECK2 and pCS2 plus (pCS2+) vectors, respectively.
For universal rab10 overexpression, full ORF and 3 UTR of rab10 was inserted to pCS2+ vector. To specifically overexpress rab10 in PGCs, the ORF of rab10 and gfp as well as the 3 UTR of nanos3 were inserted to pCS2+ vector to generate the pCS2-rab10-gfp-nos 3 utr plasmid.
All the point mutants including miR-202-5p binding sites within the 3 UTRs of rab10 or prdm12b, and the inactive or active rab10 mutants were constructed with hieff mut tm multi-sitedirected mutagenesis kit (Yeasen, Shanghai, China) as previously described [15,45].
For mRNA synthesis, recombinant pCS2+ plasmids were linearized with NotI and transcribed using SP6 mMESSAGE mMACHINE Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.
To rescue PGC migration in embryos injected rab10 mRNA or rab10 siRNA, 50 pg rab10 siRNA or 100 pg rab10 mRNA were co-injected into 1-cell embryos of Tg kop: egfp-UTR-nos3 line, and PGC migration was observed at 24 hpf. Furthermore, 300 pg miR-202-5p mimics were co−injected to rescue PGC migration in rab10 mRNA injected embryos.
All the PGC migration was examined at 24 hpf with the stereo fluorescence microscope (SMZ800N, Nikon, Kyoto, Japan). All experiments were performed in more than three independent experiments in triplicate. Each independent experiment was performed in triplicate.

PGC Phenotype Observation
At 24 hpf, PGCs of different groups of embryos were observed with the stereoscopic fluorescence microscope. The delineation of normal and mislocalized phenotype were distinguished as following: embryos with less than 3 ectopic PGCs were regarded as normal, whereas embryos with more than 3 ectopic PGCs were regarded as mislocalization. At least four repeated groups of embryos (30-50 embryos in each group) were observed.

Statistics Analysis
The statistics were calculated and analyzed with SPSS version 20. The qPCR results were analyzed by Student's t-test, and the phenotypes in different experimental groups were analyzed by one-way ANOVA. p < 0.5 was considered a statistically significant difference.

Conflicts of Interest:
The authors have declared no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.  Table A1. Sequences of primer used in this paper.