regeneration factors expressed on myeloid expression in macrophage-like cells is required for tail regeneration in Xenopus laevis tadpoles Open Access

Xenopus laevis tadpoles can regenerate whole tails with muscles, notochord and spinal cord in one week (Beck et al., 2009; Slack et al., 2008). It has been suggested that, after tail amputation, lineage-restricted progenitor cells, which are produced from stem cells of each type of tissue differentiate into their own tissues to form a regenerated tail (Cesare and Slack, 2004). In our previous study, we found that il11, the expression of which is induced after tail amputation, is required and sufficient for producing several types of progenitor cells; thus, it is essential for tail regeneration (Tsujioka et al., 2017). We found that even progenitor cells in which il11 receptors are knocked out contribute to the formation of the regenerated tail, suggesting that the progenitor cell induction function of Il11 is elicited in a non-cell autonomous manner (Suzuki et al., 2022). These findings led us to hypothesize that Il11 induces progenitor cells via a specific cell type that supports tail regeneration in response to Il11. In this study, we aim to identify the putative cell type by screening genes that function in the putative cells.

To identify the putative supporting cells, we searched for genes thought to function in the cells using published RNA-sequencing (RNA-seq) data of amputated tails from il11 knockdown (KD) tadpoles (Tsujioka et al., 2017) in order to screen for genes that function in response to Il11. We obtained 1266 genes that were downregulated in il11 KD groups. Next, to screen for genes that function during tail regeneration but not during development, we analyzed the RNA-seq data of proliferating cells from developing tailbuds and proliferating cells and non-proliferating cells from regeneration buds (Tsujioka et al., 2015), and narrowed down the candidates to seven genes that are rarely expressed in the proliferating cell fraction of the tailbud. Next, to screen for genes expressed in the putative supporting cells other than the proliferating progenitor cells, we further narrowed down the candidates to those with higher expression in the non-proliferating cell fraction than in the proliferating cell fraction and identified four genes (Table S1). Among them, we focused on Xetrov90002578m.L and Xetrov90002579m.S because their functions have not been reported and they are rarely expressed in early development (Session et al., 2016) (Fig. S1A). X. laevis is an allotetraploid species and many genes have homologous genes on L and S chromosomes (Session et al., 2016). We identified Xetrov90002578m.L and Xetrov90002579m.S to be homologous, as they have very similar sequences and are located on homologous L and S chromosomes, respectively, with conserved synteny (Fig. S1B). Hereafter, we refer to them as regeneration factors expressed on myeloid.L (rfem.L) and rfem.S, respectively. There is a putative rfem.L and rfem.S orthologue in Xenopus tropicalis (accession number KAE8633781), a closely related species of X. laevis, but no orthologues have been identified in fish and rodents. The functions of Rfem.L and Rfem.S were not predictable from their amino acid sequences because there were no remarkable domains, although they showed weak similarity to γ-crystallin, and have no putative signal peptide sequences, suggesting their intracellular function.

Temporal expression analysis of tail stumps (Fig. 1A) revealed that the expression level of rfem.L and rfem.S at the tail stump became significantly higher at 1 to 3 days post amputation (dpa) than that at 0 dpa (Fig. 1B). The temporal expression pattern correlated well with the formation of the regeneration bud, which forms within 2 dpa (Beck et al., 2003). Expression of rfem.L and rfem.S is decreased at 5 to 7 dpa (Fig. 1B), when tail regeneration is almost complete. These results indicate that the expression of rfem.L and rfem.S is induced by tail amputation and correlates well with the progression of tail regeneration. We performed in situ hybridization on sections of intact and 3 dpa tadpoles, which revealed that rfem.L- and rfem.S-expressing cells were scattered around the regeneration bud and adjacent regions at 3 dpa (Fig. 1C), but were rarely observed in intact tadpole tails (Fig. 1D). We also observed scattered rfem.L- and rfem.S-expressing cells around the head and abdomen in both intact and amputated tadpoles (Fig. S1C). The number of rfem.L- and rfem.S-expressing cells at the regeneration bud and adjacent regions at 1 dpa increased significantly compared with the corresponding region in the intact tail (Fig. 1E), it is possible that the upregulation of rfem.L and rfem.S expression after tail amputation is due to the migration of rfem.L- and rfem.S-expressing cells to the amputation site or regeneration bud.

Next, we examined the function of rfem.L and rfem.S in tail regeneration by KD using the CRISPR/Cas9 system. We injected cas9 mRNA and guide RNAs (gRNAs) that target the coding sequence of rfem.L and rfem.S into one-cell stage embryos to introduce mutations in the rfem.L and rfem.S sequences. Tadpoles obtained by this method are mosaics of gene-edited and unedited cells because gene editing occurs in each cell independently after cleavage. We refer to the mosaic tadpoles as KD. To improve the efficiency of KD by gene editing, two gRNAs with different target sites were used for rfem.L and rfem.S, respectively (Fig. S2A). Tadpoles injected with cas9 mRNA and gRNAs targeting tyrosinase, a gene responsible for albinism and unnecessary for tadpole survival, were used as controls (tyr KD). Injected embryos were maintained for 4 days (stage 41), and normally developed tadpoles were used for the experiments (Fig. 2A and Fig. S2B). The development rates did not differ between the rfem.L and rfem.S KD and tyr KD groups (Table S2). To confirm that gene editing had occurred, some of the rfem.L and rfem.S KD tadpoles were randomly selected for Inference of CRISPR Edits-Analysis (ICE analysis) (Hsiau et al., 2018 preprint), a method to estimate the proportion of cells with frameshifts or insertions/deletions ≥21 bp as a KO score. At each gRNA target site of rfem.L and rfem.S KD tadpoles, gene editions that would result in loss of function were detected (Fig. S2C). The KO scores did not differ significantly between rfem.L and rfem.S KD tadpoles with developmental abnormalities and those with normal development (Fig. S2C), suggesting that the developmental abnormalities exhibited by some rfem.L and rfem.S KD tadpoles were not due to rfem.L and rfem.S KD. Considering that rfem.L and rfem.S are rarely expressed in early development (Fig. S1A), we estimate that the effect of rfem.L and rfem.S KD on development is limited. We amputated the tails of injected tadpoles at 4 days post-fertilization (4 dpf), and the regenerative ability depending on the morphology of the regenerated tails was evaluated at 7 dpa. Several tadpoles in the rfem.L and rfem.S KD groups showed defects of tissues or regenerated tails that were bent in strange directions (Fig. 2B and Fig. S2D), and the rfem.L and rfem.S KD groups exhibited a reduced regeneration rate compared with the tyr KD groups (Fig. 2C and Fig. S2E, and Table S2). Measurement of the area and length of regenerated tail revealed that the rfem.L and rfem.S KD tadpoles regenerated significantly smaller and shorter tails (Fig. 2D,E and Fig. S2F-H). These results indicated that rfem.L and/or rfem.S are required for normal tail regeneration. The morphology of regenerated tails of rfem.L and rfem.S KD tadpoles remained abnormal as they grew (Fig. S2J) and the tadpoles showed poor regeneration outcomes after re-amputation (Fig. S2K).

To estimate the function of rfem.L and rfem.S, we performed RNA-seq of tail stumps of amputated rfem.L and rfem.S KD tadpoles. The RNA-seq detected six genes that were significantly downregulated in rfem.L and rfem.S KD tail stumps (Table S3), including three hemoglobin genes, suggesting that rfem.L and rfem.S KD affects wound repair, including blood vessel formation. We also attempted to identify the cell types that express rfem.L and rfem.S using scRNA-seq on enzymatically dispersed intact tails and 2 dpa regeneration buds (Fig. S3A). Cell clusters that reflect most of the tissues constituting the tail, such as the epidermis, muscles, notochord, nerves and blood cells, were detected (Fig. 3A and Fig. S3B,C), suggesting that the data covered a broad range of tissues in the tail. rfem.L and rfem.S were expressed in a few cells, and their expression was well correlated (Fig. 3B). The rfem.L- and rfem.S-expressing cell fraction was found in both intact tails and regeneration buds (Fig. S3D), and the fraction belonged to a cluster of leukocytes, indicating that rfem.L- and rfem.S-expressing cells are a type of leukocyte. We examined genes that are characteristically expressed in the fraction and found that expression of c1qc.L, csf1r.S, trem2.S, art5.L and crp.4.L was enriched (Fig. 3C). In particular, the expression of c1qc.L and art5.L correlated well with that of rfem.L and rfem.S, and these two genes were identified in the first screening of this study (Table S1). Complement component 1 subcomponent C chain (C1q) is produced in tissue phagocytes such as macrophages and dendritic cells (Thi et al., 2017). colony stimulating factor-1 receptor (csf1r) and triggering receptor expressed on myeloid cells 2 (trem2) are known to be expressed on myeloid cells such as macrophages and dendritic cells (Stanley and Chitu, 2014; MacDonald et al., 2005; Hickman and El, 2014). These results revealed that rfem.L- and rfem.S-expressing cells had a macrophage-like gene expression profile.

Because rfem.L and rfem.S are identified as genes that are downregulated in il11 KD tadpoles (Fig. S3E), we assessed whether the rfem.L- and rfem.S-expressing cells respond to Il11 by analyzing their il11 receptor subunit alpha (il11ra.L) expression, and found that il11ra.L expression was not detected (Fig. S3F). Therefore, it is possible that rfem.L- and rfem.S-expressing cells do not receive Il11 directly and thus are regulated indirectly by Il11 signaling. We could not, however, exclude the possibility that the rfem.L- and rfem.S-expressing cells express il11ra.L under the detection threshold because of the relatively low sensitivity of detection of gene expression by scRNA-seq.

To assess the relationship between tail regenerative capacity, and rfem.L and rfem.S expression in leukocytes, we examined the correlation between the area of the regenerated tail and the KO score in the peripheral blood cell (PBC) fraction of rfem.L and rfem.S KD tadpoles, but failed to detect a tendency for the regenerated area to be smaller in tadpoles with higher KO scores in the fraction (Fig. S2I), possibly because of the small proportion of rfem.L- and rfem.S-expressing leukocytes in PBCs. To investigate the role of the rfem.L- and rfem.S-expressing leukocytes in tail regeneration directly, we attempted to examine the effects of depletion of rfem.L- and rfem.S-expressing leukocytes on tail regeneration. The major fraction of rfem.L- and rfem.S-expressing cells co-expressed csf1r.S (Fig. 3C and Fig. S4A), a receptor for colony stimulating factor 1 (csf1). Csf1 is a cytokine that is essential for differentiation and survival of monocytes and macrophage lineage cells in mammals (Nakamichi et al., 2013), and Csf1 is also reported to function in X. laevis (Grayfer and Robert, 2013). We therefore assumed that csf1 KD depletes Csf1-dependent cells, including the rfem.L- and rfem.S-expressing leukocytes. X. laevis has only one csf1 gene on the genome (csf1.S) and we designed two gRNAs (gRNA#4 and gRNA#5; Fig. S4B) for csf1.S and generated csf1 KD tadpoles using the CRISPR/Cas9 system. Tadpoles injected with only cas9 mRNA were used as controls (cas9 groups). ICE analysis confirmed that gene editing was successful at the target site of each gRNA (Fig. S4C). csf1 KD did not decrease normal development rates (Table. S4). We quantified the rfem.L and rfem.S expression levels (Fig. 4A), and the number of rfem.L- and rfem.S-expressing cells (Fig. 4B) in the tail stumps of csf1 KD tadpoles, and found that both were significantly decreased, suggesting that csf1 KD depleted the rfem.L- and rfem.S-expressing leukocytes. We also quantified genes with expression patterns similar to that of rfem.L and rfem.S in the scRNA-seq; expression levels of c1qa.L, c1qb.L, c1qc.L, art5.L and crp.4.L (Fig. S4D, E); and the number of c1qa.L-expressing cells (Fig. S4F). We and observed a significant reduction or a trend towards reduction of the expression levels and number in the tail stumps in csf1 KD tadpoles. The lack of a statistically significant reduction in the cell number may be due to the difficulty in detecting clqa.L-expressing cells because of their low expression level. Several csf1 KD tadpoles showed tail regeneration defects with the tails becoming thinner and bending in strange directions (Fig. S4G), and the regeneration rates of csf1 KD groups were lower or tended to be lower than those of the control groups (Fig. S4H and Table S4).

To confirm that rfem.L and/or rfem.S expression in leukocytes is required for regeneration, we performed cell type-specific rescue experiments using the zebrafish mpeg1 promoter, which is reported to work in X. laevis macrophage-like cells (Paredes et al., 2015). We designed constructs that express rfem.L and acgfp1 (Aequorea coerulescens gene encoding green fluorescent protein) bicistronically, or only acgfp1 under control of the mpeg1 promoter (mpeg1:rfem and mpeg1:gfp respectively, Fig. S4I) to perform macrophage-like cell-specific rescue in rfem.L and rfem.S KD tadpoles (Fig. 4C). Scattered gfp-expressing cells were observed in regenerating tails of tadpoles injected with the construct (Fig. 4D and Fig. S4J), suggesting that the construct worked as expected. Compared with control tadpoles, the area and length of the regenerated tails in the mpeg1:rfem-rescued 7 dpa rfem.L and rfem.S KD tadpoles were significantly recovered (Fig. 4E,F and Fig. S4K,L), indicating that rfem.L or rfem.S expression in macrophage-like cells has an indispensable role in successful tail regeneration.

In this study, we identified rfem.L and rfem.S as genes required for tail regeneration, and indicated that rfem.L and rfem.S have an essential role in the regeneration-promoting function of macrophage-like cells. Leukocytes are recruited to the wound site to contribute to wound healing by killing pathogens and producing various cytokines (Koh and DiPietro, 2011; Wynn and Vannella, 2016). On the other hand, an excessive inflammatory response can lead to regenerative failure. In X. laevis, it is suggested that there are inflammatory and reparative myeloid cells that impair and promote tail regeneration, respectively (Aztekin et al., 2020). The rfem.L- and rfem.S-expressing macrophage-like cells identified by this study define a (sub)population of these reparative myeloid populations. We propose that rfem.L and rfem.S provide a clue to unveiling the cellular and molecular mechanisms that promote regenerative ability in leukocytes.

X. laevis adults were purchased from domestic breeders (Hamamatsu Seibutsu Kyouzai, Shizuoka, Japan; Watanabe Zoushoku, Hyogo, Japan). Tadpoles were obtained by mating X. laevis adults or by artificial fertilization. Adult frogs and tadpoles were kept at 20°C. Surgical procedures such as dissections and amputations were performed under 0.02% ethyl 3-aminobenzoate methanesulfonate salt (MS-222, Millipore Sigma) or ice anesthesia. Tail amputation was performed at 4 days post-fertilization (4 dpf; stage 41). Tadpoles were randomly assigned to experimental groups. Animals used in this study were grown under identical conditions without blinding. Animal experiments were performed in accordance with the Guidelines for Proper Conduct of Animal Experiments of Science Council of Japan. The protocols of experiments using living modified organisms in this study were approved by the Committee on Living Modified Organism Experiments of the Graduate School of Science at the University of Tokyo (DNA Exp 17-3).

No statistical methods were used to predetermine the sample size.

To quantify the rfem.L- and rfem.S expression in the tail stumps after tail amputation, we amputated tadpole tails at 4 dpf and sampled four batches of 20 tail stumps at 0, 1, 3, 5 and 7 dpa. Total RNA was extracted using an RNeasy Mini Kit (Qiagen) and Micro Smash MS-100 (Tomy Seiko). cDNAs were synthesized from the same amounts of total RNAs per sample using a PrimeScript RT reagent Kit with gDNA Eraser (Perfect Real Time) (Takara). A group without reverse transcription (RT-) was also prepared as a control. Real-time-PCR was performed with SYBR premix ExTaq II (Takara) or TB Green Ex Taq II (Tli RNaseH Plus) (Takara) and LightCycler 480 (Roche). Sequences of primers used are shown in Table S6. Because the sequences of rfem.L and rfem.S are quite similar and it was difficult to design primers specific to rfem.L or rfem.S, respectively, and the amplicon sizes are within 80-150 bp for qRT-PCR, we designed primers for the consensus sequences of rfem.L and rfem.S to amplify both of them. The threshold cycle was calculated using the 2nd derivative maximum method. The relative expression levels were normalized to those of elongation factor 1 alpha. We electrophoresed the PCR products and confirmed that no amplification was detected in the RT− group. Tukey-Kramer's test was performed using R (v3.6.3).

To quantify the gene expression in csf1 KD tadpoles, we amputated the tails of csf1 KD and cas9 tadpoles, and sampled three batches of 12 tail stumps at 1 day post amputation, and qRT-PCR was performed as described above.

For cloning the rfem.L and rfem.S partial sequences, cDNA was synthesized from the total RNA of thymi of X. laevis J strain stage 52 tadpoles with TRIzol reagent (Thermo Fisher Scientific) and SuperScript III Reverse Transcriptase (Thermo Fisher Scientific); then, the partial sequences of rfem.L and rfem.S were amplified by PCR with the cDNA, and the PCR products were cloned into a pGEM-T Easy vector (Promega). Sequences of the primers used are shown in Table S6. To synthesis RNA probes, the cloned vectors were amplified by PCR, and then PCR products were purified using FastGene Gel/PCR Extraction Kit (Nippon Genetics). Digoxigenin (DIG)-labeled RNA probes were synthesized using DIG RNA labeling Mix (Roche) and T7 or SP6 RNA polymerase (Roche) with the purified PCR products as templates. The probes were purified by ethanol precipitation. We synthesized probes for rfem.L and rfem.S, and the synthesis was confirmed by denaturing gel electrophoresis.

In situ hybridization was performed as described previously (Tsujioka et al., 2017; Hatta-Kobayashi et al., 2016) with several modifications. Intact 4 dpf tadpoles, and tadpoles amputated at 4 dpf at 3 dpa were fixed in MEMFA fixative [0.1 M MOPS (pH 7.4), 2 mM EGTA (pH 8.0), 1 mM MgSO₄ and 10% formalin) at room temperature for 48 h. They were dehydrated in an ethanol and Clear Plus (Falma) series and embedded in Paraplast plus (Leica); 8 µm sections were prepared and adhered to MAS-coated slide glass (Matsunami Glass). Sections were deparaffinized and rehydrated with a Clear Plus and ethanol series, permeabilized with 20 µg/ml proteinase K (Roche) and 1% Triton X-100, followed by acetylation with 0.25% acetic anhydride. The mixture of DIG-labeled rfem.L and rfem.S probes was hybridized in buffer containing 50% formamide, 5× SSC (75 mM trisodium citrate and 750 mM NaCl), 240 µg/ml torula yeast RNA (Millipore Sigma), 500 µg/ml salmon sperm DNA (FUJIFILM Wako Chemical) and 50× Denhardt's solution (FUJIFILM Wako) at 58°C for 16 h. Slides were washed with 0.2× SSC at 58°C. Sections were incubated with 1/5000 diluted alkaline phosphatase-conjugated anti-DIG antibody (catalogue code 11175041910, lot 44053100, Roche). Signals were detected with NTM buffer [100 mM Tris (pH 9.5), 100 mM NaCl and 50 mM MgCl₂] supplemented with 1/100 volume of NBT/BCIP stock solution (Roche). The endogenous melanin pigments were bleached by placing the slides in 0.5× SSC containing 0.3% H₂O₂ under fluorescent light overnight. The sections were mounted with 50% glycerol. The slides were observed using a differential interference microscope.

KD experiments using the CRISPR/Cas9 system were performed essentially as described previously (Tsujioka et al., 2017; Kato et al., 2021). Guide RNAs were synthesized as follows. The rfem.L, rfem.S, tyr.L and tyr.S sequences were obtained from Xenbase (RRID:SCR_003280; Xetrov90002578m.L for rfem.L and Xetrov90002579m.S for rfem.S, respectively). The csf1.S sequence (|JX418294.1) (Grayfer and Robert, 2013) was obtained from the NCBI (https://www.ncbi.nlm.nih.gov/). The guide RNAs were designed with CRISPRdirect (http://crispr.dbcls.jp/) (Naito et al., 2015). Two guide RNAs per gene were designed to improve the KD efficiency. The target sequences are shown in Table S6. The guide RNA #1 targets the same sequences of rfem.L and rfem.S (Fig. S2A). The target sequences were inserted into the DR274 plasmid (Hwang et al., 2013) (for gRNAs #1, #2 and #3) or the DR274-T7dG1 plasmid [for gRNAs #4 and #5; modified DR274 plasmid was used for synthesis of gRNAs that have one guanine at the 5′ end; DR274 was digested with BsaI (New England Biolabs) and the GAGACCGAGAGAGGGTCTCA sequence was inserted downstream of the T7 promoter to generate the DR274-T7dG1], and guide RNA was synthesized using an AmpliScribe T7-Flash Transcription Kit (Lucigen) then purified with an RNeasy Mini Kit. The synthesis was confirmed by denaturing gel electrophoresis.

For cas9 mRNA synthesis, pXT7-hcas9 (China Zebrafish Resource Center) (Chang et al., 2013) was digested with Xba I (Takara) and then mRNA was synthesized using mMESSAGE mMACHINE T7 ULTRA Kit (Thermo Fisher Scientific) and purified with an RNeasy Mini Kit. The synthesis was confirmed by denaturing gel electrophoresis.

Artificial fertilization was performed as follows. Unfertilized eggs were obtained by squeezing an adult female frog that was injected with 500 units of gonadotropin (ASKA Pharmaceutical) the previous night. A sperm suspension was prepared by shredding a part of a testis in 0.33× De Boer solution (18.3 mM NaCl, 217 μM KCl, 73 μM CaCl₂ and 695 μM NaHCO₃). Eggs were fertilized by mixing with sperm suspension. Fertilized eggs were dejellied with 3% cysteine (pH 7.6) and chilled to 12°C in 0.1×MMR [10 mM NaCl, 200 µM KCl, 100 µM MgSO₄, 200 µM CaCl₂, 500 µM HEPES (pH 7.4)].

We injected 18.4 nl of the guide RNA and cas9 mRNA solution into the animal hemisphere of healthy one-cell stage embryos in 0.1×MMR supplemented with 2% Ficoll PM 400 (Cytiva) at 12°C using a Nanoject II (Drummond Scientific Company). Concentrations of cas9 mRNA and guide RNA were as follows: guide RNA, 80 ng/µl each; cas9 mRNA, 700 ng/µl for rfem.L and rfem.S KD, and 1000 ng/µl for csf1 KD. Embryos of control group were injected with guide RNA#6 and #7 (tyr KD) and cas9 mRNA at the same concentration as the experimental group. Injected embryos were kept separately in 96-well plates (one embryo/well) in 0.1× MMR at 12°C for 24 h and then reared at 20°C.

Rates of normally developed tadpoles were counted at 3 dpf (stage 36-39), then healthy tadpoles with a normal morphology and motility were transferred to 0.1× Steinberg solution [5.8 mM NaCl, 67 µM KCl, 34 µM Ca(NO₃)₂, 83 µM MgSO₄ and 300 µM HEPES (pH7.4)] in a 90 mm dish. We used any mRNA and gRNA-injected, surviving and healthy tadpoles at 4 dpa for each experiment. Tadpole tails were amputated at 4 dpf, maintained for 7 days and then classified into two groups depending on the morphology of the regenerated tails, as follows: perfect, regenerated tail with normal morphology of muscle, notochord and fin; imperfect, no tail regeneration or regenerated tail with defects in any tissues and/or vent axis. Dead tadpoles or tadpoles with severe malformation of the whole body were excluded from counting. Statistical significance of the ratio of perfect/imperfect was assessed by two-sided Fisher's exact test using R (v3.6.3). Measurement of the regenerated tails were performed as described previously (Suzuki et al., 2022); we took photos of the regenerated tails of tadpoles and the parameters (area and length of regenerated tail) were measured using Fiji (Schindelin et al., 2012).

To estimate the KO scores, genomic DNA was extracted from the whole body, amputated tail or blood cells (sample types depend on each experiment) by boiling them in 50 mM NaOH at 98°C for 10 min, followed by neutralization with 1/10 volume of 1 M Tris-HCl (pH 7.5).

PBCs were collected as described previously (Naora et al., 2013) with minor modifications. After tail amputation, each tadpole was immersed in 0.6% sodium chloride with 2.5 mM EDTA in a 0.2 ml PCR tube and left to bleed for 45 min. The tadpoles were then removed and the resulting blood fraction was centrifuged at 500 g for 5 min to pellet blood cells. The supernatant was removed and pelleted blood cells were resuspended in 10 µl of 50 mM NaOH, and genomic DNA was extracted as described above.

PCR and Sanger sequencing were performed with the primers listed in Table S6. Sequence data were subjected to ICE analysis (Hsiau et al., 2018 preprint) (https://ice.synthego.com/) to estimate the KO scores (proportion of cells with either a frameshift, or at least a 21 bp insertion/deletion.).

Tails of intact tadpoles at 4 dpf and tadpoles at 2 dpa whose tails were amputated at 4 dpf (as ‘regeneration bud’) were subjected to scRNA-seq. Intact tails were divided into anterior (as ‘tail stump’) and posterior (as ‘tail tip’) parts (Fig. S3A). Tails were shredded using a razor. Cells were then enzymatically dispersed by treating them in 1× PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄ and 1.8 mM KH₂PO₄) supplemented with 0.4 mg/ml of Liberase TH Research Grade (Roche) and 177.7 units/μl DNase I (Thermo Fisher Scientific) for 40 min on a rotator at 30°C. Dispersed cells were washed and suspended in 1× sorting buffer [1× PBS supplemented with 1 mM EDTA, 25 mM HEPES and 1% BSA (Nacalai Tesque)], then filtered through a 55 μm opening nylon mesh (AS ONE, Osaka, Japan). Dead cells were stained by adding 1/100 volume of 7-AAD (TONBO Biosciences). Live cells in the cell suspension (7-AAD negative) were sorted using a FACS Aria III (Becton, Dickinson and Company). Doublet or multiplet events were gated out using height and width parameters of forward and side scatter. Sorted cells were used to generate scRNA-seq libraries using Chromium Single Cell 3′ Reagent Kits v3 (10X Genomics), followed by sequencing on a Novaseq 6000 (Illumina).

Output data were processed using Cell Ranger v3.0.2 to generate a gene count matrix. We used a genome sequence file generated by combining the X. laevis genome version 9.1 sequence (Xla.v91.repeatMasked.fa) and mitochondrial genome sequence of genome version 9.2 (XL9_2.fa) for mapping, and a gene model file generated by combining the gene model for genome version 9.1 (XL_9.1_v1.8.3.2.primaryTranscripts.gff3) and mitochondrial gene annotation of the gene model for genome version 9.2 (XENLA_9.2_Xenbase.gff3) for read counting. The genome sequence files and the gene model files were obtained from Xenbase. The generated count matrix was analyzed using Seurat package v3.2.2 (Stuart et al., 2019). The cells whose percentage of mitochondrial gene counts per total gene counts exceeded 20, or whose total gene counts were outside a 500-10,000 range were excluded from the analysis. The numbers of cells used for analysis were as follows: tail stump, 6622; tail tip, 7600; regeneration bud, 7400. The gene counts were normalized by the total expression of each cell and scaled; linear dimensional reduction was then performed by principal component analysis. For clustering the cells, we used FindNeighbors and FindClusters functions on Seurat using PC1-PC100, which divided cells to 50 clusters, and processed data was visualized using Uniform Manifold Approximation and Projection (UMAP) (Mcinnes et al., 2018 preprint).

The RNA-seq data of the X. laevis embryo developmental expression (Session et al., 2016) and tail stumps at 48 h after amputation of il11 KD and control tadpoles (Tsujioka et al., 2017) were analyzed using HISAT2 v2.1.0 (Kim et al., 2019) for mapping, and HTSeq v0.11.2 (Anders et al., 2015) for read counting, with the genome sequence file (Xla.v91.repeatMasked.fa) and the gene model file (XL_9.1_v1.8.3.2.primaryTranscripts.gff3). Differential expression analysis was performed using DESeq2 package v1.32.0 (Love et al., 2014).

rfem.L and rfem.S KD tadpoles were obtained as described above. For the control group, tadpoles injected with only cas9 mRNA were used. We amputated the tails of tadpoles at 4 dpf, then 13-20 tail stumps of tadpoles were sampled 48 h after amputation. We collected samples from three different batches. The samples were homogenized using TRIzol reagent (Thermo Fisher Scientific), and extraction of total RNA and the following RNA-seq were performed at Tsukuba i-Laboratory LLP (Ibaraki, Japan). Extracted total RNA was subjected to mRNA isolation using NEBNext Poly(A) mRNA Magnetic Isolation Module (New England Biolabs). cDNA libraries were produced using a NEBNext Ultra Directional RNA Library Prep Kit for Illumina (New England Biolabs). Sequencing was performed using a NextSeq500 (Illumina) to generate ∼4×10⁷ paired-ends reads (36 bp ×2) from each cDNA library. The RNA-seq data were analyzed as described above.

The rescue construct was generated as follows. Two I-SceI meganuclease recognition sites were inserted between SphI and PstI sites, and KpnI and SacI sites, respectively, of the pUC19 plasmid to generate the pIS2 plasmid. A DNA fragment containing the zebrafish mpeg1 promoter sequence (Paredes et al., 2015), rfem.L-coding sequence, P2A, acgfp1-coding sequence, and SV40 polyadenylation signals were synthesized by Twist Bioscience. Synonymous substitutions were introduced into the rfem.L sequence at gRNA#1 and #2 target sites [#1, GGGATtTAttctTCtATcCC; #2, GGAGCcTGGGTttTaTAcCA (substitutions are indicated with a lowercase letter)]. A DNA fragment without the rfem.L-coding sequence and P2A was also synthesized for the control experiment. pIS2 plasmid was digested with XbaI and the DNA fragment was inserted using an In-Fusion HD cloning kit (Takara) to generate mpeg1:rfem and mpeg1:gfp constructs, respectively.

Cell type-specific rfem.L forced-expression in rfem.L and rfem.S KD tadpoles was performed as follows. The constructs had two I-SceI sites to perform I-SceI mediated transgenesis, which would facilitate the expression by stably inserting into the genome, in addition to the free constructs remaining in the cells of the tadpole at the time of tail amputation. I-SceI mediated transgenesis was performed as described previously (Ogino et al., 2006) with several modifications. We prepared a solution containing 1.25 µM of Cas9 protein (PNA Bio), 13.3 ng/µl of rfem.L/S gRNA #1, #2 and #3, respectively, 0.5 units/µl of I-SceI (New England Biolabs), 75 ng/µl of mpeg1:rfem or mpeg1:gfp constructs, and 1×CutSmart Buffer (New England Biolabs). The solution was incubated at 37°C for 30 min, then 18.4 nl was injected into one-cell stage embryos as described above. Injected embryos were maintained at 20°C. Tails of tadpoles were amputated at 4 dpf, followed by assessment of the regenerative capacity by measuring of the area and length of the regenerated tails at 7 dpa. A two-sided Fisher's exact text was performed using R (v3.6.3)

All statistical analyses were conducted in R (v3.6.3). The statistical method and sample size are provided in the corresponding figure legends.

Cell sorting and flow cytometry analysis for this study were performed on instruments in the FACS Core Laboratory at The Institute of Medical Science, The University of Tokyo, with technical support of FACS Core Laboratory staff. scRNA-seq work was supported by the Japan Society for the Promotion of Science (KAKENHI 16H06279 to the Platform for Advanced Genome Science).

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.200467.reviewer-comments.pdf.