Ethics oversight
All animal experiments were approved by the Magistrate of Vienna (Genetically Modified Organism Office and MA58, City of Vienna, Austria), under licences GZ: 51072/2019/16, GZ: MA58-1432587-2022-12 and GZ: MA58-1516101-2023-21.
Axolotl husbandry
Axolotls (A. mexicanum) were raised in individual aquaria in Vienna tap water. Axolotl matings were performed by the animal-care team at the IMP. Axolotl surgery, live imaging and tissue collection were performed under anaesthesia in 0.015% benzocaine (Merck, E-1501) diluted with Vienna tap water, using the benzocaine preparation described previously60. All limb amputations were performed through the middle of the lower arm (zeugopod), unless indicated otherwise. Axolotl sizes are reported in centimetres, measured from snout to tail. Axolotls up to 16 cm in length were used in experiments (an age of approximately 8 months). Axolotls were randomly allocated into experimental or control groups and housed under identical conditions, except in experiments in which the control assay was performed on one limb and the experimental perturbation on the other limb of the same axolotl. Experiments were not blinded, except to the sex of the animals.
Axolotl genome and transcriptome reference
We used genome assembly AmexG_v6.0-DD and transcriptome assembly AmexT_v47 (ref. 61).
Isolation of anterior and posterior dermal cells
Axolotl embryos of genotype tgSceI(Mmu.Prrx1:TFPnls-T2A-ERT2-Cre-ERT2; Caggs:loxP-GFP-loxP-mCherry)Etnka were treated with 4-OHT (Merck, H7904) as described62 to permanently label connective tissue cells with mCherry, then raised individually until 12 cm long. Skin (containing dermal connective tissue cells) was removed from the lower arms (zeugopods) and then dissected into anterior and posterior halves, leaving a gap between them. We prepared two replicates for anterior and three for posterior, with each replicate deriving from eight axolotls (16 lower arms). Anterior and posterior samples were dissociated into single-cell suspensions using Liberase TM enzyme (Merck, 5401119001) as described63, with the following modifications: enzymatic digestion was done for 50 min and the cells were filtered through a 50-μm Filcon filter (BD Biosciences, 340630). The mCherry-positive cells were purified from each replicate by FACS (FACSAria III Cell Sorter, BD Biosciences) using a 100-μm low-pressure nozzle and collected into separate tubes of cold amphibian culture medium64. Each replicate was pelleted at 300 × g for 4 min at 4 °C, then resuspended in 500 μl TRIzol reagent (Thermo Fisher Scientific, 15596026). RNA was extracted according to the manufacturer’s protocol and stored at −70 °C until required.
QuantSeq library preparation and RNA sequencing
Libraries for dermal-cell RNA sequencing were prepared using QuantSeq 3′ mRNA-Seq Library Prep Kit FWD (Lexogen) with 20.25 ng of input RNA per sample. Input was 4.5 μl of input RNA plus 0.5 μl of ERCC RNA Spike-In Mix (Thermo Fisher Scientific, 4456740) pre-diluted 1:10,000 in water. Samples were multiplexed for sequencing, using i7 indices 7023 (CACACT, anterior replicate 1), 7025 (TTTATG, anterior replicate 2), 7022 (GGAGGT, posterior replicate 1), 7024 (CCGCAA, posterior replicate 2) and 7026 (AACGCC, posterior replicate 3). Each replicate was sequenced to a depth of 120 M reads, in SE 100 mode, distributed over 3 lanes of a HiSeq 2500 with v4 chemistry (Illumina). Sequencing was performed by the Next Generation Sequencing Facility at the Vienna BioCenter Core Facilities (VBCF), a member of the Vienna BioCenter (VBC), Austria. Dermal-cell RNA sequencing data have been deposited with the Gene Expression Omnibus (GEO) with accession number GSE243137.
Gene expression analysis
Adaptor sequences were trimmed from the raw sequencing reads using Trimmomatic (v.0.39)65, with parameters ILLUMINACLIP:Adapters.fa:2:30:7 SLIDINGWINDOW:4:20 MINLEN:40 in single-end mode. Trimmed sequenced reads were mapped to axolotl genome AmexG_v6.0-DD with HISAT2 (ref. 66), with parameters –no-unal –summary-file Output.log -k 5 –very-sensitive -x DBGenome -U Reads.fq.gz > Alignment.sam. We used featureCounts67 to generate a read counts table. Differential expression analysis was performed on two anterior replicates and three posterior replicates using R v.4.1.2 and DESeq2 (ref. 68) v.1.34.0 with an FDR cutoff of P < 0.01. Volcano plots were generated using ggplot2 v.3.3.6 (ref. 69). Heatmaps were generated using the pheatmap package v.1.0.12 (R. Kolde). Gene Ontology analysis was done using the topGO package v.2.46.0 (Alexa A, Rahnenfuhrer J) using parameters ontology = “BP”, geneSelectionFun = topDiffGenes, annot = annFUN.org, mapping = “org.Hs.eg.db”. To calculate significant Gene Ontology terms, Fisher’s exact test was used with the “elim” algorithm. To enable interpretation of the differential expression results, we generated a custom gene nomenclature derived from the AmexT_v47 transcriptome. We concatenated each axolotl gene identifier with the gene symbol for the direct human homologue where available or, if not available, the closest homologue from the NCBI non-redundant database.
Axolotl transgenesis
Plasmids for axolotl transgenesis were assembled by Gibson Assembly, amplified using Plasmid Maxi Kits (Qiagen, 12163) and verified by Sanger sequencing before egg injection. One-cell-stage axolotl eggs were surface sterilized twice for 5 min with about 0.004% sodium hypochlorite solution (Honeywell, 71696) diluted with Vienna tap water, then washed well with fresh tap water. The following steps were performed as described in ref. 60. Eggs were de-jellied using sharp forceps in 20% Ficoll (Merck, GE17-300-05)/1X MMR/Pen-Strep (Merck, P0781) solution, then held in 10% Ficoll/1X MMR/Pen-Strep solution until microinjection. For microinjections, borosilicate glass capillary needles with filament (Harvard Apparatus, GC100F-15) were pulled using a Flaming/Brown Micropipette Puller P-97 (Sutter Instrument) with settings P = 500, heat = 530 (Ramp test + 30), pull = 100, velocity = 120, time = 150. Then 5 nl of the appropriate injection mix was injected into each de-jellied egg, delivered in two 2.5-nl shots. Egg injections were performed using an Olympus SZX10 microscope using a PV830 pneumatic Picopump (World Precision Instruments) with settings vacuum eject, regulator 25, range 100 ms, timed, duration 10-0. Injected eggs were transferred to 5% Ficoll/0.1× MMR/Pen-Strep solution overnight. The next morning, healthy eggs were transferred to individual wells of a 24-well multiwell plate (Thermo Fisher Scientific, 142475) filled with 0.1× MMR/Pen-Strep solution. Embryos were screened for fluorescent transgene expression at embryo stage 42 using an AXIOzoom V16 widefield microscope (Zeiss). Axolotl lines are named according to the convention established in ref. 70.
The following axolotl lines were generated by random insertion I-SceI meganuclease-mediated transgenesis: ZRS>TFP (tgSceI(ZRS:TFPnls-T2A-ERT2-Cre-ERT2)Etnka); Prrx1>mCherry (tgSceI(Mmu.Prrx1:mCherry)Etnka); and Prrx1>mCherry–Hand2 (tgSceI(Mmu.Prrx1:mCherry-Hand2)Etnka), which expresses an mCherry–Hand2 fusion protein. The injection mix was prepared according to ref. 71: transgene plasmid 1 μg, I-SceI enzyme (NEB R0694) 5 units, CutSmart buffer (NEB) 1×, water to 10 μl. In Prrx1>mCherry–Hand2, we fused mCherry to the amino terminus of Hand2, connected by a glycine–serine-rich linker of the sequence SGGGGSGGGGS. In ZRS>TFP, the following CMV minimal promoter was used:
>CMV minimal promoter (56 bp)
GGCGTGTACGGTGGGAGGTCTATATAAGCAGAGCTGGTTTAGTGAACCGTCAGATC
The following axolotl lines were generated by NHEJ-mediated CRISPR/Cas9 knock-in: Hand2:EGFP (tm(Hand2t/+:Hand2-T2A-EGFP)Etnka); Hand2 lineage tracer (tm(Hand2t/+:Hand2-P2A-EGFP-T2A-ERT2-Cre-ERT2)Etnka); and Alx4:mCherry (tm(Alx4t/+:Alx4-T2A-mCherry)Etnka). We followed the protocol in ref. 72, using the following injection mix: Cas9-NLS protein, 5 μg; gRNA, 4 μg; targeting construct, 0.5 μg; Cas9 buffer 1×; water to 10 μl. Cas9-NLS protein and buffer were synthesized by the Vienna Biocenter Core Facilities.
The following transgenic axolotls were published previously: tgSceI(Caggs:loxP-GFP-dead(Stop)-loxP-mCherry)Etnka (ref. 9), tgSceI(Caggs:loxP-GFP-loxP-mCherry)Etnka (ref. 62), tgSceI(Mmu.Prrx1:TFPnls-T2A-ERT2-Cre-ERT2)Etnka (ref. 62).
Further details on generating 3′ knock-in axolotls
We generated and characterized efficient sgRNAs targeting the last intron of Hand2 or Alx4 following the protocol in (ref. 72). The sgRNAs were produced by assembly and in vitro transcription of a synthesized DNA template. The following forward oligos, harbouring the sgRNA target sequence in the Hand2/Alx4 intron (lower case) plus T7 promoter (underlined), were at 100 μM concentration (Merck). They were PCR amplified with the universal oligo_reverse, also at 100 μM, then purified to generate DNA templates for in vitro transcription:
>Hand2 sgRNA oligo_forward
GAAATTAATACGACTCACTATAGGatgctgtcctctaaaccgGTTTTAGAGCTAGAAATAGC
>Alx4 sgRNA oligo_forward
GAAATTAATACGACTCACTATAGGttcactactggtaaatacGTTTTAGAGCTAGAAATAGC
>Universal oligo_reverse
AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTA AAAC
In vitro transcription was performed using a MEGAscript T7 transcription kit (Thermo Fisher Scientific, AM1334) and 500 ng of purified template in a 20-μl reaction. The transcription reaction was done overnight at 37 °C. The DNA template was removed by adding 1 μl TURBO DNase (Thermo Fisher Scientific, AM2238) for 15 min. RNA was purified by LiCl precipitation. Purified sgRNA was re-suspended in water at a concentration of 1 μg μl−1 and stored at −70 °C.
Knock-in constructs were generated by PCR amplification and Gibson assembly of the following components: around 400 bp of the last intron (harbouring the sgRNA target sequence) plus the complete last exon of Hand2/Alx4 minus the stop codon; the transgenes to be knocked in (for example, T2A-EGFP) plus a stop codon; and poly(A) sequence from SV40 (Hand2 knock-ins) or rabbit β-globin (Alx4 knock-in).
>Hand2 last intron plus last exon minus stop codon (650 bp)
GGCCGCGGACATTAGGCGACGTAAAGAAAGGCCCATCGCAGCCGCGGCCTGTATTtTCGCGGATAATGCCTGCGCCGCGTCTGGAGGGGCAGATATAATCCCCAGCTCCACGGCAGCCCTTCAGATGTGGCGATTGCCTCGGTTTAGAGGACAGCATTTACATAGCTTTCAGGTGAACTTGAGTATGAATCGCAATCACTCGTGTTGTCTTTCTCTCTCTCTGTGTATCCCCCTCCCCCTCTCTCTTTTTATATATATATATATATATATATTGCaGTTTCGCCTACAACTGTGGCCCTGTCTGTCTGCTAAAAAGGGGGGAATTGGCAAGTGCGTGTTGCTGAAGGCTGTAGTGCGGTGTGTGTGCGTGTATATATGTTACGTAGAGATACATAGATACATATCCGTGTTACGTGTTACGAATTCGTGCGTGTGTGTGTGCGCGATTATCCGCGTTGGTTGTGAACACATGTTTGGGTCTGCAGCAAATCAACATTCAATTGTGAGATATTGAGTTCTCTTTGCTTTTGTCTCCCTTCCCGCTCTCTTGCCAGAATGAACTCTTGAAAAGTACCGTCGGCAGCAACGACAAGAAGAGCAAGGGCAGGACTGGCTGGCCTCAGCACGTCTGGGCCTTGGAGCTCAAGCAG
>Alx4 last intron plus last exon minus stop codon (679 bp)
GACATGTAGGGGCAATCTGAAGTCCCACTCAAAGCCCACCTAGAACCGTCCCTGCTCAGCTGGGGGAAGGCAGAATCAAATTTTGTGGAAGGCAGTCCTGTAACTCGCACCCAGAACTCTACAGCCTGTACACTGAAATATAATCAAATGGTGTTGATAATTCACAATGTGATTCACTACTGGTAAATACCGGTAACACTGAACCGCTGAGCGACATCATACAACATATTTCAAATTGGTATTAATTGTATAATGTTCCTATACTCGTCTCTTGCTGTAAATCTTATTTATTGCCTCCAGCCCTCCAAATAGTGCCACTTTCTCATTCCTTGTCTACTTTTGTCTTCTCCTGTTACAGATCCAGAACCCAACATGGATTGGAAACAACAGCGGGGGCTCTCCGGTGGCAGCCTGTGTGGTCCCCTGTGACACCGTCCCATCCTGCATGTCTCCTCATGGCCACCCCCATGCAAGTGGAGGTGTTTCTGAATTCCTGAGCGTGCCTAGCTCAGGAAGCCACATGGGTCAGGCACACATGGGTAACCTCTTTGGCACTGCTGGGCTCAGCACAGGCATCAATGGCTACGACCTCAACGTGGAGCCAGACCGCAAGACCTCCAGCATCGCAGCCCTGCGGATGAAGGCCAAGGAGCACAGTGCCGCCATCTCCTGGGCCACA
Codon alteration
The following sequences were codon-altered to enable them to be distinguished from endogenously expressed mRNA: axolotl Hand2 in Prrx1>mCherry–Hand2 and axolotl Shh in BV-Shh baculovirus. An axolotl codon usage table was generated using the first transcript isoform from each gene annotated in axolotl transcriptome assembly AmexT_v47. Optimizer (http://genomes.urv.es/OPTIMIZER/)73 was used with the method ‘guided random’ to alter codons while still reflecting axolotl codon usage.
>Codon-altered axolotl Hand2 ORF (645 bp)
ATGAGCCTGGTGGGCGGTTTCCCACACCACCATCCTGGCGTGCATCACCACCATGAGGGCTACCCCTTTTCGGCCGCCGCAGCAACGGGGAGATGCCACGAAGACTCGCCATACTTTCATGGTTGGCTTATCGGTCATCCGGAGCTCTCGCCTCCCGATTATGGTCCAGGAGCACCCTACAGTCCTGAATATGGAGGGGGGGGCGGCCTTGAACTATGCGGGCCTGGGGGCGCGCCAGGGGGAGGAGCCGGAGCGCTTCTCTCAACTAGACCTGTGAAGCGGCGAGGCACCGCTAATAGGAAGGAGCGGCGGAGAACCCAAAGCATCAACAGTGCTTTCGCTGAGCTCCGGGAATGTATCCCGAATGTGCCAGCCGACACGAAGTTGTCAAAGATCAAAACTTTGCGTCTAGCCACTTCTTATATCGCCTACCTGATGGATTTGCTTGCCAAGGATGAGCAGTCTGAAGCCGAAGCTTTCCGGGCAGATCTGAAACAGAGGGGAGGGGGTGGGGCTGAGTGTAAGGAAGATAAAAGAAAGAAGGAGTTGAATGAATTGCTGAAGTCCACAGTCGGGAGTAACGACAAGAAATCCAAGGGTCGCACCGGTTGGCCACAGCATGTGTGGGCGCTAGAGCTCAAGCAG
>Codon-altered axolotl Shh ORF (1,269 bp)
ATGCGTCTCCTCCTTCGCCGGCTACTGCTGGGTACCTTGGTTTGGGCACTGCTAGTGCCCAGCGGCCTGACTTGCGGCCCGGGGCGTGGTATCGGTAAAAGGAGACAGCCTAAAAAACTGACACCCCTCGCGTACAAGCAGTTTATCCCCAACGTCGCGGAGAAGACACTGGGAGCATCTGGACGTTATGAGGGGAAGATCACTAGGAACTCTGACCGTTTCAAGGAGCTCACTCCTAATTACAACCCCGACATCATTTTTAAGGACGAGGAGAATACAGGAGCTGACCGACTGATGACTCAGAGGTGCAAAGACAAACTGAATGCCCTGGCTATTAGCGTAATGAATCAGTGGCCGGGCGTGAAACTGCGGGTGACGGAAGGCTGGGATGAAGATGGTCATCACAGTGAGGAGAGTCTGCATTACGAGGGCCGAGCCGTGGATATCACAACCTCTGACCGTGACAGGTCTAAGTATGGAATGCTGGCACGTCTGGCCGTGGAGGCAGGCTTTGATTGGGTCTACTTCGAGTCCAAGGCCCACATACATTGCAGCGTGAAGGCGGAGAACAGTGTGGCAGCCAAGTCGGGAGGATGTTTTCCGGCCAGTGCTAAGGTTACACTGGAACATGGCGTTACGAGACCAGTGAAGGATCTGCGACCCGGAGACCGTGTGCTAGCAGCAGATGGACAAGGTCGACTGGTTTATAGCGACTTTCTTATGTTTCTCGACAAAGAAGAGGCAGTGACAAAGGTCTTTTACGTCATTGAGACGGAGAGACCAAGGCAGAGGCTAAGGTTGACAGCAGCCCACCTCCTGTTCGCCGCAAGGCATCCCGCAAACTCATCTAGCTCCACCGGGTTCCAAAGTATCTTCGCATCAAGGGTTCGACCTGGGCACCGGGTGCTTACTGTCGACCAGGAAGGACGGGGGCTTCAGGAGGCTACTGTCACTCGCGTGTACCTGGAGGAGGGTGCCGGAGCCTACGCCCCCGTTACCAGTCATGGAACCGTTGTGATTGACAAGGTACTCGCCAGTTGCTACGCAGTGATCGAGGAGCATTCCTGGGCCCACTGGGCTTTTGCCCCTCTGCGACTTGGCTACGGCATACTGAGCATCTTTTCCCCTCAAGATTACAGCCCACATAGTCCCCCCGCGCCTAGCCAGAAAGAAGGCGTGCATTGGTACTCAGAAATCCTGTATCATATAGGGACATGGGTGCTGCATAGCGACACTATTCACCCCTGGGGCATGGCCGCCAAGTCGAGT
Generation of Hand2 CRISPant axolotls
To generate Hand2 CRISPants lacking a translational start, the following mix was microinjected into fertilized one-cell-stage eggs, following the protocol in ref. 72: Cas9-NLS protein, 5 μg; sgRNA1, 2 μg; sgRNA2, 2 μg; Cas9 buffer 1×; water to 10 μl. The target sequences for sgRNA1 and sgRNA2 flank the Hand2 translational start. The control mix contained all these components except the sgRNAs. The sgRNA1 and sgRNA2 were produced by in vitro transcription using the PCR-based strategy described above.
>Hand2 CRISPant sgRNA1 target
GCGGCCCCTGGGAGGCCC
>Hand2 CRISPant sgRNA 2 target
ACCCCAGGGTGGTGGTGA
To assess whether Hand2 CRISPants lose Hand2 expression, an sgRNA1/sgRNA2 mix was injected into the left or right blastomeres of four-cell-stage Hand2:EGFP eggs. In this manner, half of the animal acts as an internal control (EGFP is not affected) and the other half acts as the test (to test whether EGFP is reduced). We chose to inject four-cell-stage eggs rather than two-cell-stage eggs because the latter have incomplete cell cleavage, which leads to the risk of sgRNAs leaking into the opposite half of the egg. Each injected blastomere received half the dose of the regular egg injection. EGFP fluorescence was measured when the larger of the two limb buds reached stage 46/47. Control animals were injected with the control mix lacking the sgRNAs. Hand2:EGFP intensity in the brighter limb bud was divided by the intensity in the dimmer limb bud to yield a fluorescence ratio.
To generate Hand2 M146 CRISPants targeting sequences close to M146, the following sgRNAs were injected individually instead. M146 sgRNAs were produced by in vitro transcription using the PCR-based strategy described above.
>Hand2 M146 sgRNA1 target
CCTACCTCATGGACCTGC
>Hand2 M146 sgRNA2 target
TCATGGACCTGCTGGCCA
>Hand2 M146 sgRNA3 target
CCAAGGACGAGCAGAGCG
Estimation of indel frequency in Hand2 CRISPants
Genomic DNA was individually extracted from Hand2 CRISPant or control sibling limbs. Amputations were performed through the middle of the lower arm and each off-cut was placed in 50 μl of 50 mM NaOH. Collected tissue was heated to 95 °C for 12 min, then cooled to 4 °C in a thermocycler. Next, 5 μl of 1 M Tris, pH8 was added, then the extracted DNA was stored at 4 °C until genotyping. For genotyping, 1 μl of extracted DNA was PCR amplified using KAPA HiFi HotStart ReadyMix (Roche 07958927001) and primers that generate a roughly 750 bp amplicon surrounding the Hand2 translational start. Then 30 ng of PCR-amplified and purified DNA was Sanger sequenced using Hand2_sequencing primer at the IMP/IMBA Molecular Biology Service. The indel frequency was estimated from the Sanger-sequencing results using the ICE Analysis Tool (Synthego, https://ice.synthego.com/).
>Hand2_genotyping_F
GAAGTAGCAGGGATGGACGAG
>Hand2_genotyping_R
AAGGCGCTGTTGATGCTCT
>Hand2_sequencing
CACAGGCCAGGACTTCAAGAA
3-Primer genotyping strategy to identify Hand2
Δ64 mutants
Genomic DNA was extracted individually from tail clips as described above in the section above. PCR amplification was performed with the following 3-primer mix and KAPA2G Fast HotStart Genotyping mix (Roche 2GFHSGKB). The PCR reaction produces a roughly 752 bp amplicon as a positive control and also a 431 bp amplicon if the Hand translational start is intact. Homozygous Hand2Δ64 mutants fail to amplify a 431 bp amplicon but amplify the larger amplicon.
>Hand2_genotyping_F
GAAGTAGCAGGGATGGACGAG
>Hand2_genotyping_R
AAGGCGCTGTTGATGCTCT
>Hand2_genotyping_R2
CCCCCCACCAAGCTCATG
Determination of axolotl ZRS enhancer
The axolotl ZRS enhancer was determined by multiple species alignment of the following genome sequences using mVISTA74 and PipMaker75:
axolotl (A. mexicanum) assembly, AmexG_v6.0-DD chr2p:694366863-694689506;
human (Homo sapiens) assembly, hg38 chr7:156769228-156790956;
mouse (M. musculus) assembly, mm10 chr5:29292950-29323801;
chick (G. gallus) assembly, Gal6 chr2:8538956-8559114;
fugu (Takifugu rubripes) assembly, fr3 chr10:5739579-5747090.
>Axolotl ZRS enhancer, AmexG_v6.0-DD chr2p:694613160-694613980 (821 bp)
ACCTTAATATCCATCTTTGCATTTGAAGTTGTTGCATAAAATGTACCACGAGCGACAGCAACATCCTGACTAATTAGCCAAATTACCCAGACATCCCTCCAAAAAAGCCGCGAAACAGAGAGCATGTCTGTCGGATTAAAAGGTTGTAACTCCTAAAACATCAAACGGAGCGCCAGATAATAAAAGCCAATCGTACAGAAATTTGAGGTAACTTCCTTGCTTAATTAATTAGCTAGGCCAGTTGGAGCGAGGAGGCCAACGCGGGCGCGTAGAACGCCCATAAAGCTGAACAACTCGACAGCACAAAAGTGGAGAAACAAAGATTTTTTAATATGCGTCTATCCTGTGTCACAGTTTGAAATTGTCCTGGTTTATGTCCCTTTTGGCAAAGTTACAATAAAAGTGACCCTGTACTGTATTTTATGGCCAGACGACTTTTCGTTTTGTTCCCGGTGACTAATTTGACTCAGGCCCCCATCTTGAATAGACACAGAAAGGGGCCGGGGGAATGAGGCTGTCTGTCTCGCTTGGGTTTCATTGCATTTTTTCATTATTCGGGCTCGTTTTTCGCCACAGATCATCCATAAATTGTTGGAAATGAGTGATTAAGGAAGTGCTGCTTAATGTTAGTAGCACACATTCTTTGTGCGTTTCACCCTCCCGCCCCCTCCATTTTGTGGGTGAGAGGAAATCAAGTAATGCAGAAACAATAAGGAAGCCTCCTGCTGGGAACCTTTCAAGGAAATGTAACCTGCATACTGTTTTGATCTCGGTGTTCCTTTCAGAGTATGCCGCGATGTTTCAACAGCTATTTTCATGTG
Genetic lineage tracing (ZRS/Hand2)
For lineage tracing, either ZRS>TFP axolotls or Hand2 lineage-tracing axolotls were mated with loxP–mCherry fate-mapping axolotl of genotype tgSceI(Caggs:loxP-GFP-dead(Stop)-loxP-mCherry)Etnka (ref. 9). To induce Cre–loxP recombination during development, stage-42 ZRS or Hand2 lineage-tracing embryos were bathed overnight in the dark in 500 ml of 2 μM 4-OHT, as described in the water-based method of ref. 60. We calculated the overlap between embryonic Shh cells (mCherry+) and regeneration Shh expression (TFP+) using wide-field microscopy. To induce Cre–loxP recombination in Hand2 cells of the mature limb, 7-cm axolotls were bathed individually and overnight in the dark in 100 ml of 5 μM 4-OHT. After treatment, the axolotls were transferred to tap water and allowed to recover for one week. The same 5 μM 4-OHT treatment and one week recovery was repeated twice more, for a total of three treatments. Animals were screened for Cre–loxP recombination and mCherry expression using an AXIOzoom V16 wide-field microscope (Zeiss). Regeneration experiments on ZRS>TFP lineage tracings were performed on 11-cm axolotls. Regeneration experiments on Hand2 lineage tracings were performed on 6-cm axolotls.
Quantification of ZRS>TFP fluorescence
An integrated fluorescence score was calculated for each Hand2 CRISPant limb bud (mean TFP intensity × area), then normalized to the mean of the control cohort (set to 1). Quantifications were performed on stage-42 limb buds imaged using wide-field microscopy.
Assessment of leaky mCherry expression
ZRS/Hand2 lineage tracing animals were generated as described above and genotyped individually to ensure that they carried both Cre and loxP–Stop–loxP-mCherry genetic cassettes. Primers used for genotyping were:
>Cre_Fw
ATCCGAAAAGAAAACGTTGA
>Cre_Rv
ATCCAGGTTACGGATATAGT
>Cherry_Fw
GGATAACATGGCCATCATCAAGGAGTTC
>Cherry_Rv
GTCTTGACCTCAGCGTCGTAGTG
Half of the animals were treated at limb-bud stage 42 with 2 μM 4-OHT overnight to induce recombination and mCherry labelling. The other half were left untreated, to assess for nonspecific mCherry expression (leakiness). When the animals reached 6 cm, their limbs were removed, dissociated with Liberase TM enzyme (see the section ‘Quantifying Hand2:EGFP expression during regeneration’ below) and analysed for mCherry fluorescence using a flow cytometer, a 100-μm low-pressure nozzle and FLOWJO software (BD Biosciences). Untreated and treated samples were analysed in the same session. At least 130,000 events were recorded for each sample.
Surgical depletion of embryonic Shh cells
ZRS lineage-traced axolotls were prepared by treating stage-42 embryos with 2 μM 4-OHT, as described in the section ‘Genetic lineage tracing (ZRS/Hand2)’, above. At an axolotl size of 6 cm, the left arm of each axolotl was depleted for ZRS-lineage cells by using microscissors to excise tissue posterior to the ulna in the lower arm. Successful depletion was confirmed by imaging loss of mCherry fluorescence using an AXIOzoom V16 wide-field microscope (Zeiss). The right arm of each animal was treated as a control and depleted for an equivalent amount of tissue anterior to the radius instead. Two days after surgery, each arm was amputated through the distal part of the lower arm, distal to the depleted region. Images were acquired every few days after amputation to assess the onset of ZRS>TFP expression in depleted versus control limbs. The mCherry depletion efficiency was estimated by comparing the area of mCherry-positive tissue in the blastema in wide-field images acquired from control and depleted animals.
Tissue preparation, staining and imaging
Samples were fixed overnight in 4% paraformaldehyde, pH 7.4, at 6 °C. The next morning, samples were washed well with cold PBS then equilibrated with the following solutions at 6 °C for one overnight each: 20% sucrose/PBS; 30% sucrose/PBS; and a 1:1 mix of 30% sucrose/PBS and Tissue-Tek O.C.T. compound (Sakura). Samples were mounted in Tissue-Tek O.C.T. compound, frozen on dry ice and stored at −70 C until sectioning. Cryosections of 16 μm thickness were prepared using a Cryostar NX70 (Thermo Fisher Scientific) and collected on Superfrost Plus adhesion microscope slides (Epredia, J1800AMNZ). Slides were stored at −20 °C until required. For staining, slides were brought to room temperature then washed well with PBS to remove O.C.T. compound before proceeding to the following steps. For DAPI staining only, slides were incubated with DAPI 1:1,000 in PBS + 0.2% Triton X-100 for 1 h at room temperature, then washed well with PBS + 0.2% Triton X-100 before mounting. For staining with anti-Prrx1 or anti-Col1A1 antibody, slides were blocked for 30 min at room temperature with PBS + 0.2% Triton X-100 + 1% normal goat serum (NGS), then incubated overnight at 6 °C with rabbit anti-Prrx1 antibody62 diluted 1:500 in PBS + 0.2% Triton X-100 + 0.1% NGS or mouse anti-Col1A1 antibody (SP1.D8, DSHB) diluted 1:50 in PBS + 0.2% Triton X-100 + 0.1% NGS. The following day, slides were washed well with PBS + 0.2% Triton X-100 then incubated for 2 h at room temperature with Alexa 647-conjugated anti-rabbit (Invitrogen, A-21244) or anti-mouse (Invitrogen, A-21240) secondary antibody diluted 1:500 in PBS + 0.2% Triton X-100. Slides were washed well with PBS + 0.2% Triton X-100 before mounting. For HCR (hybridization chain reaction) in situ hybridization, slides were stained according to the HCR RNA-FISH protocol for fixed frozen tissue sections (Molecular Instruments), omitting post-fixation and Proteinase K treatment. Probe hybridization buffer, wash buffer and amplification buffer were from Molecular Instruments. Samples were mounted in Abberior Mount liquid antifade mounting media (Abberior) for imaging. Images were acquired with an LSM980 AxioObserver inverted confocal microscope with ZEN software (Zeiss), plus AiryScan 2 for HCR experiments only.
Whole-mount sample preparation, staining and imaging
Samples were fixed overnight in 4% paraformaldehyde, pH 7.4, at 6 °C. The next morning, samples were washed well with cold PBS then dehydrated progressively through ice-cold 25% methanol/PBS, 50% methanol/PBS, 75% methanol/PBS and 100% methanol (30 min each). Samples were kept for one night in 100% methanol at −20 °C. The next day, samples were rehydrated progressively through ice-cold 75% methanol/PBS, 50% methanol/PBS, 25% methanol/PBS and PBS (30 min each). Samples were washed twice more with cold PBS then stained for Shh transcripts using the HCR RNA-FISH protocol for whole-mount zebrafish embryos and larvae (Molecular Instruments), starting at the section ‘Detection stage’. After completion of the HCR protocol, samples were stained overnight at 6 °C in DAPI 5 mg ml−1 (Sigma) diluted 1:1,000 in 5× SSC + 0.1% Tween 20. The next day, samples were washed well with 5× SSC + 0.1% Tween 20 then optically cleared overnight at room temperature on an aerial rotator in clearing-enhanced 3D (Ce3D) solution, refractive index 1.50, prepared as described previously76. Images were acquired in Ce3D solution using a LightSheet.Z1 microscope with ZEN software (Zeiss) and custom imaging chamber, as described77.
HCR probe design and detection
HCR probes targeting axolotl Shh mRNA, probe hybridization buffer, wash buffer, detection hairpins and amplification buffer were from Molecular Instruments. Sequences unique to Shh mRNA were identified by BLAST alignment against axolotl transcriptome assembly Amex.T_v47 (ref. 61). Sequences were considered not unique if they exhibited homology to non-target transcripts at more than 36 of 50 nucleotides. Shh HCR signal was detected using B5 hairpins conjugated to Alexa-647 fluorophore. HCR probes targeting TFP mRNA were purchased at the 50-pmol scale from IDT (Integrated DNA Technologies, oPools), suspended in water and stored at −20 °C. TFP HCR signal was detected using B1 hairpins conjugated to Alexa-546 fluorophore.
Quantifying Hand2:EGFP expression during regeneration
For whole-tissue measurements
Axolotls (4.5 cm) harbouring Hand2:EGFP were amputated through the middle of the lower arm and imaged throughout regeneration with identical acquisition settings using an AXIOzoom V16 wide-field microscope (Zeiss). Longitudinal imaging of 6 limbs was done on days 0, 2, 4, 7, 10, 15, 21, 28 and 39 after amputation. Mean EGFP fluorescence intensity was measured in Fiji software78 by manually drawing a region of interest in the EGFP-positive region of the blastema. At 0, 2 and 4 d.p.a., no or little blastema had formed, so measurements were instead taken at 500 μm behind the amputation plane. The 500-μm source zone for lower-arm regeneration was established in ref. 79.
For single-cell measurements
The Hand2:EGFP intensity of mature arm cells, 7 d.p.a. blastema cells and 14 d.p.a. blastema cells were compared by flow cytometry. Lower-arm tissue was removed from 6-cm Hand2:EGFP axolotls. The entire lower arm was removed for mature measurements. Blastemas were generated by amputating through the middle of the lower arm 7 or 14 days before flow cytometry. Tissues removed were then dissociated into single-cell suspensions using Liberase TM enzyme (Merck, 05401127001) as described previously63, with the following modifications: dissociation was done for 55 min (mature sample) or 45 min (blastema samples), and the cells were filtered through a 70-μm MACS SmartStrainer (Miltenyi Biotec, 130-098-462). Cells were analysed by FACS (using a FACSAria III Cell Sorter, BD Biosciences) with a 100-μm low-pressure nozzle. Mean fluorescence intensities were quantified using FLOWJO software (BD Biosciences).
ALM
ALM experiments were performed on the upper arm of the axolotl, as described80. For Hand2 CRISPant ALMs, donor axolotls (F0 Hand2 CRISPant) were 9–10 cm in size and host axolotls (GFP-expressing controls) were 13–14 cm. Donor skin grafts (1 mm × 1 mm) were transplanted distal to the deviated nerve on host animals. Donor grafts were removed from Hand2 CRISPants deemed to have a high mutation rate, as judged by regeneration of a hypomorphic spike after a previous lower-arm amputation. As controls, skin grafts were removed from sibling axolotls injected with Cas9 but no sgRNA. Hand2 CRISPant ALM and control grafts were done on opposite arms of each host axolotl. Then, 58 days after surgery, ALMs were removed and fixed for skeletal staining using Alcian blue and Alizarin red. For Hand2 M146 CRISPant ALMs, the procedures were similar, except that donor axolotls (F0 Hand2 M146 CRISPants generated using sgRNAs 1–3, or sibling controls injected with control mix) were 7 cm in size and host axolotls (d/d) were also 7 cm. ALMs were analysed 27 days after grafting. Hand2 misexpression ALMs were performed similarly. Donor axolotls (strong Prrx1>Hand2) were 5 cm and host axolotls (unlabelled controls) were 8 cm. As controls, skin grafts were removed from weak Prrx1>Hand2 siblings. ALM formation was deemed to have been completed by 34 days after surgery.
Alcian blue and Alizarin red staining
Skeletal staining of fixed accessory limbs was performed as described previously81. Stained limbs were imaged in 70% ethanol/water using an AxioCam ERc 5s colour camera (Carl Zeiss Microimaging) mounted on an Olympus SZX10 microscope. Alcian blue (A3157), Alizarin red (A5533) and Trypsin (85450C) were from Merck.
Cell transplantation by injection of FACS-sorted cells
Upper and lower arms were removed from 4-cm double-reporter axolotls (Alx4:mCherry_Hand2:EGFP) and dissociated into single-cell suspensions using Liberase TM enzyme (Merck, 05401127001) as described63, with the following modifications: dissociation was done for 55 min and the cells were filtered through a 70-μm MACS SmartStrainer (Miltenyi Biotec, 130-098-462). Anterior cells (mCherry-positive plus EGFP-negative) or posterior cells (EGFP-positive) were purified by FACS (FACSAria III Cell Sorter, BD Biosciences) using a 100-μm low-pressure nozzle and collected into separate tubes of amphibian culture medium64. Pelleting and injection of FACS-sorted cells into the arms of 4-cm unlabelled sibling axolotls was done as described63, using a Nanoject II injector (Drummond Scientific Company). Anterior cells were injected into the posterior lower arm, whereas posterior cells were injected into the anterior lower arm. We injected 9,000 cells per experiment. Host arms were imaged 2 days after injection to confirm successful transplantation using an AXIOzoom V16 wide-field microscope (Zeiss). At 15 days, host arms were amputated distal to the transplant. Regenerating blastemas were imaged at 6, 8, 10, 15 and 26 d.p.a. until the limb was considered fully regenerated. At 26 d.p.a., a second amputation was done through the regenerated part of the limb to generate a second blastema and test whether cells had altered their positional memory. This second blastema was removed at 8 d.p.a. and fixed and processed for whole-mount HCR staining against Shh transcripts.
Isolation of cells for RNA sequencing
For Hand2-misexpression experiments
We amputated 16-cm Prrx1>mCherry–Hand2 (test) or Prrx1>mCherry (control) axolotls through the mid-zeugopod. The animals also had ZRS>TFP in the genetic background, but this was not used in this experiment. We analysed Prrx1>mCherry–Hand2 limbs with weaker phenotypes, because strong misexpression resulted in no limb. At 14 d.p.a., the anterior part of each blastema was removed and dissociated into single cells using Liberase TM enzyme, as described above. We FACS-purified mCherry-positive cells into amphibian culture medium64 using one animal (two blastemas) per replicate. There were three replicates for Hand2 misexpression blastemas and two replicates for mCherry controls.
For A→A and A→P cells
We took upper and lower arms from uninjured 7.5-cm double-reporter axolotls (Alx4:mCherry_Hand2:EGFP) and dissociated them into single cells using Liberase TM enzyme, as described above. We purified mCherry-positive, EGFP-negative anterior cells for injection. We did cell injections as described above. About 25,000 sorted events were injected into the anterior or posterior zeugopod of unlabelled, uninjured sibling animals. Two days after injection, limbs were amputated through the distal part of the graft to induce regeneration. After 47 days, the regenerated limb was re-amputated through the distal part of the graft to induce a second blastema. This second blastema was removed at 12 d.p.a. Blastemas were dissociated into single cells as described above and prepared for FACS. In A→A experiments, we purified mCherry-positive anterior cells regardless of EGFP expression. In A→P experiments, we purified EGFP-positive posterior cells regardless of mCherry expression. Cells were sorted into amphibian culture medium64. One injected arm was used per replicate. By the time of removal, animals were approximately the same size as those used for anterior and posterior cell isolation (below).
For anterior and posterior cells
We amputated 11-cm double-reporter axolotls (Alx4:mCherry_Hand2:EGFP) at the mid-zeugopod. At 15 d.p.a., whole blastemas were removed and dissociated into single cells using Liberase TM enzyme, as described above, and prepared for FACS. To isolate anterior blastema cells, we purified mCherry-positive, EGFP-negative cells. To isolate posterior blastema cells, we purified EGFP-positive cells, regardless of mCherry expression. Anterior and posterior blastema cells in the same replicate were isolated from the same blastema preparation (paired samples). We used seven animals to generate four replicates each of anterior and posterior blastema cells. Cells were sorted into amphibian culture medium64.
RNA library preparation, sequencing and analysis
We proceeded immediately to RNA isolation after cell sorting. RNA was purified into 20 μl nuclease-free water using an in-house magnetic bead-based isolation kit. The following RNA inputs were used to construct libraries: Hand2 overexpression and mCherry control (0.7 ng each); anterior/posterior high input (45 ng each); anterior/posterior low input (4.5 ng each); and A→A and A→P (4.5 ng each). Sequencing libraries were constructed using QuantSeq 3′ mRNA-Seq V2 REV kit with unique dual indices (Lexogen). The low-input protocol was used for samples with less than 10 ng input, and RNA removal was reduced to 5 min for the 0.7 ng libraries, as suggested by the manufacturer. Sequencing libraries were purified into 18 μl nuclease-free water. Each replicate was sequenced to a depth of around 50 M reads, in PE 150 mode (read 2 + CSP-read1), distributed over two lanes of a NovaSeq X 1.5 B flowcell. Sequencing was done by the Next Generation Sequencing Facility at Vienna BioCenter Core Facilities (VBCF), part of the Vienna BioCenter (VBC), Austria. These sequencing data have been deposited at the Gene Expression Omnibus (GEO) with accession number GSE284768.
Gene expression analysis was done using DESeq2, similarly to the dermal cell samples (see the ‘Gene expression analysis’ section, above). A threshold of 100 counts was used. PCA was performed using anterior/posterior low-input libraries (to match the inputs of the A→A and A→P samples), and the other gene expression comparisons were done against the anterior/posterior high-input libraries. We used a threshold of padj < 0.05 for significant differential expression.
Dilution and storage of BMS-833923 and SAG
BMS-833923 (Cayman Chemical, 16240) was dissolved to 10 mM in ethanol and stored as single-use aliquots at −70 °C. InSolution Smoothened Agonist (SAG, Merck 566661) for intraperitoneal injections was obtained at 10 mM in water and stored as single-use aliquots at −20 °C. SAG for bathing experiments was purchased from Merck (566660), diluted to 40 μM and stored as single-use aliquots at −20 °C.
Skin transplantation plus intraperitoneal delivery of BMS-833923
Several modifications were made to the assay relative to the anterior-to-posterior and posterior-to-anterior transplantations by cell injection. One limitation of the Alx4:mCherry_Hand2:EGFP double reporter was that cells would lose fluorescence if they lost anterior and posterior identity. To avoid this, we substituted Alx4:mCherry for Prrx1>mCherry, which expresses mCherry regardless of positional identity. This enabled continuous monitoring of the transplant, and Hand2:EGFP labelled posterior identity. Owing to limited animal availability, we used hindlimbs as donors instead of forelimbs (positional memory in forelimb and hindlimb are compatible82). To reduce the number of donors required, we transplanted cells by skin grafting instead of cell injection.
Skin areas (about 1 mm × 1 mm) were transplanted from the anterior lower leg of 7-cm double-reporter axolotls (Prrx1>mCherry_Hand2:EGFP) to the posterior lower arm of 8-cm unlabelled d/d hosts, maintaining dorsal–ventral and proximal–distal directionality. Four days after transplantation, the host arm was amputated through the distal third of the transplant, and blastema outgrowth was monitored every 1–2 days using an AXIOzoom V16 wide-field microscope (Zeiss). At 6 d.p.a. (the conical blastema stage), test axolotls were injected intraperitoneally with 25 μl of BMS-833923 diluted to 1 mM in water. Control axolotls instead received the appropriate dilution of ethanol in water injection. Injection mix contained Fast Green dye (Thermo Fisher Scientific) for visual contrast. Blastemas were further imaged 4 and 15 days after injection to assess for changes in positional identity.
Assessing blocking Shh signalling on Hand2:EGFP expression
Hand2:EGFP axolotls (7 cm) were amputated at the top of the lower arm. Every 3 days from 0 d.p.a. until 21 d.p.a., test axolotls were injected intraperitoneally with 20 μl of BMS-833923 diluted to 1 mM in water, whereas control axolotls were instead injected with 20 μl of water. Injection mix contained Fast Green dye (Thermo Fisher Scientific) for visual contrast. Blastemas were imaged every three days using an AXIOzoom V16 wide-field microscope (Zeiss). Mean Hand2:EGFP fluorescence was quantified from manually defined regions of interest in the posterior blastema.
SAG positional-memory experiment
Hand2 lineage-traced axolotls were prepared by treating stage-42 embryos with 2 μM 4-OHT, as described in the section ‘Genetic lineage tracing (ZRS/Hand2)’, above. At a size of 8 cm, each axolotl had the right arm amputated through the middle of the lower arm (blastema assay). The left arm was left intact (mature assay). At 8 d.p.a., test axolotls were injected intraperitoneally with 20 μl of SAG diluted to 1.5 mM in water. Control axolotls instead received a water injection. Injection mix contained Fast Green dye (Merck F7258) for visual contrast. Both the blastema limb and the mature limb were imaged every few days until 25 days after injection using an AXIOzoom V16 wide-field microscope (Zeiss). On day 25, both limbs were amputated through the hand-plate region to assay for effects on positional memory from the expression of Hand2:EGFP.
SAG positional-memory experiment (Shh HCR assay)
Hand2:EGFP axolotls (5 cm) were amputated through both lower limbs, then bathed in water (control) or 10 nM SAG (test) for the first 21 days of regeneration. Bathing volume was 40 ml, and solution was prepared and exchanged daily, following the protocol of ref. 1. Regeneration was deemed to be complete at 30 d.p.a. Axolotls were raised for a further 30 days in water, to ensure complete washout of SAG from test animals. Subsequently, axolotls were re-amputated through the hand-plate region to generate a new blastema in the reprogrammed part of the limb. At 9 d.p.a., the new blastemas were removed and fixed for whole-mount HCR staining against Shh mRNA, tissue clearing and light sheet imaging.
Baculovirus production and injection
Pseudotyped baculovirus was produced as described in ref. 83. BV-mCherry, a control baculovirus to misexpress mCherry, was published previously as chBV83. The cytomegalovirus immediate-early promoter (pCMV) drives expression of mCherry in infected cells. BV-Shh, to misexpress axolotl Shh, was generated for this study. pCMV drives the expression of nuclear-localized mCherry T2A axolotl Shh. Co-translational cleavage in the T2A sequence releases full-length axolotl Shh protein. Axolotl Shh was codon-altered to enable the distinction of virally expressed mRNA from endogenous axolotl Shh mRNA.
Either BV-mCherry or BV-Shh was injected into the anterior lower arm of 4-cm Hand2:EGFP axolotls. The injection mix contained Fast Green dye (Merck, F7258) for visual contrast. Then, 18 days after infection, limbs were amputated through the middle of the lower arm. The regenerating blastema was imaged every few days using an AXIOzoom V16 wide-field microscope (Zeiss). At 11 d.p.a., blastemas were removed for fixation, whole-mount tissue clearing and imaging using a LightSheet.Z1 microscope (Zeiss).
Image analysis
Microscope images were analysed using ZEN software (Zeiss) or Fiji software78.
Statistical analysis and data representation
Statistical analyses and graph plotting were done using Prism software (GraphPad). Data were tested for assumptions of normality and equality of variance to determine the appropriate statistical tests to perform. Measurements were taken from distinct samples unless indicated otherwise. No data were excluded. Mean values are reported ± s.d. Statistical significance was defined as P < 0.05. All figures were made using Adobe Illustrator.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
