Evolutionary dynamics of whole-body regeneration across planarian flatworms

Abstract

Regenerative abilities vary dramatically across animals. Even amongst planarian flatworms, well-known for complete regeneration from tiny body fragments, some species have restricted regeneration abilities while others are almost entirely regeneration incompetent. Here, we assemble a diverse live collection of 40 planarian species to probe the evolution of head regeneration in the group. Combining quantification of species-specific head-regeneration abilities with a comprehensive transcriptome-based phylogeny reconstruction, we show multiple independent transitions between robust whole-body regeneration and restricted regeneration in freshwater species. RNA-mediated genetic interference inhibition of canonical Wnt signalling in RNA-mediated genetic interference-sensitive species bypassed all head-regeneration defects, suggesting that the Wnt pathway is linked to the emergence of planarian regeneration defects. Our finding that Wnt signalling has multiple roles in the reproductive system of the model species Schmidtea mediterranea raises the possibility that a trade-off between egg-laying, asexual reproduction by fission/regeneration and Wnt signalling drives regenerative trait evolution. Although quantitative comparisons of Wnt signalling levels, yolk content and reproductive strategy across our species collection remained inconclusive, they revealed divergent Wnt signalling roles in the reproductive system of planarians. Altogether, our study establishes planarians as a model taxon for comparative regeneration research and presents a framework for the mechanistic evolution of regenerative abilities.

Fig. 1. The MPI-NAT planarian collection.

a, Sampling sites (black dots) and collection location (red). b–d, Live images of flatworms that were characterized as part of this study. b, Planarians order Tricladida, Suborder Continenticola. From left to right. First row, Bdellocephala angarensis; Bdellocephala cf. brunnea; Dendrocoelum lacteum; Crenobia alpina; Polycelis felina; Polycelis tenuis; Polycelis nigra; Seidlia sp.; cf. Atrioplanaria. Second row, Phagocata gracilis; Hymanella retenuova; Phagocata pyrenaica; Planaria torva; Cura pinguis; Cura foremanii; Schmidtea lugubris; Schmidtea polychroa (dark strain); Schmidtea polychroa (unpigmented strain). Third row, Schmidtea nova; Smed (asexual strain); Dugesia tahitiensis; Dugesia sicula; Dugesia sp.; Dugesia japonica; Girardia dorotocephala; Girardia tigrina; Spathula sp. 3. c, Planarians order Tricladida, Suborder Maricola: Camerata robusta; Procerodes littoralis; Procerodes plebeius; Bdelloura candida; Cercyra hastata. d, Order Prolecithophora: Plagiostomum girardi; Cylindrostoma sp. Scale bar, 1 mm unless otherwise noted; measured (pixel resolution; solid line) or approximated during live imaging (graph paper; dotted line).

Fig. 2. Quantitative analysis of head-regeneration abilities across the flatworm collection.

a, Cartoon of the serial head-regeneration assay (left) and representative outcomes in the indicated species. Dashed lines: amputation planes and typical regeneration outcomes (small images) and their relative frequency (number pairs; ‘Dead’ in case of no survivors) at the respective A–P position. Colours designate the head-regeneration classification scheme used throughout this study: group A (green): robust regeneration (efficient head regeneration at all A–P axis positions); group B (orange): restricted regeneration (position-dependent head-regeneration defects); group C (red): poor regeneration (no head regeneration at any A–P position). Missing images in the C. hastata panel correspond to pieces that did not regenerate a head and died before imaging (No head). b, Graphical representation of the data in the form of head frequency curves. The percentage of successful head regeneration is plotted either as a fraction of the initial (black line) or surviving (red line) fragment numbers. Dashed line, percentage of fragment survival. The Roman numerals below designate the more fine-grained head-regeneration scheme by Šivickis. c, Left, head frequency curves (percentage of surviving fragments) for all analysed species and colour-coded as in a. Right, number of species in each head-regeneration category. d, Unclassified regeneration defects and their relative frequencies in the indicated collection species, including bipolar double-heads (left), double-tails (centre) and mediolateral axis duplication (right). Amputation paradigms are cartooned to the left; the respective regeneration outcome in Smed is shown for reference. e, Phylogenetic overview of flatworm groups related to planarians (left); amputation paradigm (centre); and representative images and quantifications of head-regeneration failures in the indicated Prolecitophora species (right). Scale bar, 1 mm unless otherwise noted.

Fig. 3. Phylogenetic analysis of head-regeneration abilities in planaria.

a, De novo high-quality transcriptome assembly pipeline and unbiassed gene sequence selection workflow for multigene phylogeny reconstruction. RNA-seq, RNA sequencing. b, BUSCO quality comparison between the indicated transcriptome sources, plotting BUSCO gene sequence completeness (x axis) versus fragmentation (y axis). ‘Perfect’ BUSCO representation: bottom right corner. c, Presence (black)/absence (turquoise) analysis of individual BUSCO genes (y axis) in flatworm transcriptomes (x axis, in phylogenetic order) and representative outgroups. Names of representative species are indicated (see Supplementary Table 2 for detail). The blue frame designates BUSCO genes absent from >90% of the transcriptomes. The red inset shows a selection of likely planarian-specific BUSCO loss events. d, Maximum-likelihood tree on the basis of the transcriptomes with representatives of all the suborders of the Tricladida. Numbers indicate outgroups and the preferred habitats of major planarian taxonomic groups are indicated by colour shading (see legend). e, Phylogenetic map of planarian head-regeneration abilities and ASR. Pie charts summarize the most likely regeneration ability at selected internal nodes (for other nodes, see Supplementary Fig. 3f). Arrows indicate a parsimonious history of gains and losses of regeneration ability given the ASR. Nodes represent the proportion of character histories with the indicated state for regenerating capability (Supplementary Fig. 3f and Supplementary Tables 3, 4 and 5; Methods). The colour coding of the columns to the right indicates species- and clade-specific head-regeneration abilities quantified by this study or extracted from the literature (both used for clade annotation). See Supplementary Table 1 for the species abbreviations used in the figure. O, Order; S.O., Suborder; S.F., Superfamily; F., Family.

Fig. 4. Canonical Wnt pathway inhibition rescues head-regeneration defects across planarian phylogeny.

a, Fluorometric Western blot demonstration of anti-ß-CATENIN-1 antibody G78 crossreactivity with tail tip lysates of different planarian species. Species and amount of lysate/lane as indicated. b, RNAi-mediated Wnt pathway activity modulation in the indicated species from regeneration groups A, B or C (colour coding). Animal cohorts were treated with enhanced green fluorescent protein (eGFP) (RNAi) (control), APC(RNAi) (gain of Wnt signalling) or ß-catenin-1(RNAi) (loss of Wnt signalling; Methods). Top, representative quantitative Western blots for ß-CATENIN-1 and histone H3 as loading control. Bottom, bar graph representation of G78 signal intensity relative to the control in each species (eGFP(RNAi)). Data are presented as mean ± s.d. of the mean of four technical replicates (dots). c, Head-regeneration rescue assay upon Wnt inhibition/ß-catenin-1(RNAi) in the indicated category B or C species. Live images at the indicated day postamputation (dpa) and RNAi conditions; number pairs represent the observed frequency of the shown phenotype; head or eye regeneration. Red triangles, regenerated eyes. Scale bar, 500 µm. ph, pharynx. d, Phylogenetic representation of documented head-regeneration rescued by ß-catenin-1(RNAi) (green check-mark) (phylogeny extracted from Fig. 3d). This study (red): C. pinguis, P. torva, C. robusta and B. candida. Previous studies (blue): D. lacteum and P. fluviatilis. Phagocata kawakatsui, for which head-regeneration rescue was also reported, was excluded due to the lack of a publicly available transcriptome. ß-Cat-1, ß-catenin-1.

Fig. 5. Wnt signalling functions in the Smed reproductive system.

a, Co-occurrence of robust head regeneration (Group A, red) with fissiparous reproduction across planarian clades. b, Live images illustrating the sexual (egg-laying; left) and asexual (fissiparous; right) reproduction modes of Smed laboratory strains. c, Cartoon (top) and colorimetric whole-mount in situ hybridizations (indicated markers; middle) of Smed sexual strain reproductive system components; reproductive system ablation under ophis(RNAi) (bottom). d, ß-CATENIN-1 amounts/Wnt signalling activity in A–P sections (1, head; 6, tail) quantified via Western blotting with the G78 mAB. Smed strain and RNAi conditions as indicated. Error bars, s.d. of n = 4 biological replicates, each representing the mean of four technical replicates (blots). e, Wnt component expression in the Smed reproductive system by colorimetric or fluorescent whole-mount in situ hybridization. Number pairs, specimen fraction displaying the pattern shown; red arrowhead, testes lobules; black arrowhead, expression in non-reproductive tissues; H3P, Histone H3 Ser10 phosphorylation immunolabelling; DAPI, nuclei. f, Colorimetric whole-mount in situ hybridizations of the indicated reproductive system markers under the indicated RNAi conditions; number pairs, specimen fraction displaying the pattern shown; dashed lines, approximate position of the sagittal sections in Fig. 5g. g, Relative yolk gland cross-sectional area quantifications in histological sections of the indicated RNAi conditions. Top, bar graph, n = 2 individuals (dots)/condition; error bars, s.d. of the mean. Bottom, representative Mallory-stained sagittal sections. White outline, yolk glands; red boxes, zoom views. y, yolk; vnc, ventral nerve cord. h, Coomassie-stained SDS–PAGE of lysates of the indicated sources. White box, major yolk protein. i, Fluorescent Western blot of the indicated lysates, probed with the indicated antibodies. j, EO95 specificity analysis in Smed sexual strain lysates under the indicated ferritin(RNAi). Asterisk, EO95 signal loss in ferritin-C(RNAi). k, Control (eGFP)-normalized quantification of FERRITIN-C and ß-CATENIN-1 amounts in lysates of the indicated RNAi condition via EO95 and G78 immunoblotting. Error bars, s.d. of the mean of four technical replicates (dots) in n = 1 biological replicate. l, Colorimetric whole-mount in situ hybridization of the four yolk ferritins in sexual Smed under the indicated RNAi treatments. Scale bar, 1 mm unless otherwise noted.

Fig. 6. Model and model testing.

a, Model; see text for details. b, Top, quantitative Western blot analysis of yolk content (EO95 mAb; H3, loading control) in the indicated species and RNAi conditions. Bottom, bar graph representation of RNAi-control-normalized EO95 signal. Error bars, s.d. of two to four technical replicates (dots) of n = 1 biological replicate. c, Yolk gland cross-sectional area quantifications in sagittal sections of the indicated species and RNAi treatments. White oulines, yolk glands; red frames, zooms. d, Bar graph representation of c. Error bars, s.d. of the mean of n = 4 individuals (dots), each representing the mean of five technical replicates. e, Correlation between yolk gland cross-sectional area and regenerative abilities in the indicated egg-laying species. Species ordering by relative yolk content quantified as in c,d; colours: head-regeneration abilities (green, robust/group A; orange, restricted/group B; red, poor/group C). Error bars, s.d. of the mean of n = 4 individuals (dots), each representing the mean of five technical replicates. f, Statistical analysis of the data in e, replotted as log₁₀ of mean yolk content/individual in group A versus B and C species. Red line, distribution mean. Significance assessment via linear mixed model; the indicated P value implies statistical significance. g, Calibration of the G78 mAb for interspecies comparisons. Western blot of recombinant His-tagged ß-CATENIN-1 fragments of the indicated species and probed with anti-penta-His Ab (top) or G78 mAb (bottom); ratio, species-specific correction factor. Molecular weight marker as indicated. h, Correlation between Wnt pathway activity, regenerative abilities (colour coding as in e) and reproduction mode (bottom) across the indicated species. Species ordering by ß-CATENIN-1 concentrations in tail lysates, as per corrected G78 mAB signal. Error bars, s.d. of the mean of n = 3 biological replicates (dots), each representing the mean of four technical replicates. i, Statistical analysis of the data in h, replotted as log₁₀ of the mean ß-CATENIN-1 tail tip concentration in fissiparous versus egg-laying strains/species (left) or robust versus restricted or poor regeneration groups (A versus B and C). Red line, distribution mean. Significance assessment as in f; the indicated P values imply lack of significance.