Smed-βcatenin-1 Is Required for Anteroposterior Blastema Polarity in Planarian Regeneration

See allHide authors and affiliations

Science  18 Jan 2008:
Vol. 319, Issue 5861, pp. 327-330
DOI: 10.1126/science.1149943


Planarian flatworms can regenerate heads at anterior-facing wounds and tails at posterior-facing wounds throughout the body. How this regeneration polarity is specified has been a classic problem for more than a century. We identified a planarian gene, Smed-βcatenin-1, that controls regeneration polarity. Posterior-facing blastemas regenerate a head instead of a tail in Smed-βcatenin-1(RNAi) animals. Smed-βcatenin-1 is required after wounding and at any posterior-facing wound for polarity. Additionally, intact Smed-βcatenin-1(RNAi) animals display anteriorization during tissue turnover. Five Wnt genes and a secreted Frizzled-related Wnt antagonist-like gene are expressed in domains along the anteroposterior axis that reset to new positions during regeneration, which suggests that Wnts control polarity through Smed-βcatenin-1. Our data suggest that β-catenin specifies the posterior character of the anteroposterior axis throughout the Bilateria and specifies regeneration polarity in planarians.

The ability to regenerate is widespread and varies in degree within the animal kingdom. Planarian flatworms are dramatic examples, capable of regenerating a complete animal from nearly any fragment of the organism. Amputation induces the formation of a regeneration blastema, which differentiates to produce missing structures. Blastema formation requires adult stem cells called neoblasts (1). More than a century ago, Randolph (2) and Morgan (3) noted that transverse planarian pieces can regenerate a head and a tail at the anterior and posterior wound sites, respectively. Furthermore, any body region can become involved in either head or tail regeneration depending upon the identity of the missing tissue (Fig. 1A). Therefore, planarians possess a robust system throughout the anteroposterior (A-P) axis for specifying the identity of missing tissue at wounds. This system has been named “polarity.” Morgan observed that very thin transverse pieces occasionally regenerate animals with two heads and hypothesized that a gradient of material establishes regeneration polarity. Although these features of planarian regeneration are frequently described in textbook examples of polarity, no previous explanation exists for the underlying genetic mechanisms that specify regeneration polarity.

Fig. 1.

Inhibition of Smed-βcatenin-1 results in posterior blastema polarity reversal. Day 10 of regeneration. (A) Regeneration polarity: cells between planes “I” and “II” can become involved in head or tail regeneration (27). (B) Cartoon, amputations (red); arrow, indicates trunk fragments shown. (C) Anterior, top. PR, photoreceptors. Red dots, approximate blastema/old tissue boundary. A control unc-22(RNAi) trunk produced a posterior tail (n = 114/114), and a Smed-βcatenin-1(RNAi) trunk fragment produced a head at the posterior wound (n = 105/107). (D and E) Anterior, left. (D) The photoreceptors (PR), cephalic ganglia (cg), and ventral nerve cords (vnc) of regenerating trunks were labeled with antibodies (VC-1, anti-arrestin; cg, anti-synaptotagmin, SYT). A Smed-βcatenin-1(RNAi) posterior blastema showed VC-1 and SYT staining (n = 6/6). (E) A Smed-βcatenin-1(RNAi) posterior head lacked the normal posterior expression of H.1.3b (n = 6/6). Arrow, posterior blastema. Scale bars, 500 μm (C) and 200 μm [(D) and (E)].

The planarian Schmidtea mediterranea has emerged as a powerful molecular genetic system for studying regeneration because of genome sequence availability, the ability to use RNA interference (RNAi) (4, 5), and the evolutionary position of the organism within an understudied protostome superphylum, the Lophotrochozoa (6). To study polarity in planarians, we performed an RNAi screen involving surgically removed heads and tails (Fig. 1B). We identified a single gene required for polarity and named the gene Smed-βcatenin-1. The trunks of Smed-βcatenin-1(RNAi) animals regenerated apparent heads at both posterior-facing and anterior-facing wounds (Fig. 1C). This dramatic phenotype is reminiscent of the polarity transformations observed in Morgan's classic surgical manipulation experiments. Double-stranded RNA (dsRNA) from Caenorhabditis elegans unc-22, a gene not present in planarians, served as the RNAi control (Fig. 1C). The predicted SMED-βCATENIN-1 protein is homologous to β-catenin proteins found in other animals (fig. S1A). β-catenin proteins mediate canonical Wnt signal transduction to regulate many developmental events (7). SMED-βCATENIN-1 possesses at least nine predicted armadillo repeats and a candidate N-terminal GSK-3β phosphorylation site (GSK-3β–mediated phosphorylation can regulate β-catenin stability) (fig. S1B) (8). Because armadillo repeats can be very degenerate (9), more armadillo repeats may exist and not be recognized in this protein. Smed-βcatenin-1 mRNA was broadly expressed in in situ hybridizations and reduced in Smed-βcatenin-1(RNAi) animals (fig. S2).

Smed-βcatenin-1(RNAi) posterior blastemas contained two apparent photoreceptors (Fig. 1C) and possessed headlike stretching behavior (n = 105/107). These posterior blastemas also contained cell types normally associated with the head, including photoreceptor neurons and cephalic ganglia (Fig. 1D). Posterior blastemas displayed local headlike polarity, with photoreceptors being more distal to the wound site than cephalic ganglia. Anterior-specific transcripts can be observed in the posterior Smed-βcatenin-1(RNAi) blastemas by 72 hours after amputation (fig. S3). Additionally, Smed-βcatenin-1(RNAi) posterior blastemas lacked the tail pattern of H.1.3b-zonadhesin expression (Fig. 1E). Together, these data indicate that posterior Smed-βcatenin-1(RNAi) blastemas are heads.

Smed-βcatenin-1(RNAi) animals regenerated posterior-facing heads at multiple A-P locations, indicating a requirement for Smed-βcatenin-1 throughout the A-P axis (Fig. 2A, fig. S4). We injected Smed-βcatenin-1 dsRNA into freshly amputated wild-type transverse fragments to inhibit Smed-βcatenin-1 only during regeneration. Because these animals displayed a 100% penetrant polarity reversal, Smed-βcatenin-1 is required during regeneration for the polarity decision (Fig. 2B). In addition, because these pieces only contained a small region of preexisting tissue, Smed-βcatenin-1 can be required locally.

Fig. 2.

Smed-βcatenin-1 inhibition affects polarity throughout the A-P axis, locally after wounding, and can result in supernumerary head regeneration. Amputations (red). Anterior, left. Scale bars, 500 microns. (A) Smed-βcatenin-1(RNAi) fragments produced posterior-facing heads at three anteroposterior locations (amputation planes I, II, and III). (B) Transverse fragments were generated 1 hour before dsRNA injection. A Smed-βcatenin-1(RNAi) animal formed a posterior head (n = 10/10). (C) Multiple side incisionsand tail removalweremadetogenerateasix-headed Smed-βcatenin-1(RNAi) animal (asterisks, supernumerary heads). (D) Smed-βcatenin-1(RNAi) did not impair head tip regeneration (arrows, blastema; red dots, approximate blastema/old tissue boundary; n = 10/10). Day of regeneration: (A) and (B), day 11; (C), day 20; (D) day 5.

Multiple incisions made in Smed-βcatenin-1(RNAi) animals resulted in the production of side-facing heads, dramatically illustrated by the six-headed Smed-βcatenin-1(RNAi) animal in Fig. 2C. Heads grew from side incisions at locations far from either pole of the animal (Fig. 2C) and in animals that retained their original posterior pole (fig. S4). Thus, Smed-βcatenin-1 is not required only to specify the posterior-most pole during regeneration. Given this observation, we asked whether Smed-βcatenin-1 is simply required to suppress a complete head-production program at any wound. Smed-βcatenin-1(RNAi) animals could regenerate their head tips, rather than producing an additional head following such injury. This indicates that these animals retained the ability to repair only the extent of tissue missing (Fig. 2D). Therefore, Smed-βcatenin-1 likely does not simply suppress a complete head formation program.

After transverse amputations, animals are challenged to form a new anterior or posterior pole. Planarians can also regenerate other tissues, such as tissue along the entire A-P axis during lateral regeneration of missing sides. After sagittal amputations to remove the lateral half of animals, Smed-βcatenin-1(RNAi) fragments regenerated multiple anterior photoreceptors and side-facing head-like protrusions (Fig. 3A). These animals possessed expanded cephalic ganglia and anterior marker expression in blastemas, as well as ectopic cephalic ganglia in side-facing protrusions (Fig. 3A). Thus, Smed-βcatenin-1(RNAi) animals displayed anteriorization of blastemas in which an entire A-P axis must be patterned. Together, our amputation experiments suggest Smed-βcatenin-1 normally promotes posterior aspects of A-P axis regeneration: tail rather than head formation at posterior poles and A-P regionalization in blastemas that produce tissue along the entire A-P axis.

Fig. 3.

Smed-βcatenin-1 inhibition causes anteriorization during lateral regeneration and homeostatic tissue maintenance. PR and white arrows, photoreceptors. Anterior, left. Scale bars, 500 μm. (A) Smed-βcatenin-1(RNAi) longitudinal fragments regenerated multiple eyes or a side-facing head (e.g., 32/37 and 8/37, respectively, after dsRNA injection and amputation within hours). (Red box inset) Enlargement of photoreceptors. In situ hybridizations with PC2 or PDS riboprobes showed anteriorization (red brackets) and side-facing heads (red asterisks) in Smed-βcatenin-1(RNAi) lateral blastemas. Animals are at day 12 of regeneration. Black arrows, side with blastema. (B to D) Animals between 29 and 35 days of RNAi are shown. Unamputated Smed-βcatenin-1(RNAi) animals displayed (B) posterior heads (see red inset) (n = 47/48), (C) multiple head-like peripheral protrusions (n = 12/25), (D) ectopic photoreceptors in headlike protrusions (VC-1-labeled, white arrows), and [(B) to (D)] protrusions at the pharynx (n = 24/25, yellow arrows) containing photoreceptors (n = 11/25).

Uninjured planarians constantly replace aged differentiated cells. To determine whether Smed-βcatenin-1 is required for instructing the posterior fates of new cells, uninjured Smed-βcatenin-1(RNAi) animals were observed during homeostasis. A slow transformation of the body was observed, in which head-like tissue appeared at the posterior pole (Fig. 3B) and, ultimately, around the periphery of the animals and at the pharynx (Fig. 3C, D). Therefore, Smed-βcatenin-1 is required for the homeostatic maintenance of the A-P axis.

Because β-catenin proteins are well-established effectors of Wnt signaling, we hypothesized that Wnts could provide the A-P positional information necessary for blastema polarity decisions. We characterized five Wnt-family genes in S. mediterranea and named them Smed-wntP-1 through -3, Smed-wnt2-1, and Smed-wnt11-1 (10). As assayed by whole-mount in situ hybridization, these Wnt genes were expressed in distinct domains along the A-P axis and may therefore encode signaling molecules that specify planarian A-P axis polarity (Fig. 4). Expression was detected for Smed-wntP-1 in one to four cells near the tail tip (Fig. 4A) and for Smed-wnt11-1 in a graded fashion from the posterior (Fig. 4B). Expression was detected for Smed-wntP-2 in the posterior half of the animal and internally around the pharynx (Fig. 4C), for Smed-wntP-3 at the anterior pharynx end (Fig. 4D), and for Smed-wnt2-1 laterally in the anterior half of the animal (Fig. 4E). RNAi of Wnt genes did not result in posterior head regeneration (n > 25 animals each), possibly because of redundancy among the many posteriorly expressed Wnt genes. Wnt signaling is known to be inhibited by secreted frizzled-related proteins (sFRP) (11). We cloned an S. mediterranea sFRP gene (Smed-sFRP-1) (10) and found it to be expressed at the anterior pole (Fig. 4F). The Smed-wnt genes Smed-wntP-1 through -3, as well as Smed-wnt11-1, became expressed near posterior-facing wounds (Fig. 4G-J); Smed-wnt11-1 and Smed-wntP-2 expression also were more posteriorly restricted in tail fragments (Fig. 4H, I). By contrast, Smed-wnt2-1 was expressed near anterior-facing wounds and restricted anteriorly in heads (Fig. 4K). Following amputation, Smed-sFRP-1 was expressed at new anterior poles (Fig. 4L). These homeostatic and regeneration expression patterns suggest that the A-P planarian axis is regulated by anterior Wnt inhibition and posterior Wnt activation.

Fig. 4.

Planarian Wnt genes and Smed-sFRP-1 are regionally expressed along the A-P axis during homeostasis and regeneration. Anterior, top. Scale bars, 200 μm. Arrow, Smed-wnt or Smed-sFRP signal. (A to F) Dorsal views of in situ hybridizations of planarian Wnt genes in intact and (G to L) in day 4 to 5 regenerating head, trunk, and tail fragments. (G to L) Cartoons depict expression patterns (purple dots), as well as surgical strategy (red lines). Smed-wnt genes were expressed in overlapping anteroposterior domains, with the majority in the posterior; Smed-sFRP-1 was expressed at the anterior pole. New expression was observed near wounds in regeneration. Images were taken with gamma = 0.67.

Our data suggest that a genetic program—canonical Wnt signaling mediated by Smed-βcatenin-1—controls planarian A-P tissue identity in homeostasis and blastema polarity in regeneration. We suggest that Smed-βcatenin-1 acts at posterior-facing wounds to effect the posterior rather than anterior polarity regeneration decision and that polarity is controlled by the A-P location of expression of Wnt genes and Wnt signaling antagonists. Wnt signaling has come to be used for a myriad of patterning and cell proliferation roles in the course of the evolutionary diversification of the three superphyla comprising bilaterally symmetric animals: the deuterostomes and two groupings from the protostomes, the Ecdysozoa and the Lophotrochozoa (6, 7, 1215). Among these roles, our data provide evidence that an ancestral role of Wnt signaling is to promote posterior aspects of A-P axis specification and regulation. This hypothesis is well supported by data from other organisms (16). For example, in multiple deuterostomes, Wnt and secreted Wnt inhibitor genes are expressed in a manner that resembles the A-P polarity of expression seen in planarians. Specifically, Wnt genes are expressed in the posterior of Xenopus embryos after the midblastula transition (17, 18), as well as in the posterior of cephalochordate embryos (19). Furthermore, Wnt genes and a secreted Wnt antagonist are expressed at opposite (oral and aboral) poles of radially symmetric cnidarian larvae (12, 20), and pharmacologic inhibition of the β-catenin–inhibitory GSK-3β protein in hydra leads to expanded oral fates (21). The conservation of Wnt axial expression features in these diverse animals, combined with the requirement for Smed-βcatenin-1 in A-P regeneration polarity, suggest a functionally critical and deeply conserved role for β-catenin in A-P axis polarity. In ecdysozoans, β-catenin controls anteroposterior axon polarity and cell division polarity in C. elegans (22, 23), anteroposterior segment polarity in Drosophila (24), posterior segment patterning in the intermediate germ insect Gryllus bimaculata (25), and animal-vegetal cell polarity during embryogenesis of the lophotrochozoan annelid Platynereis dumerilii (26). We propose that WNT signaling controls A-P axial polarity features of most if not all bilaterally symmetric animals and determines the polarity of regeneration.

Supporting Online Material

Materials and Methods

Figs. S1 to S4


References and Notes

View Abstract

Navigate This Article