Src acts with WNT/FGFRL signaling to pattern the planarian anteroposterior axis

Robust pattern control is fundamental to the process of regeneration (Wolpert, 1969). Animals must be able to re-establish tissue identity and proper polarity after injury for regeneration to proceed normally. Furthermore, regardless of regeneration abilities, many animals must also maintain regional identity throughout adult life as they replace and specify new cells to replenish old tissue. Planarians present a powerful system for studying these patterning control mechanisms, as they possess a remarkable ability to regenerate any missing body part and are in a state of constant cellular turnover to replace aged tissues (Elliott and Sanchez Alvarado, 2013; Reddien, 2018; Rink, 2018). Planarian regeneration abilities extend from a population of pluripotent stem cells, termed neoblasts, which continuously produce all adult cell types (Wagner et al., 2011; Zeng et al., 2018). Planarian muscle cells harbor positional information used in controlling neoblast differentiation and targeting through expression of regionalization determinants termed positional control genes (PCGs) (Witchley et al., 2013; Scimone et al., 2017). PCGs include signaling molecules in the Wnt, FGF and BMP pathways that control tissue identity along the anteroposterior (AP) axis (from head to tail), the dorsoventral (DV) axis (from back to belly) and the mediolateral (ML) axis (from midline to lateral edge). These factors are expressed in regional territories in uninjured animals that are reset during the regeneration process, and their inhibition leads to mispatterning phenotypes. However, the signaling mechanisms controlling positional information domains in muscle are not yet fully understood.

Significant progress has been made in understanding the regeneration of the planarian AP axis, which is driven by Wnt signaling. Several of the nine planarian Wnt genes are expressed in overlapping domains from the posterior, whereas Wnt inhibitors demarcate nested anterior domains. In recent years, functions of many of these factors have been elucidated in the planarian Schmidtea mediterranea. A canonical β-catenin-dependent Wnt signaling pathway controls head-versus-tail identity of blastemas after transverse amputation. Downregulation of Wnt pathway components β-catenin-1, wnt1, Evi/wntless, Dvl-1/2 or teashirt causes regeneration of ectopic heads (Petersen and Reddien, 2008, 2009; Gurley et al., 2010; Iglesias et al., 2011; Owen et al., 2015; Reuter et al., 2015) whereas upregulation of canonical Wnt signaling via RNAi inhibition of Wnt negative regulators notum and APC causes regeneration of ectopic tails (Gurley et al., 2008; Petersen and Reddien, 2011). wnt1 and notum are both transcriptionally induced by injury, during which they likely participate in control of polarization or orientation of outgrowing blastemal tissue. An activin-dependent process restricts the initial 6-18 h of notum expression to anterior-facing wounds, resulting in a low Wnt environment leading to head regeneration (Cloutier et al., 2021). At later times in regeneration (by 24-72 h) and throughout homeostasis, stem cell-dependent processes (Hayashi et al., 2011; Currie and Pearson, 2013; März et al., 2013; Scimone et al., 2014; Vasquez-Doorman and Petersen, 2014; Vogg et al., 2014; Tejada-Romero et al., 2015; Schad and Petersen, 2020) generate muscle cells expressing wnt1 and notum at the posterior and anterior midline termini (termed poles), respectively, where they may control region-specific patterning or act at the tip of a hierarchy of AP regulatory factors (Adell et al., 2009; Petersen and Reddien, 2009; Gurley et al., 2010; Stuckemann et al., 2017; Schad and Petersen, 2020).

Other Wnt-dependent pathways may function downstream or in parallel to pole identity and tissue polarization. wnt11-6 (also known as wntA) and associated factors limit the regionalization of head tissue. Inhibition of wnt11-6 or the fzd5/8-4 Wnt receptor causes posterior expansion of the brain and formation of ectopic posterior eyes (Kobayashi et al., 2007; Adell et al., 2009; Hill and Petersen, 2015; Scimone et al., 2016). Similarly, RNAi of nou darake (ndk), a member of the FGFR-like (FGFRL) family of putative FGF decoy receptors, also results in a brain expansion phenotype along with ectopic posterior eyes (Cebria et al., 2002). The Wnt inhibitor notum also acts oppositely in the head regionalization pathway and independent of its roles in wnt1/polarity signaling. notum(RNAi) decapitated animals that succeed in regenerating a head form a miniaturized brain with elongated eyes, and also notum(RNAi) regenerating head fragments attain a reduced sized brain and form an ectopic set of anterior eyes (Hill and Petersen, 2015). notum likely acts mainly through wnt11-6 for anterior patterning because co-inhibition of wnt11-6 suppresses the notum(RNAi) phenotypes of small brain and ectopic anterior photoreceptors, whereas co-inhibition of wnt1 does not modify these phenotypes (Hill and Petersen, 2015, 2018; Atabay et al., 2018). The restricted anterior expression of notum also suggests that head patterning is accomplished in part by maintaining a low-Wnt environment in the far anterior. fzd5/8-4 and ndk expression is also restricted to anterior regions, whereas wnt11-6 expression is prominent in the posterior brain and also present in body wall muscle across much of the AP axis. These factors are expressed constitutively, and their inhibition in uninjured animals leads to mispatterning phenotypes similar to those in regenerating animals (Hill and Petersen, 2015). Therefore, planarians use ongoing Wnt/FGFRL positional information to maintain anterior regionalization.

A separate set of Wnt-related and FGFRL genes control trunk identity in planarians. Inhibition of ndl-3 (a FGFRL protein), ptk7 (a kinase-dead Wnt co-receptor), wntP-2 (Wnt ligand also called wnt11-5) (Gurley et al., 2010) or fzd1/2/7 (Wnt receptor) causes posterior trunk duplication, with animals forming secondary mouths and ectopic pharynges (Lander and Petersen, 2016; Scimone et al., 2016). Similar to the anterior signals discussed above, these trunk patterning factors are required homeostatically and expressed regionally. wntP-2 is expressed in an animal-wide posterior-to-anterior gradient, ptk7 is expressed in a trunk-centered gradient and ndl-3 is expressed prepharyngeally. The head and trunk Wnt/FGFRL systems appear to act distinctly, because inhibition of each system does not influence the phenotypic output of the other. Together, these findings suggest that body-wide systems of Wnt-FGFRL signaling convey positional information needed for regeneration and homeostatic tissue maintenance. Additional factors have been identified, such as nr4A and pbx, regulating patterning at the termini (Blassberg et al., 2013; Chen et al., 2013; Li et al., 2019), prep transcription factor regulating the anterior (Felix and Aboobaker, 2010) and sp5 transcription factor regulating territory within the tail (Tewari et al., 2019). In addition, Wnt and Activin signaling regulate fissioning behavior as well as latent transverse regions prone to scission under pressure which mark sites of fissioning (Arnold et al., 2019). However, the downstream signaling factors important for body regionalization along the AP axis have not been fully resolved. In addition, it is not clear how signals from Wnt/FGFRL signaling along the AP axis relate to the canonical Wnt signaling used at the axis termini. Here, we identify src-1 as a global suppressor of anterior identities that can operate independently of pole formation. Our analysis indicates that src-1 likely acts in parallel or downstream of pathways involving Wnt and FGFRL factors to restrict anterior tissue identities in planarians.

To identify new regulators of regeneration patterning in planarians, we conducted an RNAi screen of 175 genes enriched for intracellular and receptor kinase activity (Table S1). Inhibition of 58 of these genes caused regeneration defects spanning from reduced blastema formation (21 genes), to aberrant photoreceptor formation (16 genes), to impaired movement behavior (5 genes). From this set, we identified a rare phenotype of ectopic eye formation in regeneration, due to inhibition of dd3147, an Src-family kinase that we named src-1 (Fig. 1A; Fig. S1A). A Src-like gene had been cloned from planaria previously but its functional roles in regeneration were unknown (Burgaya et al., 1994). We isolated the src-1 clone and further analyzed this phenotype by staining to examine src-1 requirements in body patterning. qPCR verified the effectiveness of the src-1 dsRNA for src-1 knockdown (Fig. S1B). Control animals regenerated two eyes as observed in live animals and measured by double fluorescence in situ hybridization (FISH) detecting both opsin (eye photosensory neurons) and tyrosinase (eye pigment cup cells) (Fig. 1A). By contrast, src-1(RNAi) animals formed ectopic posterior eyes in addition to their normal eyes (Fig. 1A-C). Ectopic posterior photoreceptors formed in src-1(RNAi) regenerating fragments at a gradation of penetrance, highest in regenerating head fragments (89%, 180/203 animals), lower in regenerating trunk fragments (56%, 114/204 animals) and lowest in regenerating tail fragments (26%, 49/189). Therefore, src-1 was most strongly required in anterior regions to prevent their posteriorization. In addition, animals inhibited homeostatically for src-1 also formed ectopic posterior eyes (Fig. S1C), indicating that src-1 activity also maintains pattern through tissue turnover.

The src-1(RNAi) eye phenotype was reminiscent of the phenotypes observed for ndk, wnt11-6 and fzd5/8-4 RNAi that also resulted in the formation of a larger brain (Cebria et al., 2002; Hill and Petersen, 2015; Scimone et al., 2016). Therefore, we sought to determine whether src-1(RNAi) animals similarly formed a larger brain. We investigated the size of the brain in src-1(RNAi) animals by examining the expression of several brain markers. src-1(RNAi) animals indeed formed a larger brain than controls that was posteriorly expanded in both regenerating head fragments stained with gluR or cintillo riboprobes (Fig. 1B) and regenerating trunk fragments stained with chat or cintillo (Fig. 1C; Fig. S1C). Likewise, animals inhibited homeostatically for src-1 also formed expanded brains (Fig. S1C). Thus, we conclude that src-1 acts to suppress head identity in general, similar to wnt11-6 or ndk factors.

Given the requirement for src-1 in regionalizing head identity, we tested whether it acted specifically in this process versus more generally in other AP patterning. Several factors have been implicated in restricting trunk identity to a more anterior position, and these do not appear to influence head patterning: fzd-1/2/7, ndl3, ptk7 and wntP-2 (Sureda-Gomez et al., 2015; Lander and Petersen, 2016; Scimone et al., 2016). Similar to inhibition of these regulators, src-1(RNAi) regenerating trunk fragments formed a secondary posterior pharynx (marked by laminin expression), at ∼30% penetrance (Fig. 1C). Similar to previous observations for inhibition of ndl-3, ptk7 and wntP-2 (Sureda-Gomez et al., 2015; Lander and Petersen, 2016; Scimone et al., 2016), ectopic pharynx phenotypes in src-1(RNAi) animals were only observed in regenerating trunk fragments possessing a pre-existing pharynx and not, for example, in regenerating head fragments (Fig. 1B). Together, we conclude that src-1 anteriorly limits both trunk and head domains and can act similarly to both the anterior and posterior Wnt/FGFRL systems.

We next investigated whether src-1 could control PCG expression as part of its function to regulate anterior patterning. We found src-1 mRNA to be broadly expressed throughout the animal without gradient-like properties, differing from known PCGs (Fig. S2A). src-1 expression was detected in both muscle and non-muscle cells (Fig. S2B) as measured by co-FISH with the muscle marker collagen. These observations are consistent with single-cell RNA sequencing which found src-1 as widely expressed in a wide variety of cell types, including muscle (Wurtzel et al., 2015) (Fig. S2C). Therefore, it is possible that src-1 could act in muscle cells to regulate anterior identity in planarians, or alternatively influence patterning in some other way.

Muscle cells themselves are required for positional information, because selective depletion of muscle subtypes causes mispatterning phenotypes. For example, longitudinal muscle is lost after myoD RNAi and circular muscle is lost after nkx1-1 RNAi, resulting in altered PCG expression (Scimone et al., 2017). Therefore, we sought to determine whether src-1 RNAi phenotypes could be explained by the absence of muscle cell bodies or their fiber projections. Immunostainings showed that src-1(RNAi) animals possessed muscle fibers stained with the 6G10 antibody and we could not detect any consistent differences in this pattern compared with control conditions (Fig. S2D). In addition, muscle cell bodies labeled by collagen mRNA were also present in apparently normal distributions in regenerating src-1(RNAi) animals (Fig. S2E). Although it remains possible that src-1 influences muscle fiber orientation and/or muscle cell bodies in a subtle way, these results suggest that src-1 regulates anterior patterning not through affecting muscle formation but instead by signaling within muscle or in other cell types.

We next tested whether expansion of the head region in src-1(RNAi) animals may result from changes in the notum-expressing anterior pole. We found that notum was asymmetrically expressed in src-1(RNAi) animals at 18 h after amputation, similar to controls (Fig. S3A) and consistent with the observation that, under these conditions, src-1(RNAi) animals did not have impaired axis polarization. At 72 h, notum expression was anterior but localized more broadly and with apparently reduced intensity along the midline after src-1 RNAi, suggestive of an early disturbance in pole formation (Fig. S3B). By 14 days after amputations, however, all src-1(RNAi) animals had succeeded in regenerating a notum+ anterior pole, which was mildly expanded laterally, and they regenerated pole-expressed foxD (Fig. S3B-E). notum is also expressed in anterior midline neurons of the brain (Hill and Petersen, 2015; Scimone et al., 2020), and in src-1(RNAi) animals brain-associated notum expression expanded posteriorly in concert with the expanded brain (Fig. S3C,D).

In contrast to notum, wnt1 is expressed at both the anterior- and posterior-facing wounds and is required for formation of the posterior pole in regenerating animals. Regenerating src-1(RNAi) animals had normal wound-induced wnt1 expression at 18 h and formed a posterior pole by 72 h post amputation (Fig. S3A,B). Furthermore, after 14 days regeneration or homeostatic inhibition, src-1(RNAi) animals had a normal wnt1+ posterior pole (Fig. S3C,D). Thus, src-1 inhibition did not strongly affect establishment or maintenance of the posterior pole under conditions that nonetheless led to brain expansion. Posterior and anterior pole formation depends strongly on β-catenin-1 and APC, suggesting that src-1 can act independently of these factors.

Given the expansion of anterior tissues in src-1(RNAi) animals, we tested whether the domains of PCGs were similarly modified. Both regenerating and uninjured src-1(RNAi) animals had expanded domains of anterior PCGs ndk and ndl-5 (Fig. 2; Fig. S4). We next investigated possible src-1-dependent regulation of trunk patterning factors ndl3, ptk7 and wntP-2 (Lander and Petersen, 2016; Scimone et al., 2016). src-1 inhibition resulted in the reduction of the anterior boundary of ndl3 and ptk7 within the pre-pharyngeal region but did not impact their posterior boundary (Fig. 2). These observations suggest src-1 acts to restrict the anterior domain in planarians and allows for the possibility that src-1 could be activating ndl-3 and ptk7 expression in order to control trunk identity. We then examined the effect of src-1 inhibition on the trunk PCG wntP-2, expressed in a posterior-to-anterior gradient. wntP-2 expression was unchanged in src-1(RNAi) uninjured animals (Fig. 2). axinB is a negative regulator of Wnt/β-catenin signaling in planarians, inhibition of which results in two-tailed planarians, and it is expressed similarly to wntP-2 in a posterior-to-anterior gradient (Iglesias et al., 2011). Axins are feedback inhibitors of β-catenin signaling, and thus axin expression marks locations of canonical Wnt pathway activity. Unlike β-catenin-1 RNAi (Lander and Petersen, 2016), src-1 inhibition did not eliminate axinB expression, but because of the low expression of the axinB transcript in these experiments, we could not unambiguously rule out the possibility that src-1 inhibition mildly modifies axinB expression in some way. However, this analysis suggests that src-1 inhibition likely does not eliminate β-catenin signaling along the body axis, similar to previous observations made after wntP-2 and ptk7 RNAi (Fig. 2) (Lander and Petersen, 2016). Likewise, expression of posterior markers fzd-4 and wnt11-1 was unchanged in src-1(RNAi) animals (Fig. 2). Together, these observations point to a primary role for src-1 in controlling anterior and central PCG expression domains.

Srcs are intracellular tyrosine kinases that can act as a signaling hub of multiple pathways and influence many cellular processes (Parsons and Parsons, 2004). Given that src-1 inhibition shifts anterior PCG domains and results in brain expansion and posterior ectopic eye phenotypes reminiscent of ndk and wnt11-6 RNAi (Cebria et al., 2002; Hill and Petersen, 2015), we sought to determine whether src-1 might signal downstream of either factor. To begin to address this question, we designed epistasis experiments using double RNAi. notum(RNAi) head fragments form an ectopic set of eyes within the head tip anterior to the pre-existing photoreceptors, whereas wnt11-6(RNAi) head fragments form an ectopic set of eyes posterior to the pre-existing eyes. Concurrent inhibition of notum and wnt11-6 has been shown to suppress the anterior ectopic photoreceptor, arguing that wnt11-6 likely acts downstream and oppositely to notum in head patterning (Hill and Petersen, 2015). We reasoned that if src-1 acted primarily downstream of wnt11-6, and therefore of notum in the anterior, then dual inhibition of notum and src-1 should produce the posterior ectopic eye and enlarged brain phenotypes seen in single inhibition of src-1 while suppressing the notum(RNAi) anterior eye phenotypes. Instead, simultaneous inhibition of notum and src-1 in amputated head fragments produced several different phenotypes: 24 of 42 animals exhibited a synthetic phenotype with both posterior and anterior photoreceptors, 9 of 42 animals had a notum(RNAi) phenotype with only anterior photoreceptors, 6 of 42 animals exhibited a src-1(RNAi) phenotype of only posterior photoreceptors, and 3 of 42 animals appeared normal (Fig. 3A). In a majority of cases, the synthetic phenotypes occurred with only a single posterior ectopic eye but two anterior ectopic eyes as shown, but we identified a small number of cases of such animals in which two posterior eyes were symmetrically formed (2/42 animals). The observation of a synthetic phenotype after inhibition of both src-1 and notum at this frequency (in ∼50% of animals) indicates these factors can exert distinct influences and strongly suggests that src-1 can act independently of notum, and therefore likely of wnt11-6, for controlling anterior identity. In support of this model, simultaneous inhibition of notum and src-1 in amputated head fragments led to a brain size (as measured by cintillo+ cell number) that was neither small like notum(RNAi) nor large like src-1(RNAi) but instead a size in between the two RNAi phenotypes (Fig. 3B,C).

We next sought to test the possibility that src-1 might transduce signaling through ndk. NDK/FGFRL receptors have ectodomains capable of binding FGFs but lack intracellular kinase domains to transduce signals, so have been proposed to act as FGF pathway decoy receptors. However, planarian FGF ligands have not been implicated in AP patterning, so it is unclear what other pathway components signal via planarian FGFRLs. The short intracellular domain of FGFRLs could be capable of recruiting other types of signaling models, so we considered the possibility that FGFRLs might signal through src-1 by a close examination of the ndk(RNAi) versus src-1(RNAi) phenotypes. The ndk RNAi phenotype typically involves production of ectopic eyes at a more posterior location than src-1 RNAi, giving some support for the distinct action of these factors.

We further probed the characteristics of the ectopic eyes in each RNAi condition, taking advantage of a newly identified distinction between wnt11-6(RNAi) and ndk(RNAi) conditions in controlling the location of eye regeneration after eye removal (Atabay et al., 2018; Hill and Petersen, 2018). Planarians under normal conditions can regenerate their eyes within ∼7 days after surgical removal. However, pattern disruption phenotypes resulting in ectopic eyes have distinct properties with respect to the location of regeneration after removal of ectopic versus pre-existing eyes. If the original, pre-existing photoreceptors in wnt11-6(RNAi) animals were surgically removed, they did not regenerate (11/11 animals). By contrast, when the ectopic photoreceptors of wnt11-6(RNAi) animals were surgically removed, new photoreceptors regenerated in that location a majority of the time (6/9 animals) (Fig. 4). These results are consistent with previous studies (Atabay et al., 2018; Hill and Petersen, 2018) showing similar behavior for eyes in animals inhibited simultaneously for wnt11-6 and fzd5/8-4. These results suggest that wnt11-6 and fzd5/8-4 signals control a target location for eye regeneration at a particular AP position in the animal. By contrast, eye regeneration in ndk(RNAi) animals does not share this property, because in these animals removal of the pre-existing eyes still allowed for regeneration at that position most of the time (62%, 8/13 animals), whereas removal of ectopic photoreceptors did not lead to eye regeneration in a majority of animals (85%, 11/13 animals) (Fig. 4) (Hill and Petersen, 2018). Therefore, ndk RNAi alters the overt pattern of animals without modifying the wnt11-6-dependent system directing the position of new eye regeneration.

We reasoned that if src-1 acted downstream of ndk to promote anterior identities, then as observed in ndk(RNAi), the ectopic eyes in src-1(RNAi) animals would be incapable of regenerating, but regeneration of pre-existing eyes would succeed. To test this model, we resected ectopic and pre-existing eyes from a cohort of homeostatic src-1(RNAi) animals then tracked each animal and eye profile individually over 22 days of recovery. Regeneration failed at the location of nearly all src-1(RNAi)-transected pre-existing eyes (10 of 11 animals), and by contrast regeneration succeeded at the sites of transected ectopic eyes at a frequency (5 of 11 animals) close to that seen in wnt11-6(RNAi) animals (Fig. 4). These results indicate that the site of eye regeneration is stably shifted after either wnt11-6 or src-1 RNAi but not after ndk RNAi. Therefore, src-1 likely acts independently from ndk. Taken together, the double-RNAi and eye regeneration tests suggest that wnt11-6, src-1 and ndk likely control separate processes important in head patterning.

Given these findings of broadly parallel action, along with the role of src-1 in maintaining anterior PCG domains, we reasoned that src-1 inhibition might be capable of exerting a broader influence on positional signaling. To test this possibility, we carried out a series of double-RNAi experiments between src-1 and other PCGs and examined effects on head and trunk patterning. We first simultaneously inhibited src-1 with the head patterning factors wnt11-6, ndk and fzd5/8-4, each known to restrict eye cell number and brain size from more posterior regions, but which do not normally influence trunk identity in planarians (Scimone et al., 2016). Strikingly, when src-1 was simultaneously inhibited with wnt11-6, ndk or fzd5/8-4, animals had dramatically increased numbers of ectopic posterior eyes compared with any single gene RNAi conditions (Fig. 5A). In addition, the double-RNAi animals had more severely expanded brains as measured by counting cintillo+ chemosensory cells (Fig. S5). Thus, inhibition of Src with anterior-specialized Wnt and FGFRL signals leads to enhanced anterior transformations.

Next, we tested the effects of simultaneous src-1 inhibition with the patterning factors known to restrict trunk but not head identity in planarians (Lander and Petersen, 2016; Scimone et al., 2016). We used laminin and Hoechst staining to test double-RNAi animals for their ability to form a secondary or tertiary pharynx. src-1 inhibition enhanced the penetrance of the ectopic pharynx phenotype after inhibition of ndl-3 [from 11% in ndl-3(RNAi) to 40% in ndl-3+src-1(RNAi)] and ptk7 [from 5% in ptk7(RNAi) to 58% in ptk7+wntP-2(RNAi)]. Under these conditions, single-gene inhibition of wntP-2 led to an already highly penetrant trunk duplication phenotype (88%), and although src-1(RNAi) did increase this slightly (94%), src-1+wntP-2 RNAi caused a higher expressivity of forming two ectopic pharynges (Fig. 5B). Together these results indicate the src-1 inhibition sensitizes animals to disruption of either head or trunk control systems.

The head and trunk PCG systems are thought to act independently, so we investigated whether src-1 co-inhibition with PCGs might reveal hidden dependencies in the outputs to these systems. To test this, we inhibited src-1 along with ndl-3, ptk7 or wntP-2 and tested for effects on head patterning by counting cintillo+ cells (Fig. S5) and likewise inhibited src-1 along with ndk and wnt11-6 and tested for effects on trunk patterning by staining for laminin (Fig. 5B). src-1 co-inhibition with ndl-3 mildly enhanced the increased brain cell number phenotype, but co-inhibition with ptk7 and wntP-2 had no additional effect on brain expansion (Fig. S5). Therefore, after src-1 inhibition, the activities of these trunk PCG regulators remained broadly tied to regulating trunk identity. Surprisingly, however, co-inhibition of src-1 along with PCGs that ordinarily only regulate head identity (ndk, wnt11-6) dramatically enhanced the penetrance of ectopic pharynx formation in regenerating trunk fragments (Fig. 5B). Whereas 0% of control and 7% (1/13) of src-1(RNAi) animals formed extra pharynges, 75% (12/16) of ndk+src-1(RNAi) animals and 40% (6/15) of wnt11-6+src-1(RNAi) animals formed extra pharynges. By contrast, we have never observed that individually inhibiting wnt11-6 or ndk leads to formation of ectopic pharynges. Therefore, in the context of src-1 RNAi but not under normal circumstances, wnt11-6 and ndk function is important for trunk patterning. These results point to an unexpected interplay of patterning signals otherwise known to be associated with distinct regions. This interaction could arise from a role for src-1 in assigning PCGs to individual outputs. Alternatively, the PCG system might be set up in a way that allows region-to-region control such that any sufficiently strong head expansion could ultimately also push the trunk territory further posterior. These results identify src-1 as a strong modifier of positional control used for whole-body regeneration.

Together, these results suggest a distinct role for src-1 in planarian regeneration in controlling anterior patterning (Fig. 6). We found that src-1 acts as a global negative regulator of anterior patterning, because its inhibition resulted in the expansion of both head and trunk identity. These effects were largely independent of hallmarks of head-tail AP axis polarization: injury-induced wnt1 or notum, or the formation of wnt1 and notum poles. As tyrosine kinases activated many upstream signals and are capable of regulating many downstream factors, Src-related factors control both signaling and morphogenesis to regulate many aspects of tissue formation and maintenance, including cell proliferation, differentiation, migration, survival, polarity and cell mechanical properties, with activating mutations to Src capable of driving cancer progression (Thomas and Brugge, 1997; Guarino, 2010; Kohlmaier et al., 2015; Espada and Martin-Perez, 2017; Anton et al., 2018; Tamada et al., 2021). Therefore, planarian src-1 could, in principle, exert its patterning function in a variety of ways. Given that Src is an intracellular kinase known to act downstream of multiple receptors (Erpel and Courtneidge, 1995; Thomas and Brugge, 1997; Abram and Courtneidge, 2000; Lemmon and Schlessinger, 2010), we focused our analysis on determining whether src-1 could regulate anterior patterning downstream or in parallel to planarian Wnt and/or FGFRL signals also known to regulate the AP axis.

Srcs have varied relationships to Wnt pathways described across several systems. In the canonical Wnt signaling pathway, Wnt binding to Frizzled receptors recruits Dishevelled (Dvl), sequestering Axin and preventing GSK3 phosphorylation of β-catenin that leads to its proteolysis through the destruction complex, thus allowing β-catenin accumulation and nuclear translocation to activate gene expression via TCF/LEF transcription factors (Gao and Chen, 2010). In mammalian F9 carcinoma cells, Src knockdown led to reduced canonical Wnt3a-stimulated TCF/LEF reporter output, an affect attributed to the ability of Src to bind and phosphorylate Dvl2, potentiating activation of canonical Wnt downstream signals (Yokoyama and Malbon, 2009). Srcs can also act downstream of noncanonical Wnt pathways, such as the Derailed/Ryk receptors transducing Wnt5 family signals, important for neuronal development (Wouda et al., 2008; Petrova et al., 2013). Srcs can also function negatively in Wnt signaling, for example through targeting the Wnt co-receptor Lrp6/arrow for inactivation in developing zebrafish embryos (Chen et al., 2014). Src can also function downstream of the kinase-dead co-receptor Ptk7 (Andreeva et al., 2014), which acts as a Wnt co-receptor for either activating or inactivating canonical Wnt signaling (Peradziryi et al., 2011; Hayes et al., 2013). Together with these observations, the identification of a planarian Src that collaborates with Wnt-dependent processes supports a potentially ancient connection between these factors for axis formation.

We considered the possibility that src-1 could act broadly downstream of several planarian Wnts. Some lines of evidence from this and previous work could be consistent with this interpretation. First, inhibition of planarian dishevelled-2 caused the simultaneous formation of a secondary ectopic pharynx and posterior ectopic photoreceptors, similar to the src-1(RNAi) phenotype (Almuedo-Castillo et al., 2011). Second, the src-1(RNAi) phenotype resembled a combination of wnt11-6 and wntP-2 RNAi phenotypes and also caused a permanent shift to the site of eye regeneration similar to wnt11-6 RNAi. While the apparent distinction between src-1 and β-catenin-1 RNAi phenotypes suggests differences in their activities, the downstream factors in wntP-2 and wnt11-6 signaling are not fully understood, in part because of pleiotropic effects from β-catenin-1 inhibition. Although reduced doses of β-catenin-1 dsRNA have been reported to result in the formation of an ectopic posterior mouth and pharynx primordium (Almuedo-Castillo et al., 2011), the most prominent effects of β-catenin-1 are highly penetrant formation of posterior and ectopic heads (Gurley et al., 2008; Petersen and Reddien, 2008; Adell et al., 2009). src-1 appears to operate more primarily with the Wnt/FGFRL signals that pattern that AP axis versus the wnt1/notum pole signals responsible for determining head-versus-tail polarity. Therefore, it is unlikely that src-1 transduces all Wnt signals in the animal or is involved in all instances of β-catenin signaling. In principle, src-1 could have interactions with other genes reported to cause ectopic photoreceptor formation posteriorly when knocked down, for example the nuclear receptor nr4A (Li et al., 2019). However, nr4A RNAi causes additional defects such as loss of muscle from the anterior and a shift in the wnt1 expression domain not detected in src-1 RNAi, suggesting these factors likely do not obligately regulate each other in all situations. Instead, we suggest there may be multiple inputs into eye patterning reflective of patterning as a multi-step process.

However, our data is not consistent with a model in which src-1 acts exclusively downstream of notum, and therefore its Wnt targets, to control anterior identity. Simultaneous inhibition of src-1 and notum generated a synthetic phenotype in which animals displayed elements of both phenotypes, as opposed to an outcome indicative of genetic epistasis (Fig. 3A). In previous work, wnt11-6 inhibition fully suppressed the notum(RNAi) ectopic eye phenotype, suggesting that notum primarily acts through wnt11-6 for controlling eye placement (Hill and Petersen, 2015), and that src-1 is unlikely to act primarily downstream of wnt11-6. However, we cannot rule out the possibility that src-1 could act downstream of any Wnts that can act independently of notum and influence head regionalization or downstream of Wnts with involvement in other patterning roles. Testing this model would require future work to identify patterning roles for other negative regulators of specific Wnts or methods to detect Src activation. Interestingly, Dishevelled has also been shown to act in non-canonical Wnt signaling and to mediate a Wnt5-derailed/Related to tyrosine kinase (RYK)-dependent signal, which can signal through Src (Gao and Chen, 2010). Planarian wnt5 defines the lateral-medial axis in planarian regeneration (Gurley et al., 2010), and inhibition of dishevelled-1 in planarians has been shown to recapitulate aspects of the wnt5(RNAi) phenotype, such as lateral separation of the planarian brain lobes (Almuedo-Castillo et al., 2011). However, wnt5 is not believed to regulate AP patterning, so it is unlikely that src-1 acts mainly downstream of wnt5 to control AP head and trunk regionalization.

We also considered the possibility that src-1 could act downstream of an unidentified receptor or FGFRLs. FGFRLs such as ndk and ndl-3 have been shown to regulate regional identity in planarians (Cebria et al., 2002; Lander and Petersen, 2016; Scimone et al., 2016), but the mechanism by which this signaling occurs is unclear. FGFRLs have been shown to act as decoy receptors in Xenopus embryos. The FGFRL1 ectodomain is shed from the cell membrane and binds to some FGF ligands with high affinity, including FGF2, FGF3, FGF4, FGF8, FGF10 and FGF22 to regulate FGF signaling (Steinberg et al., 2010). However, inhibition of FGFs or FGFRs in planarians have so far not resulted in any reported patterning phenotypes (Wagner et al., 2012; Auwal et al., 2020). The intracellular domain of human FGFRL1 can interact with the SPRED1 signaling molecule which could allow for downstream intracellular signaling (Zhuang et al., 2011). Furthermore, in β-cell insulin granules, the intracellular domain of FGFRL1 can bind SHP-1 via a SH2 domain to activate ERK signaling (Silva et al., 2013). Given the synergistic effects of src-1 inhibition with ndk and ndl-3 (Fig. 5A,B) and the similar synergistic RNAi phenotypes seen with ndk and fzd5/8-4 (Scimone et al., 2016) it was possible that cross-regulation between WNT and FGFRL signaling pathways controls body regionalization in planarians and that this may be mediated through src-1. However, the examination of the extra photoreceptor defect in src-1 RNAi revealed more similarities with the wnt11-6 phenotype than the ndk phenotype. First, ndk RNAi generated posterior photoreceptors located more distantly than either wnt11-6 or src-1 RNAi. Second, both wnt11-6 and src-1 RNAi tended to shift the location of eye regeneration more posteriorly, unlike ndk RNAi. These observations suggest that src-1 acts distinctly from ndk. Other signaling factors known to interface with Src have been described in planarians but are unlikely to explain the patterning roles of src-1. For example, integrins are well known to signal through Src for adhesion and signaling (Huttenlocher and Horwitz, 2011) but inhibition of the single planarian integrin-β led to tissue disorganization and excess neurogenesis in general rather than specific AP patterning effects seen in src-1 RNAi (Bonar and Petersen, 2017; Seebeck et al., 2017), but it remains possible that AP patterning involves some input from integrin signals. Future systematic analysis of the many possible upstream Src receptors could help resolve the role of Src for AP patterning with respect to Wnts. Thus, taken together, we propose a role for src-1 in globally suppressing anterior identity and regulating posterior determination in parallel to the action of Wnts and FGFRLs, perhaps using alternative signal inputs.

Our analysis of dual inhibition phenotypes between src-1 and known AP patterning regulators further suggests this model of parallel action. The expressivity of any patterning phenotype we examined that involved anteriorization of posterior tissue was enhanced dramatically after src-1 RNAi, which would be expected if src-1 and Wnts act independently to control pattern. Other explanations are possible, for example that src-1 acts downstream of all Wnt genes and that the relatively weaker Wnt RNAi phenotypes represent incomplete knockdown. The ability for src-1 RNAi to reprogram the outputs of anterior PCGs into controlling trunk identity suggests src-1 may act in a buffering process that helps channel PCG factors into controlling distinct outputs. In addition, the progressive nature of the src-1 RNAi phenotype to affect anterior regions more strongly than posterior regions was consistent with the finding that PCG domains were shifted rather than eliminated by src-1 RNAi. The outcomes of double-RNAi between src-1 and head or trunk PCGs are also suggestive of an anterior bias to src-1 function in which, at least within the context of src-1 inhibition, anterior factors such as ndk can be reprogrammed to influence more posterior identity, but more posterior factors such as ptk7 and wntP-2 are not reprogrammed to influence more anterior identity. This could ultimately reflect the overlapping uses of the two Wnt/FGFRL systems to define non-anterior in successive domains. It is also possible that the hypothesized parallel actions of src-1 and Wnts could arise from distinct signaling to regulate PCG domains within the muscle versus interpretation of domain identity via neoblast-dependent tissue formation. Reagents to examine Src activation status, along with systematic tests of the many possible receptors upstream of Src, will be helpful in resolving these and other mechanisms. It is intriguing to note that the Src family kinase SRC-1 in Caenorhabditis elegans acts in parallel to Wnt signaling in order to regulate spindle orientation along the AP axis of the very early embryo (Bei et al., 2002). These observations suggest there may be deep ancestry to the use of joint activities for Src family kinases and Wnt signals in forming the primary body axis. Together, our results identify src-1 as a new factor regulating positional information along the planarian AP axis that is used for specifying the proper identity of missing tissues in regeneration.

Asexual strain CIW4 of the planarian S. mediterranea were maintained in 1× Montjuic salts at 19°C as previously described (Petersen and Reddien, 2011). Planarians were fed a liver paste and starved for at least 7 days before experiments.

src-1 (dd_Smed_v6_3147_0_1) was identified through blast searching the planarian transcriptome at https://planmine.mpibpc.mpg.de/ (Brandl et al., 2016; Grohme et al., 2018). Primers used for cloning src-1 were 5′-AAGCTTGGTGGCTTGCTTTA-3′ and 5′-TGCGATCAACCAATGAAAAA-3′. Primers for cloning the genes from the screen are indicated in Table S1.

Riboprobes and double-stranded RNA (dsRNA) for src-1 were generated by in vitro transcription (NxGen, Lucigen) as previously described (Petersen and Reddien, 2011). Riboprobes and dsRNAs for src-1 were cloned by reverse transcription polymerase chain reaction (RT-PCR) into pGEM-T-easy using the primers 5′-AAGCTTGGTGGCTTGCTTTA-3′ and 5′-TGCGATCAACCAATGAAAAA-3′.

Other riboprobes (chat, cintillo, gluR, opsin, tyrosinase, collagen, laminin, notum, wnt1, ndk, ndl-5, fzd4, wntP-2, axinB, ptk7) were as previously described (Oviedo et al., 2003; Cebria et al., 2007; Reddien et al., 2007; Wang et al., 2007; Collins et al., 2010; Gurley et al., 2010; Wenemoser and Reddien, 2010; Petersen and Reddien, 2011; Lapan and Reddien, 2012; Currie and Pearson, 2013; März et al., 2013; Vu et al., 2015).

RNAi was performed by dsRNA feeding. For RNAi, dsRNA was synthesized from in vitro transcription reactions (NxGen, Lucigen). dsRNA corresponding to C. elegans unc-22, not present in the planarian genome, served as a negative control. For the RNAi screen (Table S1), animals were fed a mixture of liver paste and dsRNA three times over 6 days, then amputated transversely to generate head, trunk and tail fragments. Animals were scored for regeneration defects after 14 days of regeneration (Table S1). For other experiments, unless noted otherwise, animals were fed a mixture of liver paste and dsRNA six times in 14 days before amputation of heads and tails 4 h after the final feeding. For Fig. 4, animals were fed dsRNA 12 times over 6 weeks and starved for 3 days before eye resection. For all comparisons between double RNAi and single RNAi conditions, an equal amount of control competing dsRNA was mixed with the single RNAi condition so that animals across treatments received the same overall amount of dsRNA.

Colorimetric (NBT/BCIP) or fluorescence in situ hybridizations were performed as previously described (Lander and Petersen, 2016), after fixation in 4% formaldehyde and bleaching (Pearson et al., 2009) using blocking solution containing 10% horse serum and western blot blocking reagent (Roche) (King and Newmark, 2013). Digoxigenin- or fluorescein-labeled riboprobes were synthesized as previously described (Pearson et al., 2009) and detected with anti-digoxigenin-HRP (1:2000, Roche/Sigma-Aldrich, 11207733910, lot 10520200), anti-fluorescein-HRP (1:2000, Roche/Sigma-Aldrich, 11426346910, lot 11211620) or anti-digoxigenin-AP (1:4000, Roche/Sigma-Aldrich 11093274910, lot 11265026). Hoechst 33342 (Invitrogen) was used at 1:1000 as a counterstain. For immunostainings, animals were fixed in Carnoy's solution as previously described (Hill and Petersen, 2015), using tyramide amplification to detect labeling with rabbit anti-6G10 (1:3000, Cell Signaling Technology, D2C8, lot 3377S).

Live animals and NBT/BCIP-stained animals were imaged using a Leica M210F dissecting microscope and a Leica DFC295, with adjustments to brightness and contrast using Adobe Photoshop. Whole animal fluorescence imaging was performed on either a Leica DM5500B compound microscope with Optigrid structured illumination system or a Leica laser scanning SPE confocal microscope at 40× or 63×, and presented images are maximum projections of a z-series with adjustments to brightness and contrast using ImageJ and Photoshop. Plots were generated in Microsoft Excel or R (ggplot2).

cintillo⁺ cells in the brain were counted manually and normalized to the square root of the animal area determined using Hoechst staining and CellProfiler (Lamprecht et al., 2007).

Total RNA was extracted by mechanical homogenization in Trizol (Life Technologies), treated with DNase (TURBO DNAse, Ambion) and reverse transcribed with oligo-dT primers (Multiscribe reverse transcriptase, Applied Biosystems), and qPCR was performed using Eva Green PCR Master Mix (Biotium) from nine regenerating fragments in four biological replicates. Relative mRNA abundance was calculated using the delta-Ct method after verification of primer amplification efficiency, normalizing to Ubiquilin expression. P-values below 0.05 using an unpaired two-tailed t-test were considered as significant.

The following primer sets were used: src-1 – 5′-ATGACGTGTATAACGCCGACAC-3′, 5′-TGAGGACAGGACAGTGTTAATTTG-3′; ubiquilin – 5′-ATTCGTCGGAATTGGAAACA-3′, 5′-GCGTTCACATCTCCAAAGGT-3′.

Modified from Hill and Petersen (2018), briefly, worms were immobilized on ice for resection and eyes were removed using a hypodermic needle. All animals were tracked individually and imaged 1 day before eye removal, 1 day after eye removal to confirm resection of eye tissue, and 22 days post-surgery to determine the regenerative outcome.

We thank members of the Petersen lab for critical comments and K. Lo for help with riboprobes.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200125.