Meis1 Controls the Differentiation of Eye Progenitor Cells and the Formation of Posterior Poles during Planarian Regeneration

As a member of TALE family, Meis1 has been proven to regulate cell proliferation and differentiation during cell fate commitment; however, the mechanism is still not fully understood. The planarian, which has an abundance of stem cells (neoblasts) responsible for regenerating any organ after injury, is an ideal model for studying the mechanisms of tissue identity determination. Here, we characterized a planarian homolog of Meis1 from the planarian Dugesia japonica. Importantly, we found that knockdown of DjMeis1 inhibits the differentiation of neoblasts into eye progenitor cells and results in an eyeless phenotype with normal central nervous system. Furthermore, we observed that DjMeis1 is required for the activation of Wnt signaling pathway by promoting the Djwnt1 expression during posterior regeneration. The silencing of DjMeis1 suppresses the expression of Djwnt1 and results in the inability to reconstruct posterior poles. In general, our findings indicated that DjMeis1 acts as a trigger for the activation of eye and tail regeneration by regulating the differentiation of eye progenitor cells and the formation of posterior poles, respectively.


Introduction
Meis1 protein is a member of the three amino acid loop extension (TALE) transcription factors family. Previous studies have demonstrated that Meis1 plays essential roles in the embryonic development across metazoans. For example, Meis1 participates in the embryonic cortical development by promoting neuronal proliferation and migration [1]. Cooperating with Tbx and Hox factors, Meis1 acts as a trigger for limb initiation. Knockdown of Meis1 results in severe limb agenesis [2]. In addition, Meis1 is involved in organ patterning, including the formation of hearts [3], eyes, and lungs [4,5]. With the exception of embryonic development, biological functions of Meis1 have a close correlation with the tumorigenesis [6,7]. Upregulation of Meis1 and associated co-factors can promote the skin tumorigenesis. Knockdown of Meis1 induces a significant decrease in benign and malignant tumors in mice [8]. Meanwhile, Meis1 also serves as a negative regulator in the occurrence and development of cancers, such as non-small cell lung cancer and prostate cancer [9,10]. In general, these studies suggest that Meis1 has broad functions in the regulation of cell proliferation and differentiation during embryonic development. However, the mechanism of Meis1 family, which is responsible for stem cells proliferation and differentiation, remains unclear.
Planarians can regenerate a complete individual from any fragment of their body [11][12][13]. This strong regenerative ability depends on an abundance of stem cells (neoblasts) that exist in the adult body [14][15][16][17][18]. The regeneration of heads or tails in planarians, known as anterior-posterior (AP) axis regeneration, is an important model for studying the mechanism of stem cells proliferation, differentiation, migration, and apoptosis [19,20]. The position control genes (PCGs) are genes with a regional expression pattern and are constitutively expressed in muscles. By encoding positional information, PCGs determine the cell 2 of 14 fate along the anterior-posterior axis, medial-lateral axis, and dorsal-ventral axis. Knockdown of PCGs will lead to mis-patterning of the phenotypes [19]. At present, Wnt signaling is known to regulate the tissue identity along the AP axis during regeneration [21][22][23]. Wnt signaling pathway components, Djwnt1 and β-catenin, act as promoters for the construction of posterior poles [24,25]. Knockdown of Djwnt1 can cause planarians that are unable to regenerate tails or regenerate heads in the place of tails. β-catenin acts as a downstream gene of Djwnt1. Silencing of β-catenin can result in the reversal of posterior poles and cause planarians to regenerate posterior heads from posterior blastema [26]. Furthermore, notum serves as an inhibitor of Djwnt1 during anterior poles regeneration [27]. Notum RNAi can upregulate Wnt signaling and cause animals to regenerate tails in the place of heads [28]. Recently, Djpbx and Djislet have been proven to be positive regulators of Djwnt1 during posterior poles regeneration. Knockdown of Djpbx and Djislet can cause planarians that are unable to regenerate their tails [29,30]. In general, these studies suggest that PCGs involved in the Wnt signaling pathway act as a determinant of the cell fate along the AP axis. However, the underlying mechanisms in regulating the Wnt signaling are not yet fully understood.
In our work, we identify a homolog of Meis1 in planarian Dugesia japonica. We observe that DjMeis1 RNAi planarians cannot form mature eye cells or reconstruct the posterior poles. DjMeis1 RNAi decreases the number of eye progenitor cells and suppresses the expression of Djwnt1. Herein, we propose that DjMeis1 serves as a positive factor in the regulation of the differentiation of eye progenitor cells and the establishment of posterior poles during planarian regeneration.

DjMeis1 RNAi Inhibits the Eye and Tail Regeneration
To better investigate the role of Meis1 in the tissue identity determination, we identified three homologous proteins of Meis in planarian Dugesia japonica based on our previous transcriptome [31]. Two of the Meis homologous DjMeis2 and DjMeis3 have been previously reported as SmedMeis and SmedMeis-like in planarian Schmidtea mediterranea (Figures S1 and S2) [32,33]. RNAi experiments were performed by injecting dsRNA to the planarians. At 24 h after the last injection, planarians were amputated from the anterior and posterior sites of pharynx into three fragments: Head, trunk, and tail fragments ( Figure 1A). We observed that the fragments of control groups regenerated into complete individuals at 7 days after amputation (7 dpa) ( Figure 1B). Consistent with the phenotypes previously reported, silencing of DjMeis2 in Dugesia japonica resulted in the formation of smaller eyes ( Figure 1C). DjMeis3 RNAi induced the formation of a squared head with elongated eyes or cyclops ( Figure 1D and Figure S3). Importantly, we identified a rare eyeless combined with tailless phenotype ( Figure 1E), due to the inhibition of Meis1, a Meis1 homolog that we named DjMeis1. The DjMeis1 RNAi eyeless phenotype was reminiscent of the phenotypes observed from DjMeis2 and DjMeis3 RNAi animals, which suggested a conserved role for Meis in the eye development. In contrast to the DjMeis2 and DjMeis3 RNAi phenotypes, DjMeis1 RNAi animals displayed more sever inhibition in the eye regeneration. No eyes could be observed in newly regenerated heads in DjMeis1 RNAi animals. Furthermore, head fragments of DjMeis1 RNAi animals failed to regenerate tails ( Figure 1E). Trunk fragments could regenerate eyeless heads from anterior blastema, but were unable to develop tails from posterior blastema ( Figure 1E). These phenotypes indicate that DjMeis1 may have broad requirements in the eye and tail regeneration. Planarians were amputated into three fragments from the anterior and posterior sites of pharynx at 24 h after the last injection. (B) The head, trunk, and tail fragments of control planarians developed into normal and complete individuals at 7 dpa. (C) The newly regenerated eyes (red arrows) in trunk (30/30) and tail (30/30) fragments of DjMeis2 RNAi planarians were smaller than those in control groups at 7 dpa. (D) The trunk (24/30) and tail (23/30) fragments of DjMeis3 RNAi planarians regenerated a squared head with elongated eyes (white arrows) at 7 dpa. (E) The head (30/30) and trunk fragments (30/30) of DjMeis1 RNAi planarians failed to regenerate tails (red boxes). The trunk (30/30) and tail fragments (30/30) regenerated eyeless heads (white boxes) at 7 dpa. The images on the right were an enlarged view of the fragments in white and red boxes. Scale bars: 400 μm.

DjMeis1 Is Required for the Eye Regeneration by Regulating the Formation of Eye Progenitor Cells
It has been proven that PCGs expressed in muscles determine the tissue identity, including the specification of neoblasts and the correct location of progenitor cells [19,34,35]. To verify whether the eyeless phenotype was due to the effect of DjMeis1 RNAi on body regionalization, we detected the expression of ndl2, a PCG that was expressed in the prepharyngeal regions and defined the head patterning [36,37]. Compared to the control animals, DjMeis1 RNAi animals displayed a normal expression pattern of ndl2, suggesting that DjMeis1 RNAi might not impair the expression of anterior PCGs (Figure 2A). The brain nerves in planarian could be formed within 24 h after amputation and continued to develop into a complete bilobed brain. The first photoreceptor neurons regenerated in the dorsal side of the brain within 3 days [38]. In the view of the close connection between the development of eyes and brain nerves in planarians, we speculated that knockdown of DjMeis1 might impair the reconstruction of brain nerves, and then lead to the failure of eye regeneration. To test this possibility, we performed WISH with sert (sert is expressed in serotonergic neurons) and PC2 (PC2 is expressed in the central nervous system (CNS)) probes to detect the reconstruction of brain nerves [39][40][41]. We observed that the bilobed brain of DjMeis1 RNAi animals was normally reconstructed compared to the control animals at 7 dpa ( Figure 2B,C), suggesting that knockdown of DjMeis1 had no effect on the regeneration of brain nerves, and DjMeis1 acted independently of the brain nerves in the regulation of eye development.

DjMeis1 Is Required for the Eye Regeneration by Regulating the Formation of Eye Progenitor Cells
It has been proven that PCGs expressed in muscles determine the tissue identity, including the specification of neoblasts and the correct location of progenitor cells [19,34,35]. To verify whether the eyeless phenotype was due to the effect of DjMeis1 RNAi on body regionalization, we detected the expression of ndl2, a PCG that was expressed in the prepharyngeal regions and defined the head patterning [36,37]. Compared to the control animals, DjMeis1 RNAi animals displayed a normal expression pattern of ndl2, suggesting that DjMeis1 RNAi might not impair the expression of anterior PCGs (Figure 2A). The brain nerves in planarian could be formed within 24 h after amputation and continued to develop into a complete bilobed brain. The first photoreceptor neurons regenerated in the dorsal side of the brain within 3 days [38]. In the view of the close connection between the development of eyes and brain nerves in planarians, we speculated that knockdown of DjMeis1 might impair the reconstruction of brain nerves, and then lead to the failure of eye regeneration. To test this possibility, we performed WISH with sert (sert is expressed in serotonergic neurons) and PC2 (PC2 is expressed in the central nervous system (CNS)) probes to detect the reconstruction of brain nerves [39][40][41]. We observed that the bilobed brain of DjMeis1 RNAi animals was normally reconstructed compared to the control animals at 7 dpa ( Figure 2B,C), suggesting that knockdown of DjMeis1 had no effect on the regeneration of brain nerves, and DjMeis1 acted independently of the brain nerves in the regulation of eye development. To clarify the effect of DjMeis1 RNAi on the formation of eyes, we performed W with Djovo probe (Djovo is expressed in the eye progenitor cells) to examine the regen tion of eye progenitor cells in DjMeis1 RNAi animals [42]. Compared to the control mals, the number of Djovo + cells in DjMeis1 RNAi animals was significantly reduced ure 2D), suggesting that the deletion of DjMeis1 suppressed the differentiation of eye genitor cells from neoblasts, and DjMeis1 was required for the formation of eye proge cells. Previous studies on the development of the visual system demonstrated Djsix1/2 and Djeya were also essential for the regeneration of eye progenitor cells, these specific transcription factors were co-expressed with Djovo [32,43,44]. To furthe plore the molecular functions of DjMeis1 on the development of eye progenitor cells detected the expression pattern of DjMeis1 after knockdown of Djovo, Djsix1/2, and D respectively. We found that planarians with the silencing of Djovo, Djeya, and Djs greatly reduced the DjMeis1 signal compared to the control groups ( Figure S4). Thes sults suggest that DjMeis1 is required for the differentiation of eye progenitor cell pressing Djovo, and in turn, the formation of eye progenitor cells promote the expres of DjMeis1. Furthermore, to examine whether the differentiation from eye progenitor To clarify the effect of DjMeis1 RNAi on the formation of eyes, we performed WISH with Djovo probe (Djovo is expressed in the eye progenitor cells) to examine the regeneration of eye progenitor cells in DjMeis1 RNAi animals [42]. Compared to the control animals, the number of Djovo + cells in DjMeis1 RNAi animals was significantly reduced ( Figure 2D), suggesting that the deletion of DjMeis1 suppressed the differentiation of eye progenitor cells from neoblasts, and DjMeis1 was required for the formation of eye progenitor cells. Previous studies on the development of the visual system demonstrated that Djsix1/2 and Djeya were also essential for the regeneration of eye progenitor cells, and these specific transcription factors were co-expressed with Djovo [32,43,44]. To further explore the molecular functions of DjMeis1 on the development of eye progenitor cells, we detected the expression pattern of DjMeis1 after knockdown of Djovo, Djsix1/2, and Djeya, respectively. We found that planarians with the silencing of Djovo, Djeya, and Djsix1/2 greatly reduced the DjMeis1 signal compared to the control groups ( Figure S4). These results suggest that DjMeis1 is required for the differentiation of eye progenitor cells expressing Djovo, and in turn, the formation of eye progenitor cells promote the expression of DjMeis1. Furthermore, to examine whether the differentiation from eye progenitor cells into mature eye cells was also affected after knockdown of DjMeis1, we detected the formation of mature eye cells by WISH for opsin [41,45]. We observed that no mature eye cells were regenerated in DjMeis1 RNAi animals at 7 dpa ( Figure 2E), indicating that DjMeis1 also played an important role in the formation of mature eye cells from eye progenitor cells. To test this model, we performed eye resection to planarians, which retained the pre-existing brain nerves. We observed that DjMeis1 RNAi animals failed to regenerate their eyes at 7 days after surgical removal ( Figure 2F). Considered together, we conclude that DjMeis1 acts as a trigger for the differentiation of eye progenitor cells and promotes the formation of mature cells without affecting the regeneration of brain nerves.

DjMeis1 Control of the Tail Regeneration Is Independent of Cell Proliferation
Given the requirement of DjMeis1 in tail regeneration, we first detected the epidermal cells of planarian with LaminB probe (the epidermal boundary marker) to determine whether the posterior wound in DjMeis1 RNAi animals could be healed [46]. We observed that DjMeis1 RNAi animals showed a normal expression pattern of LaminB compared to the control groups at 7 dpa ( Figure 3A), suggesting that DjMeis1 RNAi animals retained the ability to heal the posterior wound despite failing to regenerate their tails.
motes the formation of mature cells without affecting the regeneration of brain nerves.

DjMeis1 Control of the Tail Regeneration Is Independent of Cell Proliferation
Given the requirement of DjMeis1 in tail regeneration, we first detected the epidermal cells of planarian with LaminB probe (the epidermal boundary marker) to determine whether the posterior wound in DjMeis1 RNAi animals could be healed [46]. We observed that DjMeis1 RNAi animals showed a normal expression pattern of LaminB compared to the control groups at 7 dpa ( Figure 3A), suggesting that DjMeis1 RNAi animals retained the ability to heal the posterior wound despite failing to regenerate their tails.
As planarians remodeling their missing tissues rely on the cell resources provided by the continuous proliferation of neoblasts [16,19], next, we attempted to analyze whether the tailless phenotype caused by DjMeis1 RNAi was due to the effect on cell proliferation. It has been reported that there are two waves of proliferative response of neoblasts after amputation. The first wave commences at 6 h after injury and shows an increase in the cell proliferation throughout the body. The second wave commences at 48 h and tends to be restricted to the wound site [47]. Therefore, we performed whole-mount immunofluorescence for phosphorylated histone H3 (H3p) and bromodeoxyuridine (BrdU) labeling experiments, and quantified the mitotic density at 6 and 48 h to determine whether the proliferative ability of neoblasts was impaired after DjMeis1 RNAi [48,49]. However, we observed that DjMeis1 RNAi animals showed normal proliferative ability compared to the control groups at both 6 and 48 h after amputation ( Figure 3B-E). In conclusion, DjMeis1 RNAi has no effect on the cell proliferation, and the failure of tail regeneration is not caused by the downregulation of the proliferative ability of the neoblasts.  As planarians remodeling their missing tissues rely on the cell resources provided by the continuous proliferation of neoblasts [16,19], next, we attempted to analyze whether the tailless phenotype caused by DjMeis1 RNAi was due to the effect on cell proliferation. It has been reported that there are two waves of proliferative response of neoblasts after amputation. The first wave commences at 6 h after injury and shows an increase in the cell proliferation throughout the body. The second wave commences at 48 h and tends to be restricted to the wound site [47]. Therefore, we performed whole-mount immunofluorescence for phosphorylated histone H3 (H3p) and bromodeoxyuridine (BrdU) labeling experiments, and quantified the mitotic density at 6 and 48 h to determine whether the proliferative ability of neoblasts was impaired after DjMeis1 RNAi [48,49]. However, we observed that DjMeis1 RNAi animals showed normal proliferative ability compared to the control groups at both 6 and 48 h after amputation ( Figure 3B-E). In conclusion, DjMeis1 RNAi has no effect on the cell proliferation, and the failure of tail regeneration is not caused by the downregulation of the proliferative ability of the neoblasts.

DjMeis1 Is Required for the Re-Establishment of Posterior Poles by Regulating the Expression of Djwnt1
Previous studies have demonstrated that signals of posterior poles are required for tail regeneration [29,30]. Therefore, we speculated that DjMeis1 RNAi might induce the tailless phenotype by affecting the reconstruction of posterior poles. To test this possibility, we performed whole-mount immunostaining and fluorescence in situ hybridization experiments to detect the ventral nerve cords (VNC) reconstruction. We observed that DjMeis1 RNAi animals failed to regenerate VNC posteriorly, which was consistent with the result that DjMeis1 RNAi animals healed the posterior wound at the amputated site without activating the regeneration of tails ( Figure 4A and Figure S5) [50,51]. Due to the defects on the regeneration of VNC, we were interested in whether the regeneration of the posterior organ was affected following DjMeis1 RNAi. Therefore, we detected the regeneration of pharynx (mhc-1 + cells) of the head fragments and observed that DjMeis1 RNAi animals regenerated incomplete pharynx or failed to regenerate pharynx ( Figure 4B) [37,52], suggesting that DjMeis1 RNAi inhibited the proper formation of organs during posterior regeneration. Moreover, we performed in situ detection on fz4 [29], a posterior poles marker of planarian. We found that DjMeis1 RNAi animals failed to express fz4 in posterior blastema ( Figure 4C). In general, these data suggest that DjMeis1 acts as an important regulator in the reconstruction of posterior poles.

DjMeis1 Is Required for the Re-Establishment of Posterior Poles by Regulating the E sion of Djwnt1
Previous studies have demonstrated that signals of posterior poles are requ tail regeneration [29,30]. Therefore, we speculated that DjMeis1 RNAi might ind tailless phenotype by affecting the reconstruction of posterior poles. To test this p ity, we performed whole-mount immunostaining and fluorescence in situ hybrid experiments to detect the ventral nerve cords (VNC) reconstruction. We observ DjMeis1 RNAi animals failed to regenerate VNC posteriorly, which was consiste the result that DjMeis1 RNAi animals healed the posterior wound at the amputa without activating the regeneration of tails (Figures 4A and S5) [50,51]. Due to the on the regeneration of VNC, we were interested in whether the regeneration of th rior organ was affected following DjMeis1 RNAi. Therefore, we detected the regen of pharynx (mhc-1 + cells) of the head fragments and observed that DjMeis1 RNAi regenerated incomplete pharynx or failed to regenerate pharynx ( Figure 4B) [37,5 gesting that DjMeis1 RNAi inhibited the proper formation of organs during poste generation. Moreover, we performed in situ detection on fz4 [29], a posterior poles of planarian. We found that DjMeis1 RNAi animals failed to express fz4 in posteri tema ( Figure 4C). In general, these data suggest that DjMeis1 acts as an important tor in the reconstruction of posterior poles.  During planarian regeneration, Wnt signaling is known to determine the posterior poles reconstruction [24]. Knockdown of Djwnt1 resulted in the inability of planarians to regenerate their tails or to regenerate posterior heads in the place of tails after amputation. Next, we attempted to examine whether the silencing of DjMeis1 impaired the reconstruction of posterior poles by affecting the expression of Djwnt1. It has been reported that there are two phases of Djwnt1 expression during regeneration. The first phase is detected in the wound site at 24 h after amputation, and then the second phase is observed in the posterior blastema tip at 96 h after amputation [30]. Therefore, we detected the expression pattern of Djwnt1 at 24 and 96 h after amputation, respectively. We observed that DjMeis1 RNAi significantly decreased the expression of Djwnt1 compared to the control groups at 24 h after amputation and completely inhibited the expression of Djwnt1 at 96 h after amputation ( Figure 4D). These results suggest that DjMeis1 plays an essential role in the expression of Djwnt1, which is required for the re-establishment of posterior poles. To test this model, we performed in situ detection on the Djwnt1 downstream factor Djwnt11-2 [29,30]. As expected, we observed that the expression of Djwnt11-2 was lost after knockdown of DjMeis1 ( Figure 4E). From these data, we conclude that DjMeis1 is an important factor in posterior fate specification and determines the re-establishment of posterior poles by triggering the expression of Djwnt1.

β-Catenin/DjMeis1 RNAi Induces the Posterior Poles of Planarians Reversal
β-catenin is previously reported as a down-stream gene of Djwnt1 and is required for the reconstruction of posterior poles. β-catenin RNAi animals displayed anteriorization of posterior poles and regenerated heads from posterior blastema [25,53]. In the view of the fact that tailless animals caused by DjMeis1 RNAi retained the proliferative ability of neoblasts, we were interested in whether the silencing of β-catenin could still cause the reversal of the posterior poles of DjMeis1 RNAi animals. To verify these effects, we performed β-catenin and DjMeis1 double RNAi and observed that double RNAi animals regenerated eyeless heads in the place of tails ( Figure 5A). This phenotype confirms that DjMeis1 functions to promote the reconstruction of posterior poles independently of cell proliferation.
To determine the reversal of posterior poles of double RNAi animals, we detected the expression of sFRP1 (sFRP1 is an anterior poles marker of planarian) at 7 dpa and found that sFRP1 + cells existed both in anterior and posterior poles in double RNAi animals ( Figure 5B) [37], suggesting that the event of posterior poles reversal also occurred in double RNAi animals. Since the posterior head of double RNAi animals could not regenerate eyes, we proposed that this eyeless phenotype was caused by the loss of function of DjMeis1, which acts on the eye development. Therefore, we detected the reconstruction of CNS and observed that double RNAi animals regenerated another group of brain nerves in the posterior head similar to the β-catenin RNAi animals ( Figure 5C). Meanwhile, we performed in situ detection on opsin and found that the newly regenerated head from posterior blastema of β-catenin RNAi animals displayed a clear opsin expression pattern. In contrast, no opsin + cells were detected in the posterior head of double RNAi animals ( Figure 5D). These data suggest that DjMeis1 has a conservative function on the formation of eyes, even in the ectopic heads. In general, our data prove that β-catenin and DjMeis1 double RNAi can cause the reversal of posterior poles of DjMeis1 RNAi animals without affecting the proliferation of neoblasts, but they retain the requirement of DjMeis1 for eye regeneration.
versal of the posterior poles of DjMeis1 RNAi animals. To verify these effects, we formed β-catenin and DjMeis1 double RNAi and observed that double RNAi anima generated eyeless heads in the place of tails ( Figure 5A). This phenotype confirms DjMeis1 functions to promote the reconstruction of posterior poles independently o proliferation.

Discussion
Meis1 is an important transcription factor that is initially discovered in leukemic mice and its biological functions have been extensively studied in leukemia, organogenesis, embryonic development, and tumorigenesis [54]. Recently, Meis1 has been found to be involved in the cell cycle regulation of cardiomyocytes and endothelial cells [4,55]. Based on the previous studies, Meis1 plays an important role in cell differentiation during cell fate commitment, although the mechanism remains unclear. In this study, using the planarian as a model for regeneration, we found that DjMeis1 acts as a trigger for the activation of eye and tail regeneration by regulating the differentiation of eye progenitor cells and the formation of posterior poles, respectively.
Previous studies in mice demonstrated that Meis1 played essential roles in the eye development, as smaller eye lenses were observed in Meis1 mutant embryos [5]. Consistently, in this study, we found that DjMeis1 RNAi planarians exhibited an eyeless phenotype ( Figure 1E). It has been reported that there is a close correlation between the formation of eyes and brain nerves in planarians [38]. However, we found that DjMeis1 RNAi planarians regenerated normal brain nerves compared to the control groups ( Figure 2C), suggesting that DjMeis1 did not play an important role in the reconstruction of brain nerves, and promoted the formation of eyes independently of brain nerves ( Figure 6A). By detecting different eye cells lineage during the eye development [41,42], we observed that DjMeis1 RNAi clearly reduced the number of Djovo + eye progenitor cells and inhibited the formation of mature eye cells, which differentiated from the early progenitor lineage ( Figure 2D,E). Meanwhile, the expression of DjMeis1 was suppressed after knockdown of Djovo, Djsix1/2, and Djeya, which are essential factors for the regeneration of eye progenitor cells ( Figure S4), suggesting that DjMeis1 participated in the regulation of eye progenitor cells differentiation, and in turn, the formation of eye progenitor cells promoted the DjMeis1 expression. Furthermore, we observed that no eyes could be formed in DjMeis1 planarians at 7 days after eye resection ( Figure 2F). Considered together, we conclude that DjMeis1 is required for the differentiation of eye progenitor cells by promoting the Djovo expression and promotes the proper maturation of eye cells in planarians ( Figure 6B). knockdown of DjMeis1 ( Figure 4D). The expression of Djwnt11-2, which acted as a downstream gene of Djwnt1, was also lost in DjMeis1 RNAi planarians (Figures 4E and 6C). Considered together, we propose that DjMeis1 determines the specification of posterior fate by triggering the expression of Djwnt1 ( Figure 6D). In support of this hypothesis, we performed β-catenin and DjMeis1 double RNAi. Knockdown of β-catenin could result in the reversal of posterior poles and cause planarians to regenerate posterior heads [25]. In our study, we observed that β-catenin and DjMeis1 double RNAi induced an eyeless head from posterior blastema ( Figure 5A), which supported our findings that DjMeis1 RNAi had no effect on the proliferation of neoblasts, but inhibited the expression of Djwnt1. In addition, we observed that double RNAi planarians normally regenerated a pair of brain nerves in the posterior head, but failed to produce mature eye cells ( Figure 5C,D), which further confirmed the conserved role of DjMeis1 in the differentiation of eye progenitor cells. Another TALE transcription factor, Djpbx, has been previously reported to be required for the establishment of anterior-posterior axis [29]. Knockdown of Djpbx caused the failure of planarians to regenerate their tails or heads. In contrast to Djpbx, which is required for both anterior and posterior patterning along the body axis, DjMeis1 mainly regulates the posterior patterning. However, both Djpbx and DjMeis1 function as regulators that interpret signals along the AP axis, suggesting that TALE transcription factors have a conservative function along the AP axis.  Tailless is another phenotype caused by DjMesi1 RNAi in planarian regeneration ( Figure 1E). We envision at least two possible mechanisms by which DjMeis1 RNAi inhibits the tail regeneration. DjMeis1 could act as a promoter of cell proliferation or alternatively be required for neoblasts differentiation. It has been reported that Meis1 is involved in the proliferation and migration of pulmonary artery smooth muscle cells (PASMCs) in hypoxia [56]. Recently, Meis1 has been proven to act as a transcription factor to promote hair matrix cell proliferation [57], suggesting that Meis1 has a unique function on the regulation of cell proliferation. However, in this study, we found that DjMeis1 RNAi planarians displayed normal mitotic activities compared to the control groups ( Figure 3B-E), indicating that knockdown of DjMeis1 had no effect on the proliferation of neoblasts. Previous studies have reported that Meis1 is involved in several steps of limb AP pre-patterning and the elimination of Meis1 leads to severe standstill of limb development [2]. Next, we tested the possibility that DjMeis1 might regulate the differentiation of neoblasts by promoting the reconstruction of posterior poles. Indeed, we found that DjMeis1 RNAi planarians could not properly regenerate VNC and pharynx during posterior reconstruction and failed to express the posterior poles marker fz4 ( Figure 4A-C), suggesting that DjMeis1 acted as a regulator for the proper reconstruction of posterior poles. Furthermore, the expression of the previously known and important regulator of posterior regeneration, Djwnt1, was significantly suppressed at both 24 and 96 h after amputation following the knockdown of DjMeis1 ( Figure 4D). The expression of Djwnt11-2, which acted as a downstream gene of Djwnt1, was also lost in DjMeis1 RNAi planarians (Figures 4E and 6C). Considered together, we propose that DjMeis1 determines the specification of posterior fate by triggering the expression of Djwnt1 ( Figure 6D). In support of this hypothesis, we performed β-catenin and DjMeis1 double RNAi. Knockdown of β-catenin could result in the reversal of posterior poles and cause planarians to regenerate posterior heads [25]. In our study, we observed that β-catenin and DjMeis1 double RNAi induced an eyeless head from posterior blastema ( Figure 5A), which supported our findings that DjMeis1 RNAi had no effect on the proliferation of neoblasts, but inhibited the expression of Djwnt1. In addition, we observed that double RNAi planarians normally regenerated a pair of brain nerves in the posterior head, but failed to produce mature eye cells ( Figure 5C,D), which further confirmed the conserved role of DjMeis1 in the differentiation of eye progenitor cells. Another TALE transcription factor, Djpbx, has been previously reported to be required for the establishment of anterior-posterior axis [29]. Knockdown of Djpbx caused the failure of planarians to regenerate their tails or heads. In contrast to Djpbx, which is required for both anterior and posterior patterning along the body axis, DjMeis1 mainly regulates the posterior patterning. However, both Djpbx and DjMeis1 function as regulators that interpret signals along the AP axis, suggesting that TALE transcription factors have a conservative function along the AP axis.
In conclusion, our findings generally support the deep evolutionary functional conservation of Meis1 in the eye development and re-establishment of body patterning. Our work suggests that DjMeis1 has a broad requirement for eye regeneration and the reestablishment of posterior poles by promoting the differentiation of eye progenitor cells and inducing the expression of Djwnt1, respectively, which further reveal the function of Meis1 in tissue identity determination during the regeneration process.

Species and Culture Conditions
Animals used in all experiments were a clonal strain of the planarian Dugesia japonica and were fed in autoclaved stream water at 20 • C [58]. Before all the experiments, animals were starved for at least 1 week [59], and those animals with a total of 5-8 mm in length were used for all experiments.

Gene Cloning
The ORF of DjMeis1, DjMeis2, DjMeis3, β-catenin, Djovo, Djsix1/2, and Djeya were identified in the planarian transcriptomic data [31]. Total RNA was extracted from 5 adult planarians by TRIzol (Vazyme, Nanjing, China). cDNA was synthesized from 1 µg of total RNA using Hiscript II reverse Transcriptase (Vazyme, Nanjing, China) and HiScript qRT Supermix II (Vazyme, Nanjing, China) by reverse transcription PCR. Sets of specific primers were designed to amplify the DjMeis1 sequence from cDNA by PCR. All primers used to clone and synthesize dsRNA are shown in Table S1.

RNAi
All the dsRNA (DjMeis1, DjMeis2, DjMeis3, β-catenin, Djovo, Djsix1/2, Djeya) used for the RNAi experiments were synthesized by in vitro transcription as previously described [31,60]. Briefly, T7 polymerase was used to synthesize dsRNA via in vitro transcription. DsRNA was denatured at 68 • C and annealed at 37 • C. Finally, dsRNA was extracted by ethanol precipitation. Next, dsRNA of each gene was diluted to 2 µg/µL and was injected to animals in the dorsal side one time per day for 1 week continuously using a Drummond microinjector. Then, 100 nL of dsRNA was injected to each animal every day. In double RNAi experiments, β-catenin and DjMeis1 maintained a concentration of 2 µg/µL dsRNA. The water treated by DEPC was injected to the control group animals. Head, trunk, and tail fragments were amputated from the animals from the anterior and posterior sites of the pharynx at 24 h after the last injection.

Whole-Mount In Situ Hybridization
Whole-mount ISH (WISH) was performed as previously described [61]. In each experiment, ten animals were used in each group. Animals were killed in PBS (phosphate dehydration in a methanol dilution series in PBST, the animals were maintained in 100% methanol overnight at −20 • C. Then, the animals were rehydrated in 50% methanol for 10 min and washed in PBST and 1× SSC for 5 min, respectively. Formamide-bleaching solution (5% non-deionized formamide, 0.5× SSC, and 1.2% H 2 O 2 in deionized water) was used to bleach the animals under bright light for 2 h. After washing in 1× SSC and PBST for 5 min and incubating in Proteinase K (20 mg/mL in PBST) for 10 min at 37 • C, the animals were fixed in 4% paraformaldehyde for 20 min and washed in PBST for 10 min. Then, the animals were hybridized with DIG-labeled PC2 probe at 56 • C for 16-17 h and washed in 2× SSC and 0.2× SSC three times for 20 min, respectively. PBST with 5% goat serum and 5% Western Blocking Reagent (Roche, Basel, Switzerland) was used to block the animals for 5 h. Next, the animals were incubated in 1:500 anti-Digoxigenin-POD (Roche, Basel, Switzerland) overnight. Finally, the animals were treated with tyramide conjugated to Alexa568 (Molecular Probes, Shanghai, China) and were observed by NIS element software (version 4.2.0, Olympus, IX73P1F, Tokyo, Japan). Different concentrations of SSC were prepared by dissolving 20× SSC in deionized water.

Statistical Analyses
Data were shown as means ± SD, and statistical analyses were performed by students. One-way analysis of variance (ANOVA) was used to analyze the data of two groups. A statistically significant difference was defined as p < 0.05.