Expression pattern and level of Djago2 in planarian

To examine the expression of Djago2 in intact planarians, I used WISH to detect the mRNA of Djago2. WISH for Djago2 was performed with a digoxigenin-labeled antisense RNA probe. The RNA probe with ~1500 nucleotides in sequence length included the 3’ UTR and a partial sequence of the PIWI domain. Animals that were 5~8 mm long were used for this experiment. A sense RNA probe was used as the control,

26

and no signal was detected after hybridization (Fig. 3A). This showed that the background signal was undetectable in our WISH system. The expression pattern of Djago2 was detected in the whole body and was readily detected in central nervous

system (CNS) and middle dorsal line. The expression pattern of Djago2 in intact planarians was similar to Djcbc-1 (Dμgesia japonica chromatoid body component 1) (Yoshida-Kashikawa et al., 2007). Djcbc-1 is a homolog of human RCK/p54. In planarians, Djcbc-1 is located in chromatoid bodies of neoblasts and chromatoid body-like structures in the CNS (data not shown). DjpiwiA is known to be a neoblast-specific gene, which is expressed throughout the planarian body. It is also expressed in cell clusters in the middle dorsal line (Fig. 3B).

3.2.2 Djago2 is highly expressed in the regenerating region.

I next examined the expression of Djago2 in regenerating animals. After amputation anterior and posterior to the pharyngeal region, the head, trunk, and tail were allowed to simultaneously regenerate. Expression patterns of Djago2 in the head, trunk, and tail were observed on day 3 after transverse amputation. Djago2 was expressed in cells of the CNS, middle dorsal line, and regenerating tissues (Fig. 4A).

This showed that Djago2 might be expressed in neoblasts. To observe Djago2 levels at

27

wound sites during planarian regeneration, worms were transversely amputated posterior to the pharynx, harvested and analyzed by WISH. Animals were harvested at various time points of 8 h, and 1, 2, 3, 5, and 7 days after amputation. Expression of Djago2 was increased in the regenerating region during regeneration until 3 days, and

the signal was lower on day 5 and 7 (Fig. 4B).

The expression levels of Djago2 at various regeneration time points were quantified by RT-qPCR. During planarian regeneration, the regenerating tissue, called the blastema, contains a mass of cells which is capable of replacing lost or damaged body parts. In previous studies, DjpiwiA-labeled neoblasts proliferated in the region of the post-blastema and then migrated to and differentiated into various cell types in the blastema (Umesono et al., 2011). Quantitative analysis of Djago2 was performed for tissues of the blastema and post-blastema during regeneration. Blastemas were harvested at various time points of 3, 5, and 7 days after transverse amputation at the post-pharyngeal region. Post-blastema samples of three more time points at 8, 24, 48 h post-amputation were harvested in the regenerating region. Expression levels of Djago2 were not examined in the blastema at 3, 5, or 7 days post-amputation because the newly formed tissues were too small. Control samples of the blastema and post-blastema were the tail and the region next to the first amputation, respectively. In total, 2 µg of RNA in

28

each sample was collected for reverse transcription to produce cDNA and was subjected to a qPCR analysis. In the post-blastema, expression of Djago2 was increased during regeneration, and Djago2 was respectively 2.5- and 3.2-fold higher than the control at 8 and 72 h post-amputation (Fig. 5A). In the blastema, expression of Djago2 was 6-fold higher than the control at 72 h post-amputation (Fig. 5B). Our results showed that the expression of Djago2 was increased in both the blastema and post-blastema during regeneration. Then I examined expression levels of DjpiwiA, a marker for neoblasts, in the blastema at 3, 5, and 7 days post-amputation. Expression of DjpiwiA did not rise up at 72 h after the first amputation (Fig. 5C). On days 5 and 7 post-amputation, expression levels of DjpiwiA were lower than the control. This confirmed that neoblasts had accumulated in the post-blastema but not in the blastema during regeneration.

3.3.1 γ-irradiation eliminates the Djago2-positive cells.

Since the increasing expression of Djago2 in regenerating tissues was synchronized with that of DjpiwiA, I hypothesized that Djago2 is expressed in neoblasts, and Djago2 expression is higher as a result of proliferation and differentiation of neoblasts. To check if Djago2 is expressed in neoblasts, I examined expression levels of Djago2 in both normal and neoblast-depleted animals. I depleted neoblasts in planarians by 90 Gy

29

γ-irradiation. Fourteen days after γ-irradiation, all animals displayed degeneration (Fig.

6A). All animals were dead at 3~4 weeks after irradiation, indicating a failure of tissue homeostasis. I used FACS to confirm if γ-irradiation eliminated neoblasts (Fig. 6B).

Single-cell suspensions were collected and filtered from dissociated animals. Staining with Hoechst 33342 reflects the DNA content in nuclei. The cytoplasm of living cells was stained with vital calcein-AM. In FACS, I first selected a population of particles which contained Hoechst 33342/calcein-AM double-positive cells. I separated those cells into three populations according to the DNA content and cell size. G2/M- (Fig 6B, p3) and G1/G0-phase cells (Fig. 6B, p4 and p5) could be distinguished by the Hoechst 33342 fluorescence signal. Our data showed that G2/M-phase cells (p3 in Fig 6B), which are highly proliferating neoblasts, had been eliminated at day 3 post γ-irradiation (Fig. 6B).

RT-qPCR analysis was used to examine the either Djago2 or DjpiwiA expression levels between irradiated and un-irradiated animals. Total RNA was collected from three animals in different treatments. A total 3μg RNA was collected from three animals and used for cDNA preparation. Our results showed that γ-irradiation eliminated the expression of DjpiwiA and reduced the level of Djago2 down to 23% as compared to the control. This demonstrated that γ-irradiation had eliminated neoblasts and reduced

30

Djago2 expression in intact animals (Fig. 6C). I also compared Djago2 expression

patterns between irradiated and un-irradiated animals by WISH (Fig. 7A). Djago2 was expressed in the CNS in both irradiated and un-irradiated animals. But the detected signals were weaker in the intact bodies of irradiated worms. Interestingly, signals in cell clusters of the dorsal middle line were totally eliminated by γ-irradiation, suggesting that irradiation indeed depleted Djago2-positive neoblasts in the dorsal middle line. Comparing the expression of Djago2 in irradiated and un-irradiated animals on day 3 post amputation, I found that the increased expression of Djago2 near the wound site had disappeared as well (Fig. 7B).

3.3.2 Djago2 is highly expressed in neoblasts.

To confirm the expression of Djago2 in highly proliferating neoblasts, I collected G2/M-phase cells and analyzed their Djago2 levels by an RT-qPCR. Single-cell suspensions were prepared from 60 worms at ~1.5 cm in length for the FACS analysis.

Cells in population 3 (in the G2/M phase) were sorted as neoblasts. Cells in population 5 (in the G0/G1 phase) were sorted as differentiated cells. Cells in population 4, which contained part of the neoblasts, were not sorted. In total, 300 ng RNA was collected and subjected to cDNA synthesis. I compared Djago2 and DjpiwiA expressions between

31

neoblasts and differentiated cells. The Djago2 expression level was higher in neoblasts (Fig. 8).

3.4.1 Failure to regenerate in Djago2-silenced animals.

From previous data, I already knew that the Djago2 expression level is higher in regenerating tissue, and Djago2 is highly expressed in neoblasts. To check the function of Djago2 in planarian regeneration, I knocked down Djago2 by administering dsRNA.

I observed loss of the head and tail formation in planarian regeneration with one round of Djago2 dsRNA feeding, but not in control animals (Fig. 9B). Pleiotropic regeneration defects were observed in Djago2-knocked down planarians. Fifteen of 30 trunks were observed to have no blastema formation on day 7 after transverse amputation in the anterior and posterior regions of the pharynx (Fig. 9C). Seven of 30 trunks were observed to have limited regeneration with a single photoreceptor or no photoreceptor formation on day 7 after amputation. I extended the observation period to 14 days.

Three of seven trunks had formed a complete photoreceptor, others had formed an incomplete photoreceptor, while five of 30 animals died. In control animals, 29 of 30 trunks regenerated to intact animals, and one trunk died (Fig. 9C). Higher knockdown efficiency was observed by increasing the feeding times of Djago2 dsRNA (data not

32

shown), while I also found that animals began degenerating at the head region. However, most trunks did regenerate and died when I fed them Djago2 dsRNA more than five times.

I examined the expression of Djago2 in animals that had failed to regenerate by an RT-qPCR. The expression level of Djago2 was reduced to 11% after Djago2 silencing.

Expression levels of DjpiwiA, DjpiwiC, and Djpcna were also reduced to 3%, 16%, and 17% at 7 days after amputation (Fig. 9C). This suggested that depletion of Djago2 may eliminate a large population of neoblasts and also that Djago2 is required for maintaining the homeostasis of neoblasts.

I therefore checked if neoblasts were indeed eliminated in Djago2-depleted worms using a FACS analysis. Surprisingly, the FACS results showed that cells of population 3 (in the G2/M phase) were not depleted in animals that had failed to regenerate (Fig. 9D).

Moreover, the percentage of G2/M-phase cells had increased in animals that had failed to regenerate (Fig. 9E), suggesting that an undefined mechanism may be involved in the regenerating deficiency phenotype of Djago2-depletion.

For those animals which showed incomplete regeneration with Djago2 RNAi, I further checked the pattern of formation of the head region to understand if DjAgo2 is also required for differentiation. Djndk is a head marker in planarians used to show the head

33

formation in control and limited-regeneration (RNAi) animals 14 days after amputation.

I examined expression patterns of Djndk by WISH. Compared to control RNAi animals, incomplete head formation was observed in Djago2 RNAi-treated animals (Fig. 10).

This result suggested that the differentiation potential from neoblast progeny was restricted in Djago2 RNAi-treated animals.

3.4.2 Tissue homeostasis defect was observed in long-term Djago2 RNAi-treated

animals.

As mentioned above, I observed the head-regression phenotype when I increased the Djago2 dsRNA feeding rounds. I therefore increased the dsRNA concentration in the RNAi food (6 mg of dsRNA in 25 μl of liver) to check if this degeneration phenotype resulted from the complete knockdown of Djago2. The head-regression phenotype was observed when RNAi food was given 3 to 6 times (Fig. 11A). This phenotype was similar to that of irradiated animals (Fig 6A). I examined the expression of Djago2 in head-degenerated animals by an RT-qPCR. The Djago2 expression level was reduced to 57%. Respective expression levels of DjpiwiA and Djpcna were reduced to 43% and 40% (Fig. 11B). These results suggested dysfunction of neoblasts in Djago2 RNAi-treated animals.

34

Since neoblasts are the only mitotic cells in planarians as far as is known, I labeled mitotic cells with anti-phospho-histone 3-serine 10 (H3P), to confirm if the depletion of Djago2 affected the M-phase of mitosis. As shown in Fig. 11C, numbers of

H3P-positive cells dramatically decreased in Djago2-silenced animals. However, cell cytometric data showed that G2/M-phase cells were not significantly reduced in degenerating animals (Fig. 11D). Similar to animals that had failed to regenerate (Fig.

9E), the ratio of G2/M-phase cells even increased in degenerating animals compared to the control (Fig. 11E). A summary of these results indicated that RNAi of Djago2 blocked dividing neoblasts.

35

4. Discussion

4.1 DjAgo2 is a member of the Argonaute subfamily.

Argonaute proteins are highly conserved in many organisms. Members of the Argonaute protein family are characterized by the presence of PAZ (Piwi-Argonaute-Zwille), Mid, and PIWI domains. The protein sequence of DjAgo2 is most similar to Drosophila Ago1 and human AGO2, and DjAgo2 has homology to the Argonaute subfamily (Fig. 1). The PAZ domain of DjAgo2 has 79.7% identity with Drosophila Ago1 and 75.7% identity with human AGO2. The PIWI domain of DjAgo2

has 84.8% identity with Drosophila Ago1 and 85.8% identity with human AGO2. The PAZ domain is bound to the 3’ end of small RNA. The PIWI domain is bound to the 5’

end of small RNA. This shows that DjAgo2 might have small RNA-binding affinity.

The PIWI domain of DjAgo2 contains the catalytic residue, DDH, which is an essential motif for cleavage activity. This indicates that DjAgo2 potentially has mRNA-cleavage activity. In Drosophila, Ago1 functions in the miRNA-mediated translation repression pathway (Miyoshi et al., 2005), and Ago2 plays major roles in the RNAi pathway. Both of them have slice activity, and are involved in different pathways. The protein sequence of DjAgo2 is more closely related to Drosophila Ago1 than to Ago2, suggesting that the

36

function of DjAgo2 is more related to miRNA pathways. In addition, I found that when YFP-tagged DjAgo2 was expressed in HeLa cells, it also localized to P-body and co-localized with human RCK, a P-body component (Fig. 2). Since human Ago2 is localized to cytoplasmic P-bodies which are mRNA decapping and degradation sites (Hutvagner and Simard, 2008). Our result suggests that DjAgo2 might have a role in miRNA-mediated mRNA degradation.

4.2 Djago2 is up regulated in regenerating tissue during planarian regeneration.

The expression level of Djago2 increased at the wound site (Fig. 4A). During regeneration, the expression level of Djago2 increased in the regenerating region from 8 hours to 3 days after amputation, and then decreased (Fig. 4B). In a quantification analysis, regenerating tissue was separated into two regions: the blastema and post-blastema. Post-blastema tissue contains many neoblasts which proliferate and differentiate into progenitor cells. These cells migrate to the blastema, and then differentiate into various types of differentiated cells. In the blastema, progenitor cells do not undergo mitosis. During regeneration, the expression of Djago2 was increased 2.5-fold at 8 h post-amputation, and then had declined to 1.7-fold higher than the control at 48 h post-amputation (Fig. 5B). The expression of Djago2 reached a peak at

37

72 h post-amputation (Fig. 5B). This phenomenon is similar to the change in neoblast proliferative activity during regeneration(Salo and Baguna, 1984). The increased level of Djago2 is caused by the aggregation and proliferation of neoblasts. In the blastema, the expression of Djago2 had increased to 6-fold higher than the control at 72 h post-amputation, and then dropped back to the control level (Fig. 5C). In the blastema, elevation of Djago2 may be caused by a wound response of differentiated cells or the differentiation of progenitor cells. To clarify if the increase of Djago2 is a result of wound response or a signal from progenitor cells, we examine the expression level of Djago2 at 72 h post amputation in irradiated animals. The expression of Djago2 was not

higher in irradiated animals on day 3 post-amputation (Figs. 4A, 7B). Therefore, the increased level of Djago2 did not result from a wound response of differentiated cells, but from differentiation of progenitor cells.

4.3 Expression level of Djago2 is higher in neoblasts than in differentiated cells.

My data show that Djago2 is expressed in clustering cells which are located in the middle dorsal line (Fig. 2A). DjpiwiA is also expressed in the middle dorsal line (Fig.

3C). FACS data showed that 90 Gy of γ-irradiation eliminated neoblasts (Fig. 7B), while Djago2-positive cells in the middle line were also eliminated. RT-qPCR data

38

confirmed these results. The eliminated cells were neoblasts. The expression level of Djago2 was lower in irradiated animals (Fig. 6C) and was higher in neoblasts (Fig. 8).

4.4 Djago2 is required for proliferation of neoblasts and differentiation of

progenitor cells.

In my data, three phenotypes were observed in Djago2-silenced animals. First, failure of regeneration was observed after Djago2 RNAi food was administered 3 times (Fig. 9B). Parts of the trunk could not form a blastema at the wound site, and animals died within 14 days after amputation. This phenotype was similar to silencing of Djbruli, Smedwi-2, Smed-smB (Fernandez-Taboada et al., 2010), DjPCNA (Orii et al., 2005),

and DjRbAp48(Bonuccelli et al., 2010). No blastema formation means no neoblasts or no proliferation. In Djago2 RNAi worms, expression levels of DjpiwiA, Djpcna, and DjpiwiC were lower on day 7 post-amputation (Fig. 9D) with no blastema formation

compared to control animals. DjpiwiA and DjpiwiC were used as neoblasts markers. The number of neoblasts decreased in Djago2-silenced animals.

In the second phenotype, parts of trunk fragments could form a blastema, but failed to form a pair of normal photoreceptors. This phenotype was similar to those of Djcbc-1 and Djtrans-2a. This indicates that Djago2 may be required for differentiation of

39

progenitor cells. Animals exhibiting limited head regeneration with a single photoreceptor were used to examine the ability of progenitor cells to differentiate. In a comparison of expression patterns of Djndk between normal animals and Djago2 RNAi-treated animals, an incomplete brain was observed in animals with only one photoreceptor (Fig. 10). This indicated that Djago2 silencing limited the differentiation capability of progenitor cells. However, we still need to check which types of differentiated cells can’t be formed from progenitor cells in Djago2 silenced animals.

The regeneration of many organs other than brain also needs to be monitored to confirm that Djago2 silencing limits the ability of progenitor cells to differentiate to other terminally differentiated cells (Oviedo and Levin, 2007).

In the third phenotype, homeostasis defects were observed when dsRNA was fed more than 3 times. Levels of DjpiwiA, Djpcna, and DjpiwiC decreased in both degenerating and regeneration-failed animals. Surprisingly, our FACS data showed that the number of G2/M-phase cells was increased in failed-regeneration animals (Figs. 9F, 11E). Since the mitotic cells can be also labeled by anti-H3P antibody, we found that the amounts of anti-H3P-positive cells were reduced in homeostasis-defective animals (Fig.

11C). This shows that Djago2 silencing blocked the mitosis of neoblasts. It implied that, in Djago2-silenced animals, the cell cycle of neoblasts was stopped in the G2 phase.

40

These data suggest that head regression is resulted from a deficiency of neoblasts. My data showed that Djago2 silencing led to animal regression and death, which might have been caused by apoptosis (Naoghare et al., 2011).

Taken together, our data suggest that Djago2 is required for planarian regeneration and tissue homeostasis. Djago2 is also required for proliferation and self-renewal of neoblasts and differentiation of progenitor cells (Fig. 12). In previous studies, inhibition of many RNA granule components, including Djbruli, Djvas-1, Djcbc-1, Djupf-1, Djxrn-1, Djedc-4, Djdicer and translation initiation factors, leds to lethality or limited

regeneration (Anderson and Kedersha, 2006; Rouhana et al., 2010). This suggests that post-transcriptional regulators are required for stem-cell proliferation and planarian regeneration. Since Djago2 is an orthologue of hsago2 with high conservation, it is possible that Djago2 also mediates small RNA-mediated gene silencing that regulates neoblast proliferation and progenitor cell differentiation. In the future, involvement of Djago2 in planarian miRNA- or siRNA-mediated gene regulation should be further

confirmed. We expect to identify the miRNAs regulated by Djago2 in planarians by the immunoprecipitation of Djago2. In my preliminary data, several miRNAs were found to be up-regulated in regenerating tissue. We will further confirm if these miRNAs are associated with Djago2, and are required for planarian regeneration. Many following

41

experiments, such as using miRNA inhibitors to block the function of miRNA in planarians, will be conducted in the near future.

42

5. References

Adell, T., Salo, E., Boutros, M., and Bartscherer, K. (2009). Smed-Evi/Wntless is required for beta-catenin-dependent and -independent processes during planarian regeneration. Development 136, 905-910.

Anderson, P., and Kedersha, N. (2006). RNA granules. J Cell Biol 172, 803-808.

Aukerman, M.J., and Sakai, H. (2003). Regulation of flowering time and floral organ identity by a MicroRNA and its APETALA2-like target genes. Plant Cell 15, 2730-2741.

Baguna, J., and Romero, R. (1981). Quantitative-Analysis of Cell-Types during Growth, Degrowth and Regeneration in the Planarians Dugesia-Mediterranea and

Dugesia-Tigrina. Hydrobiologia 84, 181-194.

Baguna, J., Salo, E., and Auladell, C. (1989). Regeneration and Pattern-Formation in Planarians .3. Evidence That Neoblasts Are Totipotent Stem-Cells and the Source of Blastema Cells. Development 107, 77-86.

Bardeen, C.R., and Baetjer, F.H. (1904). The inhibitive action of the Roentgen rays on regeneration in planarians. Journal of Experimental Zoology 1, 191-195.

Bartel, D.P. (2004). MicroRNAs: genomics, biogenesis, mechanism, and function. Cell