Main

In mammals, proapoptotic BH3-only protein Noxa is a member of Bcl-2 family and a candidate mediator of p53-dependent apoptosis through mitochondrial dysfunction.1, 2 Noxa also may perform functions in the immune system.3, 4, 5 However, a recent study has demonstrated a glucose-mediated control of the Noxa signaling pathway, in which cytoprotective effect is imparted by inactivating Noxa apoptotic function through phosphorylating Noxa and then enhancing glucose utilization for survival.6, 7 The above findings implicated that Noxa might be in charge of death and survival, depending on its modification.

So far, there is no detailed knowledge on zNoxa, the homolog of Noxa in zebrafish (Danio renio). Similar to mammalian Noxa, previous reports have demonstrated that zNoxa also is a candidate mediator of p53-dependent apoptosis, both in embryos and adults.8, 9, 10 Nevertheless, studies on zebrafish embryos have revealed that the apoptotic response firstly appears at mid-gastrula stage.11 Interestingly, most of the proapoptotic Bcl-2 family members, such as zBid, zNoxa and zPuma, have been observed to express before mid-gastrula stage.8 However, their exact functions have remained unclear in early embryonic development of zebrafish.

Here we have characterized zebrafish Noxa (zNoxa) and provided evidence that zNoxa is essential in early embryonic development of zebrafish, and regulates mitosis through zWnt4b and apoptosis through zBik at different development stages. Furthermore, zMcl1a and zMcl1b have been revealed to suppress the ability of zNoxa to control mitosis and apoptosis. Interestingly, mNoxa also can regulate mitosis in zebrafish embryos. The current findings indicate that zNoxa is a novel regulator of mitosis during early embryogenesis except its mediator role for apoptosis.

Results

zNoxa is expressed throughout embryonic development

We analyzed the temporal expression pattern of zNoxa during embryogenesis by semi-quantitative RT-PCR. As shown in Figure 1a, its maternal transcript existed from one-cell stage and held to shield stage until the zygote expression was initiated from this stage. Interestingly, there were two cDNA sequences of zNoxa, and both of them encoded a 45 amino-acid protein, although an extra 49 pairs of repetitive ‘GT’ sequence existed in the 3′UTR of the longer cDNA (Supplementary Figure S1). Moreover, a mouse polyclonal antibody against zNoxa-mature protein was generated (Supplementary Figure S2a), and used it to analyze the zNoxa protein expression pattern by western blot. As shown in Figure 1b, the maternal zNoxa protein also existed from one-cell stage and held to shield stage until the zygote zNoxa protein was expressed from 75% epiboly stage. Following this analysis, the spatial distribution of zNoxa transcript was detected by in situ hybridization, in which zNoxa transcript was ubiquitously distributed in early embryonic cells until shield stage, and then, the zygote zNoxa transcript was mainly localized in YSL of the embryos (Figure 1c).

Figure 1
figure 1

Expression of zebrafish Noxa (zNoxa) during embryogenesis. (a) Transcription of zNoxa in embryonic development stages was detected by semi-quantitative RT-PCR and normalized to β-actin transcript. (b) The extracted proteins from different stage embryos were subjected to western blot detection by the mouse polyclonal antiboies against zNoxa-mature protein and β-actin, respectively. (c) Expression pattern of zNoxa was detected by in situ hybridization using specific antisense or sense probes. No staining was present in embryos hybridized with sense probe (256 cell and shield representative stages shown). Embryo in ‘24 h.p.f.’ panel was lateral view, with dorsal toward the top and anterior to the left. The embryos in other panels were lateral views with animal pole toward the top and dorsal to the right

Full size image

Alterations of zNoxa expression lead to abnormal development of early embryos

To determine the biological function of zNoxa in early embryonic development, we examined its effects in early embryos through blocking translation by morpholino (MO) microinjection or through overexpression by synthetic mRNA microinjection. Firstly, two morpholinos (zNoxa MO1 and zNoxa MO2) were designed, and the optimal delivering does of zNoxa MOs were determined by a series of different concentration injections. As the injected embryos with a high dose (0.3 mM) of zNoxa MOs did not survive to bud stage, so lower doses (0.2 mM) of MOs were adopted. When 0.2 mM of MO1 or MO2 was respectively injected into one-cell embryos, there was about 80% level of zNoxa protein reduction. When the two MOs were co-injected, almost no zNoxa protein was detected. Significantly, while the two MOs were singly or together co-injected with 5 ng/μl zNoxa-rescued mRNA (rcRNA), which was not targeted by the zNoxa MO1 and MO2 through mismatching five nucleotides in zNoxa mRNA (Supplementary Material), the zNoxa protein expression was restored (Figure 2a).

Figure 2
figure 2

Altered expression of zNoxa results in deformed embryos. (a) Protein was extracted from embryos at the shield stage and 30 μg of each sample was immunoblotted for zNoxa and β-actin. β-Actin served as an internal control. (b–u) The WT embryos (b–e), the zNoxa MO1+2-injected embryos (f–i), the zNoxa mRNA-injected embryos (j–m), the zNoxa MO1+2 plus rcRNA-injected embryos (n–q) and the zNoxa mismatch MO1+2-injected embryos (r–u) were shown at shield stage, 75% epiboly stage, bud formation stage and 24 h.p.f. WT and the injected mismatch-MO embryos were identical in appearance. Lateral views of embryos at shield and 75% epiboly stages, animal pole toward the top and dorsal to the right. Lateral views of embryos at bud and 24 h.p.f. stage, antetior toward the top and dorsal to the right

Full size image

Compared with WT (Figures 2b–e) and the injected mismatch-MO embryos (Figures 2r–u), the injected embryos with an optimal dose of zNoxa MO1+2 (0.2 mM total) resulted in epiboly defect before shield stage and shortened anterior–posterior axis during 75% epiboly and bud stages (Figures 2f–h), and the affected embryos died at 24 h.p.f. (Figure 2i). Injection of 5 ng/μl zNoxa mRNA led to abnormal embryos with epiboly defects, in which the organizer was thickened from the beginning of shield stage, the anterior–posterior axis was elongated during 75% epiboly and bud stages (Figures 2j–l) and the embryonic development was arrested at 24 h.p.f. (Figure 2m). Moreover, co-injection of zNoxa MO with 5 ng/μl zNoxa rcRNA rescued most of the embryonic defects in the morphants (Figures 2n–q). Thus, the above data indicate that the zNoxa expression alterations including specific translation blocking and its overexpression all result in abnormal development of early embryos.

zNoxa deficiency arrests apoptosis at the onset of 75% epiboly stage

zNoxa was revealed to be a member of proapoptotic Bcl-2 family,8 so we further observed cell death status in the zNoxa morphant embryos by TUNEL. As shown in Figure 3a, there were a few of apoptotic cells in WT (Figures 3a and i, v and b) and the injected mismatch-MO (Figures 3a and iv, viii and b) embryos, whereas there were much less apoptotic cells in the zNoxa morphant embryos (Figures 3a and ii, vi and b), indicating that zNoxa deficiency arrests apoptosis at the onset of 75% epiboly stage. Moreover, injection of 5 ng/μl zNoxa mRNA induced extensive cell apoptosis (Figures 3a and iii, vii and b). Interestingly, almost no apoptotic cells existed in all of these embryos until 75% epiboly stage, and even injection of zNoxa mRNA could not trigger apoptosis (data not shown). This result is very similar to a previous observation.11 However, the current data indicate that zNoxa protein exists throughout embryogenesis, especially in early embryos (Figure 1b), and alteration of zNoxa expression leads to developmental defects from shield stage (Figure 2). Why it has nothing to do with apoptosis? So it is suggested that zNoxa might have another important role in early embryonic development before 75% epiboly stage.

Figure 3
figure 3

zNoxa controls mitosis and apoptosis during early embryogenesis. (a) Confocal z-stacks of apoptotic cells assessed by TUNEL in embryos which were WT (i, v) and injected with zNoxa MO1+2 (ii, vi), zNoxa mRNA (iii, vii) or zNoxa mismatch MO1+2 (iv, viii) at 75% epiboly and bud stages. Embryos were lateral views with animal pole toward the top and dorsal to the right. (b) Quantification of numbers of TUNEL-positive cells with approximately four embryos for each group in a. Error bars represent S.E.M. (n=4). (c) Confocal z-stacks of mitotic cells assessed by pH3 (phosphorylated histone H3) staining in embryos which were WT (i, vi) and injected with zNoxa MO1+2 (ii, vii), zNoxa mRNA (iii, viii), combined zNoxa MO1+2 and rcRNA (iv, ix) or zNoxa mismatch MO1+2 (v, x) at shield and 75% epiboly stages. Embryos stained by pH3 were lateral views with animal pole toward the top and dorsal to the right. (d) Quantification of numbers of pH3-positive cells with five embryos for each group in c. Error bars represent S.E.M. (n=5 embryos each). (e) A typical FACS analysis of PI (propidium iodide)-labeled cells in embryos for each group in c. (f) The FACS statistical data of three independent experiments in e. Error bars represent S.E.M. (n=3). (g) Confocal z-stacks of mitotic cells assessed by pH3 staining in embryos which were injected with zNoxa MO1+2 plus mNoxa mRNA or mNoxa mRNA at shield stage. (h) Quantification of numbers of pH3-positive cells with five embryos for each group in g. Error bars represent S.E.M. (n=5 embryos each). (i) A typical FACS analysis of PI-labeled cells in embryos for each group in g. (j) The FACS statistical data of three independent experiments in i. Error bars represent S.E.M. (n=3)

Full size image

zNoxa regulates mitosis during early embryonic development

To reveal the important role of zNoxa in early embryos, we examined whether there were any defects in mitotic index of the zNoxa-depleted embryos by staining the phosphorylated form of histone H3 (pH3), which had served as a G2/M phase marker.12, 13 As shown in Figures 3c and d, in comparison with WT and the injected mismatch-MO embryos (Figures 3c and i, v, vi, x and d), the zNoxa-depleted embryos exhibited a significant decrease in the number of mitotic cells (Figures 3c and ii, vii and d), and co-injection with zNoxa rcRNA partially rescued the decrease (Figures 3c and iv, ix and d). In contrast to the depletion, there were more mitotic cells in the injected embryos with 5 ng/μl of zNoxa mRNA at shield stage (Figures 3c and iii and d). At 75% epiboly stage, however, there were a little less mitotic cells in the embryos injected with zNoxa mRNA (Figures 3c and viii and d). The reason may be that zNoxa would not regulate mitosis while taking part in apoptosis and some embryonic cells processed apoptosis rather than mitosis from 75% epiboly stage. To confirm the above data, we further analyzed DNA contents of the dissociated cells from these embryos by FACS. As shown in Figures 3e and f, a, higher accumulation of G1-phase cells was observed at shield stage in the zNoxa morphant embryos (Figures 3e and ii and f) than in WT and the injected mismatch-MO embryos (Figures 3e and i,v and f), and this defect could be rescued by zNoxa rcRNA co-injection (Figures 3e and iv and f). Furthermore, higher accumulation of G2/M-phase cells was detected in the embryos injected with zNoxa mRNA (Figures 3e and iii and f). Again, however, no phase-specific cell blocks were observed at 75% epiboly stage (Figures 3e and vi, vii, viii, ix, x and f). These data together prove that zNoxa cannot simultaneously have dual roles in cell mitosis and apoptosis, and a smallamount of decrease of mitotic cells at 75% epiboly stage of the zNoxa morphants might be explained by the irreversible early damage caused by injection of zNoxa MOs. All data demonstrate that zNoxa regulates mitosis in early embryonic development.

Moreover, we attempted to ask whether mammalian Noxa could regulate mitosis in early development, as the critical BH3 domain of zNoxa was revealed to have a high degree of mammalian homology.8 Interestingly, the reduction of mitotic cells and higher accumulation of G1-phase cells caused by depletion of zNoxa could be rescued by co-injection with 10 ng/μl mNoxa mRNA. In addition, more mitotic cells (Figures 3g and h) and higher accumulation of G2/M-phase cells (Figures 3i and j) were also detected in the injected embryos with mNoxa mRNA. Therefore, these results indicate that mammalian Noxa can regulate mitosis in early zebrafish development, and may also control mitosis in early embryonic development of mammals.

zMcl1a and zMcl1b suppress the ability of zNoxa to regulate mitosis and apoptosis

Previous report indicated that Noxa might be a primary p53-response gene in mammals,1, 2 so we firstly examined whether p53 was the regulator of zNoxa in early embryonic development. Significantly, the zNoxa expression at shield stage was not affected by the p53 MO14 injection (Figure 4a); however, the expression of p21, a known p53-target gene, was largely reduced in the p53 MO-injected embryos (Supplementary Figure S3a), and almost no apoptotic cells were observed at bud stage in the p53 morphants (Supplementary Figure S3b and c). The data suggested that other regulators of zNoxa might exist in early embryonic development. Previous biochemical analysis revealed the selective association of Noxa with Mcl1 in mammals,15 so we tested whether zNoxa interacted with zMcl1a and zMcl1b by co-immunoprecipitation assays in zebrafish. We found that the overexpressed HA-zMcl1a and HA-zMcl1b were co-immunoprecipitated with zNoxa in embryonic cells, indicating that both zMcl1a and zMcl1b were binding partners of zNoxa (Figure 4b). Because zMcl1a and zMcl1b were revealed to be prosurvival proteins,8 so we guessed that zMcl1a and zMcl1b might be negative regulator of zNoxa. To check this hypothesis, a co-injection experiment of zNoxa mRNA and zMcl1a&b mRNA was performed. Obviously, the co-injection with zMcl1a&b mRNA suppressed the increase of mitotic cells caused by overexpression of zNoxa (Figures 4c and ii, iii and d), and largely rescued the cell cycle progression defects in the injected embryos with zNoxa mRNA (Figures 4e and ii, iii and 4f). Moreover, the co-injection with zNoxa MO1+2 (Figures 4c and v and d) inhibited the proliferation of mitotic cells in the zMcl1a&b-depleted embryos (Figures 4c and iv and d), and rescued cell cycle defects in the injected embryos with zMcl1a&b MOs (Figures 4e and iv, v and f). These data demonstrate that the ability of zNoxa is negatively regulated by zMcl1a&b. Furthermore, zMcl1a&b also suppressed the ability of mNoxa to regulate mitosis (Figures 4c and vi and d) and cell cycle progression (Figure 4e and vi and f) in zebrafish embryos, suggesting that Noxa might also cooperate with Mcl1 to control mitosis in mammalian embryos.

Figure 4
figure 4

The ability of zNoxa is suppressed by zMcl1a and zMcl1b. (a) Expression of zNoxa in WT and p53 MO-injected embryos was evaluated by real-time PCR and normalized to β-actin mRNA. (b) pcDNA3.1-HA-zMcl1a and pcDNA3.1-HA-zMcl1b were injected into zebrafish embryos to assess an interaction with endogenous zNoxa. At shield stage, zMcl1a-HA and zMcl1b-HA were immunoprecipitated (IP) and then the co-precipitation of zNoxa was assessed by immunobloting. (c) Confocal z-stacks of mitotic cells assessed by pH3 staining in embryos which were WT (i) and injected with zNoxa mRNA (ii), combined zNoxa mRNA and zMcl1a&b mRNA (iii), zMcl1a&b MOs (iv), zNoxa MO1+2 plus zMcl1a&b MOs (v) or mNoxa mRNA together with zMcl1a&b mRNA (vi) at shield stage. Embryos stained by pH3 were lateral views with animal pole toward the top and dorsal to the right. (d) Quantification of numbers of pH3-positive cells with five embryos for each group in (c). Error bars represent S.E.M. (n=5). (e) A typical FACS analysis of PI-labeled cells in embryos for each group in c. (f) The FACS statistical data of three independent experiments in e. Error bars represent S.E.M. (n=3). (g) Confocal z-stacks of apoptotic cells assessed by TUNEL in embryos which were WT (i), injected with zNoxa mRNA (ii), combined zNoxa mRNA and zMcl1a&b mRNA (iii), zMcl1a&b MO (iv), zNoxa MO1+2 together with zMcl1a&b MO (v) at bud stage. Embryos were lateral views with animal pole toward the top and dorsal to the right. (h) Quantification of numbers of TUNEL-positive cells with four embryos for each group in g. Error bars represent S.E.M. (n=4)

Full size image

To reveal whether zMcl1a&b could regulate the apoptosis caused by zNoxa, we examined the cell death status by TUNEL. As shown in Figures 4g and h, co-injection with zMcl1a&b mRNA suppressed the increase number of TUNEL-positive cells in the injected embryos with zNoxa mRNA (Figures 4g and ii, iii and h). In contrast, co-injection with zNoxa MO1+2 inhibited the increase of apoptotic cells in the zMcl1a&b-depleted embryos (Figures 4g and iv, v and h). Above data indicate that zMcl1a&b also suppresses the ability of zNoxa to regulate apoptosis.

zNoxa regulates activities of zWnt4b

To dissect the gene regulatory network controlled by zNoxa, we examined activities of key early developmental signaling pathways after depletion or ectopic expression of zNoxa. Surprisingly, we found that the expression of zebrafish Wnt4b (zWnt4b) was negatively regulated by zNoxa (Figures 5a and ii, iii, iv and b) and mNoxa (Figures 5a and v, vi and b). In contrast, as negative regulators of zNoxa, zMcl1a&b controlled the expression of zWnt4b positively (Figures 5a and vii, viii and b). Previous study had shown that zWnt4b was expressed exclusively in the floor plate,16 but its expression pattern in early-developing zebrafish embryos was not described, so we determined the expression of zWnt4b by WISH and real-time quantitative PCR. As shown in Figures 5c and d, the zWnt4b transcription begun from high-blastula stage and reached to the peak at shield stage, and then kept at a relatively stable level. The zWnt4b transcript was uniformly distributed in early embryos (Figure 5c). Moreover, a rabbit polyclonal antibody against amino acids 1–70 of zWnt4b was generated (Supplementary Figure S2b) and was used to analyze zWnt4b protein level in these early embryos. Western blot detection showed that the zygotic zWnt4b protein was translated at sphere stage and its peak was brought from shield stage (Figure 5e). Therefore, these data reveal a functional relationship between zNoxa and zWnt4b because their expression distribution in early embryos is basically consistent since high stage (Figures 1c and 5c).

Figure 5
figure 5

Activities of zWnt4b is regulated by zNoxa. (a) Expression of zWnt4b in embryos, which were WT (i) and injected with zNoxa MO1+2 (ii), zNoxa mRNA (iii), zNoxa MO1+2 plus zNoxa rcRNA (iv), mNoxa mRNA (v), together mNoxa mRNA with zNoxa mRNA (vi), zMcl1a&b mRNA (vii), or zMcl1a&b MO (viii), was detected by in situ hybridization. (b) Expression of zWnt4b in embryos for each group in a. (c) Expression pattern of zWnt4b was detected by in situ hybridization using specific antisense or sense probes. No staining was present in embryos hybridized with sense probe (the shield stage shown). Embryo were lateral views with anterior toward the top. (d) Expression of zWnt4b in embryonic development stages was checked by real-time PCR and normalized to β-actin mRNA. Error bars represent S.E.M. (n=3). (e) Zebrafish embryos were collected at different stages and protein was extracted for immunoblot analysis of zWnt4b and β-actin after discarding the yolk. β-actin served as a loading control

Full size image

zNoxa regulates mitosis through zWnt4b in early embryos

Subsequently, we attempted to understand whether zWnt4b was the mediator of mitosis regulated by zNoxa. Firstly, two morpholinos (zWnt4b MO1 and zWnt4b MO2) were designed to block the translation of zWnt4b. As the injected embryos with a high dose (0.8 mM) of zWnt4b MOs did not survive to bud stage, so lower doses (0.5 mM) of MOs were adopted. When 0.5 mM of MO1 or MO2 was injected, about 70% reduction in zWnt4b protein level was observed. When the two MOs were co-injected, almost no any zWnt4b protein was detected (Figure 6a). The co-injections with 20 ng/μl of the rescued zWnt4b mRNA (zWnt4b rcRNA), which was not targeted by the zWnt4b MO1 and MO2 by mismatching five nucleotides in zWnt4b mRNA (Supplementary Material), could restore the zWnt4b protein level (Figure 6a).

Figure 6
figure 6

Mitosis regulated by zNoxa is dependent on zWnt4b. (a) At the shield stage, protein was extracted from embryos and 30 μg of each sample was immunoblotted for zWnt4b and β-actin. β-Actin served as an internal control. (b) Confocal z-stacks of mitotic cells assessed by pH3 staining in embryos which were injected with zWnt4b Mis MO1&2 (i), zWnt4b MO1&2 (ii), zWnt4b mRNA (iii), zNoxa MO1+2 plus zWnt4b MO1&2 (iv), zNoxa mRNA together with zWn4b mRNA (v), mNoxa mRNA plus zWnt4b mRNA (vi), combined zMcl1a&b mRNA and zWnt4b MO1&2 (vii), or zMcl1a&b MO coupled with zWnt4b mRNA (viii) at shield stage. Embryos stained by pH3 were lateral views with animal pole toward the top and dorsal to the right. (c) Quantification of numbers of pH3-positive cells with five embryos for each group in (b). Error bars represent S.E.M. (n=5). (d) A typical FACS analysis of PI-labeled cells in embryos for each group in b. (e) The FACS statistical data of three independent experiments in d. Error bars represent S.E.M. (n=3). (f) Confocal z-stacks of apoptotic cells assessed by TUNEL in embryos that were wt and injected with zNoxa MO1+2 plus zWnt4b MO1&2 at bud stage. Embryos were lateral views with animal pole toward the top and dorsal to the right. (g) Quantification of numbers of TUNEL-positive cells with four embryos for each group in e. Error bars represent S.E.M. (n=4)

Full size image

Moreover, we observed the mitotic cells and cell cycle progression in these knockdown and overexpression embryos. Compared with WT (Figures 3c–i and d) and the injected mismatch-MO embryos (Figures 6b–i, c and d–i, e), the injection of zWnt4b MO1 and 2 (0.5 mM total) led to the increase of mitotic cells (Figures 6b–ii and c) and a higher accumulation of G2/M phase cells (Figures 6d and ii and e). In contrast, overexpression of zWnt4b resulted in reduction of mitotic cells (Figures 6b and iii and c) and a higher accumulation of G1 phase cells (Figures 6d and iii and e). The reduction in number of mitotic cells and the defects of cell cycle progression caused by zNoxa MO1+2 could be rescued by co-injection with zWnt4b MO1&2 (Figures 6b and iv, c and d-iv, e). Similarly, co-injection of zNoxa mRNA or mNoxa mRNA with zWnt4b mRNA suppressed the increase of mitotic cells (Figures 6b and v, vi and c) and the accumulation of G2/M phase cells (Figures 6 d-v, vi and e). Furthermore, we analyzed the effect of zMcl1a&b and zNoxa on zWnt4b. Both injection of zMcl1a&b mRNA plus zWnt4b MO1&2 and zMcl1a&b MO together with zWnt4b mRNA resulted in nearly normal cell mitosis (Figures 6b and vii, viii and c) and cell cycle progression (Figures 6d and vii, viii and e). In addition, we found that zWnt4b was not involved in the apoptosis mediated by zNoxa, because co-injection with zWnt4b MO1&2 in the zNoxa-depleted embryos could not rescue the decrease of apoptotic cells (Figures 6f and g). All these data indicate that zNoxa regulates mitosis through zWnt4b in early embryos, and zMcl1a&b and zNoxa have contrary role in mitosis mediated by zWnt4b, in which zMcl1a&b suppress the ability of zNoxa. mNoxa also regulates mitosis through zWnt4b in zebrafish, suggesting that mammalian Noxa might regulate mitosis mediated by members of Wnt family in early development.

zBik is the key mediator of apoptosis induced by zNoxa

Previously, zBik, a member of proapoptotic Bcl-2 family, was characterized to express from 75% epiboly stage in zebrafish embryos,8 and its apoptosis-related function was suggested.8 In vitro, Bik and Noxa were revealed to cooperate to induce mitochondrial fission.17 Here, we tried to determine whether zBik was the key mediator of zNoxa. We observed that the ectopic expression of zBik (Figures 7a and ii and b) induced apoptosis before 75% epiboly stage, but not in the zNoxa-depleted embryos (Figures 7a and iii and b). Moreover, co-injection with zBik mRNA could not rescue the induction of apoptotic cells in the zNoxa-depleted embryos (Figures 7a and v and b), and co-injection with zBik MO8 inhibited the increase of apoptotic cells in the injected embryos with zNoxa mRNA (Figure 7a and vi and b) at 75% epiboly stage. Furthermore, the expression of zBik was regulated by zNoxa at 75% epiboly stage, and also mediated by zMcl1a and zMcl1b at that stage (Figure 7c). All these data indicate that zNoxa firstly regulates the expression of zBik, and then cooperates with zBik to induce apoptosis. In addition, we wanted to know whether zNoxa could regulate mitosis if apoptosis was depressed. Compared with WT, the injected embryos with zNoxa mRNA plus zBik MO did not result in significant increase of mitotic cells at 75% epiboly stage (Figures 7d and e), indicating that zNoxa would not control mitosis after 75% epiboly stage, even apoptosis was inhibited.

Figure 7
figure 7

zBik mediates the apoptosis induced by zNoxa. (a) Confocal z-stacks of apoptotic cells assessed by TUNEL in embryos which were WT (v, iv) and injected with zBik mRNA (ii), zNoxa MO1+2 together with zBik mRNA (iii, v) or combined zNoxa mRNA and zBik MO (vi) at different developmental stages. Embryos were lateral views with animal pole toward the top and dorsal to the right. (b) Quantification of numbers of TUNEL-positive cells with four embryos for each group in a. Error bars represent S.E.M. (n=4). (c) Expression of zBik in embryos, which were WT and injected of zNoxa MO1+2, zNoxa mRNA, zNoxa MO1+2 plus zNoxa rcRNA, zMcl1a&b mRNA together with zNoxa mRNA, or combined zMcl1a&b MO and zNoxa MO1+2, was evaluated by real-time PCR and normalized to β-actin mRNA. Error bars represent S.E.M. (n=3). (d) Confocal z-stacks of mitotic cells assessed by pH3 staining in WT and injected of zNoxa mRNA plus zBik MO embryos at 75% epiboly stage. Embryos were lateral views with animal pole toward the top and dorsal to the right. (e) Quantification of numbers of pH3-positive cells with five embryos for each group in d. Error bars represent S.E.M. (n=5)

Full size image

Discussion

Noxa was originally isolated from an adult T-cell leukemia line as a phorbol 12-myristate-13-acetate inducible protein (PMAIP), and was considered as a canonical tumor suppressor,18 but its function was not reported in regulating embryonic development except proapoptotic function. In this study, we have characterized the zebrafish homolog zNoxa of mammalian Noxa, and observed its ubiquitous expression and distribution in early embryo stages, suggesting that zNoxa might be important in early embryonic development. Thereby, we have found for the first time that zNoxa is able to cooperate with zMcl1a and zMcl1b, and regulates mitosis through zWnt4b in early zebrafish embryonic development before 75% epiboly stage. In subsequent embryogenesis after 75% epiboly stage, zNoxa firstly induces the expression of zBik, and then cooperates with zBik to regulate apoptosis. Moreover, we have further revealed that zMcl1a and zMcl1b suppress the ability of zNoxa to regulate mitosis and apoptosis at different developmental stages (Figure 8). In vitro, Noxa was revealed to associate with a growth function regulated by glucose in human leukemia cells,6 and the proapoptotic proteins Bax and Bak were reported to have a role of survival rather than death in regulating T-cell proliferation,19 but no any related in vivo studies were performed so far. Here, we have firstly demonstrated a new proliferation function of zNoxa in zebrafish development.

Figure 8
figure 8

The proposed function relationship among zNoxa, zMcl1a, zMcl1b, zWnt4b and zBik in regulating mitosis and apoptosis before and after 75% epiboly in zebrafish embryos

Full size image

As a proapoptotic protein, Noxa was demonstrated to interact almost exclusively with prosurvival protein Mcl1 to repress survival function and then to induce apoptosis.15, 20 The Noxa/Mcl1 interaction might have major roles in glucose-dependent survival.6 A previous study also demonstrated the critical role of zMcl1a&b in curtailing apoptosis induction of early zebrafish embryos,8 but its relation between zMcl1a&b and zNoxa was not reported. In this study, we revealed interaction of both zMcl1a and zMcl1b with zNoxa, and the interaction inhibits proliferation function of zNoxa before 75% epiboly stage. After 75% epiboly stage, zMcl1a&b might exert a constraining role for apoptosis induced by zNoxa. Therefore, zMcl1a&b might function as negative modulators of zNoxa.

A significant finding in this study is the association and interaction between zNoxa and zWnt4b. zWnt4b was shown to express in the floor plate from 18 h.p.f.,16 but our current observation has indicated that zWnt4b expression is initiated at high stage, and its spatial distribution is similar to that of zNoxa. Moreover, depletion and overexpression studies show that zWnt4b is negatively regulated by zNoxa, which regulates mitosis through inhibiting the activation of zWnt4b in early embryonic development. Previous studies had revealed that Wnt4 might be positive regulator of mitosis21, 22, 23 and key mediator of apoptosis24, 25 in mammals. However, our current finding indicates that zWnt4b has a negative role for survival in early development of zebrafish embryos, and it is not involved in the apoptosis induced by zNoxa. Multiple amino-acid alignment of Wnt4 members in different species reveals more than 80% amino-acid identities (Supplementary Figure S4a), but phylogenic tree analysis shows two branches except fruitfly Wnt4; and zWnt4b and Medaka Wnt4b are clustered in one branch, whereas other sequences form another branch (Supplementary Figure S4b). These data may explain the functional differences of zWnt4b from other species Wnt4.

Another significant finding is about the cooperation between zBik and zNoxa to regulate apoptosis. Previous observation has found that in vitro delivery of Noxa to the mitochondria just results in a weak release of cytochrome c, but Bik and Noxa can cooperate to induce efficient cytochrome-c release, and then trigger apoptosis.17 In zebrafish embryos, zBik is not expressed until induced by zNoxa at 75% epiboly stage, and then the cooperation between zBik and zNoxa induces apoptosis. The finding also can provide an explanation for the absence of apoptosis in zebrafish embryos before 8 h.p.f.

In addition, the possible mechanisms of zNoxa effect on expression levels of zWnt4b and zBik are worth for further investigation. Lowman et al.6 have revealed that the phosphorylated Noxa promotes its cytosolic sequestration and suppresses its apoptotic function.7 Potential phosphorylation site analysis also observes a phosphorylation site in zNoxa (Supplementary Figure S5). Therefore, we think that the effect of zNoxa on expression levels of zWnt4b and zBik might be related the phosphorylated status. Of course, further studies for clarifying the mechanism need to be performed.

Here, we have demonstrated that mNoxa also can regulate mitosis in zebrafish embryos, although the function during embryogenesis has not been revealed in mammals. In mouse embryos, mNoxa is expressed at least from day 9,26 and apoptosis can be initially induced in day 13,27 so we believe that mNoxa may as well have dual roles in cell mitosis and apoptosis in embryonic development.

In conclusion, this study demonstrates that zNoxa is a novel regulator of cell mitosis and also a key mediator of apoptosis during zebrafish embryogenesis, and it has distinct roles at different developmental stages. Obviously, the new finding is very important for us to understand genetic basis and physiological roles of significant function genes in fish and even in vertebrates.28, 29

Materials and Methods

Maintenance of zebrafish

Zebrafish were maintained at 28.5 °C under a reproduction regime (14-h light/10-h dark cycle). All embryos were collected after natural spawning and staged in terms of previous report.30

Semi-quantitative PCR and real-time quantitative PCR

Total RNAs were extracted by RNeasy plus mini kit (Qiagen, Hilden, Germany), and the quality and concentration were determined by agarose electrophoresis. Then, RNAs were respectively reverse-transcribed with Goldscript cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) as described by the manufacturer. Semi-quantitative RT-PCR was performed as described previously31 in a volume of 25 μl as follows: 94 °C for 4 min, 94 °C for 30 s, 60 °C for 30 s and 72 °C for 1 min for 33 cycles followed by 72 °C for 10 min. Real-time quantitative PCR was performed by CFX96 Optics Module (Bio-Rad, Singapore) with SYBR Green I Dye as described previously.32 All results were analyzed according to a previous report.33 β-Actin was used as the internal control gene.

Plasmids and primers

All constructs and primer sequences were described in Supplementary Materials and Methods. All DNA constructs were verified by sequencing.

Whole-mount in situ hybridization

Embryos at distinct stages were fixed and pretreated as described.34 DIG-labeled (Roche, Mannheim, Germany) antisense and sense probes for zNoxa and zWnt4b were generated from PCR products, which included T7 RNA polymerase-binding sequence at the 3′ end. In situ hybridization was performed as described previously.35, 36

MO microinjection

All MOs were designed and synthesized by Gene-Tools (Philomath, OR, USA), and resuspended in nuclease-free water to give a solution of 2 mM and stocked at −80 °C. MO sequences are given as below.

Translation-blocking zNoxa MO1: 5′-CAGCGGTTTGCTCTTTCTTCGCCAT-3′

Five-base-mismatch zNoxa Mis MO1: 5′-GAGCCGTTTGTTCCTTCTTGGCCAT-3′

Translation-blocking zNoxa MO2: 5′-CAGCAAGTTTCTGTCACAGGTTCGC-3′

Five-base-mismatch zNoxa Mis MO2: 5′-GAGCAACTTCCTCTCACACGTTCGC-3′

Translation-blocking p53 MO: 5′-GCGCCATTGCTTTGCAAGAATTG-3′

Translation-blocking zMcl1a MO: 5′-GCCTAAAATCCAAACTCAGAGCCAT-3′

Translation-blocking zMcl1b MO: 5′-TGTCGTTGTTTCTTCCAGCGAACAT-3′

Translation-blocking zWnt4b MO1: 5′-CAGTCACAGAGGAGACTGTTGGCAT-3′

Five-base-mismatch zWbt4b Mis MO1: 5′-CGGTGACAGACGACACTGTAGGCAT-3′

Translation-blocking zWnt4b MO2: 5′-GTTGGCATCAGATTGCCTGTCTGTC-3′

Five-base-mismatch zWbt4b Mis MO2: 5′-CTTGCCATCACATTGCGTGTCCGTC-3′

Translation-blocking zBik MO: 5′-CTACAAACAAGGACACAATGGTGGA-3′

MOs were diluted to the required concentration and microinjected into naturally fertilized zebrafish eggs using a PLI-100 Pressure Injector (Harvard Apparatus, Greenvale, NY, USA). The injected embryos were raised at 28.5 °C in Danie’s medium37 and photographed by CCD as previous report.38

cDNA cloning and the capped mRNA synthesis

Full-length cDNAs enconding zNoxa, zMcl1a, zMcl1b, mNoxa and zWnt4b were amplified by RT-PCR using Taq DNA Polymerase (Fermentas, York, UK). The resulted cDNAs were subcloned into the pCS2+ vector. Then, the above vectors were linearized with NotI digestion to generate the capped mRNA using the mMessage mMachine SP6 kit (Ambion, Austin, TX, USA) according to the manufacturer’s instructions, and injected as described.39

Antibody production

The cDNA coding zNoxa- and zWnt4b-mature protein were subcloned into the pGEX-4T-1 expression vector. Fusion proteins were induced to express in BL21 (DE3) as described,40 and detected only in inclusion body. Proteins were purified with GST•Bind Kit (Novagen, Schwalbach, Germany) and immunized mouse as described previously.41, 42

Co-immunoprecipitation and western blotting

Embryos were boiled as described.43 Western blots were performed using antibodies except anti-zNoxa antibody according to the previous report,44 and blotting analysis using anti-zNoxa antibody was carried out as previously described.45 For co-immunoprecipitation experiments, embryos were collected and deyolked at the shield stage. The deyolked samples were lysed and incubated with anti-HA agarose conjugate (Sigma-Aldrich, St. Louis, MO, USA) according to the previous report.46 The anti-actin antibody was used at 1 : 1000 dilution, both anti-zNoxa antibody and anti-zWnt4b antibody was used at 1 : 200 dilution and the anti-HA antibody (Santa Cruz Biotechnology, Dallas, TX, USA) was used at 1 : 500 dilution.

Apoptosis detection

Apoptosis in zebrafish whole mounts was detected by TUNEL using In Situ Cell Death Detection Kit, Fluorescein (Roche) as described,14 and then photographed on a confocal microscope (NOL-LSM 710, Carl Zeiss, Germany) as previously described.47

Whole-mount immunohistochemistry

Embryos were fixed overnight in 4% PFA at 4 °C, then rinsed twice for 5 min in PBS and permeabilized in PBST(0.2% Triton X-100) at least 50 min at RT. After rinsing thrice for 5 min in PBS, embryos were incubated in blocking solution (1% BSA, 5% goat serum, 0.1% Triton X-100 in PBS) for at least 4 h prior to incubation with the primary antibody (1 : 1000 rabbit, Cell Signaling, Beverly, MA, USA). Following incubation at 4 °C overnight with gentle rocking and washing six times for 10 min with PBST(0.1% Triton X-100), embryos were incubated with secondary antibody-labeled Fluorescein (1 : 200 rabbit, Thermo, Pierce Antibody, Rockford, IL, USA) for at least 8 h at 4 °C with gentle rocking. Subsequently embryos were washed six times for 15 min in PBST, and then photographed on a confocal microscope (NOL-LSM 710, Carl Zeiss, Germany) as described.47

Fluorescence-activated cell sorting (FACS) analysis

The effect of cell cycle by injection of MO or mRNA was quantified by measuring the DNA content of each single embryo cells using a flow cytometer (Epics-XL Cytometer, Beckman Coulter, Indianapolis, IN, USA) as described previously.14, 48