Research ArticleCalcium signaling

The Bcl-2 Homolog Nrz Inhibits Binding of IP3 to Its Receptor to Control Calcium Signaling During Zebrafish Epiboly

See allHide authors and affiliations

Science Signaling  11 Feb 2014:
Vol. 7, Issue 312, pp. ra14
DOI: 10.1126/scisignal.2004480

Abstract

Members of the Bcl-2 protein family regulate mitochondrial membrane permeability and also localize to the endoplasmic reticulum where they control Ca2+ homeostasis by interacting with inositol 1,4,5-trisphosphate (IP3) receptors (IP3Rs). In zebrafish, Bcl-2–like 10 (Nrz) is required for Ca2+ signaling during epiboly and gastrulation. We characterized the mechanism by which Nrz controls IP3-mediated Ca2+ release during this process. We showed that Nrz was phosphorylated during early epiboly, and that in embryos in which Nrz was knocked down, reconstitution with Nrz bearing mutations designed to prevent its phosphorylation disrupted cyclic Ca2+ transients and the assembly of the actin-myosin ring and led to epiboly arrest. In cultured cells, wild-type Nrz, but not Nrz with phosphomimetic mutations, interacted with the IP3 binding domain of IP3R1, inhibited binding of IP3 to IP3R1, and prevented histamine-induced increases in cytosolic Ca2+. Collectively, these data suggest that Nrz phosphorylation is necessary for the generation of IP3-mediated Ca2+ transients and the formation of circumferential actin-myosin cables required for epiboly. Thus, in addition to their role in apoptosis, by tightly regulating Ca2+ signaling, Bcl-2 family members participate in the cellular events associated with early vertebrate development, including cytoskeletal dynamics and cell movement.

INTRODUCTION

The Bcl-2 protein family has both pro- and antiapoptotic functions, including the control of mitochondrial membrane permeability to cytochrome c, which leads to caspase activation (1). In addition, Bcl-2 proteins localize to the endoplasmic reticulum (ER) where they participate in Ca2+ homeostasis by binding to the Ca2+ channel, inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) (2). IP3 binding to IP3Rs promotes Ca2+ efflux from the ER into the cytosol, a process known as IP3-induced Ca2+ release (IICR). IP3Rs are composed of five types of domains (3): the suppressor domain (SD), the IP3-binding domain (IP3BD), the modulatory and transducing domain (MTD), the channel-forming domain (CFD), and the coupling domain (CD). The SD and IP3BD mediate binding to IP3, the MTD binds regulatory molecules that modify IP3R activity and transduces the signal from the IP3BD to the CFD, and the CD is implicated in IP3R tetramerization and channel opening (4). Bcl-2 inhibits IICR by interacting with the MTD (5), whereas Bcl-xL promotes Ca2+ release at low concentrations of IP3 by interacting with the CD (6).

Ca2+ release through IP3Rs mediates many cellular and physiological processes, such as cell proliferation, differentiation, apoptosis, fertilization, and embryonic development (7). In the zebrafish embryo, Ca2+ signaling plays a role in early stages of development (8). At the onset of the blastula stage, blastomeres that are in contact with the yolk release the contents of their cytoplasm into the yolk cell forming the yolk syncytial layer (YSL). At the end of blastulation, the embryo comprises blastomeres on top of the yolk cell, which contains the YSL. Before gastrulation, the blastomeres and the YSL begin to migrate from the animal to the vegetal pole in a process known as epiboly (9). In the YSL, IP3R-dependent Ca2+ signaling occurs at the onset of epiboly (10) and again once the blastomeres have passed the equator of the embryo (11). A contractile actin-myosin ring forms at the blastoderm margin (12) and is likely required for epiboly progression.

We recently demonstrated that the antiapoptotic protein Nrz, the zebrafish ortholog of Nrh-Bcl2l10, plays a crucial role in the progression of zebrafish epiboly. During early development, zygotic nrz is specifically expressed in the YSL, and knockdown of Nrz by morpholino (MO) injection results in epiboly arrest, constriction of the blastoderm margin, and detachment of the embryo from the yolk cell (13). This phenotype does not require caspase activity; instead, Nrz binds to IP3R1 on the surface of the ER and decreases IICR (14). Knockdown of nrz increases the concentration of Ca2+ at the blastoderm margin and promotes phosphorylation of myosin light chain (MLC) and premature formation of the circumferential actin-myosin ring, resulting in the detachment of blastomeres from the yolk (14).

Here, we investigated how Nrz regulates Ca2+ signals in the YSL during epiboly. We demonstrated that Nrz interacted with the IP3BD of IP3R1 and disrupted its binding to IP3. Furthermore, we showed that phosphorylation of Nrz prevented this interaction and promoted Ca2+ oscillations and formation of the actin-myosin ring during epiboly.

RESULTS

The BH4-BH1 region of Nrz is required to inhibit IICR

We previously demonstrated that Nrz interacts with IP3R1 via its N-terminal Bcl-2 homology (BH) 4 domain (14). Other members of the Bcl-2 family, including Bcl-2 and Bcl-xL, interact with IP3R1 (5, 6) and control IICR. Although the BH4 domain of Bcl-2 is sufficient to suppress IICR (15), the BH4 domain of Bcl-xL is not (16). To determine whether the BH4 domain of Nrz was required for IICR inhibition, we fused the BH4 domain of Nrz to the ER-targeting sequence of cytochrome b5 (NrzBH4Cb5) (Fig. 1A) and transiently overexpressed the fusion protein in HeLa cells. Similar to the Nrz BH4 domain that localizes in the cytoplasm (14) or full-length Nrz targeted to the ER (NrzCb5), NrzBH4Cb5 coimmunoprecipitated with endogenous IP3R1 (Fig. 1B). To assess whether NrzBH4Cb5 affects IICR, we monitored changes in the concentration of cytosolic Ca2+ using a Ca2+-sensitive fluorescent dye. In human umbilical vein endothelial cells, histamine triggers IP3 accumulation and increases cytosolic Ca2+, which can be reversed by inhibition of protein kinase C (17). We found that treating HeLa cells with histamine induced a transient increase in the concentration of cytosolic Ca2+ and that overexpression of NrzCb5, but not of NrzBH4Cb5, partially inhibited this effect (Fig. 1, C and D), suggesting that the BH4 domain is not sufficient to inhibit IICR.

Fig. 1 Nrz requires its BH4-BH1 region to bind IP3R1 and regulate IICR in HeLa cells.

(A) Schematic of full-length Nrz and the Cb5 domain of cytochrome b. Lines and colored boxes represent α helices and BH domains, respectively. (B) Representative Western blot of immunoprecipitation (IP) between endogenous IP3R1 and FLAG-NrzBH4 or FLAG-NrzCb5. n = 3. (C) Representative Ca2+ response curve of cells stimulated with 5 μM histamine at the indicated times. Cells were transfected with empty vector, FLAG-NrzBH4Cb5, or FLAG-NrzCb5. (D) Histogram showing the mean amplitude (±SD) of the histamine-induced Ca2+ peak. n ≥ 5 fields from biological replicates. (E) Representative Western blot of IP between endogenous IP3R1 and FLAG-tagged C-terminal Nrz deletion mutants. n = 3. (F) Representative Ca2+ response curve of cells stimulated with 5 μM histamine at the indicated times. Cells were transfected with empty vector, FLAG-Nrz1-67Cb5, or FLAG-Nrz1-94Cb5. (G) Histogram showing the mean amplitude (±SD) of the histamine-induced Ca2+ peak. n ≥ 5 fields from biological replicates. *P < 0.005 (Mann-Whitney U test) for (D) and (G).

Nrz contains a total of four BH domains, numbered BH4, BH3, BH1, and BH2, from the N to the C terminus (Fig. 1A). Thus, we asked whether longer N-terminal fragments of Nrz containing other BH domains could affect IICR. Consistent with the fact that the BH4 domain of Nrz mediates its interaction with IP3R1, we found that all forms of C-terminally truncated Nrz coimmunoprecipitated with endogenous IP3R1 (Fig. 1E). However, overexpression of the Nrz BH1, BH3, and BH4 domains (Nrz1-94Cb5) significantly reduced histamine-induced increased cytosolic Ca2+, whereas overexpression of the Nrz BH3 and BH4 domains (Nrz1-67Cb5) did not (Fig. 1, F and G), suggesting that the BH1 domain is required for IICR.

Nrz inhibition of IICR is independent of its interaction with the proapoptotic protein Bax

The fact that the BH1 domain of Nrz was required for its interaction with IP3R1 could mean that the ability of Nrz to inhibit IICR is related to its ability to bind proapoptotic members of the Bcl-2 family. The BH1 domain of Bcl-2 family proteins contains a conserved Asn-Try-Gly-Arg motif. Mutation of Gly145 to Ala in Bcl-2 prevents its heterodimerization with the proapoptotic protein Bax (18). To test whether the analogous residue in Nrz mediated dimerization with Bax, we mutated Gly85 to Ala in otherwise wild-type Nrz (NrzG85A). We found that full-length wild-type Nrz, but not NrzG85A or Nrz1-94, coimmunoprecipitated endogenous Bax (Fig. 2A) and reduced the abundance of cleaved poly(adenosine 5′-diphosphate–ribose) polymerase (PARP) induced by overexpression of Bax in HeLa cells (Fig. 2B), suggesting that the Gly85 residue in the BH1 domain is important for Nrz-dependent inhibition of Bax-mediated apoptosis. Thus, we asked whether Gly85 was required for Nrz to inhibit IICR. We fused the NrzG85A to the ER-targeting Cb5 domain (NrzG85ACb5) and expressed it in HeLa cells. We found that NrzG85ACb5 significantly reduced histamine-induced increases in cytosolic Ca2+ (Fig. 2, C and D). Collectively, these results suggest that the function of Nrz to inhibit IICR is most likely independent of binding to Bax.

Fig. 2 Nrz regulates IICR in HeLa cells independent of its interaction with Bax.

(A) Representative Western blot of IP between endogenous Bax and FLAG-tagged Nrz, Nrz1-94, or NrzG85A. n = 3. (B) Representative Western blot of cleaved PARP (cPARP) in cells transiently expressing zBax (zebrafish Bax) and FLAG-tagged full-length Nrz, Nrz1-94, or NrzG85A. n = 3. (C) Representative Ca2+ response curve of cells stimulated with 5 μM histamine at the indicated times. Cells were transfected with empty vector or FLAG-NrzG85ACb5. (D) Histogram showing mean amplitude (±SD) of the histamine-induced Ca2+ peak. n ≥ 5 fields from biological replicates. *P < 0.005 (Mann-Whitney U test).

Nrz binds to the IP3BD of zIP3R1

The BH4 domain of several antiapoptotic Bcl-2 family proteins, including Bcl-2 and Bcl-xL, interacts with either the MTD or CD of IP3R1, which is determined by the primary sequence of the BH4 domains (5, 6, 16, 19, 20). The primary sequence of the Nrz BH4 domain differs considerably from those of zBcl-2 and zBcl-xL (fig. S1). To determine which domain of Nrz interacts with IP3R1, we expressed fragments of zebrafish IP3R1 (zIP3R1) (Fig. 3A) in HeLa cells. We found that the IP3BD, but not the SD, MTDs, or CD, of zIP3R1 coimmunoprecipitated NrzCb5 (Fig. 3B) but not NrzCb5 lacking the BH4 domain (NrzΔBH4Cb5) (fig. S2), suggesting that the interaction between Nrz and IP3R1 is mediated by the specific binding of the IP3BD of IP3R1 to the BH4 domain of Nrz.

Fig. 3 Nrz binds the IP3BD of zIP3R1 in HeLa cells.

(A) Schematic of zIP3R1 domains. (B) Representative Western blot of IP between HA-tagged zIP3R1 domains and FLAG-NrzCb5. Bands corresponding to the zIP3R1 domains are indicated by a black arrow. n = 3. (C) Representative Western blot of IP between HA-tagged IP3BD or IP3BD binding site mutants and FLAG-NrzCb5. n = 3. (D) Docking model of the BH4 domain of Nrz (modeled by multi-template homology with Phyre2) binding to the IP3BD of mouse IP3R1 (PDB: 1N4K). The red dashed circle represents the IP3 binding site. (E) Representative Western blot of IP between HA-tagged IP3BD domain or IP3BD mutants (E255A, E410A, Y576F) and FLAG-NrzCb5. n = 3.

Nrz binds outside the IP3 binding pocket of IP3R1

A group of residues in the IP3BD form a positively charged pocket in the IP3-binding site (21). In mouse IP3R1, point mutations in these residues abolish the binding of both IP3 (21) and IRBIT (IP3R-binding protein released with IP3), a competitive inhibitor of IP3 (22, 23). These residues are conserved in zebrafish; therefore, we asked whether they were required for binding to Nrz. Individual mutations to each of the corresponding amino acids did not affect the ability of the IP3BD of zIP3R1 to coimmunoprecipitate NrzCb5 (Fig. 3C) when coexpressed in HeLa cells, suggesting that, unlike IRBIT, Nrz does not act as a competitive inhibitor of IP3 binding to IP3R1.

To identify which residues could govern the interaction of IP3R1 and Nrz, we performed molecular docking simulation between the IP3BD of mouse IP3R1 [Protein Data Bank (PDB): 1N4K] and the BH4 domain of Nrz (Fig. 3D). This analysis suggested that Nrz interacts with IP3R1 outside of the IP3 binding pocket, consistent with the mutational analysis (Fig. 3C), and identified three residues in IP3R1 that could mediate its interaction with Nrz. We mutated the corresponding residues in zIP3R1 (Glu255, Glu410, and Tyr576) and performed coimmunoprecipitation experiments in HeLa cells. We found that E225A prevented the interaction of NrzCb5 with the IP3BD (Fig. 3E), indicating that Glu255 of zIP3R1 was critical for binding to Nrz. Analysis of docking solutions also indicated that the interaction of Nrz and IP3R1 may depend on the interaction between zIP3R1 Glu255 and Cys20 of Nrz. Analysis of primary sequence alignments demonstrated that Cys20 is highly conserved across Nrz orthologs (fig. S3), suggesting that this residue could be important for binding between Nrz and zIP3R1 and that this interaction could be evolutionarily conserved.

Nrz prevents IP3 from binding to IP3R1

We found that Nrz inhibits IICR and binds to the IP3BD of IP3R1, suggesting that it could interfere with IP3 binding to IP3R1. We verified that Nrz does not affect the steady-state concentration of Ca2+ in the ER. In HeLa cells, expression of NrzCb5 did not affect increased cytosolic Ca2+ caused by treatment with thapsigargin (fig. S4), which inhibits the sarco-endoplasmic reticulum Ca2+ adenosine triphosphatase (ATPase) and thereby flushes Ca2+ from the ER (24).

To test whether Nrz could prevent binding of IP3 to IP3R1 to inhibit IICR, we used the IP3R-based IP3 sensor (IRIS), which contains the IP3BD flanked by enhanced cyan fluorescent protein (ECFP) and the yellow fluorescent protein Venus. When IP3 binds to IRIS, it induces a conformational change in the IP3BD that decreases fluorescence resonance energy transfer (FRET) from ECFP to Venus (25). Treating HeLa cells expressing IRIS with histamine induced a rapid increase in the ratio (R) between ECFP and Venus fluorescence intensities, indicating reduced FRET (Fig. 4A). Transient overexpression of full-length wild-type Nrz or Nrz1-94, but not NrzBH4 or NrzΔBH4, prevented histamine-induced changes in the IRIS FRET ratio (Fig. 4, A and B). We also directly assessed the binding of IP3 to IP3R1 using fluorescence polarization (26) to measure the interaction between IP3 covalently bound to fluorescein isothiocyanate (IP3-FITC) and a recombinant protein composed of the SD and IP3BD of IP3R1 (IP3R1-NTD). When we incubated IP3-FITC with IP3R1-NTD in the presence of recombinant Nrz, we found that Nrz reduced fluorescence polarization in a concentration-dependent manner (half-maximal inhibitory concentration, 2.96 × 10−8 ± 1.77 × 10−8 M) (Fig. 3C). Collectively, these results demonstrate that Nrz inhibits the binding of IP3 to IP3R1.

Fig. 4 Nrz prevents IP3 binding to IP3R1.

(A) Representative curve of the FRET signal change of IRIS in HeLa cells stimulated with 5 μM histamine and expressing the indicated proteins. The FRET signal was calculated as the ratio (R) of fluorescence emission of ECFP to fluorescence emission of Venus with excitation at 405 nm. (B) Histogram showing the mean FRET signal change (±SD) induced by histamine stimulation. n = 30 cells from at least 3 biological replicates. **P < 0.0001 (Mann-Whitney U test); n.s., nonsignificant (P > 0.1). (C) Curve showing the mean fluorescence polarization (±SD) of IP3-FITC in the presence of IP3R1 NTD and the indicated concentration of recombinant Nrz protein. n = 3 independent triplicate measurements.

Phosphorylation of Nrz inhibits its interaction with IP3R1

We hypothesized that the interaction between endogenous Nrz and IP3R1 could be regulated by signal transduction mechanisms such as phosphorylation. Phosphorylation of Bcl-2 on a Thr and two Ser residues in the loop between its BH4 and BH3 domains increases Ca2+ release from the ER (27). Similar to Bcl-2, Nrz has a Thr and two Ser residues in the loop between the BH4 and BH3 domains (Fig. 5A) that could be phosphorylated. We mutated these three residues to generate phosphomimetic (T26D, S29D, S31D, NrzDDD) and nonphosphorylatable (T26A, S29A, S31A, NrzAAA) amino acid versions of Nrz. When transiently expressed in HeLa cells, NrzAAACb5, but not NrzDDDCb5, significantly decreased histamine-induced increased cytosolic Ca2+ (Fig. 5, B and C). Moreover, NrzAAACb5, but not NrzDDDCb5, coimmunoprecipitated the IP3BD of IP3R1 (Fig. 5D). To identify which of these three amino acids was important for Nrz interaction with IP3R1, we generated single and double mutants at each position. We found that Nrz mutants containing S31D did not coimmunoprecipitate the IP3BD of IP3R1 (Fig. 5E), suggesting that Ser31 phosphorylation by an unknown kinase could play a critical role in inhibiting the interaction between Nrz and IP3R1 and promoting IICR.

Fig. 5 Phosphomimetic mutant Nrz does not bind IP3R1 or inhibit IICR in HeLa cells.

(A) Primary sequence of the Nrz protein including the BH4 and BH3 domains. Putative phospho-serines and phospho-threonine are shown in red. (B) Representative Ca2+ response curve of cells stimulated with 5 μM histamine at the indicated times. Cells were transfected with empty vector, FLAG-NrzAAACb5, or FLAG-NrzDDDCb5. (C) Histogram showing the mean amplitude (±SD) of the histamine-induced Ca2+ peak. n ≥ 5 fields from biological replicates. *P < 0.005 (Mann-Whitney U test). (D) Representative Western blot of IP between HA-tagged IP3BD and FLAG-tagged NrzAAACb5 or NrzDDDCb5. n = 3. (E) Representative Western blot of IP between HA-tagged IP3BD and FLAG-tagged Nrz or phosphomimetic Nrz mutants.

Domains of Nrz required for the inhibition of IICR are also required for zebrafish epiboly

We previously reported that Nrz knockdown by antisense MO injection results in the constriction of the blastoderm margin and the detachment of the embryo from the yolk cell during epiboly (13). The injection of nrz MO induces a robust increase in cytosolic Ca2+ and embryonic lethality, which can be prevented by co-injecting mRNA encoding NrzCb5 but not NrzΔBH4Cb5 (14). Here, we found that both the BH4 and BH1 domains of Nrz were required to inhibit IICR in HeLa cells (Fig. 1). Therefore, we tested whether these domains were sufficient to reverse the phenotype induced by nrz MO injection. We co-injected embryos with nrz MO and in vitro–synthesized mRNA encoding various truncated forms of Nrz and monitored the percentage of embryos showing abnormal margin constriction during epiboly (Fig. 6A). Co-injection of nrzbh4cb5 did not prevent epiboly defects induced by nrz MO (Fig. 6, A and B), consistent with the observation that the BH4 domain was not sufficient to reduce IICR in HeLa cells. Moreover, co-injection of nrz1-94cb5, but not of nrz1-67cb5, also prevented epiboly defects induced by nrz MO (Fig. 6C), demonstrating a requirement for the BH1 domain of Nrz in this process. We also found that the ability of Nrz to prevent IICR was independent of its ability to bind to Bax in HeLa cells (Fig. 2C). Similarly, co-injection of nrzG85A, which encodes a form of Nrz that does not bind Bax (Fig. 2A), prevented epiboly defects induced by nrz MO (Fig. 6D). Finally, we found that Nrz phosphorylation could impair its interaction with IP3R1 in HeLa cells (Fig. 5D). In zebrafish, co-injection of nrz MO with nrzAAAcb5, but not nrzDDDcb5, reduced the percentage of embryos with epiboly defects induced by nrz MO (Fig. 6E). Thus, the domains and residues of Nrz required to inhibit IICR in cultured cells are also required for zebrafish epiboly.

Fig. 6 Nrz requires its BH4-BH1 region to regulate zebrafish epiboly.

(A) Representative bright-field images at 50% epiboly of the phenotype (constriction at the blastoderm margin) induced by the injection of nrz MO alone or in combination with nrzcb5 or nrzbh4cb5 mRNA. Scale bar, 100 μm. (B to E) Histograms showing the mean percentage (±SD) of embryos displaying the phenotype in (A) at 8 hours after injection. n ≥ 3 independent injections (80 to 100 embryos per injection per condition). *P < 0.005 (Mann-Whitney U test).

Phosphorylation of Nrz enables Ca2+ signaling, actin-myosin ring formation, and epiboly

During epiboly, IP3-dependent cyclic increases in cytosolic Ca2+ concentration, referred to as Ca2+ waves, occur in the YSL (10), where nrz is expressed (13). Because Nrz inhibits IP3 binding to IP3R1 in HeLa cells (Fig. 4, A to C), we hypothesized that Nrz activity is negatively regulated to enable the generation of Ca2+ waves. In HeLa cells, we found that mutants of Nrz that mimic phosphorylated Thr and Ser do not bind to IP3R1 (Fig. 5D) or inhibit IICR (Fig. 5C). To examine whether Nrz is phosphorylated during epiboly in vivo, we injected nrzcb5 mRNA and immunoprecipitated NrzCb5 from ER fractions prepared from the YSL of embryos at different developmental stages. Western blot analysis revealed that Nrz Ser phosphorylation was barely detectable before the onset of epiboly and increased robustly by 30% epiboly (Fig. 7A), suggesting that Nrz phosphorylation could be important to enable IP3-dependent Ca2+ signaling in vivo during this process.

Fig. 7 Nrz phosphorylation is required for generation of cyclic Ca2+ waves in the YSL.

(A) Representative Western blot of IP for FLAG-NrzCb5 from ER of YSL cells isolated from embryos at the indicated stages. n = 3. (B) Ca2+ concentration ([Ca2+]) in the external YSL of uninjected wild-type embryos or embryos injected with nrz MO and nrzcb5 or nrzAAAcb5 mRNAs. Scale bars, 20 μm. (C) Representative graph of the [Ca2+] variation shown in (B). (D and E) Histograms showing the amplitude (D) or period (E) (mean ± SD) of YSL [Ca2+] as in (C). *P < 0.005 (Mann-Whitney U test). n ≥ 6 embryos per condition.

We tested whether replacing endogenous Nrz with a nonphosphorylatable form of Nrz could disrupt Ca2+ signaling. In wild-type embryos injected at the 128-cell stage with a Ca2+-sensitive dye, we observed cyclic Ca2+ waves in the YSL at 30% epiboly (Fig. 7, B to E, and movie S1). To evaluate the contribution of Nrz phosphorylation to the generation of these Ca2+ waves, we co-injected nrz MO and nrzAAAcb5 or nrzcb5. We found that nrzcb5, but not nrzAAAcb5, restored normal Ca2+ waves in the absence of endogenous Nrz (Fig. 7, B to E, and movies S2 and S3), suggesting that the phosphorylation of Nrz enables the generation of YSL Ca2+ waves at the beginning of epiboly.

An actin-myosin ring forms at the blastoderm margin in zebrafish embryos at 40% epiboly, and localized disruption of actin in the YSL at 60% epiboly delays cell movements associated with epiboly progression (12). Similarly, treating zebrafish embryos at 50% epiboly with cytochalasin B, which inhibits actin network formation, or with a Ca2+ chelator disrupts actin ring formation and epiboly (28), suggesting that Ca2+ waves are important for the formation of the actin-myosin ring. Given that nrzAAAcb5 disrupted Ca2+ waves when co-injected with nrz MO, we examined the formation of the actin-myosin ring. Co-injection of nrz MO with nrzAAAcb5, but not nrzcb5 or nrz MO injection alone, reduced F-actin staining at the blastoderm margin at 50% epiboly (Fig. 8, A and B), suggesting that the expression of nonphosphorylatable Nrz inhibits YSL Ca2+ dynamics and thereby disrupts the formation of the actin-myosin ring.

Fig. 8 Nrz phosphorylation is required for epiboly progression and actin ring formation.

(A) Phalloidin-rhodamine staining of F-actin in embryos at the shield stage injected with nrz MO alone or in combination with nrzcb5 or nrzAAAcb5 mRNA. White arrows indicate the actin-myosin ring. (B) Histogram showing the relative actin fluorescence (mean ± SD). Relative fluorescence is indicative of the amount of F-actin at the blastoderm margin and was calculated by using the ratio of the rhodamine fluorescence at the blastoderm margin to the rhodamine fluorescence in the blastomeres. n = 30 embryos per condition. (C) Histogram showing the mean percentage (±SD) of epiboly progression at 8 hours post-fertilization (hpf). n = 90 embryos from 3 independent injections. (D) Representative images of the phenotypes (epiboly arrest) of embryos at 5 hpf injected with nrz MO and nrzcb5 or nrzAAAcb5 mRNA. NrzCb5 restores epiboly progression (50% epiboly), in contrast to NrzAAACb5 (epiboly arrest). n = 3 independent injections (80 to 100 embryos per injection per condition). Scale bars, 20 μm. *P < 0.005 (Mann-Whitney U test) for (B) to (D).

The above data suggested that the phosphorylation of Nrz could be important for epiboly progression. In embryos injected with nrz MO, co-injection of nrzAAAcb5, but not of nrzcb5, resulted in epiboly delay (Fig. 8C) or arrest (Fig. 8D). Collectively, these results suggest that by 50% epiboly, endogenous Nrz is mostly phosphorylated and cannot bind to IP3R1 and is thus permissive to Ca2+ signaling required for actin assembly and contractility.

DISCUSSION

We found that Nrz binds outside the IP3 binding site on the IP3BD of IP3R1 and prevents IP3 binding, revealing a new regulatory mechanism for IP3Rs by Bcl-2–related proteins. Bcl-2 reduces Ca2+ release by interacting with the MTD of IP3R1 (5, 15), and Bcl-xL sensitizes IP3R1 to low IP3 concentrations by interacting with the CD (6). The differential binding of Bcl-2 proteins to various domains of IP3R1 may be determined by divergent amino acid sequences in the BH4 domains (16, 20). Here, we found that the BH4 domain of Nrz was sufficient for it to interact with IP3R1, suggesting that residues within this domain could be critical for determining the interaction with the IP3BD. Molecular docking simulation identified a conserved Cys (Cys20 in Nrz) as a potential factor in mitigating the binding of Nrz to the IP3BD of IP3R1. We found that Glu255 of zIP3R1 most likely contacts Nrz Cys20 and is essential for binding between these proteins. Whether Nrz Cys20 is required to bind IP3R1 and whether analogous residues confer specificity of binding of Bcl-2 family and other proteins to IP3BDs remain to be determined.

Several proteins interact with the IP3BD or decrease IP3 binding to the IP3R1. Beclin-1 binds to the IP3BD of IP3R1, but Beclin-1 knockdown has no effect on Ca2+ homeostasis (29). Sodium-potassium ATPase binds to the IP3BD of IP3R1 (30) and promotes Ca2+ release in the absence IP3 (31). Carbonic anhydrase–related protein (CARP) inhibits IP3 binding to IP3R1 by reducing the affinity of the receptor for IP3 (32). CARP interacts with the MTD of IP3R1 (32), suggesting that Nrz and CARP act by distinct mechanisms. IRBIT binds the IP3BD of IP3R1 and directly competes with IP3 for binding to its receptor (22, 23). We found that mutations in residues of IP3R1 required for interaction with IRBIT and IP3 were not required for binding to Nrz, suggesting that Nrz and IRBIT act by distinct mechanisms to inhibit IP3R1. The Nrz BH4 domain binds outside the IP3 binding site of IP3R1, and the inhibition of IICR by Nrz requires the 94 most N-terminal amino acids including the BH4, BH3, and BH1 domains. The BH4 and BH3 domains combined were not sufficient to inhibit IICR without the BH1 domain, despite still being able to bind to IP3R1. These findings imply that Nrz is an allosteric inhibitor. Binding of Nrz could produce a conformational change in the IP3BD of IP3R1 that decreases the affinity of the IP3 binding site for IP3. Alternatively, Nrz, potentially involving its BH1 domain, could hinder access of IP3 to the IP3 binding site on IP3R1 without direct competition for key residues.

The mutation of putative phosphorylated residues to phosphomimetic amino acids in Nrz prevented its ability to inhibit IICR and bind IP3R1. Mutation of analogous residues in Bcl-2 to nonphosphorylatable amino acids enhances its ability to reduce Ca2+ release from the ER (27). However, this effect is indirect. Phosphorylation of Bcl-2 reduces its interactions with proapoptotic proteins, such as Bax, which increases ER Ca2+ content by reducing the Ca2+ leak (33). Nrz did not alter basal ER Ca2+ concentration, and inhibition of IICR by Nrz did not require interaction with Bax. Moreover, phosphomimetic mutations in Nrz prevented interaction with the IP3BD of IP3R1, suggesting that Nrz acts directly on IP3R1-dependent Ca2+ release in a manner that depends on Nrz phosphorylation.

We found that Nrz was phosphorylated during early zebrafish epiboly, a process that requires Ca2+ signaling (28), and expression of Nrz with nonphosphorylatable amino acids suppressed cyclic Ca2+ waves at the beginning of epiboly, compromised actin-myosin ring formation at the blastoderm margin, and delayed or disrupted epiboly. Ca2+ waves begin at the time when the actin-myosin ring is formed (12, 28), and our data are consistent with the fact that Nrz must be inhibited to enable this process. We observed that the period of Ca2+ waves in the YSL was about 7 min. During myofibrillogenesis, actin-myosin network assembly is regulated by Ca2+ transients with a similar period (34, 35). At the blastoderm margin, actin associates with phospho–myosin-II (36). We previously demonstrated that increased cytosolic Ca2+ concentrations in embryos injected with nrz MO promotes MLC phosphorylation, which leads to premature actin-myosin ring formation at 30% epiboly (14). Thus, Ca2+ waves during early epiboly could time MLC phosphorylation and the progressive assembly of the actin-myosin ring.

Our results identify the role of Nrz in IICR and emphasize the role of Bcl-2 family proteins in the regulation of intracellular Ca2+. We recently demonstrated that Bcl-wav, a newly characterized Bcl-2 family member, is essential for convergence and extension movement in zebrafish gastrulation. During this process, Bcl-wav regulates the formation of actin protrusions via its role in mitochondrial Ca2+ uptake (37). Together, these studies emphasize the diversity of functional roles of Bcl-2 family proteins in numerous cellular processes.

MATERIALS AND METHODS

MOs, reagents, and antibodies

The nrz MO was designed according to the manufacturer’s recommendations (Gene Tools, LLC) and had the following sequence: 5′-CATTTTCCTCCCAGCGATGTCAGAC-3′. A second MO that contained four mismatches relative to nrz MO was used as a negative control (control MO). The sequence was as follows: 5′-CATTATCCTGCCAGCCATGTGAGAC-3′ (13).

The following antibodies were used: IP3R1 (PA1-901, Thermo Scientific), rabbit FLAG (F-7425, Sigma-Aldrich), mouse FLAG (F-3165, Sigma-Aldrich), rabbit hemagglutinin (HA) (Ab75640, Abcam), mouse HA (MMS-101P, Covance), PARP (Ab6079, Abcam), vinculin (SC-55465, Santa Cruz Biotechnology), Bax (5023, Cell Signaling), and phospho-serine (05-1000, Millipore).

Vector construction and mRNA in vitro transcription

The open reading frames of nrz and zbax were cloned into the pCS2+ expression vector as previously described (13). The transmembrane domain of Cb5 was cloned in pCS2+ as previously described (14). Mutants of Nrz were inserted between the Cla I and Xho I restriction sites in the pCS2+Cb5 or pCS2+ vectors. For cloning of zip3r1, total RNA from embryos was extracted using TRIzol reagent (Invitrogen) following the manufacturer’s instructions, and complementary DNA (cDNA) was generated with a cDNA synthesis kit (iScript cDNA synthesis kit, Bio-Rad) with 500 ng of total RNA and gene-specific primers. An HA tag was added to the N terminus by polymerase chain reaction. Gene fragments corresponding to the IP3R1 domains were amplified and cloned between the Cla I and Xho I or between the Eco RI and Xho I sites in pCS2+. pcDNA3.1-IRIS was provided by K. Mikoshiba (25). In vitro transcription for the different constructs was performed with the SP6 mMESSAGE mMACHINE kit according to the manufacturer’s recommendations (Ambion).

Zebrafish husbandry, manipulation, injection, and analysis

Zebrafish (crosses between AB and TU or between AB and TL strains) were raised and maintained according to standard procedures (38). Zebrafish embryos were grown in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4, pH 7.6) at 28.5°C.

MO (0.5 mM) and mRNA (100 ng/ml) injections were performed between the one- and four-cell stages. Embryos were then observed every 15 min, and embryos showing margin constriction and cell detachment from the yolk were recorded as “embryos with a phenotype.” Calcium Green-1 Dextran (Invitrogen) at 250 μM was injected at the 128-cell stage into the top of the yolk, close to the margin of the yolk and blastoderm.

Images of embryos at 8 hpf were acquired with a Nikon TE30 inverted microscope and MetaMorph software. Epiboly was quantified with Image J software by measuring the distance of the blastoderm margin from the animal pole (D) and the size of the embryos (S). Epiboly progression data were obtained by using the ratio D/S. F-actin staining of embryos was performed as previously described (36). In brief, embryos at the desired stage were fixed overnight in 4% paraformaldehyde at 4°C and washed in PBT [0.1% Triton-X 100 in phosphate-buffered saline (PBS)]. They were then permeabilized for 1 hour in 0.5% Triton-X in PBS and subsequently incubated in block solution (10% goat serum, 1% dimethyl sulfoxide, 0.1% Triton-X 100 in PBS) for 5 hours. The embryos were then incubated in rhodamine-phalloidin (Invitrogen) overnight at 4°C and washed three times in PBT.

Embryos used for Ca2+ imaging were injected with Calcium Green-1 Dextran (Invitrogen) at 250 μM into the top of the yolk at the blastoderm margin of 128-cell stage embryos (10), dechorionated, and mounted in low–melting point 0.7% agarose in E3 medium on the stage of a Zeiss LSM 780 microscope and incubated at 28.5°C. Stacks of 4 μm were acquired, and the sum of the first three stacks was used to visualize variations in Ca2+ concentration in the external YSL.

ER from the YSL was isolated as described previously (14). The 100,000g pellet was resuspended in TNE buffer [10 mM tris-HCl, 200 mM NaCl, 1 mM EDTA (pH 7.4), 1 mM β-glycerophosphate, 1 mM orthovanadate, 0.1 mM sodium pyrophosphate containing protease inhibitors] and was used for FLAG immunoprecipitation.

Cell culture, immunoprecipitation, Ca2+ imaging, and FRET

HeLa cells were grown in Dulbecco’s modified Eagle’s medium (high glucose) (Gibco) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin (Gibco) at 37°C in a 5% CO2 humidified atmosphere.

For endogenous IP3R and Bax coimmunoprecipitation experiments, 6 × 106 HeLa cells were transfected with the indicated vectors. After 24 hours, cells were lysed in TNE buffer. Extracts were precleared with protein G–Sepharose beads for 1 hour at 4°C and then incubated overnight with 5 μg of IP3R or Bax primary antibody. The extracts were then incubated with protein G–Sepharose beads for 2 hours. Immunoprecipitated fractions were washed three times with TNE and analyzed by immunoblotting. For HA and FLAG immunoprecipitation, 1 × 106 HeLa cells were transfected with the indicated vectors and lysed and processed as above. One microgram of primary rabbit HA or rabbit FLAG was used.

For Ca2+ imaging, HeLa cells cultured in Nunc Lab-Tek chambered cover glass were incubated with 5 μM FluoForte (Enzo Life Sciences) in a Ca2+-free balanced salt solution (BSS) [121 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl2, 6 mM NaHCO3, 5.5 mM d-glucose, 25 mM Hepes (pH 7.3)] for 1 hour at 37°C. Fluorescence values were collected using a Zeiss LSM 780 confocal microscope. After 10 s of measurement, 5 μM histamine (Sigma) or thapsigargin (Enzo Life Sciences) in Ca2+-free BSS was injected. For IRIS FRET, HeLa cells cultured in Nunc Lab-Tek chambered cover glass were cotransfected with pcDNA3.1-IRIS (25) and pCS2+ encoding the various Nrz mutants (nmole ratio, 1:3). Before imaging, the medium was replaced with BSS. FRET was measured with a Zeiss LSM 780 confocal microscope by excitation at 405 nm. ECFP and Venus fluorescence were detected at 450 to 510 nm and 525 to 565 nm, respectively. After 10 s of measurement, 5 μM histamine in BSS was injected. Changes in the FRET signal were calculated using the FECFP/FVenus ratio (R).

IP3-FITC fluorescence polarization measurement

IP3-FITC, IP3, and the recombinant N-terminal domain protein were purchased with the HitHunter IP3 assay kit from DiscoveRx. Recombinant Nrz protein was produced as previously described (13). IP3-FITC at 0.5 nM was incubated with 4 nM of recombinant NTD and the various recombinant Nrz proteins for 30 min at 25°C. Fluorescence polarization was recorded at 25°C with 485-nm excitation and 530-nm emission filters, using a Mithras LB 940 multimode microplate reader (Berthold Technologies).

Sequence alignment

Nrz ortholog sequences were found in the National Center for Biotechnology Information protein database. Sequence alignment was performed with the Clustal Omega tool at http://www.ebi.ac.uk/Tools/msa/clustalo/. An image of the alignment was obtained with Jalview software (39).

Molecular docking simulation

The Nrz protein structure was obtained by analyzing the Nrz amino acid sequence on the Phyre2 (40) homology modeling server. The Nrz BH4 domain structure was extracted, and a docking experiment was performed against the IP3R1 binding domain crystal structure (PDB: 1N4K) using the PatchDock server (41). Images were acquired with UCSF Chimera software (42).

Statistical analysis

Statistical significance was analyzed using the Mann-Whitney U test; P < 0.005 was considered significant.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/7/312/ra14/DC1

List of primers

Fig. S1. Sequence alignment of the BH4 domains of Nrz, zBcl-2, and zBcl-xL.

Fig. S2. Interaction of Nrz with IP3BD requires the BH4 domain.

Fig. S3. Sequence alignment of the BH4 domains of Nrz orthologs.

Fig. S4. Nrz does not reduce the ER Ca2+ content.

Movie S1. Ca2+ waves in the YSL of a wild-type uninjected embryo.

Movie S2. Ca2+ waves in the YSL of an nrzcb5-injected embryo.

Movie S3. Ca2+ waves in the YSL of an nrzAAAcb5-injected embryo.

REFERENCES AND NOTES

Acknowledgments: We thank K. Mikoshiba for providing the IRIS construct and S. Chabaud (UBET, Centre Léon Bérard), C. Vanbelle (CeCILE—SFR Santé Lyon-Est), and C. Bouchardon (CeCILE—SFR Santé Lyon-Est) for their technical assistance. Funding: This work was supported by AFMTéléthon, Ligue nationale contre le cancer (comité de la Drôme), Fondation ARC pour la recherche sur le cancer. B.B. and A.N. are fellows of the Ministère de la Recherche. N. Peyriéras received support from France BioImaging ANR-10-INBS-04 (BioEmergences platform). Author contributions: B.B., A.N., J.P., and N. Popgeorgiev performed the experiments. B.B., N. Peyriéras, R.R., and G.G. designed the experiments. B.B. and G.G. wrote the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article