Copyright Ó 2007 by the Genetics Society of America DOI: 10.1534/genetics.107.079517

Expression of the GADD45 Homolog (CG11086) Affects Egg Asymmetric Development That Is Mediated by the c-Jun N-Terminal Kinase Pathway

Gabriella Peretz, Anna Bakhrat and Uri Abdu1 Department of Life Sciences and the National Institute for Biotechnology in the Negev, Ben-Gurion University, Beer-Sheva, 84105 Israel Manuscript received July 26, 2007 Accepted for publication September 21, 2007

ABSTRACT The mammalian GADD45 (growth arrest and DNA-damage inducible) family is composed of three highly homologous small, acidic, nuclear : GADD45a, GADD45b, and GADD45g. GADD45 proteins are involved in important processes such as regulation of DNA repair, cycle control, and apoptosis. Annotation of the Drosophila melanogaster revealed that it contains a single GADD45-like (CG11086; D-GADD45). We found that, as its mammalian homologs, D-GADD45 is a nuclear protein; how- ever, D-GADD45 expression is not elevated following exposure to genotoxic and nongenotoxic agents in Schneider cells and in adult flies. We showed that the D-GADD45 transcript increased following immune response activation, consistent with previous microarray findings. Since upregulation of GADD45 proteins has been characterized as an important cellular response to genotoxic and nongenotoxic agents, we aimed to characterize the effect of D-GADD45 overexpression on D. melanogaster development. Overexpression of D-GADD45 in various tissues led to different phenotypic responses. Specifically, in the somatic follicle cells overexpression caused apoptosis, while overexpression in the germline affected the dorsal–ventral polarity of the eggshell and disrupted the localization of anterior–posterior polarity determinants. In this article we focused on the role of D-GADD45 overexpression in the germline and found that D-GADD45 caused dor- salization of the eggshell. Since mammalian GADD45 proteins are activators of the c-Jun N-terminal kinase ( JNK)/p38 mitogen-activated protein kinase (MAPK) signaling pathways, we tested for a genetic interac- tion in D. melanogaster. We found that eggshell polarity defects caused by D-GADD45 overexpression were dominantly suppressed by mutations in the JNK pathway, suggesting that the JNK pathway has a novel, D-GADD45-mediated, function in the Drosophila germline.

HE GADD45 (growth arrest and DNA-damage in- 1988, 1989; Kastan et al. 1992). It has been shown that T ducible) gene family plays an important role in reg- all the GADD45 proteins mediate cell-cycle regulation ulation of DNA repair, cell cycle control, and apoptosis, through interactions with PCNA (Kelman and Hurwitz although its involvement in development is largely un- 1998; Azam et al. 2001), the cyclin-dependent kinase in- known. The GADD45 gene family is composed of three hibitor (Kearsey et al. 1995), and the Cdk/cyclin B highly homologous (55–58% overall identity at the amino complex (Zhan et al. 1999; Jin et al. 2002; Vairapandiet al. acid level), small, acidic, nuclear proteins: GADD45a 2002). The potential role of GADD45 proteins in apopto- (Fornace et al. 1989), GADD45b (MyD118, Abdollahi sis emanates from the observation that GADD45 expres- et al. 1991), and GADD45g (CR6, cytokine response gene sion is enhanced during apoptosis following induction by 6) (Beadling et al. 1993). In recent years, evidence has a variety of genotoxic agents (reviewed in Sheikh et al. emerged that the proteins encoded by these play 2000). Several studies have shown that GADD45 proteins similar but not identical roles in terminal differentiation may play a role in apoptosis via activation of the c-Jun and negative growth control, including growth suppres- N-terminal kinase ( JNK) and/or p38 mitogen-activated sion and apoptotic cell death (Azam et al. 2001; Zhang protein kinase (MAPK) signaling pathways (Takekawa et al. 2001; Vairapandi et al. 2002). and Saito 1998; Harkin et al. 1999). GADD45 proteins One of the well-described responses to genotoxic and physically interact with the MAPKKK, MTK1 (synonym nongenotoxic stresses is the rapid upregulation of dif- MEKK4), and the ensuing interactions result in the ferent GADD45 proteins, which in turn affect cell-cycle activation of MTK1. Activated MTK1 is thought to further regulation, cell survival, and cell death (Fornace et al. activate its downstream targets JNK and p38 (Takekawa and Saito 1998). It was shown that the N-terminal of MTK1 auto-inactivates its kinase activity and binding of 1Corresponding author: Department of Life Sciences, The National ita Institute for Biotechnology in the Negev, Ben-Gurion University, POB GADD45 proteins to MTK1 relieves this inhibition (M 653, Beer-Sheva, 84105 Israel. E-mail: [email protected] et al. 2002, Miyake et al. 2007). It was proposed that in

Genetics 177: 1691–1702 (November 2007) 1692 G. Peretz, A. Bakhrat and U. Abdu response to genotoxic stress, p53 is activated, which 1994, 1997), P[w1 lac-Z]BB142)(Schu¨pbach and Roth 1994), causes transcriptional upregulation of GADD45a,and and licH6 (Suzanne et al. 1999, kindly provided by S. Noselli). GADD45a interacts with MTK1 to initiate the JNK/p38- Flies were cultured in standard cornmeal/agar medium at 25°. The balancer chromosomes used in this study, as well as marker mediated apoptotic pathway. mutations have been described in (Lindsley and Zimm 1992) Several model systems have been used to analyze and in FlyBase (http://flybase.bio.indiana.edu). theroleofGADD45proteinsduringdevelopment. Transgenic flies: To make the pUASp-GADD45 construct GADD45a-null mice exhibit several phenotypes includ- the entire coding sequence of D-GADD45 (CG11086) was am- ing genomic instability, increased radiation carcinogen- plified from an EST template (RH70774) by PCR using mod- ollander ified primers to create a KpnI restriction site at the 59 end esis, and a low frequency of exencephaly (H (GGTACCATGGTCGTCGAGGAGAACTGC) and a NotI site at et al. 1999). GADD45g-deficient mice develop normally the 39 end (GCGGCCGCCTACACAGCCGGCAGC). The result- and are indistinguishable from their littermates, possi- ing PCR product was cut using KpnI and NotI and was cloned bly due to functional redundancy among the GADD45 into pUASp. P-element-mediated germline transformation of family members (Hoffmeyer et al. 2001). In the fish, this construct was carried out according to standard protocols (Spradling 1986). Five independent UAS-D-GADD45 lines Oryzias latipes, ectopic expression of GADD45g leads to were established. cell cycle arrest without inducing apoptosis. Loss of Preparation of the eggshell for dark-field microscopy: Freshly function of GADD45g causes a significant increase in laid eggs were collected on apple juice plates and placed in a apoptosis, suggesting that GADD45g is an important drop of Hoyer’s mount (Nusslein-Volhard and Wieschaus component of the molecular pathway that coordinates 1980). After an overnight incubation at 65° the slides were ex- amined by dark-field microscopy. cell cycle vs. apoptosis decisions during vertebrate devel- b-Galactosidase staining: Ovaries were dissected in PBS opment (Candal et al. 2004). The zebrafish GADD45b transferred to glutaraldehyde fixative and rocked for 8 min, genes were found to be periodically expressed as paired followed by washes (3 3 10 min) in PBST (PBS with 1% Triton stripes in the anterior presomitic mesoderm. Both knock- X-100). The ovaries were then placed in staining buffer ½0.05 m m down and overexpression of GADD45b genes caused K3(Fe(CN)6), 0.05 K4(Fe(CN)6) in PBS with X-gal stock solution (10% in N,N-dimethylformamide) added to a final somite defects with different consequences for marker concentration of 1% and stained in the dark, overnight at 37°. , indicating that the regulated expres- The next day the ovaries were washed (3 3 10 min) in PBST sion of GADD45b genes is required for somite segmen- and mounted on slides in 50% glycerol. tation (Kawahara et al. 2005). The possible functional In situ hybridization and antibody staining: RNA in situ redundancy among the GADD45 proteins in these model hybridization on ovaries was performed using digoxigenin- labeled probes (Roche, Indianapolis) according to standard systems makes the analysis of the molecular function of protocols (Roth and Schu¨pbach 1994). Antibody staining of GADD45 difficult. Annotation of the Drosophila melanogaster ovaries was carried out as described (Queenan et al. 1999), genome revealed that it contains only one GADD45-like using mouse anti-Grk ID12 1:10, rabbit anti-Oskar 1:3000 protein (CG11086, D-GADD45). (kindly provided by P. MacDonald), mouse anti-a-tubulin Since upregulation of GADD45 proteins may affect 1:200 (Sigma, St. Louis), rabbit anti b-Gal 1:200 (Promega), cell cycle regulation, cell survival, and cell death, we mouse anti-BR-C 1:100 (Hybridoma Bank, University of Iowa), mouse anti-Lamin 1:50 (Hybridoma Bank, University of Iowa). aimed to study the effect of D-GADD45 overexpression Secondary antibodies used were goat a-rabbit 1:500 (Molecu- on D. melanogaster oogenesis. We found that overexpres- lar Probes, Eugene, OR) and Cy3 goat a-mouse 1:100 ( Jackson sion of D-GADD45 in the somatic follicle cells led to ImmunoResearch, West Grove, PA). Oregon green 488 and apoptosis of the entire egg chamber. On the other hand, Alexa Fluor 546 phalloidin (Molecular Probes) were used at m overexpression of D-GADD45 in the germline did not 1:500 and Hoechst was used at 1 g/ml (Molecular Probes). Egg chambers were photographed on a Zeiss LSM510 Laser- cause apoptosis but affected the dorsal–ventral polarity scanning confocal microscope. of the eggshell. Moreover, D-GADD45 also affected Scanning electron microscopy: To prepare scanning elec- anterior–posterior polarity determinants. However, an- tron microscopy (SEM) samples, fly heads were dehydrated terior oocyte nuclear migration and bcd localization through an ethanol series (25, 50, 75, and twice with 100%) for were unaffected. Finally, we found that mutations in the 10 min each followed by a hexamethyldisilazane (HMDS, Elec- tron Microscopy Sciences) series (50, 75, and twice with 100%) MAPK–JNK pathway dominantly suppressed the egg for 2 hr each. Flies were air dried overnight under the hood, asymmetric defects in D-GADD45 overexpression ova- sputter coated with gold, mounted onto SEM stubs, and ana- ries, suggesting a novel, D-GADD45-mediated function lyzed with a JSM-5610LV scanning electron microscope. for the JNK pathway in the germline. Yeast culture: The entire coding sequence of CG11086 was amplified by PCR using modified primers to create a SpeI restriction site at the 59 end (ACTAGTATGGTCGTCGAGGA MATERIALS AND METHODS GAACTGC) and a SacI site at the 39 end (GAGCTCCTACA CAGCCGGCAGCTGG). The resulting PCR product was cut Drosophila strains: Oregon-R was used as wild-type control. using SpeI and SacI and was cloned as a GFP fusion in the The following transgenic and mutant flies were used: CY2- pHY315–GFP yeast expression vector (kindly provided by S. Gal4 (Queenan et al. 1997); GR1-Gal4 (kindly provided by Elledge). Transfection to yeast was carried out as described T. Schu¨pbach), nanos-Gal4-VP16 (Van Doren et al. 1998); Act5C- elsewhere (Adams et al. 1997). For DNA staining cells were GAL4, 32B-Gal4, GMR-GAL4, Tub-Gal4, hepG0208 and hepr75 incubated with DAPI at 1 mg/ml (Molecular Probes) for 30 min. (Bloomington Stock Center), kinesin b-GAL insertion line Cell culture: Drosophila Schneider 2 (S2) and S2R1 cells KZ503 and the Nod-b GAL insertion line NZ143.2 (Clark et al. were grown in Schneider cell medium (Biological Industries, Drosophila Egg Development and GADD45 1693

Beer-Sheva, Israel) supplemented with 10% fetal calf serum and periments to normalize for differences in total cDNA between penicillin–streptomycin–amphotericin (Biological Industries). samples. All quantitative PCR analyses were performed in Cells were cultured at 25°. For transfection, four million cells triplicate. were plated in 1 ml Schneider medium. Cells were transfected with 1 mg plasmid DNA using the ESCORT IV transfection re- agent (Sigma) according to the manufacturer’s protocol. RESULTS To make the UAS–GFP–GADD45 construct for the localiza- tion experiment the entire coding sequence of EGFP (Invitro- The CG11086 gene encodes the Drosophila gen) was amplified by PCR using modified primers to create a GADD45 homolog that is localized to the cell nucleus: KpnI restriction site at the 59 end (GGTACCATGGTGAGCA AGGGCGAGGAGC) and an XbaI site at the 39 end (TCTAG On the basis of BLAST results it was found that the ACTTGTACAGCTCGTCCATGCCG). The resulting PCR prod- D. melanogaster gene CG11086 is related to the mamma- uct was cloned into pBluescript. D-GADD45 (CG11086) was lian growth arrest and DNA damage inducible (GADD45) amplified by PCR using modified primers to create an XbaI gene family. This Drosophila GADD45-like protein restriction site at the 59 end (TCTAGAATGGTCGTCGAGGA (D-GADD45) shows strongest amino acid sequence iden- GAACTGC) and a NotI site at the 39 end (GCGGCCGCCTA CACAGCCGGCAGC). The resulting PCR product was cloned tity with human GADD45g (31%).D-GADD45corresponds into pBluescript-GFP. The sequence composed of D-GADD45 to a single exon with an ORF encoding a 163-amino acid and GFP was cut using KpnI and NotI and cloned into pUASp protein. Protein features include a ribosomal protein as a GFP fusion. For immunostaining, S2 cells were fixed with domain that starts at position 33 and ends at position 134 4% formaldehyde for 15 min. After permeabilization with (Koonin 1997). 0.2% Triton-X100, cells were incubated with mouse anti Lamin antibody 1:10 (Hybridoma Bank, University of Iowa) in 0.2% As a first step we wanted to assess whether D-GADD45 FSG. Cy3 goat a-mouse 1:1000 (Jackson ImmunoResearch) was localizes to the nucleus as its mammalian homologs used for detection. Cells were photographed on a Zeiss LSM510 (Azam et al. 2001; Zhang et al. 2001; Vairapandi et al. Laser-scanning confocal microscope. 2002). For this purpose the green fluorescent protein Stress experiments: To study the effect of stress on D-GADD45 (GFP) tag was fused to the N terminus of D-GADD45 transcription, different treatments were applied to S2 cells and Drosophila flies. For toxic metal stress S2 cells were left untreated and we tested the localization of the fused protein in two or received 200 mm NaAsO2 (Sigma–Aldrich) and were harvested cell systems, the yeast Saccharomyces cerevisiae and in the 2 hr later. For UVradiation, S2 cells were irradiated with UVC to a Drosophila embryo-derived Schneider S2 cell line. In dose of 300 Jm2 and were harvested 2 hr later. For paraquat both S2 and yeast, the protein showed a distinct local- treatment, Drosophila wild-type adult males, aged 3–4 days were ized distribution within the cell nucleus, demonstrating starved for 2 hr and then transferred to vials containing only filter paper soaked with 20 mm paraquat in 5% sucrose solution. conservation of this property (Figure 1, A–E). Control vials contained only 5% sucrose solution. RNA was ex- D-GADD45 is not upregulated following different tracted from flies 24 hr later. For X-ray treatment, Drosophila stress stimuli: One of the well described responses to wild-type adult males aged 3–4 days were treated with 2500 rads genotoxic and nongenotoxic stresses is the rapid upreg- in the Faxitron RX650 (Faxitron X-Ray, Wheeling, IL). RNA was ulation of GADD45 proteins, which in turn affect cell extracted form flies 2 hr later. Each assay was repeated twice cycle regulation, cell survival, and cell death (Fornace et al. more. astan Infection experiments: For septic injury experiments, we 1988, 1989; K et al. 1992). To examine D-GADD45 used Drosophila wild-type adult males, aged 3–4 days at 25°. gene expression in response to exposure to different Septic injury was performed by pricking the thorax of the flies stressful agents, we used real-time PCR. We tested the with a needle previously dipped into a concentrated bacterial response in both S2 cells and in the whole fly using a culture of the gram-negative strain Escherichia coli and gram- wide range of stressful conditions. positive strains Micrococcus luteus and Enterococcus faecalis (kindly provided by B. Lemaitre). Then flies were incubated In mice, GADD45a was identified as a gene whose at 25° and collected 2 hr after infection. expression is increased following arsenic-induced stress Real-time RT–PCR: Total RNA was extracted from flies and (Liu et al. 2001). To study the effect of this stress on S2 cells using the NucleoSpin RNA II kit, including DNase treat- D-GADD45 transcription, we treated S2 cells with the ment, according to the manufacturer’s instructions (Macherey- toxic metal, arsenite. We found that treatment did not Nagel). cDNA was transcribed in 1–3 mg of total RNA using reverse transcriptase and oligo(dT) (ABgene) in a 20-ml reac- lead to D-GADD45 induction (Figure 2A). GADD45a tion according to the manufacturer’s instructions (Bio-Lab, was identified as a gene whose expression is induced Jerusalem). Real-time PCR was performed using the Mx3000p following treatment with UV radiation (Fornace et al. (Stratagene, La Jolla, CA). Reverse-transcribed total RNA (200 ng) 1988). S2 cells were therefore also treated with UV m was amplified in a 20-ml reaction containing 200 n of each radiation. D-GADD45 mRNA levels in UV treated cells primer and 10 ml of SYBR Green PCR Master Mix (Stratagene). Forward and reverse primer sequences were as follows: Rp49 were similar to those found in the control (Figure 2A). Fwd, CCGCTTCAAGGGACAGTATCTG; Rp49 Rev, CACGT GADD45a was also found to be rapidly and strongly TGTGCACCAGGAACTT; GADD45 Rev, CGTAGATGTCGT induced in human lymphoblast cell lines following TCTCGTAGC; GADD45 Fwd, CATCAACGTGCTCTCCAAGTC; X-ray irradiation (Papathanasiou et al. 1991). For this Drosomycin Fwd, TCAAGTACTTGTTCGCCCTC; and Droso- reason, we treated flies with X rays, which causes DNA mycin Rev, ATCCTTCGCACCAGCACTTC. The amount of gene product in each sample was determined breaks. In this experiment we also found that the mRNA by the comparative quantification method using the MxPro level of D-GADD45 in irradiated flies was similar to that software (Stratagene). Ribosomal protein 49 was used in all ex- found in nonirradiated flies (Figure 2B). 1694 G. Peretz, A. Bakhrat and U. Abdu

Figure 1.—D-GADD45 is a nuclear protein. (A–C) Expression of D-GADD45 in Saccharomyces cerevisiae. (A) DNA staining; (B) GFP-tagged D-GADD45 expression; (C) merged pictures. (D and E) Expression in D. melanogaster embryo- derived Schneider S2 cells, Lamin in red. (D) GFP expression; (E) GFP-tagged D-GADD45 ex- pression. D-GADD45 expression in both yeast and S2 cells revealed that the protein is distinctly localized within the cell nucleus.

All three GADD45 isoforms displayed a significant D-GADD45 was identified as a gene whose expression is increase in expression in young mice following injection induced following microbial infection (De Gregorio of paraquat (Edwards et al. 2003). This is why we chose et al. 2001). In another study aimed to identify clusters to feed flies with paraquat and examine the level of of high-affinity Dorsal binding sites in the Drosophila D-GADD45 mRNA. We found that the mRNA levels of genome, a cluster of three optimal Dorsal binding sites D-GADD45 did not differ significantly between treated was identified 3 kb upstream to the D-GADD45 tran- and untreated extracts (Figure 2B). scription start site (Markstein et al. 2002). Dorsal is an D-GADD45 is upregulated upon activation of the NF-B-like transcription factor belonging to the Toll path- immune response: In a genomewide microarray analysis way that is activated mainly by gram-positive bacteria and fungi infections (Hoffmann and Reichhart 2002; Tzou et al. 2002). To strengthen the possible role for D-GADD45 in the immune response we conducted septic injury experi- ments in which wild-type flies were pricked in the thorax with a needle previously dipped into a concentrated gram-positive (Micrococcus luteus and Enterococcus faeca- lis) and gram-negative (Escherichia coli) bacterial culture. We analyzed the change in mRNA levels of D-GADD45 by real-time PCR. As a positive control we also checked the mRNA level of Drosomycin, which is a known AMP that is induced by septic injury. Following septic injury the level of D-GADD45 mRNA was significantly in- creased as compared with uninfected flies (Figure 2B). The mRNA level of our positive control, Drosomycin, also increased significantly, in accordance with previous results of others (data not shown). D-GADD45 overexpression in somatic cells causes apoptosis: To study the effect of overexpression of D-GADD45 in vivo we used the UAS-Gal4 binary system (Brand and Perrimon 1993) in which ubiquitous or tissue-specific expression can be induced. We cloned the Figure 2.—Quantitative RT–PCR on Schneider S2 cells and full length D. melanogaster GADD45-like protein (CG11086) D. melanogaster flies. Total RNA was isolated, and quantitative into the pUASp vector allowing expression in both so- RT–PCR was carried out to determine the expression of matic and germline tissues (Rorth 1998). Five transgenic D-GADD45. Data shown are the mean from three indepen- lines harboring the construct were generated. Several dent experiments. (A) S2 cells were untreated, treated with Gal4 driver lines were used to express D-GADD45 at arsenite, or irradiated with UVC. (B) Wild-type flies were un- treated, received X-ray treatment, were fed paraquat, or were various times and in different somatic tissues. We found pricked with a needle previously dipped in a mixture of bac- that the ubiquitous overexpression using Act and Tub teria, causing septic injury. Gal4 lines caused lethality in all the transgenic lines. Drosophila Egg Development and GADD45 1695

Figure 3.—D-GADD45 overexpression in the follicle cells causes apoptosis of the egg chamber D-GADD45 overexpression in the follicle cells and in the eye. (A–D) Hoechst DNA staining in blue. (A and C) Wild type and (B and D) D-GADD45 follicle cell overexpression using CY2-Gal4. Over- expression of D-GADD45 in the follicle cells caused apoptosis of the egg chamber, as seen from the typical condensation and breaking up of DNA content in both the nurse and follicle cells (B and D). (E and F) SEM images. (E) Control GMR-Gal4 line. (F) D-GADD45 eye overexpres- sion using GMR-Gal4. No gross morphological changes were observed.

Next, we studied the effects of overexpression of eggs that have two broadened dorsal appendages (Figure D-GADD45 in the somatic follicle cells surrounding the 4B), and moderately dorsalized with one broadened dorsal egg chamber using CY2-Gal4 (Figure 3, A–D) and GR1-Gal4 appendage on the dorsal midline (Figure 4C). This obser- (data not shown) drivers. Overexpression of D-GADD45 vation suggests that overexpression of D-GADD45 in the causes apoptosis of the egg chamber as revealed by DNA germline affects the dorsal–ventral (D/V) polarity of staining, which shows condensation and fragmentation the oocyte. Relatively lower percentages of the eggshell of nurse cell DNA followed by apoptosis of the entire egg showed an additional phenotype, where the eggs had a chamber (Figure 3D). This pattern of apoptosis of fol- more posteriorly located patch of dorsal appendage mate- licle cells that is followed by germline degeneration was rial (Figure 4D). This phenotype could arise from defects previously reported by Chao and Nagoshi (1999). Since in dorsal–ventral pattering but were harder to classify. We we found that overexpression of D-GADD45 in the so- analyzed the severity of dorsalized phenotypic class matic follicle cells led to apoptosis, we tested whether among the five different transgenic lines (Table 1). this is general to somatic cells or rather a tissue-specific D-GADD45 overexpression in germline affects grk response. For this purpose, D-GADD45 was overexpressed RNA localization: Since overexpression of D-GADD45 in other somatic tissues. No apoptosis was observed in in the germ cells affects the dorsal–ventral polarity of the somatic tissues such as the compound eye (Figure 3, E oocyte (Figure 4, Table 1) we decided to check whether and F) and wing imaginal discs (data not shown). the overexpression has an effect on factors involved in D-GADD45 overexpression in germline affects the the establishment of this embryonic axis. dorsal–ventral patterning of the eggshell: To assess the The anterior–posterior (A/P) and dorsal–ventral (D/V) potential role of D-GADD45 during D. melanogaster oo- axes of the D. melanogaster embryo are established du- genesis, we overexpressed D-GADD45 in the germline. ring oogenesis by a communication between the oocyte Overexpression in the germline using nanos-Gal4 pro- and the surrounding follicle cells (Schu¨pbach 1987; moter leads to the production of a broad range of Ruohola-Baker et al. 1994). This communication is me- abnormal eggshell phenotypes, ranging from normal diated by a single signaling system involving gurken (grk) morphology to dorsalized (Figure 3; Table 1). The egg- and Epidermal growth factor receptor (Egfr) genes (Gonza´lez- shell phenotypes vary from being wild-type like, bearing Reyes et al. 1995; Roth et al. 1995). grk encodes a TGF-a-like two dorsal appendages (Figure 4A), to weakly dorsalized protein that is expressed in the germline and is secreted 1696 G. Peretz, A. Bakhrat and U. Abdu

TABLE 1 Dorsalized eggshell phenotypes of different transgenic lines with D-GADD45 overexpression in the germline using nanos-GAL4

Wild-type-like Weakly Moderately Genotype eggshell (%) dorsalized (%) dorsalized (%) n D-GADD45#1/nanos-Gal4 9 26 65 118 D-GADD45#3/nanos-Gal4 14 35 51 115 D-GADD45#4/nanos-Gal4 3 3 94 154 D-GADD45#6/nanos-Gal4 10 37 53 102 D-GADD45#8/nanos-Gal4 41 17 42 145 Classification of eggshells on the basis of the dorsal–ventral polarity was performed according to the following criteria: wild type, two distinct dorsal appendages; weakly dorsalized eggshells, two broadened dorsal appen- dages; moderately dorsalized, with one broadened dorsal appendage on the dorsal midline. from the oocyte (Neuman-Silberberg and Schu¨pbach translated, we analyzed Grk protein expression by whole- 1993, 1996). Egfr is expressed throughout the follicular mount antibody staining. We did not observe egg cham- epithelium (Kammermeyer and Wadsworth 1987; Sapir bers where Grk was mislocalized (Figure 5D), possibly et al. 1998), where it acts as a receptor for Gurken. due to insensitivity of this assay. D-GADD45 overexpression in the germline affects D-GADD45 overexpression in germline affects Grk asymmetric development during oogenesis, with phe- activity: To gain further insight into the effect over- notypic consequences on D/V patterning of the egg- expression of D-GADD45 has on the grk–Egfr pathway we shell suggesting that it might affect the grk–Egfr pathway. tested grk activity using the kekon (kek)-lacZ enhancer Therefore we examined the localization of grk RNA and trap line and BR-C patterning. The kek gene is a target of protein. In wild-type ovaries, grk RNA accumulates in the the Egfr in the follicle cells and thus serves as an indirect oocyte during early stages of oogenesis. At mid-oogen- and sensitive assay to measure grk activity in the oocyte. esis, it is localized within the oocyte, first transiently in In wild-type egg chambers, kek is expressed in a char- an anterior–cortical ring, and then to a dorsal–anterior acteristic graded dorsal–anterior pattern reflecting both domain in the cytoplasm directly overlying the oocyte nu- the intensity and localization of the underlying grk sig- cleus (Figure 5A). Grk protein is restricted to the oocyte nal (Figure 6A; Musacchio and Perrimon 1996; Sapir and in mid-oogenesis, is localized to the dorsal–anterior et al. 1998; Suzanne et al. 1999). Overexpression of corner of the oocyte, similar to the RNA (Neuman- D-GADD45 disrupts the kek-lacZ graded pattern. In 72% Silberberg and Schu¨pbach 1996; Figure 5C). of the egg chambers kek signal is significantly expanded At stage 9, 57% of D-GADD45 overexpression egg throughout the follicle cells that span the oocyte, losing chambers display a mislocalized grk RNA pattern. grk its high point in the dorsal–anterior corner (Figure 6B). mRNA is detected as a circular ring along the anterior This result suggests an increase of grk activity consistent cortex (Figure 5B, arrows). Thus, it seems that over- with the dorsalization of the chorion. expression of D-GADD45 affects the proper localization An additional marker, Broad-Complex (BR-C), belongs of grk mRNA to the dorsal–anterior region of the oocyte. to a family of transcription factors that are expressed in To determine if the mislocalized RNA in these ovaries is the follicle cells in a dynamic pattern, the late pattern

Figure 4.—Eggshell dorsal–ventral de- fects in flies with D-GADD45 overexpression in the germline. D-GADD45 overexpres- sion in the germline using nanos-Gal4. Overexpression affected the dorsal–ventral polarity of the oocyte and led to the pro- duction of a broad range of abnormal egg- shell phenotypes, ranging from normal morphology to dorsalized: (A) eggs resem- bling wild-type eggs bearing two dorsal ap- pendages, (B) slightly dorsalized eggs with two short and broadened dorsal appen- dages, (C) eggshells with one broadened dorsal appendage on the dorsal midline, and (D) eggs that had a more posteriorly located patch of dorsal appendage material. This type of egg may also be due to a dorso–ventral problem of the oocyte, but is harder to classify. Drosophila Egg Development and GADD45 1697

Figure 5.—gurken mRNA is mislocalized in flies with D-GADD45 overexpression in the germ- line. grk mRNA and Grk protein localization in wild-type and in D-GADD45 overexpression ova- ries. (A and B) grk RNA in situ localization in stage-9 egg chambers. (A) Wild-type egg cham- ber. (B) D-GADD45 germline overexpression egg chamber. In most egg chambers grk RNA ab- normally accumulates as a circular ring around the anterior of the oocyte. (C and D) Grk protein expression in stage-9 egg chambers, Grk in red. (C) Wild-type egg chamber. (D) D-GADD45 over- expression egg chamber. No difference in Grk protein localization was observed.

being defined by two groups of dorsal–anterior follicle expression is detectable only in the dorsal appendage- cells at stage 10B of oogenesis. BR-C was also shown to associated follicle cells (Figure 6E, arrows). Once again be involved in dorsal appendage morphogenesis. The with germline D-GADD45 overexpression there is no dorsal–anterior expression pattern is specified by the detectable dorsal gap (Figure 6F), which may corre- Grk–Egfr and decapentaplegic (DPP) signaling pathways spond with the broadened, fused, dorsal appendages and can therefore serve as an additional marker for Grk– described previously. In fact, there is in general a very Egfr signaling (Deng and Bownes 1997). Using an anti- good correspondence between our observations of kek body for BR-C we can examine the distribution pattern or BR-C expansion and loss of dorsal gap during oo- of BR-C. In a wild-type stage-11 egg chamber two groups genesis and the high percentage of dorsalized eggshells of lateral–dorsal–anterior follicle cells are heavily stained produced by the females (Table 1). (Figure 6C, arrows), while the posterior and the ventral D-GADD45 overexpression in germline affects anterior– follicle cells are weakly stained. Expression of D-GADD45 posterior polarity determinants: The localization of bicoid disrupts the BR-C and leads to a range of patterns. Stage-11 (bcd)andoskar (osk) mRNA to the anterior and posterior egg chambers display a complete loss of the dorsal–anterior poles of the oocyte define the A/P axis of the embryo. osk gap between the two groups of the lateral–dorsal cells mRNA, at the posterior pole, directs assembly of the pole (Figure 6D). Eighty-seven percent of the egg chambers plasm, containing determinants of the abdomen and showed a laterally expanded staining (Figure 6D, bracket). germline (Ephrussi and Lehmann 1992; Reichmann In late stages of oogenesis, wild-type stage-13 egg chamber and Ephrussi 2001). Since osk is fundamental for oocyte

Figure 6.—D-GADD45 overexpression in germline affects Grk activity as seen by staining of the kek-lacZ enhancer trap line and BR-C anti- body staining. kek-lacZ enhancer trap line and BR-C antibody staining in wild-type and in D-GADD45 overexpression egg chambers. (A and B) b-Gal staining of the kek-lacZ enhancer trap line. (A) Wild-typeeggchamber;(B)D-GADD45overexpres- sion egg chamber. The normal dorsal–anterior graded expression of kek-lacZ is disrupted. In 72% of the egg chambers the kek signal is signifi- cantly expanded throughout the follicle cells that span the oocyte, losing its high point in the dorsal– anterior corner. (C–F) BR-C protein expression in egg chambers, BR-C in red. (C and E) Wild-type egg chamber, (D and F) D-GADD45 overexpres- sion egg chamber. In 87% of stage-11 egg cham- bers BR-C is laterally expanded with a complete loss of the dorsal–anterior gap. Stage-13 egg cham- bers also have no detectable dorsal gap. 1698 G. Peretz, A. Bakhrat and U. Abdu

Figure 7.—Posterior end markers are mislocal- ized in flies with D-GADD45 overexpression in the germline. osk mRNA, Osk protein, and Kin:b-gal localization in wild-type and in D-GADD45 overex- pression ovaries are shown. (A and B) osk RNA in situ localization in stage-10 egg chambers. (A) Wild-type egg chamber; (B) D-GADD45 overex- pression egg chamber. osk RNA transcript is mislo- calized and can be detected in a more diffuse pattern at the posterior pole. (C and D) Osk pro- tein expression in stage-10 egg chambers, Osk in green.(C)Wild-typeeggchamber;(D)D-GADD45 overexpression egg chamber. Osk is not restricted to a cortical crescent at the posterior pole similarly to the mislocalized RNA pattern. (E and F) Kin:b-gal localization in stage-9 egg chambers with Kin:b-gal in green. (E)Wild-type egg chamber;(F) D-GADD45 overexpression egg chamber. Kin:b-gal appears less compact and at a more central region of the oocyte.

polarization, we decided to check the localization of osk To analyze transport to the anterior pole, we used the RNA and protein in egg chambers of flies overexpressing microtubule minus-end marker Nod: b-gal (Clark et al. D-GADD45 in the germline. 1997). We found that Nod:b-gal localization in flies with In situ hybridization of D-GADD45 overexpression egg D-GADD45 overexpression in the germline is similar chambers revealed that the osk RNA transcript is mislo- to that observed in wild-type stage-9 egg chambers, the calized. In stage-9 egg chambers osk mRNA was found fusion protein is concentrated along the anterior cortex along the anterior cortex and in a more diffused pattern of the oocyte (data not shown). These results are con- at the posterior pole (data not shown). The majority of sistent with the proper localization of bcd mRNA as de- stage-10 egg chambers were also missing the tightly lo- scribedabove.Wewerealsoabletoobservenormal calized crescent (Figure 7B) seen in wild-type egg cham- oocyte nucleus migration to the anterodorsal position at bers (Figure 7A). Immunofluorescent staining of stage-10 stage 8 of oogenesis in these egg chambers (Gonza´lez- egg chambers with anti-Oskar antibody revealed that Osk Reyes et al. 1995; Roth et al. 1995; data not shown). protein was localized to the posterior pole of the oocyte Next, we asked whether the actin and/or the micro- but was not restricted to a cortical crescent (Figure 7D) tubule network are altered during oogenesis. To analyze as in wild-type egg chambers of this stage (Figure 7C). the organization of the microtubule network in egg Next we examined the localization of bcd mRNA, which chambers with D-GADD45 overexpression in the germ- is localized at the anterior pole and later on is translated line, we used a-tubulin staining. Using this tool, we were after fertilization to produce a morphogen gradient that unable to detect any defect in the microtubule network patterns the anterior region of the embryo (Driever 1993). (data not shown). Also, to examine the actin network, We found that bcd mRNA is correctly localized along the ovaries were stained with Rhodamine-conjugated phal- anterior cortex of egg chambers with D-GADD45 over- loidin; however, no obvious defects in the organization expression (data not shown). of the actin network and ring canal structure were found Since the localization of osk and grk RNA’s are de- in the transgenic flies’ ovaries. pendent on cytoskeleton organization in the oocyte, we D-GADD45 overexpression in germline affects the asked whether the actin and/or the microtubule net- MAPK pathway: In eukaryotes from yeasts to mammals, work are altered during oogenesis. To investigate the various cellular stresses generate intracellular signals polarity and transport to the posterior pole of the oo- that converge on the MAPK pathways (Mita et al. 2002). cyte, we used the plus-end microtubule-associated motor Three main MAPK families are the extracellular signal- protein, kinesin, fused to the bacterial b-galactosidase regulated kinase (ERK), p38, and JNK. Several studies enzyme (Kin:b-gal) as a reporter (Clark et al. 1994, 1997). have shown that the GADD45 proteins may play a role Immunofluorescence assay with anti-b-Gal demonstrated in the JNK and/or p38 MAPK signaling pathways a less compact staining at a more central region of the (Takekawaand Saito 1998; Harkin et al. 1999). The ge- oocyte in egg chambers with D-GADD45 overexpression nome of the fruit fly D. melanogaster possesses all MAPK in the germline (Figure 7F) as compared to wild-type families present in the mammalian genome, but repre- ovaries (Figure 7E). sented, typically, by fewer genes (Ryabinina et al. 2006). Drosophila Egg Development and GADD45 1699

TABLE 2 Suppression of the dorsalized eggshell phenotypes of D-GADD45 overexpression in a partially deficient JNK background

Wild-type-like Dorsalized Genotype eggshell (%) eggshell (%) n FM7; D-GADD45#1/nanos-Gal4 20 80 280 Df{hep,lic}; D-GADD45#1/nanos-Gal4 71 29 269 hepG0208; D-GADD45#1/nanos-Gal4 60 40 589 hepr75; D-GADD45#1/nanos-Gal4 81 19 439

Therefore, D. melanogaster makes an excellent, simple terning defects caused by D-GADD45 overexpression model system for studying signaling in evolutionary con- were dominantly suppressed (Table 2). The level of sup- served pathways. pression was similar to that of the heterozygous defi- Human GADD45 proteins were shown to physically ciency that uncovers both hep and lic, suggesting that interact with the MAPKKK, MTK1 (synonym MEKK4), hep is the main D-GADD45 activator in the germline. resulting in the activation of MTK1 which further ac- tivates both JNK and p38 (Takekawa and Saito 1998). DISCUSSION Since GADD45 proteins were shown to activate the JNK and/or p38 MAPK signaling pathways, we searched for a In this article we characterized the gene CG11086, genetic interaction between the D-GADD45 and p38/ which has been identified as the D. melanogaster GADD45 JNK kinases. It was previously shown that flies mutant homolog. We showed that D-GADD45 preserves the nu- for licorne (lic, synonym D-MKK3), the D. melanogaster clear localization property, but unlike its mammalian p38 MAPKK, display a ventralized eggshell phenotype homologs its expression is not elevated following expo- (Suzanne et al. 1999). Loss of function of lic in the sure to different stress stimuli. This result is supported germline leads to reduced Egfr activity in the follicle by the D. melanogaster whole genome microarray analysis cells due to a reduction of Grk protein in the oocyte. Lic which did not identify D-GADD45 as a gene whose ex- is also required for maintenance of osk mRNA localiza- pression is increased following various genotoxic and tion during midoogenesis. In stage-9 lic germ-line clones nongenotoxic treatments (Brodsky et al. 2004; Girardot osk mRNA is mislocalized, diffusing through the whole et al. 2004; Qin et al. 2005; Terashima and Bownes 2005). oocyte in a gradient from the posterior to the anterior Although we tried a number of stress treatments, it is pos- pole. Suzanne et al. (1999) proposed that the patterning sible that D-GADD45 expression would rise only following of the egg relies on activation of the p38 MAPK pathway. exposure to other stressful conditions. We hypothesized that D-GADD45 overexpression, which D-GADD45 was identified as a gene whose expression caused dorsalization of the eggshell, could be an effect is induced following microbial infection (De Gregorio mediated through interactions of D-GADD45 with the et al. 2001). It was also shown that D-GADD45 expression MAPK protein(s). may be regulated by the NF-B-like transcription factor, To elucidate the nature of D-GADD45 regulation of Dorsal, which has an optimal binding site 3 kb upstream the MAPK pathway, we generated flies with D-GADD45 to D-GADD45 transcription start site (Markstein et al. overexpression in a partially deficient p38 and JNK 2002). Our results are consistent with those found by De background. The fly stock used was doubly mutant for Gregorio et al (2001) and further strengthen a possible both lic and hemipterous (hep), the D. melanogaster JNKK function for D-GADD45 in the immune response. Given gene (Suzanne et al. 1999). We found that reducing the that Drosophila is devoid of an adaptive immune system level of both lic and hep significantly suppressed the dor- and relies only on innate immune reactions for its de- salized eggshell phenotype (Table 2). Next, we analyzed fense, D-GADD45 may play an important role during whether reducing the level of p38 and JNK kinases also infection. affects the mislocalization of Osk protein in D-GADD45 We found that ubiquitous overexpression of D-GADD45 overexpression ovaries. We found that the defects in Osk was lethal, most likely due to apoptosis, as we directly protein localization were also considerably reduced from demonstrated in the follicle cells. However, our results 84 to only 33% in these flies. suggest that apoptosis induced by overexpression of Next we intended to pinpoint the dominant suppres- D-GADD45 is tissue specific since overexpression of sion to either the p38 or the JNK pathway. Since there D-GADD45 in other somatic tissues, such as the eye and are no flies singly mutant for lic, we examined the effect wing, did not lead to apoptosis. Also, overexpression in of D-GADD45 overexpression in a single hep deficient the germline did not cause cell death; rather, it affected background. To our surprise we found, using two dif- egg chamber asymmetric development. The apparent ferent hep mutant alleles, that the dorsal–ventral pat- phenotypic differences in overexpression of D-GADD45 1700 G. Peretz, A. Bakhrat and U. Abdu in the germline as opposed to somatic derived tissues the JNKK, hemipterous, dominantly suppressed the dor- probably reflect the complexity of the biological func- salized eggshell phenotype. This genetic interaction is tions of GADD45, which may be subject to tissue- and/or supported by the finding that in human cells GADD45 signal-specific regulation that ultimately dictate their proteins act as initiators of JNK/p38 signaling via their output. Similarly, it has been shown that individual mem- interaction with the MAPKKK, MTK1 (Takekawa and bers of the GADD45 family play critical roles in negative Saito 1998). It was shown that the N-terminal of MTK1 growth control in some tissues while in others they are inhibits its C-terminal kinase domain by preventing the associated with uncontrolled cell growth and tumor de- kinase domain from interacting with its substrate, MKK6, velopment. GADD45a was identified as an important and binding of GADD45 proteins relieves this auto- mediator of tumor suppression in human ovarian can- inhibition (Mita et al. 2002; Miyake et al. 2007). cer cells ( Jiang et al. 2003). While in pancreatic ductal Up until now the only roles attributed to the JNK adenocarcinoma GADD45a was found to be overex- pathway during oogenesis were in the follicle cells and pressed at the mRNA and protein level. Downregulation included morphogenesis of the dorsal appendages and of GADD45a by means of RNAi reduced proliferation the micropyle (Suzanne et al. 2001). It was also reported and induced apoptosis in pancreatic cancer cells imply- that the JNK pathway is involved in the morphogenetic ing that GADD45a contributes to pancreatic cancer cell process of dorsal closure during embryogenesis (Ip and proliferation and viability (Schneider et al. 2006). Davis 1998; Noselli 1998; Agne`s et al. 1999; Noselli We found that overexpression of D-GADD45 in the and Agne`s 1999). To our surprise, we found that egg- germline resulted in dorsalization of the chorion due shell patterning defects caused by D-GADD45 overex- to defects in grk localization and translation. We also pression are dominantly suppressed in a hep deficient showed that the posterior markers osk and Kin:b-gal background suggesting an additional role for the JNK were mislocalized during mid-oogenesis. On the other pathway in the germline. This novel function may have hand, D-GADD45 overexpression did not affect the lo- gone unnoticed in the past while studying JNK loss-of- calization of anterior end markers such as bcd and Nod: function alleles due to redundancy with some other b-gal and also the anterior oocyte nuclear migration is pathway. In our study overexpression of the JNK activator, unaffected. Similar results were reported in mutants of D-GADD45, may have unmasked this new role during squid (sqd) which encodes a heterogeneous nuclear ri- oogenesis. bonucleoprotein (hnRNP). In these mutants grk mRNA We thank Paul MacDonald, Ste´phane Noselli, Stephen Elledge, is mislocalized along the anterior ring, leading to dorsal- Bruno Lemaitre, Trudi Schu¨pbach, and the Bloomington stock center ization of the eggshell (Kelley 1993; Neuman-Silberberg for generously providing fly strains and reagents. We also thank Trudi and Schu¨pbach 1993, 1996; Matunis et al.1994).Further- Schu¨pbach for comments on the manuscript and Natalie Denef for the more, loss of sqd function causes an aberrant localization of help with the transgenic flies. This work was supported by the Research Career Development Award from the Israel Cancer Research Fund osk and Kin:b-gal, but does not affect bcd localization and to U.A. oocyte nucleus migration (Norvell et al. 2005; Steinhauer and Kalderon 2005). It was shown that in sqd mutant oocytes short microtubules (MTs) around the entire oo- cyte cortex are retained, including at the posterior pole, LITERATURE CITED unlike wild-type MTs which emanate mostly from the Abdollahi, A., K. A. Lord,B.Hoffmann-Liebermann and anterior. It was suggested that the primary MT defect in D. Liebermann, 1991 Sequence and expression of a cDNA en- sqd mutants is the failure to eliminate cortical sites of MT coding MyD118: a novel myeloid differentiation primary response teinhauer alderon gene induced by multiple cytokines. Oncogene 6: 165–167. nucleation beyond stage 7 (S and K Adams, A., D. E. Gottshling,C.A.Kaiser and T. Stearns, 2005). It is possible that D-GADD45 overexpression also 1997 Methods in Yeast Genetics, Cold Spring Harbor Laboratory affects MT organization in the oocyte. This possibility is Press, Cold Spring Harbor, NY. Agne`s, F., M. Suzanne and S. Noselli, 1999 The Drosophila JNK further supported by the finding that GADD45a interacts pathway controls the morphogenesis of imaginal disc during with elongation factor 1a (EF-1a), a microtubule-severing metamorphosis. Development 126: 5453–5462. protein that plays an important role in maintaining mi- Azam,N.,M.Vairapandi,W.Zhang,B.Hoffman and D. A. Liebermann, ong 2001 Interactionof CR6(GADD45g) with proliferatingcell nuclear crotubule cytoskeletal stability (T et al. 2005). To test antigen impedes negative growth control. J. Biol. Chem. 276: 2766– whether D-GADD45 affects MTorganization, we stained 2774. the ovaries with anti-tubulin. Using this tool, we were Beadling, C., K. W. Johnson and K. A. Smith, 1993 Isolation of in- terleukin 2-induced immediate-early genes. Proc. Natl. Acad. Sci. unable to detect any gross morphological changes in the USA 90: 2719–2723. MT network. Given that the patterning defect seen in Brand, A. H., and N. Perrimon, 1993 Targeted gene expression as a D-GADD45 overexpression is weaker than that in sqd means of altering cell fates and generating dominant pheno- types. Development 118: 401–415. mutants, it could be that this kind of staining is not sen- Brodsky, M. H., B. T. Weinert,G.Tsang,Y.S.Rong,N.M.McGinnis sitive enough to identify the MT network alterations in et al., 2004 Drosophila melanogaster MNK/Chk2 and p53 regulate D-GADD45 overexpression flies. multiple DNA repair and apoptotic pathways following DNA dam- age. Mol. Cell. Biol. 24(3): 1219–1231. We found a genetic interaction between D-GADD45 Candal, E., V. Thermes,J.S.Joly and F. Bourrat, 2004 Medaka as and proteins of the MAPK–JNK pathway. Mutations in a model system for the characterisation of cell cycle regulators: a Drosophila Egg Development and GADD45 1701

functional analysis of Ol-Gadd45g during early embryogenesis. Kearsey, J. M., P. J. Coates,A.R.Prescott,E.Warbrick and Mech. Dev. 121: 945–958. P. A. Hall, 1995 Gadd45 is a nuclear cell cycle regulated pro- Chao, S. H., and R. N. Nagoshi, 1999 Induction of apoptosis in the tein which interacts with p21Cip1. Oncogene 9: 1675–1683. germline and follicle layer of Drosophila egg chambers. Mech. Kelley, R. L., 1993 Initial organization of the Drosophila dorsoven- Dev. 88: 159–172. tral axis depends on an RNA-binding protein encoded by the Clark, I., E. Giniger,H.Ruohola-Baker,L.Y.Jan and Y. N. Jan, squid gene. Genes Dev. 7(6): 948–960. 1994 Transient posterior localization of a kinesin fusion pro- Kelman, Z., and J. Hurwitz, 1998 Protein-PCNA interactions: a tein reflects anteroposterior polarity of the Drosophila oocyte. DNA-scanning mechanism? Trends Biochem. Sci. 23: 236–238. Curr. Biol. 4: 289–300. Koonin, E. V., 1997 Cell cycle and apoptosis: possible roles of Clark, I. E., L. Y. Jan and Y. N. Jan, 1997 Reciprocal localization of Gadd45 and MyD118 proteins inferred from their homology to Nod and kinesin fusion proteins indicates microtubule polarity ribosomal proteins. J. Mol. Med. 75: 236–238. in the Drosophila oocyte, epithelium, neuron and muscle. Devel- Lindsley, D. L., and G. G. Zimm, 1992 The Genome of Drosophila mel- opment 124: 461–470. anogaster. Academic Press, New York. De Gregorio, E., P. T. Spellman,G.M.Rubin and B. Lemaitre, Liu, J., M. B. Kadiiska,Y.Liu,T.Lu,W.Qu et al., 2001 Stress-related 2001 Genome-wide analysis of the Drosophila immune re- gene expression in mice treated with inorganic arsenicals. Toxicol. sponse by using oligonucleotide microarrays. Proc. Natl. Acad. Sci. 61(2): 314–320. Sci. USA 98(22): 12590–12595. Markstein, M., P. Markstein,V.Markstein and M. S. Levine, Deng, W. M., and M. Bownes, 1997 Two signalling pathways specify 2002 Genome-wide analysis of clustered Dorsal binding sites localised expression of the Broad-Complex in Drosophila egg- identifies putative target genes in the Drosophila embryo. Proc. shell patterning and morphogenesis. Development 124(22): Natl. Acad. Sci. USA 99: 763–768. 4639–4647. Matunis, E. L., R. Kelley and G. Dreyfuss, 1994 Essential role for Driever, W., 1993 Maternal control of anterior development in the a heterogeneous nuclear ribonucleoprotein (hnRNP) in oogen- Drosophila embryo, pp. 301–324 in The Development of Drosophila esis: hrp40 is absent from the germ line in the dorsoventral mu- melanogaster, Vol. I, edited by M. Bate and A. Martinez-Arias. tant squid. Proc. Natl. Acad. Sci. USA 91(7): 2781–2784. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Mita, H., J. Tsutsui,M.Takekawa,E.A.Witten and H. Saito, Edwards, M. G., D. Sarkar,R.Klopp,J.D.Morrow,R.Weindruch 2002 Regulation of MTK1/MEKK4 kinase activity by its N-terminal et al., 2003 Age-related impairment of the transcriptional re- autoinhibitory domain and GADD45 binding. Mol. Cell. Biol. 22: sponses to oxidative stress in the mouse heart. Physiol. Genomics 4544–4555. 13: 119–127. Miyake, Z., M. Takekawa,Q.Ge and H. Saito, 2007 Activation of Ephrussi, A., and R. Lehmann, 1992 Induction of germ cell forma- MTK1/MEKK4 by GADD45 through induced N-C dissociation tion by oskar. Nature 358: 387–392. and dimerization-mediated trans autophosphorylation of the Fornace,Jr., A. J., D. W. Nebert,M.C.Hollander,J.D.Luethy, MTK1 kinase domain. Mol. Cell. Biol. 27(7): 2765–2776. M. Papathanasiou et al., 1989 Mammalian genes coordinately Musacchio, M., and N. Perrimon, 1996 The Drosophila kekkon regulated by growth arrest signals and DNA-damaging agents. genes: novel members of both the leucine-rich repeat and immu- Mol. Cell. Biol. 9: 4196–4203. noglobulin superfamilies expressed in the CNS. Dev. Biol. 178: Fornace,Jr., A. J., I. Alamo,Jr. and M. C. Hollander, 1988 DNA 63–76. damage inducible transcripts in mammalian cells. Proc. Natl. Neuman-Silberberg,F.S.,andT.Schu¨pbach, 1993 The Drosophila Acad. Sci. USA 85: 8800–8804. dorsoventral patterning gene gurken produces a dorsally localized Girardot, F., V. Monnier and H. Tricoire, 2004 Genome wide RNA and encodes a TGF alpha-like protein. Cell 75: 165–174. analysis of common and specific stress responses in adult Drosoph- Neuman-Silberberg,F.S.,andT.Schu¨pbach, 1996 The Drosophila ila melanogaster. BMC Genomics 5(1): 74. TGF-alpha-like protein Gurken: expression and cellular localiza- Gonza´lez-Reyes, A., H. Elliott and D. St Johnston, 1995 Po- tion during Drosophila oogenesis. Mech. Dev. 59: 105–113. larization of both major body axes in Drosophila by gurken- Norvell, A., A. Debec,D.Finch,L.Gibson and B. Thoma, torpedo signalling. Nature 375: 654–658. 2005 Squid is required for efficient posterior localization of os- Harkin, D. P., J. M. Bean,D.Miklos,Y.H.Song,V.B.Truong et al., kar mRNA during Drosophila oogenesis. Dev. Genes Evol. 215(7): 1999 Induction of GADD45 and JNK/SAPK-dependent apopto- 340–349. sis following inducible expression of BRCA1. Cell 97: 575–586. Noselli, S., 1998 JNK signaling and morphogenesis in Drosophila. Hoffmann, J. A., and J. M. Reichhart, 2002 Drosophila innate im- Trends Genet. 14: 33–38. munity: an evolutionary perspective. Nat. Immunol. 3: 121–126. Noselli, S., and F. Agne`s, 1999 Roles of the JNK signaling pathway Hoffmeyer, A., R. Piekorz,R.Moriggl and J. N. Ihle, in Drosophila morphogenesis. Curr. Opin. Genet. Dev. 9: 466–472. 2001 Gadd45gamma is dispensable for normal mouse develop- Nusslein-Volhard, C., and E. Wieschaus, 1980 Mutations affect- ment and T-cell proliferation. Mol. Cell. Biol. 9: 3137–3143. ing segment number and polarity in Drosophila. Nature 287(5785): Hollander, M. C., M. S. Sheikh,D.V.Bulavin,K.Lundgren, 795–801. L. Augeri-Henmueller et al., 1999 Genomic instability in Papathanasiou, M. A., N. C. Kerr,J.H.Robbins,O.W.McBride, Gadd45a-deficient mice. Nat. Genet. 23: 176–184. I. Alamo,Jr. et al., 1991 Induction by ionizing radiation of Ip, Y. T., and R. J. Davis, 1998 Signal transduction by the c-Jun the gadd45 gene in cultured human cells: lack of mediation by N-terminal kinase ( JNK)–from inflammation to development. protein kinase C. Mol. Cell. Biol. 11(2): 1009–1016. Curr. Opin. Cell Biol. 10: 205–219. Qin, W., S. J. Neal,R.M.Robertson,J.T.Westwood and V. K. Jiang, F., P. Li,A.J.Fornace,Jr., S. V. Nicosia and W. Bai, Walker, 2005 Cold hardening and transcriptional change in 2003 G2/M arrest by 1,25-dihydroxyvitamin D3 in ovarian can- Drosophila melanogaster. Insect Mol. Biol. M(6): 607–613. cer cells mediated through the induction of GADD45 via an ex- Queenan, A. M., A. Ghabrial and T. Schu¨pbach, 1997 Ectopic ac- onic enhancer. J. Biol. Chem. 278: 48030–48040. tivation of torpedo/Egfr, a Drosophila receptor tyrosine kinase, Jin, S., T. Tong,W.Fan,F.Fan,M.J.Antinore et al., 2002 GADD45- dorsalizes both the eggshell and the embryo. Development induced cell cycle G2-M arrest associates with altered subcellular 124(19): 3871–3880. distribution of cyclin B1 and is independent of p38 kinase activ- Queenan, A. M., G. Barcelo,C.Van Buskirk and T. Schu¨pbach, ity. Oncogene 21: 8696–8704. 1999 The transmembrane region of Gurken is not required Kammermeyer, K. L., and S. C. Wadsworth, 1987 Expression of for biological activity, but is necessary for transport to the oocyte Drosophila epidermal growth factor receptor homologue in mi- membrane in Drosophila. Mech. Dev. 89: 35–42. totic cell populations. Development 100: 201–210. Reichmann, V., and A. Ephrussi, 2001 Axis formation during Dro- Kastan, M. B., Q. Zhan,W.S.El-Deiry,F.Carrier,T.Jacks et al., sophila oogenesis. Curr. Opin. Genet. Dev. 11: 374–383. 1992 A mammalian cell cycle checkpoint pathway utilizing p53 Rorth, P., 1998 Gal4 in the Drosophila female germline. Mech. and GADD45 is defective in ataxia-telangiectasia. Cell 71: 587–597. Dev. 78: 113–118. Kawahara, A., Y. S. Che,R.Hanaoka,H.Takeda and I. B. Dawid, Roth, S., and T. Schu¨pbach, 1994 The relationship between ovar- 2005 Zebrafish GADD45b genes are involved in somite segmen- ian and embryonic dorsoventral patterning in Drosophila. Devel- tation. Proc. Natl. Acad. Sci. USA 102(2): 361–366. opment 120(8): 2245–2257. 1702 G. Peretz, A. Bakhrat and U. Abdu

Roth, S., F. S. Neuman-Silberberg,G.Barcelo and T. Schu¨pbach, oogenesis for egg asymmetric development. Genes Dev. 13: 1995 Cornichon and the EGF receptor signaling process are 1464–1474. necessary for both anterior-posterior and dorsal-ventral pattern Suzanne, M., N. Perrimon and S. Noselli, 2001 The Drosophila JNK formation in Drosophila. Cell 81: 967–978. pathway controls the morphogenesis of the egg dorsal appen- Ruohola-Baker, H., L. Y. Jan and Y. N. Jan, 1994 The role of gene dages and micropyle. Dev. Biol. 237: 282–294. cassettes in axis formation during Drosophila oogenesis. Trends Takekawa, M., and H. Saito, 1998 A family of stress-inducible Genet. 10: 89–94. GADD45-like proteins mediate activation of the stress-responsive Ryabinina, O. P., E. Subbian and M. S. Iordanov, 2006 D-MEKK1, MTK1/MEKK4 MAPKKK. Cell 95: 521–530. the Drosophila orthologue of mammalian MEKK4/MTK1, and Terashima, J., and M. A. Bownes, 2005 A microarray analysis of Hemipterous/D-MKK7 mediate the activation of D-JNK by cad- genes involved in relating egg production to nutritional intake mium and arsenite in Schneider cells. BMC Cell Biol. 7: 7. in Drosophila melanogaster. Cell Death Differ. 12(5): 429–440. Sapir, A., R. Schweitzer and B. Z. Shilo, 1998 Sequential activa- Tong, T., J. Ji,S.Jin,X.Li,W.Fan et al., 2005 Gadd45a expression tion of the EGF receptor pathway during Drosophila oogenesis induces Bim dissociation from the cytoskeleton and transloca- establishes the dorsoventral axis. Development 125: 191–200. tion to mitochondria. Mol. Cell. Biol. 25(11): 4488–4500. Schneider, G., A. Weber,U.Zechner,F.Oswald,H.M.Friess et al., Tzou, P., E. De Gregorio and B. Lemaitre, 2002 How Drosophila 2006 GADD45a is highly expressed in pancreatic ductal adeno- combats microbial infection: a model to study innate immunity carcinoma cells and required for tumor cell viability. Int. J. Can- and host–pathogen interactions. Curr. Opin. Microbiol. 5: 102–110. cer 118: 2405–2411. Vairapandi, M., A. G. Balliet,B.Hoffman and D. A. Liebermann, Schu¨pbach, T., 1987 Germ line and soma cooperate during oogen- 2002 GADD45b and GADD45g Are cdc2/CyclinB1 kinase in- esis to establish the dorsoventral pattern of egg shell and embryo hibitors with a role in S and G2/M cell cycle checkpoints induced in Drosophila melanogaster. Cell 49(5): 699–707. by genotoxic stress. J. Cell. Physiol. 192: 327–338. Schu¨pbach, T., and S. Roth, 1994 Dorsoventral patterning in Dro- Van Doren, M., A. L. Williamson and R. Lehmann, 1998 Regula- sophila oogenesis. Curr. Opin. Genet. Dev. 4(4): 502–507. tion of zygotic gene expression in Drosophila primordial germ Sheikh, M. S., M. C. Hollander and A. J. Fornace,Jr., 2000 Role cells. Curr. Biol. 8: 243–246. of Gadd45 in Apoptosis. Biochem. Pharmacol. 59: 43–45. Zhan, Q., M. J. Antinore,X.W.Wang,F.Carrier,M.L.Smith et al., Spradling, A. C., 1986 P element-mediated transformation, 1999 Association with Cdc2 and inhibition of Cdc2/Cyclin B1 pp. 175–197 in Drosophila—A Practical Approach, edited by D. B. kinase activity by the p53-regulated protein Gadd45. Oncogene Roberts. IRL, Oxford. 18: 2892–2900. Steinhauer, J., and D. Kalderon, 2005 The RNA-binding pro- Zhang, W., B. Hoffman and D. A. Liebermann, 2001 Ectopic ex- tein Squid is required for the establishment of anteroposterior pression of MyD118/Gadd45/CR6 (Gadd45beta/alpha/gamma) polarity in the Drosophila oocyte. Development 132(24): 5515– sensitizes neoplastic cells to genotoxic stress-induced apoptosis. 5525. Int. J. Oncol. 18: 749–757. Suzanne, M., I. Kenji,G.Bruno,A.Francois,M.Eiji et al., 1999 The Drosophila p38 MAPK pathway is required during Communicating editor: T. C. Kaufman