|Home | About | Journals | Submit | Contact Us | Français|
Regulated adhesion between cells and their environment is critical for normal cell migration. We have identified mutations in a gene encoding the Drosophila hydrogen peroxide (H2O2)-degrading enzyme Jafrac1, which lead to germ cell adhesion defects. During gastrulation, primordial germ cells (PGCs) associate tightly with the invaginating midgut primordium as it enters the embryo; however in embryos from jafrac1 mutant mothers this association is disrupted, leaving some PGCs trailing on the outside of the embryo. We observed similar phenotypes in embryos from DE-Cadherin/shotgun (shg) mutant mothers and were able to rescue the jafrac1 phenotype by increasing DE-Cadherin levels. This and our biochemical evidence strongly suggest that Jafrac1-mediated reduction of H2O2 is required to maintain DE-Cadherin protein levels in the early embryo. Our results present in vivo evidence of a peroxiredoxin regulating DE-Cadherin mediated adhesion.
During development, embryos carefully manage cell adhesion to permit the cellular movements required to achieve morphogenesis, and in adults, disseminating tumor cells can reactivate these same developmental mechanisms during cancer metastasis. Extensive studies have shown that the homophilic adhesion molecule, E-Cadherin, serves as a key modulator of cell adhesion and migration during tumor metastasis and epithelial to mesenchymal transitions (EMTs) (Thiery and Sleeman, 2006). A large body of work suggests that E-Cadherin regulation is essential for cell reorganization and migration during tumor spreading, and indicates the importance of understanding how E-Cadherin levels are controlled. E-Cadherin is regulated both at the transcriptional and post-transcriptional level. The conserved transcriptional repressor, Twist, can repress E-Cadherin, facilitating metastasis (Yang et al., 2004). E-Cadherin can also be regulated post-transcriptionally by phosphorylation and endocytosis (Fujita et al., 2002; Palacios et al., 2005). In cell culture, E-Cadherin and β-Catenin relocalization can be triggered by oxidants, through the action of tyrosine kinases (Rao et al., 2002). Yet how oxidants affect E-Cadherin localization or stability in vivo is unknown.
Dynamic regulation of DE-Cadherin and cell adhesion is an essential aspect in the control of PGC behavior in Drosophila (Kunwar et al., 2008). Furthermore, PGC migration provides an excellent in vivo model to study regulated adhesion independently of transcription since early germ cells are transcriptionally silent (Hanyu-Nakamura et al., 2008; Martinho et al., 2004). PGCs form at the posterior pole of the embryo directly abutting the future posterior midgut primordium. As the midgut internalizes during gastrulation, PGCs are carried along into the embryo. Live imaging suggests that PGCs undergo a striking transition in their adhesive behavior during these early stages. Upon formation, PGCs display aspects of active motility; subsequently during gastrulation, they pack into a tight monolayer cluster and adhere closely to the invaginating midgut. Once inside the embryo, however, at the onset of active migration, DE-Cadherin and other adherens junction (AJ) components localize to the lagging tail of PGCs. This reorganization of DE-Cadherin facilitates loss of PGC adhesion and promotes migration of individualized PGCs through the midgut epithelium (Kunwar et al., 2008).
In a genetic analysis of germ cell expressed genes in Drosophila, we found that mutations in the jafrac1 gene cause an early PGC adhesion defect. Jafrac1 is a member of the antioxidant peroxiredoxin family, which catalyzes the reduction of H2O2 and alkyl hydroperoxides through the oxidation and subsequent reduction of catalytic cysteine residues (Chae et al., 1994a; Chae et al., 1993; Chae et al., 1994b). In addition to functioning as antioxidants, it has recently been discovered that peroxiredoxins also have chaperone activity and act as redox sensors that regulate gene expression (Karplus and Hall, 2007; Veal et al., 2007). Analysis of the C. elegans peroxiredoxin, PRDX-2, supports its conserved role as both an antioxidant and chaperone protein in multicellular organisms (Olahova et al., 2008). Null mutations in the mouse peroxiredoxin, Prdx1, result in decreased viability because of a reduction in erythrocytes and an increase in lymphomas, carcinomas, and sarcomas (Neumann et al., 2003). Increased tumor incidence is even seen in Prdx1 +/− mice. Prdx2 null mice are subject to hemolytic anemia, but an increase in tumor formation was not reported (Lee et al., 2003). Demonstrating a role in signaling, Prdx2 has been shown to negatively regulate platelet-derived growth factor (PDGF) (Choi et al., 2005). The peroxidase activity of Jafrac1, an ortholog of Prdx2, is functionally conserved in Drosophila (Bauer et al., 2002; Lee et al., 2009; Radyuk et al., 2001; Radyuk et al., 2003; Rodriguez et al., 2000) but its in vivo role has only begun to be elucidated.
Here, we present evidence that a peroxiredoxin regulates cell adhesion. During gastrulation, PGCs form a tight cluster and are rapidly internalized by the movements of the underlying soma (Kunwar et al., 2008). jafrac1 mutant PGCs can lose adherence with the midgut during gastrulation and be left outside of the midgut. Live imaging reveals that jafrac1 mutant PGCs fail to properly associate with each other as gastrulation initiates. We show that PGC internalization is a DE-Cadherin dependent adhesion process that depends on the regulation of AJ components by H2O2 and Jafrac1.
To identify new genes important to germ cell formation and function, we tested the function of genes whose RNA is present in early germ cells. jafrac1 mRNA is maternally deposited in the embryo and is protected from degradation in germ cells until embryonic stage 9 (Figure 1A–C). jafrac1 mRNA is not detected in germ cells after this stage, but is later expressed in crystal cells, a special subset of Drosophila immune cells (Figure 1D). To analyze the distribution of the Jafrac1 protein, we generated a specific antibody (Figure S1A, B), which detected Jafrac1 in both PGCs and somatic cells throughout the embryo (Figure 1E, E’, E’’), in the cytoplasm (Figure 1F). The expression pattern is consistent with jafrac1 being a maternal effect gene that acts in germ cells.
Previous studies assumed that the first regulated step of germ cell migration occurs at embryonic stage 9 when germ cells migrate through the midgut epithelium (Moore et al., 1998). Our work identifies an earlier point of regulation during stage 7 when germ cells are carried inside the embryo during gastrulation movements. Two independent transposon insertions in the jafrac1 locus (Figure 1G) cause a maternal effect defect; germ cells in embryos from homozygous mutant mothers are left outside of the midgut during gastrulation (Figure 1H, I). Such a phenotype was rarely observed in wild type embryos (4% of embryos). However, 73% of the embryos from mothers homozygous for the P element allele jafrac1KG05372, hereafter referred to as jafrac1KG, displayed this phenotype, with an average of 6 germ cells left outside the midgut per embryo (Figure 1J). 25% of embryos from mothers homozygous for the piggyBac allele jafrac1F08066, hereafter referred to as jafrac1F0, failed to enter the midgut and showed a PGC adherence phenotype with an average of 0.7 PGCs left outside (Figure 1J). To determine the nature of the two alleles, we analyzed the levels of jafrac1 RNA by RT-PCR (Figure S1C) and Jafrac1 protein levels by immunoblotting (Figure S1D). We found that jafrac1KG is a null, lacking both jafrac1 mRNA and protein (Figure S1C, D). The hypomorphic jafrac1F0 allele showed reduced jafrac1 mRNA expression by RT-PCR (Figure S1C). The germ cell adhesion defect in these alleles correlates well with the reduction in RNA expression (Figure 1J). Complementation analysis of the two alleles showed an intermediate phenotype (Figure 1J). Precise transposon excision of the jafrac1F0 piggyBac allele (Thibault et al., 2004) completely reverted the phenotype (Figure 1J). Excisions of the P-element inserted into jafrac1KG caused deletions in the promoter region. We therefore tested for rescue by expressing a UASp-jafrac1 transgene in the jafrac1KG background using the maternal Gal4 driver NGT40 (Tracey et al., 2000) (Figure 4A). This led to a reduction in the penetrance of the PGC adhesion defect phenotype in jafrac1KG from 73% of embryos to 8% of embryos, and a reduction in the average number of PGCs left outside the midgut from 6 to 0.4 (Figure 4A, E). Complementation analysis of jafrac1KG with a deficiency chromosome uncovering the jafrac1 locus showed that an average of 7.2 PGCs failed to enter the midgut with 92% of embryos affected (Figure 1J). The difference in the number of germ cells left outside the midgut between the homozygous jafrac1KG and jafrac1KG over deficiency is not statistically significant (Figure 1J); we therefore conclude that jafrac1KG is a genetic null and jafrac1F0 is a hypomorph. Our results define a previously unrecognized step in germ cell migration during which Jafrac1 regulates germ cell internalization during gastrulation.
Peroxiredoxins act as antioxidants to protect organisms from oxidative damage (Morgan and Veal, 2007). The peroxidase activities of many of the Drosophila peroxiredoxins have been demonstrated in vitro and in Drosophila cell culture assays (Radyuk et al., 2001; Radyuk et al., 2003). To determine if Jafrac1 acts as an antioxidant in vivo, we fed wild type and jafrac1KG adult flies the redox cycler paraquat, which generates H2O2 endogenously within Drosophila (Cochemé and Murphy, 2009), and assessed their survival over time. jafrac1KG mutant flies were less viable compared to wild type flies suggesting an increased sensitivity to oxidative stress in the mutant (Figure 2A). To demonstrate that this increased paraquat sensitivity is caused by loss of Jafrac1, we compared paraquat resistance in jafrac1KG mutant flies to those that also ubiquitously expressed Jafrac1 using actin-Gal4, UASp-jafrac1. Jafrac1 rescued the paraquat sensitivity compared to sibling controls that did not express Jafrac1 (Figure 2A). Furthermore, overexpressing Jafrac1 using the RU486 inducible Tubulin::GeneSwitch (Tub::GS) system (Osterwalder et al., 2001) was sufficient to enhance resistance to oxidative stress (Figure 2B). Our results demonstrate that Jafrac1 protects Drosophila from oxidative stress, consistent with other work demonstrating an antioxidant role for Jafrac1 in neurons (Lee et al., 2009).
The loss of Jafrac1 antioxidant protection could have other deleterious consequences. However, despite their sensitivity to oxidative stress, jafrac1 mutants are homozygous viable and fertile, consistent with studies in other model organisms (Isermann et al., 2004; Lee et al., 2003; Neumann et al., 2003). We also found that adult jafrac1 mutant flies can tolerate dry starvation better than wild type controls (Figure 2C). Indeed, jafrac1 RNA is down regulated in starved larvae (Li et al., 2009) suggesting that regulation of Jafrac1 protein levels contributes to nutritional homeostasis and that a certain amount of oxidative stress may be beneficial during starvation.
As Jafrac1 acts as an antioxidant, we wanted to determine whether loss of the gene could affect the viability, specification, or motility of PGCs. We determined the total number of PGCs in wild type embryos and in embryos from jafrac1KG mutant mothers (hereafter referred to as “jafrac1 mutant embryos” or “jafrac PGCs”) by counting PGCs that migrated through the gut (inside the embryo) as well as those left outside the embryo during embryonic stage 10 and 11. Embryos from wild type and jafrac1KG/+ heterozygous mothers formed on average 25.8 PGCs (25.8 PGCs inside/0.04 PGCs outside) and 26.2 PGCs (26.2 PGCs inside/0 PGCs outside), respectively (Figure S2A). These numbers were similar to those of embryos from jafrac1 mutant mothers in which on average 24.3 PGCs form in jafrac1KG (16.64 PGCs inside/7.64 PGCs outside,) and 24.8 PGCs form in jafrac1KG in trans to a deficiency (17.6 PGCs inside/7.2 PGCs outside,) (Figure S2A). We conclude that Jafrac1 is not required for PGC survival.
We next asked whether Jafrac1 affected PGC specification. Transcriptional quiescence is a hallmark of PGCs (Hanyu-Nakamura et al., 2008; Martinho et al., 2004; Van Doren et al., 1998b). In mutants that affect transcriptional silencing, genes normally only expressed in the soma are now also expressed in PGCs. Using mRNA probes for the zygotic genes tailless (tll) and slow as molasses (slam), which in the wild type are expressed in somatic tissues adjacent to PGCs but not in PGCs, we did not detect any change in transcriptional repression in jafrac1KG mutant PGCs versus wild type controls (Figure S2B, B’, C, C’) (Martinho et al., 2004) (Stein et al., 2002). Thus jafrac1 is not required to maintain transcriptional quiescence in PGCs. Finally, we observed PGC migration in jafrac1KG mutant embryos. jafrac1KG mutant PGCs that enter the midgut during gastrulation polarize prior to migrating through the midgut epithelium (Figure S2D, E) and subsequently move into the mesoderm and reach the somatic gonadal precursors similarly to wild type or heterozygous controls (Figures S2F, G). In addition jafrac1KG mutant PGCs that are left outside the embryo during gastrulation remain there throughout embryogenesis and retain their normal morphology. In summary, we conclude that Jafrac1 is not required for the formation, specification, motility, or survival of PGCs.
During gastrulation, PGCs cluster, are enveloped by the soma, and are brought inside the embryo by somatic tissue reorganization (Jaglarz and Howard, 1995; Kunwar et al., 2008). We visualized PGC internalization in vivo by expressing an actin binding domain of Moesin fused to the enhanced green fluorescent protein in PGCs using the gene regulatory elements of nanos (nos) (Pnos–egfp::moe–nos 3'UTR) (Supplemental Movies 1 & 2) (Sano et al., 2005). PGCs are the first cells formed in the Drosophila embryo and subsequently rest on a layer of later forming epithelial blastoderm cells, the midgut primordium. Before gastrulation begins, PGCs produce protrusions toward the underlying blastoderm (Kunwar et al., 2008). We failed to detect any obvious difference between wild type and jafrac1KG PGCs in protrusion formation at this early stage (Figure 3A, A’). Once gastrulation initiated, wild type PGCs stopped making protrusions and maintained stable contacts with the underlying soma and with each other to form a tight cluster in every embryo imaged (n=5) (Figure 3B, Supplemental movie 1). In contrast, jafrac1KG PGCs were more amoeboid in character and less tightly associated with the soma and each other, preventing stable cluster formation in 5 of the 6 embryos imaged (Figure 3B’, Supplemental movie 2). As gastrulation proceeded, PGCs that were not close to the center of the invaginating midgut primordium in jafrac1KG mutant embryos remained in the hindgut or outside the embryo, instead of being clustered within the midgut (Figure 3C, C’). These results demonstrate an early requirement for jafrac1 in PGC-PGC and/or PGC-midgut adhesion.
Jafrac1 could be required either in the PGCs, the soma, or in both tissues for normal PGC internalization, given that it is expressed in both places (Figure 1E). We rescued the phenotype by expressing the gene maternally (Figure 4A, E), but this method expresses the gene throughout the early embryo in future somatic and germ cells. We therefore generated a transgene that directs Jafrac1 translation specifically in germ cells under the control of the pgc 3’UTR (Figure S3A, B) (Rangan et al., 2009). This transgene partially rescues the Jafrac1 mutant phenotype resulting in an average of 2.5 PGCs left outside of the midgut with 47% of embryos showing a phenotype, suggesting that Jafrac1 acts in PGCs (Figure 4B, E). In contrast, somatic expression of Jafrac1 protein using the nullo-Gal4/UASt-jafrac1 (Figure 4C, E) did not significantly alter the jafrac1KG mutant phenotype (Kunwar et al., 2003; Rose and Wieschaus, 1992; Simpson and Wieschaus, 1990), with an average of 7.2 PGCs left out of the midgut. Taken together with our finding that a maternally provided transgene (soma & PGCs) complemented the phenotype fully, we propose that Jafrac1 acts in PGCs during their internalization, but that maternally provided Jafrac1 in early somatic cells also contributes to this function.
Our data suggest that PGC association with the soma during gastrulation is a regulated step and that adhesion proteins are likely targets for this regulation. Drosophila E-Cadherin (DE-Cadherin encoded by the shotgun (shg) gene in Drosophila) is a conserved AJ component that is maternally provided to PGCs and has been shown to be important for PGC migration and gonad formation (Jenkins et al., 2003; Kunwar et al., 2008; Van Doren et al., 2003). As DE-Cadherin is required for normal oogenesis, we analyzed the maternal effect of loss of function and dominant negative DE-Cadherin alleles in trans to a wild type allele as well as in germ line clones homozygous for a hypomorphic allele (Tepass et al., 1996; Uemura et al., 1996; Broihier, 1998; Kunwar et al., 2008). In embryos from females heterozygous for the dominant negative allele shgG317, an average of 3.7 PGCs were left outside the midgut with 64% of embryos exhibiting a phenotype (Figure 5B, G). An average of 0.6 PGCs remained outside the midgut in embryos from mothers carrying one copy of the null allele shg1H with 29% of embryos exhibiting a phenotype (Figure 5C, G). The PGC adhesion defects seen in embryos from heterozygous shg mothers correlated with the strength of the allele (Broihier, 1998; Tepass et al., 1996) (Figure 5G). We also analyzed embryos from germline clones using the hypomorphic allele, shgA9-49 (Kunwar et al., 2008). An average of 0.4 PGCs were left outside of the midgut with 25% of the embryos showing PGC adhesion defects in these embryos (Figure 5G). This phenotype is dependent on the maternal genotype, as a paternal wild type copy did not alleviate the phenotype. Thus, similar to Jafrac1, DE-Cadherin is necessary for PGC adhesion during gastrulation. In particular, our results suggest that PGC adherence during gastrulation is sensitive to the maternally provided levels of DE-Cadherin.
To test whether jafrac1 and shg genetically interact, we performed gene dosage experiments. While embryos from jafrac1KG heterozygous mothers have normal PGC adhesion during gastrulation, embryos from jafrac1KG, shg doubly heterozygous mothers display a significantly enhanced PGC adherence defect compared to single shg mutants (Figure 5D, G). jafrac1KG/FM7c; shgG317/CyO mothers produced embryos with an average of 6.2 PGCs left outside of the midgut with 91% of embryos affected (Figure 5D, G). Similarly, embryos from jafrac1KG/FM7c; shg373/CyO mothers showed 93% penetrance of the phenotype with an average of 5.4 PGCs left outside of the midgut, a statistically significant enhancement of the shg373/CyO germ cell adherence defect (Figure 5G). Embryos from jafrac1KG/FM7c, shg1H/CyO mothers had an average of 0.9 PGCs outside the midgut with 26% embryos affected (Figure 5G), a statistically insignificant enhancement when compared to embryos from shg1H/CyO mothers (Figure 5G). Our results show that jafrac1 dominantly interacts with the shg G317 and shg 373 alleles leading to an enhancement of the shg PGC adhesion defect. This genetic interaction suggests that Jafrac1 and DE-Cadherin act in concert to regulate PGC adhesion.
To determine whether the requirement for Jafrac1 for normal PGC adhesion was due mainly to an effect on DE-Cadherin, we over-expressed DE-Cadherin specifically in PGCs using nos gene regulatory elements (nos–shg) in a jafrac1 mutant background. Using this method, we were able to rescue the jafrac1KG PGC adhesion defect to an average of 0.5 PGCs left outside of the midgut with 21% of embryos affected (compare Figure 5E and F, G). This result is statistically similar to heterozygous mutant controls with an average of 0.3 PGCs left outside of the midgut with 2% of embryos affected (Figure 5G) and indicates that DE-Cadherin acts downstream of Jafrac1 in PGCs to regulate their adhesion during gastrulation.
To determine whether Jafrac1 directly regulates DE-Cadherin activity or levels, we compared the amount of DE-Cadherin protein in embryos from wild type and jafrac1KG mutant mothers by quantitative fluorescent western blotting (Figure 6A) (Oda et al., 1994). DE-Cadherin protein levels were reduced to 26% of wild type controls in 0 to 5 hour jafrac1KG mutant embryo collections (Figure 6A). Another protein, Rho1, which like DE-Cadherin is present in early PGCs (Kunwar et al., 2008; Kunwar et al., 2003; Magie et al., 2002) was unchanged (Figure 6B). This suggests that the effect on DE-Cadherin levels is specific and not due to a general reduction in protein stability or production in embryos from jafrac1 mutant mothers. Consistent with our immunoblotting results, immunohistochemistry shows that Jafrac1 is required to maintain DE-Cadherin in both PGCs and the midgut until embryonic stage 9 (Figure S4A). jafrac1’s effect on DE-Cadherin protein levels is specific to the embryo as no difference was seen in ovary protein extracts (Figure S4B). Rho1 protein levels were also unaffected by the loss of jafrac1 in the ovary (Figure S4C). Maternal reintroduction of jafrac1 into jafrac1KG mutants rescued DE-Cadherin protein levels to 75% of wild type (Figure 6C), confirming that the absence of jafrac1 causes the observed changes. Taken together, our results indicate that Jafrac1 regulates the level of DE-Cadherin protein.
Jafrac1 could affect DE-Cadherin levels by regulating its protein or RNA stability. We used quantitative RT-PCR to compare shg mRNA levels in jafrac1KG and jafrac1KG/FM7c embryo collections normalized to wild type controls before the onset of zygotic transcription. shg RNA levels in jafrac1KG mutant embryos were not significantly different from heterozygous controls (Figure S4D). Thus, our results support a model in which Jafrac1 is required for the regulation of DE-Cadherin protein translation or degradation during early embryogenesis.
AJ components, like E-Cadherin, are important mediators of cell adhesion. We have shown that Jafrac1 can regulate DE-Cadherin protein levels. This opens up the possibility that Jafrac1 might regulate other AJ complex components (Niessen and Gottardi, 2008). Using quantitative fluorescent westerns, we examined the levels of α-Catenin and β-Catenin/armadillo in jafrac1KG versus wild type 0 to 5 hr embryo collections (Oda et al., 1993; Riggleman et al., 1990). We found a reduction of α-Catenin and β-Catenin proteins in jafrac1KG embryos compared to wild type controls (Figure 6D, E). Quantitation of the western blots revealed that α-Catenin was reduced by 56% and β-Catenin was reduced by 64% compared to wild type when normalized to Vasa loading controls (Figure 6D, E). Our evidence suggests that Jafrac1 stabilizes AJ components to facilitate PGC adhesion during gastrulation.
We hypothesized that Jafrac1 might stabilize DE-Cadherin through an ability to degrade H2O2, thus preventing H2O2 from down-regulating AJ components. To test this, we permeabilized wild type embryos (stages 1 and 2), and exposed them to H2O2 for 2 hours, which resulted in a significant decrease in the protein levels of DE-Cadherin and β-Catenin (Figure 7A, B) when compared to α-Tubulin and Rho1 (Figure 7C). These results demonstrate that H2O2 specifically down-regulates AJ components, and suggest that Jafrac1 stabilizes AJ components by modulating H2O2 concentrations.
Germ cell migration is a conserved developmental process necessary for the formation of a functional gonad (Kunwar et al., 2006). Studies of germ cell migration have identified novel signaling pathways that mediate cell polarization, transepithelial migration, and chemotaxis through multiple tissues to reach specific cell targets. G protein-coupled receptor (GPCR) signaling, lipid phosphate phosphatase activity, and isoprenylation of a proposed attractant by the HMGCR pathway have all been implicated in PGC migration (Hanyu-Nakamura et al., 2004; Kunwar et al., 2008; Kunwar et al., 2003; Renault et al., 2004; Ricardo and Lehmann, 2009; Santos and Lehmann, 2004; Starz-Gaiano et al., 2001; Van Doren et al., 1998a; Zhang et al., 1997). In this study we have uncovered a novel PGC adherence phenotype in which PGCs fail to be incorporated into the midgut during gastrulation. Previously, it was thought that PGCs were passively carried inside the embryo by the movements of the soma (Moore et al., 1998). We now show that PGC adhesion is an active, genetically regulated process.
We show that a peroxiredoxin, Jafrac1, regulates a maternally provided factor essential for PGC adhesion. Our genetic and biochemical analysis has identified DE-Cadherin/Shg as an important downstream target of Jafrac1, and suggests that Jafrac1 stabilizes maternally provided DE-Cadherin protein levels by controlling H2O2. We have also shown that other AJ components, α-Catenin and β-Catenin, require Jafrac1 to maintain their protein expression. The loss or reintroduction of shg alone can modulate PGC adhesion, suggesting that DE-Cadherin levels may affect the stability of the AJ complex as a whole, consistent with the previous report that a decrease in DE-Cadherin levels affects the stability of β-Catenin (Herzig et al., 2007). Taken together, this study represents evidence of a peroxiredoxin redox-regulating adhesion through an effect on AJ.
Our evidence demonstrates that DE-Cadherin can be post-transcriptionally regulated in the early embryo by Jafrac1. Localizing and achieving the appropriate level of DE-Cadherin activity may require more subtle methods of DE-Cadherin regulation than transcriptional control can provide. Post-transcriptional regulation of DE-Cadherin in germ cells is not unprecedented. We have shown previously that the Tre1 GPCR regulates the localization of DE-Cadherin during PGC transepithelial migration (Kunwar et al., 2008). In cell culture, removal of E-Cadherin from the cell membrane can occur by phosphorylation, ubiquitination by the E3-ubiquitin ligase, Hakai, and endocytosis (Fujita et al., 2002). After internalization, ubiquitinated E-Cadherin can be targeted to the lysosome where it is degraded (Palacios et al., 2005). While our data strongly suggest that Jafrac1 influences DE-Cadherin protein levels by modulating H2O2 concentrations, we do not know exactly how H2O2 causes loss of AJ components. We envision two possible mechanisms. First, Jafrac1 could be required for translation of AJ components. Recently, it has been shown that the yeast typical 2-Cys peroxiredoxin Tsa1 acts as a ribosome-associated antioxidant and its peroxidase activity is required for ribosomal function (Trotter et al, 2008). Second, Jafrac1 could negatively regulate H2O2 signaling to protect AJ components from endocytosis and lysosomal degradation. Consistent with our data, oxidants have also been shown to regulate the localization of E-Cadherin/β-Catenin complexes (Rao et al., 2002).
Peroxiredoxins are important regulators of oxidative stress (Bozonet et al., 2005; Ross et al., 2000; Veal et al., 2004; Vivancos et al., 2005) which has been correlated with carcinogenesis; yet how oxidative stress influences the development of tumors is unknown (Benz and Yau, 2008). One model is that reactive oxygen species (ROS) mediated DNA damage leads to oncogenic mutations. Oxidant production also accompanies signal transduction, thus, changes in the cytoplasm may precede chromosomal abnormalities. In support of a causative role for oxidative stress in cancer development, peroxiredoxins have been shown to prevent tumor formation possibly by reducing ROS levels (Neumann et al., 2003). Classifying peroxiredoxins as tumor suppressors is still a subject of debate (Neumann and Fang, 2007). Our findings that a peroxiredoxin is required to maintain normal levels of DE-Cadherin shed new light on the previous results that mutations in mouse Prdx1 lead to increases in tumorigenesis (Neumann et al, 2003). The down-regulation of E-Cadherin is a crucial step during EMT or cancer metastasis (Jang et al., 2007; Thiery and Sleeman, 2006). In cancer, primary tumors grow as a tight cluster. As the disease progresses, cancer cells lose adherence and disperse looking for other niches to colonize. E-Cadherin regulators are molecular switches that may influence metastatic potential. If like Jafrac1, Prdx1 is required to maintain E-Cadherin levels, its loss would lead to a reduction in adhesion between cells, thus promoting metastasis. Reduced Prdx1 and E-Cadherin expression correlate with tumor spread in a human metastatic cancer model, suggesting that peroxiredoxins could have a conserved role in regulating the level of E-Cadherin protein in human cells (Jiang et al., 2003). We speculate that orthologs of the peroxiredoxin we identified in this study may suppress tumor metastasis by protecting E-Cadherin from degradation.
For Drosophila strains and transposon excision see supplemental materials and methods
Whole mount immunohistochemistry was performed on embryos as previously described using the rabbit anti-Vasa antibody (1/2500, a gift of A. Williamson and H. Zinszner), rabbit anti-Jafrac1 (1/500, generated from a peptide by SDIX) and rabbit anti-β-galactosidase (Cappel, 1/20,000) (Stein et al., 2002). RNA in situ hybridization was performed as described earlier using jafrac1, tailless (tll), and slow as molasses (slam) cDNA (Lehmann and Tautz, 1994).
The jafrac1 open reading frame was cloned into P element transformation vectors containing either UAS sequences (UASt or UASp) or a pgc expression cassette (Rangan et al., 2009). P element mediated transformation was conducted by Genetic Services. For more details see the supplementary materials and methods.
Following separation of protein extracts on 4–20% SDS-polyacrylamide gradient gels (Bio-Rad), proteins were transferred to either PVDF or nitrocellulose membranes (Millipore) using the Bio-Rad Mini Trans-Blot Cell system. Membranes were probed with the following antisera: rabbit anti-Jafrac1 SSP51111 (1:100, from Juan Santarén (Rodriguez et al., 2000)), rabbit anti-Vasa (1:2500, A. Williamson and H. Zinszner or our laboratory), DE-Cadherin (1:100, DCAD2 from the Developmental Studies Hybridoma Bank (DSHB)), α-Catenin (1:250, DCAT1 DSHB), Rho1 (1:50, DSHB), Armadillo (1:250, N2 7A1 DSHB), β-Tubulin (1:2500, Covance), or α-Tubulin (1:10000, Sigma). Blots were then incubated with the appropriate secondary antiserum and bands were visualized using either enhanced chemiluminescence or infrared. For more details see the supplementary materials and methods.
jafrac1KG05372 mutant flies were backcrossed to yellow1 white1 (y1w1) flies for 10 generations. jafrac1KG05372/ y1w1 females were then crossed to y1w1 males and the adult progeny was allowed to mate 3 days. y1w1 and hemizygous jafrac1 KG05372 males were separated and starved for 6 hours. The males were then refed with 15 mM paraquat in 5% sucrose. The viability of the flies was then scored over time. The same procedure was followed in the rescue assay using the progeny of jafrac1 KG05372/FM7c; UASp–jafrac1 crossed with actin–Gal4 homozygous males. For over expression, UASp–jafrac1 was driven by the RU486 inducible Tub::GS (Osterwalder et al., 2001). Tub::GS homozygous females were crossed with w1118 or UASp-jafrac1 males and the adult progeny were collected and allowed to mate for 3 days. Males were separated and kept on food supplemented with 200 µM RU486 (dissolved in 80% ethanol) or control food for 2 days. Flies were then starved for 6 hours and fed 15 mM paraquat in 5% sucrose containing 1 mM RU486 or ethanol alone.
jafrac1KG05372 mutant flies were backcrossed to yellow1 white1 (y1w1) flies for 10 generations. jafrac1KG05372/ y1w1 females were then crossed to y1w1 males and the adult progeny was allowed to mate 3 to 5 days. y1w1 and hemizygous jafrac1 KG05372 males were separated and starved over a period of 36 hours, during which the viability of the flies was scored.
Live imaging was performed as described previously (Sano et al., 2005). Briefly, embryos from mothers carrying the Pnos–egfp::moe–nos 3'UTR transgene either in a wild type or a jafrac1 mutant background were collected at room temperature, dechorionated with bleach for 5 minutes and staged by light microscopy. Embryos were mounted on a coverslip in Halocarbon 200 oil and placed on an oxygen permeable membrane (YSL Inc.). Time-lapse images were acquired every 2 minutes via 2-photon microscopy (Prairie Technologies, Inc.). These images were processed to make videos using Imaris software (Bitplane).
Embryos from w1118 mothers were collected for 1 hour at 25°C, dechorionated with 50% (v/v) Clorox bleach for 5 minutes, and then air-dried for 5 minutes at room temperature. The embryos were then immersed in 1 ml octane for 5 seconds before adding an equal volume of buffer containing various concentrations of H2O2. The embryos were incubated for 2 hours at room temperature before immunoblotting (see above).
p values were calculated using two-tailed Student’s t-tests on three or more independent biological replicates.
We would like to thank Patrick B. Burke, Kean Jaime, Brian Richardson, and Allison Blum for their helpful comments on the manuscript. Jessica Seifert kindly provided us with the pnos::shg transgene. We would like to acknowledge the Developmental Studies Hybridoma Bank for antibodies to AJ components, the Bloomington stock collection for flies, and the Drosophila Genomics Resource Center for cDNA clones. We are grateful to Juan Santaren for his kind gift of the Jafrac1 antibody. This work was supported by NIH grant RO1 HD49100 to RL and DS and NIH grant RO1 AG028127 to HJ. RL is an HHMI investigator and a member of the Helen L. and Martin S. Kimmel Center for Stem Cell Biology.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.