The Genomic Response of the Drosophila Embryo to JNK Signaling

The Genomic Response of the Drosophila Embryo to JNK Signaling

Developmental Cell, Vol. 1, 579–586, October, 2001, Copyright 2001 by Cell Press The Genomic Response of the Drosophila Embryo to JNK Signaling Hein...

407KB Sizes 0 Downloads 94 Views

Developmental Cell, Vol. 1, 579–586, October, 2001, Copyright 2001 by Cell Press

The Genomic Response of the Drosophila Embryo to JNK Signaling Heinrich Jasper,1,2,3 Vladimir Benes,1 Christian Schwager,1 Silvia Sauer,1 Sandra Clauder-Mu¨nster,1 Wilhelm Ansorge,1 and Dirk Bohmann1,2,3 1 European Molecular Biology Laboratory Meyerhofstr. 1 69117 Heidelberg Germany 2 Center for Cancer Biology University of Rochester Medical Center Rochester, New York 14642

Summary During Drosophila development, the Jun N-terminal kinase signal transduction pathway regulates morphogenetic tissue closure movements that involve cell shape changes and reorganization of the actin cytoskeleton. We analyzed the genome-wide transcriptional response to activation of the JNK pathway in the Drosophila embryo by serial analysis of gene expression (SAGE) and identified loci encoding cell adhesion molecules and cytoskeletal regulators as JNK responsive genes. The role of one of the upregulated genes, chickadee (chic), encoding a Drosophila profilin, in embryogenesis was analyzed genetically. chicdeficient embryos fail to execute the JNK-mediated cytoskeletal rearrangements during dorsal closure. This study demonstrates a transcriptional mechanism of cytoskeletal regulation and establishes SAGE as an advantageous approach for genomic experiments in the fruitfly. Introduction In the fruitfly Drosophila melanogaster, the JNK signal transduction pathway is best known as a mediator of major morphogenetic movements, including embryonic dorsal closure and adult thorax closure (Kockel et al., 2001; Stronach and Perrimon, 1999). Impaired JNK signaling in the developing embryo is associated with insufficient actin condensation and failure to generate actinbased protrusions from the cells of the leading edge of the lateral epidermis, resulting in defective dorsal closure (Edwards et al., 1997; Jacinto et al., 2000; MartinBlanco and Knust, 2001; Ricos et al., 1999). Striking similarities at the tissue-morphological, the cellular, and the molecular levels can be observed between wound repair in vertebrates and the aforementioned morphogenetic processes (Grose and Martin, 1999; Jacinto et al., 2001), suggesting that Drosophila dorsal closure might provide a genetically tractable model system for studying the molecular biology of this important regenerative event. The signaling components that are required for dorsal 3 Correspondence: [email protected], [email protected]

closure encompass, among others, the JNKK Hemipterous (Hep), the JNK basket (Bsk) and its substrates, the transcription factors Jun and Fos, as well as their transcriptional target, the secreted TGF␤-type protein Decapentaplegic (Dpp) (reviewed in Kockel et al., 2001; Stronach and Perrimon, 1999). Whether changes in gene expression programs in addition to dpp upregulation might affect reorganization of the actin cytoskeleton is not known. To gain a deeper understanding of the biology and complexity of this and other JNK-regulated processes in the context of an organism, a comprehensive description of the gene expression programs involved is of key importance. Thus, we embarked on a genome-wide analysis of JNK-responsive genes in the developing Drosophila embryo using SAGE (serial analysis of gene expression; Velculescu et al., 2000; Velculescu et al., 1995; Velculescu et al., 1997). Results Transcriptome Analysis in Drosophila by SAGE A functional genomic analysis of JNK target genes in Drosophila is facilitated by extensive bioinformatics resources and the powerful genetic tools that are at hand in this system. Different approaches to map quantitative changes of the transcriptome have been developed in recent years. Microarray-based techniques offer unprecedented throughput with the parallel probing of thousands of transcripts. A main disadvantage is that the generated data represent expression ratios, which only indirectly reflect absolute mRNA abundances. Furthermore, microarrays are intrinsically limited to genes that have been previously identified or predicted using computational methods. With SAGE, virtually every transcript in a sample RNA population can be identified and quantitated by generating specific 14 bp sequence tags from a defined position. Concatemers of such tags are then sequenced, and the frequency with which a given tag is detected represents a direct measure of the abundance of the corresponding mRNA (Velculescu et al., 1995; Velculescu et al., 1997). SAGE in Drosophila has become particularly powerful with the availability of the Drosophila genomic sequence (Adams et al., 2000). Due to the comparatively small size of the fly genome (1.2 ⫻ 108 bases of euchromatin), the 14 bases of a SAGE tag (2.7 ⫻ 108 possibilities) are typically sufficient to locate the respective transcript in the genome without having to rely on further information. The dorsal closure process takes place between embryonic stages 13 and 16, corresponding to 10–16 hr of development at 25⬚C. To capture the relevant changes of gene expression involved in setting up and executing dorsal closure, we generated SAGE libraries from 4–16 hr old Drosophila embryos in which the JNK pathway was either repressed by the ubiquitous expression of a dominant-negative form of Basket (BskDN; Weber et al., 2000) or ubiquitously activated due to expression of a constitutively active form of Hemipterous (Hepact; Weber

Developmental Cell 580

Figure 1. Identification of JNK-Responsive Genes in the Drosophila melanogaster Embryo by SAGE (A–E) Cuticular phenotypes of the embryos used for the preparation of SAGE libraries. Embryos expressing the armadillo GAL4 (armG4) driver alone (A) display a wild-type cuticle phenotype. UAS hepact/⫹; armG4/⫹ embryos (C) resemble mutants homozygous for the puckered null allele pucE69 (B). Embryos expressing bskDN under the control of armG4 (E) show classical dorsal closure phenotypes as seen for example in embryos devoid of Hep function (hepr75/hepr75) (D). (F) Graph showing the distribution of up- or downregulated genes when comparing the hepact to the wild-type library. Note that the majority of tags are present in similar numbers in both libraries (values between ⫺3 and ⫹3). Correlation coefficients between libraries are as follows: hepact/wt: 0.86; hepact/bskDN: 0.78; wt/bskDN: 0.93.

et al., 2000). A library from wild-type embryos was prepared as a reference. The expression of the transgenes was dependent on Gal4 that was expressed ubiquitously under the control of the armadillo promoter starting at around 4.5 hr after egg laying (Sanson et al., 1996). The effect of the BskDN and Hepact molecules on the transcription of target genes was therefore limited to the period of embryogenesis relevant for dorsal closure. The developmental defects generated by expression of BskDN phenocopied mutants lacking components of the JNK pathway (Figures 1D and 1E; Stronach and Perrimon, 1999). Similarly, embryos expressing Hepact resembled puckered (puc) mutants and displayed a JNK gain-of-function phenotype (Figures 1B–1C; MartinBlanco et al., 1998). Thus, we expected transcripts regulated by JNK signal transduction to be differentially represented in these embryos. Employing dominantly acting transgenes (instead of homozygosity for a lossof-function allele) made it possible to use large numbers of isogenic embryos for the analyses. Approximately 10,000 tags from each of the three libraries were sequenced, and among the combined data 8,180 different tags were identified. A comparison of the datasets indicated that most transcripts were represented at very similar levels in all three libraries (Figure 1F). We considered a tag to be significantly up- (or down-) regulated when its number differed by 3-fold or more between the two compared libraries. When a given tag was absent in one library, but present three times or more in the comparison library, that tag was also considered to be upregulated. In comparing the hepact and the bskDN libraries, 312 tags were upregulated and 322 were

downregulated in the hepact library. Of these tags, 44 were upregulated and 27 were downregulated more than 10-fold. A similar picture arose when comparing the hepact library to the wild-type library: 278 tags were upregulated (36 more than 10-fold), while 267 were downregulated (with 29 more than 10-fold; Figure 1F). Most SAGE tags that were differentially represented in the bskDN and hepact libraries were present at intermediate levels in the wild-type control (see Table 1 and Supplemental Data [ content/full/1/4/579/DC1]). For the complete set of SAGE data generated, see Supplemental Data. Excel workbook 1 includes tables of tags and their frequency of occurrence in each library. Workbook 2 contains a table comparing the frequencies of tags across all three libraries. Workbook 3 carries tables of up- and downregulated tags with their annotation. For the annotation of SAGE tags in Drosophila, we generated a database containing the predicted tags for the 13,601 genes annotated by the Berkeley Drosophila Genome Project (Adams et al., 2000). Additionally, we employed the USAGE website (van Kampen et al., 2000) to generate a Drosophila tagmap from GenBank data (both databases included in Supplemental Data [http:// DC1] as Microsoft Access files). Using these databases, we could annotate more than 50% of all tags present in the three libraries. Interesting (i.e., differentially expressed) nonannotated tags were submitted to a BLAST search against the Drosophila genome to identify unpredicted transcription units. This analysis permitted the additional annotation of around 50% of these tags.

SAGE Analysis of JNK Response in Drosophila 581

Figure 2. Expression of JNKK-Responsive Genes in the Embryo RNA in situ hybridization and RT-PCR on selected mRNAs confirm their inducible expression upon ubiquitous JNK signaling. Whole-mount in situ hybridizations were performed using probes specific for IMP-L1 (A–D), CG5080 (F–I), and chickadee (chic, K–N). The genotypes of the embryos subjected to this analysis are indicated in the upper left corner of each row. The number of tags corresponding to the respective transcript among the approximately 10,000 analyzed in each library are shown in the upper right corner of each panel. Tag numbers shown for chic are derived from two tags corresponding to transcripts with alternative 3⬘ ends. Note the increased expression of the examined genes in Hepact-expressing embryos (C, H, and M) and their reduced expression in BskDN expressing (B, G, and L) and in hep mutant (D, I, and N) embryos. (E, J, and O) RT-PCR analysis using whole embryo RNA from the described genotypes (DN: armG4/UASbskDN; Hep: UAShepact/⫹;armG4/⫹). As an internal control, RNA levels of rp49 were analyzed. -RT designates control reactions performed in the absence of reverse transcriptase.

These figures indicate that the annotation error rate with regard to exon boundaries in the present release of the Drosophila genome data is at least 25%, which is consistent with estimates by others (Reese et al., 2000). The information provided by SAGE experiments will, therefore, facilitate genome annotation, especially concerning the 3⬘ ends of transcribed regions. Roughly 25% of all tags could not be annotated possibly due to polymorphisms or sequencing errors and gaps in the published genomic sequence. In order to independently verify the SAGE data, we performed RT-PCR analyses of some of the identified mRNAs (see Figure 2 for examples). The quantitation of 25 representative mRNAs confirmed the SAGE data in all cases but one (see Table and data not shown). To further corroborate the results of the SAGE analysis, RNA in situ hybridization experiments for several of the identified Hep-responsive mRNAs were performed. In all cases where an informative signal could be detected, a clear increase in the expression of the putative

target genes in Hepact-expressing embryos relative to wild-type or BskDN-expressing embryos was apparent. Three examples are shown in Figure 2. IMP-L1 (Figures 2A–2E), a secreted molecule of unknown function, is inducible by ecdyson (Natzle et al., 1992) in addition to the JNK responsiveness shown here. CG5080 (Figures 2F–2J) has homologies to profilaggrin and paramyosin and thus might encode an actin binding protein or a protein involved in intermediate filament assembly. The third gene shown, chickadee (Figures 2K–2O), encodes a Drosophila profilin, an actin binding protein involved in the regulation of cytoskeletal dynamics (Pollard et al., 2000). Consistent with the SAGE data, all three genes show increased expression upon ubiquitous activation of the JNK pathway and reduction of their endogenous expression in BskDN expressing embryos (Figure 2). Importantly, the expression of these genes is also clearly reduced in hep mutant embryos (Figures 2D, 2I, and 2N), confirming that the downregulation is a consequence of impaired

Developmental Cell 582

Table 1. JNK-Responsive Genes Encoding Cytoskeletal Regulators and Cell Adhesion Molecules or Gene Products Previously Known to Interact with JNK Pathway Components Tag Sequence

bskDN WT hepact Gene Name

RT-PCR Function/Note

GadFly Ann.


0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 3 2 0 0 2

⫹ ND ⫹⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹⫺ ⫹⫺ ⫹ ND ND ⫹ ND ND ND ⫹ ND ND ND ND ND ⫹ ND ND ND

Fban0010130 Fban0007116 FBan0008827 FBan0001388 FBan0005080

1 0 0 0 3 1 1 0 2 0 1 1 0 0 0 0 2 2 2 0 1 0 0 3 9 5 1 0 1

7 4 6 3 12 7 7 12 5 5 4 4 4 4 4 4 7 6 3 3 3 3 3 3 14 9 3 3 7

Sec61beta IMP-E1 Angiotensin conv. enzyme (Ance) TGF␤-activated kinase (tak1) CG5080 chickadee (chic/alternative 3⬘UTR) gp150 Tropomyosin2 (Tm2) CG9211 (neogenin-like protein) Amalgam EG:8D8.7 Cortactin RalA CG12191 (adhesion molecule L1.1) CG3599 myospheroid (mys) myosin alkali light chain 1 (mlc1) CG12908 (nidogen-like) wings-upA/troponin I (wupA) CG9128 (synaptojanin) CG1017 (microfibrillar-associ. prot.) Troponin C at 73F (TpnC73F) Tropomyosin1 (Tm1) myosin light chain 2 (mlc-2) myosin ligh chain 2 (mlc-2) chickadee (chic) IMPL2 tiggrin steamer duck (stck)

Suppressor of cJunAsp JNK-inducible gene Fos target in amnioserosa JNKK activating kinase cytoskeletal regulation cytoskeletal regulation cell adhesion cytoskeletal regulation cell adhesion cell adhesion cell adhesion cell adhesion/signal transduction cytoskeletal reg./signal transduction cell adhesion cell adhesion cell adhesion molecular motor cell adhesion cytoskeletal regulation signal transduction/endocytosis cell adhesion cytoskeletal reg./calcium binding cytoskeletal regulation cytoskeletal regulation cytoskeletal regulation cytoskeletal regulation cell adhesion cell adhesion/integrin binding cytoskeletal reg./transcription factor

FBan0005820 FBan0004843 FBan0009211 FBan0002198 FBan0011415 FBan0003637 FBan0002849 FBan0012191 FBan0003599 FBan0001560 FBan0005596 FBan0012908 FBan0007178 FBan0009128 FBan0001017 FBan0007930 FBan0004898 FBan0002184 FBan0002184 FBan0009553 FBan0015009 FBan0011527 FBan0007954

For each tag, the number of hits in the three libraries and its corresponding gene are listed. Using the Gadfly annotation number, comprehensive information on gene function, alleles, interacting genes, etc., can be retrieved at Genes known to interact with JNK pathway components are printed in bold. Where experimental data are not available, function was implied by homology. Semiquantitative RT-PCR was used as an assay to validate SAGE data. Validation is indicated by ⫹ (ND, not determined; ⫹⫺, weak upregulation seen in hepact). See additional information for literature describing the functional involvement of the listed molecules and their connection to JNK signaling.

JNK signal transduction and not an unspecific side effect of BskDN expression. These experiments also showed that, although induced by a ubiquitous JNK signal, the patterns in which increased expression could be detected were different for many of the genes examined. Evidently, the responsiveness of a given gene to Hep activity is highly dependent on the cellular context and developmental stage. The Hepact-induced overexpression shows spatial and temporal restrictions similar to the expression pattern of the respective mRNAs in a wild-type situation (Figure 2 and data not shown). These results suggest that JNK-inducible transcription factor(s) act in concert with other regulators to contribute in a combinatorial manner to the spatiotemporal domain of JNK-responsive gene expression. A similar combinatorial control of gene expression in Drosophila has recently been reported for other signaling pathways (Guss et al., 2001). Regulation of the Actin Cytoskeleton and Cellular Adhesiveness by JNK Signaling Among the upregulated genes identified, several were known from previous studies to interact genetically or biochemically with components of the JNK pathway or to be required for embryonic dorsal closure (Supplemental Data, Table 1 [ content/full/1/4/579/DC1]). However, dpp and puc, until

now the only known JNK-responsive genes, were not among them. Despite a significant upregulation of both dpp and puc in Hepact-expressing embryos as detected by RT-PCR (data not shown), significant numbers of the corresponding tags were not obtained in the SAGE experiment (0/0/1 for puc and 0/0/0 for dpp in the bskDN, wt, and hepact libraries, respectively). The absence of dpp and puc in our analysis is due to the low expression levels of these regulatory transcripts. To generate statistically relevant SAGE data for such rare messages, greater numbers of tags will have to be sequenced. Consistent with the proposed role of JNK signaling in reorganization of the actin cytoskeleton, our analysis identified several cytoskeletal regulators as JNK target genes (Table 1; Supplemental Data, Table 2 [http:// DC1]). One example is the Drosophila homolog of the profilin gene, chickadee (chic), which is strongly upregulated in Hepact-expressing embryos (Table 1; Figure 2). Although mutants for chic have been described as defective in a number of actin-dependent processes (Verheyen and Cooley, 1994), a role in dorsal closure has not yet been reported. We therefore examined mutants lacking chic function. In addition to pleiotropic defects observed in cuticle preparations, about 30% of these embryos secreted a very thin cuticle with obvious dorsal closure defects (a strong example is shown in Figure

SAGE Analysis of JNK Response in Drosophila 583

Figure 3. chic as an Effector of JNK Signaling during Dorsal Closure and Actin Reorganization Approximately 30% of all chic221/chic221 embryos exhibited mild or strong dorsal closure defects in cuticle preparations (A). Note the thin cuticle of this embryo, indicating a later requirement of chic for cuticle deposition. Immunostaining of chic heterozygous (B) and chic homozygous embryos (C) at stage 14–15 with anti-Fasciclin 3 (outlining individual cells) reveals incomplete cell stretching and misalignment of the two lateral epidermal cell sheets during closure of the mutants. (D and E) Male embryos from crosses of hep1 homozygous females to wild-type males die with mild dorsal closure defects (arrows in [D] and [E] demarcate the length of the dorsal hole). Only 30% of these embryos fail to close the four posterior segments (phenotype as in [E]). This phenotype increases to over 60% in dead embryos from crosses of hep1 homozygous females to chic221/CyO males. A quantification of these genetic interaction experiments is shown in (F). Averages and standard deviations are shown for three independent crosses. Total number of cuticles scored was n ⫽ 885 for hep1/hep1 ⫻ chic221 and n ⫽ 463 for hep1/hep1 ⫻ OreR. The mutant phenotype of Hepact-expressing embryos (G) is suppressed in a chic heterozygous background (H). (I–O) TRITC Phalloidin-staining visualizing the distribution of filamentous actin throughout the embryo. The prominent actin cable present in both wildtype (I and M) and hepact/⫹;armGal4/⫹ (K) embryos is much less pronounced in embryos with decreased JNK activity (L and O) as well as in chic221 mutants (J and N). Note the ectopic foci of actin polymerization throughout the lateral epidermis in embryos with increased JNK activity (arrows in [K]). Actin-based filopodia emanate from leading edge cells in wild-type embryos (arrows in [M]) but not in chic or hep mutant embryos (N and O).

3A). The lateral epidermis failed to stretch normally during the closure process, confirming that the defects detected were not secondary effects caused by the weak cuticle (Figures 3B and 3C). To further investigate the proposed role of chic downstream of JNK signaling in dorsal closure, we analyzed

whether chic and hep mutations interact genetically. When female flies homozygous for the X-linked hypomorphic mutation hep1 are crossed to wild-type males, the male offspring dies with mild dorsal closure defects: only 30% of these embryos are completely open (as shown in Figure 3E), whereas in the remaining 70%,

Developmental Cell 584

the segments a5–a8 close normally (they are “anterior open”; see Figure 3D). In contrast, when hep1 homozygous females are crossed to chic221 heterozygous males, the dorsal open phenotype is enhanced and the number of completely open embryos in the offspring is increased to around 65%. Thus, the gene dose of chic becomes critical in embryos with compromised hep function (Figures 3D–3F). In an independent experiment, we found that in chic heterozygous embryos the phenotypic effects of Hepact expression are suppressed (Figures 3G and 3H). Together with the molecular data, these genetic interactions suggest that chic is required downstream of hep for normal dorsal closure. Profilin Is Required for Cytoskeletal Rearrangements during Dorsal Closure To understand the cellular function of Hep-induced chic expression in the embryo in more detail, we analyzed the actin cytoskeleton in the relevant genotypes (Figures 3I–3O). Hepact-expressing embryos display ectopic foci of actin polymerization among the lateral epithelial cells and show increased actin polymerization in leading edge cells, resulting in a stronger actin cable compared to wild-type embryos (Figures 3I and 3K). In contrast, the leading edge of embryos lacking JNK activity shows a less prominent actin cable overall, which is occasionally disrupted (Figures 3L and 3O). Significantly, the same phenotype can be observed in chic mutant embryos (Figure 3J), consistent with Chic and Hep acting in the same pathway. Actin-based filopodia that extend dorsally from wildtype leading edge cells have been proposed to mediate the movement and proper alignment of the lateral epidermal sheets during dorsal closure (Jacinto et al., 2000; reviewed in Martin-Blanco and Knust, 2001). In hepdeficient embryos, these structures do not form. chic mutants share this phenotype and show an almost complete lack of these filopodia (Figure 3N), indicating that this defect is a consequence of insufficient profilin expression in JNK pathway mutants. Discussion We analyzed the genome-wide transcriptional response to JNK signaling in the Drosophila embryo by SAGE and identified several hundred genes that are either up- or downregulated by this signal transduction pathway. The finding that this set includes genes known to be involved in JNK-dependent processes based on prior genetic studies strongly supports the validity of the generated data. The genes that are identified by this analysis to be upregulated by the Drosophila JNK pathway can be grouped into several functional classes. One of these encompasses genes that, by homology, can be implicated in cellular stress responses (see Supplemental Data, Table 3 [ content/full/1/4/579/DC1]). This finding confirms a conserved role for JNK in the cellular defense against exogenous insults as previously predicted (Goberdhan and Wilson, 1998) and shows that, similar to its vertebrate homologs, Drosophila JNK mediates multiple biological functions.

Another group of upregulated genes encodes cytoskeletal regulators and cell adhesion molecules that represent good candidates for effectors of the JNKmediated morphogenetic processes during Drosophila development. This suggests that the coordinated induction of these genes by JNK is crucial for the execution of dorsal closure. The induced gene products might directly control the dorsal closure process or create a permissive state that allows the morphogenetic events to occur. We functionally characterized one example, the profilin-encoding gene chic, in more detail. Remarkably, the results of the SAGE experiment correctly predicted the function of profilin as an effector of JNK in dorsal closure. The phenotype of chic mutant embryos indicates that profilin acts downstream of JNK signaling to induce filopodia formation in the leading edge cells and to generate or stabilize the actin cable during dorsal closure. Both of these events are regarded as essential for the dorsal closure process. These findings suggest that the upregulation of profilin by the JNK pathway is an essential event of dorsal closure and provides an example of how, at least in the developmental context studied here, cytoskeletal organization can be influenced by a transcriptionally controlled mechanism. Detailed analyses of the other identified JNK-responsive cytoskeletal regulators and cell adhesion molecules will provide a more comprehensive view of the mechanisms by which cell sheets approach each other and fuse in a signal-dependent, directed, and ordered manner. This study illustrates the possibilities of applying novel genomic methods to the analysis of signaling during development of multicellular organisms. The synergism of combining the “awesome power of genetics,” as applied in Drosophila, with functional genomic approaches including SAGE will continue to expand our understanding of the development and function of organisms at the molecular level. Experimental Procedures Fly Strains and Genetics Fly strains used in this study were the following: w;;UASbskDN and w;UAShepact (Weber et al., 2000), gifts from M. Mlodzik; chic221 cn1/ CyO; ry506 (Verheyen and Cooley, 1994), obtained from the Bloomington stock center; hep1/FM6 and hepr75/FM6 (Glise et al., 1995), gift from S. Noselli. Crosses were performed at 25⬚C. After overnight deposition on apple juice plates, embryos were aged at 25⬚C. Embryonic cuticles were prepared using standard methods. For the generation of SAGE libraries, we used embryos derived from crosses of homozygous w;UAShepact or w;;UASbskDN females with homozygous w;;armG4 males to avoid any heterogeneity in the embryo population and to have a sufficient quantity of isogenic embryos for RNA isolation. For the preparation of the wild-type library, embryos derived from w1118 females crossed to w;;armG4 homozygous males were used. The analysis of hep mutants (Figures 2 and 3) was performed with embryos derived from crosses of hepr75/hep1 females with hep1/y males. Serial Analysis of Gene Expression Serial analysis of gene expression was performed as described (St Croix et al., 2000; Virlon et al., 1999), with minor modifications. A detailed protocol is included in Supplemental Data [http://www.]. Briefly, total RNA from dechorionated embryos was purified with Trizol (GIBCO). Dynabeads pdT (100 ␮l; Dynal) were used to isolate pA⫹ RNA from

SAGE Analysis of JNK Response in Drosophila 585

100 ␮g of total RNA. Double-stranded cDNA was synthesized on the beads and digested with the anchoring enzyme (NlaIII; NEB). After linker ligation, digestion with the tagging enzyme (BsmFI, NEB), and ligation of the ditags, PCR amplification (29 cycles) was carried out with 20% of the ligation product as template. The 100 bp PCR products were purified and submitted to a secondary PCR (10–12 cycles) with biotinylated primers to generate enough material for the concatemerization. After NlaIII digestion, the released ditags were purified by polyacrylamide gel electrophoresis and subsequently incubated with 100 ␮l of Dynabeads Streptavidin to eliminate any remaining biotinylated linkers. Concatemerization was carried out for four hours. Concatemers were cloned into the SphI site of pZero1 (Invitrogen), and resulting colonies were screened for inserts by PCR and submitted for sequencing. SAGE Sequencing Sequencing was performed on the ARAKIS sequencing system with array detectors, developed at EMBL (Erfle et al., 1997). Gel images were processed with Lane Tracker, and raw sequencing data were viewed with Gene Skipper. Both software packages were developed at EMBL.

References Adams, M.D., Celniker, S.E., Holt, R.A., Evans, C.A., Gocayne, J.D., Amanatides, P.G., Scherer, S.E., Li, P.W., Hoskins, R.A., Galle, R.F., et al. (2000). The genome sequence of Drosophila melanogaster. Science 287, 2185–2195. Edwards, K.A., Demsky, M., Montague, R.A., Weymouth, N., and Kiehart, D.P. (1997). GFP-moesin illuminates actin cytoskeleton dynamics in living tissue and demonstrates cell shape changes during morphogenesis in Drosophila. Dev. Biol. 191, 103–117. Erfle, H., Ventzki, R., Voss, H., Rechmann, S., Benes, V., Stegemann, J., and Ansorge, W. (1997). Simultaneous loading of 200 sample lanes for DNA sequencing on vertical and horizontal, standard and ultrathin gels. Nucleic Acids Res. 25, 2229–2230. Glise, B., Bourbon, H., and Noselli, S. (1995). hemipterous encodes a novel Drosophila MAP kinase kinase, required for epithelial cell sheet movement. Cell 83, 451–461. Goberdhan, D.C., and Wilson, C. (1998). JNK, cytoskeletal regulator and stress response kinase? A Drosophila perspective. Bioessays 20, 1009–1019. Grose, R., and Martin, P. (1999). Parallels between wound repair and morphogenesis in the embryo. Semin. Cell Dev. Biol. 10, 395–404.

Analysis of SAGE Data and Annotation of SAGE Tags Sequenced SAGE concatemers were analyzed using the SAGE2000 program obtained from The Johns Hopkins University (see also The database linking SAGE tags to data of the Berkeley Drosophila Genome Project was built using datasets downloaded from the BDGP site ( and extracting the 10 bp sequence downstream of the 3⬘-most CATG site. These putative tags were linked to the GadFly site of the corresponding gene. This and other databases can be downloaded in Microsoft Access and HTML format from the following sites: http://www.urmc. and http://www. Annotation of experimental data was performed using Microsoft Access to link the experimental dataset and the Tag annotation database. RT-PCR and In Situ Hybridization RT-PCR was performed using standard protocols (Steinbach and Rupp, 1999). Transcript levels of rp49 were monitored as internal controls. See Supplemental Data (http://www.developmentalcell. com/cgi/content/full/1/4/579/DC1) for detailed PCR protocols and primer sequences. In situ hybridizations were performed using probes prepared from cloned fragments (400–600 bp) of the examined genes using a standard protocol (Tautz and Pfeifle, 1989). All compared genotypes were incubated with the same amount of probe and developed side-by-side for the same period of time in order to ensure comparable results. Phalloidin Staining and Confocal Microscopy For phalloidin stains, embryos were treated as described (Ricos et al., 1999), mounted in Moviol/DABCO, and analyzed with a TCSSP1 Confocal Microscope (Leica). Acknowledgments We thank M. Mlodzik, L. Cooley, S. Cohen, S. Noselli, the Bloomington Stock Center, and the Developmental Studies Hybridoma Bank for gifts of fly stocks and antibodies. V. Velculescu, J.-M. Elalouf, and U. Wirkner provided invaluable experimental help and advice on SAGE. K. Kinzler and The Johns Hopkins University are acknowledged for sharing their SAGE analysis software. M. Mlodzik, M. Boutros, L. Kockel, W. Li, U. Gritzan, L. Ciapponi, C. Weiss, J. Brennecke, J. Homsy, and C. Ovitt improved the quality of this manuscript with their critical and insightful comments. H.J. was supported by a fellowship from the DFG Graduiertenkolleg No. 484, “Signalling Systems and Gene-Expression in Developmental Model Systems”. Received June 6, 2001; revised July 30, 2001.

Guss, K.A., Nelson, C.E., Hudson, A., Kraus, M.E., and Carroll, S.B. (2001). Control of a genetic regulatory network by a selector gene. Science 292, 1164–1167. Jacinto, A., Wood, W., Balayo, T., Turmaine, M., Martinez-Arias, A., and Martin, P. (2000). Dynamic actin-based epithelial adhesion and cell matching during Drosophila dorsal closure. Curr. Biol. 10, 1420– 1426. Jacinto, A., Martinez-Arias, A., and Martin, P. (2001). Mechanisms of epithelial fusion and repair. Nat. Cell Biol. 3, E117–E123. Kockel, L., Homsy, J.G., and Bohmann, D. (2001). Drosophila AP-1: lessons from an invertebrate. Oncogene 20, 2347–2364. Martin-Blanco, E., and Knust, E. (2001). Epithelial morphogenesis: filopodia at work. Curr. Biol. 11, R28–R31. Martin-Blanco, E., Gampel, A., Ring, J., Virdee, K., Kirov, N., Tolkovsky, A.M., and Martinez-Arias, A. (1998). puckered encodes a phosphatase that mediates a feedback loop regulating JNK activity during dorsal closure in Drosophila. Genes Dev. 12, 557–570. Natzle, J.E., Robertson, J.P., Majumdar, A., Vesenka, G.D., Enlow, B., and Clark, K.E. (1992). Sequence and expression of IMP-L1, an ecdysone-inducible gene expressed during Drosophila imaginal disc morphogenesis. Dev. Genet. 13, 331–344. Pollard, T.D., Blanchoin, L., and Mullins, R.D. (2000). Molecular mechanisms controlling actin filament dynamics in nonmuscle cells. Annu. Rev. Biophys. Biomol. Struct. 29, 545–576. Reese, M.G., Hartzell, G., Harris, N.L., Ohler, U., Abril, J.F., and Lewis, S.E. (2000). Genome annotation assessment in Drosophila melanogaster. Genome Res. 10, 483–501. Ricos, M.G., Harden, N., Sem, K.P., Lim, L., and Chia, W. (1999). Dcdc42 acts in TGF-beta signaling during Drosophila morphogenesis: distinct roles for the Drac1/JNK and Dcdc42/TGF-beta cascades in cytoskeletal regulation. J. Cell Sci. 112, 1225–1235. Sanson, B., White, P., and Vincent, J.P. (1996). Uncoupling cadherinbased adhesion from wingless signalling in Drosophila. Nature 383, 627–630. St Croix, B., Rago, C., Velculescu, V., Traverso, G., Romans, K.E., Montgomery, E., Lal, A., Riggins, G.J., Lengauer, C., Vogelstein, B., and Kinzler, K.W. (2000). Genes expressed in human tumor endothelium. Science 289, 1197–1202. Steinbach, O.C., and Rupp, R.A. (1999). Quantitative analysis of mRNA levels in Xenopus embryos by reverse transcriptase-polymerase chain reaction (RT-PCR). Methods Mol. Biol. 127, 41–56. Stronach, B.E., and Perrimon, N. (1999). Stress signaling in Drosophila. Oncogene 18, 6172–6182. Tautz, D., and Pfeifle, C. (1989). A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosma 98, 81–85.

Developmental Cell 586

van Kampen, A.H., van Schaik, B.D., Pauws, E., Michiels, E.M., Ruijter, J.M., Caron, H.N., Versteeg, R., Heisterkamp, S.H., Leunissen, J.A., Baas, F., and van der Mee, M. (2000). USAGE: a web-based approach towards the analysis of SAGE data. Serial Analysis of Gene Expression. Bioinformatics 16, 899–905. Velculescu, V.E., Zhang, L., Vogelstein, B., and Kinzler, K.W. (1995). Serial analysis of gene expression. Science 270, 484–487. Velculescu, V.E., Zhang, L., Zhou, W., Vogelstein, J., Basrai, M.A., Bassett, D.E., Hieter, P., Vogelstein, B., and Kinzler, K.W. (1997). Characterization of the yeast transcriptome. Cell 88, 243–251. Velculescu, V.E., Vogelstein, B., and Kinzler, K.W. (2000). Analysing uncharted transcriptomes with SAGE. Trends Genet. 16, 423–425. Verheyen, E.M., and Cooley, L. (1994). Profilin mutations disrupt multiple actin-dependent processes during Drosophila development. Development 120, 717–728. Virlon, B., Cheval, L., Buhler, J.M., Billon, E., Doucet, A., and Elalouf, J.M. (1999). Serial microanalysis of renal transcriptomes. Proc. Natl. Acad. Sci. USA 96, 15286–15291. Weber, U., Paricio, N., and Mlodzik, M. (2000). Jun mediates Frizzledinduced R3/R4 cell fate distinction and planar polarity determination in the Drosophila eye. Development 127, 3619–3629.