The Toll gene in Drosophila pattern formation

Toll is an exclamation of amazement that roughly translates to ‘Wow’ in English. In the dominant Toll-phenotype, the ventral segmental bands circle all around the Drosophila larva. Toll is a member of the ‘dorsal-group’ genes, named after the previously found dorsal gene, the phenotype of which displays only dorsal pattern elements.The 12 dorsal-group genes define a single system determining the entire dorsoventral axis of the Drosophila embryo. Most components of the Toll-Dorsal pathway and much of the logic of their interactions were discovered by the powerful combination of genetics and transplantation experiments, subsequently confirmed at the biochemical level.The Toll-Dorsal pathway also acts in the microbial defence system of Drosophila, whereas Toll-like receptors in mammals function as pattern recognition receptors covering almost the entire spectrum of innate immunity. Toll-like receptors (TLRs) play a crucial role in innate immunity in animals. Their discovery was rewarded a Nobel Prize to Jules Hoffmann and Bruce Beutler in 2011. The name Toll stems from a Drosophila mutant that was isolated in 1980 by Eric Wieschaus and myself as a byproduct of our screen for segmentation genes in Drosophila for which we received the Nobel Prize in 1995. It was named Toll due to its amazing dominant phenotype displayed in embryos from Toll/+ females. The analysis of Toll by Kathryn Anderson in my laboratory in Tübingen and subsequently in her own laboratory in Berkeley singled out Toll as a central component of the complex pathway regulating dorsoventral polarity and pattern of the Drosophila embryo. The Drosophila Toll story provides a striking example for the value of curiosity-driven research in providing fundamental insights that later gain strong impact on applied medical research. Toll-like receptors (TLRs) play a crucial role in innate immunity in animals. Their discovery was rewarded a Nobel Prize to Jules Hoffmann and Bruce Beutler in 2011. The name Toll stems from a Drosophila mutant that was isolated in 1980 by Eric Wieschaus and myself as a byproduct of our screen for segmentation genes in Drosophila for which we received the Nobel Prize in 1995. It was named Toll due to its amazing dominant phenotype displayed in embryos from Toll/+ females. The analysis of Toll by Kathryn Anderson in my laboratory in Tübingen and subsequently in her own laboratory in Berkeley singled out Toll as a central component of the complex pathway regulating dorsoventral polarity and pattern of the Drosophila embryo. The Drosophila Toll story provides a striking example for the value of curiosity-driven research in providing fundamental insights that later gain strong impact on applied medical research. The Drosophila egg is a huge cell that is built during oogenesis in the female organism by two cell types, the somatic follicle cells surrounding the germ line-derived nurse cell–oocyte complex. Upon fertilization, it develops rapidly to form a larva by 24 h. In the first 3 h, the cleavage nuclei are not separated by cell membranes until there are about 6000 cleavage nuclei, which get included into cells forming the cellular blastoderm (Figure 1). The cuticle pattern of the Drosophila larva is richly endowed with markers for position and polarity (Figure 2A–C ), and a blastoderm fate map of the larval hypoderm has been constructed [1.Lohs-Schardin M. et al.A fate map for the larval epidermis of Drosophila melanogaster: localized cuticle defects following irradiation of the blastoderm with an ultraviolet laser microbeam.Dev. Biol. 1979; 73: 239-255Google Scholar] (Figure 2D). This pattern develops under the control of maternal genes that have been active during oogenesis and placed morphogenetic cues into the egg and the zygotic genes of the embryo itself. The long history of Drosophila genetics allowed systematic mutational approaches for the identification of maternal as well as zygotic genes affecting patterning of the larva by means of a visible phenotype displayed in the larval cuticle.Figure 2Cuticle pattern of a Drosophila larva.Show full caption(A) Ventral view displaying the prominent denticle bands of the abdomen and thorax. Dark field, anterior up. (B) Detail from the anterior abdomen and posterior thorax. Ventral view showing the prominent denticle belts marking the segment boundaries. (C) Dorsal view displaying a fine pattern of unpigmented hairs (phase contrast). (D) Fate map of the larval hypoderm at the blastoderm stage cut open along the dorsal midline and flattened. Midventral is the anlage of the mesoderm (me), flanked by the ventral hypoderm (vhy) and the dorsal hypoderm (dhy). The transverse lines indicate the position of the anlagen of the hypoderm, the thoracic segments t1–t3, and the abdominal segments a1–a8, as determined in laser irradiation experiments. The hypoderm developing from this region covers the entire larva from anterior to posterior. All other regions give rise to internal organs and are involuted during development. Abbreviations: aen, anterior endoderm; sto, stomodeum; cns, central nervous system; pro, proctodeum; pen, posterior endoderm. Reproduced, with permission, from [8.Nüsslein-Volhard C. Pattern mutants in Drosophila embryogenesis.in: Le Douarin N. Cell Lineage, Stem Cells and Cell Determination, INSERM Symposium No. 10. 1979: 69-82Google Scholar].View Large Image Figure ViewerDownload Hi-res image Download (PPT) (A) Ventral view displaying the prominent denticle bands of the abdomen and thorax. Dark field, anterior up. (B) Detail from the anterior abdomen and posterior thorax. Ventral view showing the prominent denticle belts marking the segment boundaries. (C) Dorsal view displaying a fine pattern of unpigmented hairs (phase contrast). (D) Fate map of the larval hypoderm at the blastoderm stage cut open along the dorsal midline and flattened. Midventral is the anlage of the mesoderm (me), flanked by the ventral hypoderm (vhy) and the dorsal hypoderm (dhy). The transverse lines indicate the position of the anlagen of the hypoderm, the thoracic segments t1–t3, and the abdominal segments a1–a8, as determined in laser irradiation experiments. The hypoderm developing from this region covers the entire larva from anterior to posterior. All other regions give rise to internal organs and are involuted during development. Abbreviations: aen, anterior endoderm; sto, stomodeum; cns, central nervous system; pro, proctodeum; pen, posterior endoderm. Reproduced, with permission, from [8.Nüsslein-Volhard C. Pattern mutants in Drosophila embryogenesis.in: Le Douarin N. Cell Lineage, Stem Cells and Cell Determination, INSERM Symposium No. 10. 1979: 69-82Google Scholar]. In the early 1970s, the first maternal mutants in Drosophila had been isolated, and transplantation experiments had supported the existence of ‘cytoplasmic determinants’ [2.Garen A. Gehring W. Repair of the lethal developmental defect in deep orange embryos of Drosophila by injection of normal egg cytoplasm.Proc. Natl. Acad. Sci. U. S. A. 1972; 69: 2982-2985Google Scholar, 3.Illmensee K. Mahowald A.P. Transplantation of posterior polar plasm in Drosophila. Induction of germ cells at the anterior pole of the egg.Proc. Natl. Acad. Sci. U. S. A. 1974; 71: 1016-1020Google Scholar, 4.Wright T.R. The genetics of embryogenesis in Drosophila.Adv. Genet. 1970; 15: 261-395Google Scholar], which reinforced the idea that a genetic approach to an understanding of development is feasible. Although experimental embryology, mostly done in ascidian, nematode, sea urchin, amphibian, and insect embryos, was a field with long tradition [5.Kühn A. Vorlesungen über Entwicklungsphysiologie.2nd edn. Springer-Verlag, Berlin Heidelberg1965Google Scholar], no molecules with morphogenetic roles had yet been identified in any system, and the postulated morphogen gradients, which would determine position in the developing embryo depending on their concentration, remained elusive [6.Wolpert L. Positional information and the spatial pattern of cellular differentiation.J. Theor. Biol. 1969; 25: 1-47Google Scholar]. I joined the laboratory of Walter Gehring at the Biozentrum in Basel in 1975 with the long-term goal to discover morphogens in Drosophila. My aim was to screen for maternal mutants with defects in larval patterning and use transplantations of wild-type cytoplasm into mutant embryos as an assay for the isolation of the gene products. In a small-scale mutagenesis experiment, I fortuitously isolated the mutant dorsal with a very exciting dose-dependent phenotype; embryos from dorsal/+-mutant females lacked ventral pattern elements, the anlage of the mesoderm, and parts of the ventral denticle belts (Figure 3C ), whereas the dorsal-recessive phenotype is a fully dorsalised embryo displaying dorsal tissue all around the embryo (Figure 3B). These dramatic phenotypes, which result from a complete change of the larval fate map as ventral egg regions develop structures normally derived from more lateral or even dorsal regions, specified dorsal as the first good maternal patterning mutant. The phenotypic series suggested a morphogen gradient with a maximum at the ventral side of the egg determining the dorsoventral axis (Figure 3D) [7.Nüsslein-Volhard C. Maternal effect mutations that alter the spatial coordinates of the embryo.in: Subtelny S. Konigsberg I.R. Determinants of Spatial Organization. Academic Press, 1979: 185-211Google Scholar, 8.Nüsslein-Volhard C. Pattern mutants in Drosophila embryogenesis.in: Le Douarin N. Cell Lineage, Stem Cells and Cell Determination, INSERM Symposium No. 10. 1979: 69-82Google Scholar, 9.Nüsslein-Volhard C. et al.A dorso-ventral shift of embryonic primordia in a new maternal-effect mutant of Drosophila.Nature. 1980; 283: 474-476Google Scholar]. Pedro Santamaria later demonstrated that transplantation of wild-type cytoplasm could indeed partially rescue the mutant phenotype [10.Santamaria P. Nüsslein-Volhard C. Partial rescue of dorsal, a maternal effect mutation affecting the dorso-ventral pattern of the Drosophila embryo, by the injection of wild-type cytoplasm.EMBO J. 1983; 2: 1695-1699Google Scholar]. By that time, however, the advent of gene technology and germ line transformation in Drosophila meant that the road to the gene product would be via cloning the gene and not via protein purification. My first independent position was at the European Molecular Biology Laboratory in Heidelberg, where I shared a small laboratory for 3 years with Eric Wieschaus, whom I had met in Basel. While developing screening protocols for maternal mutants, Eric Wieschaus and I also started collecting zygotic mutants affecting segmentation, which we obtained from various sources. The observed phenotypes were spectacular, and we decided first to do large-scale screens for more zygotic mutants, because they were easier than maternal screens and extremely rewarding. Indeed, these projects, in which we were later joined by an excellent geneticist, Gerd Jürgens, as a postdoc, resulted in a large and very exciting collection of patterning mutants [11.Nüsslein-Volhard C. Wieschaus E. Mutations affecting segment number and polarity in Drosophila.Nature. 1980; 287: 795-801Google Scholar,12.Wieschaus E. Nüsslein-Volhard C. The Heidelberg screen for pattern mutants of Drosophila: a personal account.Annu. Rev. Cell Dev. Biol. 2016; 32: 1-46Google Scholar], which provided much of the material basis for a new field of developmental genetics. We did not lose sight of the maternal mutants, however, and by chance we picked up three dominant Toll alleles by their singular ventralised phenotype (Figure 4), as well as two dominant cactus alleles and one easter allele, which shared its phenotype with dorsal. This was very encouraging, but the phenotypes were puzzling and at the time quite difficult to interpret. In September 1980, Kathryn Anderson applied as a postdoc in my laboratory. She was a student in the laboratory of Judith Lengyel at the University of California, Los Angeles, and graduated with a thesis on RNA synthesis during Drosophila embryogenesis [13.Anderson K.V. Lengyel J.A. Rates of synthesis of major classes of RNA in Drosophila embryos.Dev. Biol. 1979; 70: 217-231Google Scholar]. She must have been extremely courageous to apply to work with me, coming to Germany without knowing the language, leaving family and friends behind! I, who had worked on bacterial transcription for a thesis, was not yet well established in the field and had only published four papers on Drosophila (the screen with Eric and the rescue experiments with Pedro had not yet been published). Kathryn was fascinated by the dorsal mutation and the possibility to expand on the transplantation experiments. Although trained as a biochemist, she was keen on following a genetic approach to analyse development. She started in 1981 in my new laboratory at the Friedrich Miescher Laboratory (FML) of the Max Planck Society in Tübingen, with Gerd Jürgens as a research associate soon followed by Ruth Lehmann and Hans Georg Frohnhöfer as graduate students. At FML, we set out to do a large-scale screen for maternal mutants on the third chromosome of Drosophila in the hope to find more genes with properties similar to dorsal. Trudi Schüpbach and Eric Wieschaus, now in Princeton, screened the second chromosome, and screens on the first chromosome had been done by Madeleine Gans and coworkers in Gif-sur-Yvette. Our screen provided us with an overwhelmingly rich yield of exciting mutants: we isolated alleles of eight genes that shared the dorsalised phenotype, together with torso-like, oskar, pumilio, and bicoid. One striking result that became obvious when we exchanged information with the Princeton screen group was that there was a much smaller set of observed phenotypes than identified genes, and several of these shared a common or similar phenotype. Kathryn chose the ‘dorsal-group’ of mutants for further analysis, while Ruth focused on the ‘posterior-group’ of genes, such as oskar and pumilio, and Hans Georg worked on the mutants with head defects, bicoid and torso-like. The dorsal-group was the largest of the phenotype classes we identified. We found a recessive dorsalised phenotype in alleles of nudel, pipe, snake, easter, spätzle, pelle, tube, and also Toll [14.Anderson K.V. Nüsslein-Volhard C. Information for the dorsal-ventral pattern of the Drosophila embryo is stored as maternal mRNA.Nature. 1984; 311: 223-227Google Scholar]. In the Princeton screen, one more dorsal-group gene, windbeutel, as well as cactus, which as a recessive shared with Toll the ventralised phenotype, was discovered [15.Schüpbach T. Wieschaus E. Female sterile mutations on the second chromosome of Drosophila melanogaster. I. Maternal effect mutations.Genetics. 1989; 121: 101-117Google Scholar]. The mutant gastrulation defective (gd), which had been isolated already in the first chromosome screen [16.Gans M. et al.Isolation and characterization of sex-linked female-sterile mutants in Drosophila melanogaster.Genetics. 1975; 81: 683-704Google Scholar], also turned out to have a dorsalised phenotype. These common phenotypes suggested that each gene product is an essential component of a single system of positional information determining the entire dorsoventral axis. To understand how polarity and pattern in this axis is established, it was necessary to define the function of individual components of the system, to attribute them with ‘personalities’, as Kathryn put it. Kathryn’s transplantation experiments revealed that partial or complete normality of the pattern could be restored upon transplantation of wild-type cytoplasm or from embryos mutant for any of the other dorsal-group genes in mutant embryos for snake, easter, Toll, spätzle, pelle, and tube but not for gd and pipe (nudel could not be tested, as the eggs frequently were collapsed). In all cases, with the exception of dorsal, the rescuing principle included maternal mRNA, which presumably was translated upon injection to produce the protein products that would find their way in the process. The response to injection, however, varied considerably among the different genes; in the case of snake, Toll−, easter, and tube, a strong rescue could be obtained by injecting only about 1% of the egg volume, whereas in spätzle and dorsal, only few dorsolateral structures could be induced [14.Anderson K.V. Nüsslein-Volhard C. Information for the dorsal-ventral pattern of the Drosophila embryo is stored as maternal mRNA.Nature. 1984; 311: 223-227Google Scholar]. The transplantation assay provides a means of analysing the spatial distribution of, and requirement for, the products of the different dorsal-group genes. In contrast to the experiments on mutants affecting the anterior–posterior axis studied by Ruth and Hans Georg, who identified localised activities at the anterior and posterior pole of the egg [17.Nüsslein-Volhard C. et al.Determination of anteroposterior polarity in Drosophila.Science. 1987; 238: 1675-1681Google Scholar], no prelocalisation of the rescuing activity could be detected in the early embryo for any of the dorsal-group genes. The results of the transplantation experiment singled out two genes, dorsal and Toll. In the case of dorsal, cytoplasm from the ventral side of older preblastoderm embryos is more active than that from the dorsal side, and the cytoplasm has to be delivered at the ventral side to be active [10.Santamaria P. Nüsslein-Volhard C. Partial rescue of dorsal, a maternal effect mutation affecting the dorso-ventral pattern of the Drosophila embryo, by the injection of wild-type cytoplasm.EMBO J. 1983; 2: 1695-1699Google Scholar]. Supported by the strong dosage dependence, dorsal appeared the best candidate for encoding the morphogen determining the dorsoventral axis. At this time, dorsal was already being investigated by Ruth Steward, who I had met in Basel as a graduate student of Walter Gehring; she spent some time in the laboratory in Tübingen to isolate X-ray-induced dorsal alleles, which she used to isolate the dorsal gene by chromosomal walking in Princeton in Paul Schedl’s laboratory [18.Steward R. et al.Isolation of the dorsal locus of Drosophila.Nature. 1984; 311: 262-265Google Scholar,19.Steward R. Nüsslein-Volhard C. The genetics of the dorsal-Bicaudal-D region of Drosophila melanogaster.Genetics. 1986; 113: 665-678Google Scholar]. Kathryn found that for tube, snake, easter, spätzle, and pelle, the normalised patterns upon injection of wild-type cytoplasm always form in the normal orientation as given by the curvature of the egg, no matter where the cytoplasm is removed from the wild-type donors and no matter where it is injected. Toll was an exception; in the dorsalised embryos lacking the Toll product, a complete dorsoventral pattern can be locally induced by the injection of wild-type cytoplasm, and, in this case, the site of injection determines the ventral side of the embryo. This very clear response indicated that Toll− embryos lack any residual dorsoventral polarity [20.Anderson K.V. et al.Establishment of dorsal-ventral polarity in the Drosophila embryo: the induction of polarity by the Toll gene product.Cell. 1985; 42: 791-798Google Scholar]. Small amounts of cytoplasm were sufficient to produce a complete dorsoventral pattern, implying a vast excess of the Toll product in wild-type embryos. The rescue was spatially restricted to the site of injection, indicating that the Toll product had a limited ability to diffuse. Cytoplasm from both the dorsal and ventral sides of wild-type embryos was equally able to induce ventral structures when transplanted into Toll− embryos. Thus, in wild-type embryos, Toll at the dorsal side was inhibited, and this inhibition depended on Toll being active at the ventral side. Kathryn, in discussions with our next-door neighbour, the renowned theoretical biologist Hans Meinhardt, speculated that Toll was a component of a patterning system explaining the formation of morphogen gradients from weakly polarized starting conditions. This was consistent with the Gierer-Meinhardt model, which is based on local activation by autocatalysis coupled to long-range inhibition [21.Gierer A. Meinhardt H. A theory of biological pattern formation.Kybernetik. 1972; 12: 30-39Google Scholar], and predicted that a second activated site could not be induced in the same embryo. Siegfried Roth, a graduate student in my laboratory, later tested this by performing injections of wild-type cytoplasm into two sites of Toll− embryos and found, to his and Meinhardt’s dismay, that a second Toll activation peak could be easily induced at any distance to the first peak [22.Roth S. Mechanisms of dorsal-ventral axis determination in Drosophila embryos revealed by cytoplasmic transplantations.Development. 1993; 117: 1385-1396Google Scholar], ruling out this elegant interpretation. In a ‘tour de force’, Kathryn Anderson and Gerd Jürgens genetically analysed the properties of Toll [23.Anderson K.V. et al.Establishment of dorsal-ventral polarity in the Drosophila embryo: genetic studies on the role of the Toll gene product.Cell. 1985; 42: 779-789Google Scholar] and provided remarkably accurate insights into how Toll functions in the complete absence of molecular information. Toll alleles showed various phenotypes, a ventralised dominant phenotype but lateralised or dorsalised patterns in recessive alleles. It turned out that the first Toll alleles had already been isolated in 1973 in an unpublished screen by Tom Rice in the laboratory of Alan Garen at Yale (T. Rice, PhD thesis, Yale University, 1973), without recognition of these phenotypes. Thankfully, the mutants were still available, and Kathryn identified two recessive Toll alleles with lateralised phenotypes among them together with alleles of easter, tube, and pelle. The dominant Toll phenotype is characterised by an expansion of ventral pattern elements at the expense of dorsal pattern elements (Figure 4). This pattern aberration is caused by a gain of function of the gene, as revertants of the dominant sterility, which are caused by a second mutation inactivating the gene, display the dorsalised recessive phenotype. Surprisingly, some dominant alleles in trans with a null allele displayed a dorsalised null phenotype, which means that the dominant phenotype required the presence of a wild-type gene product. Kathryn and Gerd predicted that the Toll protein should dimerise, a prediction that was later confirmed at the biochemical level. The dominance is not due to overproduction, as duplications carrying an extra copy of the Toll gene partially suppressed the phenotype. Kathryn and Gerd proposed that Toll exists in an active and an inactive state, with the dominant alleles producing constitutively activated product that does not require activation. It was likely that some of the dorsal-group genes were involved in activating Toll, whereas others acted downstream and were induced by activated Toll to produce the ventral and lateral pattern elements. To determine which of the dorsal-group genes was required for Toll activation, double mutants of a dominant Toll with loss-of-function alleles of the dorsal-group genes were constructed. Creating double mutants with a dominant female sterile is a challenge, and as many of the dorsal-group genes including Toll are located on the third chromosome, recombination had to be induced by X-rays in males (in Drosophila there is no recombination during male meiosis). It turned out that combinations of dominant Toll with loss-of-function nudel, pipe, gd, snake, and easter alleles produced some lateral pattern elements, whereas Tl;dorsal double mutants were completely dorsalised. This was interpreted in a genetic cascade model with Toll in the centre, requiring activation via the upstream genes nudel, pipe, gd, snake, and easter, whereas dorsal was placed downstream, absolutely required to produce ventral and lateral pattern elements. The genes spätzle, pelle, and tube could not be tested as no recombinants could be obtained because they are very closely located to Toll; in later experiments in the Anderson laboratory, based on transplanting transcripts of dominant alleles, spätzle was placed upstream and pelle and tube downstream of Toll between Toll and dorsal. In retrospect, all these genetic parameters pointed to Toll as a membrane-bound receptor, but, at the time, there was still no paradigm for signalling cascades, and the roles of the other dorsal-group genes could not yet be distinguished. Kathryn presented this work at a Gordon Conference in 1985. Subsequently, she accepted an offer for an assistant professorship at Berkeley to clone Toll and the other dorsal-group genes that could be rescued in her own laboratory. In my laboratory in Tübingen, we focused on the ventralised mutant cactus, which we generously obtained from the Princeton group [15.Schüpbach T. Wieschaus E. Female sterile mutations on the second chromosome of Drosophila melanogaster. I. Maternal effect mutations.Genetics. 1989; 121: 101-117Google Scholar], as well as nudel, pipe, and gd that could not be rescued. At this time the first Drosophila patterning genes had been cloned, among them the bithorax complex and the first segmentation genes, which, excitingly, turned out to encode transcription factors. The first dorsal-group gene cloned was snake [24.DeLotto R. Spierer P. A gene required for the specification of dorsal-ventral pattern in Drosophila appears to encode a serine protease.Nature. 1986; 323: 688-692Google Scholar]. During their project to clone the bithorax complex of homeotic genes, Pierre Spierer in Dave Hogness’s laboratory had created a genomic walk across the rosy-Ace-region of the third chromosome, where snake happened to map. snake homologs in vertebrates suggested that it was a member of a protease cascade similar to blood clotting enzymes. A protease cascade might be involved in activating Toll by proteolytic cleavage. Excitingly, in 1987, dorsal was shown to encode a protein with homology to the avian oncogene v-rel [25.Steward R. dorsal, an embryonic polarity gene in Drosophila, is homologous to the vertebrate proto-oncogene, c-rel.Science. 1987; 238: 692-694Google Scholar], which later turned out to be a transcription factor with homology to NF-κB. Toll was cloned by the Anderson laboratory in 1988 [26.Hashimoto C. et al.The Toll gene of Drosophila, required for dorsal-ventral embryonic polarity, appears to encode a transmembrane protein.Cell. 1988; 52: 269-279Google Scholar]. The Anderson laboratory used a standard procedure for positional cloning of Drosophila genes by producing a Toll allele due to the insertion of a transposable P-element; in the case of Toll it was straightforward to screen for revertants of the dominant female sterility. The P-insertion site and adjacent sequences were cloned and used to screen for mRNAs. A long transcript of 5.3 kb was found to rescue the dorsalised phenotype of Toll alleles in the same polarity-inducing manner as cytoplasm from wild-type donor embryos. The amino acid sequence revealed a signal peptide as well as a membrane-spanning domain separating the transcript into an 803-amino acid extracellular domain and a 269-amino acid cytoplasmic domain. This suggested that Toll might be a receptor protein that was activated by a molecule present in the extracellular compartment surrounding the egg cell, the perivitelline space. At that time, only a limited repertoire of cloned genes, mostly from mammalian cell culture systems, was known. The leucine-rich sequences in the extracellular domain were similar to those in some membrane proteins, without much explanatory value. However, after this publication, a strong homology of the cytoplasmic domain of Toll with that of the interleukin receptor (IL-1R), later called TIR domain, which has an important role in immunity in mammals, was found [27.Gay N.J. Keith F.J. Drosophila Toll and IL-1 receptor.Nature. 1991; 351: 355-356Google Scholar]. In tissue culture cells, IL-1R activates nuclear uptake of the transcription factor NF-κB after its release from the inhibitor IκB [28.Ghosh S. Baltimore D. Activation in vitro of NF-κB by phosphorylation of its inhibitor IκB.Nature. 1990; 344: 678-682Google Scholar], thus suggesting an interesting homology of the Toll-Dorsal and the IL-1R-NF-κB pathway. In my laboratory, Siegfried Roth and Robert Geisler investigated the recessive ventralising cactus mutants, and cactus turned out, as expected, to encode the Drosophila homolog of IκB [29.Geisler R. et al.cactus, a gene involved in dorsoventral pattern formation of Drosophila, is related to the IκB gene family of vertebrates.Cell. 1992; 71: 613-621Google Scholar,30.Roth S. et al.cactus, a maternal gene required for proper formation of the dorsoventral morphogen gradient in Drosophila embryos.Development. 1991; 112: 371-388Google Scholar]. Double mutants of cactus with the dorsal-group genes indicated that none of them, with the exception of dorsal itself, abolishes the faculty in cactus mutants to produce lateral and ventral pattern elements, supporting the notion of cactus acting immediately upstream of dorsal, the latter encoding the morphogen. In the laboratory, Wolfgang Driever had just discovered the morphogen gradient of the transcription factor Bicoid, produced by the anteriorly localised mRNA that spreads posteriorly and determines the pattern in a concentration-dependent manner [31.Driever W. Nüsslein-Volhard C. The bicoid protein determines position in the Drosophila embryo in a concentration-dependent manner.Cell. 1988; 54: 95-104Google Scholar,32.Nüsslein-Volhard C. The bicoid morphogen