Maternally for Embryonic Dorsal-Ventral Pattern Formation

Maternal expression of the Toll gene is required for the production and the correct spatial organization of all lateral and ventral structures of the Drosophila embryo. We show here that the Toll gene is transcribed zygotically in the embryo and that zygotic expression is important for the viability of the larva. Both genetic and molecular data indicate that the zygotic Toll product has the same biochemical activity as the maternal product. The spatial distribution of the Toll transcript in the embryo was analyzed. In contrast to the uniform distribution of the maternal RNA, the zygotic Toll RNA is present in a complex spatial and temporal pattern in the embryo. A striking feature of this pattern is the correlation of the regions of invaginating cells with sites of accumulation of zygotic TOU RNA. S YSTEMATIC screens for mutations that alter the spatial pattern of the cuticle of the Drosophila larva have been remarkably successful in identifying gene products that play specific roles in embryonic pattern formation (N~SSLEIN-VOLHARD, WIESCHAUS and KLUDING 1984; JCRGENS et al. 1984; WIESCHAUS, NCSSLEIN-VOLHARD and JCRGENS 1984). Both zygotically and maternally expressed genes have been identified, and in many cases it has been possible to assign the gene a unique function in the generation of embryonic pattern (reviewed in SCOTT and O'FARRELL 1986; MAHOWALD and HARDY 1985). The Toll gene is one of these genes that has been shown to play a specific role in defining the spatial organization of the dorsal-ventral pattern in the early embryo. Toll was identified on the basis of a series of maternal effect alleles which cause dramatic changes in the mbryonic dorsal-ventral pattern, independent of the genotype of the fertilizing sperm (ANDERSON, JCRGENS and NCSSLEIN-VOLHARD 1985). Absence of Toll function in the mother causes a complete dorsalization of the embryo, such that cells at all positions of the embryo differentiate like cells in the dorsal regions of the wild-type embryo. Dominant gain-of-function alleles of Toll cause a ventralization of the embryonic pattern, with loss of the dorsal epidermal structures and expansion of the remaining ventrolateral pattern elements. These opposing phenotypes suggest that the Toll product is not only necessary for the development of lateral and ventral structures, but that its activity is also important in defining the position at which those structures are produced. The importance of the Toll product in defining the spatial organization of the dorsal-ventral pattern was most directly shown in phenotypic rescue Genetics 119: 123-133 (May, 1988). experiments in which Toll+ cytoplasm was injected into Tollembryos (ANDERSON, BOKLA and N~SSLEINVOLHARD 1985). In those experiments, the high local concentration of the Toll' product at the injection site defined the polarity of the entire embryonic dorsal-ventral pattern. These genetic and rescue studies established that the product of the maternally expressed Toll gene plays a specific and central role in early embryonic pattern formation. For the majority of maternal effect mutations, only single maternal effect alleles at a locus have been isolated, and these represent leaky, partial loss-offunction alleles of essential genes (PERRIMON et al. 1986). For those loci, like Toll, where several maternal effect alleles have been isolated, it has been supposed that those genes are expressed only in the female during oogenesis and that their only function is to promote normal embryonic development. A combination of genetic and molecular data, presented here, clearly demonstrates that the Toll gene is expressed and is functional zygotically as well as maternally. Almost all individuals homozygous for null mutations in Toll die as larvae despite the Toll+ genotype of the mother, indicating the importance of zygotic Toll product. To help define the zygotic function of Toll, we have examined the temporal and spatial pattern of accumulation of zygotic Toll RNA during embryogenesis. The distribution of the zygotic Toll transcript in the embryo is surprisingly complex. Intriguingly, the sites of zygotic Toll RNA accumulation are correlated with regions undergoing morphogenetic movements. This correlation suggests the hypothesis that the Toll protein may be a component of the cellular machinery that promotes morphogenetic movements. 124 S. Gerttula, Y . Jin and K. V. Anderson MATERIALS AND METHODS Mutant alleles: The maternal effect phenotypes of Toll alleles are described in Table 1. Most of the Toll alleles used in these studies have been previously described (ANDERSON, URGENS and N~SSLEIN-VOLHARD 1985a). The Df(3R)ros/b, isolated by P. LEWIS, deletes complementation groups both proximal and distal to Toll and is the standard Tolldeficiency used in these experiments. Three new alleles, TrB1, TrB2 and TlLB', were isolated in a screen designed to identify both lethal and maternal alleles of Toll and of tube, another dorsal group gene. Briefly, rucuca (tu h th st cu ST e" cu) males were fed 30 mM EMS according to the procedure of LEWIS and BACHER (1968) and then mated with tu stIIn(3LR) 361, DTS 4 th st Sb e females (MARSH 1978). Single F, ro en rucuca*/In(3LR) 561, DTS4 were mated with tub2' T?QR!st e cuITM3 flies, and the crosses were kept at 28" to kill off DTS4-bearing progeny. 5340 F2 lines were scored for lethal and maternal effect alleles. Potential lethal alleles of tube or Toll were identified on the basis of the absence of st e cu progeny. Lines with viable rucuca*Itub T1 progeny were put in egg-laying blocks, and embryonic phenotypes were scored 24-48 hr after egg laying in living, differentiated embryos (NCSSLEIN-VOLHARD 1977). No alleles of tube were recovered. One allele of Toll,TP', was recovered on the basis of lethality; on retest it was found to be semilethal. Although the maternal effect phenotype of T~B1/Df(3R)ro80b is strongly dorsalizing, TILB1 is not a null allele, because it has an antimorphic effect on the maternal effect henotype of other non-null alleles. Two Toll alleles, T$' and TPB2, were recovered on the basis of their recessive maternal phenotypes, which in each case results in a moderate dorsalization of the embryo. Marker mutations and balancer chromosomes are described in LINDSLEY and GRELL (1968). Viability measurements and temperature-sensitive period determination: The viability of flies carrying heteroallelic combinations of Toll alleles was determined relative to the viability of the Toll alleles when heterozygous with a balancer chromosome. That is, in the cross TPlbalancer X Tl%alancer, the number of TP/T16 progeny relative to the number of TPhalancer or Tl'lbalancer progeny was defined as the relative viability of TPlTl'. For those crosses in which it was possible to distinguish between the TPI balancer and the Tlblbalancer on the basis of markers, no significant differences between the viability of the two balancer classes was observed; that is, no dominant lethality of particular Toll alleles was observed. In other crosses, it was not possible to distinguish between these two classes, so it was assumed that they were equally viable. In addition, in the many reciprocal crosses carried out, no difference in viability was observed if a particular Toll allele was maternally or paternally contributed. The temperature sensitive period for lethality was determined by letting the parents lay eggs for 12 or 24 hr at the appropriate temperature, and then shifting the progeny in bottles up or down in temperature at the desired time after egg laying. Progeny were not staged by morphological criteria. RNA analyses: Total embryonic poly (A)'RNA was purified and assayed by blot hybridization using standard procedures (MANIATIS, FRISCH and SAMBROOK 1982). The hybridization probe for the Toll RNA was the 1.8-kb EcoRI fragment at the 5' end of the Toll transcript (HASHIMOTO, HUDSON and ANDERSON 1988). Polysomal RNAs were prepared as previously described (ANDERSON and LENGYEL 1981), except that in order to prevent pelleting of the polysomes containing Toll RNA it was necessary to add a b c d e f g FIGURE 1.-The amount of Toll RNA as a function of embryonic age. Poly(A)+RNA was prepared from staged embryos, electrophoresed in an agarose gel, transferred to nitrocellulose and probed with the 1.8-kb EcoRI fragment of the T p l clone (see MATERIALS ASD METHODS) that hybridizes to the 5.3-kb Toll RNA (uppm b u d ) and with actin DNA (lower b u d ) . The actin signal comes from a reprobing of the blot with actin DNA, followed by a shorter exposure time than that needed to see a strong Toll signal, and was photographically superimposed here, so the relative signal intensities do not reflect the relative abundance of the two RNAs. a; 0-3-hr embryos; b, 3-6 hr; c, 6-9 h; d, 9-12 hr; e, 1215 hr; f, 15-18 hr; g, 18-21 hr (age in hours at 22"). Triton X-100 to 0.5% and deoxycholate to 0.5% to the homogenate before it was layered on the sucrose gradient. In situ hybridization to embryonic tissues: In situ hybridizations were carried out essentially according to the method of MAHONEY and LENGYEL (1987). Briefly, embryos were fixed in paraformaldehyde, and the whole embryos were hybridized with '%-labeled SP6 antisense transcript from the 2.4-kb EcoRI fragment of the Toll transcription unit (HASHIMOTO, HUDSON and ANDERSON 1988) cloned into pGeml (Promega). After washing to remove unhybridized label, embryos were embedded in methacrylate, 2p, sections were cut, plastic was removed with xylene, and slides were dipped in 1 : 1 diluted Kodak NTB2 emulsion. Slides were exposed for 7-14 days, developed and stained with Giemsa. No signal was detected in control hybridizations with "Slabeled sense RNA transcribed from the T7 promoter of pGem 1.