oroshigane, a New Segment Polarity Gene of Drosophila ... - NCBI

4 downloads 10610 Views 9MB Size Report
HARD et al. 1984) where ... pressed in the posterior compartment of the disc, pro- motes the .... patterns of dead embryos in the resultant stocks. In both ... day of recovery, the males were mated with females of the. B"" stock ..... The size of the eye is less than one-half of that of R' heterozygotes (data not shown). There are.
Copyright 0 1997 by the Genetics Society of America

oroshigane, a New Segment Polarity Gene of Drosophila mlanogmter, Functions in Hedgehog Signal Transduction Janet L. Epps, Jessa B. Jones* and Soichi Tanda* Molecular and Cell Biology Graduate Program and *Department of Zoology, University of Maryland, College Park, Maryland 20742 Manuscript received June 18, 1996 Accepted for publication January 8, 1997

ABSTRACT Here we describe a new segment polarity gene of Drosophila melanogaster, mxhigane ( m o ) . Identified as a dominant enhancerof Bar ( B ) , oro is also recessive embryonic lethal, and homozygous oro embryos show variable substitution of naked cuticle with denticles. These patterns are distinctly similar to those of hedgehog ( h h ) and wingless ( w g ) embryos, which indicates that oro functions in determining embryonic segment polarity. Evidence that or0 function is involved in Hh signal transduction during embryogenesis is provided byits genetic interactions with the segment polarity genes patched ( p t c ) and fused ( f u ) . Furthermore,pt8' is a dominant suppressor of the oro embryonic lethal phenotype, suggesting a close and dose-dependent relationship between ora and ptc in Hhsignal transduction. oro function is also required in imaginal development. The oro' allele significantly reducesdecapentaplepc ( d p p ) ,but not hh, expression in the eye imaginal disc. Furthermore, oro enhances the fu' wing phenotype in a dominant manner. Based upon the interactions of ora with hh, ptc, and fu, we propose that the or0 geneplays important roles in Hh signal transduction.

C

OMMUNICATION between neighboring cells is one of the most important processes necessary for undifferentiated cells to acquire appropriate developmental fates and is common among many multicellular organisms ranging from insects to mammals. In Drosophila,communication between cells abutting each other is very important during embryogenesis as well as in imaginal development. One such communication system in Drosophila is Hh signaling. hh and several ofits downstream components were identified in D. melanogaster during a large scale screen for mutations and causing embryonic lethality (NUSSLEIN-VOLHARD WIESCHAUSE 1980; JORGENS et al. 1984; NCJSSLEIN-VOLHARD et al. 1984) where they were shown to function in segment polarity determination. The hh gene was cloned, and its deduced protein sequence suggested that it is a secreted protein, as subsequently confirmed by antibody staining (LEEet al. 1992; MOHLERand VANI et al. 1992; TASHIROet al. 1993). Target 1992; TABATA genes for hh signal transduction are thought to be zug and ptc inthe embryos, and wg, dpp and ptc in the imaginal discs. Recent genetic and molecular studies of Hh signaling have established a basic molecular structure of this signal transduction pathway (reviewed in KORNBERGand TABATA1993). The list ofitsdownstream components consistsof ptc, smoothened ( s m o ) , fused Suppressor offused [Sulfu)], costal (cos), cubitus interruptus (ci), and CAMP-dependent protein kinase A (PKA). A striking outcome of these molecular studies

vu),

Corresponding author: Soichi Tanda, Department of Zoology, University of Maryland, College Park, MD 20742. E-mail: [email protected] Genetics 145: 1041-1052 (April, 1997)

is that most components in Hh signaling are conserved in evolution (INGHAM 1995). ptc and ci have been shown SCOTT to possess vertebrate homologues (HOOPER and 1989; NAKANOet al. 1989; ORENIC et al. 1990; GOODRICH et al. 1996), and Fu and PKA are serine/threonine kinases. In the case ofptc, expression patterns and responsiveness to Hh suggest its functional conservation from Drosophila to vertebrates (GOODRICH et al. 1996). This indicates that the Hh signal transduction pathway is a universal component in development in a wide variety of eukaryotes from insects to mammals. In Drosophila, Hh signal transduction plays important roles both in embryogenesis and in imaginal development (reviewed in DINARDOet al. 1994; PERRIMON 1994, 1995).The molecular mechanism of Hh signaling seems to be similar in both systems. However there are two major differences between these systems. First of all, there seems to be no reciprocal regulation loops involved in the stabilization of hh expression and its downstream targets across the anterior and posterior compartment border inimaginal discs at this moment. Secondly, dpp also plays an important role as aHhdownstream target(HEBERLEINet al. 1993; MA et ai. 1993; BASLERand STRUHL 1994; TABATA and KORNBERC 1994). In the leg disc, the Hh protein, which is expressed in the posterior compartment of the disc, promotes the expression of wg in the anterior-ventral section and dpp in the anterior compartment along the anterior-posterior compartment border (A/P border). These two genesareimportantfor establishing the proximal-distal axis in the leg discs by promoting the expression of Distal-less and aristaless (CAMPBELL et al. 1993;DIM-BENJUMEA et al. 1993; COHENet al. 1994).

1042

J. L. Epps, J. B. and Jones

Hh activates dpp expression in the anterior compartment along the A/P border of the wing disc, where dfip is thought to act as a long range morphogen that organizes the discs' gross architecture and epidermal fates (e.g., ZECCA et al. 1995). In these cases, hh expression appears to depend solely on engrailed and inuerted (TABATA et al. 1992). In the eye imaginal disc, hh regulates the expression of d# in the morphogeneticfurrow (MF), which is thought to be equivalent to the A/P border in the wing and leg discs, to maintain the progression of eye morphogenesis (HEBERLEIN et al. 1993; MA et al. 1993). hh is expressed in differentiating photoreceptors behind the MF, and the cells in the MF respolld to the Hh signal by expressing dpp, which in turn triggers events that contribute to continuing eye morphogenesis in a proper manner. The molecular mechanism of Hh signal transduction seems to be complex. ptc, cos and PKA have negative effects on Hh signaling while smv, ,fu, and ci have positive effects. Although there areseveral different models for how these components interact with one another to transmit the Hh signal, ptr, smo, fu, and r i have been proposed to work in alinear pathway, according to epistasisanalysis in the embryo (FORBES et ut. 1993), and Su(fu)and cos likely function with fu and ci. As in embryogenesis, ptc represses wg and dpp expression in anterior regions of these imaginal discs (INGHAMand HIDALGO 1993; CAPDEVIIA et d . 1994; MA and MOSES 1995;WEHRLI andTOMLINSON1995). PKA, like ptr, functions as a constitutive negative factor for u g and dpf expression in the anterior compartment andin the region anterior to the MF (JIANG and STRUHI.1995; LEPAGE et al. 1995; LI et al. 1995; PANand RUBIN1995; STRUTTet al. 1995). Most recent studies in Drosophila and mammals have shown that the Ptc protein physically binds to Hh (CHEN andSTRUHI.1996; MARIGOet al. 1996; STONEet al. 1996). Smo, a G protein-linked receptor (ALCEDO et al. 1996; VAN DEN HEWEL and INGHAM 1996), islikely a signaling componentthat functions immediately downstream of Ptc. There is, however, no definitive and conclusive scenario of how the Hh signal is transmitted into the nucleus. To gain more insight into Hhsignal transduction, one approach is to use genetics to identify new genes in the Hhsignaling pathway using a sensitized system and to understand how newly identified genes function in Hh signaling. This study is a report of such an attempt in which we used the Bar ( B ) mutant background as a sensitized system for mutagenesis. We identified a new gene oroshigane ( o w ) initially asan enhancer of B. Further analysis showed that vro behaves as a segmentpolarity gene functioning in Hh signal transduction during embryogenesisaswellas imaginal development. vro is a recessive lethal mutation and enhances multiple B alleles and the hh' allele in ahaploinsufficient manner. Its recessive embryonic phenotype resembles those of wgtype segment polarity genes. Like other segment polarity genes,

S. Tanda

orofunction is required in imaginal development as well as embryogenesis. Strong genetic interactions with ptc and fu suggest that vro functions primarily in Hh signal transduction. MATERIALS AND METHODS

Fly stocks and culture conditions: Flies were raised in uncrowded conditions on a standard cornmeal, yeast, molasses and agar medium at -22" unless otherwise indicated. Phenotypes of mutations are described in LINDSLEY and ZIMM (1992) or FlyBase. A set of second chromosome deficiencies and PZ element-induced lethals were obtained from the Bloomington Stock Center. The following mutant stocks were used in the present study: w$ rn' b7u' W $ ' ~ ' ~ ~ ~ " ' / hh""/CyO, C~O, hh'", hh', pt(\/CyO, w$- l i 6' pr / CyO, smu' cn' bw' s '/CyO, smo' 6' pd/CyO, cn Tn(H1 - 1); ryiOh (BS3.0), Injl)B"f."j:In(l]fl5,fu'/ ClB, 7 u ' I f S P{ry+ = hsF12P}1;Adv'/CyO, P{ry+ = neoFK7)40A; ryxl',, y { w + m'.' = o v o l ~ l - ' " ~ I , l p/w+"'(.= ouol)l-lX121.2 I ' l r > + l i 2 - new FRT}40A/llp(3;2)bd', S'Sp'Ms(Z)M'bJ'/CyO. A heat inducible Om(1D)gene of D.ananassae, the B homologue, was constructed (TANDA and C0Rc:es 1991). We refer to this transgene as a h d construct (or transgene) for simplicity in this study since the ananassue transgene works the same as the D. melanogaster B transgene in the D. melanogaster genome (KOIIMAet al. 1991; TANDA and CORCES1991). The presence of multiple mutations in asingle second chromosome was confirmed by complementation tests using appropriate mutant alleles. Itshouldbe mentioned that the presence of oro' and pt?" alleles on a single second chromosome was confirmed by separating these alleles onto individual chromosomes, followed by complementation tests using the parental oro' and pt8' alleles and examination of cuticle patterns of dead embryos in the resultant stocks. In both cases, the results indicated that the oro' pt8'chromosome that we created does carry these two alleles. oru' homozygous female germline clones were generated by using an ouo" transgene on thesecond chromosome (CHOU and PERRIMON 1992; SIEGFRIED et al. 1994) and the FLP/FRT system (Xu and RL~BIN 1993). Females of the 7 ~ ' 'P{?+ ~ ~ = hsF'LP/ 1; A d d / Cy0 stock were crossed with P/w+'"'' = ouol'l.'s )2I*Jp{w+'"': = o u o I ) I ~ I . Y )2L2 P{ry+'" = neoFRT/40A/CyO males segregating-in the P/7u+'"' = ova"'-")2LI P{ru""" = ovd)'.'N }2L2 P{?+""' = n~oE'K7')40A/Up(3;2)b7J',S'Sp'M,s(2)M'bd'/ Cy0 stock. At the next generation, 7~1'"'~P{ry+ = fzsFLP)I / Y; uUol)l-lH 1 2 1 ~ P/zu+"I(~= oud" ' " ) ~ L P{?+'" z = n w PIwi."" = FKf'{4OA/f$O males were selected and mated with females of the oro' P{ry' = neoFRT/40A/CyO stock. The animals in the last cross were heat shocked at 37" for 2 hr three times during the third instar larval sta e U on their eclosion, 7 ~ " ' ~ P{v+= hsFLP}l/+; ora' P{$'.'' = p n e o ~ ~ 7 { 4 0 A / P / , o t =" : o u o i ~ i - l x)2LI P{7U+""'= ov0""")2L2 P{vy+"' = nroFRT/40A fe-

y',

males were collected and mated with oro'/CyO males. Unhatched embryos from these females were examinedafter being cleared with the Hoyer's mount. In this condition, most control females [ w i l l xP{ry+ = hsFU)l/+; oro+P{~y+"~~ = neoFR"fOA/p(w'"'' = o~o1'""J2I,1P { ~ ~ + '=" 'o7id"~"}2L2 : p{ry+" = neoFK7)40A] became fertile to produce the eggs. Muta enesis: Three-to five-day-oldvirgin males of the 5 In(1)Y" stock were fed 5 mM diepoxybutane over Zn(l)R"", night (OI.SENand GREEN1982) after 8-hr starvation. After 1 day of recovery, the males were mated with females of the B"" stock and progeny were collected for 3 days. They were raised at 18" to enhance the B'"phenotype and were screened for enhancers or suppressers of this B phenotype. Chromosomes derived from selected GI individuals, which were selected for enhanced B " ~phenotypes for several generations,

Transduction 1043Signal Hh

or0 in Function

were established as stocks later using the Cy0 balancer chromosome. The p 2allele is associated with a heterochromatic inversion (SUTTON1943),so that there is one B gene instead of two copies as in B’. Thus, this screening was also useful for isolating recessive B alleles as well. Histochemistry: For0-galactosidase (0-gal) staining, late third instar larvae were dissectedand the imaginal discs were collected in ice cold PBS until we finished collecting a sufficient number ofdiscs. The discswere then fixedin 0.1% glutaraldehyde in PBS on ice for 3 min and washed with PBS several times. @-galstaining was performed according to the protocol describedin ASHBURNER (1989).To enhance weaker 0-gal staining in some cases,the staining solutionwas concentrated three times the normal concentration used. After staining, the discswererinsedwithPBS and mounted in 100% glycerol. The specimen were observed with DIC optics. Cobalt sulfide staining was done by the methodof MELAMED and TRUJILLO-CENOZ (1975).Larvae carrying a hs-Om(ID) (= a hs-B) transgene (TANDA and CORCES1991)were heat treated and kept at-22“ at 37” for 2 hr at late third instar larval stage for the next10 hr. The treated larvae were dissected toobtain the eye-antennal imaginal discs forcobaltsulfidestaining. The stained discswere mounted in water and immediately observed. Embryonic cuticlepreparation: The embryoswerecollected every 24 hr and cultured for an additional 24 hr. Unhatched embryoswere collected, dechorionated with 50% bleach and washed with PBS. The embryos were then fixed in a 1:4 solutionof glycerol and acetic acid for 30 min at 65” and mounted in Hoyer’smount. The embryos wereincubated at 65” until cleared.The specimens were observed with phase contrast for dark field and DIC for bright field. For each genotype, 500-1000 embryos were examined. Scanning electron microscopy:The fly heads were fixed in 2% glutaraldehyde in 0.12 M phosphate buffer (pH. 7.2) at room temperature. After several washes with the phosphate buffer, the headswere further fixed in 1% O s 0 4 in the phosphate buffer,which was followed by a brief rinse with ddH20. After dehydration with an ethanol series, the specimens were dried by a critical point drying methodand coated with gold (60%)-palladium(40%).They were observed with an Amray 1820D scanning electron microscope. RESULTS

Ectopic Bar expression stops progression of the morphogenetic furrow: The B’ allele is associated with a tandem duplication of the division 16A of the X chromosome (SUTTON1943), is neomorphic by definition (MULLER 1932), and causes the overexpression of the B gene (TANDA and CORCES1991). Ithas been reported that dpp expression in the MF is abolishedin the B background, and therefore that MF progressionhas prematurely ceased resulting in its characteristic barshaped compound eye (HEBERLEIN et al. 1993). However, this hypothesis has not been directly tested. We wanted to test if the B overexpression is sufficient to stop MF progression and to exploreits molecular mechanism, becausethe B mutant backgroundwould be convenient and useful to search for genes involved in MF progression if the B protein directly prevents the MF from advancing anteriorly. We used a hs-B transgene (see MATERIALS AND METHODS for detailed information) to expose eye imaginal

discs to excess B protein. The eye-antennalimaginal discs were dissected from late third instar larvae 10 h r after the end of heat-shock treatment and stainedwith cobalt sulfide tovisualize the MF. Strikingly, no MF was detected in eye imaginal discs that have accumulated the B protein, while a clear MF was visible in wild-type discs (Figure 1, A and B ) , indicating that the overexpressed B protein prevents MF progression. Furthermore, the staining patternof the most anterior clusters of photoreceptors in the &treated disc are similar to those of 10-hr-old photoreceptor clusters, which indicates that no new ommatidial clusters had been initiated for -8 hr in the discs overexpressing B protein. This observation also suggests that the inhibitory effect of the overexpressed B protein is apparent soon after the protein is induced. At the same time, the presence of these more maturedpreclusters at the normaldevelopmental schedule indicates that the overexpressed B protein has little deleterious effect on differentiation of photoreceptor clusters that have already started to develop behind the MF, which coincides with the previous report (Figure 10, D and E in TANDAand CORCES 1991). Failure of MF progression likely triggersprogrammed cell death in the undifferentiatedcells ahead of the MF (TANDAet al. 1993; TANDA et al. 1994). To explore the mechanism of the cessation of MF progression caused by the overexpressed B protein, hh and dpp expression was followed as hh-lac2 (hh”30,LEE et al. 1992) and dpp-lacZ (dpp”’) expression patterns (Figure 1) afterthe B protein was inducedinthird instar larvae carryinga hs-B transgene by heat-shock treatment. hh-lac2 expression first began to fade 8 hr after the end of heat induction of the B protein and reached the lowest levels in the region just behind the MF (Figure 1D). dpp-ZacZexpressionstarted to decrease in the MF 10 h r after the treatment (Figure 1F). This result indicates that hh-ladexpressionis another target for B inhibition in addition to that of dpp-lac2 expression. SinceP-galactosidase (P-gal) stays active for a fairly long time, the above time course of the B inhibition of hh and dpp expression is likely to be artificial to some extent, which may explain the difference between timing of the cessation of MF progression and of abolishment of hh- and dpplucz expression. Although the functionof B, which encodes a homeoand CORCES domain protein (KOJIMAet al. 1991; TANDA 1991), is required for R1, R6 and primary pigment cell differentiation behind theMF (HIGASHIJIMA et al. 1992), the overexpressed B protein in the eye imaginal disc apparently interferes with MF progression by inhibiting dpp expression in the MF and hh expression just behind the MF. As expected, the adulteye phenotype of hh’, a hypomorphic hh allele, isvery similar tothose of U mutations (Figure 5B), which indicates that theB mutations produce phenotypes in eye imaginal discs that resemble those of partial loss-of-function hh mutations. Therefore, the B mutant backgrounds provide us with

1044

J . L. Epps, .J. B. Jones andS. Tanda

r-

1 '

1

~

1-

.

..

-

- . ._l

..,

._

.-.,..",

--/r

..'$ . . .-. .

.

-':.;.", .,,

-

FIGI~RE 1.-Effects of the overexpressed Bar protein on morphogenetic furrow progression i n the eye imaginal disc. Anterior is to the left, and dorsal is to the top. Approximate locations of the morphogenetic furrow are indicated by arrows. (A) Wild-type eve imaginal disc stained with cobalt sulfide. The morphogenetic furrow is clearly seen as a vertical stripe of cells with narrow apical surfaces. (€3) C o b a l ts u l f i d e stained eve imaginal disc that was fixed 1 0 hr after the end of heat induction of the B protein. Note thatthere is no morphogenetic fL1rrow. Filled triangles in A and I3 indicate photorecep tor preclusters with similar morphology, which means that they areat similar developmental stages during eye morphogenesis. (C) Ith expression (hh")"j inwildtypeeye imaginal disc. ith is expressed just behind the morphogenetic furrow in developing photoreceptor preclusters. (D) Ith expression in B treated eye imaginal disc. Overall hh expression includingtheantennal disc and ocelli region is suppressed,but the region justbehind the morphogenetic furrow is most afrected. (E) d j expression ~ ~ (dpp"'552) in wild-type eyeimaginal disc. dpp is expressed in or just behind the morphogenetic furrow and the edges of the eye disc, which is characteristic to this dpp/mZIine. (F) dpp expression in Rtreated eye imaginal disc. Note that there is no dpp expression in the morphogenetic furrow, except lor slightly higher background staining in the regionposterior tothe firrow.

FIGL‘KK ‘L.”Dcnticlc belt patterns of wild-type, / A , and om embryos. Anterior is t o the top. ( A ) Denticle I d t s i n the abdominal segments of a wild-type embryo. Note a periodic pattern o f denticle belts and naked crlticle. (B) Pattern o f a / I / / ” ” embryo. There is n o obvious naked cuticle and most abdominal segments are covered w i t h tlcnticles instead. ( C ) The most extreme cuticle pattern of an o m ’ embryo.Note that the pattern is very similar to that of hh”” shown in B. In addition, there are orohomozvgous embryos with some other cuticle patterns. About one-half of or0 homozygotes show cuticle patterns, and I O t o 1.5% of tl;kse embryos show rug (or ///+)-likecuticie patterns.

a suitable genetic background foridentifjring genes involved in MF progression, specifically the ones which ftmction in Hh signal transduction. Enhancers of Barwere isolatedby diepoxybutane mutagenesis: To identi5 genes that are involved in MF progression, we conducted diepoxybutane mutagenesis. All F, progeny were screened for enhanced orsuppressed 13.”’ eye phenotypes. I?”’homozygotesshow only a slight indentation at the anterior edge of the eve, therefore this mutant phenotype turned out to be especially suitahle to find enhancers of B. Four dominant enhancers of Rnrwere found in 10,321 oK$pring from the treated I?”’males. While one of the identified enhancers was lost before detailed analysis, the other threemutations wereestablishedasstocks and subjected to genetic analvsis. The three isolated enhancers of B are all recessive lethal, dominantly enhance multiple B alleles and hh’ (Figure 5, C and D) and linked to the second chromosome. Two of them fail to complement each other and were named oroshiganr (om: oro‘ and ora') because their embryonic cuticle lethal phenotypes are very similar to those of hh and 7ug (Figure 2) and resemble the surface of an “oroshigane”, Japanese for a grater. The third enhancer of 11 was shown by complementation tests to be an allele of S/nr. Neither oro allele shows any eye morphology defect when heterozygous with a wild-type chromosome, but theoro’ allele shows stronger enhancing effect on 13 than does the oro’ allele. Both alleles enhance the hh’ mutant eye phenotype, which suggests a general involvement in MF progression rather than a spccific interaction with B.

The recessive lethal phenotypes of om‘ and om’ map to the region uncovered in the deficiency Df(2L)aZand fail tocomplementthe lethality of the deficiencv Df(Z1,)aZ. In addition, the deficiency 1?f(2I2)a1is an apparent strong dominant enhancer of I3 and combinations of these oro alleles with the deficiency IIJ(Z1,)aZ result in more frequent occurrence of stronger segment polarity lethal phenotypes. These results suggest that the oro gene is located in the deficiency Df(Zl.)aZ, that oro is a loss-of-function mutation enhancing I3 phenotypes in a dominant manner, and that both oro alleles are likely hypomorphic. m is a new segmentpolarity gene: To infer thefunction of oro, we examined cuticle phenotypes of homozygous oro embryos. Both oro’ and MO’ homozygous embryos show denticle belt patters very similar to those of hh and 7ug alleles, but varying in severity so that a large portion of oro’ and om2 embryos appear wild type (- 1015% of oro homozygotesshow hklike cuticle phentr type). In the most extreme cases, om’ embryos lookjust like those of theIZIL?’~’ allele, which we used as an example of a strong hh allele (Figure 2).In addition to these 7ugtype cuticlephenotypes,some embryos show no head structure, which is also seen in a 7ug null allele, 7ug’-’’ (formally 7ug‘”). Some alsoindicatefailure of germband retraction, which suggest that oro function is also required in later stages of embryogenesis.The low penetrance of zugtype cuticle patterns could be due to maternal contributionof orofunction, but this maternal contribution may not be enoughto compensate lack of o m function at later stages of embryogenesis, resulting in lethality of oro- homozygotes. Alternatively, oro may

1046

J. L. Epps, J. and B. Jones

have some redundant function that is shared with other genes. To test whether there is a maternal contribution of the wo function, we created wo- female germline clones (see MATERIALSAND METHODS for detail).We did not observe any significant maternal contribution of wo function to the oro homozygous embryonic lethal phenotypes although the number of the embryos observedwere small. The oro- female germline clones were difficult to create and the rareoro- eggs weremore spherical than the wild-typeeggs. The cuticle phenotypes of the embryos from such oro- eggs varied from wildtype to mildly affected segment polarity phenotypes. This observation suggests that variable oro- phenotypes are not attributed to a strong maternal contribution of the or0 function. Furthermore, the rare woegg production and the abnormal shape of those eggs also suggest a role of the oro gene during oogenesis, which may be supported by the fact that the treated oro'/ovo" heterozygous females (possibly carrying ora'/ wo' germline) often accumulate necrotic nuclei in the egg chambers at the early (2-4) stages. The embryonic phenotypes of oro' and wo2 strongly suggest that wo is a segment polarity gene. Since both oro alleles are likely hypomorphic and map in the deficiency Df(2L)nZ,it is possible that oro might be an allele of smo, a segment polarity gene, which also maps near the deficiency Qf(2L)nl (LINDSLEYand ZIMM 1992). However, the two smo alleles examined, smo' and smo3, showed only weak dominant enhancing effects on B mutations. To further examine this possibility, we performed complementationtests oftwo oro alleles with the above smoalleles. In allfour combinations, the expected number of oro/smo animals survived to the adult stage, indicatingcompletecomplementation. The progeny from these crosses show a typical Mendelian segregation for the Cy wing phenotype; the ratio of CJLwing versus non-Cywing flies was 2:l. Further evidence that smo and wo are distinct genes is indicated by the response to temperature. Since smo is reported as a coldsensitive mutation (LINDSLEYand ZIMM 1992), we tested at two different temperatures, 18" and 25". The results were the same at both temperatures. In contrast to smo, there is no evident temperature effect of oro alleles with respect to embryonic cuticle phenotypes. Furthermore, jltc is epistatic to wo (see below) while smo is epistatic to ptc (LINDSLEYand ZIMM1992; FlyBase; Figure 3, A and B). These lines of evidence clearly indicate that oro is not allelic to smo, and that wo is a new segment polarity gene. ptc is epistatic to mo and acts as a dominant suppressor of the o m lethality: To determine which pathway (Hh or Wg) wo functions, we first examined the relationship between oro and ptc. For this purpose, we combined the wo' allele with the p~d"allele. The existence of these alleles on a single second chromosome was verified bytwo different methods (see MATERIALS AND

S. Tanda

FIGURE3.-Epistatic relationships among o m , ptc, and smo. Anterior is to the top. (A) Denticle belt pattern of the smo' allele raised at 18". This is a moderate pattern of this allele, and a typical wgqpe segment polarity pattern is seen in this example. (B) Denticle pattern of the smo' pt8" doubly homozygous embryo. An obvious fused denticle belt pattern is seen, which indicates that smo is epistatic to ptc. (C) Denticle pattern of a pt8"embryo. ptcembryos are characterized by duplication of the first and second rows of denticles. (D)Denticle pattern of the OTO' pt8" doubly homozygous embryo. Note that the pattern is not different from that of the ptr mutation alone, indicating that ptc is epistatic to oro. METHODS). The embryos homozygous for both mutations die andshow a ventral cuticle pattern identical to that of the ptd'homozygous embryos (Figure 3, C and D), which is based on the observation of -1000 doubly mutant embryos. This relationship is the same as that of ptc and hh, but opposite to that of rug and ptc, where 7ug is epistatic to ptc (Bgsov~c and WIESCHAUS 1993). This observation implies that wo function is not essential for Wg signal transduction since rug is responsible for producing naked cuticle. oro should be epistatic to plc if wofunction is required in M'g signaling to generate naked cuticle. Therefore, the wo' Pt? cuticle phenotype suggests that wo works in Hh signal transduction, and that wo acts upstream of p t c in a possible scenario where hh, plc, and oro function in a linear Hh pathway during embryogenesis. While the embryos doubly homozygous for wo' and ptd' die, many wo' ptd'/wo' + embryos develop to the adult stage. This is also the casewhen the wo' fit?''

or0 Function in Hh Signal Transduction

chromosome is combined with the oro' allele and the deficiency Df(2L)nl. Thus, lethality of the homozygous oro' allele can be rescued to a great extent by reduction of ptc activity. Combinations of oro', oro', and Df(2L)al with the oro' pt8' chromosome resulted in 96, 81 and 46% viability, respectively.ptc functions as an inhibitor of 7ugexpression in the process of parasegment border establishment. The input from Hh signal transduction opposes Ptc inhibition in the cells just anterior to the hh expressing cells. In the oro- background, Hh signal transduction is likely to be greatly reduced. A reduction of Ptc activity may restore the balance of Ptc and Hh signaling input, which results in promoting enough wg expression in the cells at the posterior end of parasegments. As a result, many oro' pt("'im-0- + embryos survive to adults. This result further emphasizes the strong and specific relationship between or0 and ptc. m works downstream of hh, but upstream of dplb to enhance the B phenotypes: To obtain more evidence to support that or0 functions in Hh signal transduction, we examined how oro mutations enhance the B phenotype with respect to d m and hh expression in the eye imaginal disc. This may also provide some clue to help us infer oro function in eye morphogenesis, specifically in MF progression since B mutations suppress hh and dpp expression in the eye imaginal disc (Figure 1). Therefore, we examined the expression of dpp and hh in the eye imaginal disc. In the B heterozygous background, dpp expression is abolished in the middle of the eye imaginal disc (or a half way through MF progression). In a similar manner, dpp expression was shown to be abolished in B / + ; ora'/ + eye imaginal discs, but ata much earliertime period of MF progression (Figure 4D),which suggests that or0 is involved in positively regulating dpp expression in the eye imaginal disc. Furthermore, dpp expression was also observed to be decreased to an even greater extent in B+; oro' heterozygotes (Figure 4B). dpp expression in the eye imaginal disc ofthe oro' heterozygotes is approximately one-half of that in wild type, which appears to be sufficient to support MF progression to the anterior end of the eye imaginal disc ( o r 0 alone does not cause a Mike phenotype). Thedifferences in dpp expression in different genetic backgrounds suggest that wo regulates dpp expression in a dosedependentmanner, which is effectivelyvisualized in thecompound eye when the amount of the Hh protein is limited like that in the B background. On the other hand, hh expression in wo heterozygotes remains the same as that in wild-type disc (data not shown). There is no indication of reduced hh expression even in the region just posterior to the MF, where hh expression is most affected in B mutants (Figure 1D). This result indicates that the wo' mutation affects dpp expression, but not hh expression in the eye imaginal disc during MF progression in a dosedependent manner. Thus, we hypothesize that wo regulates

1047

FIGURE4.-dpp expression in the eye imaginal discs with combinations of I3 and oro. Anterior is to the left, and dorsal is to the top (except for C). Approximate locations of the morphogenetic furrow are indicated by arrows. All discs were stained at the sametime and with the same staining solution to minimize possible differences among preparations. (A) dpplacZ expression pattern (dpplacZBS3.0, B~ACKMAN et al. 1991) in thewild-type eye imaginal disc. Strong dppexpression is seen at the morphogenetic furrow. (B) dpplncZ pattern of an oro' heterozygote. Note a significant reduction of dpplaci! expression in the disc. (C) dpp-lucZ pattern in the heterozygous R'. This disc shows that dpp expression starts to diminish midway through eye morphogenesis. The equator region is first affected, but there is still strong dpp expression in the peripheral region. (D) dpplnci! expression in the disc of a R ' / + ; oro'/+ individual. Note that there is no dppexpression in the disc except for the peripheral region.

dpp expression either downstream of or parallel to hh during MF progression. ptc suppresses the phenotypes of Band hh' enhanced by m: p t c is epistatic to oro in embryonic ventral epidermis patterning. In addition, ptc is a dominant suppressor of B and hh' and negatively regulates d$p expression in the eye imaginal disc (MA and MOSES 1995; U 7 ~ ~ r u 1 and TOMLINSON 1995). To examinethe relationship between oro and ptc during eye morphogenesis, we introduced the pt8'allele into the Bandhh' backgrounds with and without the oro' allele. or0 is a dominant enhancerof B and hh', while ptc is a dominant suppressor of these same alleles (Figure 5). The oro' allele enhances the hh' phenotype although its effect is a little weaker than its effect on weak B phenotypes. On the other hand, the pt8'alIele strongly

1048

.J. I.. Epps,J. B. J o n e s and S. Tanda

FI(;I:KE:5.-l