Development 112, 371-388 (1991)Printed in Great Britain © The Company of Biologists Limited 1991

371

cactus, a maternal gene required for proper formation of the dorsoventral

morphogen gradient in Drosophila embryos

SIEGFRIED ROTH1*, YASUSHI HIROMI2*, DOROTHEA GODT3 and CHRISTIANE NUSSLEIN-

VOLHARD1

1 Max-Planck-Institut filr Entwicklungsbiologie, Spemannstrasse 35/111, 7400 Tubingen, FRG2Department of Molecular and Cell Biology, University of California, Berkeley, CA 94720, USA3Institut filr Entwicklungsphysiologic, Universitat Kdln, Gyrhofstrasse 17, 5000 Kb'ln 41, FRG

•Present address: Department of Molecular Biology, University of Princeton, Princeton, NJ 08544, USA

Summary

The dorsoventral pattern of the Drosophila embryo ismediated by a gradient of nuclear localization of thedorsal protein which acts as a morphogen. Establish-ment of the nuclear concentration gradient of dorsalprotein requires the activities of the 10 maternal 'dorsalgroup' genes whose function results in the positiveregulation of the nuclear uptake of the dorsal protein.Here we show that in contrast to the dorsal group genes,the maternal gene cactus acts as a negative regulator ofthe nuclear localization of the dorsal protein.

While loss of function mutations of any of the dorsalgroup genes lead to dorsalized embryos, loss of cactusfunction results in a ventralization of the body pattern.Progressive loss of maternal cactus activity causesprogressive loss of dorsal pattern elements accompaniedby the expansion of ventrolateral and ventral anlagen.However, embryos still retain dorsoventral polarity,even if derived from germline clones using the strongestavailable, zygotic lethal cactus alleles. In contrast to theloss-of-function alleles, gain-of-function alleles of cactuscause a dorsalization of the embryonic pattern. Geneticstudies indicate that they are not overproduces ofnormal activity, but rather synthesize products withaltered function. Epistatic relationships of cactus with

dorsal group genes were investigated by double mutantanalysis. The dorsalized phenotype of the dorsalmutation is unchanged upon loss of cactus activity. Thisresult implies that cactus acts via dorsal and has noindependent morphogen function. In all other dorsalgroup mutant backgrounds, reduction of cactus functionleads to embryos that express ventrolateral patternelements and have increased nuclear uptake of the dorsalprotein at all positions along the dorsoventral axis.Thus, the cactus gene product can prevent nucleartransport of dorsal protein in the absence of function ofthe dorsal group genes. Genetic and cytoplasmictransplantation studies suggest that the cactus product isevenly distributed along the dorsoventral axis. Thus theinhibitory function that cactus product exerts on thenuclear transport of the dorsal protein appears to beantagonized on the ventral side. We discuss models ofhow the action of the dorsal group genes mightcounteract the cactus function ventrally.

Key words: developmental genetics, dorsoventral pattern,dorsal nuclear gradient, inhibition of nuclear transport,maternal-effect gene, Drosophila, cactus.

Introduction

In Drosophila, polarity and primary subdivisions of theembryonic body axes are dependent on the activities ofmaternal-effect genes (Nusslein-Volhard et al. 1987;Anderson, 1987; Nusslein-Volhard and Roth, 1989).One integrated system of twelve components, encodedby the eleven dorsal group genes and cactus, is requiredto establish the dorsoventral pattern of the embryo(Anderson and Nusslein-Volhard, 1986; Schiipbach andWieschaus, 1989). Females lacking the activity of any ofthe dorsal group loci produce completely dorsalized

embryos, whose cells behave and differentiate like thedorsal cells of wild-type embryos.

For many of the dorsal group genes, hypomorphicalleles have been isolated which cause only a partialdorsalization (Anderson and Nusslein-Volhard, 1986).The phenotypes produced by such alleles can bearranged in a hypomorphic series, which covers acontinuous spectrum of pattern alterations rangingfrom wild type to complete dorsalization. The coordi-nated and continuous fate map shifts underlying thesephenotypes suggested that the positional information ofthe dorsoventral pattern is graded and that the action of

372 S. Roth, Y. Hiromi, D. Godt and C. Nilsslein-Volhard

the dorsal gTOup genes culminates in the establishmentof a morphogen gradient (Niisslein-Volhard, 1979a,b;Anderson et al. 1985).

Classical as well as molecular experiments led to afunctional dissection of the group of eleven genes withidentical phenotypes. The double mutant combinationsbetween dominant ventralizing Toll alleles and dorsal-izing mutants demonstrated that Toll acts downstreamof gastrulation defective, nudel, pipe, snake and easter,but upstream of dorsal (Anderson et al. 1985). Thus, theproduct encoded by the dorsal gene functions at the endof the developmental pathway, dorsal is distinct from allother dorsal group genes in two further respects. First,it is dosage-sensitive. At high temperature, embryosproduced by females lacking one dorsal copy areweakly dorsalized (Niisslein-Volhard, 1979a; Niisslein-Volhard et al. 1980). The dosage-sensitivity indicating aconcentration-dependent action might be characteristicof morphogens as it was also observed for bicoid, whichencodes the anterior morphogen (Frohnhofer andNiisslein-Volhard, 1986). Second, cytoplasmic trans-plantation experiments revealed a ventral enrichmentof rescuing activity, indicating an asymmetric distri-bution of active dorsal product (Santamaria andNusslein-Volhard, 1983; Niisslein-Volhard and Roth,1989). Together, these observations suggested thatdorsal encodes the morphogen of the dorsoventral axis.

In agreement with this assumption, dorsal proteinwas found to form a nuclear concentration gradientalong the dorsoventral axis with a maximum at theventral side (Steward et al. 1988; Steward, 1989;Rushlow et al. 1989; Roth et al. 1989). The analysis ofdorsal protein distributions in mutant embryos showsthat different concentration ranges of nuclear dorsalprotein correspond to the main subdivisions of thedorsoventral pattern: the mesoderm; the ventral neur-oectoderm, which gives rise to the central nervoussystem and the ventral epidermis; the dorsolateralectoderm; and a dorsal region, from which dorsalepidermis and amnioserosa are derived. The cell fatesof these regions are determined by the correspondingnuclear dorsal protein concentrations in a largelyautonomous manner (Roth et al. 1989). dorsal proteinprobably acts as a transcriptional regulator. It hassequence similarities with the nuclear proto-oncogenec-rel and with the transcription factor NF-JCB/KBFI(Steward, 1987; Kieran etal. 1990; Ghosh et al. 1990). Acomparison of the dorsal protein distribution with theexpression pattern of zygotic dorsoventral genessuggests that it exerts both activating and repressingfunctions. The highest nuclear dorsal protein levels arerequired to initiate directly or indirectly the transcrip-tion of the zygotic genes twist and snail, both needed tospecify the mesoderm (Thisse et al. 1987; Leptin andGrunewald, 1990). Low nuclear levels cause a direct orindirect repression of genes, like zen or dpp, requiredfor the development of dorsal epidermis and amnioser-osa (Rushlow et al. 1987; Irish and Gelbart, 1987; St.Johnston and Gelbart, 1987; Roth et al. 1989).

No overall asymmetry in the concentration of dorsalprotein is observed along the dorsoventral axis of

syncytial blastoderm embryos. Therefore, the forma-tion of the nuclear gradient must occur via a spatialregulation of the nuclear localization of dorsal protein(Steward, 1989; Rushlow et al. 1989; Roth etal. 1989).Mutations in any of the 10 dorsal group genes lead to acomplete cytoplasmic dorsal protein localization, indi-cating that they are involved in the nuclear uptake ofdorsal protein at ventral positions. The products of thedorsal group genes form a signal transduction pathway.Toll encodes a transmembrane protein that is evenlydistributed in the plasma membrane of syncytialblastoderm embryos (Hashimoto et al. 1988; C.Hashimoto, personal communication). Its ventral acti-vation requires the activities of two serine protease-likemolecules encoded by snake and easter (DeLotto andSpierer, 1986; Chasan and Anderson, 1989), which areprobably secreted into the perivitelline space. Theactivated Toll protein stimulates the spatially restrictednuclear localization of the dorsal protein.

While 10 genes are involved in the positive regulationof the nuclear localization of dorsal protein, only onegene, cactus, has been identified that has opposingeffects. Loss-of-function mutations at the cactus locusresult in a partial ventralization of the embryonicpattern (Schiipbach and Wieschaus, 1989), ac-companied by an extension of the nuclear dorsalprotein gradient towards the dorsal side (Steward, 1989;Roth etal. 1989). Here, we describe the effects of cactusmutations which include both ventralizing loss-of-function and dorsalizing gain-of-function alleles. Thedosage sensitivity of cactus and the continuity ofphenotypic alterations produced by hypomorphicalleles demonstrate that the concentration of activecactus product is crucial for the determination of thedorsoventral anlagen. However, using double mutantcombinations, we show that the concentration-depen-dent action of cactus occurs entirely via dorsal. Basedon genetic studies and cytoplasmic transplantationexperiments, we propose that the cactus product isevenly distributed along the dorsoventral axis. Thedorsalizing function of cactus is normally suppressed inthe ventral half of the egg circumference (EC) by theaction of dorsal group genes. This process is probablyblocked in dorsalizing gain-of-function cactus alleles.The dorsal protein distributions exhibited by mutantand doubly mutant embryos suggest that cactus encodesan inhibitor of the nuclear concentration of dorsalprotein. We propose a direct interaction of cactus anddorsal proteins.

Materials and methods

Fly strainsThe wild-type stock was Oregon R. The references for thecactus alleles used in this study are given in Table 2. Thecactus deficiencies Df(2L)E10RN2 and Df(2L)TE116GW21are described in Ashburner etal. 1990. dl' (Nusslein-Volhard,1979a). In(2L)dlT (Steward and Niisslein-Volhard, 1986). Thealleles of dorsal group genes spzm7, plF™8, TF™9 (designatedas mel(3)7, mel(3)8, mel(3)9 by T. Rice. 1973); TlSBRE,T,9QRE 7-rO2 Df(3L)ro*63t nd^6t ndfa, ^ ( f o r m e r

Developmental genetics of cactus 373

designation ea"R6s\ ea2, tub"8, tub238, snk073, snk229, pip386,pip , pll078, spz , spz67 are described in Anderson andNusslein-Volhard (1984, 1986), Anderson et al. (1985). ea5 "(Chasan and Anderson, 1989). windRP (Schiipbach andWieschaus, 1989).

All mutant chromosomes carried visible markers allowinggenotypic identification. Flies were grown and eggs collectedunder standard conditions (Nusslein-Volhard et al. 1984).Staging of embryos was according to Campos-Ortega andHartenstein (1985).

Double mutant analysisThe choice of the cactus allele is crucial for the investigation ofthe epistasis relations of cactus and the dorsal group genes. Ifweak (V4) or medium (V3) alleles are used, all doublemutants with completely dorsalizing alleles of dorsal groupgenes result in completely dorsalized embryos. Althoughthese cactus alleles cause a deletion of dorsal and dorsolateralpattern elements on their own they do not, if simultaneously adorsal group gene is removed. Therefore, it is necessary toperform the double mutant analysis with strong cactus alleles(V2). Even in this case the phenotype exhibited by the doublemutant combination is very sensitive to the alleles or alleliccombinations used, as demonstrated in Fig. 8. To getconsistent results, we constructed all double mutants with thestrongest viable allele, cact42:

cact^/cact*2; ndf93 / ndf46

cac^/cact*2; pip664'/pip386

C4ctA2windRP/cactA2windRP

cart*21 cad*2; spzm7/spz197

gd2/gd2; cac^/cac^cacfi/aict*2; snk<"3 / snk229

cac^/cact*2; eaJ/ea2

cad*21 cad*2; T?QRE / Df(3R)wXB3

cac^/cact*2; plf78[pir8

cac^/cact*2; tub^/tub"8

cacf^dl'/cact^dl'

X-ray mutagenesis for reversion of cactE1°cacfi10 was isolated by excision of the P-element from aP-induced loss-of-function allele cact"5 (Table 1). cact810/CyO males were X-ray irradiated and mated to In(2L)dlT b pren sea/CyO females. Approximately 12000 cacf10'/'In(2L)dlT

females were screened at 18°C for the production of viableprogeny. Putative revertant lines were established from maleprogeny exhibiting cn+ Cy phenotype.

Temperature-shift experimentsTo determine the temperature-sensitive periods of cactHE andcacf8 eggs were collected from cactHE or cact*18 homozygousfemales on yeasted agar plates for 3h at 29°C or for 6h at18°C and shifted to the other temperature at the end of theegg collection period. The eggs were covered with Voltalef 3Soil and visibly staged by selecting gastrulating embryos (stage6; Campos-Ortega and Hartenstein, 1985) at regular intervalsafter the shift (15min at 29°C; 30min at 18°C). The selectedembryos developed at 18 °C (down-shift) or 29 °C (up-shift).Cuticle preparations of the differentiated embryos wereexamined.

Cytoplasmic transplantationCytoplasmic transplantations were performed according toSantamaria and Nusslein-Volhard, 1983.

Pole cell transplantationThe protocol for the transplantation of pole cells has been

described (Lehmann and Nusslein-Volhard, 1986). Pole cellswere transplanted into the progeny of a cross between OregonR females and ovoDI v males (Busson et al. 1983). In thisexperimental design, all female progenies are expected to besterile unless they have received functional pole cells bytransplantation.

AntibodiesThe production of antibodies against twist and dorsal proteinis described in Roth et al. 1989. Antibodies against zen proteinwere obtained from C. Rushlow (Rushlow et al. 1987).Immunological staining of whole-mount embryos with biotin-ylated HRP-avidin complexes bound to biotinylated secondantibody (Vector Laboratories, Avidin/Biotin ABC system)was carried out as described by Macdonald and Struhl (1986),with the modification that during the washes we added 100 DIMNaCl to the solutions. For sectioning, stained embryos weredehydrated (lOmin 70% ethanol, 2xl0min 100% ethanol,2x100% acetone) and mounted in Durcupan-ACM (Fluka).A complete series of transverse sections (10 /an) was preparedfor each embryo to study changes of the staining pattern alongthe anteroposterior body axes.

Cuticle preparations of embryosFor the observation of cuticular structures, differentiatedembryos with vitelline membrane or dissected out of thevitelline membrane were mounted in a mixture of Hoyer'smedium (Van der Meer, 1977) and lactic acid (1:1).

Results

Cactus has the meiotic position 2-51,7 (Roth, 1990). andmaps to the chromosomal region 35F1.2;36A1.2 (Ash-burner et al. 1990). In the course of several mutagenesisexperiments, 45 cactus alleles have been isolated(Table 2). They include homozygous viable as well aszygotic lethal alleles. In the following, we describe thematernal effects cactus mutations exert on the embry-onic dorsoventral pattern. The mutant phenotypes arecharacterized by the cuticle patterns, the morphogen-etic movements during gastrulation, the expressionpatterns of the zygotic genes zen (Rushlow et al. 1987)and twist (Thisse et al. 1988) and dorsal proteindistribution. We apply the phenotypic classification ofAnderson et al. 1985 with minor modifications assummarized in Table 1.

The haploinsufftcient dominant phenotypeDeficiencies of the cactus locus as well as lethal allelesand strong viable loss-of-function alleles cause adominant maternal effect. Depending on the geneticbackground, 50-90% of the larvae produced byheterozygous females do not hatch. In contrast to thedominant effect of the dorsal gene (Nusslein-Volhard,1979a), the dominant phenotype of cactus is notdependent on temperature. 20-40% of the non-hatching larvae have cuticle phenotypes resemblingthose of weakly ventralizing zygotic mutations such aszen (Wakimoto et al. 1984) or twisted gastrulation(Zusmann and Wieschaus, 1985). They form allcuticular pattern elements found in wild-type embryos,

374 5. Roth, Y. Hiromi, D. Godt and C. Niisslein-Volhard

Table 1. Classification of mutant phenotypes

DOD lD2D3V4V3V2VI

voLI (apolar)LI (polar)L2 (apolar)L2 (polar)L3

Amnioserosa

++++

partial-—-—---

partialpartial

Dorsal epidermis

++++

partial-—-—---

partial+

Filzkdrper, antennalsense organs

—+++

partial-—----++

- / +

Lateral epidermis,anal plates

—++++

partial—-—--++-

Ventral epidermis

-—+++++

partial-++---

Mesoderm

—-—

partial+++++-

partial---

The classification of phenotypes is according to Anderson et al. 1985 with modifications. Deleted structures are designated as ' - ' .'partial' means that the structure is reduced in size or deleted at some positions along the anteroposterior axis. In most cases, thestructures designated as ' + ' are extended to compensate for the loss of those structures designated as ' —'. Dorsauzed embryos (D) arecharacterized by the loss of ventral pattern elements and an accompanying expansion of more dorsally derived structures. In the mostextreme case (DO), only dorsal epidermis and amnioserosa are expressed along the entire embryonic circumference (EC). Ventralizedembryos (V) reveal an expansion of lateral and ventral pattern elements at the expense of more dorsally derived structures. In contrast toV2 embryos, VI embryos exhibit an expansion of the mesoderm at all positions along the anteroposterior axis, a feature that cannot beshown in the representation used in the table. A complete ventralization (VO) is characterized by the uniform expression of mesoderm.Lateralized embryos (L) are those that have lost both dorsauy and ventrally derived structures. They may be completely apolar atdifferent positional levels (e.g. ventrolateral, dorsolateral) or they exhibit residual polarity.

but head involution does not occur and the telson isfrequently pulled inside the posterior abdominal region(Fig. 1A). The dorsoventral extent of the zen ex-pression domain is reduced compared to wild type andfrequently no expression can be detected anteriorly at75% EL (egg length; anterior tip=100% EL) andposteriorly at 15 % EL (Table 3a, Fig. 3B). This revealsthat the reduction of cactus function causes a partial lossof dorsal or dorsolateral anlagen. Thus, cactus is ahaploinsufficient gene and the normal amount of itsproduct appears to be crucial for the determination ofdorsal structures.

The loss-of-function allelesA phenotypic continuum ranging from wild type tostrong ventralization is produced in embryos derivedfrom females that are homozygous or transhetero-zygous for different loss-of-function cactus alleles. Forsimplicity, we divide the ventralized phenotypes intofour classes of increasing strength (V4, V3, V2 and VI;Table 1). The class of weakest phenotypes, designatedas V4, also includes the haploinsufficient phenotype ofcactus. The class of strongest phenotypes, designated asVI, can be observed only in gerrnline mosaics withlethal alleles (see below).

The cuticle phenotype caused by weak alleles showsslightly wider than normal ventral denticle belts. Alldorsally and dorsolaterally derived pattern elements arestill present (V4, Fig. 1A,B). Stronger mutations resultin the replacement of the dorsal epidermis by nakedcuticle characteristic of more lateral positions in wild-type embryos. In addition, the dorsolaterally derivedfilzkorper and the sense organs of the head are lost (V3,Fig. 1C). The strongest viable alleles cause a complete

lack of all dorsally and laterally derived structuresaccompanied by the expansion of the ventral epidermisaround the entire circumference (V2; Fig. ID). Oftenthe cuticle is poorly differentiated and may be reducedin amount.

The alterations of morphogenetic movements incactus mutant embryos reflect early changes of thedorsoventral anlagen. During gastrulation even weakcactus mutations cause a shift of the cephalic fold,normally a lateral structure, to more dorsal positionsindicating an expansion of lateral at the expense ofdorsal anlagen (Fig. 2C,D). Germ band extension canbe used as a measure of ventralization. In wild type, thegerm band extends up to 65% EL, in weaklyventralized embryos up to 30% EL, and in stronglyventralized embryos no extension occurs. However,ventral furrow formation is normal even in the strongestmutations investigated (Fig. 2E).

Fig. 1. Dark-field photographs of the cuticle produced bycactus mutant embryos. (A) cacfi6 j cacfi6, 22°C (=V4).This phenotype results also from Df cact~/+. (B) cacrE/cacr*-, 18°C (=V4). Note that the embryos shown in Aand B have their Fk (arrow) and 8th abdominal segmentpulled inside the abdomen. (C) cactPD/cactPD (=V3).(D) cactHE/cactHE, 29°C (=V2). (E) cact013/Df(2L)E10RN2 (=V1). The embryo is derived from agerm line chimeric female (see Table 4). (F) Wild type.(G) cacf101cact*010 (=D3). This phenotype results from apartial loss of mesoderm as shown in Fig. 3L. (H) cact8'0/cact°"(=D2). The embryo has smaller ventral denticlebelts as compared to wild type indicating that it not onlylacks mesoderm, but also exhibits reduced ventrolateralanlagen. (I) cacf101 caa013 (=D1). (K) cact30/cact80

(=D0). For phenotypic classification, see Table 1. ASO,antennal sense organs; Fk, filzkorper; VE, ventralepidermis.

Developmental genetics of cactus 375

* w

/ 'I'

376 S. Roth, Y. Hiromi, D. Godt and C. Nusslein-Volhard

Table 2. cactus alleles

Allele Zygotic Phenotypicdesignation Mutagen lethality* class Referenced

Table 3. zen and twist expression

(a) zen expression in embryos with reduced cactusactivity

AA58AB10H8Q6YllPIP3P6P7

PDSGHE09OilH4P5

A2

meFUE19

99G8SIU7D12D13AA51AB13X6255E2E4E6E8E13E15El 7E20E25E29E10RJ1E10R01EP

E10BQ

EMSEMSEMSEMSEMS

P dysgenesisP dysgenesisP dysgenesisP dysgenesis

EMSEMSEMSEMSEMSEMS

P dysgenesisEMSEMSEMS

P dysgenesis

EMSEMSEMSEMSEMSEMSEMSEMSEMS

P dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesisP dysgenesis

X rayX ray

P dysgenesisP dysgenesisP dysgenesis

WildtypeWildtype (ts)c

V4 (ts)c

V4 (ts)c

V4 (ts)c

V4V4V4V4

V3V3

V3 (ts)c

V3V3V3V3

V2V2V2V2

V2*V2*V2*V2*V2*V2*V2*V2*V2*V2*V2*V2*V2»V2*V2*V2*V2*V2»V2*V2*V2*V2*V2*

L2-D1*L2-D0

442223333

1112223

2025

0222224445a5b5b5b5b5b5b5b5b5b5b776

56

+++++++++++++++++++++

'Zygotic lethality of individuals when homozygous ortransheterozygous with a cactus deficiency.

b Classification of the phenotypes of embryos denved fromhomozygous females or from females carrying the allele in trans tocacr*2 (*) according to the criteria given in Table 1.

cTemperature-sensitive allele.dThe known cactus alleles derive from eight different mutant

screens:(0) Nusslein-Volhard et al. 1984(1) Schupbach and Wieschaus, 1989(2) U. Mayer, R. Lehmann and C. NUsslein-Volhard, unpublished(3) S. Richstein and C. Nusslein-Volhard, unpublished(4) Ashburner et al. 1990, Roth 1990(5) Y. Hiromi, unpublished; 5a, original insertional mutant; 5b,dysgenic revertants of cact255

(6) D. St Johnston, unpublished(7) X-ray revertants of the gain-of-function allele cacf0, see textfor further details.

c cacfi10 is zygotic lethal when homozygous, but viable whentransheterozygous with lethal cactus alleles, cactus deficiencies orrevertants of cac^10. Therefore, cacr610 exhibits a gain-of-functionzygotic lethality.

% egg length(number of cellsaround EC)

90(59)75 (71)60(85)45 (91)30(88)15 (81)5(65)

zen-domain (% EC)

Wild type

34%40%42%42%40%35%

100%

cacr/+*(V4)

0%21%**25%28%28%21%**

100%

(b) twist expression in cactus mutant embryos

% egg length(number of cellsaround EC)

85(55)70(80)55(94)40(100)25(96)10 (67)

Wild type

40%29%26%25%26%32%

nvisr-domain (% EC)

cacfDlcacf6

(V3)

48%28%24%24%29%6 1 %

cact*2/ cact013/cact" DfcacC***(V2)

7 1 %28%29%30%32%86%

(VL)

100%32%30%30%49%

100%

Wild-type and mutant embryos were stained using anti-zerc oranti-nvir( antibodies as described in Materials and Methods. Afterembedding, the embryos were sectioned to yield complete series oftransverse sections each 10/on thick. At distances of 50/mi thetotal number of cells around the embryonic circumference (EC)and number of zen or twist expressing cells were counted. Theposition along the anteroposterior axis is indicated as % egglength (0%, posterior end). The results of 7-10 embryos wereaveraged. The width of the expression domains (%EC) arecalculated by relating the averaged numbers of twist or zenexpressing cells to the averaged number of cells around the EC ata given egg length. The values varied by 3% for differentembryos. The phenotypic classification (V4-V1) is indicated inbrackets (see Table 1).

* cacfl+ : cactA2/b pr.** At these positions the zen expression is very weak and

sometimes hard to detect in embryos derived from cacf/+females.

***The embryos are derived from germ line chimeric females(see Table 4). The genotype of the germ line was: cact°13/Df(2L)E10RN2.

Although the weakly ventralized phenotypes classi-fied as V4 include cuticular defects, similar or moresevere than those caused by zen mutations (Fig. IB),they always show residual zen expression between 30 %and 60 % EL (Fig. 3B) and differentiate some amnio-serosa (data not shown). This suggests that structuresderived from the dorsalmost positions do not disappeartotally before more lateral regions are affected. Acomplete deletion of the dorsal zen expression domainis first seen in ventralized embryos which also deletedorsolaterally derived structures (V3; Rushlow et al.1987; Fig. 3C).

As in the case of zen expression, twist expression isnot uniformly affected along the anteroposterior axis.Weak alleles show no deviation from wild type (data

Developmental genetics of cactus 377

PMG

Fig. 2. The morphogenetic movements of cactus mutant embryos. (A,B) Wild type. (C,D) cacfi6 / cacfi6. 22°C (=V4).(E,F) cac^/cact*2 (=V2). (G,H) cart*10[Df(2L)TE116GW21 (=L2 with polarity). The germ band extension visible in Hdemonstrates residual polarity. (I,K) cacrQ/cactBQ (=D0). For each genotype, the same embryo is shown at gastrulation(stage 6: A,C,E,G,I) and lOmin later at midgut invagination, stage 7: B,D,F,H,K). Staging is according to Campos-Ortegaand Hartenstein (1985). For phenotypic classification, see Table 1. CF, cephalic fold; DF, dorsal folds; AMG, anteriormidgut; PMG, posterior midgut.

not shown); intermediate alleles cause an expansion ofthe twist stripe only in terminal regions (Table 3b). Thiseffect is more pronounced in strong alleles where thetwist stripe is slightly wider also in the abdominal region(Table 3b; Fig. 3H). The observation that the twistexpression and therefore the anlagen of the mesoderm

are only weakly altered by cactus mutations suggeststhat the deletion of dorsal and dorsolateral anlagen iscompensated by an expansion of predominantly ventro-lateral anlagen.

Transheterozygous combinations of the loss-of-func-tion alleles including cactus deficiencies reveal no

378 5. Roth, Y. Hiromi, D. Godt and C. Nusslein-Volhard

Developmental genetics of cactus 379

Fig. 3. zen and twist expression in cactus mutant embryos.Embryos at blastoderm stage were stained using anti-zen oranti-twist antibodies as described in Materials and methods.Whole-mount preparations were photographed usingNomarski optics. Panel G, K and L show ventral surfaceviews, all others show optical sagittal sections.(A-E) Embryos stained with anti-zen antibodies.(F-L) Embryos stained with anti-twist antibodies.(A,F,G) Wild type. (B) cact"8/ cact"8, 22°C (=V4). Similarexpression patterns are seen in embryos derived from DfcacC/+ females. (C) cact"8/cact"8, 28°C (=V3). Note thatthe embryos shown in B and C have no change of zenexpression in terminal regions (marked by stars). Theterminal zen expression is not affected by alterations of thedorsoventral anlagen; it rather depends on terminal groupgenes (Rushlow et al. 1987). (D) cacP0/Df(2L)TE116GW21 (=L2 with polarity). (E) cact80/cact80

(=D0). The arrows in B and D mark a domain of residualzen expression. (H) cact*2'/'cact*2. (I,K) cact013/Df(2L)E10RN2 (=V1). The embryos are derived fromgerm line chimeric females (see Table 4). (L) cact*-10/cact*810 (=D3). Similar expression patterns as in L areseen in embryos derived from dl~/+ females at 28 °C. Thearrows in F, H, I and L mark the extent of twistexpression. For phenotypic classification, see Table 1.

unusual complementation behavior (Roth, 1990). Weakalleles (giving rise to V4 phenotypes) when tranhetero-zygous with strong alleles (giving rise to V2 pheno-types) result in an intermediate ventralization (V3). Intransheterozygous combinations, some strong allelesbehave like cactus deficiencies and thus may representamorphic mutations. However, the group of mutationsdenned by this criterium is not uniform, but includesviable as well as zygotic lethal alleles. Therefore, thedetermination of the amorphic phenotype requires theanalysis of the maternal effect of zygotic lethal cactusmutations.

Germline clones with lethal cactus allelesThe zygotic lethal cactus alleles cause late larval andpupal death associated with the formation of melanotictumors (Sparrow, 1978; Roth, 1990). In order to analysetheir maternal effects, we produced females with wild-type soma and mutant germline by pole cell transplan-tation. About one quarter of the fertile, chimeric

females produced strongly ventralized embryos whilethe others produced either wild-type or weakly ventra-lized (haploinsufficient phenotype) embryos (Table 4)reflecting the distribution of genotypes expected fromthe donor cross. This shows that the genotype of thegermline determines the embryonic phenotype. Thus,cactus, like most of the dorsal group genes (Schiipbachand Wieschaus, 1986; Seifert et al. 1987; Konrad et al.1988; Stein et al. in preparation) is a germline-dependent maternal-effect gene. We have producedgermline clones with three different lethal alleles(Table 4). Embryos derived from a germline of thegenotype cact° /Df cact~ produce only small frag-ments of poorly differentiated ventral cuticle (Fig. IE).Their ventral twist domain is wider at all anteropos-terior positions and they express twist uniformly inbroad terminal regions (Fig. 31,K, Table 3b), indicatingthat they are more strongly ventralized (VI) than theembryos derived from the strongest viable allelesdescribed above. The localized expression of the twistprotein (Fig. 3I,K) demonstrates, however, that theyare not completely ventralized and retain polarity.Germline chimeras with other lethal cactus allelesresulted in weaker phenotypes (V2, Table 4). Thus,some of the lethal alleles have residual cactus activity.

In summary, the loss-of-function alleles reveal thatcactus activity is required in wild-type embryos tospecify the development of dorsal, dorsolateral and, inpart, ventrolateral structures. The hypomorphic allelesproduce a continuous phenotypic spectrum character-ized by coordinated alterations of the entire dorsoven-tral anlagen. The ventralmost part of the pattern, themesoderm, is the least affected structure. The regionmost sensitive to a reduction of cactus activity is not thedorsalmost anlage of amnioserosa and dorsal epider-mis, but the dorsolateral and ventrolateral ectoderm.Despite the various cuticular defects of head, thoraxand telson and the non-uniform changes of zen and twistexpression along the anteroposterior axis, cactusmutant embryos have normal segmental and terminalanteroposterior anlagen (Caroll et al. 1987; Roth,1990).

The gain-of-function allelesWe have identified two mutations (cact*-'0 and

Table 4. Germ line chimeras with lethal cactus alleles

Donor Cross Adults"Fertilefemales

Genotype ofgerm line Phenotypec

cact0'2 y cact51

CyO CyO

cact013 Dfcacr*CyO CyO

104

218 17

cact 1 CyOCyO/CyO '

cact 1 cact: 2

Df/CyO .CyO/CyO '

cacti Df: 3

5

14

WT, V4

V2

WT, V4

VI

'Pole cells of embryos derived from donor crosses were transplanted into ovoD/+ recipients. The number of adult flies derived frominjected embryos was recorded.

"Dfcocr: Df(2L)E10RN2.cThe phenotype of embryos derived from chimeric females was characterized using cuticle preparations (Fig. IE) and staining with twist

antibodies (Fig. 31,K). For phenotypic classification see Table 1.

380 S. Roth, Y. Hiromi, D. Godt and C. Niisslein-Volhard

which in trans to loss-of-function cactus alleles producedorsalized, rather than ventralized embryos(Fig. 1G-K). These two mutations also fail to comp-lement amorphic mutations of dorsal, giving rise topartially dorsalized embryos (data not shown). In orderto distinguish whether E10 is a mutation in cactus,dorsal, or another interacting gene, we attempted torevert the maternal lethality of E10 in transheterozy-gotes with dorsal. Six X-ray revertants of E10 wereobtained, which produce viable progeny in trans todorsal. Of these, two behave as strong cactus alleleswhile four are deficiencies that uncover cactus, but notdorsal (Table 2; Ashburner et al. 1990). This exper-iment proves that E10 is indeed a cactus allele. For BQ,a spontaneous revertant was isolated, which exhibitsthe loss-of-function phenotype, suggesting that BQ is acactus allele, too.

When transheterozygous with weak, intermediateand strong cactus alleles, cacf10 and cart8® show aprogressive deletion of ventral anlagen compensated byan expansion of dorsal anlagen (Fig. 1G-K). The weakdorsalization exhibited by embryos derived fromcactE10/cactAB females is characterized by the partialloss of mesoderm (D3; Fig. 31). In transheterozygoteswith cacf11, the mesoderm is completely and theventral epidermis is partially deleted (Fig. 1H). Finally,in trans to cacf13 or a cactus deficiency the embryosdifferentiate only dorsal and dorsolateral cuticularstructures (Fig. II). The results for cactE1° and cactBQ

are similar. However, while cactE1°/Df cact~ leads to alateralization at dorsolateral level (L2, visible ingastrulation phenotype, Fig. 2G,H and zen expression,Fig. 3D), 40% of the embryos produced by cacfQ/DfcacC or cacf QI cacf Q females are completely dorsa-lized. Cuticle phenotype (Fig. IK), gastrulation(Fig. 21,K) and zen expression extending around theentire EC (egg circumference) (Fig. 3E) resemble thosecaused by completely dorsalizing mutations of dorsalgroup genes.

The observation that the dorsalizing effect of cactE1°and cad8® is dependent on a reduction of cactus activityand can be suppressed almost completely by a wild-typecopy of cactus indicates that these alleles do notrepresent overproducers. Rather, they lead to productswhich at ventral positions exert the dorsalizing functionnormally restricted to dorsal and lateral positions. Theyare therefore classified as neomorphic alleles thatinterfere with the spatial regulation required for normaldorsoventral pattern formation.

As shown earlier, following loss of cactus activitymore nuclei take up dorsal protein and consequentlythe nuclear dorsal protein gradient extends furthertowards the dorsal side than in wildtype (Roth et al.1989, Fig. 4A,B). The opposite effect results fromneomorphic cactus alleles. In agreement with theirdorsalized phenotype they cause more nuclei to excludethe dorsal protein. In the most severe cases nuclearlocalization is abolished around the entire EC and thedorsal protein is predominantly cytoplasmic (Fig. 4C).The protein distribution cannot be distinguished fromthat caused by strongly dorsalizing alleles of dorsal

Fig. 4. dorsal protein distribution in cactus mutantembryos. Embryos at blastoderm stage were stained usinganti-dorsal antibodies as described in Materials andmethods. Whole-mount preparations were photographedusing Nomarski optics. The maternal genotypes are:(A) Wild type. (B) cac^/cact*2. (C) cacf'^/cacf'3.

group genes (Roth et al. 1989). In summary, loss-of-function and gain-of-function alleles of cactus act viaopposing effects on the nuclear localization of dorsalprotein.

The time of action of the cactus gene productMost of the EMS-induced weak cactus alleles are heatsensitive: females homozygous or transheterozygousfor cacr48, cactQ6, cactY", cacf8 and cactHE produceembryos that are more strongly ventralized at 29°C thanat 18°C. The alleles cactQ6, cactY", cacf8 lead tohatching larvae at 18°C when homozygous. We usedcacf8 and cacfE, which show the strongest tempera-ture-dependent phenotypic changes, and performedtemperature-shift experiments to determine thephenocritical period (Fig. 5). For both alleles thetemperature-sensitivity is largely restricted to thesyncytial blastoderm stage indicating that the cactusproduct is acting at the same time when the nucleardorsal protein gradient is established and when nucleardorsal protein exerts its function on zygotic gene

Developmental genetics of cactus 381

100-

80-

60-

40-

20-

o-

I8-C-.29-C 29"C -> I8-C

Developmenta] stage at shift

Fig. 5. The temperature-sensitive period of cactHE andcacf18. The diagrams show the percentage of embryos withstrong phenotypes (V2 for cacflE • • and V3 forcact1^ O O) for the different developmental stages atwhich the temperature-shift occurred (see Materials andmethods). For each time point, the phenotypes of 50-100differentiated embryos were scored. The developmentalstages are indicated by numbers according to Campos-Ortega and Hartenstein (1985): 1,2, preblastoderm; 3, polecell formation; 4, syncytial blastoderm; 5, cellularization; 6,gastrulation.

expression (Roth et al. 1989; Rushlow et al. 1989;Steward, 1989).

Phenotypic rescue by injection of wild-type cytoplasmThe cactus mutant phenotype is only weakly suppressedby injections of wild-type cytoplasm. Embryos derivedfrom cactPD/cact011 females exhibit a ventralization ofmedium strength and do not produce filzkorper (V3).After transplanting wild-type cytoplasm into posteriordorsal positions, 25-60 % of the recipient embryos formpatches of filzkorper material (Table 5; Fig. 6). Therescue response occurs only locally, as the injectedembryos exhibit no weakening of the ventralization atmore anterior positions.

To investigate the spatial distribution of the rescuingactivity the transplanted cytoplasm was taken fromeither the dorsal or the ventral side of wild-typeembryos. No significant spatial asymmetry of therescuing activity was observed (Table 5). Despite thevariability in our results the experiments demonstrate

Table 5. Partial rescue of cactus by the injection ofwild-type cytoplasm

Donor:

Origin ofcytoplasm"

Dorsal

Ventral

Wlldtype

Age

CleavageSyncytial

CleavageSyncytial

Recipient: cacf"/cactPDb

Number of recipient embryos

Developed Rescuedc

87 2279 «

101 JK

116 m

% Rescued

2560

4030

"The cytoplasm was taken from either dorsal or ventralpositions of wild-type embryos.

b Recipient embryos (stage 2-4b) derived from cac/"'/cactPI>

females were injected into the posterior dorsal side in allexperiments. 50-80% of the injected embryos developed cuticularstructures.

cThe criterion for rescue was the production of filzkorper.

Fig. 6. Cuticle phenotype of injected cactus mutantembryos. Cleavage stage cactp /cactPD embryos wereinjected with cytoplasm from wild-type cleavage-stageembryos. The amount of transplanted cytoplasmcorresponded to approximately 3% of the total eggvolume. The cytoplasm was transplanted to the dorsal sideof the posterior region. The injected embryos were allowedto differentiate. (A) The dark-field photograph of aninjected embryo shows that the rescue response is weakand locally restricted. (B) Phase-contrast magnification ofthe posterior, dorsal region of the same embryo showsnonextended filzkorper (Fk) and spiracles (Sp) never seenin uninjected control embryos.

that the cactus product, although predominantlyrequired in the dorsal half of the EC (as inferred fromthe loss-of-function phenotype), is also present ven-trally.

Double mutants of cactus and completely dorsalizingmutantsTo analyse the relationship between cactus and the 11dorsal group loci, we constructed females simul-taneously homozygous for a strong cactus allele and anamorphic allele of a dorsal group gene (for completedescription of genotypes, see Materials and methods).In ten of the eleven possible combinations, thecompletely dorsalized phenotype caused by the dorsalgroup mutation is changed upon the loss of cactusactivity. Embryos are produced that differentiateventral epidermis around the entire EC (Fig. 7A). Incontrast to cactus embryos, they lack polarity: they donot form a ventral furrow during gastrulation and the

382 5. Roth, Y. Hiromi, D. Godt and C. Niisslein-VolhardAnti-dorsal

Fig. 7. Double mutants of cactus with amorphic alleles of dorsal group loci. (A-D) cacr42 / cacr42; TfQRE/Df(3R)roXB3

(=L1). (E) ndf93/ndf46 (=D0). (F) cac^/cact42; ndf93'/ndf46 (=L1). (G,H) cacr42 dl'/cacr42 dl' (=D0). (A,G) Dark-field photographs of cuticle. While the embryo in A shows ventral denticle belts surrounding the entire embryoniccircumference, the embryo in G shows only dorsal epidermis and cannot be distinguished from embryos derived fromdorsal homozygous females. (B) Living embryo during gastrulation (stage 6). The laterally derived CF is visible at bothdorsal and ventral sides. No ventral furrow formation occurs. Compare to ventralized gastrulation shown in Fig. 2E,F.(C) Blastoderm embryo (stage 5) stained with anti-twist antibodies. (D,H) Blastoderm embryo (stage 5) stained with anti-zen antibodies. (E,F) Blastoderm embryos (stage 5) stained with anti-dorsal antibodies. While the embryo in E reveals anexclusively cytoplasmic dorsal protein localization, all nuclei of the embryo in F contain medium levels of dorsal protein.Compare to Fig. 4A,B. CF, cephalic fold; VE, ventral epidermis.

laterally derived cephalic fold is present dorsally andventrally (Fig. 7B). They express neither zen nor twist(outside the terminal regions of the anteroposterioraxis) (Fig. 7C,D), which demonstrates the loss of bothdorsalmost and ventralmost pattern elements ac-companied by the expansion of ventrolateral structuresas characteristic for a lateralized phenotype (LI).

All double mutant combinations share the completelack of polarity with the amorphic single mutants of thedorsal group loci. Therefore, the dorsal group genes areall needed to polarize the embryonic dorsoventralpattern, but they are not absolutely necessary to

express fates different from that of the dorsalmostanlagen. There is only one exception from this rule: theembryos produced by cact~dl~ females are completelydorsalized and indistinguishable from embryos pro-duced by dl~ females. This can be seen by the cuticlephenotype, exhibiting only dorsal epidermis (Fig. 7G),and the uniform expression of zen (Fig. 7H). Hence,only the activity of the dorsal gene is an absoluteprerequisite for the production of more ventral struc-tures.

As shown earlier, dorsal protein is present inembryos mutant for dorsalizing alleles of all dorsal

Developmental genetics of cactus 383

Fig. 8. Double mutants of hypomorphic cactus alleles with amorphic alleles of dorsal group loci. Maternal genotypes:(A,B) cactHE/cacHE; spzm7/spz197 (29°C). (C-F) cac^/cact42; spzm7/spz197. (A,C) Dark-field photographs of cuticlepreparations. (B,D) Phase-contrast photographs of cuticle preparations. (E) Gastrulating embryo (stage 6). (F) Blastodermembryo (stage 5) stained with anti-zen antibodies. The arrows mark a domain of residual zen expression. The phenotypedepicted in A,B is an apolar lateralization at a dorsolateral level (L2). Dorsolaterally derived Fk material is visible at allpositions of the embryonic circumference. The phenotype shown in C-F resembles that of completely dorsalized embryoswith respect to the cuticle pattern, however, in 30% of the embryos antennal sense organs are visible (D). The deepcephalic fold during gastrulation (E) and the partial lack of zen expression (F), features visible in all mutant embryos,demonstrate that this phenotype is lateralized, however, at a more dorsal level (L3) compared to L2. See Table 1 forclassification of phenotypes. ASO, antennal sense organs; CF, cephalic fold; Fk, filzkorper; Sp, spiracles.

group genes (Steward, 1989; Roth et al. 1989).However, it is excluded from the nuclei and remainslocalized in the cytoplasm. The exclusively cytoplasmiclocalization is due to the presence of cactus product,because in doubly mutant embryos lacking both cactusand a dorsal group activity nuclei at all dorsoventralpositions take up dorsal protein equally (Fig. 7E,F).The intermediate nuclear dorsal protein levels corre-spond to the uniform development of ventrolateralstructures (Roth et al. 1989).

If, instead of a strong cactus mutation (cacr42),weaker alleles (e.g. cact^E or cactPD) are combinedwith amorphic alleles of dorsal group genes, the level of

apolar lateralization exhibited by the doubly mutantembryos is shifted to more dorsal positions. Thus, thedevelopmental fate can be altered at all positions of theEC as a consequence of a change of cactus activity andhence, of the level of nuclear dorsal protein. We areable to distinguish three types of apolar lateralizedembryos resulting from such double mutant combi-nations: (1) the ventrolateral level described above (LI,Fig. 7A-D,F) , (2) a dorsolateral level with rings offilzkorper material (L2, Fig. 8A,B), (3) a dorsolaterallevel with dorsal epidermis, but reduced zen expressionand a lateralized gastrulation phenotype (L3,Fig. 8C-F).

384 S. Roth, Y. Hiromi, D. Godt and C. Niisslein-Volhard

Fig. 9. Double mutants of cactus with partially dorsalizing and ventralizing mutants. Embryos at blastoderm stage werestained using anti-twist antibodies as described in Materials and methods. Whole-mount preparations were photographedusing Nomarski optics. The maternal genotypes are: (A) spz67/spzm? (22°C, =D2). (B) cact*2/cact*2; spz^/spz"'7 (22°C,= L1 with polarity). The partial reduction of twist expression exhibited by the embryo shown in B is not visible in opticalsagittal sections. (C) easl3/+ (=L1 with polarity). (D) cact*2 / cacf42; ea*'3/+ (=V1). (E) Tlrm9/Tl""9 (=L1). (F) cact*2/cacf*2; jirm91Tirm9 (=V0). The arrows demarcate the domain of twist expression. See Table 1 for classification ofphenotypes.

In summary, the double mutant analysis reveals thatcactus and dorsal act downstream of all other dorsalgroup genes and that cactus is not an independent factorrequired for the determination of dorsal and lateralfates. Instead, it exerts its function via the dorsalproduct, dorsal can act as a morphogen in the absenceof all other dorsal group genes. The only requirementfor its morphogen function is the nuclear localization ofthe dorsal protein, whose extent is closely linked to theresidual amount of cactus activity.

Double mutants of cactus and partially dorsalizingand ventralizing mutantsA variety of alterations of the dorsoventral pattern canbe produced using partially dorsalizing, ventralizingand lateralizing alleles of different dorsal group genes(Anderson et al. 1985; Anderson and Niisslein-Volhard,1986). The different phenotypes correspond to differentdorsal protein distributions (Steward, 1989; Roth et al.1989). We wondered how these aberrant patternschange if cactus activity is reduced. A weakly dorsal-izing mutation (e.g. spz67/spz67) leads only to thedeletion of mesoderm (D2). In a double mutant with a

strong cactus allele (cacf*2 / cacr42; spz67/spz67), theability to form mesoderm is partially restored(Fig. 9A,B). A dominant easier allele (ea5 , Chasan etAnderson, 1989) can be used to produce lateralizedembryos that still have residual polarity although theyhave lost dorsalmost and ventralmost structures andreveal only a weak ventral twist expression (LI withpolarity). The reduction of the cactus activity incactA2/cactA2; ea513/+ females leads to a dramaticincrease of the twist domain. However, the doublymutant embryos still retain polarity as revealed bydifferences in the level of twist expression along theirdorsoventral axis (VI, Fig. 9C,D). Finally, the removalof the cactus activity in a mutant background thatalready causes an apolar lateralized phenotype {Tlrm9/Tlrm9, LI) leads to a complete ventralization (cacr42/cact42; Tlrm9/Tlrm9; Roth et al. 1989). The embryosexpress the twist protein evenly along their dorsoventralaxis (Fig. 9E,F). They show no polarity during develop-ment and do not produce any cuticular structures (VO).

In these three double mutant combinations, thereduction of cactus activity preserves the polarity of thepattern present in the single dorsal group mutant. Thereduction of cactus activity affects all positions of the

Developmental genetics of cactus 385

EC and it leads to the formation of mesoderm not oronly weakly present in the single mutant.

Discussion

Among the 12 maternal-effect genes that encodecomponents of the dorsoventral pattern formationprocess cactus is unique in that only loss-of-functionmutations of cactus lead to ventralized phenotypes. Theobservation that the ventralization exhibited by cactusembryos is accompanied by increased levels of nucleardorsal protein suggested that cactus is a negativeregulator of the nuclear localization of dorsal protein(Roth etal. 1989; Steward, 1989). This hypothesis raisesseveral questions, which we address in this paper usingmainly a genetic approach. (1) Does cactus indeed exertits function entirely via dorsal or is it also requiredindependently for dorsoventral pattern formation? (2)Does cactus inhibit the nuclear localization of dorsalprotein indirectly (mediated by other components) orvia direct interactions with dorsal protein? (3) How isthe cactus product distributed in the embryo? (4) Whatis the relationship between cactus and the dorsal groupgenes?

(1) The loss-of-function alleles of cactus demonstratethat in wild-type embryos cactus activity is required (atleast) for the determination of the dorsal, the dorsolat-eral and part of the ventrolateral anlagen. The strongdosage sensitivity of the phenotype might lead to theassumption that cactus, like bicoid or dorsal, acts as amorphogen and determines the anlagen of the dorsalhalf of the EC in a concentration-dependent manner.However, cactus is not required for the formation ofdorsally derived structures if dorsal function is missing.Therefore, cactus is not an independent morphogen. Itacts exclusively via dorsal. The dosage sensitivity andthe continuity of phenotypic alterations produced byhypomorphic cactus alleles can be explained if theamount of cactus activity is closely linked to themorphogen function of dorsal.

(2) In the double mutant combinations of cactus andcompletely dorsalizing alleles of dorsal group genes, theremoval of cactus confers some potential to formventrolateral pattern elements upon all dorsal groupmutants with the only exception of dorsal. Hence, noneof the so-far-identified components acts between cactusand dorsal and the products of both cactus and dorsalfunction at the end of the developmental pathway.Although we cannot be sure that all genes of thedorsoventral pattern formation process are known, thepresented data are consistent with a close relationship,potentially a formation of a complex including cactusand dorsal products.

It has been shown that the amorphic alleles of thedorsal group loci lead to embryos that contain normalamounts of dorsal protein. However, the protein fails tobe taken up by the nuclei (Steward, 1989; Rushlow etal.1989; Roth et al. 1989). The double mutants with cactusdemonstrate that dorsal protein is not only present butalso functional in the absence of dorsal group activities.Obviously the only requirement for its function is its

nuclear localization, which can occur when cactusactivity is reduced. One simple model would be that thecytoplasmic localization of dorsal protein is controlledby the formation of a complex with cactus protein. Onlydorsal protein released from the cytoplasmic complexwith cactus enters the nucleus. The investigation ofseveral hypomorphic dorsal alleles that are defective inthe cacfMS-dependent inhibition of nuclear localizationsupports this view (Roth, 1990).

The proposed biochemical interactions are reminis-cent of those of the transcription factor NF-JCB and itsinhibitor IKB (Baeuerle and Baltimore, 1988; Zabeland Baeuerle, 1990). NF-KB is required for theinducible expression of a variety of genes in differentcell types (Lenardo and Baltimore, 1989). The active,nuclear form of NF-*B is a complex of two polypep-tides, p65 and p50 (Baeuerle and Baltimore, 1989). p50represents the DNA-binding component of the complexand, as recently shown, has sequence similarities to thedorsal protein (Kieran et al. 1990; Ghosh et al. 1990).The other polypeptide of the complex, p65, is requiredfor the interaction with the inhibitor protein, IKB(Baeuerle and Baltimore, 1989). If complex formationwith IR-B occurs, p50/p65 remains in the cytoplasm.Therefore, I-^B acts as an inhibitor of p50/p65 nucleartransport. Given the sequence similarities betweendorsal and p50, the structure of cactus might be relatedto IKB.

(3) The formation of the nuclear gradient of dorsalprotein requires that its nuclear localization is regulatedin a graded manner along the dorsoventral axis. Theinhibition that cactus exerts on the nuclear localizationshould be highest dorsally and it should continuouslydecrease towards more ventral positions. Unequalactivity of cactus along the dorsoventral axis could arisefrom an asymmetric distribution of its product, thespatially regulated inhibition of its function or otheractivities competing with its function in a spatiallycontrolled way. In cytoplasmic transplantation exper-iments, we find cactus activity at both dorsal and ventralsides of wild-type embryos. However, the rescueresponse is so weak that we could only have excluded avery strong asymmetry of distribution. In doublemutant combinations with various dorsalizing andventralizing mutations (Figs 7, 8, 9), the reduction ofcactus activity causes fate map changes at all positionsalong the dorsoventral axis. This indicates that cactusacts both dorsally and ventrally when dorsal groupactivities are absent or partially reduced. Because someof the mutations used in this study (e.g. spz67) causeonly weak phenotypes, we think that this conclusion isvalid for wild-type embryos, too. Therefore, wepropose that the cactus product is equally distributedalong the dorsoventral axis. According to this interpret-ation the phenotypic effects caused by gain-of-functionalleles of cactus are not due to a mislocalization ofcactus products. We believe that the neomorphic allelesproduce mutant cactus protein that like the wild-typeprotein is equally distributed, but leads to an unregu-lated inhibition of the nuclear localization of dorsalprotein. The phenotypes caused by these alleles

386 S. Roth, Y. Hiromi, D. Godt and C. Niisslein-Volhard

resemble those of dorsal group mutations, because theirproducts block the action of the dorsal group genes,which normally promote the graded nuclear localizationof dorsal protein.

Aside from information relating to the cactusfunction, the double mutant combinations of amorphicdorsal group alleles with cactus alleles of differentstrength (Figs 7, 8) are relevant to the interpretation ofthe morphogen function of dorsal protein. The embryosthat result from these double mutants accumulate thesame amount of dorsal protein in all nuclei as illustratedfor LI embryos (Fig. 7F). They show no polarity duringdevelopment and express the same pattern elementsaround the entire EC. Therefore, the pattern elementsdifferentiated by such embryos cannot result frominteractions of regions with different developmentalfate, but are determined by the respective nucleardorsal protein concentration in a largely autonomousway (Roth et al. 1989). Using cactus alleles of differentstrength, the nuclear dorsal protein concentrations canbe altered in an almost continuous manner. In this way,we can determine how many distinct regions of thedorsoventral axis are directly dependent on themorphogen gradient. Here, we have described threedifferent forms of lateralization (L3, L2, LI). Thus,together with the complete dorsalization (DO) andventralization (VO), so far, five different positionallevels of the dorsoventral axis have been shown todepend on different nuclear dorsal protein concen-trations.

(4) The action of the dorsal group genes in theventral half of the egg, which causes the formation ofthe nuclear concentration gradient of the dorsalprotein, can be explained in two different ways. Thedorsal group genes could act either via dorsal or viacactus (Roth et al. 1989). The first model explains theregulation of nuclear transport of dorsal proteinassuming that the dorsal group genes and cactus act inparallel (Fig. 10). In this model, cactus could be a

Mode) 1

dorsal groupgenes

cactus

nuclear transportof dorsal protein

Model 2

dorsal groupgenes cactus

nuclear transportof dorsal protein

Fig. 10. Two models explaining the relation of the dorsalgroup genes and cactus. According to model 1 cactus andthe dorsal group genes act in parallel to regulate thenuclear transport of dorsal protein. Model 2 assumes aserial action.

cytoplasmic anchor that retains dorsal protein in thecytoplasm. Upon modification of dorsal protein due tothe activity of dorsal group products (in the ventral halfof the egg) dorsal protein is released.from its binding tocactus and enters the nucleus. The second modelproposes a serial action of the dorsal group genes andcactus; this model formally represents the inhibition ofan inhibition (Fig. 10). Also in this model cactus couldbe a cytoplasmic anchor that binds dorsal protein.However, the action of the dorsal group genes wouldcause a modification of cactus that leads to the releaseof dorsal protein.

The second model proposing that the dorsal groupgenes act exclusively via cactus postulates that acomplete loss of cactus function results in totallyventralized embryos without residual polarity. Thedouble mutants between dorsal group genes (dorsalexcepted) and cactus should also lead to a completeventralization. At variance with these predictions, thestrongest cactus mutations that we found so far haveresidual polarity and the double mutants investigatedare lateralized at a ventrolateral level. However,despite the large number of available cactus alleles, it ispossible that we have not observed the amorphic cactusphenotype. Some of the zygotic lethal cactus alleleshave residual maternal activity. This could also apply tothe strongest zygotic lethal cactus allele that we tested ingermline chimeras. Furthermore, a currently unident-ified component might exist that acts similar to cactus,so that only the lack of both activities would result intotally ventralized embryos. Thus, the existing geneticdata do not exclude the possibility that the dorsal groupgenes act via cactus.

If, as assumed in the first model, the dorsal groupgenes act directly on dorsal, even in the completeabsence of cactus some residual polarity could beretained. Dorsal group activities might modify thedorsal protein not only to prevent its binding to cactus,but also to enhance the rate of its nuclear uptake. Whilethe unmodified dorsal protein, when released from itscomplex with cactus, would reach only intermediatenuclear concentrations, the highest nuclear proteinlevels would require a modification of dorsal protein. Anuclear dorsal protein gradient with a high pointventrally could exist then even in the absence of cactusactivity.

The cytoplasmic localization of dorsal protein atventral positions caused by the neomorphic cactusalleles can be explained in both models. The alteredcactus proteins derived from these alleles may lack theability to interact with dorsal group products, so thatthey cannot be modified and, hence, do not releasebound dorsal protein. Alternatively, the proteinsderived from neomorphic cactus alleles may bind thedorsal protein in a way that the modification of dorsalprotein by dorsal group activities is blocked or that inspite of the modification no release from the complexwith cactus occurs. Both alternatives explain why theneomorphic cactus alleles prevent the dorsal groupproducts from stimulating the nuclear localization ofdorsal protein.

Developmental genetics of cactus 387

Some results of our double mutant analysis imposecertain constraints on models explaining the action ofthe dorsal group genes. The double mutants ofamorphic dorsal group mutants with cactus alleles ofdifferent strength lead to apolar lateralized embryoswhich express ventrolateral or dorsolateral structures(LI, L2, L3; Fig. 7A-D, Fig. 8). These double mutantsdemonstrate that all major pattern elements present inthe dorsal half of the embryonic circumference can beexpressed in an apolar fashion upon reduction of cactusactivity without any change in the activity of dorsalgroup products. Further, if intermediate levels of dorsalgroup activity corresponding to ventrolateral structuresare present like in TFm9, the additional reduction ofcactus activity in cact;TT"9 double mutants, causes theformation of the ventralmost structure (mesoderm,Fig. 9). The examples show that the different nuclearconcentrations of dorsal protein need not be in a one-to-one correspondence with different levels of dorsalgroup activity. Rather, they can be generated in twodistinct ways, either by a change in activity of cactus orof the dorsal group. These results can be easilyexplained using the second model of the doublenegative serial action of dorsal group genes and cactus,however they do not exclude the first model. Further-more, both mechanisms do not exclude each other;thus, the possibility exists that they are realizedsimultaneously. The dorsal group genes could act onboth cactus and dorsal protein to prevent complexformation and to stimulate nuclear import respectively.

In the case of NF-KB/I -KB, it was shown that thedissociation of the cytoplasmic complex is mediated bya protein modification of the inhibitor. In response to avariety of extracellular signals, IJCB is phosphorylated,presumably by protein kinase C (Ghosh and Baltimore,1990"). This phosphorylation causes the release ofp50/p65 which then enters the nucleus and activatestranscription. The molecular analysis of cactus willelucidate whether IKB has structural similarities tocactus, and, if so, whether functional similarities alsoexist.

We are very grateful to T. Schupbach for the gift of theoriginal cactus alleles and to D. St Johnston for the discoveryof cact8®. We thank J. Habeck for help in preparing sections,E. Vogelsang for help with pole cell transplantations, C.Rushlow for the gift of antibodies. We thank R. Geisler, M.Mullins, M. Leptin, D. Stein, and L. Stevens for stimulatingdiscussion and suggestions on the manuscript. R. Groemke-Lutz prepared photographs. The contribution by Y. H. wasdone while he was a postdoctoral fellow with Corey S.Goodman. Y. H. would like to thank C. S. G. for support,him and K. V. Anderson for stimulating discussions. Thiswork was supported by the DFG (Leibniz program). Y. H.was a senior postdoctoral fellow of the American CancerSociety.

References

ANDERSON, K. V. (1987). Dorsal-ventral embryonic pattern genesof Drosophila. TIG 3, 91-97.

ANDERSON, K. V., JORGENS, G. AND NOSSLEIN-VOLHARD, C.

(1985). Establishment of dorsal-ventral polarity in theDrosophila embryo: genetic studies on the role of the Toll geneproduct. Cell 42, 779-789.

ANDERSON, K. V. AND NOSSLEIN-VOLHARD, C. (1984). Informationfor the dorsal-ventral pattern of the Drosophila embryo isstored as maternal mRNA. Nature 311, 223-227.

ANDERSON, K. V. AND NOSSLEIN-VOLHARD, C. (1986). Dorsal-group genes of Drosophila. In Gametogenesis and the EarlyEmbryo (ed. J. Gall) pp. 177-194. New York: Alan R. Liss,Inc.

ASHBURNER, M . , THOMSON, P . , ROOTE, J . , L A S K O , P. F . , GRAU,

Y., E L MESSAL, M., ROTH, S. AND SIMPSON, P. (1990) Thegenetics of a small autosomal region of Drosophilamelanogaster, including the structural gene for Alcoholdehydrogenase. VII. Characterization of "the region around thesnail and cactus loci. Genetics 126, 679-694.

BAEUERLE, P. A. AND BALTIMORE, D. (1988). IR-B: a specificinhibitor of the NF-xrB transcription factor. Science 242,540-546.

BUSSON, B. , GANS, M., KOMITOPOULOU, K. AND MASSON, M

(1983). Genetic analysis of three dominant female sterilemutations on the X-chromosome of Drosophila melanogaster.Genetics 105, 309-325.

CAMPOS-ORTEGA, J. A. AND HARTENSTEIN, V. (1985). The

Embryonic Development of Drosophila melanogaster. Berlin:Springer.

CARROLL, S. B. , WINSLOW, G. M., TWOMBLY, V. J. AND SCOTT, M.

P. (1987). Genes that control dorsoventral polarity affect geneexpression along the anteroposterior axis of the Drosophilaembryo. Development 99, 327-332.

CHASAN, R. AND ANDERSON, K. V. (1989). The role of easter, anapparent serine protease, in organizing the dorsal-ventralpattern of the Drosophila embryo. Cell 56, 391-400.

DELOTTO, R. AND SPIERER, P. (1986). A gene required for thespecification of dorsal-ventral pattern in Drosophila appears toencode a serine protease. Nature 323, 688-692.

FROHNHOFER, H. G. AND NOSSLETN-VOLHARD, C. (1986).

Organisation of anterior pattern in the Drosophila embryo bythe maternal gene bicoid. Nature 324, 120-125.

GHOSH, S. AND BALTIMORE, D. (1990). Activation in vitro of NF-fcB by phosphorylation of its inhibitor IkB. Nature 344, 678-682.

GHOSH, S., GIFFORD, A. M., RIVIERE, L. R., TEMPST, P., NOLAN,

G. P. AND BALTIMORE, D. (1990). Cloning of the p50 DNAbinding subunit of NF-*rB: homology to rel and dorsal. Cell 62,1019-1029.

HASHIMOTO, C , HUDSON, K. L. AND ANDERSON, K. V. (1988).

The Toll gene of Drosophila, required for dorsal-ventralembryonic polarity, appears to encode a transmembrane protein.Cell 52, 269-279.

IRISH, V. F. AND GELBAJTT, W. M. (1987). The decapentaplegicgene is required for dorsal-ventral patterning of the Drosophilaembryo. Genes Dev. 1, 868-879.

KIERAN, M., BLANK, V., LOCEAT, F., VANDEKERCKHOVE, J.,

LOTTSPEICH, F., LE BAIL, O., URBAN, M. B., KOURILSKY, P.,BAEUERLE, P. A. AND ISRAEL, A. (1990). The DNA bindingsubunit of NF-JCB is identical to factor KBF1 and homologous tothe rel oncogene product. Cell 62, 1007-1018.

KONRAD, K. D., GORALSKI, T. J. AND MAHOWALD, A. P. (1988).

Developmental genetics of the gastrulation defective locus inDrosophila melanogaster. Devi Biol. 127, 133-142.

LEHMANN, R. AND NOSSLEIN-VOLHARD, C. (1986). hunchback, agene required for segmentation of an anterior and posteriorregion of the Drosophila embryo. Devi Biol. 119, 402-417.

LENARDO, M. J. AND BALTIMORE, D. (1989). NF-*rB: a pleiotropicmediator of inducible and tissue-specific gene control. Cell 58,227-229.

LEPTIN, G. AND GRUNEWALD, B. (1990). Cell shape changes duringgastrulation in Drosophila. Development 110, 73-84.

MACDONALD, P. M. AND STRUHL, G. (1986). A molecular gradientin early Drosophila embryos and its role in specifying the bodypattern. Nature 324, 537-545.

NOSSLEIN-VOLHARD, C. (1979a). Maternal effect mutations thatalter the spatial coordinates of the embryo of Drosophilamelanogaster. In Determinants of Spatial Organization (ed. I.

388 S. Roth, Y. Hiromi, D. Godt and C. Nilsslein-Volhard

Koenigsberg and S. Subtelney), pp. 185-211. New York:Academic Press.

NOSSLEIN-VOLHABD, C. (1979b). Pattern mutants in Drosophilaembryogenesis. In Cell Lineage, Stem Cells and CellDetermination. INSERM Symposium No. 10, (ed. N. LeDouarin) pp. 69-82. New York: North-Holland.

N0SSLE1N-VOLHARD, C , FROHNHOFER, H . G . AND LEHMANN, R.(1987). Determination of anteroposterior polarity in Drosophila.Science 238, 1675-1681.

NOSSLEIN-VOLHARD, C , LOHS-SCHARDIN, M . , SANDER, K. ANDCREMER, C. (1980). A dorso-ventral shift of embryonicprimordia in a new matemal-effect mutant of Drosophila.Nature 283, 474-476.

NOSSLEIN-VOLHARD, C. AND ROTH, S. (1989). Axis determinationin insect embryos. In Cellular Basis of Morphogenesis. CibaFoundation Symposium 144. pp. 37-55. New York: John Wileyand Sons.

NOSSLEIN-VOLHARD, C , WIESCHAUS, E. AND KLUDING, H. (1984).Mutations affecting the pattern of the larval cuticle inDrosophila melanogaster. I. Zygotic loci on the secondchromosome. Wilhelm Roux's Arch, devl Biol. 183, 267-282.

ROTH, S. (1990). Die Rolle der maternalen Gene cactus und dorsalbei der dorsoventralen Musterbildung im Drosophila Embryo.PhD Thesis. University of Tubingen.

ROTH, S., STEIN, D. AND NOSSLEIN-VOLHARD, C. (1989). Agradient of nuclear localization of the dorsal protein determinesdorsoventral pattern in the Drosophila embryo. Cell 59,1189-1202.

RUSHLOW, C , FRASCH, M., DOYLE, H. AND LEVINE, M. (1987).Maternal regulation of zerknUllt: a homoeobox gene controllingdifferentiation of dorsal tissues in Drosophila. Nature 330,583-586.

RUSHLOW, C. A., HAN, K., MANLEY, J. L. AND LEVINE, M.(1989). The graded distribution of the dorsal morphogen isinitiated by selective nuclear transport in Drosophila. Cell 59,1165-1177.

SANTAMARIA, P. AND NOSSLEIN-VOLHARD, C. (1983). Partial rescueof dorsal, a maternal effect mutation affecting the dorso-ventralpattern of the Drosophila embryo, by the injection of wild-typecytoplasm. EMBO J. 2, 1695-1699.

SCHOPBACH, T. AND WIESCHAUS, E. (1986). Germline autonomy ofmaternal-effect mutations altering the embryonic body patternof Drosophila. Devi Biol. 113, 443-448.

SCHOPBACH, T. AND WIESCHAUS, E. (1989). Female sterilemutations on the second chromosome of Drosophilamelanogaster. I. Maternal effect mutations. Genetics 121,101-117.

SEIFERT, E., MOLLER-HOLTKAMP, F., MARCEY, D. AND JACKLE, H.(1987). Injection of wild-type cytoplasm and poly(A) + RNAprovokes phenotype rescue in sptitzle mutant Drosophilaembryos. Roux's Arch, devl Biol. 196, 78-82.

SPARROW, J. C. (1978). Melanotic 'Tumours'. In The Genetics andBiology of Drosophila. vol. 2b (ed. M. Ashburner and T. R. F.Wright), pp. 277-313. New York und London: Academic Press.

STEWARD, R. (1987). Dorsal, an embryonic polarity gene inDrosophila, is homologous to the vertebrate proto-oncogene, c-rel. Science 238, 692-694.

STEWARD, R. (1989). Relocalization of the dorsal protein from thecytoplasm to the nucleus correlates with its function. Cell 59,1179-1188.

STEWARD, R. AND NOSSLEIN-VOLHARD, C. (1986). Tha genetics ofthe dorsal-Bicaudal-D region of Drosophila melanogaster.Genetics 113, 665-^78.

STEWARD, R., ZUSMAN, S. B., HUANG, L. H. AND SCHEDL, P.(1988). The dorsal protein is distributed in a gradient in earlyDrosophila embryos. Cell 55, 487-495.

ST JOHNSTON, R. D. AND GELBART, W. M. (1987). Decapentaplegictranscripts are localized along the dorsal-ventral axis of theDrosophila embryo. EMBO J. 6, 2785-2791.

THISSE, B., STOETZEL, C , EL MESSAL, M. AND PERRIN-SCHMITT, F.(1987). Genes of the Drosophila maternal dorsal group controlthe specific expression of the zygotic gene twist in presumptivemesodermal cells. Genes Dev. 1, 709-715.

THISSE, B., STOETZEL, C , GOROSTIZA-THISSE, C. AND PERRIN-SCHMITT, F. (1988). Sequence of the twist gene and nuclearlocalization of its protein in endomesodermal cells of earlyDrosophila embryos. EMBO J. 7, 2175-2183.

VAN DER MEER, J. M. (1977). Optical clean and permanent wholemount preparations for phase-contrast microscopy of cuticularstructures of insect larvae. Dros. Info. Serv. 52, 160.

WAKIMOTO, B. T., TURNER, F. R. AND KAUFMAN, T. C. (1984).Effects in embryogenesis in mutants associated with theantennapedia gene complex of Drosophila melanogaster. DevlBiol. 102, 147-172.

ZABEL, U. AND BAEUERLE, P. A. (1990). Purified human IATB canrapidly dissociate the complex of the NF-irB transcription factorwith its cognate DNA. Cell 61, 255-265.

ZUSMANN, S. AND WIESCHAUS, E. (1985). Requirements for zygoticgene activity during gastrulation in Drosophila melanogaster.Devl Biol. I l l , 359-371.

{Accepted 20 March 1991)