Development 1989 Supplement, 13-19Printed in Great Britain © The Company of Biologists Limited 1989

13

Multiple steps in the localization of bicoid RNA to the anterior pole of the

Drosophila oocyte

DANIEL ST. JOHNSTON, WOLFGANG DRIEVER, THOMAS BERLETH, SIBYLL RICHSTEIN and

CHRISTIANE NOSSLEIN-VOLHARD

Max Planck Institut fitr EnlwickJungsbiologie, Abteilung Cenetik, Spemannstrasse 35, 7400 Tubingen, FRG

Summary

The anterior region of the Drosophila embryonic patternis determined by a gradient of the bicoid (bed) protein.The correct formation of this gradient requires thelocalization of bed RNA to the anterior pole of the egg.Here we use a wholemount in situ technique to examinethe process of bed RNA localization during oogenesis andembryogenesis. While bed protein becomes distributedin a gradient that extends throughout the anterior twothirds of the early embryo, bed RNA remains restrictedto a much smaller region at the anterior pole. Thedifference between these distributions indicates that theshape of the protein gradient must depend to some extenton the posterior movement of the protein after it hasbeen synthesized.

Four distinct phases of bed RNA localization can bedistinguished during oogenesis. Between stages 6 and 9of oogenesis, the RNA accumulates in a ring at theanterior end of the oocyte. During the second phase, instage 9-10a follicles, the RNA also localizes to the apicalregions of the nurse cells, demonstrating that the nurse

cells possess an intrinsic polarity. As the nurse cellscontract during stages 10b-l l , all of the bed RNAbecomes localized to the cortex at the anterior end of theoocyte. During a final phase that must occur betweenstage 12 of oogenesis and egg deposition, the RNAbecomes localized to a spherical region that occupies aslightly dorsal position at the anterior pole.

Mutations in the maternal-effect genes, exuperantia(exu) and swallow (sww), lead to an almost uniformdistribution of bed RNA in the early embryo, whilestaufen (stau) mutations produce a gradient of RNA atthe anterior pole, exu mutations disrupt the second stageof bed RNA localization during oogenesis, sww mutationsdisrupt the third, and stau mutations affect the fourthphase.

Key words: Drosophila, bicoid, RNA localization,oogenesis, exuperantia, swallow, staufen, gradientformation.

Introduction

The anterior-posterior pattern of the Drosophila em-bryo is determined in response to maternal factors thatare deposited in the egg during oogenesis. Three groupsof maternal genes are required to specify distinctregions of this pattern; the head and thorax (theanterior group), the abdomen (the posterior group) andthe acron and telson (the terminal group) (reviewed inNusslein-Volhard and Roth, 1989; Nusslein-Volhard etal. 1987). The anterior portion of the body plan dependson the product of the bicoid (bed) gene (Frohnhoferand Nusslein-Volhard, 1986). In bicoid~ embryos, thehead and thorax are absent, and the abdominal fatemap expands anteriorly. In addition, a duplicatedtelson forms at the anterior end of the embryo. Trans-plantation experiments have shown that bcd+ activity islocalized to the anterior pole of the egg, and whentransplanted to more posterior positions, can induce theformation of anterior structures at ectopic sites (Frohn-

h6fer and Nusslein-Volhard, 1986; Frohnhofer et al.1986).

In agreement with the results of the transplantationexperiments, bed RNA is localized to the anterior poleof the egg (Frigerio et al. 1986; Berleth et al. 1988). Theprotein that is translated from this RNA forms aconcentration gradient which extends throughout theanterior two thirds of the embryo (Driever and Nuss-lein-Volhard, 1988fl). This protein gradient appears todetermine anterior positional values in a concentration-dependent manner, since changes in the maternal bcd+

gene dosage produce complementary shifts in both theprotein gradient and the fate map (Driever and Nuss-lein-Volhard, 19886). Further support for the role ofbed protein as the anterior morphogen comes from theresults of experiments in which bed RNA synthesized invitro is injected into early embryos. When the RNA isinjected into the middle of a recipient embryo, it caninduce the development of ectopic head and thoracicstructures (Driever, Siegel and Nusslein-Volhard, in

14 D. St. Johnston and others

preparation). The most anterior pattern elements formclosest to the site of injection, with more posterior(thoracic) structures developing on either side. Thisresult indicates not only that the bed protein determinesthe anterior quality of the structures that develop, butalso that the slope of the protein gradient specifies thepolarity of the pattern. In addition, the ability of bedRNA to organize anterior pattern in the middle of theembryo shows that no other anteriorly localized mol-ecules are required for this process.

The bicoid protein contains a homeodomain, sugges-ting that bicoid encodes a DNA-binding protein (Fri-gerio et al. 1986). This makes it attractive to supposethat bed protein specifies anterior positional values byactivating or repressing zygotic target genes in a con-centration-dependent manner. Genetic and molecularevidence indicates that bed regulates the anterior zygo-tic expression of the gap gene, hunchback (Lehmannand Nusslein-Volhard, 1987; Schroder et al. 1988;Tautz, 1988). bed protein binds to several sites in thehunchback promoter, and acts as a transcriptionalactivator of hunchback (Driever and Nusslein-Volhard,1989). Driever et al. (in press) have made constructscontaining synthetic-bed-binding sites fused to thehsp70 promoter and a reporter gene. When trans-formed into flies, the bed-binding sites drive the an-terior expression of the reporter gene in the embryo.The size of this anterior domain of expression isreduced when binding sites with lower affinities for thebed protein are used. The observation that promoterswith low-affinity bed-binding sites are only activated athigh protein concentrations suggests a model for howthe bed protein gradient could activate several zygotictarget genes in distinct anterior domains, and therebydetermine several levels of anterior development.

The formation of the wild-type protein gradient, andthus the determination of a normal anterior pattern,depends upon the localization of bed RNA to theanterior pole. Mutations in the maternal genes exuper-antia (exu) and swallow (sww) disrupt this localizationduring oogenesis, and lead to an almost uniform distri-bution of bed RNA in the early embryo (Berleth et al.1988; Stephenson et al. 1988). Both of these mutationsresult in phenotypes in which anterior structures of thehead are deleted, and the thoracic region is expanded(Frohnhofer and Nusslein-Volhard, 1987). Embryosderived from females mutant for the posterior groupgene staufen show similar but weaker head defects, inaddition to the abdominal deletions characteristic of allmutations in this class (Schiipbach and Wieschaus,1986; Nusslein-Volhard et al. 1987; Lehmann, 1988).These embryos also show an anterior shift in the fatemap, which can be seen in a shift in the position of thecephalic furrow and the first fushi tarazu stripe (Schiip-bach and Wieschaus, 1986; Carroll et al. 1986; Leh-mann, 1988). The observation that the bed proteingradient is shallower in the mutant embryos suggeststhat staufen mutations may also alter the distribution ofbed RNA (Driever and Nusslein-Volhard, 19886).

In this report, we use a non-radioactive, enzyme-linked in situ technique developed by Tautz and Pfeifle

(1989) to examine the process of bed RNA localizationduring oogenesis and early embryogenesis. This tech-nique has the advantage that one can perform hybridiz-ations on wholemount preparations. We have usedthese wholemount stainings to make direct comparisonsbetween the bed RNA distribution and the proteindistribution, as revealed by antibody stainings.

Results

bed RNA localization in wild-type embryosThe localization of bed RNA in wild-type embryos haspreviously been described by Frigerio et al. (1986) andBerleth et al. (1988). We have repeated these investi-gations in order to gain a clearer understanding of thethree-dimensional distribution of bed RNA during thefirst stages of embryogenesis. In very early (stage 1)embryos, bed RNA staining resembles a flattened ballwhich is closely apposed to the anterior pole, and whichfrequently occupies a slightly dorsal position (Fig. 1A).The RNA does not seem to be specifically bound to thecortex of the egg, since much of it is in the interior. Asdevelopment proceeds through pole cell formation(Fig. IB) to syncytial blastoderm (Fig. 1C), most of theRNA becomes localized to the periphery, in the clearcytoplasm that surrounds each nucleus. This movementto the cortex sometimes results in a slight posterior-wards shift in the RNA distribution. By early nuclearcycle 14, bed RNA begins to disappear (Fig. ID) andmidway through cellularization the signal is no longerdetectable (Fig. IE).

Fig. 1. The distribution of bed RNA in wild-type embryos.The embryos were fixed, and hybridized with a random-primed probe synthesized from a bicoid cDNA clone,following the procedure of Tautz and Pfeifle (1989). In allfigures, anterior is to the left and dorsal is uppermost.(A) Early cleavage stage. (B) Late cleavage stage, afterpole cell formation. (C) Syncytial blastoderm. (D) Latesyncytial blastoderm, after the 13th nuclear division.(E) Early cellularization. These embryos have beenoverstained in order to clearly show the posterior extent ofthe bed RNA distribution. This overstating partiallyobscures the cortical localization of the RNA in syncytialblastoderm embryos (C), which can be clearly seen inunderstated embryos or in later embryos as the RNAstarts to disappear (D).Fig. 2. The distributions of bicoid RNA, bicoid protein andhunchback RNA in syncytial blastoderm embryos. (A) bedRNA (B) bed protein. The embryos were stained with apolyclonal anti-bicoid antibody, as described by Driever andNusslein-Volhard (1988a). (C) hunchback RNA (Tautz et al.1987).Fig. 4. The bed RNA distribution in mutant embryos.(A) Wild-type cleavage stage. (B) Cleavage-stage embryoderived from an exu /exu mother, showing a uniformdistribution of bed RNA. (C) Syncytial blastoderm embryoderived from an exuPJ/exuPJ mother. By this stage ashallow anterior-posterior gradient of RNA has formed.(D) Cleavage-stage embryo derived from a swwN/swwN

mother, showing the early weak RNA gradient.(E) Cleavage-stage embryo derived from a siauD3/Df(2R)PC4 mother. The RNA is distributed in a steep gradient atthe anterior end of the embryo.

LXJ

HI

6A

C Fig.

6.

The

dis

trib

utio

n of

bed

RN

A d

urin

g w

ild-ty

pe o

ogen

esis

. (A

) Pa

rt o

f an

ova

riol

e co

ntai

ning

a s

tage

-6

folli

cle

with

a r

ing

of b

ed R

NA

in

the

pres

umpt

ive

oocy

te.

The

adj

acen

t st

age

7 an

d st

age

9 fo

llicl

es a

lso

show

bed

RN

A s

tain

ing.

(B

) A

sta

ge 8

fol

licle

with

bed

RN

A l

ocal

ized

to

the

ante

rior

mar

gins

of

the

oocy

te.

A w

eak

sign

al c

an a

lso

be s

een

in t

he n

urse

cel

ls.

(C)

A s

tage

9 f

ollic

le s

how

ing

the

apic

allo

caliz

atio

n in

the

15

nurs

e ce

lls,

in a

dditi

on t

o th

e an

teri

or s

igna

l in

the

ooc

yte.

The

sta

ge 8

fol

licle

to

the

left

also

sho

ws

the

ante

rior

rin

g of

RN

A.

(D)

A s

tage

10a

fol

licle

. (E

) A

sta

ge 1

0b f

ollic

le.

The

nur

se c

ell

loca

lizat

ion

is d

isap

pear

ing,

as

the

nurs

e ce

lls c

ontr

act

and

the

bed

RN

A a

ccum

ulat

es a

t th

e an

teri

or p

ole

ofth

e oo

cyte

. (F

) A

sta

ge 1

2 fo

llicl

e. T

he n

urse

cel

ls a

re d

egen

erat

ing,

and

all

bed

RN

A i

s no

w l

ocal

ized

to

the

cort

ex a

t th

e an

teri

or o

f th

e oo

cyte

.Fi

g. 7

. T

he d

istr

ibut

ion

of b

ed R

NA

in

mut

ant

ovar

ies.

(A

) ex

uPJ,

stag

e 7-

8. T

he R

NA

for

ms

a ri

ng a

t th

ean

teri

or e

nd o

f th

e oo

cyte

, but

thi

s ri

ng a

ppea

rs m

ore

diff

use.

(B

) ex

u , s

tage

10a

. T

he R

NA

sho

ws

aun

iform

dis

trib

utio

n in

the

cyt

opla

sm o

f th

e nu

rse

cells

. (C

) sw

w14

', st

age

9. T

he R

NA

loc

aliz

es n

orm

ally

inth

e nu

rse

cells

and

the

ooc

yte.

(D

) sw

w1'4,

stag

e 10

b. T

he c

ortic

al b

ed R

NA

is

no l

onge

r lo

caliz

ed t

o th

em

ost

ante

rior

end

of

the

oocy

te b

ut i

nste

ad e

xten

ds p

oste

rior

ly.

Muc

h of

the

RN

A e

nter

ing

the

oocy

te a

tth

is s

tage

is

not

loca

lized

to

the

cort

ex,

and

lead

s to

wea

k st

aini

ng i

n th

e an

teri

or t

hird

of

the

oocy

te.

(E)

stai

i*',

stag

e 12

. (F

) st

auD

3, st

age

12. T

he b

ed R

NA

loc

aliz

atio

n is

com

plet

ely

norm

al a

t th

is s

tage

. W

eha

ve a

lso

exam

ined

exu

QR a

nd e

xw

ovar

ies

whi

ch s

how

the

sam

e ph

enot

ypes

as

exuF

J.

Localization of bicoid RNA 15

Fig. 2 compares the distributions of bed RNA andprotein at the syncytial blastoderm stage. The levels ofRNA and protein were quantified using the proceduredescribed by Driever and Nusslein-Volhard (1988a),and these results are presented graphically in Fig. 3.These measurements clearly demonstrate that the RNAis more tightly localized to the anterior end of theembryo, than is the protein, bed RNA forms a steepgradient at the anterior end of the embryo, in which90 % of the RNA is restricted to the region anterior to82% egg length (0% EL is the posterior pole). Incontrast, bed protein is distributed in a much shallowergradient, in which only 57% of the protein falls withinthis anterior region. The protein must therefore moveposteriorly after it has been synthesized. One of thefunctions of bed protein is to activate the anteriorzygotic expression of hunchback (SchrSder et al. 1988;Tautz, 1988; Driever and Nusslein-Volhard, 1989). Thisanterior hunchback domain is shown in Fig. 2C, andillustrates that the bed protein gradient must extend toat least 55 % egg length, a position at which bed RNA isnot detectable above background.

1200

1000

8 0 0 -

— 6 0 0 "

4 0 0 "

2 0 0 -

bed protein

bed RNA

80 60 40 20

position (% egg length)

Fig. 3. The distribution of bed mRNA (solid line) and bedprotein (dotted line) in the embryo. Whole wild-typeembryos were stained for bed protein, using animmunohistochemical method (Driever and Nusslein-Volhard, 1988a), or for bed mRNA, using the enzymelinked in situ detection technique. Video images of whole-mount embryos were taken and a background imagesubtracted. The distribution of the stain intensities alongthe anterior posterior midline was recorded for fiveembryos (nuclear cycle 13) and the average values at 30equidistant positions calculated. Signal linearity of the insitu detection method was tested by performing theenzymatic colour reaction for 3 and a half or 16 minutes,respectively. Values in the anteriormost 15% of theembryos appear to be nonlinear in the 16min reaction andwere corrected accordingly. The areas under the RNA andprotein curves have been made equal in order to allow acomparison of the two distribution profiles. Anterior is tothe left.

bed RNA localization in mutant embryosIn early (stage 1) embryos derived from exu homo-zygous mothers, bed RNA is uniformly distributedthroughout the egg cytoplasm (Fig. 4B). However, bysyncytial blastoderm a shallow anterior-posteriorgradient has formed (Fig. 4C). Berleth et al. (1987)have found that bed RNA remains uniformly distrib-uted in eggs laid by mothers which are mutant for bothexu and the posterior group gene vasa. This suggeststhat the late bed RNA gradient is created by thedegradation of the RNA in the posterior region of theembryo, due to the activity of the posterior organizingcentre (Nusslein-Volhard et al. 1987).

Embryos laid by sww homozygotes have a morevariable distribution of bed RNA. While some embryosshow no localization, most contain a weak anterior-to-posterior gradient (Fig. 4D). As is the case for exumutant embryos, the gradient becomes more pro-nounced as development proceeds, due to the posteriordegradation of the RNA. The variability of the bedRNA distribution in sww mutant embryos is reflected inthe variability in the final cuticular phenotype (Frohn-hdfer and Nusslein-Volhard, 1986).

In embryos derived from staufen mutant mothers,bed RNA forms a gradient in the anterior region of theembryo (Fig. 4E). This phenotype is clearly distinctfrom that produced by exu or sww mutations, andcannot depend upon the posterior degradation of theRNA, since there is no localized posterior activity instau mutant embryos (Nusslein-Volhard et al. 1987;Lehmann and Nusslein-Volhard, unpublished results).This partial mislocalization of bed RNA leads to ashallower protein gradient (Driever and Nusslein-Vol-hard, 1988b). The bed RNA distributions in exu, sww,and stau mutant embryos are compared to the wild-typedistribution in Fig. 5. Since these comparisons wereperformed using syncytial blastoderm embryos, exu andsww both show a shallow anterior-to-posterior RNAgradient.

The process of bed RNA localization during wild-typeoogenesisbed RNA can first be detected in late previtellogenicfollicles (stages 5-6 of King (1970)) accumulating in asingle, posteriorly located cell of the germ cell cluster(Fig. 6A). As oogenesis proceeds and the oocyte growslarger than the fifteen nurse cells, it becomes clear thatthis cell is the oocyte. At these stages, very little RNAcan be seen in the nurse cells, and the RNA forms a ringaround the anterior margin of the developing oocyte(Fig. 6B). This ring increases in size as the folliclegrows. During stages 9-10a, large amounts of bed RNAaccumulate in the nurse cells (Fig. 6C,D). Unlike othermaternal RNAs, bed RNA is not uniformly distributedwithin the nurse cell cytoplasm, but is concentrated in aperipheral region adjacent to each nurse cell nucleus.During stages 10b and 11, the localization within thenurse cells gradually disappears, as the nurse cellscontract and transfer their cytoplasm to the oocyte(Fig. 6E). Upon entering the oocyte, the bed RNAbecomes localized to the cortex of the anterior pole. At

16 D. St. Johnston and others

100 60 40 20

position (% egg length)

Fig. 5. Distribution of bed mRNA in wild-type embryosand in embryos from females mutant for exu, sww and stau.The distribution of bed mRNA in early nuclear cycle 13embryos was visualized and measured as described in thelegend to Fig. 3. The RNA distribution profiles of embryosderived from mutant females (solid lines; (A) exuPJ/exuPJ\(B) SWWN/SWH-"\ (C) stauD3/Df(2R) PC4) were plottedtogether with those of wild-type embryos (dotted lines;(A and B) wild-type controls stained in parallel to themutant embryos; (C) wild-type embryos stained in the samebatch as the stau mutant embryos, which were identified bythe lack of polecells).

this stage, the RNA is no longer restricted to a ringaround the anterior pole, and instead covers most of theanterior end of the oocyte. By stage 12, when the nursecells are degenerating, bed RNA is localized in a cap atthe anterior end of the egg, with more of the RNAbeing found ventrally than dorsally (Fig. 6F). During

the final few hours of oogenesis, stages 13 and 14, thefollicle cells that surround the oocyte secrete thechorion (King, 1970). We have been unable to analyzebed RNA localization during these stages because thevitelline membrane and chorion prevent the entry ofthe probe into the oocyte. However, the distribution ofthe RNA in stage 12 oocytes is different from thatobserved in very young eggs. In the oocytes, the RNA islocalized to the cortex and is more concentrated ven-trally, whereas in the early embryo, the RNA is foundin a spherical region that extends into the interior of theegg and which is often located slightly dorsally. Thesedifferences suggest that there is a redistribution of bedRNA, either during stages 13 and 14 of oogenesis, orimmediately after fertilization.

bed RNA localization in mutant ovariesIn order to understand how the mutations that alter thebed RNA distribution in the embryo affect the fourphases of RNA localization during oogenesis, we haveperformed in situ hybridizations on exu, sww and staumutant ovaries. The earliest differences from wild-typebed RNA localization are observed in exu mutantovaries. The initial accumulation of bed RNA duringstages 5-7 occurs normally. As the oocyte increases insize, the RNA still forms a ring at the anterior end, butthis ring often appears more diffuse (Fig. 7A). The firstmajor deviation from normal oogenesis becomes appar-ent during stage 10a, when bed RNA is not localized tothe apical regions of the nurse cells, and instead shows auniform distribution in the nurse cell cytoplasm(Fig. 7B). When this RNA enters the oocyte, it is notretained at the anterior pole, and from stage 10 onwardsno localization within the oocyte is visible.

In the ovaries of sww homozygous females, theprocess of bed RNA localization appears entirely nor-mal up to stage 10a. The ring of RNA forms at theanterior end of the oocyte, and RNA accumulates in theapical regions of the nurse cells (Fig. 7C). During stages10b and 11, the anterior ring of RNA seems to slipposteriorly and become more diffuse (Fig. 7D). Inaddition, much of the RNA that enters the oocyteduring this time is not localized to the cortex. By stage12, all bed RNA appears to have been released from thecortex and forms a shallow anterior-to-posterior gradi-ent.

In homozygous stau ovaries, the process of bed RNAlocalization appears completely normal up until stage12 (Fig. 7E,F), the latest stage that we have examined.Since the RNA is distributed in a gradient in earlyembryos, it must be released from the anterior poleafter stage 12 of oogenesis.

Discussion

The experiments presented in this report use a whole-mount in situ technique to provide a detailed picture ofthe process of bed RNA localization during oogenesisand embryogenesis. These results confirm and extendthe previous analyses of bed RNA localization, which

exu sww

Localization of bicoid RNA

stau

17

stage 13, 14fertilization —egg deposition

Stage 8 Stage 9 Stage 10b Stage 12 embryo

Fig. 8. A drawing showing the four phases of bed RNA localization during oogenesis and the points at which exu, sww, andstau are required. The drawings are based on King (1972).

were performed using radioactive probes on sectionedmaterial (Frigerio et al. 1986; Berleth et al. 1987;Stephenson et al. 1988). The wholemount procedureallows direct comparisons between the bed RNA andprotein distributions. As shown in Figs 2 and 3, the twodistributions are quite different. Given that no RNAcan be detected posterior to 60% egg length even inoverstained embryos, whereas the protein gradientextends at least to 30 % EL, the shape of the proteingradient must depend to some extent on the movementof protein molecules towards the posterior after theyhave been translated. Simple diffusion of the proteincan account for this distribution (Driever and Nusslein-Volhard, 1988a).

The in situ hybridizations to embryos derived fromstaufen mutant females reveal that, like exuperantia andswallow, staufen is required for the correct localizationof bed RNA to the anterior pole. However, staufenmutations only cause a partial mislocalization of bedRNA. The anterior RNA gradient produced by thesemutations results in a shallower bed protein distri-bution, in which the anterior levels of bed protein arestrongly reduced compared to wild-type, and moreposterior levels are slightly increased (Driever andNilsslein-Volhard, 1988/?). The anterior reduction inbed protein concentration accounts for the loss ofanterior head structures observed in staufen embryos(Schiipbach and Wieschaus, 1986). In addition to thehead defects, staufen mutations produce a typical pos-terior group phenotype in which the abdomen is deletedand pole cells do not form. The phenotypes produced indouble mutant combinations between staufen and othermaternal effect mutations strongly suggest that theposterior effects of staufen are due to a failure totransport pole plasm constituents to the posterior pole(R. Lehmann, personal communication). Thus thestaufen gene product is implicated in the localization ofmaternal factors to both the anterior and posteriorpoles of the egg.

Based on the geometry of the follicle, Frohnhofer

and Niisslein-Volhard (1987) have proposed a simplemodel for bed RNA localization, in which RNA enter-ing at the anterior end of the oocyte is trapped andattached to the cytoskeleton by factors that are uni-formly distributed in the egg. The present data suggestthat the process of localization is likely to be morecomplex. The in situ hybridizations to wild-type ovariesreveal at least four phases of bed RNA localization. Inthe first phase, which extends from stage 6 to early stage9 of oogenesis, bed RNA is found in a ring at theanterior end of the oocyte. During stages 9-10a, RNAis still found in this ring but bed RNA also accumulatesto high levels in the apical regions of the nurse cells.Stephenson et al. (1988) have also noticed the nonuni-form distribution of bed RNA in the nurse cells ofsectioned ovaries. The apical localization of bed RNAwithin the nurse cells is unusual since all other maternalRNAs that have been examined show a uniform distri-bution in the nurse cell cytoplasm, and no specializedcytological structures have been observed in theseregions (for example Kobayashi et al. 1988; Sprenger etal. 1989). In the third phase, during stages 10b-12, thenurse cell localization disappears and all bed RNAbecomes localized to the cortex at the anterior pole ofthe egg. A final phase of RNA redistribution must occurafter stage 12, to produce the spherical pattern of bedstaining seen in early embryos. The bed RNA distri-butions during these four phases are presented sche-matically in Fig. 8. The second and fourth phases of bedRNA localization cannot be explained by the simplemodel, since, during the second phase, the RNA islocalized in the nurse cells before it enters the egg, and,in the fourth phase, the RNA is redistributed within the

egg.Since exu mutations lead to a uniform distribution of

bed RNA in the nurse cell cytoplasm as well as in theoocyte, it seems likely that in wild-type ovaries bothlocalizations occur by similar mechanisms. The apicalregion of a nurse cell can be considered to be theanterior end of the cell, since it lies on the opposite side

18 D. St. Johnston and others

of the cell from the ring canals which connect to theother nurse cells and the oocyte. Thus bed RNA may betransiently localized to the anterior ends of the nursecells in a similar fashion to its localization within theoocyte, suggesting that the nurse cells also possess anintrinsic anterior-posterior polarity. The moleculesthat are required for bed RNA localization in the oocyteare most probably synthesized in the nurse cells. Thesemolecules could therefore mediate the transient RNAlocalization within the nurse cells, before they them-selves are transported into the oocyte. Since bed RNAis synthesized within the nurse cells, this localizationcannot depend upon the polar entry of the RNA intoone side of the cell. Thus the localization mechanismcan function, at least within the nurse cells, in theabsence of an asymmetric source of the RNA. If theprocess of localization within the egg is similar to that innurse cells, the anterior accumulation of bed RNAwithin the oocyte may involve a more active mechanismthan the simple trapping of the RNA as it enters at theanterior pole.

None of the mutations examined in this study com-pletely disrupts all phases of bed RNA localization. Theearliest phenotypes are seen in exu mutant ovaries, inwhich the RNA does not become restricted to the apicalregions of the nurse cells during stages 9-10a. exumutations also seem to have a weak effect on the firstphase of bed RNA localization, since the anterior ringoften appears more diffuse. Although we have usedthree different strong exu alleles, which all produce thesame phenotype, it is possible that none of these is anamorphic mutation. In an exu null genotype, even theearliest phase of bed RNA localization might be abol-ished. Berleth et al. (1988) have proposed that the exugene product binds directly to bed RNA, and mediatesits attachment to the cytoskeleton. Since exu mutationsabolish the localization in nurse cells, if exu does bindthe RNA, it must first do so within the nurse cells. Thisraises the possibility that bed RNA enters the oocyte aspart of a ribonucleoprotein complex that also containsexu protein. If this is the case, one might expect exuprotein to colocalize with bed RNA at the anterior poleof the embryo.

In sww mutant ovaries, the third phase of bed RNAlocalization is disrupted. As Berleth et al. (1988) andStephenson et al. (1988) have previously noted, theRNA is gradually released from the cortex of the oocyteduring stages 10b-ll of oogenesis. Since sww mutationsalso cause defects in nuclear migration and cellulariz-ation during embryogenesis, Frohnhofer and Niisslein-Volhard (1987) and Stephenson et al. (1988) havesuggested that the swallow protein is a component ofthe cytoskeleton. In sww mutants, the lack of thisprotein would cause a destabilization of the cytoskeletalelements that anchor bed RNA to the cortex of theoocyte and lead to a gradual release of the RNA. Ourobservation that sww mutations have no effect on theapical localization of the RNA in nurse cells providesupport for the idea that swallow encodes an oocyte-specific component of the cytoskeleton.

In staufen homozygotes, the process of bed RNA

localization appears completely normal up until stage12 of oogenesis, yet the RNA is not correctly localizedin early embryos. The difference between these distri-butions indicates that the RNA must be released fromthe anterior pole sometime between stage 12 and eggdeposition. This suggests that the staufen gene productmay be required for the movement of bed RNA fromthe anterior/ventral cortex to the more dorsally locatedspherical region observed in early embryos, and pro-vides further evidence for a fourth phase of bed RNAlocalization. The anterior gradient of RNA observed instaufen mutant embryos most probably results from thediffusion of the RNA during a short period of timebetween its release from the anterior pole and the startof embryogenesis. In general, there is a correlationbetween the stage at which a mutation affects bed RNAduring oogenesis and the distribution of the RNA inembryos, exu mutations show the earliest phenotype(stage 9) and result in a uniform distribution in theembryo, sww mutations disrupt localization duringstages 10-11 and lead to a very shallow embryonicRNA gradient, while staufen alleles seem to cause arelease of the RNA after stage 12, producing a muchsteeper gradient at the anterior pole.

This study has identified four phases of bed RNAlocalization during oogenesis, and has demonstratedthat exu, sww, and stau mutations each affect a differentstage of this process. At present, we still lack anyinformation on how this localization is achieved at abiochemical or cell biological level. Macdonald andStruhl (1988) have identified a 3' untranslated region ofof the bed RNA that is sufficient for localization to theanterior half of the embryo. As the molecular analysisof the trans-acting factors involved in this processadvances, we may discover how this region of the bedRNA is recognized, and what components of thecytoarchitecture of the oocyte participate in each phaseof the localization of bed RNA to the anterior pole.

We thank Maria Leptin, Leslie Stevens, Frank Sprengerand Dominique Ferrandon for their constructive criticisms ofthe manuscript, and Roswitha Gr6mke-Lutz and Doris Ederfor help with the Figures. This work was supported by anEMBO fellowship to D. St. J, and by the Leibniz program ofthe Deutsche Forschungsgemeinschaft.

References

BERLETH, T., BURRI, M., THOMA, G., BOPP, D., RICHSTEIN, S..

FRIGERIO, G., NOLL, M. AND NOSSLEIN-VOLHARD, C. (1988). Therole of bicoid RNA in organizing the anterior pattern of theDrosophila embryo. EMBO J 7, 1749-1756.

CARROLL, S. B., WINSLOW, G. M., SCHCPBACH, T. AND SCOTT. MP. (1986). Maternal control of Drosophila segmentation geneexpression. Nature. Lond. 323, 278-280.

DRIEVER, W. AND NOSSLEIN-VOLHARD. C. (1988a). A gradient ofbicoid protein in Drosophila embryos. Cell 54, 83-93.

DRIEVER, W. AND NOSSLEIN-VOLHARD, C. (1988b). The bicoidprotein determines position in the Drosophila embryo in aconcentration-dependent manner. Cell 54, 95-104.

DRIEVER, W. AND NOSSLEIN-VOLHARD, C. (1989). The bicoidprotein is a positive regulator of hunchback transcription in theearly Drosophila embryo. Nature. Lond. 337, 138-143.

FRIGERIO, G., BURRI, M.. BOPP, D., BAUMGARTNER, S. AND NOLL,

Localization of bicoid RNA 19

M. (1986). Structure of the segmentation gene paired and theDrosophila PRD gene set as part of a gene network. Cell 47,735-746.

FKOHNHOFER, H.-G., LEHMANN, R. AND NOSSLEIN-VOLHARD, C.

(1986). Manipulating the anteroposterior pattern of theDrosophila embryo. J. Embryol. exp. Morph. 97 Supplement169-179.

FROHNHOFER, H.-G. & NOSSLEIN-VOLHARD, C. (1986). Organizationof anterior pattern in the Drosophila embryo by the maternalgene bicoid. Nature, Lond. 324, 120-125.

FROHNHOFER, H.-G. AND NOSSLEIN-VOLHARD, C. (1987). Maternalgenes required for the anterior localization of bicoid activity inthe embryo of Drosophila. Genes and Development 1, 880-890.

KING, R. C. (1970). Ovarian Development in Drosophilamelanogaster. New York: Academic Press.

KOBAYASHI, S., MIZUNO, H. AND OKADA, M. (1988). Accumulation

and spatial distribution of poly(A)+ RNA in oocytes and earlyembryos of Drosophila melanogaster. Develop. Growth & Differ.30, 251-260.

LEHMANN, R. (1988). Maternal and zygotic control of Drosophilasegmentation. Development 104 Supplement, 17-27.

LEHMANN, R. AND NOSSLEIN-VOLHARD, C. (1987). hunchback, agene required for segmentation of an anterior and posteriorregion of the Drosophila embryo. Devi Biol. 119, 402-417.

MACDONALD, P. M. AND STRUHL, G. (1988). Cis-acting sequencesresponsible for anterior localization of bicoid mRNA inDrosophila embryos. Nature, Lond. 336, 595-598.

NOSSLEIN-VOLHARD, C , FROHNHOFER, H.-G. AND LEHMANN, R.

(1987). Determination of anteroposterior polarity in Drosophila.Science 238, 1675-1681.

NOSSLEIN-VOLHARD, C. AND ROTH, S. (1989). Axis determination ininsect embryos. In Cellular Basis of Morphogenesis (CibaFoundation Symposium 144), pp. 37-64. Chichester: Wiley.

SCHRODER, C , TAUTZ, D., SEIFERT, E. AND JACKLE, H. (1988).Differential regulation of the two transcripts from the Drosophilagap segmentation gene hunchback. EMBO J. 7, 2881-2887.

SCHOPBACH, T. AND WIESCHAUS, E. (1986). Maternal-effectmutations altering the anterior-posterior pattern of theDrosophila embryo. Roux's Arch devl Biol. 195, 302-317.

SPRENGER, F., STEVENS, L. M. AND NOSSLEIN-VOLHARD, C. (1989).The Drosophila gene torso encodes a putative receptor tyrosinekinase. Nature, Lond. 338, 478-483.

STEPHENSON, E. C , CHAO, Y. AND FACKENTHAL, J. D. (1988).Molecular analysis of the swallow gene of Drosophilamelanogaster. Genes and Development 2, 1655-1665.

TAUTZ, D. (1988). Regulation of the Drosophila segmentation genehunchback by two maternal morphogenetic centres. Nature,Lond. 332, 281-284.

TAUTZ, D., LEHMANN, R., SCHNORCH, H., SCHUH, R., SEIFERT, E.,KIENLIN, A., JONES, K. AND JACKLE, H. (1987). Finger protein ofnovel structure encoded by hunchback, a second member of thegap class of Drosophila segmentation genes. Nature, Lond. 327,383-389.

TAUTZ, D. AND PFEIFLE, C. (1989). A non-radioactive in situhybridization method for the localization of specific RNAs inDrosophila embryos reveals translational control of thesegmentation gene hunchback. Chromosoma 98, 81-85.