The function of rteriralized during Drosopliila development. Edward Yeh, Department of Zoolog, University of Toronto. Doctoral thesis (2001).

Abstract

Initially identificd as a group of gencs affccting Drosopliilri embryonic ncrvous

system development, the ncurogenic gcncs Iiiivc sincc been shown to encode components

of a highly conservcd signaling pathway essential for many dcvclopmental processes

(Chan and Jan, 1999; Greenwald, 1998; Lewis. 1998). Bcst characterizcd for thcir role in

cell fate determination, the neurogenic gencs havc also been shown to bc important for

cvents such as ccll differcntiation, cstablishing tissuc polarity, establishing compartmcnt

boundaiies, and regulating and controlling tissue growth. Thc fact that nciirogcnic gcncs

pariicipatc in a varicty of dcvelopmcntal proccsscs in many divcrsc spccics, cmphasizcs

the necd to undcrstand how this pathway functions.

While many of the main componcnts within the signiiling pathway havc bccn

idcntified, undcrstanding of the mcchanism of ncurogcnic signaling remains incomplctc.

Activation of the pathway involvcs the interaction of two transmcmbrane protcins, Delr(i

(DI) and Note11 (N), initiating a scries of evcnts that ultimatcly lcad to transcription of

downstrcam gencs. Studics arc bcginning to rcvcal the naturc of thcsc cvcnts and the

gcnes that are responsible for transducing thc N signal 10 the nucleus. One of thc gcncs

that appears to be ncccssary for ncurogenic signaling is the gcnc ~rcrtrrtliierl (rieri), which

encodcs a RING fingcr domain-containing protein. The function of ricil within thc

signaling pathway, howcver, is unknown. In addition, the rolc of iieii throughout

dcvclopment has not becn wcll charactcrized.

This thesis reports the rolc of rieil during Drosophilri dcvclopment, and its

function within the neurogenic signaling pathway. Using mosaic aniilysis, a role for rteii

in the development of bristle scnse organs is dcscribed in chaptcr 2. These data

dcmonstratc that rreu functions during the detcmination and differcntiation of scnsc

organs, and that rreu functions cell autonomously to transduce or propagatc N signaling.

As well, thc subccllular localization of the Ncu protcin at the plasma mcmbranc is

demonstratcd. In chaptcr 3 , various rreir constructs arc uscd to dcmonstratc thnt Ncu is an

E3 ubiquitin ligasc and that ubiquitin ligase activity is confcrrcd by thc RING fingcr

domain. Tnken togcthcr, thc rcsults of thcse studies siiggest that Ncu functions to

positivcly transduce the N signal by tnrgcting mcmbranc-associatcd componcnts of thc

ncurogenic signaling piithway for ubiquitinütion. In appendix A, the clonal analysis is

cxtcndcd to characterizc thc role of rrcir during the eyc and wing dcvclopmcnt.

Prcliminary evidence suçgcsts that ireil functions during N signaling throughout

Dro.sophilrr devclopmcnt. Howcver, an investigation of the rolc of rrerr during oogcnesis

shows that the function of rierr may be rcdundant or dispcnsablc for somc N signaling

cvents. This knowledgc providcs a basis to bcgin investigaihg not only the exact rolc of

ireri in N signaling, but how ubiquitin can be used to triinsducc signals in gencral.

Table of contents

Acknowledgements

List of figures

List of abbreviations

Chapter 1 - Introduction

Chapter 2 - iienralized functions cell autononiously to regulate Drosopliila sense organ development.

Chapter 3 - riertralizerl functions as an E3 uhiquitin ligase during Drosopliila development

Chapter 4 - Summary and Discussion

Appcndix A - rieriralizcd functions throughout Drosopliila development

References

i i i

Acknowledgements

The completion of this thesis would not have bcen possible without the assistance

and suppc.i of many people.

1 would like to thünk my collaborators on thc published work, much of which is

prcsented here. Their contributions have directly led to the evolving story which is told

in this thcsis. In addition, scveral labs have contributcd rcagents that were vital for many

of the experiments 1 perfomed. 1 have noted their contributions within thc text of this

thesis.

1 would also like to thank thosc people who may not have contributcd directly to

the work dcscribcd herc, but have without a doubt bccn part of the long proccss which

has lcd me to his point. Sabrina Kim has been a close fricnd and colleaguc, and has bccn

a source of grcat discussion on topics rclated and unrelatcd to science. 1 considcr

Mahmood Mohtashami to bc an OG ("original grad-studcnt") from the

Boulianne~Trimble crcw. He has bcen a fellow grad-studcnt from the bcginning, and has

on countlcss occasions steered mc in the right direction. Paulinc Henry and Cho Xu.

while not official mcmbers of the Bouliannerïrimblc crcw, have becn co-conspir;itors on

many scicntific and unscientific cndeavours. Their antics in and out of thc lab will ncvcr

be forgottcn. 1 am indebted 10 Christian Smith for insight that directly Icd to rcsults

describcd in Chaptcr Threc. His love for science has becn inspiring. 1 would also likc to

thank al1 the mcmbers of the Boulianne and Trimblc lab, past and prcscnt. In sonic wny

or anothcr, every membcr has made it possible for me to rcach this point.

The members of my committee - Dr. Howard Lipshitz, Dr. Dorothca Godt, and

Dr. Andrew Spcnce - arc greatly rcsponsible for thc path that my doctoral thesis hris

taken. Their knowledgc, insight, guidance and patience will alwiiys be appreciated

The pcrson who is rcsponsible for much oîwhat 1 havc accomplished during my

doctoral studies is my supcrvisor, Dr. Gabricllc Boulianne. Shc has bcen my tcachcr, my

mentor and my friend. 1 can say absolutely that my abilitics and confidence as a scientist

are a direct consequence of her guidance and inspiration. 1 am forcvcr indebtcd to hcr for

giving me my first chance to expcricncr scientific rcscarch, and subsequcntly providing

the environmcnt for me to dctcrmine what 1 was capable of achicving. 1 am cxtrcmcly

fortunatc to have encountcred her carly in my scicntiric carcer and hcr impact will surcly

endure for thc rcmüinder of it.

In addition to my collcagues and mentors, 1 would likc to thank rny parcnts and

family. 1 havc been fortunatc to havc bccn givcn the opporiunity to rcach this lcvcl of

education. Their support and cncouragcmcnt will alw;iys bc chcrishcd.

Finally, this thesis is not solcly my accomplishmcnt, as al1 this would not hovc

comc about without the love and support of my wife, Kclly McLcod. Evcry step of thc

way, she has becn a partncr when 1 was loving scicncc, and whcn I \vas loathing scicncc.

Her encouragement allowed mc to ovcrcomc timcs whcn 1 was urimotivatcd, and pcriods

of self-doubt. During the completion of this thcsis thc birih of Our daughtcr, Morgan

Mallory, was a monumental occasion. Forcver changing iny outlook on life and my

carcer, she brings me joy daily and inccntivc to strive for succcss.

List of Figures

Figure 1 - N, Dl signaling

Figure 2 - Comparison of Ncu homologs

Figure 3 - SOP dctcnnination and differentiation

Figure 4 - Eyc developmcnt

Figure 5 - Oogcnesis

Figure 6 - Ubiquitination pathway

Figure 7 - Scanning electron micrographs of r~err"'~' clones in adult cycs and nota

Figure 8 - Bristlc tufts in r~eii"'~' cloncs arc ciiuscd by supcmumcrary scnsory oigan prccursors.

Figure 9 - Supcmumcrary SOPs arc the rcsult of changcs in ccll fatc iind not aberrant mitosis.

Figure 10 - Thc amorphic allclc, i i e~r '~~ ' , causes a bald cuticlc phcnotypc.

Figure 11 - neii functions cell autonomously in ccll fatc decisions

Figure 12 - Expression of ireil during SOP dcvclopmcnt.

Figure 13 - Expression of iieir in third instar wing imagina1 discs

Figure 14 - Wild type and myc-taggcd i~cti constructs ciin rcscue thc i~err"'"' ncurogcnic phcnotypc.

Figure 15 - Ectopic expression of twrr causcs adult phenotypcs.

Figure 16 - scri-GA1A driven cxprcssion of ii C-terminal myc-tagged r1cir construct rcvcals Ncu localization at the plasma mcmbranc.

Figure 17 - Neu constructs uscd in gcnctic studics.

Figure 18 - Ncu localizes to ~ h c plasme membrane

Figure 19 - Ncu function rcquircs the RING fingcr domain.

Figure 20 - Neu functions as an E3 ubiquitin ligasc.

Figure 21 -Expression of rieu during cyc dcvelopmcnt

Figure 22 -Mutant ~teu"'~' clones affect cye devclopmcnt.

Figure 23 - Wild type pupal eyes stai:,ed with a cut antibody rcveal thc intcrommatidial bristlc SOPs

Figure 24 - rieu'""' mutant clones affect wing devclopmcnt.

Figure 25 - Ovcrcxprcssion of rieri during oogcncsis causes N likc phcnotypcs.

Figure 26 - Mutant rieri clones generatcd during oogcncsis cause somc dcfccts in cgg chamber dcvclopment.

vii

List of abbreviations

l0 S O P ANK APF bcd bib BrdU bx CSL C Y 0 DAPI DI DSL dx E(sp1) E C F ECFR EMS FlTC FLP-FRT FM7 GALé G F P Gt IR G S GST H MA HRP hsp70 II>TG kar kDa Ki kuz LNG mnm N NECD neu NHR NlCD osk

pnmary sense orjan precursor ankyrin alter pupanum formation bicoid big brain bromodeoxyuradine bithorax CBFllSuppressor of Hairlessllag-1 Curly balancer J',G-diamidino-2-phenylindolc Delta Dclta/Serratcllag-3 deltex Enhancer of split Epidcrmal growth factor Epidcrmal growth factor rcccptor cthylmethylsulfonate fluorcscein yeast FIp recombinasc-Flp rccombinasc target X chromosome balancer yeast GAL transcriptional activator green tluoresccnt protcin growth hormone rcccptor glutathione scpharosc glutathione S transfcrasc Hairless Hemaglutanin horseradish peroxidase hcat shock protein 70 Isopropyl-6-D-thiogaIactopyranosidc karmoisin kilo Dallons Kinked kuzbanian lin-12, Notch, glp-1 mastermind Notch Notch extracellular domain neuralized neuralized homology region Notch intracellular domain oskar

vi i i

PBS PBT PCR PEST PTB Pt c Pr pwn RI-R8 RING RTK 'Y Sb S C â

Ser S m SH2 SM3 S O SOI' Su(t1) Su(dx) TM3 UAS Ubc W

XGAL Y

phosphate buffered saline phosphatc buffcrcd saline with 0.1 C/o Triton-X polymcrase cliain reaction proline glutamate senne tyrosine phospho tyrosine binding patched pNnC prawn photoreceptors 1 through 8 really interesthg new gcne rcceptor tyrosine kinase rosy Stubble scabrous Scrratc shaggy src homology 2 sccond chromosome balancer scnsc organ sense organ precursor Suppressor of Hairlcss Suppressor of deltex third cliromosomc balanccr upstrcam activating scquencc ubiquitin conjugating white 5' Bromo-4-chloro-3-indolyl-P-D-galüctopyranoside yellow

Chapter 1 index - I~troduction

1.1 - Neurogenic signaling

1.2 - Components of Neurogenic signaling

Notclz is a cell surface transmembrane receptor Notch ligands Suppressor of Hairless is the primary effector of Notch activation Effectors of Notclz signaling neuralized is a RING finger domain containing protein

1.3 - Involvement of neurogenic signaling in Drosopl~ila development

Tense organ development * Eye development

Oogenesis Wing development

1.4 - Mechanism of Notch signaling

Notch Cleavage Involvement of Endocytosis

* Activation of transcription

1.5 - Ubiquitination and signaling

The ubiquitination pathway Ubiquitination involvement in signaling

1.6 - The function of neuralized during Drosophila development and within the Notclz pathway

1.1 - Neurogenic signaling

The neurogenic group genes were first identified in a screen for embryonic lethal

mutations that give rise to a hypertrophy of the embryonic nervous system at the expense

of epidermal tissue (Lehmann et al., 1983). Five mutants were identified - Notch (IV),

Delta (il:). rieiiralized (neir), mastennind (maal), and big brai11 (bib). N and Dl had

already been identified as mutants affzcting wing and embryonic nervous system

development, while ilen, niant and bib were novel mutations. Based on the neurogenic

phenotype, it was hypothesized that the neurogenic genes are involved in the choice

between neuronal and epidermal cell fate determination. Later genetic and cell

transplantation studies would establish a model where the neurogenic genes participate in

a process termed lateral inhibition (Campos-Ortega, 1985; Vissin et al., 1985). In lateral

inhibition, single cells inhibit neighbouring cells frorn adopting similar cell fate choices.

Thus during development of the embryonic nervous system, neuroblasts inhibit

neighbouring cells through a process involving lateral inhibition which appears to be

mediated by the qeurogenic genes.

It has since been determined that the neurogenic genes encode components of a

cornplex signaling pathway involved in many aspects of development, with N being the

best characterized of the group (hence, the neurogenic signaling pathway is often referred

to as the N signaling pathway - Fig. 1). N and Dl function as transmembrane receptor

and ligand respectively (Fehon et al., 1990; Heitzler and Simpson, 1991; Lyman and

Young, 1993; Simpson, 1990). rriûni appears to function as a transcriptional CO-activator

of N (Wu et al., 2000) while bib appears to function as a channel protein to promote N

signaling (Doherty et al., 1997). The function of neri within the signaling pathway

remains essentially unknown. Other essential factors required in the signaling pithway

have been identified whose func:ions are more defined. One such factor is Suppressor of

Hairlexs (Si<(H)), the main effector of N signaling (Bailey and Posakony, 1995;

Lecourtois and Schweisguth, 1995). In addition, other ligands such as Serrate (Ser) have

been identified. The current understanding of the initiation of neurogenic signaling

entails binding of ligand to the N receptor. This results in the release of the intracellular

domain of N (N"~). In a bound state with Su(H), translocates to the nucleus where

it behaves as a transcription factor. The main downstream targets include the genes of

the E~ihaiicer of split E(sp1) complex, although other genes such as ivirigless and trarlrrrak

are also regulated by N signaling (Guo et al., 1996; Rulifson and Blair, 1995).

In addition to those factors that promote N signaling, several are known to

negatively regulate neurogenic signaling. One protein, Fringe, affects the ability of Ser to

bind to N (Fleming et al., 1997). Another protein, Numb, appeirs to block N signaling

during transduction from the plasma membrane to the nucleus (Frise et al., 1996; Guo et

al., 1996). The HECT domain-containing protein encoded by Suppressor of deltex

(Su(&)) genetically downregulates N signaling (Cornell et al., 1999; Fostieret al., 1998).

Hairless (H) encodes a DNA binding protein that antagonizes N signaling by binding

sites that N " ~ ISu(H) would nomally access to activate transcription (Bang et al., 1995;

Schweisguth and Posakony, 1994). Thus regulation of neurogenic signaling, both

positive and negative, can occur at many points in the N pathway.

Figure 1 - Notch signaling. 1. Notcli signaling begins with the binding of a iigand (in this case, Delta). 2. Putative rearrangements in the receptor ligand complex cause clesvage of the Notch receptor, and the Notch intracellular domain is intemalized. It has been also shown that the Notch extracellular domain is transendocytosed with Delta. Cues for Notch endocytosis have not been determined, but Numb, which binds the Notch intracellular domain and inhibits Notch signaling, has been implicated in endocytosis. Suppressor of Hairless binds to the Notch intracellular domain to form a transcription factor. This binding may be regulated by Deltex. 3. The Notch intracellular domain1Suppressor of Hairless complex translocates to the nucleus and activates downstream genes such as those in the Enhancer of split complex.

i Deka Iniracellular w Numb m Delta extracellular A Neuralized - Notch extracellular

Notch intracellular @B Suppressor of Hairless

CS Deltex O Hairless

The N signaling pathway is now known to be required in species as diverse as

worms and humans to regulate many developmental processes including cell fate

detemination, establishment of tissue polarity, establishment of compartment

boundaries, and regulation of tissue growth. While several studies have provided much

information into the mechanisms and requirements of N signaling, a clear picture of how

N signaling proceeds has yet to be described. The following pages will try to provide an

overview of N signaling in Drosopkila development with the main focus being the

characterization of the involvement of neic within the N pathway.

1.2 - Components of the Notch signaling pathway

Many genes hiive been identified that are proposed to function within or in

concert with the N pathway. This section will describe those components that are thought

to be intimately involved in the N pathway and their proposed functions within the

pathway. While many of the main components of the N signaling pathway have been

identified both genetically and molecularly, it is clear that many components have yet to

be identified. The task of identifying these unknown genes remains daunting. However,

thoroughly understanding the function of each of the known components in the pathway

will not only provide insight into the mechanism of N signaling, but will assist in

identifying other components that have yet to be detemined.

Notck is a cell surface transmembrane receptor

The main component in the N signaling pathway is the N receptor itself.

Identified in the context of the neurogenic signaling group by Lehman et al. (1983).

mutations affecting embryonic nervous system development were characterized as early

as 1937 (Poulson, 1937). The N gene encodes a single pass transmembrane protein that

is 2703 amino acids in length (Artavanis-Tsakonas et al., 1983; Wharton et al., 1985).

Maturation of the receptor requires cleavage of the protein into two fragments within the

Golgi network. The two fragments form a heterodimer, consisting of a 180 kDa

extracellular fragment and a 115 kDa intracellular fragment, which then inserts into the

plasma membrane (Blaumueller et al., 1997). Two homologs exist in C. elegarzs (firi-12

and gip-1) (Greenwald, 1998; Weinmaster, 1997), and four homologs in mammals

(Notchl-4) (Robey, 1997) and companson of these homologs reveals that the N receptor

contains many conserved domains. A combination of studies of the vanous N homologs

has led to an understanding of the function of the conserved domains.

The large extracellular region of N contains 36 EGF repeats that are

thought to mediate ligand binding, protein stiability and cell adhesion. EGF repeats I I

and 12 have specifically been shown to be involved in binding of the Ser and DI ligands.

Repeats 24-29 have been implicated in ligand binding through characterization of the N

Abri~ptex mutations, which act as dominant ligand-dependent hypermorphs. Also present

in the extracellular domain are conserved lin-12 or LNG (lin-12, Notch, Glp-1) repeats.

The function of these repeats may be to maintain N stability and prevent signaling in the

absence of ligand.

The intracellular domain or N " ~ contains several conserved motifs. These

include the PEST sequence, cdc IOIankynn (ANK) repeats, and RAM23 domain. The

PEST sequence is thought to downregulate signaling by mediating protein degradation

and turnover. The RAM23 domain interacts with Su(H) (Tamura et al., 1995), the main

effector of N signaling. The RAM23 domain also interacts with Numb, a negative

regulator of signaling. The ANK repeats are able to bind to several proteins but the exact

role of these repeats is not clear. Initial studies suggested that Su(H) binds to the ANK

repeats. However, while the ANK repeats may actually be involved in promoting Su(H)

binding. these repeats cannot bind Su(H) alone. The ANK repeats also interact with

Deltex, which genetically is required for N signaling in some contexts (Diedench et al.,

1994; Matsuno et ai., 1995). It has been suggested that Deltex binding to the ANK

repeats displaces or renders Su(H) binding less efficient, thereby allowing Su(H) to

translocate to the nucleus. However, this has yet to be clearly demonstrated. Another

function of the ANK repeats may be to mediate homotypic interactions between N

receptors (Matsuno et al., 1997). The role of such interactions in N signaling is not clear.

Notch ligands

The primary ligands for the N receptor in Drosophila are the proteins encoded by

Dl and Ser, which also are conserved among various species, and are sometimes referred

to as the DSL (Delta, Serrate, hg-2) class of proteins. DI is a transmembrane protein

consisting of 833 amino acids (Kopczynski et al., 1988). The extracellular domain of Dl

contains several conserved domains including an N-terminal (NT) domain that contains a

signal sequence, a DSL motif, and 9 EGF-like repeats which consist of 44 amino acids

each. While the intracellular domain of DI does not exhibit any striking motifs, it appears

to be required for N signaling as constructs lacking the intracellular domain act as

dominant negatives (Sun and Artavanis-Tsakonas, 1996; Sun and Artavünis-Tsakonas,

1997). Like N, Dl may be cleaved into several peptide fragments. Four isoforms have

been identified in vivo, including a secreted form (Klueg et al., 1998). However, the

significance of these various isoforms has yet to be clearly characterized.

The protein structure of Ser is very similar to that of Dl. Consisting of 1404

amino acids, Ser is a transmembrane protein and the extracellular domain contains an NT

domain, a DSL motif and 14 EGF like repeats. In addition, the extracellular domain

contains a cysteine rich region (Fleming, 1998; Fleming et al., 1990).

Despite differences in their size, DI and Ser may be partially redundant as one

ligand can substitute for the other in certain N signaling events (Gu et al., 1995; Zeng et

al., 1998). Thus, ligand activity rnay lie within the various domains common between the

two proteins. In both proteins, the NT domain in addition to the DSL domain form an

EGF binding domain (EBD), which is necessary and sufficient for N binding

(Muskavitch, 1994). However, neither the NT domain nor the DSL domain is sufficient

to mediate cell adhesion with N expressing cells (Fleming, 1998). In addition, the EBD

appears to confer ligand specificity. Fringe (Fig) encodes a protein that blocks Serrate

but not Dl activation of N (Fleming et al., 1997). Substituting the Ser EBD with a DI

EBD results in a ligand which is unaffected by the presence of Fng (Flcming et al., 1997).

Suppressor of Hairless is the primary effector of Notcli activation

Another consewed component of N signaling is the Sii(H) class of proteins,

sometimes referred to as the CSL (CBFlIRBP-Ju, Su(H), hg-1) proteins. Su(H) is a

transcription factor consisting of 594 amino acids (Gho et al., 1996) that can bind to the

RAM23 domain of N (Tamura et al., 1995). Studies have demonstrated that Su(H) acts

as a transcriptional activator of N target genes. In addition, localization within the ceil

can be influenced by N signaling, although the exact mechanism of activation of Su(H)

by N signaling has not been clearly resolved. Details of Su(H) activation by N signaling

will be discussed later.

There is increasing evidence suggesting that Su(H) may interact with different co-

activators within the nucleus to alter the nature of its transcriptional activation. Recently,

it has been demonstrated that Mastemind is a component of the nuclear transcriptional

activator complex involving the N ' ' ~ and Su(H). tnastennind encodes a rather large

nuclear protein consisting of 1596 amino acids (Smoller et al., 1990). A human homolog,

MAMLI, has been identified and has been shown to be located in the nucleus. In

addition, MAMLl is able to interact with the intracellular domain of Notchl (ICNI) and

RBP-JK (Wu et al., 2000). Details of these findings will be discussed later.

Another protein that can bind Su(H) within the nucleus is the product of the gene

Hairless (H). However, H acts as an antagonist of Su(H). H is a large protein consisting

of 1096 amino acids and is distributed both in the cytoplasm and in the nucleus (Bang

and Posakony, 1992; Maier et al., 1999; Maier et al., 1992). Studies have shown that H

can bind directly to Su(H) and reduce the efficiency with which Su(H) can bind DNA and

activate transcription (Brou et al., 1994). As well, the H/Su(H) complex may actually

f o m a transcriptional repressor in certain contexts (Funiols and Bray. 2000). Clearly,

understanding how the output of the N signaling pathway leads to transcnptional

activation is complicated by the presence of various CO-activators and repressors.

Effectors of Notclr signaling

Through genetic screens many other genes have been identified that appear to

function within the N signaling pathway. One such gene, deltex (dx), was identified as a

component of N signaling in a screen for suppressors of N mutants (Xu and Artavanis-

Tsakonas, 1990). d x encodes a novel737 amino acid cytoplasmic protein capable of

binding the ankyrin repeats of N (Busseau et al., 1994; Diederich et al., 1994). Dx

contains a RING finger domain and an SH3 domain, and prevents the cytoplasmic

retention of Su(H) in cell culture. Genetic studies have demonstrated that dx functions to

positively regulate N signaling, although it may not be required for al1 N signaling events

(Matsuno et al., 1995). Thus the predicted role of dx is to antagonize the interaction of

Su(H) with N, allowing Su(H) to translocate to the nucleus. The recent demonstration

that RING finger domains may be involved in ubiquitination (Lorick et al., 1999) may

implicate d r in the ubiquitin pathway, but presently there is no evidence to support this.

Initially identified as a suppressor of the d.r mutant phenotype . sirppressor of

deltex (Su(&)) has since been shown to function as a negative regulator of N signaling

(Fostier et al., 1998). Recent molecular characterization of Sir(dx). a 949 amino acid

protein, has revealed that it belongs to the family of Nedd 4 E3 ubiquitin ligases. Su(dx)

contains a membrane-targeting C2 domain. proline rich binding WW domains, and a

ubiquitin ligase HECT domain (Comell et al., 1999). While there is no data

demonstrating that Su(dx) behaves as a ubiquitin ligase, a related mammülian gene, Itch,

binds and ubiquitinates N in vitro (Qiu et al., 2000).

~iitriib mutations cause a reduction in peripheral neurons. iiiinib encodes a zinc

finger protein containing several PEST sequences and a phosphotyrosine-binding domain

(PTB). During neural development, Numb acts as a cell-intrinsic factor to regulate cell

fate decisions. This role can be best seen during the lineage of the sense organ precursor

(SOP), where Numb is asymmetrically localized and distributed to one of two daughter

cells. Within this cell, Numb antagonizes N signaling. Numb can bind to the

intracellulx domain of N, specifically to the RAM23 domains and C terminal region

(Guo et al., 1996).

neiiralized encodes a RING finger domain containing protein

The molecular characterization of Drosopliila rieii revealed that it encodes a 754

amino acid protein. The translated protein predicts the presence of a carboxyl-teminal

RiNG finger, as well as possible helix-rich regions and a nuclear translocation signal

(Fig. 2) (Bouliannc et al., 1991). Homologs have been identified in C. elegaris, mouse,

and human (Moschonas, 1998; Nnkamura et al.. 1998; Wilson et al., 1994). A

comparison of rieu sequences from C. elegarzs and human reveals that the carboxyl-

terminal C3HG RING finger is conserved among species. As well, two regions of

homology temed m u homology speats (NHR) were identified (Fig. 2) (Nakamura et

al., 1998). However, the function of these domains has yet to be detemined.

neir is expressed in various tissues during development and its expression

correlates with the involvement of N signaling within the:e tissues. In addition to the

predicted expression of rieu during embryogenesis (based on its mutant neurogenic

phenotype), rieu is expressed in SOPs in imaginal tissues, in ommatidial clusters in the

eye imaginal disc, in the inner and outer proliferation centres within the optic lobes of the

larval central nervous system, and in polar follicle cells during oogenesis (Boulianne et

al., 1991). The expression pattern of neir suggests that neii functions in the development

of the embryonic nervous system and throughout development.

Figure 2 - Comparison of Neuralized homologs. Comparison of Neuralized from Drosopliila, C. elegaris and human reveals the presence of two Neuralized homology repeats (NHR) and conservation of the RING finger dornain. A helix rich region (hel- rich) is not conserved between vertebrates and invertebrates (modified from (NaIcamura et al., 1998))

ai 741

D-neu

C-neu

h-neu iüNC

1.3 - Involvement of the Notcli signaling pathway in development

The neurogenic group genes were first identified as mutations affecting

embryonic nervous system development, suggesting that their function is to mediate cell

fate decisions in the ventral neurogenic region. It is now known that these genes function

in a variety of developmental processes. To illustrate the involvement of the neurogenic

genes in some of these processes, the role of the N signaling pathway in the development

of various Drosopltila tissues will be described.

Sense organ development

The development of the bristle external sense organ in Drosopltila is well

understood at the cellular level. The lineüge that results in the mature bristle sense organ

is stereotyped for every bristle, and cell fate decisions involve both intrinsic as well as

extrinsic factors (Fig. 3). This makes this process ideally suited to study the genetics of

cell determination and differentiation. As a result, the involvement of N signaling in the

development of the sense organ is well understood.

Sense organ development begins with the determination of a field of cells that

have the potential to become a sense organ precursor (SOP). The genes responsible for

establishment of this field, or proneural cluster, are the proneural group genes (Garcia-

Bellido, 1979; Ghysen and Richelle, 1979). Members of the ucltaete-scirre contplex,

daiighterless, and asettse are pan of this group and mutations in these genes result in the

failure to produce sensory organs due to the inability to establish proneural clusters

(Cubas et al., 1991; Simpson, 1990; Skeath and Carroll, 1993). Initially, cells within the

proneural cluster fonn an equivalence group, with al1 cells capable of adopting either

neural or epidermal ceIl fates (Simpson, 1990). Subsequent to proneural cluster

Figure 3 - Sense organ determination and differentiation. A) The first step in sense organ determination is the establishment of the proneural field or cluster. N signaling within the proneural cluster is required for determination of a single sense organ precursor, which undergoes a stereotyped lineage (B) to produce al1 the cells present in the mature bristle (C - top). N signaling is required within the linîage for the proper adoption of secondary and accessory cell fates. Factors such as nirmb and Hairless antagonize N signaling asymmetrically, allowing one cell to receive more N signaling than another cell. This unequal N signaling ultimately is responsible for adoption of different cell fates. A scanning electron micrograph shows the extemal structure of an adult bristle sense organ consisting of a shaft and socket (C - bottom).

shsath neumn

O proneural cluster

sense organ precursor = Notch, Delta, neurallzed

X numb, Hairless

detemination, a single SOP is detemined within the cluster through a process of mutual

inhibition andor lateral inhibition. Mutual inhibition describes stochastic signaling

among cells within the cluster that leads to some cells yielding more signaling than

others. Eventually, a single cell dominates signaling and adopts the SOP or primary SOP

(lOSOP) fate. Through lateral inhibition, the 1" SOP suppresses neighbouring cells from

adopting the neural fate and they adopt epidermal fates. It is believed that the neurogenic

genes are responsible for this process of mutualllateral inhibition (Fig. 3).

During adult sensory bristle differentiation, the Io SOP divides to produce al1 the

cells that form a mature bristle (Fig. 3). The first division yields two secondary SOPs

teimed pIIA and pIIB (Hartenstein and Posakony, 1989). The pIIB cell then divides to

produce a glial cell, and a tertiary SOP temed pIIIB (Gho et al., 1999). The pIIIB cell

divides to produce a neuron and a sheath (thecogen) cell. The pIIA cell divides to

produce the shaft (tricogen) and socket (tomogen) cells (Hartenstein and Posakony,

1989). While it is known that intrinsic factors such a Numb are important for subsequent

SOP and accessory cell fates in the lineage, cell-cell communication involving

neurogenic signaling is also required (Hartenstein and Posakony, 1990; Parks and

Muskavitch, 1993; Rhyu et al., 1994).

Mosaic analysis of N, Dl. neu. and rriatri mutations provided the first evidence for

the involvement of these genes in sense organ development (Dietrich and Campos-

Ortega, 1984). N mutant clones produced bald cuticle lacking any bristles (when N ~ ~ . ~ ~ ,

a strong neurogenic allele, was used) and bristle tufts (when N" was used). neit and rriarti

mutant clones only produced bald cuticle, while Dl mutait clones produced bristle tufts.

Previous descriptions of allele strength, based on the severity of the embryonic phenotype

caused by each allele, predicted that strong neurogenic alleles yield bald cuticle, while

weaker neurogenic alleles cause bristle tufting. Hairless (H) when used in double mutant

combinations with DI and nert, was able to suppress the phenotypes associaied with Dl or

neic mutant clones, suggesting that H function antagonizes Dl and lieu.

Funher analysis of N using a temperature sensitive mutation clarified its

involvement during sense organ development (Hartenstein and Posakony, 1990).

Shifting developing larvae to the restrictive temperature during 1" SOP determination

resulted in supernumerary l0 SOPs. Each l0SOP divided to form a mature bristle

producing the tufting phenotype. Thus, N appears to be required for mutual inhibition

between cells that yields a single Io SOP within the proneural cluster. Unequal N

signaling among cells leads to determination of a single SOP, which then inhibits

neighbouring cells from adopting the same fate through lateral inhibition. These cells

then adopt an epidermal cell fate.

Removal of N function during the SOP divisions also led to defects (Hartenstein

and Posakony, 1990). Shifting flies to the restrictive temperature during the first SOP

division caused both daughter cells to adopt the pIIB fate. In addition, removal of N

function during the division of the secondary SOPs caused transformation of sheath cells

into neurons, and shafts into sockets. A gain-of-function N mutation gives rise to

opposite cell fate choices: pIIB cells are transformed into pIIA cells, sockets into shafts

and neurons into sheaths. These results suggest that during the first SOP cell division, N

functions to allow cells to adopt the pIIA fate. Loss of N yields two pIlB cells while

gain-of-function N yields two pIIA cells. During determination of the ûccessory cell

fates, Nsignaling allows for the adoption of the sheath ceIl fate from the pIIIB cell, and

adoption of socket cell fate from the pIIA cell. A similar approach was used to

demonstnte an identical involvement of Dl in SOP determination and differentiation

(Parks and Muskavitch, 1993). Thus the N signaling pathway appears to be involved

throughout the development of bristle sense organs, controlling the fate of every cell in

the SOP lineage.

The model of lateral or mutual inhibition involving N signaling predicts that

unequal signaling must occur between neighbouring cells. However, the exact details of

how unequal signaling in the proneural cluster is established are not known. While

intrinsic factors probably play a role, it is thought that an additional positive feedback

mechanism may also be involved to ensure that unequal cell signaling occurs within the

proneural cluster. In addition to establishing the proneural cluster, :he proneural genes

are also targets of N signaling (Hinz et al., 1994; Singson et al., 1994). Further

downstream of proneural gene activation, Dl expression is downregulated as a result of

activation of the N pathway (Parks et al., 1997). Thus one model predicts that N

signaling leads to decreased levels of Dl expression. Lowered DI levels reduce the ability

to signal to a neighbouring cell, and thus the neighbounng ce11 can continue to make

higher levels of Dl. Stochastic alterations in the levels of N signaling amongst cells

within the proneural cluster may be amplified by the feedback mechanism, allowing for

the determination of a single SOP within the developmental time period.

Identification of genes that negatively regulate N signaling has faci!itated thc

understanding of how N signaling controls cell fate decisions in the SOP lineagc.

Hairless ( H ) encodes a DNA binding protein that genetically antagonizes N signaling and

can suppress neurogenic phenotypes (Bang et al., 1995; Dietrich and Campos-Orîega,

1984). It has been demonstrated that H binds to the same DNA sequences as Sir(H)

(Brou et al., 1994). Increasing evidence that Su(H) and H form a transcnptional

repressor complex suggests that N signaling may relieve transcnptional repression by H

(Fumols and Bray, 2000). H mutants cause a pIIB to PUA cell fate transformation, as

well as shaft to socket and neuron to sheath transformations (Bang et al., 1995;

Schweisguth and Posakony, 1994). Thus, H appears to suppress N signaling in one of thc

two secondary SOP cells, allowing that cell to adopt the pIIB fate. Similarly. H

suppression of N signaling allows one cell of each pair of accessory cells, the shaft and

the neuron, to adopt the appropriate cell fate.

Another gene that negatively regulates N signaling is riirtrib. Numb binds to the

intracellular domain of N and genetically antagonizes N signaling (Frise et al., 1996; Guo

et al., 1996; Rhyu et al., 1994). Loss-of-function nrrrtib yields phenotypes opposite to

those produced by N mutations - riiorrb mutations cause transformation of pIIB into pIIA

fates, and causes transformation of neurons into sheaths. Numb is expressed in the 1"

SOP and the protein localizes to an anterior crescent prior to division of the 1" SOP into

the pIIA and pILB daughter cells (Lu et al., 1998; Rhyu et al., 1994). Division of the 1"

SOP occurs in the anterior-posterior plane, and Numb is only received by the pIIB cell

(located anteriorly of the two daughter cells). Thus like H, riirnib appears to antagonize N

signaling in only one of the two daughters, allowing the cell with reduced N signaling to

adopt the pIIB fate.

Curiously, while Sir(H) is considered to be the main effector of N signaling, Sir(H)

does not appear to be required for the determination of every cell fate within the SOP

lineage. While thc individual fates of cells within the SOP lineage require N signaling

and unequal suppression by tiitnib, epistasis experiments reveal that ~iunib mutants are

suppressed by Sii(H) mutations only with respect to shaft to socket cell fate decisions.

The neuron to sheath cell transformation is not suppressed by Sii(H), suggesting that in

this case, N signaling is tranduced via a Sii(H)-independent mechanism (Wang et al.,

1997).

Eye development

The N signaling pathway has a complex role during eye development. It has bcen

implicated in determination of cell fates (Baker and Yu, 1997; Cagan and Ready, 1989;

Parks et al., 1995). regulating tissue growth (Papayannopoulos et al., 1998),

establishment of boundaries and determination of tissue polarity. A brief description of

eye development will assist in understanding the cmcial role N has in eye development.

Signals from the dorsal and ventral sides of the eye disc direct the movement of

the morphogenetic furrow in a posterior to antenor direction. The morphogenetic furrow

is an indentation in the epithelium caused by local cell shape changes whose purpose is to

prepare cells for interactions required for cell fate determination. Ce119 immediately

posterior to the furrow begin to organize into clusters that will eventually adopt

photoreceptor cell fates (termed RI through R8) in a sequential manner. The first cell

determined in the cluster is the R8 photoreceptor, followed by the RYR5 pair, and then

the R31R4 pair. R7 and the RllR6 pair then join the cluster, which then reonentates 90'

with respect to the genetically defined central equator that lies between the dorsal and

ventral halves of the eye. The photoreceptors orientate themselves in a stereotyped

pattern with RI through R6 aligning themselves in a trapezoidal pattern, with R7 and R8

Figure 4 - Determination of photoreceptor polarity. As the morphogenetic furrow moves antenorly across the developing eye pnmordium, determination of photoreccptors occurs in a sequential manner immediately posterior to the furrow beginning with ~ 8 , then the R2lR5 pair, and then the R3lR4 pair. R7 and the RllR6 pair are determined last and join the other photoreceptors in the cluster. Each ommatidium then rotates 90" and orients itself with respect tothe dorsal ventral equator, such that there is a mirror image symmetry. In addition, photoreceptor R4 repositions itself within the cluster to form a ciooked house like of phoÏoreceptors RI-R6 surrounding R71R8 (modified from (Stmtt and Stmtt, 1999)).

dorsal pole

-" 0 @

O 00

4-------- rnorphogenetlc furrow

located in the centre (Fig. 4). Of the R31R4 pair in every ommatidial cluster, R3 is

slways located anteriorly while R4 is Iocated closer to the central equator (Fig. 4). This

arrangement provides chirality to the photoreceptor pattern, and every ommatidium is

onented such that a mirror image reflection of ommatidial polarity exists along the

central equator (Fig. 4). Ultimately the formation of a highly pattemed and organized

adult eye occurs.

Early investigations involving the temperature sensitive N allele revealed that N

signaling was involved in almost every ceIl fate decision during eye development. Loss-

of-function N could affect the determination of photoreceptors, cone cells, pigment cclls

and bristle sense organs (Cagan and Ready, 1989). Studies using a Dl temperature

sensitive mutation revealed similar results (Parks et al., 1995). Detailed studies have

revealed specific roles for N signaling in the determination of specific photoreceptors R8

(Baker et al., 1990). R3/R4 (Fanto and Mlodzik, 1999) and R7 (Fortini et al., 1993).

N has also been shown to play a role in tissue polarity during eye development.

First, N participates in defining the dorsal-ventral equator along the anterior-posterior

mis. This was first suggested following the observation that the N ligand DI is expressed

in the dorsal half of the eye imagina1 tissue, while the other N ligand Ser is expressed in

the ventral half, along with fringe (Choi and Choi, 1998; Papayannopoulos et al., 1998).

fringc interferes with N's ability to bind Ser and potentiates binding to DI in the wing

(Fleming et al., 1997). Furthermore, Sercan also induce expression of Dl but only in the

absence of fringe while Dl can induce expression of Ser but only in the presence of fringc

(Papayannopoulos et al., 1998). Nin the ventral region of the eye appears to respond to

DI binding dong the edge of Fringe expression which induces Serexpression.

Simultaneously, N on the dorsal side responds to Ser which induces Dl expression dong

the boundary. This positive feedback strengthens local N activation and establishes the

dorsal-ventral boundary (Papayannopoulos et al., 1998).

frirrge is also involved in the determination of the R 3 R 4 photoreceptors, which

orient themselves within each ommatidium confemng chiral polarity. Both

overexpression and loss-of-function fri~rge cause defects in ommatidial polarity (Fanto et

al., 1998; Zheng et al., 1995). The identification of N and Dl as enhancers of the frilzge

overexpression phenotype as well as the sensitivity of R3/R4 photoreceptor specification

to N signaling revealed an additional role for N in establishing eye tissue polanty (Fanto

and Mlodzik, 1999). It appears that the R3lR4 cells initially represent an equivalence

group in which the R3 fate is the primary fate (Zheng et al., 1995). Once one cell adopts

the R3 fate, the other is inhibited from adopting the same fate and adopts the R4 fate. An

unknown polarizing signal from the dorsal-ventral equator causes one cell of the pair, the

future R3, to obtain higher fritrge activation than the other ceIl (Zheng et al., 1995). It

tums out that a transcriptional target of fririge activation is Dl. The demonstration that

R4 specification depends on N signaling fits nicely with the model of frirzge specifying

the R 3 R 4 fates (Cooper and Bray, 1999: Fanto and Mlodzik, 1999). And thus, in this

instance the role of N signaling in cell fate decisions ultimately leads to tissue polanty.

Oogenesis

A role for the neurogenic genes in oogenesis was first suggested by the

observation that N and Dl are expressed during oogenesis, and that temperature sensitive

mutations in N and Dl gave rise to specific defects in oogenesis (Bender et al., 1993;

Ruohola et al., 1991; Xu et al., 1992). (For a brief explanation of oogenesis, see Fig. 5).

Adult females mutant for N or Dl, placed at the restrictive temperature, displayed fused

egg chambers containing too many germline nuclei caused by a failure of egg chambers

to separate from the germarium and other egg chambers. A closer examination using

markers for polar follicle cells revealed a defect in the specification of follicle cell fates.

and it was concluded that an increased number of polar follicle cells were specified at the

expense of follicular stalk cells. Thus, egg chambers failed to separate from the

germarium due to a failure of stalk cell formation (Ruohola et al., 1991). Expression of

an activated form of N during oogenesis caused the opposite phenotyl;?. Transient

expression of N lacking the extracellular domain caused abnormally long chains of stalks

of cells bctween egg chambers. Using markers for vanous follicle cell types, these giant

stalks were determined to be primarily made up of follicular stalk cells (Lürkin et al.,

1996). Thus, N signaling appears to affect the decision to adopt a stalk cell vcrsus a polar

cell fate.

A second defect in oocyte polanty was also detected in developing ovaries grown

at the restrictive temperature. Using the localization of bicoid (bcd) and oskar (osk) to

assay the polarity of developing oocytes (bcd transcripts are localized anteriorly; osk

postenorly), it was observer! that many of the oocytes without N or Dl activity lacked

anterior-posterior polarity (Ruohola et al., 1991). Expression of an activated N construct

confirmed the iiivolvement of the signaling pathway in the establishment of anterior-

posterior polarity of the oocyte, as mild expression of activated N which does not affect

follicle cell fates, can affect anterior-posterior polarity of the oocyte (Larkin et al., I996).

The demonstration that microtubules undenvent premature streaming suggests that N

Figure 5 - Drosophila oogenesis. The adult female possesses two ovaries which are each made up of approximately 20 structures called ovarioles (one is shown here; anterior is the left). Each ovariole consists of a chain of egg chambers that are sequentially more developed from anterior to posterior. The ovariole consists of a germarium which contains the most anterior terminal filament (tf), germline stem cells (gsc), follicle stem cells and early stage 1 egg chambers (Sl). Self-renewing germline stem cells produce a germline cyst, which divides without complete cytokinesis to produce a 16 cell cyst. One of these cells becomes the future oocyte (nucleus is brown) which positions itself posteriorly within each egg chamber, while the remaining 15 cells become nurse cells (grey nuclei). Proliferating follicle cells surround the 16 cell cyst to produce a stage 1 egg chamber which pinches off from the posterior end of the germarium and enters the vitellarium. Specialized follicle cells called stalk cells, which separate egg chambers (green) and polar follicle cells, a pair of cells found at the terminal ends of each egg chamber (red), differentiate approximately at this time. Egg chambers complete development within the vitellarium. The oocyte increases in size, while the nurse cells become polyploid synthesizing cytoplasmic components which will be dumped into the maturing oocyte. Further follicle cell specializations occur (such as determination of border cells) which are necessary for events such as polarity determination, and oocyte structure.

Germarium borda cclk

signaling may be involved in communication between the oocyte and posterior follicle

cells.

Recent reports have clarified that N signaling emanating from the germline is

required for differentiation of polar and stalk cell fates, differentiation of epithelial

follicle cells, cell cycle control and proper follicular migratory behavior (Grammont and

Irvine, 2001; Lopez-Schier and St. Johnsto-, 2001). By generating mutant clones of a

nuIl N allele, it was shown that loss of N function in the germarium does not result in the

overspecification of polar follicle cells, but rather, results in the failure to differentiate

polar cells. N signaling is then thought to be required from the polar cells to differentiate

stalk cells. Dunng stages 5 to 7, loss of N function resulted in a failure of epithelial

follicle cells to exit mitosis and enter endoreplication, and a failure to undergo migration.

These defects in teminal follicle cell and epithelial cell differentiation can explain both

the fused egg chamber phenotype and the defects in antenor postcrior polarity.

Wing development

The N pathway is also required during wing development to control cell fate

decisions, establish polarity and regulate tissue growth. Dorsal-ventral boundary

formation in the wing (Blair, 1997; Blair, 1999) is very similar to establishment of the

dorsal-ventral equator in the eye. Also, N involvement in sense organ precursor

determination and differentiation which occurs at the wing margin is thought to be

similar to development of sense organs on the notum. Howevcr, N signaling is nlso

important for the detemination of wing vein tissue. This was first revealcd by the

obsarration that Dl mutants have ectopic wing veins. It has been shown that N signaling

depends on a feedback mechanism that establishes regions of high and low Dl expression

(Huppert et al., 1997). Expression of DI and N during vein-intervein specification reveals

high expression and accumulation of DI in provein regions (future vein tissue). This

complements high expression and accumulation of N in lateral provein regions (future

intervein tissue). It tums out that N expression and accumulation in lateral provein

regions depend on Dl signaling while DI expression and accumulation in provein regions

is suppressed by constitutive N signaling (Huppert et al., 1997). The complementary

patterns of N and DI expression that are required to generate properly pattemed vein and

intervein regions suggests that unequal signaling from provein regions leads to unequal

activation and accrimulation of N and DI protein.

1.4 - Mechanism of Nolcli signaling

Notch Cleavage

Post-translational processing of N through proteolytic cleavage is vital for N

signaling. Much of the evidence for N cleavage has ansen from studies of mammalian N,

and to date fourcleavage events have been identified which occur both pnor to and after

ligand binding. The first cleavage in the trans-Golgi network is carried out by a furin-like

convertase (Blaurnueller et al., 1997; Logeat et al., 1998), producing a non-covalently

linked heterodimer. The site of cleavage is approximately 70 amino acids N-terminal to

the transmembrane domain resulting in fragments of 180 kDa (containing the

extracellular domain) and 120 kDa (containing the transmembrane and intracellular

domain). This cleavage is part of the maturation process required for a functional N

receptor at the cell surface. Upon ligand binding, two additional cleavage events are

thought to occur. One occurs in the extracellular domain and another in the

transmembrane domain; this latter cleavage is thought to release the (Lecourtois and

Schweisguth, 1998; Schroeter et al., 1998). The extracellular cleavage is thought to be

accomplished by a metalloprotease protein, and the transmembrane cleavage by a y-

secretase-like protein (possibly presenilin). Recently a fourth cleavage event has been

descnbed, that occurs in the intracellular domain and is catalyzed by the divergent

caspase DRONC. Interestingly, DRONC was identified due to its ability to bind Numb.

Subsequent studies revealed that DRONC binds and cleaves N in vitro and in Drosopldu

S2 cells. DRONC appears to be functionally significant during N signaling as RNA

interference expenments using DRONC doublestranded RNA yield embryonic

neurogenic phenotypes. As well, expression of a dominant-negative catalytic mutant

form of DRONC causes N loss-of-function phenotypes, whereas overexpression of wild

type DRONC causes N gain-of-function phenotypes (Petntsch et al., 2000).

DI processing also appears to be important for signaling. The demonstration that

the ADAM metalloprotease Kuzbanian (Kuz) cleaves DI (Klueg et al., 1998; Qi et al.,

1999). and that kuz mutations cause N like phenotypes (Pan and Rubin, 1997; Sotillos et

al., 1997) strongly suggests that Dl must be cleaved for N signaling. In addition, a

soluble Dl extracellular fragment can be detected in the supernatant of DI expressing

Drosophila tissue cultures. Formation of this fragment in vitro and iri vivo can be

suppressed by inhibiting Kuz. However, the functional significance of this extracellular

fragment is not clear since it possesses both antagonistic and agonistic qualities in

different experimental contexts (Qi et al., 1999; Sun and Artavanis-Tsakonas, 1997).

Involvement of Endocytosis

In addition to the above described cleavage events, N endocytosis is also thought

to be required for N signaling. Genetic evidence is provided by the demonstration that

mutations in the dynamin gene, shibire, can produce neurogenic phenotypes during

embryogenesis (Poodry, 1990) and N-like mutant phenotypes during sense organ

development (Seugnet et al., 1997). As wcll, it has been demonstrated that ligand-

dependent gain-of-function N signaling (using the Nn'mutation) is suppressed in slii

mutants (Seugnet et al., 1997). However, the requirement for endocytosis is not fully

understood. Endocytosis also appears to regulate Dl protein localization and

accumulation (Parks et al., 2000; Parks et al., 1995). In addition, trans-endocytosis and

CO-localization of N~~~ with DI, appears to be required for N signaling (Parks et al.,

2000). A model has been proposed that DI endocytosis is concurrent with and nccessary

for N cleavage.

The recent chancterization of Numb as an endocytic protein provides further

support for the notion that endocytosis is required during N signaling. Numb appears to

function downstream of N and has also been shown to bind the intracellular domain of N

(Guo et al., 1996). Interaction of Numb with EPS-15 (Salcini et al., 1997). an endocytic

protein. led researchers to study whether Numb is involved in endocytosis. Through

immunocytochemical studies, human Numb expressed in various cell lines was found to

localize to endocytic organelles including endosomes, vesicles and clathrin coated pits

(Santolini et al., 2000). Upon EGF stimulation, Numb protein was seen to translocate

over time, from the trans-Golgi network to coated pits and vesicles, and eventually

endosomes. In addition, Numb appeared to CO-traffic with intemalizing EGF receptors

and bind to a-adaptin, an endocytic adaptor protein. Finally, it was observed that

carboxyl-terminal Numb fragments behaved as dominant negatives, inhibiting

internalization of EGFR. Of particular interest, one of these Numb fragments lacked the

binding sites for EPS-15 and a-adaptin suggesting that Numb was able to block

endocytosis through some other mechanism (Santolini et al.. 2000). It is speculated that

Numb functions similarly with respect to N signaling, but such data has not been

reported.

Activation of transcription

At present, the senes of events that occur between ligand activation of the N

receptor and nuclear activation of downstream gene targets have not been described in

detail in Drosophila. Genetic data as well as biochemical data from mammalian Notch

studies have provided a model for NIC0 release from the membrane, entry into the nucleus

and activation of transcription. To date, however, endogenous N protein has never been

observed within the nucleus at any time during Drosoplzilo development, leading some to

question whether the NICD actually enters the nucleus. Researchers have addressed this

problem by establishing experimental paradigms where the outcome depends on nuclear

entry of N"'. Specifically, these studies involved using constructs where the NrCO was

fused to the DNA binding domain of GALA alone or dong with the transcriptional

activation domain of VPl6 (Lecourtois and Schweisguth, 1998; Stnihl and Adachi,

1998). In the presence of a UAS-P galactosidase reporter gene, observation of reporter

gene expression would only occur if GALA enters the nucleus where it can activüte

transcription. Researchers found that expression of the reporter occurred in a N-

dependent manner (Le. where N was expected to be activated) and in a ligand-dependent

manner (as mutations in Dl blocked activation of the reporter gene). The functional

significance of these constructs was demonstrated by a correlation between their ability to

activate transcription and their ability to rescue N mutations. Researchers also

demonstrated that when a transcriptional repressor domain was fused to the NICD,

activation of reporter gene expression and rescuing ability were abolished (Stmhl and

Adachi, 1998). These studies went on to show that activation of the reporter was

dependent on Su(H), suggesting that NICD interacts with Su(H) to form a transcriptional

activator (Lecourtois and Schweisguth, 1998).

A caveat to these experiments is that evcn with these artificial protein constmcts,

the N fusion products could not be detected in the nucleus, suggesting that very small

amounts of protein (below the level of immunodeteciion) were able to activate

transcription. This raises the possibility that small amounts of the NIC0 fusion protein can

enter the nucleus in an unregulated manner and do not reflect the normal action of N

signaling. For example, these constmcts could rescue N mutations through a mechanism

that does not involve nuclearaccess of^"^. Subsequent degradation of these fusion

constmcts produces NICD fragments, some of which might enter the nucleus and activate

transcription of the reporter. Evidence supporting a model where does not entcr the

nucleus cornes from studies where the cleavage site that generates the NICD fragment is

mutated. This constmct reduces the ability of the protein to signal but does not abolish

signaling (Schroeter et al., 1998). As well, mutations in the nuclear localization sequence

in human Notchl do not abolish N's ability to signal (Aster et al., 1997).

The involvement of Su(H) in binding to NICD may also be more complex than to

create a transcriptional activator in the nucleus. For example, there is data supporting the

idea that Su(H) can enter the nucleus and activate transcription independent of the N"~ .

Ectopically expressing wild type Su(H) can give rise to N gain-of-function phenotypes in

developing sense organs, suggesting that Su(H) by iiself can activate transcription of N

target genes (Schweisguth and Posakony, 1994). Other evidence implicates a role for N

signaling in regulating the localization of Su(H) in the cytoplasm versus the nucleus.

Su(H) can bind to the intracellular domain of N and in cell culture, Su(H) CO-localizes

with N (Fortini and Artavanis-Tsakonas, 1994). Upon ligand binding, Su(H) localizes to

the nucleus. In developing sense organs, Su(H) expression is found to be localized in the

nucleus of newly determined socket cells, where N activation is thought to occur, soon

after division of the pIIA cell. In differentiating socket cells (where presumably N

signaling is no longer active), Su(H) is found in both the cytoplasm and nucleus. In

addition, N signaling appears to regulate Su(H) localization between the cytoplasm and

nucleus in socket cells. The temperature-sensitive N allele, as well as expression of the

N"~, causes Su(H) to localize predominantly in the nucleus of differentiating socket

cells. Moreover, overexpression of deltex, whose binding site on N overlaps that of

Su(H), causes nuclear localization of Su(H) (Gho et al., 1996). These results suggest that

N signaling is involved in regulating cytoplasmic retention of Su(H).

Mam appears to be an essential component of the Su(H)- N ' ' ~ transcriptional

activator complex. As stated earlier, recent characterization of a mammalian homolog,

MAMLI, revealed that MAMLl expression can alter the localization of the intracellular

domain of Notchl (ICNI) and RBP-JK such that they CO-localize with MAMLI. In

addition, MAMLl coimmunoprecipitated with ICNl and this interaction was enhanced

by the presence of RBP-JK. MAMLl could immunoprecipitate with RBP-JK only in the

presence of ICN1. In vitro binding studies further determined that MAMLl binding to

ICNl required the ANK repeats. These results strongly suggested that MAMLl forms a

complex with ICNl and RBP-JK within the nucleus (Wu et al., 2000). Support for this

hypothesis was provided by the demonstration that MAMLl and ICNl RAWANK

region can cause an electrophoretic shift of RBP-JK bound to DNA. Neither MAMLl

alone, nor the RAWANK region possessed this ability. Furthemore, it was shown that

MAMLl potentiated N-dependent expression of a human Ettliancer of Split (FIES)

promoter construct (Wu et al., 2000). Another study has obtained similar results, and has

further demonstrated that Drosophifn Mam c m f o m a complex with and Su(H)

(Petcherski and Kimble, 2000).

As stated earlier, N signaling has both Sri(H-dependent and Sir(H)-independent

functions. It now appears that Srt(H) itself has functions that are antagonistic to and

independent of N signaling. By analysing single-niirided (sini) expression in the embryo,

researchers found that Stt(H) represses siin expression in the mesectoderm independent of

N. Nactivation antagonizes Sri(H) repression of siit~, allowing for sim expression in a

single row of cells in the mesectodem. Finally, N activation of sint expression requires

Sri(H) (Morel and Schweisguth, 2000). These data suggest that &(H) generally represses

N targets in the mesectoderm, independent of N. In cells where N is activated, this

repression is relieved and hi(H) acts with N to activate target genes. Another study used

the ectopic expression of three Su(H) constructs during wing development - a Su(H)

fusion with the activation domain of VP16 (Su(H)VP16), a Su(H) with a nuclear

localization signal (Su(H)NLS), and a Su(H) lacking the H binding domain (Su(H)AH) -

and produced results supporting this mode1 (Fumols and Bray, 2000). Ectopic

expression of the Su(H)VP16 construct could fully mimic N gain-of-function phenotypes

in the context of wing development. Ectopic expression of wild type Su(H) and

Su(H)NLS gave rise to both gain-of-function and loss-of-function phenotypes, while

expression of Su(H)AH gave nse to loss-of-function phenotypes. These results suggest

that Su(H) has two functions: a repressive role in cells with low N signaling where Su(H)

binds in a complex to repress N targets, and an inductive role where N activation causes

Su(H) to form transcnptional activation complexes (Fumols and Bray, 2000).

1.5 - Ubiquitination and Signaling

The ubiquitination pathway

Ubiquitin and the ubiquitination pathway are highly conserved between species,

and have been shown to be important for regulating many cellular events such as protein

degradation, cell cycling, DNA repair, endocytosis, receptor internalization, and cell

signaling (Hershko and Ciechanover, 1998). Thus, based on the molecular and

biochemical descriptions of the mechanism of N signaling, it is likely that ubiquitination

plays a role during N signaling both as a positive and a negative regulator.

Ubiquitin is a 76 amino acid residue protein that attaches covalently to target

proteins via lysine residues. Activation of the ubiquitin pathway begins with the covalent

attachment of free ubiquitin to the ubiquitin activating enzyme El. Ubiquitin inolecules

form a thioester bond between glycine residue 76 at the carboxyl end of the protein and a

conserved cysteine in the E l activating enzyme. The ubiquitin moleculc is then

transferred to the ubiquitin conjugating (Ubc) enzyme E2, again forming a thioester bond.

Then, through interaction with an E3 ubiquitin ligase target protein complex, the

ubiquitin molecule is transferred to the target protein, forming a bond with lysine residues

on the target protein. Further rounds of ubiquitination can lead to polyubiquitination of

the target protein as each new ubiquitin molecule forms bonds with lysines present on the

previously attached ubiquitin molecule (Fig. 6) (Hershko and Ciechanover, 1998).

Figure 6 - Ubiquitin pathway. Ubiquitination begins with the activation and covalent conjugation of free ubiquitin (Ub) to the ubiquitin activating enzyme El. The El associates with a ubiquitin conjugating E2 enzyme and transfers the ubiquitin molecule to the E? enzyme. An E3 ubiquitin ligase molecule or complex brings the target substrate to the E2 enzyme and catalyzes the transfer of ubiquitin to the target protein. Several rounds of ubiquitination may lead to apolyubiquitinated substrate, which can act as a signal for protein degradation. A putative E4 enzyme regulates the Ub chain length.

Recently, a fourth factor, E4, in the ubiquitin pathway has been proposed by the

identification of Ufd2. This protein appears to be involved in regulating ubiquitin chain

length on target proteins and represents a possible ubiquitin elongation factor (Koegl et

al., 1999). In addition, the identification of several deubiquinating proteins suggests that

the control of ubiquitin chain length is an important event in detenining the outcome of

polyubiquitination (Wilkinson, 2000).

The substrate specificity of the ubiquitination pathway lies with the E3 ubiquitin

ligase and as such, there exist many different E3 ubiquitin ligases. Identification of a

HECT domain within the protein E6-AP, and demonstration that HECT domains confer

E2 binding and ubiquitin transfer to target proteins helped to identify the subclass of

HECT domain E3 ubiquitin ligases (Huibregtse et al., 1995; Kumar et al., 1997;

Scheffner et al., 1994; Scheffner et al., 1995). Recently, the demonstration that RING

finger domains can confer ubiquitin ligase activity in vitro (Lonck et al., 1999; Joazeiro

and Weissman, 2000). and that such activity has functional significance in vivo (Joazeiro

et al., 1999; Levkowitz et al., 1999; Waterman et al., 1999). revealed an additional class

of E3 ubiquitin ligases.

Differences in how E3 ubiquitin ligases function are conferred not only by the

type of enzymatic domain present, but in their requirement for other factors. E3 ubiquitin

ligases can function either within multiprotein complexes that recognize target proteins

and associate with UbcE2 enzyme, or individually bind to targets and associate with

UbcE2. An example of a multiprotein complex E3 ubiquitin ligase is the well

charactenzed SCF E3 ubiquitin ligase complex involved in cell cycling in yeast. This

complex consisting of Skpl, cullen-1, and an F-box protein requires the RING domain-

containing Rbxl protein to ubiquitinate substrates (Skowyra et al., 1999; Tyers and

Jorgensen, 2000). In addition, the human homolog of Rbxl, Rocl, can interact with

many members of the cullen family of proteins, and by utilizing other adaptor proteins

can form new ubiquitin ligase complexes (Kamura et al., 1999; Ohta et al., 1999). Thus it

appears that various combinations of subunits can form distinct multiprotein E3 ubiquitin

ligases. In contrast to Rbxl, c-Cbl is a RING finger containing E3 ubiquitin ligase that

appears to function independently of other factors. Via its SH2 domain, c-Cbl is capable

of associating with target receptor protein tyrosine kinases including EGFR and

associates with the E2 enzyme via its RWG finger domain [Joazeiro et al., 1999;

(Waterman et al., 1999; Zheng et al., 2000). There does not appear to be a comlation

between how the ubiquitin ligase functions and the type of enzymatic domain it contains.

Thus, the composition of the enzymatic complex must be examined for individual E3

ubiquitin ligases.

The extent of ubiquitination as well as the choice of lysine may be important in

distir:guishing whether the target protein is marked for degradation by proteasomes, or

targeted for some other cellular event. The identification of the polyubiquitin receptor

protein, Rpn 10 or SSa, led to the discovery that Rpn 10 binds to polyubiquitinated chains

containing more than four ubiquitin molecules (Deveraux et al., 1994). Rpn 10 is part of

the proteasome complex responsible for identifying polyubiquitinated proteins targeted

for degradation. The ubiquitin molecule itself has seven lysine residues that cm act as

possible sites for polyubiquitination. Polyubiquitin chains have been detected on four of

these lysine residues (Pickart, 1997). Evidence that the site seiection of

polyubiquitination c m confer distinct signals comes from the observation that growing

ubiquitin chains attach ubiquiïin to the same lysine position used in the first attachment.

Thus single polyubiquitin chains contain molecules al1 attached to !he same lysine

residue. Mutations made to yeast ubiquitin revealed that only specific lysine residue sites

can confer the signal leading to protein degndation. Mutation of each of the lysine

residues in ubiquitin revealed that only lysine 48 appeared to be involved in protein

degradation. Mutation of this lysine to arginine caused defects in proteolysis, while al1

other mutations appeared to maintain normal levels of protein degrüdation (Finley et al.,

1994). Furthemore, while mutating lysine 63 does not affect levels of proteolysis, yeast

expressing this mutation were sensitive to DNA mutagens, implicating polyubiquitination

on lysine 63 in DNA repair (Spence et al., 1995).

Ubiquitination and signaling

It has been demonstrated that ubiquitination can regulate signaling by cell surface

receptors through a variety of mechanisms. One such mechanism involves the activation

of membrane bound transcription factors. This mechanism was demonstrated by the

observation that the RSP5 E3 ubiquitin ligase in yeast binds and ubiquitinates the

transcription factor SPT23. SPT23 and MGA2, membrane bound proteins, regulate the

transcription of the OLEl gene which is involved in fatty acid synthesis. Ubiquitination

by RSP5 was shown to be necessary for SPT23 and MGA2 processing and its subsequent

ability to activate OLEl transcription (Hoppe et al., 2000). This result suggests that

endoproteolytic processing by proteasomes enables the transcription activation domains

of SPT23lMGA2 to traffic to the nucleus (Hoppe et al., 2000).

Evidence exists that ubiquitination may also be a signal for receptor

intemalization through endocytosis (Strous and Govers, 1999). An example is the growth

hormone receptor (GHR) which is endocytosed upon ligand binding. Endocytosis only

occurs in the presence of ubiquitination, indicating that ubiquitination of the GHR signals

its intemalization (Govers et al., 1999; Strous et al., 1996). As well, identificatiûn of a

ubiquitin conjugation motif required for endocytosis, UbE, confims the role of

ubiquitination in GHR internalization and activation (Govers et al., 1999). Furthermore,

the ubiquitination pathway also regulates the degradation and down-regulation of the

GHR after endocytosis revealing a role for ubiquitination in regulating the strength of

GHR signaling (Alves dos Santos et al., 2001). Another example of ubiquitination

involvement in regulating signaling strength was revealed by the demonstration that c-

Cbl acts as an E3 ubiquitin ligase (Joazeiro et al., 1999). c-Cbl acts in the down

regulation of the epidermal growth factor receptor (EGFR) and requires the

ubiquitination ligase activity conferred by a RING finger domain (Levkowitz et al., 1999;

Wateman et al., 1999; Yokouchi et al., 1999; Zheng et al., 2000). One can envision that

by removing components of the pathway, including the receptor itself, the strength of the

signal can be reduced by ubiquitin-mediated proteolysis. Similarly, if inhibitory proteins

are removed, or if components in competing pathways are removed, the strength of the

pathway can be increased. Thus, ubiquitination can regulate signal transduction at the

level of receptor intemalization and within the transduction cascade itself.

Of particular interest to N signaling, evidence exists indicating the possible

involvement of ubiquitination in propagating the N signal. The molecular

characterization of Sic(&) revealed that it is a HECT domain containing protein and

predicts that Su(dx) is an E3 ubiquitin ligase (Comell et al., 1999). A related mammalian

gene, Itch, has been shown to bind and ubiquitinate N in vitro and in human Jurkat cells

(Qiu et al., 2000). Additional genetic evidence cames from the demonstration that

Drosophila mutations in the 26s proteasome 82 and subunits phenocopy N gain-of-

function mutations during sense organ differentiation. In addition, these mutations

enhanced the phenotype of expression and stabilized N~~~ protein (Schweisguth.

1999). While these are interesting results, there is no direct evidence that ubiquitination

is indeed a normal mechanism of regulating N signaling. It has yet to be demonstrated

that Su(dx) behaves biochernically as an E3 ubiquitin ligase, and the functional

significance of Itch ubiquitination of N has not been demonstrated in vivo. Similarly, the

effects of the 82/86 subunit mutations may not represent a direct consequence of altered

N processing and may instead reflect a consequence of general defects in cellular protein

degradation.

Regardless, one can certainly envision how ubiquitination may be used to regulate

or trigger events required in N signaling. N receptor internalization and activation

through endocytosis may possibly be triggered by ubiquitination. As well, ubiquitination

may be used to tag N for proteolytic cleavage, including those cleavage events required

for the putative generation and release of the N " ~ . In addition, N downregulation once

signalin2 is completed may be achieved by ubiquitin-mediated proteolysis. Finally,

endocytosis, proteolysis, and processing of the Dl ligand could very well be rnediated by

ubiquitination.

1.6 - The function of nercralized during Drosopliila development and within the

Notclr pathway

As previously described, neir encodes a RING finger domain containing protein

whose function is unknown. The Drosoplrila sequence predicts that rreir encodes a

nuclear protein with DNA binding potential. However, comparison of the Drosophila

protein to homologs in C. elegarrs and human reveals that many domains including the

putative nuclear translocation sequence are not conserved. In contrast, the carboxyl-

terminal RiNG finger is highly conserved. In addition, two regions of conservation

termed Neuralized homology repeats (NHR) are revealed. The studies described in the

next two chapters will attempt to elucidate the function of rieu during Drosopliila

development, as well as within the N signaling pathway.

To gain insight into how neir functions within the N pathway, mosaic analysis was

used to study the involvement of neit during sense organ development. Mutant clones of

two alleles, the hypomorphic rred"' allele and the nuIl allele, were generated

prior to sense organ development and the effects of these clones on the development of

sense organs were observed. From this study it was possible to determine if Neu is

involved in presenting the ligand to trigger the N signaling pathway, in which case it

functions cell non-autonomously, versus whether Neu is involved in transducing the N

signal and thus functions cell autonomously. In addition, the involvemznt of nerc during

the SOP lineage was characterized. The results indicate that rieu functions cell

autonomously to transduce or propagate the N signal throughout SOP development.

Furthermore, using a myc-tagged Neu construct, Neu was found to be localized at the

plasma membrane. These results are descnbed in Chapter Two.

The function of the Neu protein was then investigated. Recently, it has been

demonstrated that RING finger domains may function as ubiquitin ligases (Lorick et al.,

1999). raising the possibility that the RING finger domain of Neu may function in this

manner. To investigate this further. GST-fusion proteins were made with various Neu

constructs that contain or lack an intact RING finger, and the ability of these fusion

proteins to function as ubiquitin ligases in vitro was assayed. In addition, various myc-

tagged Neu deletion constructs were expressed in vivo, and the effects of deleting the

NHR and RING finger domains on protein localization and function were assayed.

Based on the results of these experiments, described in Chapter Three, it appears that Neu

functions as a ubiquitin ligase in vitro and that the ubiquitin ligase activity is conferred

solely by the RING finger domain. Genetically, Neu requires the RING finger domain to

rescue rieu mutants, and the NHR domains appear to be responsible for plasma membrane

localization. Taken together, these data provide evidence that Neu functions as a

ubiquitin ligase during N signaling.

The mosaic analysis was then extended to characterize the involvement of rteii

during the development of other tissues, particularly those in which N signaling has been

implicated. The effects of mutant neu clones during eye and wing development were

observed. The role of neu during oogenesis was determined through a combination of

ectopic expression studies and mosaic analysis. Involvement in the N signaling pathway

predicts that loss of neu function during the development of these tissues should result in

N-like phenotypes. The results of these experiments indicate that net1 does indeed

function to transduce N signal during the development of these tissues. However, neir

appears to be redundant or non-essential for some N signaling events suggesting that rieii

may not be a general component of the signaling pathway. Duc to the preliminary nature

of these results, they are provided in Appendix A.

Understanding how individual components of the N pathway function has

ultimately led to a better description of the mechanisms involved in N signaling. The

goal of the following studies was to understand the function of one such component, iterr,

during Drosopliila development. Understanding the role of neu during various

developrnental processes has provided information into how tteii functions within the N

signaling pathway. The determination of Neu's biochemical function has provided

details of lieu involvement in N signaling, and has provided insights into the mechanisms

involved in N signaling.

Chapter 2 index - Neuralized functions ce11 autonomously to regulate Drosophila sense organ development

2.2 - Materials and Methods

Fly stocks and transgenics 0 Immunocytochemistry and in situ hybridization

Rescue of neurF6' O Mosaic analysis 0 BrdU labelling

Scanning and Transmission Electron Microscopy Imaging and microscopy

2.3 - Results

Loss of neu function affects cell fate decisions during sense organ development 0 Tuïting caused by ne~r"'~' mutant clones is the result of supernumerary SOP

determination neu is also required for pIIa/pIIb and accessory cell fate determination

0 neii functions cell autonomously in sense organ development 0 Neu is associated with the plasma membrane

neir is required cell autonomously for specification of non-neuronal cell fates within the proneural cluster Cell autonomous function and neir expression

0 Neu is associnted with the plasma membrane

This chapter essentially appears as published as (Yeh et al., 2000). The contributions of the authors are as follows: L. Zhou generated the myc-tagged Neu transgenic line; N. Rudzik generated the scanning electron micrographs; E. Yeh perfonned the remainder of the studies. This work was done under the supervision of G.L. Boulianne.

2.1 - Introduction

The process of sense organ (SO) formation in Drosophila is well-characterized at

the genetic and cellular levels and provides an ideal model to study the role of cell

lineage and cell-cell interactions during development (Hartenstein and Posakony, 1989;

Huang et al., 1991; Posakony, 1994). Bristle SOS are compnsed of four cell types -

tormogen (socket), tricogen (shaft), thecogen (sheath) and neuron - that anse from a

single sense organ precursor (SOP). A fifth cell, the soma sheath cell or glia cell, is

associated with each SO. SOP determination occurs within an equivalence group called a

proneural cluster and requires the action of proneural group genes (Garcia-Bellido, 1979;

Ghysen and O'Kane, 1989; Simpson, 1990). After primary SOP determination, al1 other

cells within the proneural cluster are prevented from becoming SOPs through a process

of lateral or mutual inhibition and these other cells adopt an epidermal cell fate (Ghysen

et al., 1993). The primary SOP divides asymmetrically to produce two secondary SOPs

called pIIa and plIb. pIIa will divide to produce the shaft and socket cell (Hartenstein

and Posakony, 1989). while pIIb founds a lineage that produces the neuron and sheath

cell. Recently it has been demonstrated that the pIIb cell divides first to produce a glial

cell and a daughter cell named pIIIb. The pIIIb cell divides to produce the neuron and

sheath, indicating that the SOP lineage yields al1 five cells associated with a mature SO

(Gho et al., 1999). Although the SOP undergoes a stereotypical pattern of cell division

and cell lineage is important in cell fate detemination, cell-cell interactions have also

been shown to play an important role in the determination of SOP and daughter cell fates

(Hartenstein and Posakony, 1990).

Many of the cell-ceil interactions that are important in SO development are

thought to be regulated by neurogenic genes. Neurogenic genes, including Notcli (N),

Delta (DI). big brairi (bib), rnastemtiiid (iiianr) and neirralized (tieit), were first identified

as embryonic recessive lethal mutations that cause hyperpiasia of the embryonic nervous

system at the expense of epidennal tissue (Lehmann et al., 1983). The best characterized

members of the group. N and Dl, function as a receptor and ligand, respectively. Other

members of the group are believed to play a role in generating the signal or propagating

the signal. Besides providing the signaling pathway that is believed responsible for

lateral or mutual inhibition within proneural fields, it is thought that neurogenic genes

function together as a genetic cassette to regulate cell-cell interactions important for cell

fate decisions in a variety of tissues during development (Ruohola-Baker et al., 1994).

The role of neurogenic genes in SO development was demonstrated by generating

mutant clones of N or Dl cells during development (Dietrich and Campos-Ortega, 1984).

These studies revealed that mutant clones of N or Dl exhibit specific defects in bristle

development. In Dl clones, these defects included tufting (supemumerary SO bristles),

whereas in N clones both tufting and balding (absence of bristles) were observed. The

variability observed in N clones appeared to be due to spatial differences within the

notum; some regions yielded the tufted phenotype while other regions produced only the

balding phenotype.

A detailed analysis of mutant clones of N, Dl, and sliaggy (sgg), another

neurogenic gene, using adult epidennal markers confirmed the involvement of these

genes in SO development (Heitzler and Simpson, 1991). Analysis of the genotypes of

bristles found at the boundaries between mutant and wild type cells revealed that N is

required autonomously for receiving the neurogenic signal that prevents cells within the

proneural cluster from adopting the SOP fate. The same type of analysis revealed that Dl

is required non-autonomously to produce the signal that allows epidermal ce11

specification. sgg was found to be required cell autonomously to send and receive the

neurogenic signal. Taken together, these results provided evidence that N and Dl

function in signaling as a receptor and ligand and that sgg probably plays a role in a

feedback-based regulatory mechanism (Heitzler and Simpson, 1991).

Temperature-sensitive alleles have been used to furtlier elaborate the role of N and

Dl in SO development (Hartenstein and Posakony, 1990; Parks and Muskavitch, 1993).

These studies have shown that loss of N function pnor to or during the determination of

the pnmary SOP causes supernumerary pnmary SOPs to form. These extra SOPs

develop normally and produce bnstle tufts. When N loss-of-function is induced

subsequent to pnmary SOP determination and dunng the divisioii 2nd differentiation of

the accessory cells, al1 the cells in the SOP lineage are transformed iiito neurons, resulting

in a bald phenotype. Further analysis revealed that N is required at every step of the SO

lineage; proper determination of the pIIa and pIIb fates, as well as the accessory cell

fates, requires N signaling. These results explain the apparent spatially dependent

phenotypes caused by N mutant clones and suggest that chaetae do not develop

synchronously. This approach also revealed similar requirements for Dl dunng SOP

developrnent.

The neurogenic gene nerr has also been implicated in SO development. nerr, like

N and Dl, was first identified by means of loss-of-function mutations that cause

hyperplasia of the central and penpheral nervous system at the expense of the epidermis

(Lehmann et al., 1983). Although neic genetically interacts with other neurogenic genes,

its role within this pathway remains unclear. The function of Neu protein is also

unknown. The amino acid sequence of Neu suggests that it might encode a nuclear

protein with a putative nuclex localization signal, helix-tum-helix domain and a CJHCJ

Zn-finger ("RING") domain at the C-terminus (Boulianne et al., 1991; Price et al., 1993).

However, the subcellular distribution of Neu has yet to be determined. Homologs of iieii

have been identified in other species including human (Nakamura et al., 1998). mouse

(Moschonas, 1998) and C. elegaits (Wilson et al., 1994) suggesting that the function of

rietc in N-Dl signaling has been conserved. Comparison of these sequences reveals that

the RING finger dornain is present within al1 homologs. However, the putative nuc l~ î r

localization signal and helix-tum-helix domains have not been well conserved.

In situ hybridization studies have s h o w that tteic is broadly expressed during

early embryogcnesis, but becomes restricted to the ventral neurogenic region and

eventually to neuroblasts during neuroblast determination. During the third larval instar

stage, neic is expressed in SOP cells that will give rise to macrochaetae on the adult

notum (Boulianne et al., 1991). Dietrich and Campos-Ortega (Dietrich and Campos-

Ortega, 1984) camed out mosaic analysis to determine the role of neic in SO development

and found that iteic mutant clones gave rise to a balding phenotype which, in contras1 to

N, are spatially independent. However, these studies did no1 reveal the cellular nature of

these defects. Furthermore, in the absence of appropriate markers, these studies could not

establish whether neic was required autonomously or non-autonomously during SO

development.

To further characterize the role of neri in the neurogenic signaling pathway we

have studied its function during SO development by generating mutant clones using the

FLP-FRT system (Golic and Lindquist, 1989). Using two independent rie11 alleles we

find that ne11 is required for epidermal cell fate determination within the proneural cluster.

neii mutant clones that overlap proneural regions exhibit supernumerary determination of

SOPs. Analysis of the bristle genotypes found at clonal boundarirs reveals that neii

functions cell autonomously in receiving the signal that prevents SOP determination. As

well, loss of iieii function produces phenotypes similar to those seen in loss-of-function

mutants for N and DI, indicating that it is required for proper determination of cell fates

in the SO lineage. To gain insight into where neir functions in the neurogenic pathway,

we examined the localization of Neu within the cell. Using the GAWIUAS system

(Brand and Pemmon, 1993) to express wildtype and epitope-tagged iieii constructs

during development, we show that both constructs are able to partially rescue the

embryonic neurogenic phenotype caused by a mutation in neii and that Neu localizes to

the plasma membrane. We propose that the function of neii during SO development is to

modulate the efficacy of neurogenic signaling within the proneural cluster by affecting

the ability of cells to receive or propagate signals through N and DI.

2.2 - Materials and Methods

Fly stocks and transgenics

Fly stocks were kindly provided by the following individuals: scu-GALA (Y.N.

Jan, UC San Francisco); e22c-GAL4 UAS-FLP 1 CyO; FRT82B. RM (D. Harrison, U of

Kentucky); y iv hsflp; FRT82B. y+nM (T. XU, Yale); prpivn FLP[38] 1 CyO; Ki

ka?ryso6 and prpwn; FRT82B. ka?rym6 /FRT82B, k ~ 1 ? r y ' ~ ~ b x ' J ~ ~ ~ ( 2 : 3 ) ~ 3 2 (P.

Simpson, IGBMC). FRT82BirM and pic-GAIA were obtained from the Bloomington

Stock Center (82004 and B2017). The allele ~ieri"~' is a hypomorphic allele caused by a

lacZ enhancer trap insertion into the neir gene (Bellen et al., 1989) and is available from

the Bloomington Stock Center (B4369). nedF6.' is an amorphic allele (de la Concha, et

al., 1988; (Brand and Campos-Ortega, 1988) and was provided by Y.N. Jan. A78 is a

GAL4 line generated from an enhancer trap screen (Gustafson and Boulianne, 1996)

which expresses ubiquitously during embryonic and larval development. All fly stocks

were maintained at either room temperature or lS°C on standard corn meal agar media.

A 3.2 kb full length neri cDNA was subcloned as a K p d fragment into the vector

pUAST (Brand and Pemmon, 1993) and introduced into flies by standard P-element-

mediated transformation (Spradling, 1986). Thirteen independent insertions were

generated and balanced over FM7B. SM.5, or TM3. Seven lines expressing a C-terminal

myc-tagged neri construct (generated by PCR using specific primers) were also created.

Expression of wildtype and epitope-tagged neii during development was achieved by

crossing lines canying these constructs to various GAL4 lines includingpfc-GAIA and

sca-GAIA.

Immunocytochemistry and in situ hybridization

Imaginal discs and pupal tissues were dissected in PBS ( l x PBS recipe in

(Verheyen and Cooley, 1994)). Immunocytochemistry was perfomed essentially as

described (Patel. 1994). Primary antibodies were used at the following dilutions: mouse

anti-22ClO (C. Goodman, UC Berkeley), 1:10; rabbit anti-P-gal (Cappel), 1:2000; mouse

anti-Bromodeoxyuridine (BrdU) (Sigma), 1:1000; rabbit anti-phosphohistone H3

(Upstate Biotechnology), 1:250; mouse anti-myc. undiluted, and rabbit anti-spectrin

(kindly pmvided by D. Bnnton, Harvard), 1:250. HRP conjugated anii-mouse and anti-

rabbit secondaries (BioRad) were used at 1:200. FITC-conjugated donkey anti-rabbit and

Cy3-conjugated donkey anti-mouse (Jackson labs) were used at 1:200. FITC conjugated

sheep anti-HRP (ICN) was used at 1:200. Samples stained with DAPI were treated as

previously described (Verheyen and Coolcy, 1994).

In situ hybridization was peiîormed as previously described (Bouliannc et al..

1991). Third instar lawae were staged based on feeding behavior and the absence or

presence of bromophenol blue dyed food in the gut.

Rescue of rieitfF6'

To assay the ability of wildtype neii or myc-tagged neir to rcscue the ~ i e i r ' ~ ~ '

mutation, embryos from the cross ptc-GALAIGFP, Cy; I I ~ ~ < ' ~ ~ ~ I T M ~ and UAS-neitlUAS-

rieu; or ptc-GAJAIGFP, Cy; 1leir '~~~lTM3 and UAS-neirmyclUAS-neitrnyc;

1 r e i r ' ~ ~ ~ 1 ~ ~ 6 were collected ovemight and allowed to develop for an additional 24 hours

at room temperature. Embryos were then processed for in situ hybridization with

antisense neii probe and immunocytochemistry with FITC conjugated anti-HRP antibody.

Rescue of the neurogenic phenotype was then determined in embryos displayingptc-

GALA driven expression.

Ectopic expression of neu during third larval instar imaginai disc developrnent

was achieved by crossing UAS-nefi lines to the A78 CAL4 enhancer trap line.

Mosaic analysis

The alleles neii "Io' and neitF6j were recombined onto third chromosome arms

containing FRT sequences at 82B as outlined by (Xu and Hamson, 1994) and balanced

over TM3. Mosaic clones of neu' cells were generated using the FLP-FRT system (Golic

and Lindquist, 1989; Xu and Rubin, 1993). The source of FLP-rccombinase was either

11sJlp which consists of the& gene under Iisp70 control, or e22cGAL4 UAS-flp, which

consists of a GALA enhancer trap that is expressed ubiquitously al low levels during early

Drosopliila development when flies are reared at 2g°C (data not shown; D. Hamson,

personal communication).

To generate clones using the heat shock-FLP construct, y IV lis/lp/y IV hs/p;

FRT82 y+nM (IV*) 1 FRT82B yCnM (ivC) females were crossed to FRT82B

,leliAIOl or IF65 1 TM3 males and y i v Iisfip; FRT82B r i e ~ i " ' ~ ' ~ ~ ' ~ ~ ~ 1 FRT82B ytnM (ivC)

male progeny were then crossed back to y iv hsflp/y i v hsflp; FRT82 y+nM (ivC)

1 FRT82B yCnM (IV+) Sbuz6 females. Embryos from this cross were collecied in

vials for 48 hours at 18°C. subjected to a 1 hour heat shock at 37'C and allowed to

complete development at room temperature. To mark adult clones with the epidermal

hair markerpivn, FRT82B neik"' 1 TM3 were crossed to prpivn FLP[38] 1 CyO; Ki

ka?r$O%nd Ki, Cy+ male progeny were mated to prpivn; FRT82B. kaiVso6/ FRT82B.

k a ? r y ' " b ~ ' ~ ~ ~ ~ p ( 2 : 3 ) ~ 3 2 . Progeny from this cross were collected at room temperature

and heat shocked for 1 hour at 37"C at approximately 24 hours after egg laying.

GALA driven expression of FLP during early development was obtained by

crossing e22cGAL4, UAS-FLP 1 CyO; FRT82B. nM 1 FRT82B. nM with FRT82B.

~ieic"'~' 1 TM3. Embryos from this cross were collected in vials for 24 hours at room

temperature, and then allowed to develop at 29°C.

BrdU lnbeling

Late larval third instar imagina1 wing discs were dissected in Schneider's medium

(Schneider, 1964), transferred to fresh medium containing 0.01 mM BrdU and incabated

at room temperature with slight rotation for 2 hours. After three washes in PBS, the discs

were fixed for 20 minutes with 4% formaldehyde in PBS, washed ihree times with PBT

( l x PBS, 0.1% Triton-X) and blocked 30 minutes with blocking solution (PBT, 2%

bovine serum albumin, 1% normal goat serum). Detection of both p-gal and BrdU was

as follows. Discs were first incubated with rabbit anti-0-gal overnight at 4°C. washed

three times 10 minutes with PBT, incubated 30 minutes in blocking solution and

incubated with FITC conjugated donkey anti-rabbit antibody for 2 hours at room

temperature. After three washes of 10 minutes with PBT, a second fixation of 15 minutes

with 4% formaldehyde in PBT was performed. Discs were then washed three times 10

minutes with PBT and hydrolyzed with 2N HCI (4N HCI diluted 1: 1 with PBT) for 35

minutes. Samples were washed three times 10 minutes with PBT and incubated

overnight at 4'C with anti-BrdU. Discs were then washed three times 10 minutes with

PBT and incubated with Cy3-conjugated donkey anti-mouse for 2 hours at room

temperature. After three washes of 10 minutes with PBT, discs were further dissected

and mounted in 70% glycerol containing 2% DABCO (in PBS).

Scanning and Transmission Electron Microscopy

Adult heads were fixed and embedded for scanning electron microscopy as

described by (Carthew and Rubin, 1990). Sectioned adult eyes were fixed and embedded

as described by (Basler and Hafen, 1988).

Imaging and microscopy

Adult structures and dissected tissue samples were examined using a Nikon

Optiphot 2 equipped with Nomarski optics. Fluorescent and light images were captured

using a Sony digital camera and analyzed using Northem Exposure software. Confocal

images were obtained using a Leica confocal microscope and PC computer ~ n n i n g

Scanware (Leica). Pseudo color was added to confocal images using Adobe Photoshop.

Results

Loss of neu function affects cell fate decisions during sense organ ùevelopment

To assay the effecis of loss of neii during SO development we generated mutant

clones using FLP/FRT-mediated somatic recombination (Golic and Lindquist, 1989; Xu

and Harrison, 1994). For these purposes, two alleles of rleii were recombined onto third

chromosome arms containing FRT sequences at 82B. rteiiNm is a hypomorphic ai:c!e

resulting from the insertion of a lacZ enhancer trap into the upstream regulatory region of

the ncu locus (Bellen et al, 1989), while rteti IF'' is an amorphic EMS induced allelc (de la

Concha et al, 1988; Brand and Campos-Ortega, 1988). Both alleles produce severe

hyperplasia of the embryonic nervous system leading to lethality and both fail to

complement any other known iteii allele.

Flies heterozygous for neilMo' and carrying a source of FLP (e22cGAIA, UAS-

FLP, FRT82B. rrM 1 FRT82B. ~ te i t '~ ' ) displayed a bristle tufting phenotype afyecting

both macrochaetae and microchaetae. The severity of the phenotype ranged fror!i

duplicated bristles to tufts containing several bristles (Fig. 7). Supernumerary

macrochaetae and microchaetae were always found in characteristically normal positions.

With the exception of extreme cases of microchaetae tufting observed only at the

anterior-most part of the notum (Fig. 7C), regions between bristles and bristle tufts did

not appear to be affected. This suggests that neii functions to prevent cells from adopting

SOP fates within the proneural cluster.

Figure 7 - Scanning electron micrographs of ne$'o' clones in adult eyes and nota. A) and D) show SEMs of wildtype nota. B), C), E), and F) show nota from flies in which iiet~'"~' clones have been induced. The bristle phenotype observed is tufting of both microchaetae (mowhead) and macrochaetae (arrows). Tufts appear in locations of normal bristle formation (B) and are separated by epidermal cells (E - asterisks), although severe tufts can occur at the anterior of the notum (C). The majority of bristles within tufts contain a socket associated with each shaft (F). >leuNm clones can also cause eye defects. G) shows the repetitive uniform organization of ommatidia, and regular spacing of sensory bristles in a wildtype eye. H) shows some of the defects seen in eyes in which neit clones have been induced including irregularly shaped ommatidia, and mispositioned, missing or extra sensory bristles. Severe scamng can also be observed (1).

The effects of ize~i"'~' clones were not limited to bristles on the notum. Tufting

could also be observed with adult head sensilla surrounding the eye and ocelli. In

addition, bristle sensilla throughout the body, including the dorsal and ventral abdomen,

also appeared to form tufts. As was observed for macrochaetae, these tufts always

occurred in the location where normal bristles are formed. rieii""' clones also gave rise

to defects in the adult eye. The severity of the phenotype ranged from ectopic

interommatidial bristles and aberrant ommatidial size to scaning (Fig. 7H.I) and

defective photoreceptor development (data not shown). In addition, defects in wing

development, including irregular wing mnrgin sensory bristle formation and ectopic wing

vein formation, were observed (data not shown).

Tufting caused by i~ea"'~' mutant clones is the result of supernumerary SOP determination

To determine whether the supernumerary bristles (Le., tufting) that we observe in

mutant clones are due to commitment to the SOP fate, we took advantage of the fact that

the iieu mutant allele iieu""' is a lacZ enhancer trap line in the rieu locus that can be used

as a marker of SOP determination. Previous studies have demonstrated that r~eir""'

expression can be detected within third larval instar wing imagina1 discs in primary SOPs

that give rise to macrochaetae on the notum and sensory bristles dong the wing margin

(Huang et al., 1991). As development proceeds, expression of neil*''' can also be

detected in secondary SOPs as well as the accessory cells that are associated with each

primary SOP. >ieuA''' is similarly expressed in SOPs on the pupal notum that will give

rise to microchaetae. Since the appearance and differentiation of each macrochaete SOP

is well documented, it is possible to examine the fate of each SOP at particular

developmental time points. For example, the primary SOPs that will give rise to bristles

dong the adult wing margin are determined during late third larval instar, but do not

divide until5-10 hours APF (Huang et al., 1991). Therefore, any supemumerary p-gal

positive cells along the wing margin that are observed during third larval instar

development are most likely primary SOPs rather than secondary SOPs. Using the xMyc

marker to identify neiiN0' clones, we found that supemumerary SOPs arose from i~e i~"'~'

cells (Fig. 8). Since supemumerary SOPs are not observed in ectopic locations in the

wing disc this suggests that iieii functions normally in the proneural cluster to determine

epidermal cell fates.

To ask whether the supemumerary primary SOPs arose from determination of

excess primary SOPs within the proneural cluster or from abnomal proliferation of the

primary SOPs, BrdU labeling experiments were performed. As described earlier,

wildtype primary SOPs at the wing margin remain mitotically quiescent until after

pupariation. Thus, they do not incorporate BrdU during third instar larval development

(Fig. 9A). Similarly, the supemumerary SOPs that developed dong the wing margin in

~ieii"'~' mutant clones did not incorporate BrdU (Fig. 9B, inset). Using an antibody

against a phosphorylated form of histone H3 as a marker of mitosis (Hendzel et al., 1997)

further reveals that supemumwy SOPs are not actively dividing (Fig. 9D). These two

results demonstrate that the increase in primary SOPs was not a result of aberrant mitotic

events.

neu is also required for pIIa1pIIb and accessory cell fate determination

neit mutant clones were also generated using rieic'"'. In this case, we found that

mutant clones gave rise to a balding phenotype characterized by the absence of chaetae

Figure 8 - Bristle tufts in ~ieu"'~'clones are caused by supernumerary sensory organ precursors. A) shows the SOP expression pattern of the lacZ enhancer trap that causes the mutation A101. Wing discs in B) and C) have been double stained with anti-myc (red) to identify the ~ieii"~' clone and anti-p-gal (green) to identify SOPs. When a mutant clone (absence of rnyc staining) overlaps a proneural region, supernumerary SOPs (arrows) are determined. SOPs do not a i se in ectopic locations, suggesting that rieu functions within proneural clusters to properly determine cell fates.

Figure 9 - Supemumerary SOPs are the result of changes in cell fate and not aberrant mitotis. A) In wildtype third larval instar wing discs, SOPs (green) dong the wing margin do not divide under after pupanum formation and thus do not inco orate BrdU (red) du"g larval deuelopment. B) Supemumerary SOPs caused by mug clones 8 s o do not incorporate BrdU, indicating that extra SOPs (arrow, inset) do not anse from abnormal mitosis. C) Similarly, wildtype SOPs (red) do not stain with an antibody to phosphohistone H3 (green) indicatin that they are mitotically quiescent. D) Supemumerary SOPs caused by iteuAIo' clones do not stain with anti-phosphohistone H3 indicating that they are not mitotically active.

on the adult notum (Fig. 10B). This is consistent with a role for neii in the determination

of both pIIa1pIIb and accessory cell fates. In N and Dl mutant clones. loss-of-function

dunng secondary SOP and accessory cell fate determination causes cells to assume a

neuronal cell fate (Hartenstein and Posakony, 1990). To determine if ~ieir'~*\lones give

rise to similar alterations in secondary SOP and accessory cell fates, pupal nota (24 hours

APF) were dissected and stained with the neuronal marker 22C10. In wildtype notum at

this stage, 22ClO expression is detected in a single neuron comprising each individual

sense organ (as identified by the double axon processes; Fig. 10C). In contrast, 22C10

expression in pupal nota from clones, revealed clusters of 22C10 positive cells

(Fig. 10D) that al1 displayed the double axon processes associated with neuronal

differentiation. The presence of more than four 22C10 positive cells in some clusters

further demonstrates that supemumerary primary SOPs are determined in ueir mutant

clones and that the descendants of these mutant SOPs differentiate to assume neuronal

cell fates. Taken together, these data demonstrate that ireir affects multiple cell fate

decisions required for the proper development of sense organs. Like N and Dl, neri is

required for proper determination of the pIIa cell fate, and is also required for

determination of the thecogen cell fate in the pIIIb lineage.

neu functions cell autonomously in sense organ development

Using mosaic analysis, we have demonstrated a role for neii in sense organ

development. To determine if neii is required autonomously in this process, we

generated neu mutant clones that were genetically marked with y in a y+background.

Somatic recombination was induced by heat-shocking nies of the genotype y w lrsflp;

FRT 82B, i~eii*'~'/ FRT 82B. y'~b nM during late embryogenesis. We found that the

supemumerary bristles were non-Sb and y- (Fig. 1 IA , B) demonstrating that they arose

from riedio' cells; mixtures of wildtype and mutant bristles were never observed.

To ascertain the ability of mutant cells to send or receive the signal that prevents neural

determination, and thus delineate autonomous versus non-autonomous nerr function, adult

clones were examined using the epidermal hair markerpawtt (pion), which can be used to

identify clonal boundaries on the adult cuticle. Sincepivri affects bristle morphology

(producing tmncated bristles), mutant cells can be identified as well. If mutant bristles

can be found at the clonal boundary, and these are unaffected by neighbouring wild type

cells, then netr must be required autonomously to inhibit neuronal cell fates. In contrast,

if wildtype bristles are found at the boundary next to net< mutant cells, then rieil must act

non-autonomously within cells since they fail to suppress neighbouring cells from

becoming SOS.

We found that mutant bristles exist at clonal boundaries next to wildtype cells

more frequently than wild type bristles next to mutant cells (81% versus 19%.

respectively, n=382). Furthermore, mutant bristles at the clonal boundary were observed

as either single bnstles or tufts (Fig. 1 IC, D). Thus, tietr mutant cells are affected in their

ability to receive the signal that prevents neural determination and they form SOS at

clonal borders despite the presence of wildtype cells. The ability of mutant cells to send

a signal does not appear to be affected as mixtures of wildtype and mutant bristles were

not observed. Taken together, these results clearly demonstrate that iteir functions cell

autonomously during SOP determination to specify epidermal fates in Drosopliila.

Figure 10 -The amorphic allele, 11eu'"~, causes a bald cuticle phenotype. A) shows the notum of a wildtype fly. B) When clones of I I ~ u ' ~ ~ ~ ' are generated, bristles do not form causing a bald cuticle (arrow). Epidermal cells appear to form nonnally. The bald phenotype of clones is caused by transformation of the SOP accessory cell fates into neuronal fates. C) shows the expression of a neuronal marker 22C10 in wildtype pupal nota 24 APF. Neurons can also be unambiguously identified by their double mon projections. Clusters of cells expressing 22'210 (some containing more than 4 positive staining cells) can be detected in nota (arrowheads) in which clones have been generated, indicating that multiple primary SOPs are detennined and that each SOP divides and differentiates into the neuronal cell fate.

Figure 11 - neri functions cell autonomously in cell fate decisions. A) anC B) show that supemumerary bristles are genetically homozygous mutant as bristles are never seen within tufts. Arrowhead indicates supemumerary macrochaetae, arrow indicates supemumerary microchaetae. Adult 11eir clones identified usingpwn confimi the cell autonomous nature of rieii function. C) The majority of bristles obsemed at clonal boundruies are mutant (arrow) for neirA'". D) A smaller fraction of bristles at boudaries are wildtype (arrowhead). Also, single bristles that arc ne~r"'~' (arrow) can exist at clonal boundaries (D).

Figure 12 -Expression of net1 during SOP development. Detection of ~ieu using in situ hybridization rkeals expression within SOPs in the third instar imagina1 disc. Ëarly (A) and mid (B) larvai wing discs reveal no expression of neti in proneural r e ~ o n s , and weak expression in emerging~Ops. Late (C) third instar wing disis show expr&sion of mir within emerging SOPs. D) In situ analysis of pupal nota (24 APF) reveals neu expression within the neuron (characterized by dual processes of the axons) of each future SO.

Figure 13 - Expression of lieu in third instar wing imagina1 discs. A cornparison of wing discs probed with sense ne11 RNA (A) and anti-sense neii RNA (B) reveals iteu

expression in SOPs in the notum and future wing rnargin. Higher magnification of the noturn probed with sense rie11 RNA (C) and anti-sense rieu RNA (D) shows that expression is detected only in the SOP.

Figure 14 - Wildtype and myc-tagged neii constructs can rescue the ~ieic"'~' neurogenic phenotype. A) A wildtype embryo stained with anti-HRP at late stages of embryogenesis displays an organized central nervous system characterized by its fascicles. B) A neurogenic ne~r"~' embryo displays severe hyperplasia of the central nervous system. Fascicles do not form and the central nervous system is greatly expanded. A neurogenic nedF6' embryo that contains prc-GAIA driven expression of wildtype rieic (C) or myc- tagged neii (D) has a partially rescued nervous system. The central nervous system is not as expanded and is more organized, with partial fascicles forming.

Figure 15 - Ectopic expression of neic causes adult phenotypes. A) Wildtype macrochaetae assume stereotyped locations on the notum (white amws) . B) Ectopic expression of nerc using A78 causes missing macrochaetae (white arrows) as well as incomplete wing vein formation (D, black arrows). A wildtype wing is shown in C.

Figure 16 - sca-GAIA driven expression of a C-terminal myc-tagged rieu constmct reveals Neu localization at the plasma membrane. Localization of a-spectrin, shown in panel A (green) reveals that the protein is associated with the plasma membrane. B) Expression of myc-tagged Neu (red) in a wing disc proneural cluster reveals localization at the plasma membrane. The superimposed image shown in panel C shows they are co- localized. Myc-tagged Neu protein expressed in salivary glands exhibits the süme plasma membrane localization. E) shows a-spectrin localization (green), D) shows myc-tagged Neu localization (red), and F) shows the superimposed images.

Neu is associated with the plasma membrane

To understand how neu could function in the signaling process that allows for

epidermal ce11 fate determination, we examined the expression pattern of lieu during SOP

determination using in situ hybndization techniques on staged third larval instar wing

imagina1 discs. neic was undetectable in proneural clusters prior to SOP determination

(Fig. 12A, B; Fig. 13). The first detectable mir expression occurs within SOPs in wing

discs of late third larval instars as previously descnbed (Fig. 12C, (Boulianne et al.,

1991)). iieii expression was also examined within the notum at 24 hours after puparium

formation (APF) where its expression was found to be associated with the neuron of each

SO cluster (Fig. 12D). At this stage, al1 the accessory cells of each SO have been

determined and the neuron can be identified based on its shape.

To determine where Neu protein functions within the cell, we then generated

transgenic lines that expressed wildtype or rnyc epitope-tagged neic constructs in the

vector pUAST. To ensure that the rnyc tag did not disrupt Neu function, the ability to

rescue the ~ i e i ~ ' ~ ~ ' mutant allele with both the wildtype and myc-tagged construct was

compared. The ~ i e u ' ~ ~ ' allele produces a severe neurogenic phenotype charactenzed by

hyperplasia of the central and penpheral nervous system (Fig. 14B). and complete lack of

ventral cuticle. Using aptc-GAIA line to dnve expression, we found that both constructs

were equally able to partially rescue the neurogenic phenotype (Fig. 14C. D) and restore

ventral cuticle (data not shown). This indicates that the rnyc epitope does not disrupt the

Neu protein and that the fusion protein is functionally wildtype. In addition to being able

to rescue ~ i e i l ' ~ ~ ~ embryonic phenotypes, we found that ectopic expression of either

tagged or untagged neu constructs yielded identical adult phenotypes characterized by

missing macrochaetae and incomplete wing vein formation (Fig. 15).

The myc-tagged UAS-neri lines were then crossed to a sca-GAIA line that drives

expression of the transgene in proneural clusters in third instar larval wing imagina1 discs.

We found that myc-tagged Neu was primarily localized at the plasma membrane (Fig.

16). Double labeling with an antibody to a-spectrin, a stmctural protein found associated

with the plasma membrane (Dubreuil, et al, 1997; Lee, et al, 1997) confirmed this

localization. Myc-tagged Neu protein expressed in third instar larval salivary glands was

also found localized at the plasma membrane. Since the N signaling pathway is not

active during this stage of salivary gland development, whereas N signaling is active in

the proneural cluster, Neu localization does not appear to be affected by N signaling.

This suggests that Neu functions at the plasma membrane to affect neurogenic signaling.

Discussion

rieii is required cell autonomously for specification of non-neuronal cell fates within the proneural cluster

We have examined the role of the neurogenic gene iieu in sense organ

development in Drosoplfila to gain insight into its function within the neurogenic

signaling pathway. To observe the effects of removing neir during SOP determination,

we generated mutant clones using two ifeu alleles. We find that rieil is required within

the proneural cluster to determine epidermal cell fates; mutant neii clones give rise to

supemumerary SOPs within proneural clusters. This phenotype is very similar to the

loss-of-function N phenotype. However, while iteii appears to be needed for receiving

the signal that allows epidermal cell fate determination, it is not essential for epidermal

development. Support for this conclusion comes primarily from the observation that the

majority of ,~eii*'~' clones included intewening epidermal cells between bristle tufts.

Also, the more severe iieiiiF6' allele proauced bald clones that still contained epidermal

cells. This is in contrast to N, which has been shown to be specifically required for

epidermal cell fdes (Le., strong N alleles fail to support epidermal development (Heitzler

and Simpson, 1991). Therefore, although rieii is required for the determination of non-

neuronal cell fates within the proneural cluster, epidermal development can still occur in

the absence of iicii.

Using a number of genetic markers, we show that ncii functions cell

autonomously during the process of SOP determination. Mutant bristles were found next

to wildtype cells more frequently than wildtype bristles next to mutant cells. As well,

tufts that were generated in neiiAioi mutant clones never contained mixtures of wildtype

and mutant bristles. Taken together, these results indicate that neir mutant cells have

reduced ability to receive or propagate the signal that prevents SOP specification. neil

mutant cells, however, appear to be unaffected in their ability to send this signal.

We also found that neii is involved in the specification of accessory cells. Clones

generated by Dietrich et al. (1984) and in Our experiments using iie~i"~' produced a bald

phenotype. Loss-of-function during the division of the SOP leads to a transformation of

the pIIa daughter cell to the pIIb fate because bald regions contain no extemal SOP

structures. This suggests that the pIIa cell fails to receive the signal from the pIIb cell

that normally prevents it from adopting the same fate. Also, 22CIO staining reveals that

supemumerary neurons develop in ne~c '~~ ' clones, suggesting that the lineage derived

from the pIIIb cell is affected. Specifically, the pIIIb cell produces two neurons,

indicating a M u r e of the sheath cell to rcceive a signal from the neuron. These

phenotypes are much like the phenotypes seen when N is removed during development of

the SOP lineage (Harienstein and Posakony, 1990).

Cell autonomous function and neri expression

Careful examination of the expression pattern of neir, by in situ hybridization,

suggests that expression of neu is restricted to the primary SOP. It is currently difficult to

explain Our expression data given the cell autonomous requirement for neii in inhibiting

neuronal cell fate specification. Clonal analysis reveals that rici< is required for receiving

the signal that allows for suppression of SOP specification yet rrerr expression is

undetectable using in situ hybridization in those cells that are prevented from becoming

SOPs. While it is possible that the supemumerary SOPs that anse in 11cii mutant clones

are a result of abnormal rounds of primary SOP cell division pnor to SO differentiation,

BrdU labeling expenments and anti-phosphohistone H3 staining clearly show that

supemumerary SOPs do not result from extra rounds of mitosis. An altemative

explanation is that wildtype SOPs at the clonal boundary function normally to inhibit

neighbouring cells from becoming neuronal. Thus, tufts could only arise when mutant

SOPs were located away from the clone border. However, Our analysis clearly shows

that the rnajority of bristles at the boundary are mutant and can exist eitheras single

bristles or tufts.

This leaves the possibility that in situ hybridization is not sensitive enough to

detect low levels of neii transcripts within the proneural cluster before primary SOP

determination occurs and that this low level of neir is required for epidermal cell

specification. The high levels of neir observed in SOPs might then be attnbutable to

elevated levels of proneural gene expression observed subsequent to SOP determination.

This would be consistent with previous studies demonstrating that rieii is a target of

proneural gene expression (Hinz et al., 1994; Singson et al., 1994). The ability to

generate two phenotypes with different alleles suggests that the determination of primary

SOPs is more sensitive to levels of rieu activity than is developinent of the SO lineage. In

rie~i"~' clones, in which Neu function is reduced, elevated levels of iieir expression in the

primary SOP might compensate for decreased Neu activity dunng pIIdpIIb and

accessory cell determination. Then, supernumerary SOPs essentially develop normally

and give "se to tufts. However, low levels of n e ~ i ~ ' ~ ' expression dunng primary SOP

determination might be insufficient to compensate for reduced Neu activity. In nedF6'

clones, lack of Neu function cannot be rescued by elevated levels of expression. The

SOP lineage is affected - no pIIa cells are determined and exces neurons are produced.

Neu is associated with the plasma membrane

To gain further insight into the role of Neu in the neurogenic pathway, the

subcellular localization cf Neu was determined. The a:cino acid sequence of Neu

predicts a protein containing a C-terminal RING finger domain that is often found in

DNA binding proteins (Boulianne et al., 1991; Pnce et al., 1993). However, there has

been no evidence to demonstrate that Neu functions in the nucleus. Also, it has been

demonstrated that some RING fingers have functions outside the nucleus (Joazeiro et al.,

1999). Using the GAUIUAS system to express neii dunng embryogenesis, we are able

to partially rescue nedF6' phenotypes. ptc-GALA dnven expression of either wildtype

neu or a myc-tagged rie11 constnict reduces the hyperplasia of the nervous system seen in

nea'F65 mutant embryos. The failure to completely rescue embryos is most likeiy due to

spatial and temporal differences in prc-GAIA expression and normal neli expression.

The ability of myc-tagged Neu to rescue the ~ i e i i ' ~ ~ ~ mutation as well as wild type Neu

indicates that the epitope does not d i s ~ p t the function of the protein. Also, ectopic

expression of either construct during early development yields similar adult phenotypes.

Myc-tagged Neu was found to be closely associated with the plasma membrane. While

this does not exclude the possibility that endogenous Neu, like Notch, may exist at low

levels within the nucleus or elsewhere in the cell, it suggests that neurogenic signaling

does not require nuclear Neu.

The finding that Neu protein associates with the plasma membrane suggests a

possible role in promoting or modulating neurogenic signaling at the receptodligand

level. One possible model is that rieii affects the ability of the cell to receive or propagate

signals by affecting N, and that the function of rieii in the proneural cluster is to promote

differences in the level of N-Dl signaling activity required for mutual inhibition.

According to this model, initial low Ievels of rieir expression within the proneural cluster

would be required to promote differences in neurogenic activity. Through mutual

inhibition mechanisms that involve feedback between the proneural and neurogenic

genes, these differences would then be amplified leading to selection of a single SOP. In

the absence of neit function, the formation of multiple SOPs would be the result of loss in

the ability to receive a N-Dl signal. Expression of iieii would then be upregulated in the

SOP, and rieii would function during the SO lineage in a similar manner to allow cells to

respond to N-Dl signaling. Ectopic expression of iteii allows al1 cells within the field to

signal equally, effectively causing gain-of-function N phenotypes. Interestingly, the

RING finger in Neu shares high homology to the RING finger found in the oncogene c-

cbl (Joazeiro et al., 1999). c-cbl has been shown to have ubiquitin ligase activity and

affects the strength of receptor tyrosine kinase (RTK) signaling activity by targeting

RTKs for degradation. The ubiquitin ligase activity has been shown to be conferred by a

domain encompassing the RiNG finger domain (Joazeiro et al., 1999). Whether the

RMG finger in Neu regulates NID1 signaling by targeting either N or Dl for

ubiquitination, remains to be determined.

Chapter 3 index - Neuralized functions as an E3 ubiquitin ligase during Drosopllila development

3.2 - Materials and Methods

Fly stocks Generation of myc-tagged neu constructs Transgenics and ectopic expression

e Immunocytochemistry and in si tu hybridization Imaging and microscopy Expression of GST fusion proteins

e Ubiquitin ligase activity assay

Localization of Neu is encoded within the NHR's Rescue of neuY6' requires the RING finger domain The RING finger domain confers ubiquitin ligase activity

3.4 - Discussion

This chapter is presented essentially as accepted for publication. Edward Yeh, Matt Dermer, Cosimo Commisso, Lily Zhou, Jane McGlade, and Gabrielle Boulianne. The contribution of the authors is as follows: M. Dermer performed the in vitro ubiquitin ligase assays under the supervision of J. McGlade, L. Zhou and C. Commisso generated the transgenic lines. E. Yeh performed the remainder of the study under the supervision of G.L. Boulianne.

3.1 -Introduction

The Notch pathway is a widely studied means of intercellular signaling

responsible for the detemination of cell fate, cell differentiation and boundary formation

(reviewed in (Artavanis-Tsakonas et al., 1999; Bray, 1998). First identified in

Drosophila r~relanogaster, this pathway is conserved from nematodes Io humans and

plays important roles throughout development (Kidd et al., 1998; Lecourtois and

Schweisguth, 1998; Struhl and Adachi, 1998). A key component of the pathway is the N

receptor which resides at the plasma membrane. Interaction between N and its ligand

(the DSL class of genes) results in trmslocation of the intracellu!ar domain of N to the

nucleus and regulation of downstream genes (Kidd et al., 1998; Lecourtois and

Schweisguth, 1998; Struhl and Adachi, 1998). A series of proteolytic cleavage events are

thought to occur in N both before and after ligand binding that are essential for N

signaling (Blaumueller et al., 1997; Pan and Rubin, 1997; Schroeter et al., 1998; Sotillos

et al., 1997; Ye et al., 1999). Also, regulation by N of downstream genes in the nucleus is

dependent on another gene hppressor of Hairless (Sir(H)) (Fortini and Artavanis-

Tsakonas, 1994) although there are indications of Su(H) independent signaling (Wang et

al., 1997).

Although many aspects of the N signaling pathway have been characterized its

broad role throughout development suggests that several different mechanisms may be

required to regulate its levels and function within cells. Ubiquitination has been shown to

be an important mechanism utilized by the cell to regulate many different events such as

proteolysis, protein degradation, cell cycle control and receptor-mediated signal

transduction (Hershko and Ciechanover, 1998; Levkowitz et al., 1999; Tyers and

Jorgensen, 2000). Ubiquitination has also been shown to regulate the activation of a

membrane bound transcription factor. Specifically, the active form of the yeast SPT23

transcription factor is generated by ubiquitin-mediated processing of a membrane bound

SPT23 precursor (Hoppe et al., 2000). Polyubiquitination of proteins involves a series of

enzymatic reactions involving E l ubiquitin activating enzymes, E2 ubiquitin conjugating

enzymes, and E3 ubiquitin ligases (Hershko and Ciechanover, 1998). Free ubiquitin

molecules are covalently attached and passed from the E l enzyme to the E2 enzyme and

finally to the protein target via the E3 ubiquitin ligase. The E3 ubiquitin ligases are

responsible for confemng target or substrate specificity.

Several recent studies have implicated ubiquitination as a potential regulatory

mechanism involved in N signaling. The molecular characterization of Suppmssor of

deltexSu(&), which has been shown to genetically interact with N, revealed that it

encodes a HECT domain found in proteins that possess E3 ubiquitin ligase activity

(Comell et al., 1999). Although a direct role for Su(dx) in the ubiquitination of N has not

yet been shown, a related mammalian protein called Itch has been shown to interact with

and ubiquitinate N in virro and in human Jurkat cells (Qiu et al., 2000). Further genetic

evidence irnplicating ubiquitination in N signaling comes from observations that

mutations in the 82 and 86 subunits of the 20s proteosome cause N gain-of-function

phenotypes during sense organ development in Drosophila (Schweisguth, 1999). These

mutations also increase the stability of an ectopically expressed activated f o m of N,

suggesting that activated N is targeted to the proteasome for degradation. Taken

together, these data strongly suggest that the N pathway utilizes ubiquitination for proper

signaling during deveiopment.

Here, we show that the function of neitralized (neu), another component of the N

pathway, may also be to target proteins for ubiquitination. tleit was first identified as a

recessive loss-of-function mutation that gives rise to embryonic hypertrophy of the

nervous system similar to what is observed in Notch mutants. Recently we have shown

using mosaic analysis, that t1eu functions cell autonomously to receive or propagate the N

signal (Yeh et al., 2000). However, the biochemical function of Neu was unknown.

Comparison of Neu sequences from Drosophila, C. elegarls and humans reveals the

presence of two neitraliied homology repeats (NHR) and a conserved carboxyl-terminal

C3HC4 type RING finger, suggesting that these regions may be important functional

domains within the protein (Nakamuraet al., 1998). Interestingly, recent findings have

shown that RING finger domains of the type found in Neu may function as E3 ubiquitin

ligases (Lorick et al., 1999) raising the possibility that rleir also functions in the

ubiquitination pathway.

To determine if the RING finger of Neu is required for function and can confer

E3 ubiquitin ligase activity, we have examined the ability of rieit deletion constructs to

rescue neu mutant embryos in vivo, and directly tested the ability of Neu to function as

an E3 ubiquitin ligase in vitro. We iïnd that Neu constructs lacking the RING finger

properly localize to the plasma membrane in nelt mutant embryos but have no capacity to

rescue the nuIl phenotype, demonstrating that the RING finger is required for Neu

function. We also show, using GST fusion proteins, that Neu functions as a ubiquitin

ligase in vitro, and that the RING finger domain confers this activity. Taken together,

these results suggest that Neu may function as an E3 ubiquitin ligase in the N signaling

pathway by targeting components of the pathway for ubiquitination.

3.2 - Materials and Methods

Fly stocks

sca-GAIA was obtained from Y.N. Jan (University of California, San Francisco).

rier<1"* andprc-GALA were obtained from the Bloomington Stock Center (B2004 and

B2017). Al1 fly stocks were maintained at either room temperature or lS°C on standard

cornmeal agar media.

Generation of myc-tagged nerc constructs

To generate an N terminal rnyc tag, the following primers were used to amplify

by PCR approximately 400 bp of the N terminus of neii -

5'GTGGGATCCGGTCTATCGGATATACCA 3' and 5'

GGCCCTGCAGAAGCTCTCAAAGCAACG 3'. This fragment was then subcloned

into the neii cDNA contained in BsSK via Pst1 and BaniHI. The following myc oligos, 5'

GGCCGCGCTGAAATGGAGCAGAAGCTGATCAGCGAGGAGGACCTGAACG3'

and 5' GATCCGTTCAGGTCCTCCTCGCTGATCAGCTTCTGCTCCATITCAGCGC

3' were then ligated to the N terminus of neii via NotI and BamHI. To generate a C

terminal rnyc tag, the following primers were used to amplify by PCR approximately 400

bp of the C terminus of neic - 5' GCGGCCAGGCCAACGGCCACGGTAACCTCC 3'

and 5' CCGGATCCCGTGGTGTAGGTGCGGAT 3'. This fragment was subcloned into

the neu cDNA contained in BsSK via BarnHI and BstEII. The following myc oligos, 5'

GATCCGAGCAGAAGCTGATCAGCGAGGAGGACCTGAACTAGT3'and5'

CTAGACTAGTTCAGGTCCTCCTCGCTGATCAGCTTCTGCTCG 3' were ligated to

the C terminus of neu via Ban~Hl and XbaI. Constmcts were confirmed through

sequencing. Myc tagged neic constructs were then subcloned as a Kpril-Nor1 fragments

into the vector pUAST (Brand and Pemmon, 1993).

To make a Neu protein lacking the RING finger, the pUAST-N terminal myc tag

necc construct was digested with BsrEll and Kpril. The ends were then filled and re-

ligated. This removes the carboxyl-terminal amino acids beginning at 656.

To make Neu protein consisting mainly of the RING finger, BsSK-N terminal myc tag

rieii was digested with SiyZ and re-ligated, essentially deleting amino acids 59 to 427.

This construct was then subcloned into pUAST as a KpriI-NotI fragment.

Transgenics and ectopic expression

Myc-tagged rieii constructs under UAS control were introduced into flies by

standard P-element-mediated transformation (Spradling, 1386). Proneural expression of

neu during development was achieved by crossing lines carrying the various UAS-myc

tag rieu transgenes to ~cu-GAIA lines. Expression of neil for the purpose of nedF6'

rescue was achieved through the following scheme. Second chromosome insertions of

UAS-wild type rieic, UAS-myc-tagged rieic (T6) and UAS-myc-tagged RING finger

domain (T42) were established as [P]/[P]; r i e i c 'F65 /~~3 . An X chromosome insertion of

myc-tagged rieic N terminal region (T23) was established as T23; n e i l F 6 5 / ~ ~ 3 . Males

from these lines were then crossed to prc-GAUIGFP, CyO; 1 1 e i l ' ~ ~ ~ / T M 3 females.

Embryos were collected ovemight on gnpe juice agar plates and aged for an additional

24 hrs.

Immunocytochemistry and in situ hybridization

Immunocytochemistry was performed as described in chapter two. Primary

antibodies were used at the following dilutions: mouse anti-22C10, 1:10 (hybndoma

bank); mouse anti-myc, undiluted (hybridoma bank); rabbit anti-8-spectrin, 1:iOO @.

Brantori). FITC conjugated donkey anti-rabbit and Cy3 conjugated donkey anti-mouse

(Jackson labs) were used at 1:200.

Imaging and microscopy

Confocal images were obtained using a Leica confocal microscope and PC

computer tunning Scanware (Leica). Pseudo colour was added to confocal images using

Adobe Photoshop.

Expression of GST fusion proteins

To generate GST fusion proteins, the following primers were uscd to amplify the

individual rieic fragments from a rieic cDNA template: full length Neu - primer neuN (5'

CCCGGGGGGTCTATCGGATATA 3') and neuC (5'

GTCGACCGTACGTGGTGTAGGT 3'); NeuARiNG -primer NeuN and primer T42C

(5' GGTCGACCGTACATCTGïXCG 3'); NeuRiNG -primer T22N (5'

CCGGGCTCGCACGATATAAAC 3') and primer NeuC.

These fragments were then subcloned into the Topo TA cloning vector (Invitrogen). The

rieu sequence was then isolated and subcloned into the pGEX vector as a SrrrallSall

fragment.

The following primers were then used to create a cysteine to serine mutation at

amino acid 701 in GST-NeuRING, which is a consewed amino acid in the RING finger

domain, using the Quickchange (Stratagene) method:

5'ACCGATïCGAGTGCCGAATCCACCATCTGCTACGAGAATCCCATC3' and

3'TGGCTAAGCTCACGGCT"ïACGTGGTAGACGATGCTCïTAGGGTAGS'.

GST-Neu fusion proteins were expressed and purified using glutathione beads as outlined

in (Harper and Speicher, 1997).

Ubiquitin lignse activity iissay

Hemaglutanin (HA) tagged-UbcH5b (E2) in pT7-7 (a gift from Kazuhiro Iwai)

was produced in the BL21 strain of E. coli following induction with 0.1 mM IPTG

(isopropyl-1-thio-P-D- galactopyranoside) at 3 7 T for 3 - 5 h. Fusion protein was

isolated from bacteria by sonication in PBS (phosphate buffered saline) containing 1%

Triton X-100. The bacterial lysate was cleared by centrifugation and incubated for 1 h at

4OC with a 50% sluny of Glutathione Sepharose (GS) (Amershan Pharmacia Biotech).

The GS was washed 4 times with PBS, quantified and used directly in the ubiquitination

assay. Clarified bacterial lysates were stored in small aliquots at -20°C. The irz vitro

ubiquitination assays were camed out using 0.6-1.0ug of GS bound GST-Neu fusion

proteins, 2uL of UbcHSb(E2) lysate, 500mM yeast E l (Affinity ü W 8545). 2mM ATP,

5uM ubiquitin (Sigma-Aldrich U6253). ImM creatine phosphate (Sigma-Aldrich 27920).

7.5-15 units of creatine phosphokinase (Sigma Aldrich), 50mM tris-HCI (pH 7.4), 2.5mM

MgC12, and 0.5mM D'il". The reaction mixture was incubated at room temperature for

90 minutes and stopped by adding 2X SDS sample buffer. The reaction mixture was

resolved by SDS PAGE and ubiquitinated proteins detected by Western blot analysis

using monoclonal anti-ubiquitin antibody (CHEMICON MAB1510).

3.3 - Results

Localization of Neu is encoded within the NHR's

To determine if the RING finger domain of rieu is functionally important, we

genented transgenic flies that expressed two specific Neu deletion constructs. The first

construct deletes the C :enninus from amino acid 656 thereby removing the RING finger

(T22). A second construct deletes the NHR regions (amino acids 59-427) leaving the

RING finger intact (T42). These constructs, as well as one expressing a wildtype neil

transgene, were myc-tagged and placed under the transcriptional control of the

GALAIUAS system and used to generate transgenic flies (Fig. 17B). As previously

shown, expression of a full-length myc-tagged neii transgene using the scabroiwGAL4

line (which expresses in proneural clusters in third instar wing discs) revealed that full

length myc-tagged Neu (T6) is localized at the plasma membrane (Fig. 18A. (Yeh et al.,

2000)). This plasma membrane localization is also seen in third instar salivary glands

(Yeh et al., 2000). where N signaling has not been shown to be active, suggesting that the

localization of Neu is not influenced by N signaling. Transgenic flies that expressed a

truncated fonn of the protein containing only the N-terminal655 amino acids bu: lacking

the RING finger (T22) were also found to localize the Neu protein to the plasma

membrane (Fig. 18B). In contrast, a Neu construct in which the two NHR regions have

been deleted, leaving the RING finger domain intact (T42), was distributed throughout

the cytoplasm (Fig. 18C). Similar membrane localization of T22 and cytoplasmic

localization of T42 was observed in third instar salivary glands (data not shown)

indicating that this effect is not tissue specific or dependent on Notch signaling.

Figure 17 - Neu constructs used in genetic studies. A) Comparison of Neu from Drosoplzila, C. elegans and human reveal the presence of two NHR repeats and conservation of the RING finger domain. (modified from (Nakamura et al., 1998)). B) myc-tagged ne11 constnicts were generated under the control of the UAS promoter. TG represents full length Neu with a C-terminal myc tag. T22 and T23 contain a delrtion from amino acid 655 that removes the RING finger domain. T42 deletes the two NHR regions, leaving essentially just the C-terminal RING finger domain. T22, T23 and T42 have an N-terminal myc tag.

h-neu

B T6

Figure 18 - Neu localizes to the plasma membrane. sca-GAIA expression of full length myc-tagged Neu, T6, shows protein localization at the plasma membrane in a third instar imagina1 wing disc (A). Similarly, Neu protein that lacks the RING finger, T22, is localized at the plasma mombrane (B). ln contrast the myc-tagged RING finger domain, T42, is found throughout the cytoplasm (13). All panels show anti-myc staining in red, and anti-a-spectrin staining in green.

Taken together, these data suggest that wildtype Neu is localized to the plasma membrane

and that the NHR regions of the protein are required for proper localization.

Rescue of requires the RING finger dumain

The myc-tagged constructs were then tested to determine if they could rescue a

nuIl rietc mutation rie^'^^'). Specifically, we wanted to determine if the RING finger

domain is required for Neu function. Using aparclied-GAL4 line to drive expression

during embryogenesis, we found that both a wildtype nerc constmct and the full-length

myc-tagged rieu construct (T6) were able to partially rescue rieic'"' embryos (Fig. 19).

Whereas nerc mutant embryos are characterized by a hypertrophy of the nervous system,

the nervous system in rescued embryos was less expanded and some fasciculation was

restored (Fig. 19D). The extent of rescue obtained with the myc-tagged construct was

equivalent to that obsewed with a wild type rze~c construct demonstrating that the

presence of the myc tag does not alter the function of the protein (Fig. 19C). The failure

to obtain complete rescue with either construct likely reflects the fact that theptc-GALA

line used to drive expression does not completely mimic rietc expression during

embryogenesis. The N terminal construct (T22), which lacks the RING finger domain,

was unable to rescue rieic mutant embryos despite the fact that the protein was expressed

and localized to the plasma membrane (Fig. 19E). This suggests that the RING finger

domain of Neu is required for activity of the protein. Consistent with this, we found that

expression of the RING finger domain (T42) alone, appeared to have an anti-neurogenic

like phenotype. That is, the nervous system was greatly reduced in neic mutant embryos

rescued with this COnStNCt (Fig. 19F). The observation that the RING finger construct

was mislocalized within the cytoplasm and no longer concentrated at the plasma

membrane suggests that normal Neu function requires that the RING finger be localized

to the plasma membrane.

The RING finger domain confers ubiquitin ligase activity

The m e n t finding that RING fingers may confer E3 ubiquitin ligase activity

suggested that Neu may also function in this manner. To directly test this possibility, the

following GST fusion proteins were made: a full length GST-Neu protein; GST-

NeuARiNG (Neu N terminal region from amino acids 1-423 which lacks the RING finger

domain); and GST-NeuRING (Neu protein from amino acids 631-754 containing the

RING finger). A fourth fusion protein consisting of the RING finger domain with a

cysteine to serine mutation in the absolutely consewed cysteine residue at position 701

was also made (GST-NeuRINGC701S). These fusion proteins wcre then tested in an in

vitro assay that measures the ability of a protein to catalyze the formation of multi-

ubiquitin chains in a reaction containing E l and E2 enzymes, ubiquitin and ATP. The

addition of a protein with E3 ubiquitin ligase activity (as a GST fusion protein) leads to

multiubiquitinization of the GST-E3 fusion protein which can be detected by probing

western blots with anti-ubiquitin (Fig. 20). In this assay, both the full length GST-Neu

(Fig. 20A, lane 2) and the RING finger domain GST-NeuRING (Fig. 20B, lane 2) had E3

ligase activity as revealed by the presence of polyubiquitinated products. Reactions

lacking the essential E2 subunit do not contain polyubiquitinated proteins nor did those

containing GST atone, demonstrating the specificity of the activity conferred by the GST-

Neu fusion proteins. Neither Neu protein lacking the RING finger domain, GST-

NeuARiNG (Fig. 20A, lanefi), nor a mutant Neu RING finger, GST-NeuRINGC701S

(Fig. 20B, lane 5), had E3 ligase activity. Taken together, these results show that Neu

functions as an E3 ubiquitin ligase in vitro, and that this activity requires the RING finger

domain.

Since the C701S mutation in the RING finger abolishes ubiquitin ligase activity in

vitro and previously it was demonstrated that the RING finger is essential for Neu

activity in vivo, transgenic flies were generated canying a UAS-neu C701S construct. It

would be predicted that this construct would lack ubiquitin ligase activity in vivo, and

tiiat this activity is required for Neu function. In rescue experiments similar to the ones

described earlier, the UAS-rceiC701S construct failed to rescue nedF6' embryos (Fig.

19G). This suggests that the RING finger confers ubiquitin ligase activity, and that this

activity is essential for Neu's function in vivo.

3.4 - Discussion

We have used a combination of in vivo and in vitro assays to demonstrate that

Neu function requires the C-terminal RING finger domain and that this domain confers

ubiquitin ligase activity. The conclusion that the RING finger domain is required for

function is supported by the observation that a Neil construct that lacks this domain can

no longer rescue nerr mutant embryos despite the fact that the protein is expressed and

properly localized to the plasma membrane. Interestingly, Neu constructs that retain the

RING finger domain but lack the NHR domüins fail to rescue necc mutant embryos and

appear to behave as gain-of-function mutations, giving rise to an anti-neurogenic like

phenotype. In mutant embryos expressing this construct, the Neu protein is distributed

throughout the cytoplasm suggesting that the RING finger must be localized to the

plasma membrane to rescue ncci activity and that the NHR domains of the protein are

required for this localization. This result is comparable to those obtained in ectopic

Figure 19 - Functional Neu requires the RING finger domain. Using anti-HRP antibodies to reveal the nervous systern of embryos, both wild type neil (C) and full length myc-tagged neic (D) expression by pfc-GAIA can artially rescue the neit'"' P neurogenic phenotype (compare with wild type and ned 6' embryos in A and B). Neu protein that lacks the RING finger domain (T23) is unable to rescue nedF6' mutant embryos (E) while rescue by the RING finger domain (T42) yields a pin-of-function like phenotype (F). In addition, Neu protein containing a C701S mutation which abolishes ubiquitin ligase activity in vitro, fails to rescue neu'"' (G).

Figure 20 - Neu functions as an E3 ubiquitin ligase. GST-Neu fusion proteins were generated and tested for ubiquitin ligase activity. In both panels, there are 9 lanes per blot. Lanes 1.4 and 7 contain the GST-fusion protein only, lanes 2.5 and 8 (+) contain the in vitro reaction including E2 enzyme, and lanes 3.6, and 9 (-) contain the in vitro reaction without the E2 enzyme. A) Full length Neu (GST-Neu) functions as an E3 ubiquitin ligase, as revealed by the presence of ubiquitin positive higher molecular weight products. Neu protein lacking the RING finger (GST-NeuARWG) does not have any enzyrnatic activity. A GST control reveals no enzymatic activity. B) The RING finger domain itself (GST-RING) has E3 ubiquitin ligase activity. A specific mutation in the RING finger dornain (GST-RINGC7OIS) abolishes enzyrnatic activity, indicating that this activity is specific to the RING finger.

B 1 2 3 J 5 G 7 8 9 GST-RING CST-RlNCC7OlS CST

expression studies of various Neu constmcts; ectopic expression of a RING finger only

constmct caused phenotypes suggestive of a neomorph in some N dependent

developmental contexts (Lai and Rubin, 2001; Lai and Rubin, 2001).

Using a biochemical assay developed by Lorick et al (Lonck et al., 1999). we

have shown that Neu can catalpe the formation of multiubiquitin chains in an E2-

dependent manner demonstrating that in vitro, Neu functions as an E3 ubiquitin ligase or

as part of an E3 complex. Full length Neu, as well as the RING finger domain alone,

caused ubiquitination of proteins in the E3 assay; NeuARiNG and neuRiNGC701S had

no activity. In these assays GST fusion proteins were used and ubiquitination is believed

to occur on the GST moiety. Since Neu constmcts that delete the RING finger, or

contain a mutated RING finger, lack ubiquitin ligase activity this clearly indicates that the

Neu RING finger is essential for the ubiquitin ligase activity in this assay. Taken

together with the rescue experiments, this implies that Neu may function as an E3

ubiquitin ligase in vivo and provides evidence that ubiquitination plays a role in cell fate

determination.

While there is increasing evidence that RING fingers can directly bind to E2

conjugating enzymes, not al1 RING fingers function in this manner (Freemont, 2000).

Some RING finger containing proteins interact with E2 enzymes directly while others do

so as part of a multiprotein complex. Similar types of interactions can also occur

between the E3-ubiquitin ligases and their substrates. At present, the identity of Neu

target proteins and the E2 enzyme that interacts with Neu is unknown. Based on the

expression of various Neu deletion consînicts, it has been suggested that the NHR

domains may function as protein interaction domains and that Neu may function as part

of a multiprotein complex; however, this has not been formally demonstrated (Lai and

Rubin, 2001; Lai and Rubin, 2001). Our results suggest that the NHR domains are

required for Neu localization, and thus it may be that Neu localization requires protein

interaction with its target(s) andlor complex. These lines ûf evidence are suggestive that

Neu functions as part of a multimenc complex and additional studies will be required to

identify the critical components of this complex.

The effect of ubiquitination of target proteins in vivo by Neu is also unknown.

Several studies have previously shown that ubiquitination of proteins by E3-ubiquitin

ligases can lead to protein degradation, targeting to various subcellular locations as well

as receptor-mediated endocytosis. We have previously shown that Neu, like Notch, is

required cell autonomously to specify non-neuronal cell fates during sense organ

development (Yeh et al., 2000). Taken together with Our data demonstrating that Neu is

localized to the plasma membrane raises the possibility that the function of Neu is to

target plasma membrane components of the N signaling pathway for ubiquitination.

Since genetic data suggest that neu positively propagates N signals, we do not think that

Neu targets N protein for ubiquitin-mediated degradation. Rather Neu may play a role in

N receptor activation (perhaps through a proteolytic event) or relieve inhibition of the N

pathway by targeting an inhibitor for degradation. Several studies have recently shown

that both ubiquitination and endocytosis are important in regulating the N signaling

pathway. Mutations in the 82 and 86 subunits of the 20s proteasome cause N gain-of-

function phenotypes (Schweisguth, 1999). Several studies have also demonstrated that

endocytosis of both the N receptor and its ligand, Dl, is also an important mechanism by

which signaling is controlled during development (Klueg and Muskavitch, 1999; Parks et

I l l

al., 2000; Seugnet et al., 1997). Clearly, ubiquitination can be used as a signal for many

cellular events and identifying which components of the N signaling pathway are targeted

by Neu will aid in understanding how N signals propagate within the cell.

Chapter 4 - Summary and Future Directions

The goal of the studies described here was to provide a better understanding of the

role of neid throughout Drosopliila development. The analysis of ne11 mutant clones

generated during sense organ determination and differentiation suggests that ne11

functions in the N signaling pathway to propagate or transduce signal. Consistent wirh

this finding, neii was demonstrated to be required for proper development of the eye and

wing, in a manner similar to the known requirements of N signaling. Through genetic

and biochemical methods, the function of Neu was determined to be a ubiquitin ligase.

The ubiquitin ligase activity is conferred by a carboxyl-terminal RING finger which is

necessary for Neu function in vivo. Together, these results indicate that the role of neil

within the N signaling pathway appears to be to target proteins for ubiquitination and that

this is required for transduction of the N signal.

The involvement of rieu during SOP development is not surprising due to the

expression of neii transcripts in 1'' SOPs, and the known involvement of the N signaling

pathway during SOP determination and differentiation. However, the finding that lieu

acts cell autonomously is surprising, given the phenotype manifested in mutant clones.

Mutant neu clones gave rise to a neurogenic phenotype - bristle tufting due to

supernumerary 1' SOP determination when using the hypomorphic neitA"' allele, and

cuticle balding due to complete neurogenesis in the SOP lineage when using nedF6'

allele. Analysis of the phenotypes of bristles formed at the clonal boundaries revealed

that neit acts cell autonomously for epidermal determination even though expression of

np:i is not detected in cells within the proneural cluster. A model in which

supernumerary SOP formation was attributed to aberrant mitotic events prior to SOP

differentiation was rejected based on the lack of expression of cell mitotic markers in

SOPs within rie11 mutant clones. Thus, the cell autonomous ability of ricic to determine

epidermal cell fates most likely stems from low level rieu expression within epidermal

cells, which is undetectable by iri situ hybridization. This is not an unprecedented

explanation given that the N " ~ is thought to activate transcription of downstream genes

even though the protein cannot be detected in the nucleus. A recent study has found

similar results regarding neu involvement in sense organ development, and likewise

concluded that low levels of undetectable Neu within epidemal cells may be required for

N signaling within the proneural cluster (Lai and Rubin, 2001).

The role of rie11 to positively transduce the N signal was based on the analysis of

rieu mutant clones during sense organ development and was confirmed by the observation

of a role for net1 in other tissues known to require N signaling. Mutant rie11 clones

produced eye defects characterizcd by improper determination of photoreceptor cells in

the ommatidia, as well as improper determination of accessory cells such as cone cells

and interommatidial SOPs. Furthemore, defects in ommatidial polarity were also

observed. Failure to transduce the N signal during eye development can produce al1 these

phenotypes. While the finding that rieu mutant clones cause developmental defects in the

eye is preliminary (the requirernent of N signaling during eye development is complicated

both temporally and spatially), the phenotypes seen are consistent with a failure to

transduce the N signal.

During wing development, defects were observed in rieu mutant clones including

improper SOP determination dong the wing margin and ectopic wing vein formation.

Again, these phenotypes are consistent with a failure to transduce the N signal. However,

generation of ileu mutant clones during oogenesis failed to give rise to a consistent

phenotype. The role of N signaling during oogenesis is quite clear - N signaling is

required for specific follicle cell determination and for polarity determination in the

maturing oocyte. Yet phenotypes associated with a failure to transduce the N signal were

not seen on a consistent basis in egg chambers in which follicular neii clones were

generated. This may indicate that the role of neil during N signaling in oogenesis is

redundant. Such a possibility is consistent with a model in which some cornponents in

the N signaling pathway are used universally, while others are used in a context-specific

manner. Quite possibly, neri may function during N signaling in some cell determination

contexts, but rnay be redundant or unnecessary for signaling in other developmental

contexts. For example, while some defects associated with loss of N signaling were

observed in the wing, other defects such as those associated with N's involvement in

dorsal-ventral boundary formation were not seen. As seen in these studies, lieu mutant

clones appear to drastically affect neurogenesis in various tissues while non-neural events

are not affected or affected to a lesser extent.

The finding that neii functions as a ubiquitin ligase provides insight into how neu

functions to transduce the N signal and provides the potential to investigate how

ubiquitination is used to regulate the N signaling pathway. As discussed previously,

ubiquitination has been implicated in N signaling, but there is no direct evidence

dernonstrating an in vivo role for ubiquitination in the N pathway. Molecular

characterization of the gene Su(&) predicts that it may function as an E3 ubiquitin ligase

yet there is no direct evidence to support this hypothesis. Charactenzation of the

mammalian gene itcli and demonstration of its ability to bind and ubiquitinate N were

done in vitro but then: is no data confirming such a function in vivo. The data

demonstrating that proteasome subunit P2 and 06 mutations mimic N gain-of-function

phenotypes are suggestive of a role for ubiquitin-mediated protein degradation in N

signaling; in addition such mutations stabilize ectopically expressed N ' ' ~ protein.

However, the phenotypes obtained could be due to indirect results of global defects in

protein degradation, and not specifically due to a failure in regulating or degrading N

itself. The results presented here provide strong evidence that ubiquitination is required

in vivo for transducing N signal. The combination of the genetic rescue experiments with

the biochemical finding that rleri has ubiquitin ligase activity in vitro provide persuasive

evidence for the role of ubiquitination in vivo. Rescue constructs that lack the RING

finger domain fail to restore normal neit function; biochemically, the RING finger

behaves as a ubiquitin ligase. The ability of Neu with a mutated RING finger (C701S) to

rescue the ~ l e i r ' ~ ~ ~ mutant is currently being tested and should provide further evidence of

the role of iieii as a ubiquitin ligase during N signaling.

It remains to be demonstrated which component of the N signaling pathway Neu

actually targets for ubiquitination. The most convincing data will come in the form of

protein targets of neri ubiquitination and the demonstration that such targets are of

functional significance to N signaling. The most attractive hypothesis is that neii targets

the N receptor itself, possibly to signal receptor intemalization. Since rierr functions

positively to propagate N signal, a role for rieri in N degradation would only seem

plausible in a recycling context. It has been suggested that N signaling can only persist

by the removal of DI-N complexes from the plasma membrane and insertion of new,

uncomplexed Dl and N proteins. The finding that neii protein localizes to the plasma

membrane makes any of the identified membrane-associated N signaling components

possible targets.

An alternative hypothesis for rieu involvement in N signaling may be that it

regulates Dl protein. Although mutant rieu clones did not give rise to non-cell

autonomous defects in sense organ development, it remains possible that Neu regulates a

cell autonomous function of DI. It has been demonstrated that one such cell autonomous

effect of Dl is to alter the cell's ability to receive N signaling; high levels of Dl

expression reduce the cell's ability to respond to N. Thus by altering the levels of DI, it is

possible to affect the ability of the ce11 to respond to N. It is possible that Neu targets Dl

for degradation, thus reducing the levels of DI and increasing the ability of a cell to

respond to N. The consequence is that these altered abilities to respond to N provide a

mechanism for unequal or asymmetric signaling between neighbouring cells.

The reagents and results presented here provide a means to begin identifying

target proteins of Neu ubiquitination. The identification of targets can be approached in

various ways. One method is to assay candidate proteins. Due to the localization of Neu,

sevenl known components of the N pathway are possible candidates for ubiquitination,

including DI and Fi. The in vitro assay used to determine Neu's ubiquitin ligase activity

can be modified and used as an assay to test whether candidate proteins are actual targets

of ubiquitination. Addition of candidate proteins to the ubiquitin ligase assay has been

used to demonstrate target specificity of other E3 ubiquitin ligases in vitro. While this

assay may confirm potential targets of Neu, not al1 interactions with Neu are likely to

result in ubiquitination. Identification of these proteins is still valuable as they will

provide insight into the protein interactions required for Neu function. A method to

identify proteins that bind to Neu is the yeast two hybrid screen. Yeast two-hybrid

constructs have been genented and tested for suitability; such analyses may lead to the

identification of Neu binding proteins. Identified proteins could subsequently be tested

as potential targets of ubiquitination using the modified in vitro ubiquitin ligase assay.

While the proposed candidatelyeast Iwo hybrid approach may certainly yield

information about possible target proteins of Neu ubiquitination, consideration should be

given to the manner in which Neu functions as a ubiquitin ligase. As described earlier,

ubiquitin ligases either function within multiprotein complexes or are capable of

functioning independently of other factors. Typically, those that are able to function

individually contain amino acid motifs for protein binding, such as SH2 domains, F

boxes, WW domains, etc. There are, however, no such obvious protein binding motifs in

the Neu amino acid sequence. Some evidence suggests that the NHR domains may

function as protein interaction domains but this has not been formally demonstrated.

While not excluding the prospect that Neu can function independent of other factors, the

possibility remains that Neu may function within a multiprotein complex. If such is the

case, then a candidate or yeast two hybrid approach may not identify the targets of

ubiquitination.

In contrast, the various GST-Neu fusion constructs generated in the studies

described here can be used in pull down assays to identify Neu protein complexes. In

addition to the GST moiety, a histidine tag has also been incorporated into the various

Neu fusion constructs, making the constructs ideal for pull down assays using Drosophila

tissue lysates. Due to the nature of these assays, proteins that bind to Neu do so in the

presence of any factors that are required, and thus protein complexes containing Neu

could potentially be isolated. A common method to identify individual components

within these complexes is to separate and isolate the individual proteins through gel

electrophoresis. Isolated proteins could then be identified through a vanety of methods

including mass spectrometry. Thus, without knowing what individual proteins Neu can

bind, it is possible to identify targets of Neu ubiquitination, cofactors for Neu

ubiquitination, and potentially the UbcE2 required for ubiquitination.

Knowledge of Neu targets will provide insight into regulation of the N signaling

pathway by ubiquitination. The exact nature of the identified targets will dictate the best

approach for determining the functional significance of interactions with Neu, but

ultimately the in vivo significance of ubiquitination by Neu mus1 be addressed and

confirmed. As mentioned previously, several lines of evidence have implicated

ubiquitination in regulating N signaling. However, there has yet to be a correlation

between ubiquitination and any of the events required for N signaling. The best example

of a requirement for ubiquitination has been the observation that Itch can ubiquitinate N;

however, significance for N signaling has not been demonstrated. With the tools

generated for the studies descnbed here, the knowledge of Neu targets will allow for

investigation of how ubiquitination controls a specific molecular event in N signaling.

For example, if N turns out to be a target of Neu ubiquitination, the hypothesis that neii

signals N endocytosis could be tested directly. The staining of mutant neil tissues wiih

antibodies to N~~~ and N'" could reveal whether this is the case; one would expect to

see defects in N~~~ transendocytosis, or N " ~ endocytosis as observed in the developing

pupal eye. If ne i~ actually signals one of the N cleavage events, then assaying changes in

the electrophoretic profile of N in flies lacking neil function as well as flies

overexpressing neir may reveal such a role. In addition. quantitative compansons of the

levels of N protein within these expenmental contexts may reveal whether riea's role is in

regulating N degradation.

Besides being used to identify protein targets of ubiquitination, these reagents and

approaches could be used to identify other factors involved in regulating neic function.

For example, the NHR domains appear to function in localizing Neu at the plasma

membrane. These domains are dispensable for ubiquitin ligase function in vitro, but

clearly are required for endogenous function in vivo (Fig. 25F). This suggests that

plasma membrane localization of Neu may be required for interaction with its target

protein, and that mislocalized Neu may promiscuously associate with other proteins.

Identifying proteins that bind to the NHR domains may provide information about how

Neu becomes localized to the plasma membrane, and how it associates with its target

proteins. The constructs used in the studies descnbed here provide the starting tools to

identifying such proteins.

The identification of Neu binding proteins will also provide insight into how N

signaling can be differentially interpreted within different developmental situations. The

results descnbed in Appendix A suggest that neic functions in N signaling in only a subset

of al1 possible contexts. Elucidation of factors that function in concert with neii to

regulate N signaling may lead to the identification of context specific factors. 1 his will

ultimately lead to a better understanding of N signaling throughout development.

Appendii A index - nen functions throughout Drosophila development

A . l - Introduction

A.2 - Materials and Methods Fly stocks and transgenics Heat shock protocol lmmunocytochemistry and DAPI staining Generation of nerc mutant clones Mounting of adult wings Eye sections and scanning electron microscopy

e rieuralized functions during eye development 0 nerrralized is involved in wing development

The involvement of neuralized during oogenesis

A.4 - Discussion

A . l - Introduction

A role for nerr during sense organ development was previously demonstrated

through the generation of mosaic clones (Yeh et al., 2000). Specifically, mosaic analysis

revealed that 1iei1 functions cell autonomously to promote or propagate N signaling. To

provide further support for neic's role in N signaling, mutant tieri clones were generated in

additional tissues known to be dependent on N signaling. If iieir is a conserved and non-

redundant component of N signaling, then mutant n e ~ clones should manifest N-like

phenotypes in al1 tissues requiring N signaling.

As mentioned earlier. N signaling is important for the development of many

tissues. During eye development, N signaling is responsible for photoreceptor and

accessory cell determination, control of tissue growth, establishment of tissue patteming

and ommatidial polarity. Similarly, N controls ceIl fate determination, tissue patteming,

and regulation of growth during the development of other tissues such as the wing and

the ovarian follicle. Based on the results obtained during sense organ development, it is

expected that loss of nerr function would also give rise to N like phenotypes during the

development of these other tissues.

nerc expression is detected during development at times when N signaling is

known to occur. In addition to the expression and involvement of neil in the wing disc

and notum during sense organ development described earlier, aeir is expressed in the

developing eye. In imagina1 eye discs, neu expression can be seen in clusters of cells

posterior to the morphogenetic furrow (Fig. 21). In pupal eyes, neic expression can also

Figure 21 -Expression of rierc during the eye development. A) In situ hybridization reveals ttac expression in clusters of cells behind the morphogenetic furrow (mf) in eye imagina1 discs. B) During pupal eye development, rieri expression is observed in cone cells.

be detected in cone cells (Fig. 21). N signaling is required in these cells for proper cell

determination and patterning. During oogenesis, iieit expression has been described in

polar follicle cells (Ruohola et al., 1991) and a B-galactosidase enhancer trap into the iien

locus causes expression of the reporter in polar follicle cells (Fig. 25).

In this section, a description of the phenotypes caused by mutant rien clones in the

development of the eye, wing, and egg chambers will be presented. During eye

development, mutant i ~ e i i ~ ' ~ ' clones cause defects in photoreceptor determination and

ommatidial patterning. In addition, cone cell and sense organ bristle determination are

affected. These phenotypes are consistent with a role for rieil in promoting or

propagating N signaling during eye developmcnt. In the developing wing, rieir mutant

clones cause defects in sense organ development along the wing margin, and affect vein

and intewein determination. However, defects in establishing dorsal-ventral boundaries

as predicted by the involvement of N. Dl, and Ser signaling, werc not observed. In

addition, neir mutant clones did not produce consistent phenotypes during oogenesis,

suggesting that rieu may be redundant or non-functional in the ovary. This suggests that

nerr may not be universally utilized for al1 N signaling events.

A.2 - Materials and Methods

Fly stocks and transgenics

Fly stocks were kindly provided by the following individuals: e22c-GAIA UAS-

FIP 1 CyO; FRT82B. RM (D. Harrison, U of Kentucky); y w lisfp; FRT82B. y+nM

(T. Xu, Yale). FRT82BxM was obtained from the Bloomington Stock Center (B2004).

The allele rie$'o' is a hypomorphic allele caused by a lacZ enhancer trap insertion into

the rieii gene (Bellen et al., 1989) and is available from the Bloomington Stock Center

1s an amorphic allele (de la Concha, et al., 1988; Bnnd and Campos- (B4369). "eurF6' '

Ortega, 1988) and was provided by Y.N. Jan. Ail fly stocks were maintained at either

room temperature \Ir 18OC on standard com meal agar media.

To generate transgenic flies carrying a heat shock inducible tieri construct, a 3.2

kb full length neii cDNA was subcloned as a Kpril fragment into the vector pWHI, which

places tieu under the transcriptional control of an hsp70 promoter. This construct was

introduced into flies by standard P-element-mediated transformation (Spradling, 1986).

Heat shock protocol

Adult females in the presence of males were placed in a 35°C oven for 1 hour,

twice a day. After three days of heat shocking, ovaries were dissected acd ;?rocessed for

immunocytochemistry and DAPI staining.

Immunocytochemistry and DAPI staining

Imagina1 discs and pupal tissues were dissected and treated as previously

descnbed (see chapter two). Pnmary antibodies were used at the following dilutions:

mouse anti-cut, 1:100; mouse anti-fasciclin III, 1: 10; mouse anti-myc, undiluted; rabbit

anti-P-galactosidase, 1:2000. HRP conjugated anti-mouse and anti-rabbit secondaries

(BioRad) were used at 1:200. FITC-conjugated donkey anti-rabbit and Cy3-conjugated

donkey anti-mouse (Jackson labs) were used at 1:200. FITC conjugated sheep anti-HRP

(ICN) was used at I:200. Samples stained with DAPI were treated as previously

descnbed (Verheyen and Cooley, 1994). In situ hybridization was performed as

previously descnbed (see chapter two).

Generation of aeu mutant clones

Mutant neii clones were generated during eye and wing development as

previously described (see chapter two). Clones during oogenesis were genented by

expressing a UAS-FLP construct under the control of e22cGAL4 (which expresses in the

follicle stem cells). Clones were identified in the ovary by the expression of tke nMyc

marker located on the FRT82B third chromosome. To induce expression or the marker,

flies were placed in a 37°C water bath for 1 hour. After recovery at room temperature for

30 minutes, ovaries were dissected and processed for immunocytochemistry.

Mounting of adult wings

Adult wings were dissected and mounted in Gary's magic mountant. Mounted

wings were then baked at 60°C ovemight and viewed using a Nikon Optiphot 2

microscope using light microscopy. Images were captured using a Sony CCD camera

and Northem Image software.

Eye sections and scanning electron microscopy

Adult eye sectioning and staining was perfonned as previously described in

chapter two.

A.3 - Results

neuralized functions during eye development

Mutant t~ert'O' clones were generated to detennine the role of neii during eye

development. Scanning electron micrographs of mutant tissue in adult eyes revealed

defects in ommatidial shape and size, mislocalized or misnumbered interommatidial

bristles, and tissue scarring (Fig. 22B.C). To characterize the phenotype at the cellular

level. mutant clones were induced dunng first instar l a ~ a l development, and developing

pupal eyes were dissected and stained with an a-cut antibody 24 hours after pupanum

formation (APF). At this stage of eye development, Cut is expressed in the four cone

cells associated with each developing ommatidium as well as the bristle sense organ

precursors (Fig. 23A, C). Eyes dissected from pupae 24 hrs APF, in which clones had

been induced and stained with a-cut, revealed defects in both cone cell and bristle SOP

determination. Specifically, improper numbers of cone cells and SOPs per ommatidium

were formed. As well, the organized pattern of cone cells and latticework pattern of

SOPs was dismpted (Fig. 23B, D)

It is known that N - DI signaling is required for photoreceptor specification, and

thus it is possible that neic mutant clones may also affect photoreceptor development. To

determine if this is the case, adult eyes containing iieu mutant clones were sectioned, and

photoreceptor development was examined. Clones were made using a ivliire+ minigene

in a ivl~ire- background which enabled the unambiguous identification of the clone.

Sections through a wild type adult eye reveal seven photoreceptor rhabdorneres per

ommatidium that have a characteristic position and orientation. Only seven

photoreceptor rhabdomeres are visible in any one section because photoreceptors seven

and eight are positioned above one another. Six photoreceptors surround the seventh

photoreceptor in a characteristic chiral pattern. Sections through eyes containing rieu

mutant clones revealed that photoreceptor development is affected. These defects range

from abnormal photoreceptor specification (Fig. 23E) to severe dismptions of ommatidia

formation and organization (Fig. 23F). In addition, some ommatidia display a

symmetncal polarity indicative of defects in R3R4 specification (Fig. 23F). Taken

Figure 22 - iicicNol mutant clones affect eye development. A) A wild type eye displays uniformly sized ommatidia and interommatidial bristle positioning at altemate vertices. B) Scanning electron micrographs of eyes containing nedi'' clones reveal defects in ommatidia size and shape. In addition, interommatidial bristle location and number are affected. C) In severe cases, scaning of the eye is observed.

Figure 23 - Wild type pupal eyes stained with a cut antibody reveal the interommatidial bristle SOPs (A), as well as the four cone cells of each ommatidium (C). In pupal eyes in which ne~c"'~' clones were generated, defects in interommatidial SOP location and number are observed (B - arrows). Also, cone cell specification appears to be affected (D - arrowheads). Stained sections through adult eyes containing Ileu mutant clones reveal defects in photoreceptor formation, and ommatidial polarity (E). The ommatidium marked by the arrow displays an extra photoreceptor. The ommatidium marked by the arrowhead displays an extra photoreceptor, as well as an apparent defect in R3/R4 specification. This ornmatidium displays an orientation defect (red bars mark the orientation plane). An unstained eye section through a mutant clone as revealed by the lack of w+ expression reveals severe defects in photoreceptor determination and ommatidia organization (F).

Figure 24 - ~ted"~' mutant clones affect wing development. Bristle sense organs dong the .wing margin forin in a double row, establishing a uniform pattern (A). The dorsal rci..., consists of repetitive bristles, whereas the ventral row forms a single bristle ap~roximately every 5 dorsal bristles. In a t~eic"'~' mutant wing, defects in bristle formation affect this uniform pattern (B). Mild ectopic wing vein formation was also seen in ttetc"" wings taken from adults with mutant clones (C - arrow). Occasionally, an ectopic sense organ-like bristle was seen to form in the vein tissue (D - arrowhead).

together, these results suggest that neii functions during the specification of photoreceptor

cells and in the determination of ommatidia polarity, similar to N and Dl.

neuralized is involved in wing development

In addition to bristle and eye defects, mutant rieiiA'" clones also revealed that m i r

is involved in wing development. To examine these effects more closely, wings from

adults in which rte~i"'~' mutant clones had been induced were mounted and analyzed

using light microscopy. Wild type wings display acharacteristic wing vein and intervein

pattern. Also, sense orgaii bristles dong the anterior wing margin form a patterned

double row of small and large sense organ bristles (Fig. 24A). Wings taken from adults

containing mutant clones exhibited signs of abnormal bristle patterning and formation

dong the wing margin (Fig. 24B). In addition, some mild ectopic wing vein formation

was seen although ectopic vein formation occurred only at the L3-LA crossvein (Fig.

24C). Occasionally, an ectopic sense organ-like bristle was observed to form in the vein

region (Fig. 24D). Generation of mutant clones using the nuIl i ~ e i i ' ~ ~ ~ allele yielded

similar phenotypes (data not shown).

The involvement of neuralized during oogenesis

The expression of ne11 in polar follicle cells suggests that rieii may play a role

during oogenesis. To examine this possibility, transgenic nies were generated carrying

an hsp70-neil cDNA constmct. The effects of overexpressing neu during oogenesis were

then observed by heat shocking female adults. Stmctural defects were observed in many

of the egg chambers suggesting that overexpression of neil is affecting oogenesis. Egg

chambers containing more than 16 germline nuclei were observed as well as egg

chambers containing too few nuclei (Fig. 25). This phenotype is characteristic of the

loss-of-function N phenotype in early oogenesis, in which polar cells are overspecified at

the expense of stalk cells. To address whether the same defect was occumng in neii

overexpressing ovaries, ovaries were stained with a fasciclin III antibody, which marks

polar follicle cells in stage 3 or later egg chambers. In some egg chambers, the

expression of fasciclin III was no! limited to the polar follicles but instead was observed

in a greater numbers of cells (Fig. 25). However, the cells ectopically expressing

fasciclin III did not look morphologically like polar follicle cells. The effects of

ectopically expressing neii were also analyzed using another marker of polar follicle

cells, the ~ i e i ~ ' ~ ' enhancer trap line. ~ieii"~' contains a B-galactosidase enhancer trap

which marks the polar follicle cells. To utilize this marker, the neii'"O1 allele was crossed

into the neu overexpressing flies. XGAL staining revealed that some egg chambers

contained extra polar follicle cells (Fig. 25). These results suggest that neir expression

can induce markers of polar follicle ce11 differentiation.

Next, the involvement of neii during oogenesis was analyzed by generating

mutant neuA'" clones in adult fernales. A GALA line that drives FLP expression in

follicle ce11 precursors, e22cGALA; UAS-FLP, was used to generate mutant clones early

in oogenesis. Surprisingly, the structure of developing egg chambers appeared to be

unaffected in neii mutant clones. In almost al1 cases, egg chambers appeared normal in

structure. Antibody staining for P - galactosidase expression from the Al01 allele which

marks the polar follicle ceIl fate also showed that neri was not required for polar follicle

ce11 determination (Fig. 26).

Figure 25 - Overexpression of tieu during oogenesis causes N loss of function-like phenotypes. A) A wild type control ovariole stained with DAPI reveals normal egg chamber development. B) Overexpression of neil causes egg chambers to develop with too many (arrow) or too few (arrowheads) nurse cells. C) A wild type ovariole shows fasciclin III expression in polar follicle cells of egg chambers older than stage 3. D) Ectopic expression of neil during oogenesis can induce ectopic expression of fasciclin III in other follicle cells (arrow). However, these cells do not morphologically resemble polar follicle cells. E) Expression of neuAf0' is detected in polar follicle cells in heterozygote fernale ovaries. Only two polar follicle cells are formed at each egg chamber terminal end (inset). F) Ectopic expression of rieu can cause an increase number of polar follicle cells as detected by ~teic"'~' expression.

Since ~ieii"'~' is a hypomorphic allele, it is possible that sufficient activity remains

to function during oogenesis. To address this, mutant clones were generated using the

nuIl allele ~ r e i c ' ~ . Defects in egg chamber development were observed (Fig. 26) but not

at the rate expected from the frequency and size of clones generated. These defects

included egg chambers containing too few or too many nurse cell. e22cGAL4; UAS-FLP

consistently produces clones in approximately 50% of egg chambers and these clones can

cover up to half or more of the follicle cells per egg chamber (data not shown). Defects

were observed at a rate far less than this, suggesting that neii may have a redundant or

non-functional role during oogenesis.

A.4 - Discussion

In this chapter, the function of neii during the development of tissues known to

require N signaling was characterized. The involvement of rteii during N signaling

predicts that loss-of-function rieu should produce phenotypes in tissue development

similar to those observed in N mutants. In fact, neii mutant clones did produce

phenotypes during development that are consistent with its role in promoting or

transducing N signaling. However, neu appears to function in some but not al1 of the N

signaling events that have been described throughout development.

In the eye, neri appears to function to transduce N signaling dunng photoreceptor

and accessory cell specification. The defects observed in photoreceptor specification and

ommatidial polarity are characteristic of N loss-of-function phenotypes. N signaling is

required for R8 photoreceptor determination, which in turn affects the specification of the

remaining photoreceptors. The phenotypes detected in ne~i"'~' mutant clones indicate

Figure 26 - ~ i e i < ' ~ ~ ' mutant clones generated during oogenesis cause some defects in egg chamber development. A, B) neilA'o' mutant clones do not appear to affect the development of egg chambers. and polar follicle cell specification is normal. Mutant clones are identified by the lack of nMyc expression (red). Polar follicle cells are identified by rieuA'O' expression (green). Inset shows a higher magnification of the polar follicle cells marked by the arrowheads. Mutant clones of the nuIl allele r~eic'~~' cause some egg chamber separation defects as revealed by egg chambers containing abnormal numbers of germline nuclei. C) A wild type ovariole stained with DAPI reveals the normal structure of a developing egg chamber. Defects detected in ovaries containing ~iei<'~~''clones include egg chambers containing too few nuclei (D) and egg chambers containing too many nuclei (E - arrow).

a failure to transduce the N signal. Defects in photoreceptor determination, including

specification of the R3R4 fates, were also seen. In larger nei~*'~' clones, a complete

dismption of ommatidial organization was observed. A role for rteic in the specification

of ommatidia orientation w5s also revealed. These phenotypes are consistent with a loss

of N signaling. Later defects in cone cell and bristle sense organ specification seen in

ne$"' clones are also consistent with a loss of N function. However, it is not clear

whether these defects are due to loss of rteii function during specification of cone cclls

and bristle sense organs, or due to the earlier fiiilure of photoreceptor specification.

Further analysis whereby clones are induced at specific times during eye development

may address this.

During wing development, neu clearly functions in the specification of bristle

sense organs dong the wing margin. This is not unexpected due to the results seen

during microchaetae and macrochaetae development. However, it is not clear whether

neic also participates during the determination of the wing vein and intervein regions.

Ectopic wing vein tissue was seen, consistent with a loss of N-Dl signaling; yet in every

case, the ectopic vein phenotype was very mild, and was only observed at the L3-LA

intervein. Use of the nuIl allele, nedF6', suggests that the rote of neii during N-Dl

specification of vein and intervein regions may be redundant. It does appear that rteit is

redundant or does not function during N-Dl determination of the dorsal-ventral boundary,

as no wings displaying defects in dorsal-ventral boundary formation were observed.

The function of rteu during N signaling in oogenesis is unclear. Overexpression

of rwic yielded phenotypes similar to those associated with a loss of N signaling. This is

somewhat surprising given the overwhelming data suggesting a role for netc in

transducing N signal. However, ectopic expression data can be unpredictable and more

difficult to interpret since increased gene expression occurs both within and outside the

normal spatial domains of endogenous gene expression. The observation that mutant iieii

clones cause defects in oogenesis is more easily interpreted, and consistent with riett's

role in transducing the N signal. Egg chambers containing too many or too few germline

nuclei were detected. However, the frequency at which these phenotypes were observed

was well below the expected frequency of clones generated. The e22cGAL4; UAS-FLP

line used to induce clones expresses 'FLP recombinase in the follicle cell precursors and

cm efficiently generate clones in up to 50% of egg chambers. Most of the mature egg

chambers contained very large clones including some covering entire egg chambers.

While some defects were detected using r~eir"'~', most egg chambers exhibited normal

polar follicle cell specification when clones overlapped the terminal ends. The frequency

of mutant phenotypes was low for both i~eir"'~' and itedF6', suggesting that rteu's

involvement in N signaling during oogensis is redundant or non-functional.

It remains possible that the role of rteu in oogenesis is temporally regulated, and

that creating permanent mutant clones masks its role in N signaling. It has been

suggested that during oogenesis the role of N signaling is to keep populations of follicle

cells in an undifferentiated state, allowing for cell determination to occur in an orderly

temporal manner (Larkin et al., 1996). If neir participates in creating unequal N signaling

within these populations, then complete removal of net1 from al1 the cells may do nothing

more than delay the ability of these cells to adopt fates and differentiate. Eventually,

normal cell determination would occur, albeit in a less efficient manner. A similar

process may also be used to explain neil's non-functional role in N signaling during the

development of the wing.

A very plausible explanation for nen's differing involvement in N signaling

events is that not a11 components of N signaling are generally utilized. The generic nature

of the N pathway, being used as a signaling mechanism for a variety of developmental

processes, dictates that specific components will be used in specific contexts. The

observation that components such as Dl and Sri(H) have context specific roles in N

signaling during development contributes to the notion that a gene's involvement in N

signaling should be investigated in a context specific manner. Clearly, a thorough

cxnmination of nerc involvement in various developmental processes will be necessary to

elucidate the role of neri in the different contexts in which N signaling is known to be

required.

References

Alves dos Santos. C. M.. van Kerkhof, P.. and Strous, G. J. (2001). The signal transduction of the growth hormone receptor is regulated by the ubiquitinlproteasome system and continues after endocytosis. Journal of Biological Chemistry prrblislred ofilitre, Jaririary IO, 2001.

Artavanis-Tsakonas, S.. Muskavitch, M. A. T., and Yedvobnick. B. (1983). Molecular cloning of Norch. a locus affecting neurogenesis in Drosoyliila ffielnnognsrer. Proceedings of the National Academy of Sciences of the United States of America 80. 1977-1981.

Artavanis-Tsakonas. S., Rand. M. D., and Lake, R. J. (1999). Notch signaling: cell fate control and signal integration in developrnent. Science 284.770-6.

Aster, 1.. Robertson. E.. Hasserjian, R.. Turner. 1.. Kieff, E.. and Sklar. 1. (1997). Oncogenic forms of NOTCHI lacking either the primary binding site for RBP-JK or nuclear localization sequences retain the ability to associate with RBP-JK and activate transcription. The Journal of Biolqical Chemistry 272. 11336-1 1343.

Bailey. A. M., and Posakony. J. W. (1995). Suppressor of hairless directly activates transcription of Eiilranccr ofsplir complex genes in response Io notch receptor activity. Genes and Development 9,2609- 2621.

Baker. N.. Mlodzik. M.. and Rubin. G. (1990). Spacingdifferentiation in the developing Drosophila cye: a fibrinogen-rclated inhibitor encoded by scabrous. Science 250. 1370-1377.

Baker, N. E.. and Yu. S.-Y. (1997). Proneural function of neurogenic genes in the developing Drospliila eye. Current Biology 7, 122-132.

Bang. A. G., Bailey. A. M.. and Posakony. J. W . (1995). Hoirless promotes stable commitment IO the sensory organ precursor cell fate by negatively regulating the activity of the Norcli signaling pathway. Developmental Biology 172,479-494.

Bang. A. G.. and Posakony. 1. W. (1992). The Drosoplrilo gene Hairless encodes a novel basic protein that controls alternative cell fates in adult sensory organ development. Genes & Development 6. 1752-1769.

Basler. K., and Hafen, E. (1988). Control of photoreceptor cell fate by the sevenless protein requires a functional tyrosine kinase domain. Cel13.299-311.

Bender, L. B.. Kooh. P.J.. and Muskavitch. M. A. T. (1993). Complex Function and Expression of Delta During Drosophila Oogenesis. Genetics 133,967-978.

Blair. S. (1997). Limb development: Marginal Fringe benefits. Current Biology 7. R686-R690.

Blair. S. S. (1999). Eye development: Notch lends a handedness. Current Biology 9, R356-60.

Blaumueller, C. M., Qi, H., &gouras, P.. and Artavanis-Tsakonas. S. (1997). Intracellular Cleavage of Notch Leads to a Heterodimeric Receptor on the Plasma Membrane. Cell 90,281-291.

Boulianne, G. L., de la Concha, A.. Campos-Ortega. J. A., Jan. L. Y.. and Jan. Y. N. (1991). The Drosophila neurogenic gene nerrralired encodes a novel prolein and is expressed in precursors of larval and adult neurons. EMBO Journal 2975,2975-2983.

Brand. A. H.. and Perrirnon. N. (1993). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Developrnent. 401-415.

Brand, M.. and Campos-Onega. J. A. (1988). Two groups of intenelated genes regulate carly neurogenesis in Drosopliilo »iclnnogaster. Roux's Archives of Developmental Biology 197,457470.

Bray. S. (1998). A Notch Affair. Cell 93,499-503.

Brou. C.. Logeat, F.. Lecourtois. M., Vandekerckhove. 1.. Kouri1sky.P.. Schweisguth, F.. and Israel. A. (1994). Inhibition of the DNA-binding activity of Drosophila suppressor of hairless and of its human homolog. KBFZIRBP-J knppn. by direct protein-protein interaction with Drosophila hairless. Genes and Development 8.2491-2503.

Busseau. 1.. Diederich, R. 1.. Xu, T., and Artavanis-Tsakonas, S. (1994). A Member of the Notch Croup of lnteracting Loci. deltex. Encodes a Cytoplasmic Basic Protein. Genetics 136.585-596.

Cagan. R. L.. and Ready. D. F. (1989). Notch is required for successive cell decisions in the developing Drosophiln retina. Genes and Development 3, 1099-1 112.

Campos-Ortega, J. A. (1985). Genetics of carly neurogenesis in Drosopl~iln meln~iognster. Trends in Neuroscicnce.

Cûrthew, R.. and Rubin, G. (1990). seven in absentis a gene required for specification of R7 ccll fate in the Drosophila eye. Cell 3.561-577.

Choi. K.. and Choi. K. (1998). Fringe is esscntial for mirror symmctry and morphogenesis in the Drosophila eye. Nature 396,272-276.

Cooper, M.. and Bray. S. (1999). Friuled regulntion of Notch signalling polarizes cell fate in the Drosophila eye. Nature 397.526-530.

Cornell. M., Evans, D. A., Mmn, R., Fostier. M., Flasza. M., Monthatong, M., Artavanis-Tsakonas. S.. and Baron, M. (1999). The Drosophila melanogaster Suppressor of deltex gene. a rcgulator of the Notch receptor signaling pathway, is an E3 class ubiquitin ligase. Genetics 152.567-76.

Cubas, P., de Celis, 1.-F., Campuzano, S., and Modolell, J. (1991). Proneural clusters of ncliocre-scirfe expression and the generation of scnsory organs in the Drosophila imaginal wing d ix . Genes & Development 5.996-1008.

Deveraux. Q.. Uslrell. V., Pickart, C., and Reichsteiner. M. (1994). A 26s protease subunit that binds polyubiquitin conjugates. Journal of Biological Chemislry 269,7059-7061.

Diederich, R. J., Matsuno. K., Hing. H., and Artavanis-Tsakonas, S. (1994). Cytosolic interaction between deltex and Notch ankyrin repeats implicates deltex in the Notch signaling pathway. Development 120,473- 481.

Dietrich, U.. and Campos-Ortega, J. A. (1984). The expression of neuropnic loci in imaginal epidermal cells of Drosophila aiela~iognster. Journal of Neurogenetics 1,315-332.

Doherty, D., Jan. L. Y.. and Jan, Y. N. (1997). The Drosophila neurogenic gene big brain, which encodes a membrane-associated protein. acts cell autonomously and can act synergistically with Notch nnd Delta. Development 124,3881-3893.

Fanto. M., Mayes. C.. and Mlodzik. M. (1998). Linking cell-fate specification to planar polarity: determination of the R3IR4 photoreceptors is n prerequisite for the interpretation of the Frizzled mediated polarity signal. Mechanisms of Development 74-51-58.

Fanto. M., and Mlodzik. M. (1999). Asymmetric Notch activation specifies photoreceptors R3 and R4 and planar polarity in the Drosophila eye. Nature 397,523-526.

Fehon. R. G.. Kooh, P. 1.. Rebay. 1.. Regan, C. L., Xu, T.. Muskavitch. M. A. T.. and Artavanis-Tsakonas, S. (1990). Molecular interactions between the protein produccs of the neurogenic loci Notch and Delta. two EGF-homologous genes in Drosophiln. Cell 61,523-534.

Finley. D.. Sadis. S.. Monia, B.. Boucher, P.. Ecker, D.. Crooke. S.. and Chau, V. (1994). Inhibition of proteolysis and cell cycle progression in the multiubiquitination-deficient yeast mutant. Molecular Cell Biology 14,5501-5509.

Fleming. R. J. (1998). Stmctural conservation of Notch receptors and ligands. Seminars in Cell and Developmental biology 9,599-607.

Fleming. R. J., Gu, Y.. and Hukriede. N. A. (1997). Serrnre-mediated activation of Norcli is specifically blocked by the product of the gene fringe in the dorsal compartment of the Drosopltiln wing imagina1 disc. Development 124,2973-2981.

Fleming. R. J.. Scottgale, T. N., Diederich, R. 1.. and Artavanis-Tsakonas, S. (1990). The gene Serrnre encodes a putative EGF-like wansmembrane protein essential for proper ectodermal development in Drosoplda nrelntrogasrer. Genes and Development 4.2 188-2201.

Fortin, M. E.. and Anavanis-Tsakonis. S. (1994). The suppressor of hairless protein participates in Notch receptor signaling. Cell 79.273-282.

Fortini, M. E., Rebay, 1.. Caron. L. A., and Artavanis-Tsakonis, S. (1993). An activated Notch receptor clock cell-fate commitment in the developing Drosophila eye. Nature 365.555557.

Fostier, M.. Evans, D. A.. Artavanis-Tsakonas. S., and Baron, M. (1998). Genetic characterization of the Drosophila melanogaster Suppressor of deltex gene: A regulator of Notch signaling. Genetics 150. 1477- 1485.

Freemont. P. S. (2000). RING for destmction? Current Biology 1% R84-7.

Frise, E., Knoblich. J. A., Younger-Shepherd, S.. Jan. L. Y.. and Jan, Y. N. (1996). The Drosophila Numb protein inhibits signaling of t h e ~ o t c h receptor during cell-cell interaction in sensory organ lineage. Proccedings of the National Academy of Sciences of the United SI-tes of America 93. 1 1925-32.

Furriols, M., and Bray. S. (2Cûû). Dissecting the mechanisms of Suppressor of Hairless function. Developmental Biology 227,520-532.

Garcia-Bellido, A. (1979). Genetic Analysis of the Achnere-Scue System of Drosophila ntelanognsler. Genetics 91.491-520.

Gho. M.. Bellaiche. Y.. and Schweisguth. F. (1999). Revisiting the Drosophila microchaetc lineage: a novel intrinsically asymmetric cell division generates a glid cell. Development. 3573-3584.

Gho. M.. Lecourtois. M.. Geraud. G.. Posakony. J. W.. and Schweisguth, F. (1996). Subcellular localization of Supressor of Hairless in Drosophila sense organ cells during Notch signaling. Development 122. 1673- 1682.

Ghysen. A., Dambly-Chaudière. C.. Jan, L. Y., and Jan. Y.-N. (1993). Cell interactions and gene interactions in peripheral neurogenesis. Genes & Development 7,723-733.

Ghysen. A., and O'Kane. C. (1989). Neural enhancer-like elements as specific cell markers in Drosophiln. Development 105.35-52.

Ghysen, A., and Richelle. J. (1979). Detemination of Sensory Bristles and Pattern Formation in Drosophila II. The Achaete-Scute Locus. Developmental Biology 70.438452.

Golic. K. G.. and Lindquist. S. (1989). The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cel159.499-509.

Govers. R., ten Broeke, T., van Kerkhof, P.. Schwartz. A. L.. and Strous. G. J. (1999). Identification of a novel ubiquitin conjugation motif. required for ligand-induced internalization of the growth hormone receptor. EMBO Journal 18.28-36.

Grammont. M., and Irvine. K. (2001). fringe and Notch specify polar cell face during Drosophila oopnesis. Development 128.2243-53.

Greenwald, 1. (1998). LIN-IZNotch signaling: lessons from worms and flics. Genes & Development 12, 1751-1762.

Gu. Y., Hukriede. N. A., and Fleming, R. J. (1995). Serrafe expression can functionally replace Delfa activiiy during neuroblast segregation in the Drosoplrila embryo. Development 121. 855-865.

Guo, M., Jan. L. Y.. and Jan. Y. N. (1996). Control of daughter cell fates during asymmetric division: Interaction of Numb and Notch. Neuron 17.2741.

Gustafson. K., and Boulianne, G. L. (1996). Distinct expression patterns detected within individual tissues by the GAL4 enhancer trap technique. Genome 39,174-182.

Harper, S.. and Speicher, D. W. (1997). Expression and purification of GST fusion proteins. In Current Protocols in Protein Science. J. E. Coligan. B. M. Dunn, H. L. Ploegh. D. W. Speicher and P. T. Wingfield. eds. (Toronto: John Wiley & Sons, Inc.). pp. 6.6.1-6.6.21.

Hartenstein, V.. and Posakony. J. (1990). A dual function of the Notch gene in Drosophila sensillum development. Developmental Biology 142. 13-30.

Hartenstein, V.. and Posakony. 1. W. (1989). Development of adult sensilla on the wing and notum of Drosoplrila rrtelariogasfer. Developmenl107.389-405.

Heiizler. P., and Simpson. P. (1991). The Choice of Cell Fate in the Epidermis of Drosophila. Cell64, 1083-1092.

Hendzel, M. J., Wei, Y., Mancini, M. A.. Van Hooser, A.. Ranalli, T., Brinkley. B. R.. Bazett-Jones. D. P., and Allis, C. D. (1997). Mitosis-specific phosphorylation of histone H3 initiates primarily within pericentromeric heterochromatin during G2 and spreads in on ordered fashion coincident with mitotic chromosome condensation. Chromosoma 106,348-360.

Hershko, A.. and Ciechanover, A. (1998). The ubiquitin system. Annual Review of Biochemistry 67,425- 79.

Hinz. U., Giebel, B., and Campos-Ortega. J. A. (1994). The basic-helix-loop-helix domain of Drosopliila leilia1 of scicfe protein is sufficient for proneural function and activates neurogenic genes. Cell76.77-87.

Hoppe. T.. Matuschewski. K., Rape. M.. Schlenker, S.. Ulricli. H., and Jentsch. S. (2000). Activation of a membrane-bound transcription factor by regulated ubiquitin/proteasome-dependent processing. Cell 102. 577-586.

Huang. F.. Dambly-Chaudiere. C.. and Ghysen. A. (1991). The emergence of sense organs in the wing disc of Drosoplrila. Development 111. 1087-1095.

Huibregtse. J.. Scheffner. M.. Beaudenon, S., and Howley, P. (1995). A family of proteins stmcturally and functionally related Io the E6-AP ubiquitin-protein ligase. Proceedings of the National Academy of Science 92,2563-2567.

Huppen. S. S.. Jacobsen. T. L.. and Muskaviich. M. A. T. (1997). Fccdback reguhtion is central to Delta- Noich signalling required for Drosophila wing vein morphogenesis. Development 124. 3283-3291.

Joazeiro, C.A.. Wing, S.S.. Huang, H., Leverson, J.D., Hunter. T., and Liu. Y.C. (1999). The tyrosine kinase negative regulator c-Cbl as a RING-type. E2-dependent ubiquitin-protein ligase. Science 286,309- 12.

Joazeiro. C.A.P.. and Weissman, A.M. (20M)). RING finger proteins: Mediators of ubiquitin ligase activity. Cell 102,549-552.

Kamura. T.. Koepp. D. M.. Conrad. M. N., Skowyn. D.. Moreland. R. J . . Iliopoulos. O., Lane. W. S., Kaelin. W. G. 1.. Elledge, S. J.. Conaway. R. C.. Harper. J. W., and Conaway. J. W. (1999). Rbxl, a component of the VHL tumor suppressor complex and SCF ubiquitin ligase. Science 284.657-61.

Kidd, S.. Lieber. T., and Young. M. W. (1998). Ligand-induced cleavage and regulation of nuclear entry of Notch in Drosophila melanogaster embryos. Genes and Development 12,3728-40.

Klueg. K., and Muskaviich, M. (1999). Ligand-receptor interactions and trans-endocytosis of Delta. Serrate and Notch: members of the Notch signalling pathway in Drosophila. Journal of Cell Science 112.3289-97.

Klueg. K. M., Parody. T. R.. and Muskavitch. M. A. T. (1998). Complex proteolytic processing acts on Delta. a transmembrane ligand for Notch. during Drosophila development. Molecular Biology of the Cell 9. 1709-1723.

Koegl. M.. Hoppe. T.. Schlenker. S., Ulrich. H., Mayer. T.. and Jentsch. S. (1999). A novel ubiquitination facler. E4. is involved in multiubiquitin chain assembly. Cell 96.635-644.

Kopczynski, C. C.. Alton, A. K.. Fechtel, K., Kooh, P. J., and Muskavitch, M. A. T. (1988). Dclra. a ~r&li i la neurogenic gene, is transcripiionally complex and encodes a proiein related to blood coagulation factors and epidermal growth factor of vertebraes. Genes and Development 2. 1723-1735.

Kumar, S., Kao, W. H.. and Howley. P. M. (1997). Physical interaction between specific E2 and Hect E3 enzymes determines functional cooperativity. Journal of Biological Chemistry 272, 13548-13554.

Lai, E.. and Rubin, G. (2001). neuralized functions cell-autonomously Io regulate a subset of noich. dependent processes during adult Drosophila development. Developmental Biology 231.217-233.

Lai, E.. and Rubin, G. (2001). Neuralized is esseniial for a subset of Notch pathway-dependent cell fate decisions during Drosophila eye development. Proceedings of the National Academy of Science, USA 98.

Lai, E. C., and Rubin. G. M. (2001). neuralized functions cell-autonomously IO regulate a subset of Notch- dependent processes during adult Drosophila development. Developmenlal Biology prrblislierl onliae.

Larkin, M. K., Holder, K., Yost. C.. Gininger. E., and Ruohola-Baker, H. (1996). Expression of constitutively active Norclr arrests follicle cells nt a precursor stage during Drosopliila oogenesis and disrupts the anterior-posterior axis of the oocyte. Development 122.3639-3650.

Lecourtois. M.. and Schweisguth, F. (1998). Indirect evidence for Delta-dependent intracellular processing of notch in Drosophila embryos. Current Biology 8,771-4.

Lecourtois. M., and Schweisguth, F. (1995). The neurogenic suppressor of Iiairless DNA-binding profein mediates the transcriptional activation of the enlrartcer of splir complex genes uiggered by Notch signaling. Genes & Development 9.2598-2608.

Lehmann. R.. Jiménez. F.. Dietrich, U.. and Campos-Onega. J. A. (1983). On the phenotype and development of mutants of early neurogenesis in Drosopliila sielai~ogasrer. Roux's Archives of Dcvelopmenfal Biology 192.62-74.

Levkowitz. G.. Waterman. H.. Ettenberg. S. A.. Katz, M..Tsygankov. A. Y.. Alroy, 1.. Lavi. S.. Iwai. K., Reiss. Y.. Ciechanover. A.. Lipkowitz. S.. and Yarden. Y. (1999). Ubiquitin ligase activity and tyrosine phosphorylation underlie suppression of growth factor signaling by c-CbllSli-1. Mol Cell4. 1029-40.

Logeai. F.. Bessia. C.. Brou, C.. LeBail, O., Jarriault, S., Seidah. N. G.. and Israel. A. (1998). The Notchl receptor is cleaved constitutively by a furin-like convertase. Proceedings of the National Academy of Science USA 95.8108-8 112.

Lopez-Schier, H.. and St. lohnsfon. D. (2001). Delta signaling from the germ line conuols the proliferation and differentiation of the somatic follicle cells during Drosophila oogenesis. Genes & Development 15. 1393-1405.

Lorick. K. L., Jensen. J. P.. Fang, S., Ong. A. M., Hatakeyama, S.. and Weissman, A. M. (1999). RING fingers mediate ubiquitin-conjugating enzyme (E2)-dependent ubiquitination. Proceedings of the National Academy of Sciences of the United States of America 96. 113G4-9.

Lu, B., Rothenberg, M., Jan, L., and Jan, Y. (1998). Partner of Numb colocalizes with Numb during mitosis and directs Numb assymeuic locali~ation in Drosophila n e i d nnd muscle progeniiors. Cell95.225-235.

Lyman. D.. and Young, M. W. (1993). Further evidence for funcfion of the Drosopliila Notch protein as a transmembrane receptor. Proceedings of the National Academy of Science USA 90, 10395-10399.

Mdier, D., Nagel. A. C., Johannes, B., and Preiss. A. (1999). Subccllular localization of Hairless profein shows a major focus of activity within the nucleus. Mechanisms of Development 89, 195-199.

Maier, D., Sfumm. G.. KuFn, K.. and Preiss, A. (1992). Hairless, a Drosophila gene involved in neural development. encodes a novel, serine rich protein. Mechanisms of Development 38, 143-156.

Matsuno, K.. Diederich, R. 1.. Go. M. 1.. Blaumueller, C. M., and Artavanis-Tsakonas. S. (1995). Deltex acts as a positive regulator of Notch signaling through interactions with the Notch ankyrin repeats. Development I2I. 2633-2644.

Matsuno, K.. Go. M. J., Sun. X.. Eastman. D. S., and Artavanis-Tsakonas, S. (1997). Suppressor of Hairless-independent events in Notch signaling imply novel pathway elemcnts. Development 124,4265- 4273.

Morel. V., and Schweisgufh. F. (2000). Repression by suppressor of hairless and activation by Notch are required 10 define a single row of single-minded expressing cells in the Drosophila embryo. Genes and Development14.377-88.

Moschonas. N. (1998). direct submission (EMBL).

Muskavitch, M. A. (1994). Delta-Notch signaling and Drosophila cell fate choice. Developmenfal Biology 166.415-430.

Nakamura, H.. Yoshida. M.. Tsuiki, H., Ito. K.. Ueno. M.. Nakao. M.. Oka, K., Tada. M., Kochi. M.. Kuratsu, J., Ushio, Y., and Saya. H. (1998). Identification of a humnn homolog of the Drosophiln neuralized gene within the 10q25.1 malignant astrocyfoma deletion region. Oncogene 16. 1009-19.

Ohta. T., Michel, J. J.. Schottelius. A. J., and Xiong, Y. (1999). Rocl, a homolog of Apcll. represenü a family of culin partners with an associated ubiquitin ligase nctivity. Molecular Cell3.535-541.

Pan. D.. and Rubin, G. M. (1997). Kuzbanian Controls Proteolytic Processing of Notch and Mediates Lateral Inhibition during Drosophila and Venebrate Neurogenesis. Cell 90.271-280.

Papnyannopoulos. V.. Tomlinson. A.. Panin. V.. Rauskolb. C.. and Irvine. K. (1998). Donnl-vcntrnl signnling in the Drosophila eye. Science 281.203 1-2031.

Parks, A. L., Huppcn. S. S.. and Musknvitch, M. A. (1997). The dynnmics of ncurogcnic signalling underlying bristle development in Drosophiln melanognstcr. Mechnnisms of Dcvelopmcnt 63.61-74.

Parks. A. L., Klueg, K. M., Stout, J. R.. and Muskavitch, M. A. (2000). Ligand endocytosis drives receptor dissociation and nctivation in the Notch pnthway. Development 127. 1373-85.

Parks. A. L.. and Musknvitch. M. A. (1993). Delta function is rcquircd for bristle orgnn dcterminntion and morphogenesis in Drosophila. Developmental Biology 157.481-96.

Parks, A. L.. Turner. F. R.. and Musknvitch, M. A. (1995). Relationships between complex Delta expression and the specification of retinal cell fates during Drosophila eye development. Mechanisms of Development 50.201-16.

Patel. N. (1994). Imaging Neuronal Subsets and Other Cell Types in Whole-mount Drosophila Embryos and Larvae Using Antibody Probes. In Drosophila melanogaster: Practical Uses in Cell and Molecular Biology, L. S. Goldstein and E. A. Fyrberg. eds. (Toronto: Academic Press). pp. 445-487.

Petcherski. A. G., and Kimble. J. (2000). Mastermind is a putative aclivator of Notch. Current Biology 10. R47 1-R473.

Petnuch. C., Wang, S., Jan. L. Y., and Jan. Y. N. (2000). The divergent capase DRONC processes Notch and rcgulates neurogenesis. In 40th American Society for Cell Biology Annual Meeting, K. R. Yamamoto. ed. (San Francisco. CA: The American Society for Cell Biology).

Pickart. C. (1997). Targeting of substrates to the 26s proteasome. FASEB Journal 11, 1055-1066.

Poodry. C. A. (1990). shibire. a neurogenic mutant of Drosophila. Dev. Biol. 138.464472.

Posakony, J. W. (1994). Nature versus nurture: Asymmeuic cell divisions in Drosophila bristle development. Cell 76,415-418.

Poulson. D. (1937). Chromosomal deficiencies and the embryonic development of Drosophila melanogaster. Proceedings of the National Academy of Sciences of the United States of America 23. 133- 137.

Price. B. D.. Chang, Z.. Smith, R., Bockheim, S.. and Laughon, A. (1993). The Drosophila ncuralized p n e encodes a C3HC4 zinc RngecEmbo Journal 12,241 1-8.

Qi, H.. Rand, M. D.. Wu. X., Sestan, N., Wang. W.. Rakic. P., Xu.T., and Artnvanis-Tsakonas. S. (1999). Processing of the Notch ligand Delta by the Metalloprotease Kuzbanian. Science 283.91-94.

Qiu, L.. Joazeiro. C., Fang. N.. Wang. H., Elly, C.. Altman, Y.. Fang. D.. Hunter, T.. and Liu, Y. (2000). Recognition and ubiquitination of Notch by Itch, a hect-type E3 ubiquitin ligase. J Biol Chem 275.

Rhyu. M. S.. Jan. L. Y.. and Jan, Y. N. (1994). Asymmetric distribution of niinib protein during division of the sensory organ precursor cell confers distinct fates to daughter cells. Cell 76.477-491.

Robey. E. (1997). Notch in venebrates. Current Opinion in Genetics and Development 7,551-557.

Rulifson. E. J., and Blair. S. S. (1995). Norch regulates ivitigless expression and is not required for reception of the paracrine wirigless signal during wing margin neurogenesis in Drosopltila. Development 121.2813-2824.

Ruohola.H.,Bremer,K. A.. Baker. D., Swedlow, J. R.. Jan,L. Y.. and Jan,Y. N. (1991). Roleof the neurogenic genes in establishment of follical cell fate and oocyte polarity during oogenesis in Drosopliilo. Cell66.433-449.

Ruohola-Baker. H.. Jan. L. Y., and Jan. Y. N. (1994). The role of gene cassettes in axis formation during Drosopliila oogenesis. Trends in Genetics IO. 89-94.

Salcini. A., Confalonieri. S., Doria. M., Santolini, E., Tassi, E., Minenkova. O.. Cesareni. G., Pelicci. P.. and Di Fiore, P. (1997). Binding specificity and in vivo targeis of the EH domain. a novel protein-protein interaction module. Genes and Development 11,2239-2249.

Santolini. E.. Puri. C.. Salcini, A. E.. Gagliani. M. C., Pelicci, P. G..Tacchetti. C.. and Di Fiore. P. P. (2000). numb is an endocytic protein. Journal ofCeIl Biology 151, 1345-1351.

Scheffner, M., Huibregtse, J., and Howley, P. (1994). Identification of a human ubiquitin-conjugating enzyme that mediates the EG-APedependent ubiquitination of p53. Proceedings of the National Academy of Science. 91,8797-8801.

Schaffner, M., Nuber. U.. and Huibregtse. J. (1995). Protein ubiquitination involving an EI-EZE3 enzyme cL;quitin thioester cascade. Nature 373, 81-83.

Schroeter, E. H., Kisslinger, J. A., and Kopan, R. (1998). Notchl signaling requires ligand-induced proteolytic release of the intracellular domain. Nature 393.382-386.

Schweisguth. F. (1999). Dominant-negative mutation in the b e t d and beta6 proteasorne subunit genes affect alternative cell fate decisions in the Drosophila sense organ linwge. Proceedings of the National Academy of Sciences of the United States of America 96, 11382-6.

Schweisguth, F., and Posakony. J. W. (1994). Antagonistic activities of Siippressor of Hoirless and Hairless control alternative cell fates in the Drosopl~ila epidermis. Development 120. 1433-1441.

Seugnet. L.. Simpson. P.. and Haenlin, M. (1997). Requirement for dynamin during notch signaling in Drosophila neurogenesis. Developmental Biology 192,585-598.

Simpson, P. (1990). Lateral inhibition and the dcvelopment of the sensory bristles of the adult peripheral nervous system of Drosophila. Development 109.509-519.

Singson. A.. Leviten. M. W.. Bang, A. G.. Hua. X. H., and Posakony. J. W. (1994). Direct downstream targets of proneuwl activators in the imagina1 disc include p n e s involved in lateral inhibitory signaling. Genes & Development 8.2058-2071.

Skeath, J. B., and Carroll, S. B. (1993). The acliaete-scirre complex: Generation of cellular pattern and fate within the Drosoplda nervous system. FASEB Journal 8,714-721.

Skowyra, D., Koepp, D. M., Kamura, T., Conrad. M. N., Conawq, R. C.. Conaway, J. W., Elledge. S. J.. and Harper, J. W. (1999). Reconstitution of G1 cyclin ubiquitination with complexes containing SCFGrrl and Rbxl. Science 284,662-5.

Smoller, D., Friedel. C., Schmid. A., Bettler, D.. Lam, L.. and Yedvobnick, B. (1990). The Drosopliila neurogenic locus mastermind encodes a nuclear protein unusually rich in amino acid homopolymers. Genes and Development 4,1688-1700.

Sotillos, S.. Roch. F., and Campuzano, S. (1997). The metalloprotease-disintegrin Kuzbanian participates in Notch activation during growth and patterning of Drosophila imagina1 discs. Development 124,4769-4779.

Spence. 1.. Sadis. S., Haas, A., and Finley, D. (1995). A ubiquitin mutant with specific defccts in DNA repair and multiubiquitinntion. Molecular Cell Biology 15, 1265-1273.

Spradling. A. (1986). P element-mediated transformation. In Drosophila a practical approach. D. Roberts, ed. (Washington: IRL press). pp. 175-197.

Strous. G., van Kerhof. P.. Govers. R., Ciechanover, A., and Schwarrz. A. (1996). The ubiquitin conjugation system is required for ligand-induced endocytosis and degradation of the growth hormone receptor. EMBO Journal 15,3806-3812.

Strous, G. 1.. and Govers, R. (1999). The ubiquitin-proteasome systein and endocytosis. Journal of Cell Science 112, 1417-1423.

Struhl, G.. and Adachi. A. (1998). Nuclear access and action of Notch in vivo. Cell93.649-660.

Strutt, H., and Strutt. D. (1999). Polarity determination in the Drosophila eye. Current Opinion in Genetics and Development 9,442-446.

Sun. X., and Artavanis-Tsakonas. S. (1996). The iniracellular deletions of Delta and Serraie define dominant-negative forms of the Drosophila Notch ligands. Development 122,2465-2474.

Sun. X., and Artavanis-Tsakonas. S. (1997). Secreted forms of Delta and Serrate define antagonists of Notch signaling in Drosophila. Development 124,3439-3448.

Tamun. K., Taniguchi. Y.. Minoguchi. S., Snkai, T., Tun. T.. Furukawa. T., and Honjo, T. (1995). Physical interaction between a novel domain of the receptor Notch and the transcription factor RBP-J kappalSu(H). Current Biology 5. 1416-23.

Tyers. M.. and Jorgensen. P. (2000). Proteolysis and the cell cycle: with this RING 1 do thee destroy. Current Opinion in Genetics and Development 10.54-64.

VBssin. H., Vielmetter. J., and Campos-Onega. J. A. (1985). Genetic interactions in early neurogenesis of Drosopliila n~clariogaster. Journal of Neurogenetics 2.291-308.

Verheyen. E., and Cooley, L. (1994). Looking al Oogenesis. In Drosophila melanognsier: Practicnl Uses in Cell and Molecular Biology, L. S. Goldstein and E. A. Fyrberg. eds. (Toronto: Academic Press). pp. 545- 561.

Wang. S.. Younger-Shepherd, S., Jan, L. Y., and Jan. Y. N. (1997). Only a subset of the binary cell fate decisions mediated by NumbINotch signaling in Drosophila sensory organ lineage requires Suppressor of Hairless. Development 124.4435446.

Waterman, H.. Levkowitz, G., Alroy. 1.. and Yarden, Y. (1999). The RING finger of c-Cbl mediates desensitization of the epidermal growth factor receptor. Journal of Biological Chemistry 274.2215 1-4.

Weinmaster, G. (1997). The ins and ours of notch signaling. Molecular and Cellular Neurosciences 9.91- 102.

Wharton, K. A.. Johansen. K. M.. Xu, T.. and Artavanis-Tsakonas, S. (1985). Nucleotide sequence from the neurogenic locus Norcll implies a gene product that shares homology with proteins containig EGF-like repeats. Cell43.567-581.

Wilkinson, K. (2M)O). Ubiquitination and deubiquitination: Targeting of proteins for degradation by the proteasorne. Semin Cell Dev Biol 11, 141-148.

Wilson. R., Ainscough, R.. Anderson. K.. Baynes, C.. Berks. M., Bonfield, J., Burton, J.. Connell. M.. Copsey. T., Cooper. J., and al. e. (1994). 2.2 Mb of contiguous nucleotide sequence from chromosome III of C. elegans. Nature 368.32-8.

Wu, L., Aster, J. C.. Blacklow, S. C., Lake, R., Artavanis-Tsakonas, S., and Griffin. J. D. (2000). MAMLI. a human homologue of Drosophila Mastermind. is a transcriptional co-activator for NOTCH recep'ors. Nature Genetics 36.484-489.

Xu. T., and Arhvanis-Tsakonas. S. (1990). deltex. a locus interacting with the neurogenic genes, Notch, Delta. and mastermind in Drosophila melanogaster. Genetics 126,665-77.

Xu, T., Caron. L. A., Fehon. R. G., and Artavanis-Tsakonas. S. (1992). The involvement of the Norclr locus in Drosopliila oogenesis. Development 115,913-922.

Xu, T., and Harrison, S. (1994). Mosaic Analysis Using FLP Recombinase. In Drosophila melanogaster: Practical Uses in Cell and Molecular Biology, L. S. Goldstein and E. A. Fyrberg. eds. (Toronto: Academic Press), pp. 655-681.

Xu. T., and Rubin. G. M. (1993). Analysis of genetic mosaics in devrloping and adulr Drosopliilo tissues. Development 117,1223-1237.

Ye, Y.. Lukinova. N., and Fortini, M. (1999). Neurogenic phenotypes and altered Notch processing in Drosophila Presenilin mutants. Nature 398,525-529.

Yeh, E., Zhou, L.. Rudzik, N.. and Boulianne, G. (2MH)). Neuralized functions cell autonomously to regulatc Drosophila sense organ development. EMBO J 19,4827-37.

Yokouchi, M., Kondo. T., Houghton, A., Bartkiewicz, M.. Horne, W. C., Zhanc. H.. Yoshimura, A.. and Baron. R. (1999). ~ i~and- indu& ubiquitination of the epidermal growth factor receptor involves th- internction of the c-Cbl RING finger and UbcH7. Journal of Biological Chemistry 274.31707-12.

Zeng, C., Younger-Shepherd. S.. Jan. L., and Jan. Y. (1998). Delta and Serrate are redendnnt Notch ligands required for asymmelric cell divisions within the Drosophila sensory organ lineage. Genes and Developmenl12, 1086-1091.

Zheng, L., Zhang. J.. and Carlhew. R. (1995). frizzled regulates mirror-symmelric pattern formation in the Drosophila eye. Development 121,3045-3055.

Zheng. N.. Wang. P.. Jeffrey, P.. and Pavletich, N. (2000). Structure of a c-Cbl-UbcH7 complex: RING domain funclion in ubiquitin-protein ligases. Cell 102.533-539.