The novel Notch homologue mouse Notch 3 lacks specific epidermal growth factor-repeats and is expressed in proliferating neuroepithelium

ELSEVIER SCIENCE IREI AND

Mechanisms of Development 46 (1994) 123-136

The novel Notch homologue mouse Notch 3 lacks specific epidermal growth factor-repeats and is expressed in proliferating neuroepithelium Michael Lardelli, Jonas Dahlstrand, Urban Lendahl* Department of Cell and Molecular Biology, Medical Nobel Institute, Karolinska Institute, S-171 77 Stockholm, Sweden

(Received 17 December 1993; accepted 22 December 1993)

Abstract

In Drosophila, the Notch gene is pivotal for cell fate decisions at many stages of development and, in particular, during the formation of the nervous system. Absence of Notch results in the generation of excessive numbers of neural cells at the expense of epidermal cells. Two previously identified mammalian Notch homologues encode all the principal features of the Drosophila gene, e.g. 36 EGF-repeats and 3 Notch/lin-12 repeats extracellularly and 6 intracellular cdclO/SWl6 repeats. We report here the characterisation of a third mammalian homologue, mouse Notch 3, which shares the same remarkable conservation relative to the Drosophila gene as the two previously identified homologues, but with three important distinctions. First, Notch 3 specifically lacks the equivalent of EGF-repeat 21; second, it lacks an EGF-repeat-sized region comprising parts of EGF-repeats 2 and 3; and third, it encodes a considerably shorter intracellular domain. The Notch 3 gene is expressed at high levels in proliferating neuroepithelium and expression is downregulated at later stages. The expression patterns of the Notch 1, 2 and 3 genes are quite distinct during central nervous system (CNS) development, and all possible combinations of expression, i.e. none, one, two, or all three genes, are seen, suggesting a combinatorial code of Notch function in mammals. Considering the predominantly early expression in CNS and its distinct structural features, the Notch 3 gene is likely to contribute significantly to vertebrate Notch function during CNS development. Key words. Development; Neurobiology; Neurogenic genes: Differentiation; Neural tube: Olfactory epithelium; Opa repeats

I. Introduction

Development of a nervous system is a complex process, and in Drosophila this critically depends on the function of a group of genes called the neurogenic genes. Loss of any one of these genes results in the differentiation of excess numbers of neural cells at the expense of epidermal cells (for review see Artavanis-Tsakonas and Simpson, 1991; Greenwald and Rubin, 1992). The best characterised neurogenic gene is Notch, and in addition to its effects on neural/epidermal fate choice in Drosophila, Notch also exerts effects during later stages of nervous system development and in other tissues. For example, loss of Notch function affects axon guidance in the embryo (Giniger et al., 1993) and perturbs eye * Corresponding author.

(Cagan and Ready, 1989; Fortini et al., 1993) and bristle (Heitzler and Simpson, 1991) development in the pupa. Notch function is also required for the formation and maintenance of various embryonic epithelia (Hartenstein et al., 1992) and during oogenesis (Xu et al., 1992). The diverse effects caused by loss of Notch function are in keeping with its broad and complex pattern of expression. The Notch gene encodes a large transmembrane protein (Kidd et al., 1986; Wharton et al., 1985a) which is believed to function as a cell surface receptor mediating signals between cells in close proximity (Heitzler and Simpson, 1991). The Notch protein, in its extracellular domain, contains 36 repeats with homology to epidermal growth factor (EGF-repeats) and three lin-12/Notch repeat domains. On the intracellular side the most conspicuous region consists of six cdclO/S WI6

0925-4773/94/$07.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0925-4773(94)00236-G

124

M. Lardelli et al. / Mechanisms q[' Deveh~pment 46 (1994) 123-136

repeats, which are similar to motifs found in a number of proteins (e.g. ankyrin, bcl-3 and cactus see discussion) and are thought to mediate protein-protein binding. The putative action of Notch as a receptor is supported by several lines of evidence. Notch hypomorphic mutations behave in a cell-autonomous fashion in genetic mosaic studies while cells mutant for Delta, a neurogenic gene producing a putative ligand for Notch, behave non-autonomously (Heitzler and Simpson, 1991). Intracellularly, the protein kinase product of the shaggy gene is thought to mediate signalling by Notch to the nucleus and possibly also laterally to Delta proteins residing in the same plasma membrane (Ruel et al., 1993). Finally, Notch protein molecules probably dimerise in the plasma membrane and are subject to phosphorylation (Kidd et al., 1989) but do not show homology to tyrosine kinases. Recently, two vertebrate homologues of the Drosophila Notch gene have been identified. One of the homologues, Notch 1, has been characterised in Xenopus (Xotch) (Coffman et al., 1990), rat (Weinmaster et al., 1991), man (TANI) (Ellisen et al., 1991) and mouse (Franco del Amo et al., 1992; Reaume et al., 1992; Kopan and Weintraub, 1993; Lardelli and Lendahl, 1993 (formerly called Moteh A)). The second, Notch 2, has thus far been identified in man (hN) (Stifani et al., 1992), rat (Weinmaster et al., 1992) and mouse (Lardelli and Lendahl, 1993) (formerly called Motch B). Both homologues are remarkably conserved throughout their coding regions, and are approximately 50% identical at the amino acid level, both with respect to each other and the Drosophila Notch gene. The individual forms of particular EGF-repeats in Notch proteins are probably important for Notch function since all the hitherto characterised vertebrate homologues maintain the 36 repeats of the Drosophila protein and corresponding EGF-repeats in different Notch proteins are more homologous to each other than to adjacent repeats in the same protein (Coffman et al., 1990; Ellisen et al., 1991; Lardelli and Lendahl, 1993). The importance of the EGF-repeats is further supported by the finding that two specific repeats are essential for binding to Delta (Rebay et al., 1991), and that point mutations in the EGF-repeats of Drosophila Notch produce patterning defects (Hartley et al., 1987; Kelley et al., 1987). The involvement of individual EGF-repeats in proteinprotein interactions has also been demonstrated in other systems, e.g. in laminin binding to nidogen (Mayer et al., 1993). The cell-cell signalling mediated by Notch during Drosophila neurogenesis is required for the process of 'lateral inhibition' during which a single neuroblast differentiates and delaminates from a group of potential neural precursors in the ectodermal epithelium while inhibiting neighbouring cells from doing so (see

Artavanis-Tsakonas and Simpson, 1991 : Cabrera, 1992: Greenwald and Rubin, 1992 for review). In vertebrates, discussion of Notch function has emphasised a role in controlling cell proliferation and maintaining cells in a state in which they are receptive to instructive signals. These ideas have arisen from three findings: first, the human Notch 1 gene, when translocated and fused to the T-cell receptor gene, produces leukemia (Ellisen et al., 1991). Second, an MMTV integration into the Notchrelated mouse gene int-3 can produce mammary tumors (Robbins et al., 1992), and the same phenotype appears in transgenic mice into which the truncated #tt-3 gene has been integrated (Jhappan et al., 1992). Third, expression of a truncated Notch gene in Xenopus produces overcommitment of ectodermal cells to the neural fate (Coffman et al., 1993). Similar experiments in Drosophila further support a receptor function (Fortini et al., 1993; Rebay et al., 1993; Struhl et al., 1993). Despite the above observations and the extensive work carried out in Drosophila, it has not yet proven possible to propose a unifying model of Notch and neurogenic gene function(s) that encompasses the wide variety of mutant phenotypes observed. Important steps towards such a model clearly involve a better understanding of the complexity of the Notch gene family and the expression patterns of its individual members. With this object in mind and a general interest in understanding the molecular mechanisms controlling neural pattern formation and cell differentiation, we originally set out to characterise mouse homologues of the Drosophila Notch gene. Here we report the discovery of a third mouse Notch homologue, Notch 3. Notch 3 encodes a protein of similar size and with all the principal characteristics of Notch 1 and 2, but shows three distinct structural differences: (i) it lacks EGF-repeat 21, (ii) it lacks an EGF-repeat-sized region covering parts of EGF-repeats 2 and 3: (iii) it encodes a considerably shorter C-terminal domain. Parallel analyses of the expression patterns of Notch 1, 2 and 3 show that Notch 3 is abundantly expressed in proliferating neuroepithelium during CNS development. In different tissues we observe all possible combinations of Notch homologue expression suggesting a combinatorial code for vertebrate Notch function. The distinct structural features and the high level of Notch 3 expression in the early CNS suggest that it plays an important role in executing some of the functions attributed to the Drosophila Notch gene in CNS development. 2. Results

2.1. Identification of a novel mouse Notch homologue We previously reported the cloning of two mouse Notch homologues, Notch 1 (Motch A) and Notch 2, (Motch B) by a polymerase chain reaction (PCR) approach, using degenerate primers to clone portions of

M. Lardelli et al. /Mechanisms o/" Development 46 (1994) 123-136 A

~u~,

Not~ ) i ~ L O / ~ G~ R ~ R . . . . . . . . . . . I III I

~ro.o~

I G A * ~ P ~ C L D ~ PC ~ R I I II~

R 124kL PP~P P P ~ ~ P L L L L L A G L . . . . . . . t I I I I I I

~. ~ot~

CI'H ~ P S LB~A C LC L PO~V O~'~C ~LED ~ I I I I I~ I II

~

I

I

f

+1

II

SG~ ~ C I d I

II

100 ~ $ W A G T ~ p $ ~ C~ I II I

*~2

oP1

I~

125

I

I

I

I

IIII

I

II

II

I

I

200

II

I

II

II

I

PCAPs~(~h~Y~R~`SSDVY~bt~r~tMV~DCP~H~C~J4~CVDG~R~P[N~PP~-P~PCT~DgC.

Q~P~C~C~C~S~ATA~CPHGAT

II [~[I I P ~1 I I I I IIII III I ill II I lltll PCS P S p C Q N A G I C~. $ NGL~ Y~(~KCPKG F ] ~ G ~ EQN't~X>C L ~ L C ~ G O T C

I III ~ ~ WCQ~T

II I I I1~ II II I [~ I S[3YT C!~CI~PI~PTG~ FC ~

{l

rill

L

300

IIHI D~

;I ~

I

I H~

II IIIII I ~ IC ~ A G L ~

II I III S~T~ ~

I

II III C ~GAT

too

CH~RVASFYCACPAC~T~LLC~LDD~CVSN~CH~DAICDTN~VSG~A~CTC~PGFTC~ACDQ~VD~H~FLC~GR~PRC~CL~TCLDkI~I~ 3

I

I~1;

I

+l~llllllllll

lllJl

IIIl~l

I

I

I

I

I

I

I

1

III

I

~lll

I

C~T~3VC*S~Y~C`'~K~F/~TGLLCHLDD~`C`.tSNPCkL~d~CDTSP~C~'YAC~CA,P~:`[KGV`DCSEDIDEC~ ~H~I "EGF10

IIII ~1 I I I IIIII15 CVI~pGsy~ ~S~.~GP~C

Ill III I III I I III ~ N I NECRSH p C ~ E ~ S C L D ~ P ~ ~ "£GF12

"EGF11

S00

AGF~GTYCEVD~DECQS$PCV]~<~VC1~VNGPSCTCP~GFSC~CQL~VDECA~TPCRNGAXCV~Q~D~YEC~C~E~FEGTLC~$PDP~HH~I~F~ACA~%~C£$QVD£C~S PGFTGT~

E i DI DECQ~NpC L N D G T C H D K iN ~ F K C S C ~ G P T G ~ B C Q I

NI ~OCQSQPCRN]~G ICH[A~ i AGYSCECp p ~ G T ~

• EGF1 ~

E i NI ~

D ~ C "EGFIS

-I~GF 1 ~

II I I i I I I I NPCQ . v D ~ C~C~ VG~YY C ~ ~ , ~

I IIIII ~G~C ~

i~ $

p KC ~ D ~ T ~

~ PR c s ~

1CHCppGYTG~C..

I Illl ~ I IIII I I I PtJtl I ~1 I I I I ] ~ I II I FI I I~ I~ ii H$1¢PC ~ ATC I DG[ N S Y X C ~ C V P G P ~ C E K b ~ D E C t S $ ~ C I D~VI~YKCEC P R G F Y ~ C L S D V D ~ ~ C ~ E "ZOO17 ~ "ZO~IO " ZGVl~

LADDACB$ Q P C Q A G ~ CYSDG I G F R C ~ ~ F E LDI D E C S S N ~ Q H G O T C Y I ~ K

~ / ~

.......................................

LI~SCQCI~TQ~C~'~

I~ V ~ C G N G ~ T C

pl~R~

~

~

I i ~ i ~ C E I ~ ~N EF

gO0 CESDp[;I~L~. C~pp ~.t~pB

~ ~p$~BH~

IE~G~CVC]KV

I

,EGPI6

¥ Pc v c z ~

i C ~ Q i N ~C EL

p

s

~PLDF~

~L~G~

¥

S0o C~VD~C AGA~ PC GP H G T C T ~ P ¢~NFR C lCH I~G ~ GPFC DQ~I [X~CD p]~PC LH G~ S ~ > G V C ~ F ~ C $ C I ~ I~AOp!~~ ~ C ~ S ~ G p G. ~ T ~ ~ ~ ~ C p ~ G GFM C £ I DL p ~ $ p s ~ ~ I III III I I I II I ~I I III II I I II I II I III I i I IIII II I II I ~ li II I I I II III l CDEDI D E C S L S $ ~ C , ~ t ~ C L C ~ K G Y I ~ D C A I I ~ D O C A S p P C ~ ~ L D ~ I ( ~ y ~ c L c v ~ p I ~ K H C ~ D I ~ ~ p ~p~G I ~ E ~ ESSC ~ • I~GF23 *]~Op~4 "EGF2$ *EGF26 1ooo ~G'I~CVEGV SSFS CLC~ P ~ G T H C Q y E A DPC F S~ PC LHGG I ( ~ T H P~ p~%'T C~ L~GI ~ ~ C~ ~ CUbA.. yC IC p F G W S ~ L C DIQ~ LPCT F ~ I ~ G V R L E Q L C Q ~ G ~ C 1D

I II I I II I II I I III I I I IIII II llll I II IJ I I G~SC~DG~NGYN~SCLAGY~C4~CQ`~LN~DSNPCL2~ATCH~YTCH~P~GFT~QCS~~HQPs~A~Q~GL~Q~ "SG~27 I~00

[ ~

II I III

I

~

"IGP~I

II

I

I II~

I

I ~

"~GF29

K~S~YCVCPEGRTq~HCE~E~PCTAQPCQ~GGTCRG~4GGYvCECPAGYAGDSC~NID~CASQPC~q~IDLV~Y~SCP~T~EI~P$LDS~L~DLVGG~C~ II I I I II I I I IIII I[lll Y~5 HV~C$~¥AC~yC~KE I DEC~PC~DLI • ~GP~0

I I I I III ~YZ C ~ F ~

III II ~11111 I I I IIIIIII E~I ~ ~ H D ~ P ~ S C p ~ *[~OF~I

I

IIII III I ICEI~POA

I ....... ,RGp32

~200

I II I I II1+ I I C ~ Z ~VGGFECVC

~00

p p ~ Y TGLHCEADI NEC~ pGACk~A Hq~ DCL~DpC~H PR~'CHP~F~OP~CQI A L S P C ~ p ~ H S L ~ G L T F T C ½ C ~ P P W G ~ C~ v~ ~ EL~PV~ I ~'A~ I II IIIII I I I~ ~ I II I I I III II I I I I I II I I il ,pOF V C ~ C EGD iNEC LSNp C SNA GT L D C V Q L V ~ Y~C~C~ PG ~ H C E H KVDF CA~,~ pC ~N GC~NC N IkQS . . . . . GH HC I C b ~ F y c ~ c E L s G Q O C D~Npc~vGNCVVADEG

GpR CACP pGLSGp I II II i I pGY~ C£CP ~GT LGE

J I ztoo SC~VS~SpSC~J~SCASApCLH~OSCL~'VQSVppPI~CVCAP~W~p~¢~ ................ TpS#U~p~vps~p~CP~CQkK~GD~CC~CI~tpGCGWCOO~CSL~'~COP~pC~bCC~m F

I

i~

I

I

I

I I II

II

I

I

I I

II

III

~

~I IIII

II

I I

il

lsoo --~ LF~ ~CDpACS

SPAC L' ~ p [ ~

KI~CNREC~C

W

S~

[~RTCNPVYI~CA~pA~CD~GC~ECOW[~L~C~

:HDC.. E R K L ~

~TLF D A Y C ~ Y ~ D G p

EV. pA LLARGVLV LTVL LppE ELLR S S A D F ~

CDyQ~ItCSWDGL~C~2~T~S

p VIJ~G~I,~ V

~

L ~ I ~ T$ ~ ~ L ~

E I~ Q p ~

~ ~

~

~Vp

H

py MR pS pG

D I II~

V

z~oo t I SE . . . . . . S~V~RELGpE ...GSVVMLEIDb~LCL~D~CPPDA~Ju~DYL(~tLSAVER LDPPyp~DV~G., .EPLEApE~V~LLPLLVAGAVFLLI IFILGV~V..ARPJLREHSTL~FPEGp L I I llllll I II I II I I I l III I I I II ili I llllll pE I E D ' P D p ~ I L y T ~ H ~ f i ~ IQI y LE I D i ~ C . . . . . . T]ICPT~AV~FLAATA~QLR~D~QINSVBO II~rPGD~DNGEppANVKY~ I TGI I LV % I A ~ p ~ H~Wp pEGp

;@o0 GFT p 124LA$ p C G C ~ E

I iiii

i

1~oo ~ E p VG~DAJ,Q['U~Ib~GE$ ~ E V V T D L . . . . . . II I+ $ I I I

A L ~ DI ~ G ~ G I

I

P ~ A E EDI A ~

II

ND~ECpEA~LK ~ I II

S ~ I ISDL I C @ A Q L G A ~ T D ~ T G E T A L H

I

I

llil

III

I I

I llPl

. t/E Ep~HCIJ~ . . . . . . . . . . . . . I I I

I

II

L V ~ , D I RVAPAT ALTP P ~ I I I

L~

IIIIHIII

II~I

I II

~

D

llllll

~I III

iIIII~

pAmPA ~ % A F ~ R L F ½ A ~ R T P L ~ V ~ Q I

III

L

I I [II

L

~

I

~ pD III

II

1900 QI b I R ~ S T D L D A P ~ D G ~ T

L~yA.~A~LLDAC~Q~4~GRTPLHTAVTA~GVF

llllll

GLTP I~IAAVRGG~LDT. O~DI F2~N]~D~TAQ~ I/DL L A ~ A E L N A T H D K T ~ E T S L H

EPED~½ i I I I

J'cd~Z

ALI L A ~ LAV

IJ 4 llllllll

~

pLI L ~

LAI

~o,oo EGMVEEL l A~MADT~IAVDELG~$A LHWAAAVNNVRAT LAL L ~

Ab~D~K~]~E ET PL PLAJ~ E GSYE ~

IIIII II II I II IIIIIIIIII II II II EGNV ZDL I TADA D ZNA A~NS GKTA L ~ A A A V N ~ T E A ~N I L L ~ H A ~ •talcs

I1~

I

GPPL EG ,

I

, pyA'I~ . . . . .

I

II

I

t

I

I

II

ATAV$1J, QLC,,,I~SI~,GpLGRQp!~].. QCV .

I

I

H

I

I

....

~p

.........

o P . SI, L .

I

II

. . . . . . .

I

N

•

GH . . . . . . . . . . . . . . . . . . .

I~

. . . . . . . . . . . . . . . . .

SPG

I

T

~

G£EYp~G ......................... I I II

~

~

~SSP lid

~

V

T ...........

I

I

I

I

I

Ill

2200 eaSY..SPQ~p ~

I

p!wQQE LIA~~ O L ~

I L

I

L I ~ I p V A V p L . . . . DI~ . . . . . . . . . . . . . . . . . . . .

. . . . . .

II Q

II

I

L~Cl~OU~ I F~DIqYAy ~

I ~

. . . . .

I

~tL....laCP .................

+ I~

L

I

I ~ .ta~PGp . . . . . . . . . . . . .

ySNQSpp½ SVQ~ SLAL SpHAYLOS p S p ~ L p S 6 ~ $ I ~ H

I

I

III

~'GMG~',U., p S p y D~ SS I,Iy ~ A H , AAp LA.N. ~ * ~ M ' I ~ . , AIKQp p s y E I ~ I ] ~ I A Q ~

E I TDHL ~ L pR DVAQE~ ~ QDI ~ LL ~ p SGp ~ $ p ,

~ OPHGL.GPLL I Ill IIIIIIII IIIIIIll IIII Illllll t111 t I I S ~ A ~ b b ~ ~ A ~ E I T O I l e R L P R D V ~ g ~ LHH D I V ~ LLDE H. ~ P R S P ~ ' ¢ L J ~ T P ~ M Z~ J 2t00

III I IItHIIIIIIIIII D A ~)OKPET p L P L ~ c~S

I

LL~ ~

S

I ~

II $

L

~G

....

I~p . . . . . . .

I O

~

~ L C ~

II

I S~

KA .RF~VPSEH ..... I I I I

R..

LP PP

I

I

I E~

II pp

I

I~

p ...... J

A ..... G L ~ A ~

YZa~P . . . . . . . . . . . . . . . . . . . . .

I I

Ii

~ S ~

I ~

IS

~

YLTPSpESp£H~ASpsppsL$~.I +

IIIIIII

III

II

d III

~~L~GLEPGJ;AGLD1J~G¥CC~PDSF~S~Qi~*~P~S~SSM~S~;S~G~L~PS~Q~`~4@~AyYQYLTP$~T~LV~L~SPES~G~$ssPR~NS~ OPA replltm

I

HI

I

P

2~00

II

I

~31@

PEST d ~ n

I ~70~

Fig. 1. (A) The mouse Notch 3 gene. An alignment of the putative amino acid sequences of the mouse Notch 3 (upper) and Drosophila Notch (lower) proteins. Identical amino acid residues are indicated with vertical bars. Numbers refer to amino acid residues in the Notch 3 protein. Positions of the first amino acid residues in EGF-repeats, Notch/lin-12 (LNR) repeats, and c~h'IO/SWI6 repeats of Drosophila Notch are indicated by * and are according to Coffman et al. (1990). The EGF-repeat region, LNR-repeat region, transmembrane region, cdclO/SWl6 region, and the PEST domain are denoted by boxes, while the OPA repeats in the Drosophiht Notch protein are underlined. Bold black bars indicate the regions of missing EGF-repeats in the Notch 3 gene. Genomic D N A was sequenced between codons for amino acids 101-150 and 725-791. Vertical arrows indicate the amino acid residues in whose codons introns were found. The alignment was performed using the G A P programme (Devereux et al., 1984) with the following parameters: from amino acid (aa) 1 to 1998: gap weight = 3.0, length weight = 1.0: from aa 1999 to 2318: gap weight = 1.0, length weight = 1.0. This combination allowed optimal alignment of the PEST sequences near the C-termini of the proteins. Drosophila Notch has 297 additional amino acid residues in the region between the ccA'IO/SWI6 repeats and the PEST sequences. The start codon was selected on the grounds that it lies in a strong Kozac's sequence (yon Heijne, 1987) and is in a similar region to other Notch gene start sites, but we cannot unequivocally exclude the possibility that the Notch 3 open reading frame includes D N A sequence 5' to that which has been characterised. The Notch 3 cDNA sequence will appear in the EMBL DNA database under accession number X74760.

126

M. Lardelli et al. ' Mechanisms ~I Development 46 (1994) 123-136

the genes from eDNA (Lardelli and Lendahl, 1993). Using the same set of primers we also identified a 916 bp fragment derived from a third mouse Notch homologue, Notch 3 (Fig. IA). This fragment, which corresponds to amino acids 1094-1398, was used as a probe to screen a mouse genomic library and a number of clones were identified. After establishing a restriction map for the locus, particular restriction fragments were subcloned and partially sequenced to locate reading frames related to known Notch sequences. We tested whether these sequences were transcribed by performing PCR on eDNA from E12.5 embryos using primers corresponding to the candidate open reading frames. When successful, the same primers were then used to isolate eDNA clones from mouse prepubescent testis and El 1.5 whole embryo eDNA libraries using a modified screening procedure (Lardelli and Lendahl, 1994). The organization of the eDNA clones is depicted in Fig. lB. 2.2. Structural Jeatures o[ the Notch 3 gene product Sequencing of the new eDNA clones produced a total of 7943 bp. Alignment of the predicted Notch 3 protein sequence with that of Drosophila Notch indicates that these cDNAs probably span the entire reading frame (Fig. 1A). A comparison of the Notch 3 protein sequence with those of rat Notch 1 and 2 shows that Notch 3 shares the key features of all known Notch proteins: EGF-repeats, Notch/lin-12 repeats, a putative transmembrane domain, cdclO/SWl6 repeats and the Cterminal PEST domain that is thought to be important for protein stability (Rogers et al., 1986). Extracellularly, Notch 3 possesses only 34 EGF-repeats, in contrast to the 36 repeats found in all other Notch proteins. The high degree of conservation of Notch EGF-repeats

,-1 <

"

:z

E

.m

°,,,=

=,~ L

23-9.4-6.6m 4.4--

2 . 3 ~

2.0m

B L

E

BS

h

L

J

X

E

S

E

B

S

S

B

,

L

i

~

i

i,

I

i

Drosophila Notch

lllllllllllllllllllllllllllllllllllllm I I I "F-T'-49% 6 6 * 511* Notch 3 I I I |IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIBI[ I:.l:.l::I:.l I I LI

1,'1

Fig~ I. (B) The Drosophila Notch gene and the Notch 3 gene are schematically depicted, and the amino acid identity in the EGFregions, the cdclO/SWl6 region and the C-terminal region indicated. At the bottom the arrangement of the analysed eDNA clones from the Notch 3 gene is shown. The open rectangle denotes the region that was analysed by RT-PCR.

aa 1 7 8 4

aa 1 6 5 8

probe

aa 9 3 6

aa 8 0 0

Fig. 2. Low stringency zoo-blot using a Notch 3 probe. The Co1732 Notch 3 eDNA clone was used in low stringency Southern hybridisation on a DNA blot of Eco Rl-cleaved genomic DNA from a number of species. The hybridising DNA band at 4.9 kb in DNA from a BALB/c mouse is from the Notch 3 locus. The > 10 kb, 2.6 kb, and 1.8 kb bands represent hybridisation to loci other than Notch 3. DNA fragments of known length were run in parallel and the sizes (kb) are indicated to the left. A parallel experiment with mouse strain 129/Sv DNA produced a similar set of bands (data not shown). Below is shown a restriction map of part of the mouse Notch 3 locus (clone #29, a bacteriophage ~, EMBL 3 clone of genomic mouse strain 129/Sv DNA). Barn HI (B), Eco RI (E), Sac 1 (S) and Xba 1 (X) restriction enzyme sites are marked. Sizes are ± 5%. Arrows indicate the approximate positions of codons for particular amino acid residues (aa) in the putative Notch 3 peptide sequence. Of the 9166 pb Co1732 eDNA sequence 866 bp lie within the 4.9 kb Eco RI fragment indicated.

M. Lardelli et al. / Mechanisms 0[" Deveh)pment 46 (1994) 123-136

makes it possible to identify the regions of homology that are missing. In one case, the missing region corresponds to a near-perfect deletion of EGF-repeat 21, whereas the other region corresponds to a loss of parts of EGF-repeats 2 and 3, together equalling one EGFrepeat (Fig. 1). Thus, in both cases, the net result is a loss of approximately the same size but positioned differently. RT-PCR tests on RNA from E8.5 and E12.5 embryos revealed no evidence for alternative splicing in

A

s •

9.5 7.5

4.4

'

2.4

'

1.35

800

400

the regions of missing EGF repeat homology (data not shown). We also sequenced the genomic DNA encompassing these regions, lntrons dividing the Notch 3 reading frame were found in the codons for amino acids 115 and 767, respectively (Fig. IA). The intron in codon 115 is only 89 bp, which is too small to encode an EGFrepeat, and the 569 bp intron in codon 767 does not contain any EGF-repeat related sequences (data not shown). These data rule out the possibility that the lack of the two EGF-repeats in the Notch 3 gene is a cloning artefact. The overall degree of amino acid identity of the remaining 34 EGF-repeats is 56% with Notch 1, 55% with Notch 2, and 49% with Drosophila Notch (Fig. 1B). The three Notch/lin-12 repeats show 55, 56, and 45% amino acid identity, respectively, with the corresponding Drosophila Notch/lin-12 repeats. In common with Notch 1 and 2, the most conserved region of Notch 3 is the intracellular cdclO/SWl6 domain. In this domain, identity with Notch I and 2 and Drosophila Notch is 77, 79, and 66%, respectively. The C-terminal PEST sequence is also relatively highly conserved, but the region separating these two domains is conserved to a much lower degree. Although this region is of somewhat variable length in all the characterised Notch genes, it is considerably shorter in Notch 3 (Fig. 1). The opa repeats found in this region of the Drosophila gene are not found in Notch 2 or 3, while the small groups of glutamine codons in Notch 1 may represent relics of an opa repeat for which there is no longer any selective pressure (see Discussion). In the region between the cdclO/SWl6 and the PEST sequences, database searches revealed only low and possibly insignificant homologies to other genes such as the Drosophila gene stauJen. 2.3. Notch 3-related sequences in mouse and other species A probe corresponding to Notch 3 EGF-repeats 28-34 was hybridised to Eco Rl-digested genomic DNA from a variety of species (zoo-blot). Under low stringency conditions, hybridising DNA sequences could be detected in all mammalian species analysed but not in avian species (Fig. 2). Two strong and two weaker bands

B

C

127

adult brain m N3 N1 N3 N1 N3 N1 N 3 N 1 E8.5

E12.5

E17.5

m

Fig. 3. Analysis of m R N A from different tissues probed for Notch 3 expression. (A) The Col732 Notch 3 cDNA clone was used to probe a northern blot of polyA + RNA from a variety of mouse tissues. The probe hybridises to an approximately 9 kb m R N A species in heart, brain, lung, liver, skeletal muscle kidney and testis. R N A fragments of known length were run in parallel and the sizes (kb) are indicated to the left. (B) Rehybridisation of the same filter with an actin probe produced bands in all lanes (1.8 kb). The lower molecular weight band (1.4 kb) in heart and muscle reflects hybridisation to a muscle specific form of actin. (C) PCR analysis of Notch expression. The RT-PCR amplified products from the Notch 1 and 3 genes, 660 and 467 bp, respectively, were separated in agarose gels. The developmental stages and tissues analysed and the PCR primer combinations (Notch 1 = N I; Notch 3 = N3) are shown at the top. A 100 bp DNA marker ladder (m) was run in parallel and sizes (bp) are indicated to the left.

128

M. Lardelli et al. / Mechanisms ~1 Development 46 ~1994) 123-136

were identified in mouse, only one of which (4.9 kb) could be accounted for by Eco Rl-restriction in the Notch 3 locus (Fig. 2). Rehybridisation of the zoo-blot with probes from the homologous regions of the Notch 1 and 2 genes did not account for the other three bands, even at low stringency conditions (data not shown). It thus appears that the mouse genome contains sequences more closely related to the EGF-repeats of Notch 3 than to those of Notch 1 and 2, thereby raising the possibility

/

of the presence of additional, yet unidentified, members of the gene family. However, the observation of only one Notch 3 hybridising band in humans suggests that this may not be the general case in all mammals (Fig. 2). 2.4. Spatial and temporal expression ~[ the mouse Notch gene family in the nervous system Poly(A) + RNA from various adult tissues was examined for the presence of Notch 3 transcripts by northern

M. Lardelli et al./ Mechanisms of Development 46 (1994) 123-136

129

/

;

j

k //

I

t 1MM

J

Fig. 4. Notch gene expression patterns in mouse nervous system at El 1.5. Dark-field photomicrographs of in situ transcript hybridisation using probes for Notch 1 (A), 2 (B), and 3 (C) transcripts on transverse sections through the brain vesicles. Notch 3 is expressed at high levels in the neuroepithelium (n) of the mesencephalic vesicle (mv) and the fourth ventricle (t3 and at lower levels in the otocysts (o). D - F Sections through the spinal cord show the mantle layer (m), ventricular zone (v) and central canal (c) of the neural tube with the flanking dorsal root ganglia (drg) and perineurium (p). D, E, and F show'sections probed for Notch 1.2, and 3, respectively, in dark-field. Notch 3 is expressed at high levels in the ventricular zone and at lower levels in the dorsal root ganglia and perineurium. G is the same section as in C, and H is the same section as in F, but seen under bright-field optics, l represents a schematic depiction of an Eli.5 embryo and shows the regions from which the sections in A-H were produced. The size bars correspond to 400/~m.

blot analysis and a single hybridising species of approximately 9 kb in length was identified in heart, brain, lung, liver, skeletal muscle, kidney and testis (Fig. 3A). The similar sized transcript argues against extensive transcript variation in various tissues, and RT-PCRs covering various Notch 3 regions from different tissues and developmental stages have not revealed evidence for differential splicing (data not shown). A strong hybridisation signal was observed in all lanes when the northern filter was rehybridised with an actin probe (Fig. 3B). In RT-PCR experiments on total RNA from various timepoints during development, we detect Notch 3 RNA from at least E8.5 into adulthood (Fig. 3C). Although the Drosophila Notch gene is known to act at many timepoints in development and in many tissues, the best characterised function is the effect on early nervous system formation. We therefore decided to analyse the temporal and spatial expression patterns of the three mouse Notch genes at various critical stages of CNS development by in situ transcript hybridisation. The stages (El 1.5, E15.5, P0, P5, and adult) were selected to represent regions of both proliferating, undifferentiated CNS progenitor cells and of fully differentiated neurons

and glial cells (Jacobson, 1991). To eliminate the possibility of cross-hybridisation between our probes and the different mouse Notch gene transcripts, we used 48-mer antisense oligonucleotides from regions of the three Notch genes that show relatively low inter-gene homology. Staging and nomenclature are according to Kaufman (1992). 2.5. E l l . 5

At El 1.5, the neural tube is entirely closed and the CNS progenitor cells of the neuroepithelium are actively proliferating in most regions (Jacobson, 1991). The rostral and caudal aspects have differentiated to various extents and are therefore both included in the study. The neuroepithelium consists of a simple, pseudostratified layer, in the rostral aspects. Fig. 4 shows transverse sections through the brain vesicles hybridised with the three probes. Notch 3 RNA (Fig. 4C) is found at high levels in the neuroepithelium of the mesenscephalic vesicle and the fourth ventricle, and also in the otocyst. In contrast, Notch 1 RNA is found at low levels and Notch 2 RNA is undetectable in the same regions (Fig. 4A and B). In more caudal aspects of the spinal cord, cellular

130

M. Lardelli et al, ,' Mechanism,~ O/Development 46 ~ 1994) 123-136

differentiation has proceeded further, and the proliferating cells are found close to the lumen of the spinal cord in the ventricular (ependymal) zone, whereas more differentiated cells are found in the mantle layer. Dorsal root ganglia (drg) have formed outside the neural tube (Fig. 4 D - F and H). Notch 3 is again expressed in the ventricular zone, and also at lower levels in drg and surrounding perineurium (Fig. 4F). The Notch 1 gene is highly active in drg, and to some extent also in the ventricular zone (Fig. 4D), while Notch 2 displays an almost inverse pattern, i.e. low in drg and high in the perineurium (Fig. 4E). Amongst non-neural tissues, we observe Notch 1 and 3 expression in skin (data not shown).

of mesodermal cells in the nasal septum and surrounding mesodermal parts of the nasal cavity are labelled at birth (Fig. 6A). The distribution of the clusters outside the septum shows a tendency to bilateral symmetry. When analysed at higher magnification there is no overt morphological difference between the cells expressing Notch 1 and the surrounding non-expressing mesoderreal cells (data not shown). Notch 2 and 3 show expression patterns similar to those at E15.5, i.e. the majority of expressing cells are located in the olfactory epithelium, but two arrays of Notch 2 and 3 expressing cells can be seen in the mesoderm of the nasal septum (Fig. 6B,C). A low level of Notch 1 and 3 expression remains

2.6. E15.5

At E15.5, a major site of proliferation in the CNS is the ventricular zone of the developing telencephalon (Jacobson, 1991). Sagittal sections from the E15.5 lateral vesicle were hybridised with the three Notch probes (Fig. 5A-C). All three Notch genes are expressed at high levels in the ventricular zone, but only to a very minor extent in the subventricular, intermediate and marginal zones which contain differentiated neurons. In a transverse view over the third and telencephalic ventricles, a uniform level of hybridisation by the three probes is observed in all ventricular zones, and also in massa intermedia (Fig. 5E-G). At this stage, the developing choroid plexus expresses Notch 2 (Fig. 5F) (also observed by Weinmaster et al., 1992), but not the other two Notch genes (Fig. 5E and G). The olfactory system (which is unique in the sense that CNS progenitor cells in the olfactory epithelium persist into adulthood to produce new olfactory neurons, see Ressler et al., 1993) expresses the three Notch genes in distinct patterns (Fig. 5I-K). Notch 3 is expressed at high levels in the olfactory epithelium, and in certain regions of the underlying olfactory mesoderm, especially in the septum (Fig. 5K). The Notch 2 gene is expressed more uniformly in mesoderm and at lower levels in epithelium (Fig. 5J), whereas Notch 1 is expressed predominantly in the mesoderm (Fig. 51). Notch genes are expressed in complex patterns in many non-CNS regions in E15.5 embryos, and our findings for the Notch 1 and 2 genes are largely in agreement with previous reports (Franco del Amo et al., 1992; Reaume et al., 1992; Weinmaster et al., 1992; Kopan and Weintraub, 1993). Notch 3 expression is also observed in lung, kidney, and stomach (data not shown).

i i!;

iii ........

C

+

D

2.7. PO

When compared with the E15.5 stage, the Notch expression patterns in the olfactory system of a newborn (P0) mouse show certain interesting features. From the relatively uniform mesodermal hybridisation of the Notch 1 probe at E15.5 (Fig. 5I), only discrete dusters

iI~f

E

i

]

)

M. Lardelli et al. / Mechanisms of Development 46 (1994) 123-136

G

131

H i ¸¸

J

K

Fig. 5. Notch gene expression patterns in mouse CNS at E15.5. A - C : Sagittal sections through lateral vesicle hybridised with probes for Notch 1 (At, 2 (B), and 3 (C) transcripts. Expression of Notch 3 is observed in the ventricular zone (v). Much lower expression is seen in the subventricular (s), intermediate (i), and marginal (m) zones. D is the same section as in C, but seen under bright-field optics. The boxed region in D corresponds to C. E - G : Transverse sections through the third ventricle (th) and telencephalic ventricle (t). E, F and G show sections probed for Notch 1, 2 and 3, respectively. Note the expression of Notch 2 in choroid plexus (cp). Uniform levels of expression of Notch 3 are seen in all ventricular zones including the ganglionic eminence (striatum) (g) and in massa intermedia (mi). H is the same section as G, but seen under bright-field optics. The boxed region in H corresponds to G. ( I - K ) Transverse sections through the developing olfactory system and both eyes (e) hybridised with probes for Notch 1 (1), 2 (J), and 3 (K). Notch 3 is expressed at high levels in the olfactory epithelium (oe), in nasal septum (ns), and in other regions of the olfactory mesoderm (om). L is the same section as K, but viewed under bright-field optics. The boxed region in L corresponds to K. A - C , E - G and I - K are viewed under dark-field optics. The size bars correspond to 400 ,am.

132

M. Lardelli et al. ' Mechanisms ~I Development 46 (1994) 123-136

in the ventricular zones in the brain, and Notch 2 is expressed at high levels in choroid plexus (data not shown). An area, although not of CNS origin, which displays strikingly different patterns of Notch gene expression in the P0 animal is that of the developing teeth. The Notch 1 gene is expressed at low levels in dental epithelium (Fig. 6A), while Notch 2 is expressed at very high levels in the epithelium covering the dental mesenchyme (Fig. 6B). Notch 3 is expressed in a small region facing the oral cavity (Fig. 6C). 2.8. P5

Whereas a good correlation is seen between domains of expression of the Notch genes (especially Notch 3) and the presence of progenitor cells during CNS development in the embryo (i.e. in the neural tube, telencephalic

ventricular zone, and olfactory epithelium), Notch genes are not expressed at detectable levels in regions of cell proliferation in the P5 CNS. Thus, we fail to observe hybridisation to the external granular layer of the cerebellum or to the progenitor cells of the hippocampal granule cells (data not shown). The expression of all three genes appears to be down-regulated, with the exception of continued Notch 2 expression in choroid plexus (data not shown). 2.9. adult

The expression of Notch genes in adult mouse brain is very similar to that at P5. In sagittal and transverse sections only Notch 2 expression in choroid plexus is detectable. The choroid plexus hybridisation is seen in a sagittal section in Fig. 7, and is in keeping with previous observations (Weinmaster et al., 1992). The lack of

+

Fig. 6. Olfactory and dental expression patterns of mouse Notch genes at P0. A-C: Coronal sections through the olfactory system and developing teeth hybridised with probes for Notch 1 (A), 2 (B), and 3 (C) transcripts. Notch 1 expression is seen in discrete clusters of cells in the nasal septum (ns) and olfactory mesoderm (ore) surrounding the nasal cavity (nc). The pattern is, at least partially, bilaterally symmetrical around the nasal septurn (arrow heads). Notch 3 is expressed in the olfactory epithelium (oe) and in two cellular arrays in the mesoderm of the nasal septum (arrow heads). Notch 2 (B) is also expressed in epithelium (m) overlying the dental mesenchyme (tp), while Notch 3 (C) is expressed in a cell population facing the oral cavity (small arrow). A - C are viewed under dark-field optics, while D is the same section as C but viewed under bright-field optics. The boxed region in D corresponds to C. The size bars correspond to 400 ~m.

M. Lardelli et al. / Mechanisms of Development 46 (1994) 123-136

133

A .......

Cp

g .....

Fig. 7. Expression pattern of Notch 2 in the choroid plexus of adult mice. Dark-field (A) and bright-field (B) views of a sagittal section through the choroid plexus (cp) and hippocampus of an adult mouse hybridised with a probe for Notch 2 transcripts. Expression can be detected in the choroid plexus, but not in hippocampus. Notch 1 and 3 do not show any expression at this stage (data not shown), The dentate gyrus (dg) of the hippocampus is indicated. The size bar corresponds to 400 ~m,

detectable levels of Notch 3 RNA by in situ hybridisation contrasts with our findings of low levels of Notch 3 RNA from adult rat brain in the northern blot experiment (Fig. 3). This discrepancy may be explained by the presence of local concentrations of Notch 3 expressing cells which were missed in the in situ analysis, or by very low levels of expression that might be undetectable by our in situ hybridisation assay. In any case, it is apparent that all three Notch homologues are expressed predominantly early in CNS development, and that Notch 3 expression correlates with proliferating CNS progenitor cells during these stages. 3. Discussion

3.1. Notch 3 - - a bona fide Notch homologue with distinct structural features The data presented here show that the mammalian genome harbours and expresses a third homologue of the Drosophila Notch gene, Notch 3. Based on its overall structure, and its level of homology with the previously described vertebrate Notch 1 and 2 genes, Notch 3 may be regarded as a true member of the mammalian Notch gene family. This is in contrast to the mouse int-3 gene, which while similar in structure, represents a more distantly related member of a Notch gene super-family (Robbins et al., 1992; Weinmaster et al., 1992). The evolution of the Notch 3 gene and the other mammalian Notch homologues is remarkable, in that all the subdomains of each gene show strong homology to the Drosophila gene. This implies strong selective pressure for maintenance of the overall structure, rather than only particular domains as is commonly found (Kessel and Gruss, 1990). We previously proposed that these separate Notch gene lineages arose at a time between the divergence of arthropods and amphibia from the mammalian lineage (Lardelli and Lendahl, 1993), based on

our characterisation of mouse Notch 1 and 2. Assuming that Notch 3 is evolving at a similar rate to Notch 1 and 2, it is reasonable to infer that the duplication event that generated this gene occurred at about the same time. During evolution the Notch 3 gene has acquired certain distinct structural features: it lacks sequences corresponding to two EGF-repeat-sized regions and it encodes a considerably shorter C-terminal domain. The absence of two EGF-repeat-sized stretches of sequence spanning EGF-repeat 21 and the boundary between EGF-repeats 2 and 3 represents the greatest structural variation yet seen in a member of the Notch gene family. Two conclusions can be drawn from this. First, the two regions have most likely been deleted during Notch 3 evolution, since all other characterised Notch genes carry the repeats. Second, in both cases near-perfect EGF-repeats were deleted, although in one case not in register with the standard definition of an EGF-repeat. This suggests that the EGF-repeat is an important functional module, and there are several lines of evidence supporting that individual EGF-repeats are involved in protein-protein interaction and ligand binding. First, EGF-repeats 11 and 12 in the Drosophila Notch protein are essential and sufficient for binding to the putative ligand Delta in a cell aggregation assay (Rebay et al., 1991 ). Second, the Drosophila patterning mutations split and Abruptex have been shown to be point mutations in Notch EGF-repeat sequences (Hartley et al., 1987; Kelley et al., 1987) and may affect ligand binding as well as the postulated dimerisation occuring between Notch molecules residing in the same cell membrane (Kelley et al., 1987). Third, individual EGF-repeats have been shown to be important for protein-protein interactions in other systems, such as the binding of nidogen to laminin (Mayer et al., 1993). Thus, the missing EGF-repeats of Notch 3 suggest that its ligand-binding specificity differs from that of Notch 1 and 2 (see below). Unfor-

134

M. Lardelli et al./ Mechanisms of Development 46 (1994) 123-136

tunately, the lack of Drosophila mutants mapping to EGF-repeats 2, 3 or 21 precludes a direct insight into their functional role. Notch 3 also encodes an intracellular domain considerably shorter than those of Notch 1 or 2. This may reflect a different intracellular function of the Notch 3 protein or may simply represent a lack of selective pressure for the deleted sequences. Noticably absent from the intracellular domains of all members of the mammalian Notch family is an opa repeat region consisting of long stretches of glutamine and/or histidine residues (Wharton et ai., 1985b). opa repeat motifs are common among developmentally regulated Drosophila proteins and have also been found in mammalian proteins under cell-type specific and developmental controk although their function is still a matter of debate (Grabowski et al., 1991). The Notch 1 gene encodes some small groups of glutamine residues which may represent remnants of an opa repeat region. In general however, it appears that the mammalian Notch genes no longer require an opa function. A precedent for size variations in polyglutamine stretches are the genes for Huntington's disease and spinobulbar muscular atrophy, in which expansions of glutamine tracks, produced by trinucleotide repeat expansion, correlate with the disease (for review see Mandel, 1993). The intracellular cdclO/SWI6 repeats are, in contrast, the region of greatest conservation among the Notch genes - - the characterisation of Notch 3 further illustrates this fact. This focuses attention on the cdclO/SWI6 region as pivotal for the postulated intracellular signalling of Notch. cdclO/SWl6 repeat-like sequences were originally discovered in the yeast cdclO (Aves et al., 1985) and SWI6 genes (Breeden and Nasmyth, 1987), which are important for cell cycle control. Similar sequences have also been found in the ankyrin (Lux et al., 1990) and bcl-3 genes (Ohno et al., 1990). Recently, more direct evidence for the involvement of cdclO/SWl6 repeats in Notch function was produced by functional analysis of mutated intracellular domains in Drosophila (Fortini et al., 1993). In addition, Kidd (1993) showed that the protein encoded by the Drosophila cactus gene contains cdclO/SWl6 repeats, which are essential for binding to the dorsal protein, a transcription factor involved in dorso-ventral patterning.

3.2. Notch 3 is abundantly expressed in areas of proliferating neuroepithelium The effects of the Drosophila Notch gene are pleitropic (Artavanis-Tsakonas and Simpson, 1991) and this is probably also the case in mammals, as the three Notch homologues show complex patterns of expression in many tissues. Effects on neural/epidermal fate choice in the embryo and during subsequent steps in nervous system development (Cagan and Ready, 1989; Giniger

et al., 1993) are some of the best characterised of the various Drosophila Notch functions that have been studied. We therefore analysed the Notch 1, 2, and 3 expression patterns during mouse CNS development, to see if expression domains were associated with particular cell types. All three genes, particularly Notch 3, are predominantly expressed in regions of undifferentiated, proliferating CNS progenitor cells in the prenatal mouse. The strong correlation of expression with undifferentiated cells is in keeping with a postulated role for mammalian Notch genes in the selection of developmental fates. The choice for proliferating CNS progenitor cells may be between remaining in the proliferative state or becoming a post-mitotic neuron (Jacobson, 1991; McKay, 1989; Price, 1989). Experimental evidence that undifferentiated early CNS cells may respond to a Notch signal was recently provided by Coffman et al., (1993), who showed that expression in Xenopus embryos of a Xenopus Notch 1 gene deprived of the region encoding the extracellular domain produced neural hyperplasia and an expanded neural tube. The authors suggested that this was a result of delaying cell determination, thus allowing excessive proliferation before the cells finally differentiated (Coffman et al., 1993). These data are compatible with the observed complex expression pattern of the Notch genes in the developing, but not mature CNS. The discovery of a third Notch homologue highlights the question of why vertebrates possess multiple Notch genes when only a single Notch gene is found in Drosophila. One possibility is that the vertebrate Notch genes encode biochemically equivalent proteins but that their different patterns of transcription provide more intricate patterns of Notch function. A precedent for this could be the case of the Notch-related genes lin-12 and glp-1 of Caenorhabditis elegans that are responsible for fate choices in different cells and have been suggested to be biochemically interchangeable (Lambie and Kimble, 1991; Mango et al., 1991). Alternatively, the different vertebrate Notch genes may encode proteins that possess different ligandbinding and/or intracellular signal transmission specificities. The obvious structural differences between Notch 3 and the other Notch homologues imply that, at least for this gene, the latter hypothesis is correct. If vertebrate Notch proteins do dimerise, as appears to be the case in Drosophila (Kidd et al., 1989), then the different proteins may possess different tendencies for homo- or heterodimerisation when present in the same plasma membrane. The formation of heterodimers between non-equivalent Notch proteins could provide a cell with a more sophisticated repertoire of signalling information. Our observation that the expression patterns of Notch 1, 2 and 3 overlap in a variety of combinations suggests that such ideas may be important in future models of Notch and neurogenic gene function.

M. Lardelli et al. /Mechanisms of Development 46 (1994) 123-136 3.3. Are Notch genes establishing zones of gene expression in the CNS? If mammalian Notch genes do indeed act during fate decisions, this may be reflected in the expression of specific genes in the resulting cell type. In this regard, the similarity between the expression patterns of Notch 1 and genes for olfactory receptors is quite intriguing. The expression of Notch 1 in olfactory mesoderm undergoes a developmental change from a relatively uniform pattern at E15.5 to a pattern of discrete islands of expression. This pattern is strikingly similar to the localised expression zones of olfactory receptor genes which are distributed in a bilaterally symmetric fashion in the mouse olfactory epithelium (Nef et al., 1992; Strotmann et al., 1992; Ressler et al., 1993). It is thus possible that the localised expression of Notch 1 in the underlying mesoderm is in some way involved in organizing or maintaining zones of expression of the large number of olfactory receptor genes. In conclusion, the characterisation of the Notch 3 gene and the analysis of its expression pattern relative to Notch 1 and 2 during brain development have demonstrated a greater complexity for the Notch gene family in mammals. The unique structure of Notch 3 is the first strong indication of functional differences between different family members and will be important to consider in any models of Notch gene action. The complex overlapping expression patterns of all three Notch genes raise the possibility of combinatorial Notch homologue action during the various cell fate decisions and complex morphological movements in the developing CNS and other regions of the vertebrate embryo.

4. Experimental procedures 4.1. Isolation and sequencing of Notch 3 genomic and cDNA clones A 916-bp mouse Notch 3 cDNA fragment was amplified by PCR from 12.5 days post coitum (E12.5) mouse embryo cDNA (RT-PCR) using degenerate primers corresponding to the Drosophila and Xenopus Notch genes, as previously described (Lardelli and Lendahl, 1993). The fragment was cloned into Bluescript KS and the clone (Co1732) was used to screen a 129/Sv mouse genomic cDNA library in the bacteriophage h EMBL3 vector (Stratagene) by standard procedures (Sambrook et al., 1989). A number of positive clones were isolated and subclones of particular restriction fragments from these bacteriophages were sequenced at their termini. PCR primers were then designed from any potential Notch homologous reading frames revealed. These primers were used in PCR reactions with E12.5 mouse embryo cDNA. Primers producing PCR products of the expected size were used in a modified screening procedure (Lardelli and Lendahl, 1994) to isolate cDNA clones from an oligo-dT-primed Ell.5 mouse

135

embryo cDNA library (Clontech) and from a randomly primed cDNA library of prepubescent mouse testis mRNA. cDNA clones covering all coding regions except for the DNA sequence encoding amino acids 1384-1520 were isolated (Fig. 1). The uncloned region was sequenced from three independent RT-PCR isolates from mouse embryo cDNA as previously described (Lardelli and Lendahl, 1993). The PCR primers used to isolate this region were: 5'-CTA CCT GTA CAG AGT GTC3' and 5'-TGA GTC GCT GCA GAA AGT-3'. To isolate intron sequences in the regions of missing EGFrepeat homology, we used genomic DNA in PCR with the following primers: EGF-repeat region 21: 5'-CAG GAA CAG ACA GTC AGC-3'; 5'-TTC CGC TGT GTT TGT GAG-3'; EGF-repeat regions 2/3: 5'-TTG CCA GAG TTC AGT GGT-3'; 5'-GTC ACT TTG GCA GCT TTG-3'. Comparison of Notch 3 DNA and protein sequences with those of other Notch genes is as previously described (Lardeili and Lendahl, 1993b). 4.2. Southern and northern blot analyses The Co1732 cDNA clone that encodes amino acids 1094-1398 was 3zp-labeled by random priming to a specific activity of 1-2 x 10 9 counts/min/t~g DNA and used as a Notch 3 probe in both Southern and northern blot hybridisation analyses (Sambrook et al., 1989). The zoo-blot (8 ~g DNA in each lane) and the mouse adult multi-tissue northern blot (5 #g polyA + RNA in each lane) were obtained from a commercial supplier (Clontech). Hybridisation and washing conditions have been described previously (Lardelli and Lendahl, 1993). The integrity of the RNA in the northern blot was tested by rehybridisation with an actin probe, according to the manufacturer's instructions (Clontech). Low stringency Southern blot hybridisation was performed as described (Lardelli and Lendahl, 1993) and the washing conditions were 3 x 20 min at 65°C in 6x SSC, 0.1% SDS. 4.3. In situ transcript hybridisation In situ hybridisation was performed on sections from mouse embryos and postnatal tissue using 48-mer antisense oligonucleotides as probes. The tissue was prepared and sectioned (14 #m sections) as previously described (Ibfinez et al., 1993). The probes (mouse Notch 1: 5'-TGT CCC TCG GTG AGC TGG CAG TCG AAG CCA TCA AAG AGG CAG CCG GCC-3'; mouse Notch 2: 5'-CGT CTT GCT ATT CCT CTG GCA CTC AAA GTT GTC AAA CAG GCA CTC TGC-3'; mouse Notch 3: 5'-AAG AAA GGG ACA CTC TGT ACA GGT AGG CAC GAG CCC CCA TGC AGA CAA-3') were labeled at the 3' end with a35S-dATP using terminal transferase (IBI) to a specific activity of approximately 1 × 10 9 counts/min//~g DNA. Hybridisation (Ernfors et al., 1992) and dipping (lb~inez et al., 1993) were performed as previously described. The sections were exposed to photoemulsion for 3-6

136

M. Lardelli et al. / Mechanisms o[ Development 46 (1994) 123-136

weeks. M i c e were h o u s e d a n d c a r e d for in c o m p l i a n c e with N a t i o n a l I n s t i t u t e o f H e a l t h guidelines for the c a r e o f l a b o r a t o r y animals.

4.4. P C R tests f o r Notch 3 expression E x p r e s s i o n o f the Notch 3 g e n e was a n a l y s e d by R T P C R , s t a r t i n g f r o m total R N A f r o m a d u l t testis a n d emb r y o s as p r e v i o u s l y d e s c r i b e d ( L a r d e l l i a n d L e n d a h l , 1993). P C R p r i m e r s for Notch 3 a m p l i f i c a t i o n w e r e 5 ' A C A C T G G G A G T T C T C T G T - 3 ' and 5 ' - G T C T G C T G G C A T G G G A T A - 3 ' . P r i m e r s for the Notch 1 gene have been d e s c r i b e d p r e v i o u s l y ( L a r d e l l i a n d L e n d a h l , 1993). T h e P C R p r o d u c t s w e r e e l e c t r o p h o r e s e d beside a 100-bp D N A m a r k e r l a d d e r ( P h a r m a c i a , Cat. N o . 274001-01 ) in 2.5"/0 a g a r o s e gels a n d visualised by e t h i d i u m b r o m i d e staining ( S a m b r o o k et al., 1989).

Acknowledgements W e t h a n k Erik N i l s s o n for excellent t e c h n i c a l assistance, Dr. C h r i s t e r H 6 6 g for the r a n d o m p r i m e d m o u s e testis c D N A library a n d M a r i e - L o u i s e A l u n and M o n a S c h r 6 d e r for help in the a n i m a l facility. T h i s w o r k was s u p p o r t e d by the S w e d i s h C a n c e r Society, M a r g a r e t o c h Axel A x : s o n J o h n s o n s Stiftelse, Kjell o c h Mfirta Beijers Stiftelse, K n u t och A l i c e W a l l e n b e r g s Stiftelse, M a g n . Bergvalls Stiftelse, a n d K a r o l i n s k a institutets fonder. M . L . was s u p p o r t e d by a l o n g - t e r m E M B O fellowship.

References Artavanis-Tsakonas, S. and Simpson, P. (1991) Trends Genet. 7, 403 -408. Aves, S.J., Durkacz, B.W., Carr, A. and Nurse, P. (1985) EMBO J. 4, 457-463. Breeden, L. and Nasmyth, K. (1987) Nature 329, 651-654. Cabrera, C.V. (1992) Development 115, 893-901. Cagan, R.L. and Ready, D.F. (1989) Genes Dev. 3, 1099-1112. Coffman, C., Harris, W. and Kintner, C. (1990) Science 249, 1438-1441. Coffman, CR., Skoglund, P., Harris, W.A. and Kintner, C.R. (1993) Cell 73, 659-671. Devereux, J., Haeberli, P. and Smithies, O. (1984) Nucleic Acids Res. 12, 387-395. Ellisen, LW., Bird, J., West, D.C., Soreng, A.L., Reynolds, T.C., Smith, S.D. and Sklar, J. (1991) Cell 66, 649-661. Ernfors, P., Merlio, J.-P. and Persson, H. (1992) Eur. J. Neurosci. 4, 1140-1158. Fortini, M.E., Rebay, I., Carom LA and Artavanis-Tsakonis, S. (1993) Nature 365, 555-557. Franco del Amo, F., Smith, D.E., Swiatek, P.J., Gendron-Maguire, M., Greenspan, R.J., McMahon, A.P. and Gridley, T. (1992) Development 115, 737-744. Giniger, E., L.Y., J. and Jan, Y.N. (1993) Development 117, 43 l--440. Grabowski, D.T., Carney, J.P. and Kelley, M.R. (1991) Nucleic Acids Res. 19, 1709. Greenwald, I. and Rubin, G.M. (1992) Cell 68, 271-281. Hartenstein, A.Y., Rugendorff, A., Tepass, U. and Hartenstein, V. (1992) Development 116, 1203-1220. Hartley, D.A., Xu, T. and Artavanis-Tsakonas, S. (1987) EMBO J. 6, 3407-3417. Heitzler, P. and Simpson, P. (1991) Cell 64, 1083-1092.

Ibfinez, C.F., Ernfors, P., Timmusk, T., Ip, N.Y., Arenas, E., Yancopoulos, G.D. and Persson, H. (1993) Development 117, 1345-1353. Jacobson, M. (1991) Developmental Neurobiology (Third edition). Plenum Publishing Corp., New York, pp. 776. Jhappan, C., Gallahan, D., Stahle, C., Chu, E., Smith, G.H., Merlino, G. and Callahan, R. (1992) Genes Dev. 6, 345-355. Kaufman, M.H. (1992)q'he Atlas of Mouse Development. Academic Press Inc., London, 512 pp. Kelley, M.R., Kidd, S., Deutsch W.A. and Young, M.W. (1987) Cell 51, 539-548. Kessel, M. and Gruss, P. (1990) Science 249, 374-379. Kidd, S. (1993) Cell 71, 623-635. Kidd, S., Baylies, M.K., Gasic, G.P. and Young, M.W. (1989) Genes Dev. 3, 1113-1129. Kid& S., Kelley, M.R. and Young, M.W. (1986) Mol. Cell. Biol. 6, 3094- 3108. Kopan, R. and Weintraub, H. (1993) J. Cell Biol. 1211 631-641. Lambie, E.J. and Kimble, J. (1991) Development 112, 231-240. Lardelli, M. and Lendahl, U. (1994) BioTechniques 16, 420-422. Lardelli, M. and Lendahl, U. (1993) Exp. Cell Res. 204, 364-372. Lux, S.E., John, K.M. and Bennett, V. (1990) Nature 344, 36-42. Mandel, J.-L. (1993) Nature Genet. 4, 8-9. Mango, S.E., Maine, E.M. and Kimble, J. (1991) Nature 352, 811-815. Mayer, U., Nischt, R., P6schl, E., Mann, K., Fukuda, K., Gerl, M., Yamada, Y. and Timpl, R. (1993) EMBO J. 12, 1879-1885. McKay, R. (1989) Cell 58, 815-821. Nef, P., Hermans-Borgmeyer, 1., Arti6res-Pin, H., Beasley, L., Dionne, V.E. and Heinemann, S.F. (1992) Proc. Natl. Acad. Sci. USA 89,8948-8952. Ohno, H., Takimoto, G. and McKeihan, T.W. (1990) Cell 60, 991-997. Price, J. (1989) Trends Neurosci. 12, 276-278. Reaume, A.G., Conlon, R.A., Zirngibl, R., Yamaguchi, T.P. and Rossant, J. (1992) Dev. Biol. 154, 377-387. Rebay, I., Fehon, R.G. and Artavanis-Tsakonis, S. (1993) Cell 74, 319-329. Rebay, I., Fleming, R.J., Fehon, R.G., Cherbas, L., Cherbas, P. and Artavanis-Tsakonis, S. ( 1991 ) Cell 67, 687-699. Ressler, K.J., Sullivan, S.L. and Buck, L.B. (1993) Cell 73, 597-609. Robbins, J., Blondel, BJ., Gallahan, D. and Callahan, R. (1992) J. Virol. 66, 2594-2599. Rogers, S., Wells, R. and Rechsteiner, M. (19861 Science 234, 364-368. Ruel, L., Bourouis, M., Heitzler, P., Pantesco, V. and Simpson, P. (1993) Nature 362, 557-560. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual (second edition). Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Stifani, S., Blaumueller, C.M., Redhead, N.J., Hill, R.E. and Artavanis-Tsakonis, S. (1992) Nature Genet 2, 119-127. Strotmann. J., Wanner, 1., Krieger, J., Raming, K. and Breer, H. (1992) Neuroreport 3, 1053-1056. Struhl, G., Fitzgerald, K. and Greenwald, 1. (1993) Cell 74, 331-345. von Heijne, G. (1987) Sequence Analysis in Molecular Biology. Academic Press Inc., London, 188 pp. Weinmaster, G., Roberts, V.J. and Lemke, G. (1991) Development 113, 199-205. Weinmaster, G., Roberts, V.J. and Lemke, G. (1992) Development 116, 931-941. Wharton, K.A., Johansen, K.M., Xu, T. and Artavanis-Tsakonas, S. (1985a) Cell 43, 567-581. Wharton, K.A., Yedvobnick, B., Finnerty, V.G. and ArtavanisTsakonas, S. (1985b) Cell 40, 55-62. Xu, T., Caron, L.A., Fehon, R.G. and Artavanis-Tsakonis, S. (1992) Development 115, 913-922.