embryo brain kinase: a novel gene of the eph/elk receptor tyrosine kinase family

ELSEVIER Mechanisms of Development 52 (1995) 319-341

Embryo brain kinase: a novel gene of the epWeZk receptor tyrosine kinase family

Jonathan Ellisa.1, Qiurong Liub,*, Martin Breitman b-w, Nancy A. Jenkiw, Debra J. Gilbertc, Neal G. Copelandc, Heidi V. Tempesta, Simon Warrena, Elizabeth Muira, Heather Schillingd,

Fred A. Fletcherd-3, Steven F. Ziegler(‘j4, John H. Roger+*

aDepartment of Physiology, University of Cambridge, Cambridge CB2 3EG, UK bSamuel Lunenfeld Research Institute, Mt. Sinai Hospital, Toronto, MSG 1X5, Canada

‘Mammalian Genetics Lab, ABL-Basic Research Program, NCI-Frederick Cancer Research and Development Center, Frederick, MD 21702, USA dlmmunex Corp., Seattle, WA 98101, USA

Received 28 November 1994; revision received 8 May 1995; accepted 15 May 1995

Abstract

A new gene belonging to the EpWEcWElk receptor tyrosine kinase family has been cloned from mouse brain. The gene maps to mouse chromosome 4. In the adult brain it is expressed exclusively and abundantly in the hippocampus. We propose to name it Ebk

(embryo brain kinase), as in situ hybridisation shows expression in many parts of the developing mouse brain. The most abundant expression is in the subcommissural organ, and the earliest expression is in the forebrain neural folds, in rhombomeres 2-6, and in somites and heart. Other regions positive at various stages include the cochlear duct, trigeminal ganglion, lung, first branchial arch, and tooth primordia. Also positive are areas of mesenchyme underlying various epithelia during morphogenesis, especially in the mouth and nose, as well as in the eyelids and toes. We compare these patterns with the available data on the 12 other known members of this gene family. Most of them, like Ebk, are expressed in brain (especially adult hippocampus and embryonic rhombomeres) and in organs rich in epithelia (especially lung), although the spatial and temporal patterns differ. We suggest that combinatorial patterns of these receptors act as labels for the regional identity of neurons and epithelia, and could mediate fine control of neurite pathfinding and epithelial morphogenesis.

Keywords: Receptor tyrosine kinases; Hippocampus; Neural development; Morphogenesis; Subcommissural organ; Rhombomeres; Lung; Heart; Branchial arches

1. Introduction

The number of known receptor tyrosine kinases is rapidly growing. They include many that are involved in

Q Martin Breitman passed away on 13 February 1994, and this paper

is dedicated to his memory.

* Corresponding author, Tel: +44 1223 333865; Fax: +44 1223

333840; E-mail: [email protected].

’ Present address: Molecular Immunology Group, The Wellcome

Foundation, Building 111, Langley Court, S. Eden Park Rd., Becken-

ham, Kent BR3 3BS, UK.

’ Present address: Amgen Inst., 620 University Ave., Suite 706, To-

ronto M5G 2C1, Canada.

3 Present address: Amgen Inc., Thousand Oaks, CA 91230-1789,

USA

cell fate decisions during development, including many in the nervous system. One example is W/kit in mammals, which is necessary for migration of several types of stem cell and is also expressed in the developing nervous sys-

tem (Geissler et al., 1988; Keshet et al., 1991). Another is sevenless in insects, which is the receptor for a signal in-

ducing differentiation into the R7 type of photoreceptor (Rubin, 1991). In the vertebrate brain, receptor tyrosine kinases include the Trk family of receptors for nerve growth factor and related neurotrophins (Chao, 1992; Meakin and Shooter, 1992; Davies, 1994), and the family of receptors for fibroblast growth factors, which can also have neurotrophic activity (Heuer et al., 1990; Wanaka et al., 1990; Lai and Lemke, 1991). In the peripheral nerv-

4 Present address: Darwin Molecular Corp., Bothell, WA 98021, ous system, control of glial cell proliferation and of syn- USA. aptogenesis involves the HER2INeu and HER41ErbB4

0925.4773/95/$09.50 0 1995 Elsevier Science Ireland Ltd. All rights reserved

SSDf 0925-4773(95)00411-S

320 J. Ellis et al. I Mechanisms of Development 52 (1995) 319-341

tyrosine kinases, which form receptors for the heregulin or glial growth factor family of ligands (Lemke, 1993;

Plowman et al., 1993). In the adult brain, tyrosine kinase activity is implicated in long-term potentiation of syn-

apses (O’Dell et al., 1991; Grant et al., 1992). The con- cept of receptor-linked tyrosine kinases has recently been

broadened by the discoveries that various families of cy- tosolic tyrosine kinases transduce signals from cytokine

receptors and from immunoglobulin-related cell surface molecules. So there is increasing evidence that tyrosine

kinases are implicated in signal transduction at many stages of nervous system development, including the maintenance of precursor cell proliferation, induction of neural differentiation, neurite adhesion, maintenance of neural survival in a competitive target environment, synapse formation, and synaptic modification in the adult

brain.

There are numerous genes homologous to receptor ty-

rosine kinases for which functions have not yet been

identified. These include many members of the Eph/EcWElk family, most of which are expressed in brain.

This family now contains 12 members (Hirai et al., 1987; Lai and Lemke, 1991; Sajjadi et al., 1991; Gilardi- Hebenstreit et al., 1992; Kuma et al., 1993; Maisonpierre et al., 1993; Sajjadi and Pasquale, 1993; Henkemeyer et

al., 1994; Zhou et al., 1994; see Discussion for details). They are closely homologous in the tyrosine kinase domain, and share a distinctive set of extracellular domains

including a cysteine-rich region and two fibronectin type

III domains, one of them being interrupted by an insert

with four cysteines (Fig. 1). We will refer to this as the

ELK (Eph-like kinase) family. Several members have been localised in the adult hippocampus (Lai and Lemke, 1991) and in embryonic rhombomeres (Nieto et al., 1992; Ganju et al., 1994; Henkemeyer et al., 1994) as well as in various epithelial organs in the adult or embryo (see Sec- tion 3 for details). However, no ligands had been identified for any of them until very recently, and no function

Cys-rich Fnlll Fnlll

has yet been identified for any of these putative receptors or their ligands.

In the search for genes for new receptor tyrosine kinases, the polymerase chain reaction (PCR) has proven to be an effective method in nervous tissue as well as elsewhere (e.g. Kamb et al., 1989; Wilks, 1989a,b; Lai and Lemke, 1991; Gilardi-Hebenstreit et al., 1992; Mar-

celle and Eichmann, 1992; Ziegler et al., 1993). To search for new receptor tyrosine kinases which might be involved in neural development, we have performed PCR on cDNA from rodent brains of different ages. Here we

report a gene initially obtained from adult mouse brain cDNA, which is also expressed in many areas undergoing neural differentiation or tissue morphogenesis in the embryo. This is a new member of the ELK family. In con- formity with names for other members of the family, and

in view of the expression patterns that we describe, we propose to name it Ebk for embryo brain kinase.

2. Results

2.1. Cloning and sequencing To look for new tyrosine kinases in brain, we per-

formed PCR on mouse brain cDNA, and cloned and sequenced the products. Almost all the clones were mem-

bers of kinase families. They included 12 clones homologous to serine/threonine kinases, 15 homologous to solu-

ble tyrosine kinases (8 clones of Jak2 and 2-3 each of Jakl, Abl, and Fyn), and 13 homologous to receptor tyrosine kinases. The latter included five clones very similar

to clone tyro-3 (Lai and Lemke, 1991), four identical to Flg (FGF receptor family), one identical to Flk2, and

three apparently new ones. One of these, clone TK29, was clearly a member of the ELK family.

A longer cDNA clone was obtained by plaque hybridisation of a mouse embryo cDNA library. The 2.1 kb insert contained coding sequence for 568 amino acids, followed by 388 nucleotides of 3’ untranslated sequence and

M Tyr Kinase

/ I I

Clone TK29.4 , . , I

. . I .

I I

I 7 I Probes N

Fig. 1. (Top) Consensus structure of the Eph/Elk family receptor tyrosine kinases. (Bottom) Scale map of the 2.1 kb insert in clone pTK29.4, and

probes used for in situ hybridisation (see Materials and methods for details).

J. Ellis et ~1. I Mechanisms ofDevelopment 52 (1995) 319-341 321

10 20 30

Tk29.P FAAVSI-AAPSQVS GVMKERVX?P.--‘Q ~~~~~~~~:}1111:~ :::!:I: ::I:

tiwnhek TVI’D- EIDAVNGVSELSSPPRQPASI~AAPSPVLTIKKDRTSRNSIS 430 440 450 460 470 480

40 50 60 70 00 90

Tk29. P MEPEHPNGVITEYE IKHYEKDQRERTYSTLKTKSTSASI NNLKPGTVYVPQIRAVTA

IIIIIIIIIII:I :~~:~:~~~:::~ :I::I::::I:::I::III:I:IIIIIII 11 Humhok mEPEHmJGIILDY EVKYYEKQEQETSYTILRARcTNVTISSLKPlZI~QIRARTA

490 500 510 520 530 540

100 110 120 130 140 150

Tk29.P AGYGNYSPR~WATLEEASGKMFEATAVSSEQNPVIIIAWAVAGTIIL~GFI~GR IllI: I:::: I:1 : 1 ::I:1 ::/::I[: I::::l:l: ::: :I

Humhek AGYGTNSRKFE?- ETSPDSF-SISGESSQ~SAAVAIIIJ.TWIWL.I--G 550 560 570 + 580 590

160 170 180 190 200 210

Tk29. P RHCGYSKADQEGDEELYF -HFIVPGTKTYIDPFPYEDPNRAVHQPAKELDASCIKIE

1 III:: : :::: !:I I:l:lI :11:11:11111::111:1111111: I:!:: Humhok WCGYXStMGADEKRLHFGNGHLKLPGLRTYVD PH’IYEDPTQAVHEFAKELDATNISXDK

600 610 620 630 640 650

I 220 230 240 250 260 270

Tk29.P VIGAGEFGEVCSGPUCLPG KRDVAVi+JKTLKVGYTEKQRRDFLCEASIMGQF~

‘I:lllll1lllllll1lI:I I

::::IIIIIIIIIIIIIIIIIII IIIIIIllllll:::I Humhek WGACIEFGLVCSGWUSSKKEISVAI~G~KQ~FVIEASFDHPNII~

660 670 680 690 700 710

280 290 300 310 320 330

Tk29 .P ‘EGVVTRGKPVMIVIEF

I

HENGALDAFLRKHEGQFTVIQLVGMLRGIAAGMRYLAD!4GYVHR

lllll::~lllli~~~; Humhck ECVVTKS

lHl~~lll~ll:ll:lllllll GMKyLsDmYVHR

720 730 740 750 760 770

340 350 360 370 380 390 Tk29. P DLAARNILVNSNLVCKVSDFGLSRVI:EDDPEAVYTTTGG KI-APEAIQYRKPTSAS

1111111~:111l111111111111:111111:111 lIIll:lII:lIIl:lIIlllll Humhok DLJJARNIIJNSNLVCKVSDFGLSRVLEDDPEAAY?TRCXi

700 790 800 810 820 030

1 400 410 420 430 440 450 Tk29.P DVWSYG ~YGERPYWraYNQwl~EEG~~~Il~~R

IIIIIIII:IIIIllIIIIll:lllllllll::llllll:llllll:l:lllllllllll Humhek DVWSYGV YGERPYWEMSNQDVTKAVDEGYRLPPP~KDR

840 050 060 870 880 890

i 460 470 480 490 500 510

Tk29 .P

I ::I]IIIIII:IIII:IIII:III: :: :I1 : jIllI: I:::1 ::I:jl::::

Humhck NNRPKFEQIVSILDKLIRSLKIITSAAARPSNLLLEQSNVDISTFR~~GVFZ 900 --I 910 920 930 940 950

520 530 540 550 560 570

Tk29.P ~YKDNFTAAG~SLESIDDVMSiGITLVGHQKKI.USSIQTMRAQntBLBGTGIQV :::I: jj:::j:\ :::1::::11::::1:1:11:1111:11~::: :I:

Humhak AH-1 FI’GVEYSSCDTIAKISTD~GVIVVGPQKKZiSSI~SKNGPVPV 960 970 9.30 990 1000 1010

Fig. 2. Amino acid sequence encoded by the Ebk cDNA clone, aligned with human Hek (Wicks et al., 1992). The Ebk sequence is incomplete at the 5’ end. The cDNA sequence has been submitted to the European Bioinformatics Institute (formerly EMBL) database, accession no. X81466.


20 nucleotides of poly(A). The DNA sequence has been submitted to the European Bioinformatics Institute (formerly EMBL) database, submission no. X8 1466.

The encoded sequence is shown in Fig. 2. It matches the last extracellular fibronectin type III domain, and the

whole intracellular domain, of an ELK-family receptor

tyrosine kinase. Although the closest homologue is Hek,

this is clearly a novel gene, which we name Ebk (embryo brain kinase).

According to BLAST screens, the amino acid sequence shows similarly high degrees of homology to ELK family genes Cek-8 from chicken and Sek, Bsk, and Hek from

mammals. In Fig. 2 it is aligned with human Hek; the homology is 86% in the kinase domain, 53% in the fibronectin type III domain, and less than 50% in the transmembrane and C-terminal regions. The most notice- able difference from other ELK family members is a 7-

amino acid insertion immediately before the transmem-

brane segment.

2.2. Chromosomal location

The mouse chromosomal location of Ebk was determined by interspecific backcross analysis using progeny

from matings of [(C57BL/6J X M. spretus)F, X C57BL/ 6J] mice. This interspecific backcross mapping panel has been typed for over 1600 loci that are well distributed

among all the autosomes as well as the X chromosome (Copeland and Jenkins, 1991). C57BL/6J and M. spretus

DNAs were digested with several enzymes and analysed

by Southern blot hybridisation for informative restriction fragment length polymorphisms (RFLPs) using the Ebk

cDNA probe. Three M. spretus-specific Sac1 RFLPs (see Section 4) were used to follow the segregation of the Ebk

locus in backcross mice. The results indicated that Ebk is in the proximal region of mouse chromosome 4 linked to Mos and Cga. Although 116 mice were analysed for

every marker and are shown in the segregation analysis

(Fig. 3), up to 119 mice were typed for some pairs of markers. Each locus was analysed in pairwise combinations for recombination frequencies using the additional data. The ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analysed for each pair of loci and the most likely gene

order are: centromere - Mos - 6/l 17 - Ebk - 3/l 17 - Cga.

The recombination frequencies (expressed as genetic dis-

tances in centimorgans (CM) + standard error) are: Mos -

5.1 (k2.0) - Ebk - 2.6 (k1.5) - Cga.

We have compared our interspecific map of chromosome 4 with a composite mouse linkage map that reports the map location of many uncloned mouse mutations (compiled by M.T. Davisson, T.H. Roderick, A.L. Hil- lyard and D.P. Doolittle, and provided from GBASE, a computerised database maintained at the Jackson Labora- tory, Bar Harbor, Maine). Ebk mapped in a region of the composite map that lacks any mutations with a phenotype that might be expected for an alteration in this locus.

4

5.1

2.6

i

MOS

Ebk

Cga

UN IIIM MCI

0 3

8qll

6q14-q21

Fig. 3. Mapping of Ebk in the proximal region of mouse chromosome 4. Ebk was placed on chromosome 4 by interspecific backcross analysis. The segregation patterns of Ebk and flanking genes in 116 backcross animals that were typed for all loci are shown at the top. Each column represents the chromosome identified in the backcross progeny that was inherited from the (C57BU6J X hf. sprerus)Ft parent. The shaded boxes represent the presence of a C57BU6J allele, and white boxes, a M. sprerus allele. The number of offspring inheriting each type of chromosome is listed below each column. A partial chromosome 4 linkage map is shown at the bottom of the figure. Recombination distances in centimorgans are shown to the left and the positions of loci in human chromosomes are shown to the right. References for the human map positions can be obtained from GDB (Genome Data Base), a computerised database of human linkage information maintained by The William H. Welch Medical Library of The Johns Hopkins University (Baltimore, MD).

The proximal region of chromosome 4 shares a region of homology with human chromosomes 6q and 8q (summarised in Fig. 3). This suggests that the human ho- molog of Ebk will reside on one of these chromosomes.

2.3. Distribution of Ebk gene expression In the adult mouse, northern blots were used to exam-

ine the tissue distribution of Ebk transcripts. RNA from mouse hippocampus, where in situ hybridisation shows a

J. Ellis et al. I Mechanisms of Development 52 (1995) 319-341 323

28s

18s

TK29

Fig. 4. The 4.0 kb Ebk mRNA in mouse hippocampus. Northern blot of 2pg of poly(A)+ RNA from adult mouse hippocampus, with an RNA

probe from pTK29.4 coveting the same region as probe 7.

strong signal (see below), shows an abundant 4-kb

mRNA (Fig. 4). Other blots, using 1.5 pg of poly(A)+ RNA per lane, showed a weak band in adult brain but nothing in adult lung, kidney, muscle, small intestine,

spleen, or liver. Therefore, Ebk is not abundantly expressed in these organs in the adult mouse.

A

cx Hip Cbl

In the brain and in mouse embryos, distribution was

examined by in situ hybridisation (ISH) to tissue sections.

ISH was carried out on adult mouse brain, on postnatal l- day and E17Sd and E15Sd mouse heads, on E16d rat

brain, and on whole mouse embryos between E14.5d and E9.5d. Probes from different parts of the gene (Fig. 1) gave the same results. Brain and head are illustrated in

Figs. 6-l 1, hybridised with probe 7; whole embryos are

shown in Fig. 12, hybridised with probe 1. Unless otherwise stated, these figures show sagittal sections with dorsal side up. Paired bright-field and dark-field images are shown.

The results for each region are described working backwards in time.

Brain. In adult mouse brain, Ebk is strikingly expressed in the hippocampus and nowhere else (Figs. 5 and 6) except for a few weakly positive cells in the subiculum. The sig-

nal is strong throughout the pyramidal cell layer and dentate gyrus granule cell layer, but the pyramidal cell

layer is noticeably stronger in CA1 than elsewhere. At Pld and E17.5d, signal is also seen in hippocampus

(Fig. 7) though it is weaker in the dentate gyrus which is

incompletely formed at those stages. The signal is variable at E17.5d when the hippocampus is only just acquir- ing its characteristic shape, but weak signal can be seen in

the future hippocampal area as early as E15.5d. At Pld, E17.5d and E15.5d, many other areas of the

B Mam ATec Hip Cx

Pal

Fig. 5. ISH with Ebk probe 7 to sagittal sections of (A) adult brain, (B) postnatal l-day head. For each, the top panel shows the Nissl stain, the bottom panel an autoradiograph on X-ray film. Note that hybridisation in adult is restricted to the hippocampus, but in the neonate it is widespread in brain and weakly detectable in non-neural structures. Abbreviations: Hip, hippocampus; Cx, cortex; Cbl, cerebellum; ATec, anterior tectum; Mam, mammil-

lary area; Str, striatum; Pal, palate; M, mouth.

324 J. Ellis et al. I Mechanism of Development 52 (1995) 319-341

brain are positive (Table IA; Figs. 5,7,8). Signal is seen throughout the cerebral cortex, including the whole thick-

ness of the cortical plate and subplate. At E15Sd, this is not present in all areas, and there is no signal in telen-

cephalon at earlier stages. Signal is also seen in several parts of the thalamus, mainly dorsal; this first appears

around El4Sd or slightly earlier. The mammillary area of the hypothalamus is conspicuously positive at all these stages, and this signal can be seen as early as El0.5d or

even E9.5d, when this area of neuroepithelium is not obviously differentiated. The rostra1 half of the superior

cohiculus or tectum between Pld and El3Sd shows a broad, moderate signal, fading away diffusely in the cau- da1 direction (Figs. 5,7,10,13A).

One of the most impressive areas of Ebk expression is at the rostra1 end of the tectum: the subcommissural organ (X0). There is an extremely intense signal in a sharply

defined patch of thin, medial and dorsal neuroepithelium (Figs. 8 and 9). We have localised this most precisely in embryos of E15.5 to E13.5 days (including transverse

sections) and in the E16d rat brain. Posteriorly it is bounded by thicker tectum, anteriorly it partially surrounds the pineal epiphysis (which is negative), and it

does not extend far from the midline. This coincides precisely with the subcommissural organ, a little-known member of the class of circumventricular organs that are

thought to mediate chemical communication between the blood and the ventricles. At these stages it is distinct from

the weaker, more diffusely hybridising anterior tectum (posteriorly, the two areas of hybridisation are not adja-

cent, as the strongly positive SC0 lies deeper while only

Table 1A

Regions positive for Ebk mRNA by in situ hybridisation: head

Fig. 6. Adult hippocampus. Figs. 6-l 1 all show mouse sections hybrid-

ised with probe 7, exposed in photographic emulsion, and counter-

stained with cresyl violet. Paired bright-field and dark-field images arc

shown. All are sagittal sections with dorsal side up, unless otherwise

stated. Scale bar = 0.5 mm. Positive areas in this field are the pyramidal

cells (especially in CA 1, to the left of the pair of white dots) and den-

tate gyrus granule cells.

Adult PI El7.5 El5.5 Rat El6 El4.5 E13.5, El2.5

Hippocampus

Cortex

Thalamus (parts)

Mammil. area

SC0

Ant. tectum

Cerebellum (parts)

Hindbraina (parts)

++ + tl- + _

- + + +I- - -

- + + t tl- tl- tl- - t f t t t ti-

t++ tit +++ tit -

;+, ;+,

t t t t

t t tf- -

- (+) (+) + t t +

Rest of heud Cochlear duct

V ganglion

Mesenc. flexure

Meninges

Cartilage

Eyelidsb

Mouth and nose (parts)b

+ t tt tt tt

+ + t t +

t t t ++

t t ++

+l- +I-

+ t t

+ + tt t ++

Blanks indicate that no data were obtained; -, no hybridisation. SCO, subcommissural organ.

a Including superior olivary nucleus, parts of rapht and/or vestibulru nuclei, and large areas of postero-lateral medulla identified as spinal trigeminal

nucleus. b Mainly in mesoderm deep to the epithelium, in some places extending quite deep. See text for details.

J. Ellis et ul. I Mechanisms ofDevelopment 52 (1995) 319-341 325

Fig. 7. Postnatal 1 -day forebrain. (Anterior to right.) Positive areas: hippocampus; plate and subplate of cortex; striatum; parts of thalamus; anterior

tectum (ATec). The subventricular zone (SVZ) is negative. Scale bar = 0.5 mm.

the lateral part of the tectum is positive; Fig. 9A). Before

E135d, the SC0 is not morphologically distinct, but its primordium is included in a strongly hybridising area of

dorsal diencephalon and rostra1 tectum (see Section 2.4). The hindbrain is negative in the adult but contains sev-

eral positive structures at Pld, El7Sd, and Ei4.5d (and in E 16d rat). At all these stages, there is signal in parts of the developing cerebellum (but not in the external granule layer), and in part of the rapht region, and in several areas of the medulla (also seen earlier; Figs. 10 and 12). Trans-

verse sections at El5d (not shown) show that the most striking positive nucleus is the superior olive, while broad, strongly positive zones in the lateral medulla (and also in the lateral pons) coincide with the spinal nucleus of the trigeminal nerve.

The medullary signals develop out of a more general region of Ebk expression in the segmented hindbrain at E9.5 days (Figs. 12,13). Strong signal is seen in rhom-

bomeres 3, 4, 5, and the rostra1 part of rhombomere 6; there is also weak signal in rhombomere 2 (Fig. 13). Levels are highest in rhombomere 5. Signal is mainly in

the inner (luminal) part of the neuroepithelium, and mainly on the ventral side.

Earlier stages of brain expression were studied in intact

embryos (see below).

Other parts of head. The developing head expresses Ebk in several peripheral neural and non-neural structures

(Figs. 8,11,12). The trigeminal ganglion is positive between Pl and

E13.5d. The inner ear is one of the most persistent sites of ex-

pression. Most of the cochlear epithelium is strongly positive at all stages up to Pld (we have not looked at it in the adult). Transverse sections at El5d show the signal is also present on part of the epithelium in the utricle and


Fig. 8. El5day brain. (Transverse section). The most intense signal is in the subcommissural organ (SCO). Other positive areas of brain: areas of neo-

cortex (NCx, weak) and entorhinal cortex (ECx, stronger); part of amygdala (Amyg); large areas of dorsal thalamus; small parts of hypothalamus.

Weak signal in hippocampal area (Hipp). Outside the brain: meninges; cochlear epithelium (Coch); trigeminal ganglion (V.G); Meckel’s cartilage

(Meek); mesoderm near external auditory meatus (EAM). Scale bar = 0.5 mm.

saccule, but absent from the semicircular canals. The

cochlear expression begins in the otic vesicle at E9.G 9.5d (Fig. 13).

There is hybridisation in the meninges at most embryonic stages, and particularly in tissue within the mesencephalic flexure, which may be meninges or mesen-

thyme. Rathke’s pouch, however, is negative. Signal is seen in many areas of the developing face. It

is mainly in mesenchyme deep to various epithelia, at

places where marked morphogenesis is occurring; in

some such places it extends quite deep. It is most evident between El55 and El25 days. At these stages, there is hybridisation in the mesenchyme of the palate, of the up-

per and lower lips, of the floor of the mouth, of parts of the external auditory meatus, and of the inner sides of the eyelids. Signal in the posterior part of the palate is very intense and persists at least as late as Pld. The tongue is mostly negative although signal is sometimes seen in the ventral half.

Some follicles of nasal whiskers and of teeth (Fig. 1 IB) are positive between Pld and E14.5d. There is also rather weak signal in some regions of cartilage at E15.5

and E14.5 days, viz. Meckel’s cartilage (Fig. S), and a ventral strip of nasal cartilage, and sometimes in the otic capsule. Other areas of cartilage are negative.

These positive regions of the developing nose and mouth can be seen emerging from less differentiated stages. At El 3.5 to E10.5, there is widespread although not very strong hybridisation in the snout and the first branchial arch (Fig. 12).

Rest of the body. Ebk is transiently expressed in several tissues during organogenesis (Table IB,C; Fig. 12). These

sites have not been mapped in detail nor followed beyond E14.5d. One of the most persistent sites is the lungs and bronchi, which are positive starting with the lung rudi- ment at E10.5d. The heart shows signal which is variable between embryos, sometimes weak and incomplete but sometimes strong all over the heart, peaking between

E12.5 and E10.5d. Some major arteries also show signal. There is weak, patchy expression in the developing kid-

ney.

Between E14.5 and E12.5, there is strong expression in the mesodermal parts of various morphogenetic structures, but not in the associated epithelia. In the feet, it

Table IB

Regions positive for Ebk mRNA by in situ hybridisation: rest of body

E14.5 E13.5, El2.5

Mesenchymea +I- ++

Heart (+)/- +

Bronchi/lungs (+) + Kidney (+) (+) Genital tubercle (dorsal part)b t +

Gut (par@ +I- tl-

Toes (~art)~ t +

a Apparently undifferentiated mesoderm flanking the spinal cord and

dorsal root ganglia, possibly including somitic mesoderm.

b Mainly in mesoderm deep to the epithelium, in some places extend-

ing quite deep. See text for details.

J. Ellis et al. I Mechanisms qf Development 52 (1995) 319-341 321

Table 1C

Regions positive for Ebk mRNA by in situ hybridisation: early embryos

Sections Whole mounts

El 1.5 El05 E9.5 24 som.

turned

(E9.5)

12-16 som. lo-12 som. 28 som.

turned turning unturned

(E9.0) (E8.7) (ES.3

Bruin Forebrain (ext. midline) - -

Dorsal fore/midbrain” +++ + ++ ++ + + Neural folds

Mammillary area + +l- -

Medulla (rhom.2-6) +P +P ++ ++ + +P (+/-)P Rhombomeres

Rest qf heud Otic vesicle ++ ++ + (+I-) _ - Otic pit

Mesen. flexure ++ ++ -

Meninges + +I- -

1st branchial arch + fP -

Snout + +

Rest qf body Heart ++/+p ++/+p (+I-) + + + + Heart

Bronchi/lungbud ii ++ -

Mesenchymeb ++ ++ ++l- + + +/- Somites

El 1 Sd to E9.5d, sagittal sections; E9Sd to ESSd, intact embryos (staged by embryo morphology).

a Sequentially: forebrain neural folds; forebrain dorsal midline; enhanced at caudal end (SCO) and spreading into dorsal midbrain.

b Apparently undifferentiated tissue flanking the spinal cord and dorsal root ganglia, including some somitic and marginal mesoderm. Also intense

hybridisation around the mesonephric duct and on the ventral side of the aorta.

p, part only.

appears to be in mesenchyme where each of the toes is beginning to grow out. In the midgut loop in the um-

bilical hernia, strong signal surrounds the endoderm in

one limb of the loop but not in the other. In the genital tubercle, there is diffuse signal on the dorsal side. At slightly earlier stages, strong signals are also seen in the

wall of the mesonephric duct and on the ventral side of the aorta; these too may include adjacent mesenchymal

areas. The most conspicuous post-cranial expression, be-

tween El35 and E10.5, is in mesoderm around the spinal

cord and between the dorsal root ganglia and alongside the vertebrae. This begins at E9.5-ElOSd in undifferentiated tissue flanking the spinal cord, including part of the

somites, and possibly also diffuse mesenchyme or neural crest (Figs. 12C). These signals are weak or absent by E14.5d. The spinal neural tube is always negative.

2.4. Ebk gene expression in early embryos Expression before ElOd was studied by ISH to intact

(‘whole-mount’) embryos, to determine when Ebk is first turned on (Table 1C; Figs. 14 and 15).

The hindbrain expression first appears faintly about

E8.5d (19 somites, embryo just starting to turn). This first region is ill-defined but probably becomes rhombomeres 2 and 3. Rhombomere 5 becomes positive later (about 12 somites, turning almost complete), and rhombomere 6 very soon afterwards. By E9.5d (>20 somites, fully turned), the pattern is as in Fig. 13.

The forebrain/midbrain expression first appears faintly on the crests of the forebrain neural folds, even before

they fuse, before E8.5d (unturned embryo). As they fuse, it becomes stronger, starting at the forebrain/midbrain

boundary which becomes the SCO. (At this stage (8-13 somites), the signal is also visible extending weakly into

the optic evaginations.) The outcome is an intensely positive strip along the dorsal midline of the forebrain by E9.5d, which later concentrates in the SC0 and rostra1

tectum to give the pattern described above. In mesodermal tissues, whole-mount ISH shows that

the somites become positive around E8.5d (7-9 somites, embryo half-turned). The heart is positive even earlier, in

unturned embryos.

3. Discussion

3.1. The Eph-like kinase family Ebk is the thirteenth member of the Eph-like kinase

family to be identified, according to the listing presented in Table 2. In this table, probable species orthologs are identified by their degree of homology, largely from the dendrograms of Sajjadi and Pasquale (1993) and Mai- sonpierre et al. (1993). The numbering of genes in this table is arbitrary and there is so far no systematic nomen-

clature for the family, which we call the ELK (Eph-like kinase) family.

Since first submission of this paper, Fox et al. (1995) have reported more clones for human ELK genes, all


Table 2

Members of the Eph-like kinase family

Rodent Human Chicken Refs.

1 W 1 2 EcWSek-2/Myk-2 Eck 14.20.22

3 Eek Eek 7 4 Mek-4’ Hek Cek-4 3.6 5 Nuk/tyro-S/Sek-3 Erk Cek-5 2,4,7,9,20

6 Elk Cek-6 5,8 I Ehk-l/Bsk Cek-7 5,10,11 8 SeWtyro- 1 Cek-8 2,5,12 9 Cek-9 5

10 Sek-Wtyro-6 Hek-2 Cek-10 2,5,13,20

11 tyro- 1 l/Myk- 1 2,22

12 Ehk-2 10

13 Ebk New

a Also tyro-4 [2] may be orthologous to Mek-4, but this is uncertain

as sequence is only available from a very conserved region, This num-

bering agrees with that being proposed by members of Amgen (Fox et

al., 1995). except that they list tyro-II/Myk-I as no. 13 and their human

orthologue of Ebk (HEK-11) as no. 11. They report additional human

orthologues. Note that these tyrosine kinases are not related to kinases

in the MAP kinase cascade (ERK, MEK, SEK) nor to transcription

factors (ELK) which have been given the same abbreviations. Refer-

ences to Tables 2 and 3: 1, Hirai et al. (1987); 2, Lai and Lcmke (1991);

3, Sajjadi et al. (1991); 4, Pasquale (1991); 5, Sajjadi and Pasquale

(1993); 6, Wicks et al. (1992); 7, Chan and Watt (1991); 8, Lhotak et al.

(1991); 9, Henkemeyer et al. (1994); 10, Maisonpierre et al. (1993); 11,

Zhou et al. (1994); 12, Gilardi-Hebenstreit et al. (1992); 13, Bohme et

al. (1993); 14, Lindberg and Hunter (1990); 15, Maru et al. (1988); 16,

Pasquale et al. (1992); 17, Nieto et al. (1992); 18, Irving et al. (1994);

19, Ganju et al. (1994); 20, Becker et al. (1994); 21, Ruiz and Robert-

son (1994); 22, Andres et al. (1994); 23, Soans et al. (1994).

orthologous to genes in Table 2. Their numbering agrees with Table 2, except that they list their human orthologue of Ebk (HEK-II) as no. 11, and tyro-I I/Myk-I as no. 13.

3.2, Expression patterns of Ebk We have shown that, in adult brain, Ebk is expressed

solely in the hippocampus. The most obvious property of the hippocampus is long-term potentiation (LTP), which raises the possibility that Ebk might be involved in this

process (see below). It is unlikely to be involved in developmental specification of the hippocampus as the strong expression only appears shortly before birth as the hippocampal morphology appears.

However, in the developing brain, while neurons are migrating and forming connections, Ebk is much more widely expressed. There are two very early locations, in the forebrain and the hindbrain.

The hindbrain expression can be related to the hindbrain segments or rhombomeres (Lumsden and Keynes, 1989; Lumsden, 1990; Wilkinson and Krumlauf, 1990; Graham, 1992). Rhombomeres are compartments, in the sense that cells acquire ‘even’ or ‘odd’ identity and show only limited migration into a rhombomere of opposite identity (Fraser et al., 1990; Guthrie and Lumsden, 1991;

Guthrie et al., 1993; Birgbauer and Fraser, 1994). This implies cell-cell interactions in which a receptor tyrosine kinase might be involved. Alternatively, such a receptor

might function later, as amarker for the segmental identity of neurons (see below).

The forebrain expression, in contrast, seems more related to morphogenesis, as it appears on the crests of the neural folds and is enhanced as they fuse. (Indeed, we found one aberrant embryo of 12 somites with fusion at the rostra1 as well as caudal end of the forebrain, with Ebk in both lines of fusion but not between them; data not

shown.) The line of fusion does not develop into neural tissue, and Ebk expression fades away, except at the cau- da1 end where fusion began.

This caudal end becomes the most striking location

of Ebk in the developing brain: the subcommissural

organ (SCO), which overlies the junction between the third ventricle and the midbrain aqueduct. This is one of

the circumventricular organs, areas of midline neuroepithelium in which there appears to be chemical communication between blood and cerebrospinal fluid (Rodriguez et al., 1987). The SCO, unlike the others, has an intact blood-brain barrier, although it is highly vascular. It

appears in all vertebrates although it degenerates after birth in the human. Its function is uncertain; there is some

evidence for a role in regulation of the body’s salt and water balance in association with the adrenal cortex, al-

though the SC0 is not essential for life (Severs et al., 1987). It also secretes Reissner’s fibre, a fibrous rod which extends all the way down the lumen of the neural

tube. The SC0 ependymal cells receive serotonergic syn-

apses in the rat but not in the mouse. This innervation is established postnatally in the rat, and inhibits S-100 protein expression and induces GABA uptake in the SC0

cells (Wiklund, 1974; Wiklund et al., 1977; Didier-Bazes al., 1992). In view of these species differences, it is note-

worthy that Ebk prenatal expression is the same in rat and

mouse. The development of the SC0 in the mouse embryo

was described by Rakic and Sidman (1968). It becomes histologically identifiable on the tenth day of gestation (presumably E9Sd), when the ependyma there is thinner than in adjacent areas. The SC0 develops without evident migration of cells. It becomes morphologically distinct between El2 and E14d, at which time secretion of Reiss- ner’s fibre begins, and its cells perform their final DNA replication between El 1 and E15d. The stages between El2 and E15d are also when we see extremely high ex-

pression of Ebk in the SCO. The cochlea is another major site of expression of Ebk,

beginning as soon as the otic vesicle invaginates, and continuing until after birth. Also in the peripheral nervous system, the trigeminal ganglion is positive at all stages that we have examined. In contrast, dorsal root ganglia show little or no Ebk expression.

.I. Ellis et al. I Mechanisms of Development 52 (1995) 319-341 329

Fig. 9. The intense hybridisation in the subcommissural organ (SCO). (A) E15d. transverse section, more posterior than in Fig. 8. The SC0 is separate

from the more weakly positive anterior tectum (ATec). (B) E16d rat, sag&l section in midline (anterior to left). Note relation of SC0 to the pineal

primordium and to the ventricle of the midbrain (mid). (C) E15d mouse, parasagittal section (anterior to left). Here the SC0 lies below the posterior

commissore. Meninges are positive. Scale bar = 0.5 mm.

Ebk is also expressed in many non-neural embryonic organs and tissues, notably in mesoderm underlying epithelia at times of conspicuous morphogenesis. Such sites include the toes, genital tubercle, eyelids, ears, and many parts of the mouth and nose. There is also expression in the heart (briefly), lung, and kidney, during organogenesis; in these cases we cannot resolve which tissue layer is positive. (Conversely liver, muscle, and most skeletal structures are negative.) The induction of these morphogenetic events presumably involves reciprocal signalling

between the cell layers, and Ebk might function as a receptor in these processes.

Many of these sites of expression have also been recorded for other members of the ELK family, and to un- derstand their roles, it will be necessary to consider not just one gene, but the whole family together.

3.3. Expression patterns of the Eph-like kinase family

There is some information on the expression patterns of all the 13 members of the ELK family listed in Table 2.


Table 3

Expression of the Eph-like kinase family

Gene Adult

Brain Lung O.E. 0.0.

Adult brain

Hipp. cx. Cbl. Other

Embryos Refs.

Brain/head Body

In bruin, almost nowhere else 3 Eek if - -

4 Mek-4, Hek + - -

4? tyro-4 (+) - 6 Elk ii - -

7 Ehk-I, Bsk ++ (-) I Cek-7 (CHICK) (*) - -

12 Ehk-2 - -

13 Ebk ;+; - -

Mainly in brain

5 Nuk, tyro-5, Erk, (+) + - Sek-3

5 Cek-5 (CHICK) ++ (-) +l- 8 Sek, tyro- I ++ + -

In most organs under mainly epitheliul organs 2 Eck, Sek-2 (+) ++ ++

4 Cek-4 (CHICK) ++ + + 6 Cek-6 (CHICK) ++ ++ + 8 Cek-8 (CHICK) ++ + +I- 9 Cek-9 (CHICK)

10 Cek- IO (CHICK) t+) (+) (+) + a

10 Hek-2 (HUMAN) + ++ + 10 tyro-6, Sek-4 +

11 tyro- I 1, Myk- 1 + ++ + I W - + +

-Ia C-1 -P

(+) -

-

C-1 a

(+I

+ + +I- (+)/-/a

(+) (+) (9

+

-

CA2-4

All

All

All

++

All

All

W - - + ++

++ Phi. Purl- ++ + (+)

Phi. P.&GC ++

+ C-L) (+) (+)

- - ++ + (inc.Rh2 6) (L,H)

+ Pur ++ (-) (inc.Rh2.3.5) (H)

+I- EGJJPF + ++ ++(L) + Pur ++

fin,. Rh3,S) (+)

Pur

d,c. Rh4) (-)

++ +

++ ++

++ ++

++ +

+ +

inc. Rh3,S) +

7

2

2,8 IO,11

5

IO

New

2,7, 920

4,16

2,12, 17.18

14, 19-22

3

5

5,23

5

5

I3

2.20

2,22

1.15

Expression patterns were surveyed in rodents unless otherwise specified. Symbols for tissues: O.E., other epithelia; O.O., other organs; Hipp., hip-

pocampus (‘All’, in all major cell layers; Ehk-I/&k(7) more abundant in CA2-4); Cx., cortex (Phi., piriform cortex); Cbl, cerebellum (Pur. or P.,

Purkinje cells; GC, granule cells; EGL, external granule layer; PF, parallel tibres). Symbols for expression patterns: ++, strong; +, present; (+), weak;

(-), barely or variably detectable; -, absent. Some of the weak signals could be due to expression in ganglia or Schwann cells. Blank means not tested.

a Also in adult: ryro-6(2), Schwann cells (+); Me/r-4(4), testis +; E&(6), testis (+); SeWrym-l(l), heart (+); Cek-9 in chick, thymus ++; Cek-IO in

chick, kidney ++. Hek(4) is also transcribed in human lymphoid cell lines. Also in embryos: (L), including lung; (H), including heart. See Table 2 for

key to references.

In the interests of clarity in this discussion, we append the numbers from Table 2 to the mammalian gene names, in the form Eph( 1) and Ebk( 13); these numbers are not in-

tended as a formal nomenclature but are chosen to match those given to the chick genes Cek-4 to Cek-10. In the following discussion, we consider avian and mammalian

genes separately, giving a total of 19 genes. All have been mapped at least by northern blots on adult tissues, and some by other techniques and in embryos.

Table 3 gives a brief summary of the expression patterns so far reported, with references. This table is provi- sional because the expression patterns have been determined by different groups with different degrees of sensi- tivity and completeness. This is particularly true of embryos where Nuk(5) and Sek(8) were studied in detail in early stages whereas the ‘tyro’ and ‘Cek’ genes were only investigated by northern blots from mid-to-late stages. In the adult, most data are from northern blots but some

groups used western blots. For several genes there are

splice variants which may be differentially expressed. Ebk is the first to have been mapped systematically throughout

development. There are many similarities between expression pat-

terns of different members; however, these patterns do not appear to be strongly conserved between mammals and birds. Thus Cek-4, -5, -6 are expressed in many organs in chicken although their putative mammalian orthologs are expressed only in brain and (for Erk(5)) in lung. Conversely Cek-7 is almost switched off in the adult chicken brain while Ehk-l(7) is widespread in adult rodent brain.

In the adult, the brain is obviously the major site of expression of this family: seven of the 19 genes, including Ebk( 13), have been recorded at substantial levels only in the brain or other neural tissue, with a few exceptions. All other members are also expressed in brain except for

J. Ellis et al. / Mechanisms c?f Development 52 (1995) 319-341

Fig. 10. E13d head (sagittal; anterior to left). Positive areas in brain: anterior tectum (ATec); strips of thalamus (Thai); mammilhuy area (Mam); supe-

rior olivary nucleus (Sol); parts of rapht and medulla. Weak signal in hippocampal area (Hipp); cortex (Cx) is negative. Ventricles are indicated (fore,

mid, hind). Outside the brain: meninges, especially in mesencephalic flexure (MeFIx); large areas of mesoderm around the nasal cavity (NasC) and

buccal cavity (BucC), including the upper lip. Inset: a more lateral section, showing signal on the tip of the nose adjacent to the nostrils, and intense

signal in the posterior palate (PPal). Scale bar = 0.5 mm.

E@(l). (Patterns within the brain are discussed below.)

Of the 15 genes that have also been looked for in embryos, again, all but one are expressed in the brain or head, often more abundantly than in the adult brain. Cek-

7, ryro-4(4?), and Ebk(l3) are notable in being expressed in embryo brain but hardly at all in adult brain, while

Ehk-2(12), the only one to be very weak in embryo brain,

is expressed strongly in adult brain. Non-neural sites of expression are predominantly

epithelial organs (especially for &h(l), E&(2), Hek- 2(10), and Cek-IO), and the most frequent site is the lung. This seems surprising given its apparent simplicity as a

tissue compared to the brain. Of three genes whose pre- dominant expression is in brain, SeWryro-I(8) and Erk(5)

also show major expression in lung, and Cek-5 in various tissues including lung. We note that embryonic lung also expresses Ebk( 13).

mapping of Nuk(5) and Sek(8) in early embryos showed

much less non-neural expression than for Ebk, but there were some parallels. The heart is a major site of Ebk ex-

pression in the early embryo, and weakly expresses Sek in the adult, and the main non-neural location for Nuk is the

sinus venosus of the heart (Henkemeyer et al., 1994). Many of these genes are also expressed in early develop-

ing mesoderm and neural-crest mesenchyme, in patterns which are too complex to summarise briefly.

The other nine genes are expressed in many organs in the adult; for several of them, brain and lung are again among the most conspicuous sites (Cek-6, Cek-8, and Hek-2(10)). Cek-9 is most abundant in the thymus, and Cek-10 in the kidney. Expression of one or another ELK family gene has been seen in endodermal, mesodermal, and neurectodermal organs.

Expression patterns in the adult brain. Of the eight genes that have been mapped by ISH within the adult brain, Ebk(13) and tyro-4(4?) have the most limited expression; all others are found in hippocampus, cortex, and cerebel-

lum. Ebk appears only in the hippocampus. The hippocampus is also a major site of expression for the other

seven genes; in each case the expression is localised over the major cell layers, i.e. the pyramidal cells and/or den-

tate gyrus granule cells. Five of the seven genes are expressed in all the hippocampal sectors (including Cek-5 in chick, which does not have the characteristic topology of the mammalian hippocampus); but Ehk-I/&k(7) is mostly expressed in sectors CA2, CA3, and CA4, while Ebk(13)

is slightly less abundant in those sectors than elsewhere. In embryos, judging by the far-from-complete data The cerebellum expresses all eight genes except

published, non-neural expression is probably more com- Ebk( 13) and tyro-4(4?); five of these genes are expressed mon than in the adult, as we found for Ebk. The detailed in the Purkinje cells, Ehk-2(12) also in the granule cells,


Fig. 11. Pld head. (A) Signal in cochlea (Coch) and trigeminal ganglion (V.G). (B) Signal around the teeth in upper jaw, adjacent to buccal cavity

(BucC) and posterior palate (PPal). Scale bar = 0.5 mm.

A

V.G

Palate

Heart

Foot

Ant tectum Cochlea ,

Lung

J. Ellis et al. I Mechanisms qf Development 52 (1995) 319-341

I ST - - L Br.Arch

e.Flx

Lung

C Medulla i

1 st Br.Arch

spc

mnd spc


D

Rh tom.2

0 tic.v.

Rh om.6

SPC

Fig. 12. Four sagittal sections of whole embryos, hybridised with probe 1, exposed in photographic emulsion, and counterstained with a Nissl stain. Paired bright-field and dark-field images are shown. The lumen of the neural tube is indicated in forebrain (fore), midbrain (mid), hindbrain (hind), and spinal cord (spc). Major positive structures are indicated. (A) E14.5d. V.G, trigeminal ganglion. Lips and whisker follicles are also positive. The liver scatters light in dark field but is not positive. (B) E12.5d. MeFIx., mesencephalic flexure. Note extensive hybridisation in meninges, and in mesen- thyme flanking the spinal cord and aorta, and in first branchial arch including the origin of the tongue. (C) El I .5d. Same as at E125d; also shows strong signal in the mesonephric duct (mnd). (D) E9.5d. Rhombomeres 26 are positive; the first branchial arch is negative.

and chick Cek-5 in the granule cells (external granule layer and parallel fibres by immunohistochemistry) but not in the Purkinje cells.

Other parts of adult brain have not been systematically examined for all eight genes, but at least Cek-5, Ehk-

I/Bsk(7), and Ehk-2(12) are widely but non-uniformly expressed. Several genes are expressed in at least part of the cortex , and we note that Ebk is expressed in fetal cortex but switched off in the adult.

Expression patterns in the embryo brain. In the early embryo brain, it is interesting to compare the Ebk expression pattern with the detailed mapping of Nuk/Sek-3(5) (Becker et al., 1994; Henkemeyer et al., 1994), Sek(8) (Nieto et al., 1992), and Sek-4(10) (Becker et al., 1994). In addition, recently, Eck/Sek-2(2) has been mapped up to E9.5d, but its cerebral expression is limited to rhom-

bomere 4 (r4), as well as some cranial ganglia (Ganju et al., 1994; Becker et al., 1994).

The early distributions of Nuk, Sek, Sek-4, and Ebk

show remarkable similarities and differences. In the hindbrain, all four genes are expressed mainly in

odd-numbered rhombomeres, but with slight differences. Sek(8) appears just before E8.0d and shows rapid regulation, appearing from r2 to r6 but mainly in r3 and later in r5. Nuk(5) appears about the same time in r2, r3 and r5. Sek appears sufficiently early that it might be involved in setting up the segmental pattern and inducing Krox-20

(Nieto et al., 1992), while the immunolocalisation of Nuk protein to the dorsal surfaces of neural tube cells at El l.Sd, where they contact adjacent mesenchymal cells, might suggest a role in subsequent morphogenesis of the brain (Henkemeyer et al., 1994). Sek-4(10) becomes con- fined to the rhombomere region around E8.0-E8.5d

J. Ellis et al. I Mechanisms ofDevelopment 52 (1995) 319-341

Fig. 13. Ebk in the rhombomeres (numbered) and otic vesicle (ov). These are parasagittal sections of the hindbrain region at E95d; anterior to right.

All are hybridised with probe 2, except for C which shows stronger signal with probe 1. Note signal in rhombomeres 2 (weak), 3, 4, 5 (strongest, ex-

cept in A), and 6 (rostra1 half). The signal is mainly in the inner (luminal) part of the neuroepithelium, and is absent from rhombomere 5 in A because

the section cuts only the lateral edge of that rhombomere. In some sections, signal can also be seen in mesoderm flanking the spinal cord. Branchial

arches are at bottom.

336

EB.0 d (4 som.)

E8.5 d (8 som.)

E9.0 d (12-16 som.)

E9.5 d (24 som.)

J. Ellis et al. 1 Mechanisms of Development 52 (1995) 319-341

r2 r3 r4 r5 1-6

Fig. 14. Diagram of hybridisation patterns in the rhombomere region. Compare with expression of other ELK genes in Fig. 7 of Becker et al. (1994). Segmental pattern appears at E8.0-8.5d (Sek and Krox-20 expression) and E&5-9.Od (morphological).

(Becker et al., 1994), and Ebk comes on here at E8.5 E9.0d (8-16 somites); both genes are thus slightly later than Sek and Nuk, but show a very similar shifting pat-

tern, with rise in r3 slightly before r5. Thus Ebk and Sek- 4(10) come on just as morphological segmentation begins, and after Krox-20 expression starts. In chick, rhom-

bomere compartment boundaries form at the same time as

the morphological boundaries, between the 6-somite and 16-somite stages (Fraser et al., 1990; Lumsden, 1990). If this is also true in mouse, Ebk and &k-4(10) come on too late to be involved in segmentation, but could contribute to Iabelling cells with their rhombomeric identities.

In the otic vesicle (which is ectodermal but has many

neural features), three of these four genes are expressed. Sek and Ebk are prominent, and Ebk persists in the coch-

lea at least until birth. Nuk appears only in a small area which evaginates to form the endolymphatic duct; again,

Nuk protein was localised at membrane contacts with adjacent mesenchyme/neural crest cells, suggesting a role in sculpting the growth and form of this duct (Henke-

meyer et al., 1994). In the forebrain and midbrain of the embryo, the four

genes have almost complementary patterns. Near the ventral midline, Sek-4( 10) is expressed transiently then

disappears, while Sek-3/Nuk(S) is abundant. Nuk(5) is strongly expressed along the ventral midbrain, optic chi-

asm, and telencephalon, but not in the hypothalamus, whereas Ebk( 13) is expressed in the mammillary area of

the hypothalamus. (Is there another Eph-like kinase that labels the remainder of the ventral hypothalamus?) Dor- sally, Sek(8) labels the forebrain, becoming restricted to the telencephalon, while Ebk( 13) labels the forebrain dorsal midline, later becoming prominent at its posterior end (the SCO) and also appearing in the anterior midbrain. So the first four ELK family genes to be mapped form an

almost-complete, non-overlapping, labelling system for these regions. In addition, the anterior-posterior gradient of Ebk expression on the tectum suggests that it might have a role in retinotectal mapping.

With reference to the strong expression of Ebk in the

embryonic SC0 (we have not examined it in adult), it is worth noting that Ehk-2( 12) is strongly expressed in two other circumventricular organs, the adult pineal and pitui- tary (Maisonpierre et al., 1993).

3.4. Functions of Eph-like kinases: combinatorial patterns?

The distributions described suggest strongly that these

genes may be active in morphogenesis, including the segmentation of the neural tube and of mesoderm (Nieto et al., 1992) the subsequent development of the brain and

ganglia, the formation of many organs that involve epi-

thelial proliferation, and (notably for Ebk) the moulding of many of the orifices and protuberances of the body.

But why are there so many of these receptors with similar expression patterns?

The emerging pattern seems to be that ELKS have largely overlapping domains, but also individual differences in the places and times of expression. Thus most of

them are expressed in the brain and the lung (although for some of them, such as Ebk in the lung, only in the em-

bryo), and some are expressed in other organs, particularly epithelial ones, Within the brain, most of them are

expressed in the hippocampus, cortex, and cerebellum (including Ebk, but in the cortex and cerebellum only while immature), and some in other structures. Being cell- surface receptors, these patterns suggest that the ELKS act in combinations. Perhaps a combination of ELKS on a neuron or epithelial cell is a label for regional identity and

developmental stage, giving a precisely specified address for their interactions with other cells or matrix.

Given that receptor tyrosine kinases act as dimers, and that closely related ones can form heterodimers (Wada et

al., 1990; Bellot et al., 1991; Schlessinger and Ullrich, 1992; also see Plowman et al., 1993), it is possible that ELKS also might form functional heterodimers.

An important clue to the role of such combinatorial addresses on neurons is given by the recent immunolo-

calisation of Nuk (Henkemeyer et al., 1994). It was found that Nuk is expressed on growing axons, including the growth cones. Likewise, Cek-8 immunoreactivity is present on growing motor axons (Soans et al., 1994). We can therefore envisage each axon carrying an ‘identity code’ in the form of a combination of ELKS. The distribution of the ELK ligands is so far unknown; perhaps they too are

expressed in a mosaic, either in territories through which the axons may grow (to guide pathfinding), or in the target tissues (to specify synaptogenesis). Axons could navigate by sensing combinations of ligands. For example, if the observed rhombomere expression is shared by all the neurons, motor axons of rhombomeres 3 and 5 would

J. Ellis et (11. I Mechanisms of Development 52 (199.5) 319-341

Fig. 1.5. Whole-mount ISH with digoxigenin-labelled Ebk. x37. (A) Expression in forebrain neural folds during closure. From left to right: Unturned

with no forebrain fusion; partly turned, 7-9 somites, with single point of fusion; partly turned, 1 l-12 somites, with line of fusion; 12-13 somites; 16

17 somites. (B) Expression in rhombomeres. From left to right: partly turned, 11-12 somites; 12-13 somites; 16-17 somites (head of this specimen

somewhat flattened); 24-25 somites. (C) Lateral view of a 13-somite embryo showing Ebk in forebrain, rhombomeres, heart, and somites. Positive

regions are marked as follows: E, optic evaginations; F, forebrain neural folds, later fusing in midline; H, heart; M, midbrain; 0, otic pit or vesicle; S,

first somite; r2-r6, rhombomeres 2-6.

express Nuk, Sek, and Ebk (and very likely other ELKS also, cf. tyro-4(4?) in adult facial motor nucleus; however

S&-4( 10) mRNA has disappeared by this time). Those of r.5 forming the facial nerve would have slightly more Ebk, while those of r3 entering the trigeminal nerve have Sek

persisting longer. Similarly, dendrites could navigate using ELKS over a

much shorter range, and a system like this would be ap- propriate for dictating where synapses should form. Neu- rons do not make synapses indiscriminately, as shown for example by mammalian neurons in culture (Ruffolo et al., 1978) and by nematode mutants (White et al., 1992). Are ELKS the molecules by which neurons identify correct partners for synapsing? This could be a role for ELKS in the adult brain, where axonal growth does not occur but synaptic remodelling does. It is an attractive explanation

for the remarkable concentration of ELKS in hippocampal neurons, with their high potential for long-term potentiation (LTP). (LTP can also be elicited from neurons in the cortex, but much less robustly than in the hippocampus;

Bear et al., 1992, and references therein.) Although it is not obvious why LTP, a biochemical phenomenon at al- ready established synapses, should have anything in

common with neuronal growth, it is possible that ana- tomical recognition between the pre- and post-synaptic elements may be an element in it. There is no evidence for synaptic growth in LTP, but some physical remodelling of synapses may occur (Desmond and Levy, 1990; Geinisman et al., 1991).

The first ligands for the ELK family have only just been discovered. Eck is bound by B61 (Bartley et al., 1994), a small secreted protein (Holzman et al., 1990)

338 J. Ellis et al. 1 Mechanisms of Development 52 (1995) 319-341

which also appears to exist in a GPI-linked cell-surface form (Bartley et al., 1994). B61 is secreted from endothe-

lial cells after induction by TNFa; whether other tissues synthesise it has not yet been reported. Three other ligands, all homologous to B61 and all binding to multiple

ELKS, have also been reported (Beckmann et al., 1994;

Cheng and Flanagan, 1994; Davis et al., 1994; Shao et al.,

1994). They include LERK-2, which is expressed in many

tissues including lung but not adult brain (Beckmann et al., 1994; Fletcher et al., 1994), and ELF-l, which is ex-

pressed in mouse embryos in a complicated pattern overlapping those of its receptors (Cheng and Flanagan, 1994). It will be important to know how many of these

ligands there are, and where and when they are expressed.

Are they displayed like a multicoloured carpet across which the ELK-labelled axons can find their way?

4. Materials and methods

4.1. Cloning and sequencing Oligonucleotides and conditions for PCR were similar

to those used by others (Wilks, 1989a,b; Lai and Lemke, 1991; Ziegler et al., 1993). The oligonucleotides were designed to match the conserved coding sequences for peptides IHRDL and DVWSFG, with 5’ extensions containing BamHI sites. The sequences were: 5’-ATCGGA- TCCAC2GNGAYYT-3’ (encoding IHRDL) and 5’- ATAGGATCCA3AGGACCA4ACRTC-3’ (complemen-

tary to codons for DVWSFG), where 2 = A/C, 3 = AIT,

4 = G/C, Y = T/C, R = A/G. The template for PCR was double-stranded cDNA

made from adult mouse brain using an Amersham cDNA

synthesis kit. PCR produced a =200-bp band as expected. After cleavage with BamHI, it was gel-purified and cloned into M13-mp18. Clones were sequenced by the Sanger dideoxy chain termination procedure using the

Sequenase 2 kit (U.S. Biochemicals). To obtain longer cDNA clones, the PCR clone TK29

was used to screen an E12.5d mouse embryo cDNA library (Logan et al., 1992). The 2.1 kb insert of the longest

clone, called pTK29.4, was subcloned by PCR into

pBluescriptSK+/-. It was sequenced by the dideoxy

method. To identify the highest degrees of homology with other

genes, a gene database screen was done using the BLAST network service at NCBI (National Center for Biotech- nology Information). The highest-scoring DNA matches were to mouse Bsk (Zhou et al., 1994) human Hek (Wicks et al., 1992), and mouse Mek4 (Sajjadi et al., 1991).

4.2. Chromosomal mapping Interspecific mouse backcross progeny were generated

by mating (C57BL/6J X M.spretus)F, females and C57BL/6J males as described (Copeland and Jenkins, 1991). A total of 205 N2 mice were used to map the Ebk

locus. DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, Southern blot transfer and hybridisation were performed essentially as described (Jenkins et al., 1982). All blots were prepared with Hy-

bond-N+ nylon membrane (Amersham). The probe was

the complete insert of pTK29.4, labelled with [a-32P]- dCTP by nick translation; washing was done to a final

stringency of 1.0X standard saline citrate (SSC) with phosphate, 0.1% sodium dodecyl sulphate (SDS), 65°C.

Major fragments of 14.5, 4.2 and 2.1 kb were detected in SacI-digested C57BW6J DNA and fragments of 14.5, 11.0, 5.0, and 2.5 kb were detected in SacI-digested M,spretus DNA. The presence or absence of the 11 .O, 5.0 and 2.5 kb M. spretus-specific Sac1 fragments; which co-

segregated, was followed in backcross mice. The probes and RFLPs for the loci linked to Ebk, including Mos and Cga (Tsha), have been described previously (Ceci et al., 1989). Recombination distances were calculated as de-

scribed (Green, 1981) using the computer program

SPRETUS MADNESS. Gene order was determined by minimising the number of recombination events required to explain the allele distribution patterns.

4.3. Synthesis of probes The probes for gel blots and in situ hybridisation (ISH)

were as follows (see Fig. 1). Nucleotide numbering of clone pTK29.4 begins with 28 nucleotides of polylinker.

Probe 7. 35S-Labelled single-stranded DNA probe from

the 5’ part of the gene, encoding the fibronectin type III

domain; nucleotide homology 56% to the most closely related genes such as Bsk. pTK29.4 fragment BamHI

(nucl.14) to Hind111 (350) was subcloned into M13- mp18. The DNA probe was made by polymerisation with Klenow enzyme from universal primer in the presence of limiting amount of [a-35S]dATP (400 Ci/mmol; Amer- sham) plus a small proportion of [a-32P]dATP, followed

by denaturation and purification of the labelled strand

from an agarose gel (Rogers and Hunt, 1987).

Probe 6 (opposite strand control). Identical to probe 7

except opposite orientation.

Probe 3. 32P-Labelled single-stranded DNA probe from the 3’ part of the gene, covering the C-terminal domain (nucleotide homology 60-62% to the most closely related genes such as Bsk, Hek, MeM) and 3’ untranslated region (negligible homology). Fragment Hind111 (350) to BamHI (1925) was subcloned into M13-mp9, and used to make the DNA probe covering XbaI (1407) to BamHI (1925). Probe was made by polymerisation with Klenow enzyme

from universal primer in the presence of [a-32P]dATP (1500 Ci/mmol; Amersham), followed by chase with un- labelled nucleotides, cleavage with XbaI, denaturation, and purification of the labelled strand from an agarose gel (Akam, 1983).

J. Ellis et ul. I Mechanisms &Development 52 (1995) 319-341 339

Probe 2. 35S-Labelled RNA probe from the 3’ part of the

gene, covering similar region to probe 3. Clone pTK29.4

in pBluescriptSK+/- was cut with PvuII and used as template for T7 RNA polymerase to make a 35S-labelled

RNA probe, complementary to sequences from PvuII (1323) to the 3’ end.

Probe 1. 35S-labelled RNA probe from the 3’ part of the

gene. It covers the same region as probe 2 plus most of

the kinase domain, which has nucleotide homology 76-

77% to the most closely related genes such as Bsk, Hek, Mek4. Clone pTK29.4 was cut with BssHII and used as

template for T7 RNA polymerase, complementary to se-

quences from BssHII (1019) to the 3’end.

For whole-mount ISH, a digoxigenin-labelled single-

stranded DNA version of probe 2 was made, primed with an oligonucleotide around nucleotide 2000 of the pTK29.4 insert, labelled with digoxigenin-1 l-dUTP (Boehringer-Mannheim), and synthesised with Taq DNA

polymerase through 20 thermal cycles as in Pate1 and

Goodman (1992). All four antisense probes were judged to be specific

for Ebk on the following grounds:

(i) Probes 2, 3, and 7 cover non-conserved regions of

the gene (see above for % homologies with other genes). (ii) These probes from the 5’ and 3’ halves of the gene

gave the same ISH results; probe 1, which includes conserved kinase sequences, gave stronger signals but the

same patterns. (iii) Probes 3 and 6 were tested on a genomic Southern

blot. For each probe, two of three lanes (digested with

different restriction enzymes) showed only two bands (data not shown). This implies that there is no more than one cross-hybridising gene, and perhaps none if the two

bands represent separate exons of Ebk. Southern and northern blots were done using Hybond-

N nylon membranes (Amersham), hybridised in a mixture

which included 50% formamide and 5X SSC. The northern blot in Fig. 4 was probed with an RNA probe covering the same region as probe 7, at 63’C, and washed in 0.1 x SSC/l% SDS at 68°C. Other blots were hybridised

with single-stranded DNA probe 3, at 42°C and washed in 1 X SSC/O. 1% SDS at 53°C to give the same stringency

as for ISH.

4.4. In situ hybridisation to sections In one set of experiments (e.g. Figs. 5-I l), procedures

were similar to those used for chick embryos by Rogers

and Hunt (1987). Adult mouse brains, newborn and E17.5d mouse heads, and E14.5d embryos, were quick- frozen; E9.5 to E15d mouse embryos, and E16d rat brains, were fixed by immersion for 6 h in 4% paraformaldehyde in phosphate buffer pH 7, followed by 20% sucrose in phosphate-buffered saline (PBS) overnight, all

at 4°C. All specimens were then embedded in Tissue-Tek,

sectioned on a cryostat into 12-pm sections which were

mounted on polylysine-coated slides (without gelatin), dried, fixed in cold 4% paraformaldehyde in phosphate buffer (pH 7) for 5 min, and further processed as described (Rogers and Hunt, 1987), including Pronase digestion for the sections of E9.5 to E15d mouse embryos which had been pre-fixed by immersion. Sections were

stored at -2O’C either dry or in 95% ethanol. Before hy-

bridisation, the sections were incubated in prehybridisa- tion mixture at room temperature for a few minutes, then

this was replaced with hybridisation mixture with probe

(Rogers and Hunt, 1987) and the slides were covered with

Parafilm and incubated at 42°C overnight. Washes were in 1X SSC, with the most stringent wash being in 1 X SSC plus 20 mM dithiothreitol plus 0.1% SDS for lo-

15 min at 53”C, followed by 0.1 X SSC at room temperature, and dehydration through ethanols containing 0.3 M ammonium acetate.

In another set of experiments (e.g. Figs. 12,13), mouse

embryos at ages E7.5d to E15.5d were isolated, fixed overnight in 4% paraformaldehyde (shorter fixation time for younger embryos), dehydrated with ethanol and xy-

lene, processed for paraffin embedding, sectioned at 6 pm

thickness, and mounted on 3-aminopropyltriethoxysilane- treated slides (Sigma). After removal of paraffin, the sec-

tions were predigested with proteinase K, acetylated with acetic anhydride, dehydrated and hybridised as described

by Frohman et al. (1990) except that base hydrolysis of probe was omitted.

Structures were identified using the atlases of Kauf-

man (1992) and Schambra, Lauder and Silver (1992).

4.5. In situ hybridisation to intact embryos Whole-mount ISH was done by a variation of the met-

hod of Wilkinson (1992) (D.G. Wilkinson, pers. com-

mun.). For our DNA probe, we found that permeabilisa-

tion required proteinase K at 40pglml for 30 min, or (especially for later embryos with closed cavities) pronase

E at 250pg/ml in 50 mM Tris (pH 7.5), 5 mM EDTA, 9% sodium dodecyl sulphate, for 15 min. Hybridisation

was at 50°C. Digoxigenin detection reagents were from Boehringer Mannheim.

Note added in proof

Since this paper was submitted, the Ebk gene has

also been described by three other groups: T. Ciossek et al. (mouse gene, named MDK-I; Oncogene 9 (1995)

97); G. Fox et al. (human gene, named HEK-II; Onco- gene 97 (1995) 1573). It may be of interest that the Ebk gene maps close to the mouse mutation Wheels, which gives recessive lethality and dominant defects of the semicircular canals (P. Nolan et al., Genetics 140 (1995) 245).

340 1. Ellis et at. I Mechanisms of Development 52 (1995) 319-341

Acknowledgements

Thanks are due to the following: Aaron Loomis

(Seattle) for oligonucleotide synthesis; Mark Hanks and Alexandra Joyner (Toronto) for the cDNA library; M.E.

Barnstead (Frederick) for excellent technical assistance; Sheila Barton, Paul Scowan and Azim Surani (Cam-

bridge) for providing mouse embryos; Eduardo Torres,

Chen Quin, Julia Watkins, Paul Monks, and Peter Starling

(Cambridge) for assistance in various ways. J. Ellis was

supported by a MRC studentship and a Research Student- ship from Magdalene College, Cambridge. E. Muir was supported by the International Spinal Research Trust. The chromosomal mapping was supported in part by the Na- tional Cancer Institute, DHHS, under contract NOl-CO- 74101 with ABL.

References

Akam, M.E. (1983) EMBO J. 2.2075-2084.

Andre& A.-C., Reid, H.H., ZiIrcher, G., Blaschke, R.J., Albrecht, D.

and Ziemiecki, A. (1994) Oncogene 9, 1461-1467. Bartley, T.D. et al. (1994) Nature 368, 558-560.

Bear, M.F., Press, W.A. and Connors, B.W. (1992) J. Neurophysiol. 67,

841-851.

Becker, N., Seitanidou, T., Murphy, P., Mattei, M.-G., Topilko, P.,

Nieto, M.A., Wilkinson, D.G., Chamay, P. and Gilardi-Hebenstreit,

P. (1994) Mech. Dev. 47,3-17.

Beckmann, M.P., Cerreti, D.P., Baum, P., VandenBos, T., James, L.,

Far&, T., Kozlosky, C., Hollingsworth, T., Shilling, H., Marask-

ovsky, E., Fletcher, F.A., Lhotak, V., Pawson, T. and Lyman, SD.

(1994) EMBO J. 13,3757-3762.

Bellot, F., Crumley, G., Kaplow, J.M., Schlessinger, J., Jaye, M. and

Dionne, C.A. (1991) EMBO J. 10,2849-2854.

Birgbauer, E. and Fraser, S.E. (1994) Development 120, 1347-1356.

Bliss, T.V.P. and Collingridge, G.L. (1993) Nature 361.31-39.

Biihme, B., Holtrich, U., Wolf, G., Luzius, H., Grzeschik, K.-H.,

Strebhardt, K. and Riibsamen-Waignmann, H. (1993) Oncogene 8,

2857-2862.

Ceci, J.D., Siracusam L.D., Jenkins, N.A. and Copeland, N.G. (1989)

Genomics 5,699-709.

Ghan, J. and Watt, V.M. (1991) Oncogene 6. 1057-1061.

Chao, M.V. (1992) Neuron 9,583-593.

Cheng, H.-J. and Flanagan, J.G. (1994) Cell 79, 157-168.

Copeland, N.G. and Jenkins, N.A. (1991) Trends Genet. 7, 113-118.

Davies, A.M. (1994) Nature 368, 193-194.

Davis, S., Gale, N.W., Aldrich, T.H., Maisonpierre, P.C., Lhotak, V.,

Pawson, T., Goldfarb, M. and Yancopoulos, G.D. (1994) Science

266,816-819.

Desmond, N.L. and Levy, W.B. (1990) Synapse 5, 139-143.

Didier-Bazes, M., Chouaf, L., Hardin, H., Aguera, M., Voutsinos, B.

and Belin, M.F. (1992) Prog. Brain Res. 91, 343-347.

Fletcher, F. et al. (1994) Oncogene 9,3241-3247.

Fox, G.M., Holst, P.L., Chute, H.T., Lindberg, R.A., Janssen, A.M.,

Basu, R. and Welcher, A.A. (1995) Oncogene 10,897-905.

Fraser, S., Keynes, R. and Lumsden, A. (1990) Nature 344, 43l-

434.

Frobman, M.A., Boyle, M. and Martin, G.R. (1990) Development 110,

589-607. Ganju, P., Shigemoto, K., Brennan, 3.. Entwistle, A. and Reith, A.D.

(1994) Oncogene 9, 1613-1624.

Geinisman, Y.. de Toledo-Morrell, L. and Morrell, F. (1991) Brain Res.

566.77-88.

Geissler, E.N., Ryan, M.A. and Housman, D.E. (1988) Cell 55, 185-

192.

Gilardi-Hebenstreit, P., Nieto, M.A., Frain, M., Mattbi, M.-G., Chestier,

A., Wilkinson, D.G. and Charnay, P. (1992) Oncogene 7, 2499- 2506. (Erratum: ibid. 8, 1103.)

Graham, A. (1992) Sem. Neurosci. 4,307-3 16.

Green, E.L. (1981) In Genetics and Probability in Animal Breeding

Experiments. Oxford University Press, New York, pp. 77-l 13.

Guthrie, S., Prince, V. and Lumsden, A. (1993) Development 118, 527-

538.

Guthrie, S. and Lumsden, A. (1991) Development 112, 221-229.

Henkemeyer, M., Marengere, L.E.M., McGlade, J., Olivier, J.P., Con-

Ion, R.A., Holmyard, D.P., Letwin, K. and Pawson, T. (1994) On-

cogene 9.1001-1014.

Heuer, J.G., von Bartheld, C.S., Kinoshita, Y., Evers, P.C. and Both-

well, M. (1990) Neuron 5.283-296. Hirai, H., Maru, Y., Hagiwara, K.. Nishida, J. and Takaku, F. (1987)

Science 238, 1717-1719.

Holzman, L.B., Marks, R.M. and Dixit, V.M. (1990) Mol. Cell. Biol.

IO, 5830-5838.

Irving, C., Alldus, G., Nieto, M.A. and Wilkinson, D. (1994) British

Society for Developmental Biology, Spring Meeting, Abstract B68.

Jenkins, N.A., Copeland, N.G., Taylor, B.A. and Lee, B.K. (1982) J.

Virol. 43.26-36.

Kamb, A., Weir, M., Rudy, B., Varmus, H. and Kenyon, C. (1989)

Proc. Nat]. Acad. Sci. USA 86,4372-4376.

Kaufman, M.H. (1992) The Atlas of Mouse Development. Academic

Press, London, San Diego, CA.

Keshet, E., Lyman, S.D., Williams, D.E., Anderson, D.M., Jenkins,

N.A., Copeland, N.G. and Parada, L.F. (1991) EMBO J. 10, 2425-

2435. Kuma, K., Iwabe, N. and Miyata, T. (1993) Mol. Biol. Evol. 10, 539-

551. Lai, C. and Len&e, G. (1991) Neuron 6,691-704.

Len&e, G. (1993) Nature 362, 291-292.

Lhotak, V., Greer, P., Letwin, K. and Pawson, T. (1991) Mol. Cell.

Biol. 11,2496-2502.

Lindberg, R.A. and Hunter, T. (1990) Mol. Cell. Biol. 10,6316-6324.

Logan, C., Hanks, M.C., Noble-Topham, S., Nallainathan, D.. Provart.

N.J. and Joyner. A.L. (1992) Dev. Genet. 13,345-358.

Lumsden, A. and Keynes, R. (1989) Nature 337.424-428.

Lumsden, A. (1990) Trends Neurosci. 13,329-335.

Maisonpierre, P.C., Barrezueta, N.X. and Yancopoulos, G.D. (1993)

Oncogene 8,3277-3288. Marcelle, C. and Eichmann, A. (1992) Oncogene 7.2479-2487.

Maru, Y., Hirai, H., Yoshida, M.C. and Takaku, F. (1988) Mol. Cell.

Biol. 8,3770-3776.

Meakin, S.O. and Shooter, E.M. (1992) Trends Neurosci. 15, 323-331.

Nieto, M.A., Gilardi-Hebenstreit, P., Chamay, P. and Wilkinson, D.G.

( 1992) Development I 16, 1137-I 150. O’Dell, T.J., Kandel, E.R. and Grant, S.G.N. (1991) Nature 353, 558-

560.

Pasquale, E.B. (1991) Cell Regul. 2, 523-534.

Pasquale, E.B., Deerinck, T.J., Singer, S.J. and Ellisman, M.H. (1992)

J. Neurosci. 12, 3956-3967. Pate], N. and Goodman, C. (1992) In Nonradioactive In Situ Hybridi-

sation. Boehringer Mannheim GmbH, Biochemica, Mannheim, pp.

62-63. Plowman, G.D., Green, J.M., Culouscou, J.-M., Carlton, G.W., Roth-

well, V.M. and Buckley, S. (1993) Nature 366,473-475. Rakic, P. and Sidman, R.L. (1968) Am. J. Anat. 122, 317-336.

Rodriguez, E.M., Oksche, A., Rodriguez, S., Hein, S., Peruzzo, B..

Schoebitz, K. and Herrera, H. (1987) In Gross. P.M. (ed), Circum-

ventricular Organs and Body Fluids, vol. Il. CRC Press, Boca Ra-

ton, FL, pp. 3-4 1. Rogers, J.H. and Hunt, S.P. (1987) Neuroscience 23, 343-361.

Rubin, G.M. (1991) Trends Genet. 7, 372-377.

J. Ellis et al. I Mechanisms of Development 52 (1995) 319-341 341

Ruffolo, R.R., Eisenbarth, G.S., Thompson, J.M. and Nirenberg, M.

(1978) Proc. Natl. Acad. Sci. USA 75.2281-2285.

Ruiz, J.C. and Robertson, E.J. (1994) Mech. Dev. 46, 87-100.

Sajjadi, F.G. and Pasquale, E.B. (1993) Oncogene 8, 1807-1813.

Sa.jjadi, F.G., Pasquale, E.B. and Subramani, S. (1991) New Biologist

3.769-778.

Schambra, U.B., Lauder, J.M. and Silver, J. (1992) Atlas of the Prenatal

Mouse Brain. Academic Press, London, San Diego, CA.

Schlessinger, J. and Ullrich, A. (1992) Neuron 9, 383-391.

Severs, W.B., Dundore, R.L. and Balaban. C.D. (1987) In Gross, P.M.

(ed), Circumventricular Organs and Body Fluids, vol. II. CRC

Press, Boca Raton, FL, pp. 43-58.

Shao, H., Lou, L., Pandey, A., Pasquale, E.B. and Dixit, V.M. (1994) J.

Biol. Chem. 269.26606-26609.

Soans, C., Holash, J.A. and Pasquale, E.B. (1994) Oncogene 9, 3353-

3361.

Wada, T., Qian, X. and Greene, M.I. (1990) Cell 61, 1339-1347.

Wanaka, A., Johnson, E.M. and Milbrandt, J. (1990) Neuron 5, 267-

281.

White, J.G., Southgate, E. and Thomson, J.N. (1992) Nature 355, 838-

841.

Wicks, I.P., Wilkinson, D., Salvaris, E. and Boyd, A.W. (1992) Proc.

Natl. Acad. Sci. USA 89, 1611-1615.

Wiklund, L. (1974) Cell Tissue Res. 155, 231-243.

Wiklund, L., Lundberg, J.J. and MollgSrd, K. (1977) Acta Physiol.

Stand., Suppl. 452, 27-30.

Wilkinson, D.G. and Krumlauf, R. (1990) Trends Neurosci. 13, 335-

339.

Wilkinson, D.G. (ed) (1992) In Situ Hybridisation: A Practical Ap-

proach. IRL Press, Oxford, pp. 75-83.

Wilks, A.F. (1989a) Proc. Natl. Acad. Sci. USA 86, 1603-1607.

Wilks, A.F. (1989b) Gene 85,67-74.

Ziegler, S.F., Bird, T.A., Schneringer, J.A., Schooley, K.A. and Baum,

P.R. (1993) Oncogene 8,663670.

Zhou, R., Copeland, T.D., Kramer, L.F. and Schultz, N.T. (1994) J.

Neurosci. Res. 37, 129-143.

embryo brain kinase: a novel gene of the eph/elk receptor tyrosine kinase family

Documents