Neuron, Vol. 14, 1189-1199, June, 1995, Copyright © 1995 by Cell Press

Plexin: A Novel Neuronal Cell Surface Molecule That Mediates Cell Adhesion via a Homophilic Binding Mechanism in the Presence of Calcium Ions

Kunimasa Ohta,* t Akihito Mizutani,* Atsushi Kawakami,* Yasunori Murakami,* Yasuyo Kasuya,* Shin Takagi,* Hideaki Tanaka,~ and Hajime Fujisawa* *Department of Molecular Biology School of Science Nagoya University Chikusa-ku, Nagoya 464-01 Japan 1:Division of Developmental Neu,,obiology Department of Neuroscience and Immunology Kumamoto University Graduate School

of Medical Sciences Kuhonji, Kumamoto 862 Japan

Summary

Plexin (previously referred to as B2) is a neuronal cell surface molecule that has been identified in Xenopus. cDNA cloning reveals that plexin has no homology to known neuronal cell surface molecules but possesses, in its extracellular segment, three internal repeats of cysteine clusters that are homologous to the cysteine- rich domain of the c-met proto-oncogene protein prod- uct. The exogenous plexin proteins expressed on the surfaces of L cells by cDNA transfection mediate cell adhesion via a homophilic binding mechanism, under the presence of calcium ions. Plexin is expressed in the receptors and neurons of particular sensory sys- tems. These findings indicate that plexin is a novel calcium-dependent cell adhesion molecule and sug- gest its involvement in specific neuronal cell interac- tion and/or contact.

Introduction

During development, neurons ~re interconnected in a highly specific manner. Though a variety of mechanisms are thought to work in concert to accomplish final patterns of neuronal connections (for reviews, see Dodd and Jes- sell, 1988; Goodman and Shatz, 1993), cell-cell adhesion seems to be an important mechanism, at least in such initial phases of neuron network formation as the guidance of growing axons along particular paths, selective fascicu- lation of axons, and specific neuronal cell recognition and contact. Over the past decade, several studies have identi- fied cell surface molecules that mediate neuronal cell adhesion in both vertebrates and invertebrates and suggested their involvement in axonal growth and fascicu- lation (Neugebauer et al., 1988; Matsunaga et al., 1988; Seilheimer and Schachner, 1988; Doherty et al., 1990;

tPresent address: Division of Developmental Neurobiology, Depart- ment of Neuroscience and Immunology, Kumamoto University Gradu- ate School of Medical Sciences, Kuhonji, Kumamoto 862, Japan.

Elkins et al., 1990; Gennarini et al., 1991 ; Grenningloh et al., 1991) and in neuronal recognition (Nose et al., 1992).

On the basis of molecular structures and manners of cell binding, the foregoing cell-cell adhesion molecules found in the nervous tissues are categorized as the immu- noglobulin family, the cadherin family (for reviews see Jes- sell, 1988; Edelman and Cunningham, 1990; Grenningloh et al., 1990), or other molecules with specific cell adhesion motives (Snow et al., 1989; Nose et al., 1992). Among these, only the molecules that belong to the cadherin fam- ily require calcium ions to mediate cell-cell adhesion via homophilic molecular bindings (Takeichi et al., 1990; Ta- keichi, 1991).

We report here that a neuronal cell surface molecule named plexin (previously referred to as B2 or the B2 anti- gen; see Takagi et al., 1987; Ohta et al., 1992) is a new member of molecule that mediates cell-cell adhesion via a homophilic binding mechanism in the presence of calcium ions but is distinctly different from the cadherin molecular family.

Plexin is a 220 kDa protein recognized by a monoclonal antibody named B2 (MAbB2), which was generated against the nervous tissues of Xenopus tadpoles and is expressed in the plexiform layers of the optic tectum (Ta- kagi et al., 1987). As shown in a previous study (Ohta et al., 1992), plexin is also localized in the inner plexiform layer of the neural retina in which cellular processes of retinal neurons intermingle and contact each other. In the development of the neural retina, plexin first appears in the newly formed inner plexiform layer, then is restricted to the sublaminas a and b of the inner plexiform layer, where cellular processes of the off-center and on-center retinal neurons are contacted to make synapses, respec- tively. We also reported that a serum antibody generated against the affinity-purified plexin proteins impeded the formation of the inner plexiform layer in cultured embry- onic neural retinas (Ohta et al., 1992). These foregoing results suggest that plexin is involved in neuronal cell inter- action and/or contact. The structure and molecular roles of plexin, however, have remained obscure.

In this study, to gain further insight into the function of plexin in nervous systems, we perform cDNA cloning and sequencing of Xenopus plexin and analyze whether plexin has cell adhesion properties. Sequence analysis reveals that plexin is a type I membrane protein with no homology to known neuronal cell surface molecules. The putative extracellular segment of the plexin protein possesses three internal repeats of unique cysteine clusters, which are homologous to the cysteine-rich domain of the c-met proto-oncogene protein product, the receptor protein for the hepatocyte growth factor (Park et al., 1987; Chan et al., 1988). The putative cytoplasmic segment of the plexin protein, however, lacks the tyrosine kinase domain, which is an essential motif for growth factor receptors, sug- gesting that plexin is not a growth factor receptor. Rather, we show that the cells transfected with the cDNA encoding the plexin protein acquire cell adhesiveness, indicating

Neuron 1190

A

B ,49 I

B Sm

B,ZA i

8A :

813A I

B14A ]

827A l

I

I ~ 8,zB ' ~ 8B

[ 813B

q 81,1s

[ 827B

i l lkb

I 8z4

~iill

- -9 .5 - -7.5 - -4 ,4

Figure 1. Plexin cDNA Clones and Northern Blots (A) A schematic representation of the overall structure of the plexin cDNA. The coding region is indicated by an open box, and the noncod- ing region by solid lines. Restriction sites are as follows: B, BamHI; E, EcoRI; K, Kpnl; Sin, Smal. (B) cDNA clones whose nucleotide sequences were determined. The clones 812, 8, 813,814, and 827 are restricted with EcoRI at the same site. The open reading frame of the plexin cDNA was determined by combining the sequences of the cDNAs indicated by bold lines. (C) Northern blot. Total cellular RNA (20 p.g) isolated from Xenopus tadpole brains (stages 51-53) was loaded and hybridized with a cDNA (clone 8B). Molecular weight markers are given in kilobases. One major transcript of -10.0 kb is detected.

that plexin is a molecule with cell adhesion properties. We show that the cell adhesion occurs via a homophilic binding of the plexin proteins under the presence of cal- cium ions. Also, we perform immunohistochemistry in combination with in situ hybridization in Xenopus tadpoles and show that plexin is expressed in all neuronal elements composing the accessory olfactory, lateral line, and vestib- ular systems. On the basis of these data, we discuss the possible roles of plexin in nervous systems.

Results

Isolation of cDNA Clones Encoding Plexin To elucidate the primary structure of plexin, we screened a ;~gtl 1 cDNA library of Xenopus tadpole brains using a guinea pig anti-plexin (B2) antiserum (Ohta et al., 1992) and detected 9 positive clones among 9 x 105 recombi- nant phage. One clone containing a 1.0 kb insert (named 8; Figure 1 B) showed the strongest immunoreactMty. The insert cDNA contained an endogenous EcoRI site and was restricted to two cDNA fragments named 8A and 8B. By using the 8B insert as a probe, we performed Northern blot analysis of the total RNA of tadpole brains and detected a band corresponding to about a 10.0 kb mRNA, which is long enough to encode the 220 kDa polypeptide of plexin (Figure 1C). Thus, we selected insert 8B for further cloning.

The 8B cDNA fragment was labeled with [~-~P]dCTP using the random primed DNA labeling kit (Boehringer). The labeled cDNA was used to screen a ;~gtl0 cDNA li- brary of Xenopus tadpole brain. By repeating the screen- ings with newly isolated cDNA clones, several clones that encompass the putative entire coding region for the plexin protein were subsequently isolated (Figures 1A and 1B). Besides clone 8, clones 812,813, 814, and 827 were also restricted by EcoRI into a pair of cDNAs of different lengths. Among all these clones, the nucleotide se-

1

az

161

141

SOl

~ez

~61

1041

Z Z n

1281

ZV6Z

1841

~RPLPFHLW~L~GSWIATG~Sp~SDWSLT~I~Iy~S~G 80

PVE DNEKC¥ PPPSVQ SC pH GL I T~ t~ rNK~L ~ Dy SDk~L I AC GSASQ3 ICQFLRLDDL~K LGEp~KEHYLSS~ SG 160

~ S ~ Z I E V ~ V G T P I ~ S E Y F P T L ~ S ~ L ~ E N ~ Q D E F V S S Q ~ I p S D ~ S K F P ~ D I ~ S 240

F S SEQFVYY LTLQLD TQLT S pDS TGEQFFTSK I VRLCgDDpKF¥ S TVE F p I C-C~ DGV~ FR LI QDAYLSKPGKP.LAKELG 320

i SE~DI~FS~QK~IKpg~ S ~ I K D K I~R IQSC~G~LSL~INSP~IDD~ 400

D ~QP~G~ IE GTPLFLDKED~TSV~Y DyRGH~AG~S~ I L~L~S S H L V ~ A I ~D 480

PK~ IS~SE~A~VpDLSR~CSFE~GRI~ IyCTSP~VI p ~ G D ~ S ~ 640

K FAS~FW ~ S C L SC~GS FPC~CKy~C ~N~CS FQEGR~ ~ I L eS~ I T I P ~ p I ~TA 720

KNL~GQRNYEC IFH I ~ S V ~ S T S IQCQNTSXN~EG~I SDL~S~G~VI D N ~ N ~ 800

S~SCGLC~SDRRFEC~SE~CT~QNCp~T~SR~I~FpE~PR~LTI~ 880

LRFED IB FGgRVG MVMCVPVE SE ~ I SAEQ IVCE INDAGR TRVMEAQVEVCV~DC s QD~RAI SpK S FT~S~T~R 96O

GPLS~ I S IEG~AGSDVSVAIG~S~T~ IRCKT~pST~ IQ rL I ~ S E ~ T ~ D ~ 1040

QK IEP~S I~TPL I~T IKEPKI~yGD~CTLY~D~C~S~Np~SP~RPDE IGF IM 1120

DNV~ I V~TTS FLY I pDPVFE pLTASGtC~J~pSS pL I IKGRNL I P~P~GI~T~ IGDT PC~ SE TQDDCE S 1200

pNL~QHKVT I~XS~QI YSDS~A I IO l ~GL~L I I I ~ I A y ~ S ~ R T ~ S ] 280

R V ~ C K E ~ LOTD I ~ L ~ D ~ G I ~ ~ ~ ~ V ~ ~ IE DH ~ V Q ~ S L ~ L L ~ T 1360

F IR~EAQR SF S~DRG~AS L I ~E~YAT~QLLS DL ZE ~ SK~ ~ S V ~ 1440

KFL~ C A~ PLTM~CA I K ~ KGp I DA I T ~ y S LSEDKL i R~ I D~N~D~LSDE SC~SPQT~PE 1520

NEN~E I ~VLNCDT I ~EK ~DA~ KGVpy S QR ~ A G ~ I I ~DE D ~ X D ~ Q 1600

V~G S SV~VPKQN SAY N I SN S S ~ SL SRY~ SM~ TA S S pD 5~SRT~ I T pD~ ~ ~ G D R G 1680

SKMVS E I ¥LTR LLA'FKGTLQKFVDDLTE T I F STAERG S~LpLA IKyMFDFLDE QADKIIQ I TDY~T~E SNCLgI~dV 1760

NV I KN PQFVTD I HEN S I TDACLS WAQTT~SC STSE ~LGKD SP.SNKLLYAKD I I~TE SWV~BT YAD I AKMPV I SDQDM 1840

SAy~QSRLHLSQ~SMS~E I YSY I ~mDE I LT~EQ~Q~S~ID~QSS 1905

Figure 2. Deduced Amino Acid Sequence of the Synthetic cDNA De- termined from Clones 49, 812A, 8B, and 824 A thin line at the N-terminal indicates the putative signal sequence. The putative transmembrane domain is indicated by a thick line. The boxes indicate the cysteine-rich domains (see text). Potential N-linked glycosylation sites are indicated by dots.

quences surrounding the EcoRI restriction site were com- pletely matched. Thus, we concluded that the EcoRI site was intrinsic.

By combining the sequences of clones 49, 812A, 8B, and 824, we determined the nucleotide sequence for the putative plexin cDNA (6163 bases; see the GSDB, DDBJ, EMBL, and NCBI nucleotide sequence databases with the accession number D38175). The open reading frame of 5715 nucleotides is initiated by a ATG codon (nucleotide position 111), which is preceded by stop codon TAA (18 bases upstream) in the same reading frame and embed- ded in Kozak consensus motif (Kozak, 1989). The compos- ite cDNA predicts a polypeptide consisting of 1905 amino acid residues (Figure 2).

The amino acid sequences of the N-terminal peptides for the affinity-purified plexin (220 kDa) and its degraded product (160 kDa; Ohta et al., 1992) were A/L/M/S P/Q K/ V DIG F/LN and S P K/V DIG F/L, respectively. Though the exact N-terminal amino acid sequences for both peptides were not decided because of the impurity of the speci- mens, the consensus sequence as SerProLysAspPhe between the two peptides completely matched the N-terminal sequence predicted from the cDNA cloning (Figure 2). Moreover, cells transfected with the cloned cDNA express a 220 kDa protein that reacts with a plexin- specific antibody (see below). Taken together, these re- sults indicate that present cDNA encodes the plexin pro- tein.

Plexin Is a Type I Membrane Protein with Homology to the c-met Proto-Oncogene Protein Product The hydropathy analysis (according to the method of Kyte and Doolittle, 1982) on the amino acid sequence for the plexin protein suggests the presence of a signal peptide and a single membrane-spanning region. As shown in Fig- ure 2, the N-terminal methionine is followed by 28 hy- drophobic amino acids that appear to be a cleavable signal sequence accordingto the "(-3, -1) rules" of von Heijne

Plexin: A Novel Neuronal Cell Adhesion Molecule 1191

A c-Met

Plexin-C1

Plexin-C2

Plexin-C3

~ -B 4 ~ . o:~--RAP 47, o : ~ i - - -F'qqi~:~R~- ~ B - - w T E ] , . . . . . . . . ~:~ i~ q,, -IE SI~ O ~i~-Fsl- -I-4"~ a~ 4~1"4'- -1"4M ~ 44@-~E,~, , O -BL . ,F Z -BD O- - ~r'RIq~ ~ - I q 45~L .,'~,V 4' 4- - -"T Pl- ~--:44~'4"'~ 4"F- ~ N ~" ~.-r4. . . . . . I ~ . , : - - 1 4 ~ o ~

B Plexin

c-Met

C~ C-rich C3 domain

~ PN

Figure 3. Comparison of Primary Structures of the Plexin Protein and the c-met Proto-Oncogene Protein Product (A) The cysteine-rich domains of the mouse c-met proto-oncogene protein product (c-Met protein, 519-560 amino acid residues; see Chan et al., 1988) and of the plexin protein (designated as C1, C2, and C3, following the N- to the C-terminal of the plexin protein) were aligned to maximize the match among them. Eight cysteine residues in each domain are hatched. Residues identical in at least two domains are enclosed by boxes. (B) The structures of the plexin protein and the mouse c-Met protein are schematically represented. The cysteine-rich domains are shown as open boxes (C1, C2, and C3 for the plexin protein and a Cys-rich domain for the c-Met protein). The mouse c-Met protein has a tyrosine kinase domain (TK) in its cytoplasmic segment. PM, plasma membrane.

(1986). A classical transmembrane domain consisting of highly hydrophobic amino acids (amino acid residues 1230-1257) is also identified and is immediately followed by a cluster of hydrophilic amino acids (LysArgLys, resi- dues 1258-1260) that can serve as a cytoplasmic anchor. The putative extracellular segment has 18 potential N-linked glycosylation sites (AsnXSer/Thr; X :~ Pro).

A computer-aided homology search shows that the plexin protein has no similarity to such known neuronal cell surface molecules as the immunoglobulin family, the cadherin family, and the integrin family. The most striking feature of the plexin protein is that its putative extracellular segment possesses three cysteine-rich domains desig- nated C1, C2, and C3, which correspond to amino acid residues 506-554, 652-697, and 800-851, respectively (Figure 2 and Figure 3). Each domain contains 8 cysteines, and has a consensus frame of amino acid sequence as CysX(5-6) CysX(2)CysX(6-8)CysX(2)CysX(5)CysX(5- 6)CysX(12, 15-16)Cys (Figu re 3A). The internal homology score among the cysteine-rich domains is 33%-39% (Ta- ble 1). A homology search revealed that the C1, C2, and C3 cysteine-rich domains of the plexin protein have homol- ogies (33%-38%) to a cysteine-rich domain of the c -met

proto-oncogene protein product (c~Met protein; Park et al., 1987; Chan et al., 1988), which is the receptor for the hepatocyte growth factor (Table t; Bottaro et al., 1991). The cysteine-rich domain of the c-Met protein also consists of 8 cysteines aligned as CysX(5)CysX(2)CysX(8)CysX(2)- CysX(3)CysX(5)CysX(8)Cys (Figure 3A; see Park et al., 1987; Chan et al, 1988). The extracellular region of the plexin protein flanking the C3 domain (amino acid residues 852-1229) also has homology (25%) to the c-Met protein.

Besides the cysteine-rich domains, the ArgGlyAsp (RGD) sequence (amino acid residues 369-371), which is one of the binding sites of extracellular matrix molecules to their cell surface receptors (for review, see Hynes, 1992), and the LeuArgGlu (LRE) sequence (amino acid residues 803-805), which is an adhesion site of s-laminin (Hunter et al., 1989), are located in the extracellular seg- ment of the plexin protein.

The c-Met protein has a tyrosine kinase domain in its cytoplasmic segment (Park, et al., 1987; Chan et al., 1988), as schematically shown in Figure 3B. The putative cyto- plasmic segment of the plexin protein consisting of 648 amino acid residues (positions 1258-1905) has no signifi- cant homology with known proteins nor protein motives, except a consensus sequence for the tyrosine kinase phosphorylation site R/KX(2-3)D/EX(2-3)Y (amino acid residues 1819-1826; see Hunter, 1982; Patschinsky et al., 1982; Cooper et al., 1984). It has not been decided whether the Tyr1826 of the plexin protein is phosphory- lated.

Transfected cDNA Expresses Plexin Proteins on the Cell Surface The plexin cDNA was inserted into the eukaryotic expres- sion vector pSVL (Pharmacia) and transfected into a mouse fibroblastic cell line (L cell) that does not express endogenous plexin proteins. Cells transfected with the plexin cDNA were stained with a plexin-specific mono- clonal antibody MAbB2 (Takagi et al., 1987; Ohta et al., 1992). The antibody binds to the living transfectants (Fig- ure 4A), indicating that the proteins encoded by the plexin cDNA are expressed on the cell surface. When the transfectants contact each other, the plexin proteins accu- mulate at the sites of contact (Figure 4A), as do L cells transfected with cadherin cDNAs (Nagafuchi et al., 1987; Hatta et al., 1986),

Immunoblot analyses of the cDNA-transfected cells us-

Table 1. Sequence Homology Between the Cysteine-Rich Domain of the Mouse c-met Proto-Oncogene Protein Product (c-Met) and C1, C2, and C3 Domains of the Plexin Protein

c-met Plexin-C1 Plexin-C2 Plexin-C3

c-Met -- 38 36 33 Plexin-C1 - 35 39 Plexin-C2 - 33

Values represent percent homology.

Neuron 1192

220 kd

B 1 2 3 4 5

Figure 4. Exogenous Plexin Proteins Expressed in L Cells after Trans- fection with Plexin cDNA (A) An immunofluorescent staining. A plexin-specific monoclonal anti- body (MAbB2) binds to living L cells that had been transfected with the plexin cDNA. Note that the plexin proteins are accumulated at the sites of cell contact (arrows). Scale bar, 20 I~m. (B) Immunoblots, Affinity-purified plexin protein (0.5 p~g; lane 1), tissue extracts of Xenopus tadpole brain at stage 53 (one-tenth of a brain; lane 2), L cells transfected with the pSVL vector (1 x 104 cells; lane 3), L cells transfected with the full length of plexin cDNA (1 x 104 cells; lane 4), and L cells transfected with the mutant plexin cDNA (see text; 1 x 104 cells; lane 5) were resolved by SDS-PAGE, transferred to a nitrocellulose filter, and incubated with guinea pig anti-plexin antise- rum. The antiserum binds to the 220 kDa protein in lanes 1, 2, and 4, and also to the 160 kDa protein in lane 5, but not in lane 3.

ing a guinea pig anti-plexin antiserum (Ohta et al., 1992) show a band comigrating at around 220 kDa (Figure 4B, lane 4), which exactly corresponds to the native plexin proteins (Figure 4B, lanes 1 and 2). These results confirm that the cloned cDNA encodes the full length of the plexin molecule.

Plexin Mediates Cell Adhesion via Homophilic Binding under the Presence of Calcium Ions To test whether plexin can function as a cell adhesion molecule, we performed a cell aggregation test using the plexin-expressing transfectants. As a control, we used the L cells that were transfected with the pSVL vector. Agita- tion of the control cells in the presence of 2 mM CaCI2 at 37°C resulted in weak cell aggregation (Figures 5A and 5D). In contrast, the plexin-expressing transfectants formed large aggregates under the same conditions (Fig- ures 5B and 5D). The plexin-expressing cells were also aggregated in the presence of 2 rnM MgCI2. The aggrega- tion activity, however, was weaker than that in the pres- ence of calcium ions (data not shown). Neither the control cells nor the plexin-expressing transfectants formed ag- gregates without calcium or magnesium ions (data not shown). The cell aggregation did not occur at 4°C (data not shown). The guinea pig anti-plexin antibody (Fab frag- ments, 0.4 mg/ml) inhibited the aggregation of the plexin- expressing cells, though not completely (Figures 5C and 5D).

As the control cells themselves showed a weak cell ad- hesiveness, in the presence of calcium ions (see Figures

5A and 5D), the cell aggregation assay could not exactly define whether plexin mediates cell adhesion via homophi- lic or heterophilic molecular interactions. To clarify the nature of the cell binding mediated by plexin, we examined whether the plexin-expressing cells could adhere to immo- bilized plexin proteins. According to procedures reported (Lemmon et al., 1989; Tanaka et al., 1991), we spotted affinity-purified plexin proteins on nitrocellulose.coated plastic dishes, then applied the dissociated transfectants. The plexin-expressing cells adhered to the plexin spots in less than 30 min (Figure 5E), whereas the control cells did not (data not shown). The plexin-expressing cells ad- hered only to the plexin spots (Figure 5F). Adhesiveness of the plexin-expressing cells to the immobilized plexin proteins was dependent upon the amount of the precoated plexin protein and was partially inhibited by the plexin- specific serum antibody (data not shown). In the absence of calcium ions, the plexin-expressing cells did not adhere to the plexin spots (data not shown). These results indicate that the plexin proteins expressed on the cell surface medi- ate cell adhesion via a homophilic binding mechanism in the presence of calcium ions.

To test whether the cytoplasmic segment of the plexin protein is essential in the plexin-mediated cell adhesion, we transfected L cells with a mutant plexin cDNA in which the sequences encoding most of the cytoplasmic segment of the plexin protein (595 amino acid residues out of 648 between the position 1311 to the C-terminal; see Figure 2) had been deleted and isolated a clone that stably ex- pressed a minute plexin protein of 160 kDa (see lane 5 in Figure 4B). The living transfectants were stained with MAbB2 (data not shown), indicating that the minute plexin proteins are localized on the cell surface. The cell aggrega- tion test showed that the transfectants with the minute plexin proteins also formed aggregates in the presence of calcium ions (data not shown). There was no clear differ- ence in the ability of cell aggregation between the transfec- tants with the normal and minute plexin proteins, even though the amount of the exogenous proteins expressed was considerably higher in the cells transfected with the mutant plexin cDNA than in the cells transfected with the full-length plexin cDNA (compare lanes 5 and 4 in Figure 4B). The transfectants with the minute plexin proteins also adhered to the immobilized plexin proteins (see Figure 5G). In this assay, the more transfectants with the minute plexin proteins adhered to the plexin spot than the transfectant with the normal plexin proteins (compare Fig- ures 5G and 5F).

Expression of Plexin in Sensory Systems We have reported that the plexin is expressed in the plexi- form layers of the optic tectum (Takagi et al., 1987) and of the neural retina (Ohta et al., 1992) and suggested the involvement of the molecule in the interaction and/or con- tact of retinal neurons (Ohta et al., 1992). Plexin is also expressed in parts of the Xenopus nervous system other than the optic tectum and neural retina. Thus, to better understand the roles of plexin in the organization of ner- vous systems, we extensively analyzed the expression patterns of plexin by means of immunohistochemistry and

Plexin: A Novel Neuronal Cell Adhesion Molecule 1193

h ' t~o 0.~

0.(

D 15 30 45 6 0

~n

Figure 5. Cell Aggregation and Adhesion of the Transfectants (A-D) Cell aggregation test. The transfectants (5 x 105 cells/0.5 ml) were agitated at 80 rpm at 37°C. (A-C) show dark field micrographs of aggregates formed at 60 min of agitation. (A) The L cells transfected with pSVL vector. (B) The plexin-expressing transfectants. (C) The plexin-expressing transfectants that were preincubated with the anti-plexin antibody (see text). (D) Kinetics of cell aggregation. The number of particles was counted at 15 rain intervals. N,/N0 represents the number of particles observed at a given time (N,) relative to the number at the beginning of the assay (No). Mean and standard deviation in particle number at each time point were plotted for the L cells transfectee with pSVL vector (open squares; n = 4), the plexin-expressing transfectants (open circles; n = 4), and the plexin-exprsssing transfectants incubated with the anti-plexin antibody (closed circles; n = 5). (E-G) Cell adhesion test. The transfemants (2 x 106 cells) were added to dishes that had been spotted with the affinity-purified plexin proteins and incubated for 30 min. (E) and (F) show The binding of the plexin-expressing transfectants. (E) A phase contrast micrography. The transfectants adhere to the plexin-spotted side of the dish (asterisk). (F) Immunofluorescent staining with MAbB2. MAbB2-positive transfectants adhere only to the MAbB2-positive side of the dish (asterisk). (F) Arrows indicate the boundary of the MAbB2-positive area. (G) A phase contrast micrograph showing the binding of the transfectants expressing the minute plexin protein (see text). The transfectants adhere to the plexin-spotted side of the dish (asterisk). Bars, 100 prn (A-C, E, G); 20 prn (F).

in situ hybridization and clarified that plexin was expressed in all neuronal elements composing such sensory circuits as the accessory olfactory system, the lateral line system, and the vestibular system. The Accessory Olfactory System The accessory olfactory system consists of the vomerona- sal receptor neurons within the vomeronasal organ, their axons (the vomeronasal nerve fibers), and the accessory olfactory bulb in which the vomeronasal nerve fibers termi- nate (Scalia, 1976; Burton, 1990; Burton et al., 1990). Im- munostaining with MAbB2 indicates that the plexin proteins are localized on the epithelial cells of the vomeronasal organ (Figure 6A), the vomeronasal nerve fibers (Figure 6A and 6B), and the neuropile of the accessory olfactory bulb (Fig- ure 6D). The plexin mRNAs are detected in the epithelial cells of the vomeronasal organ (Figure 6C) and in the tar- get neurons for the vomeronasal nerve, which are located at the deeper part of the accessory olfactory bulb (Figures 6E and 6F). No in situ hybridization signals are detected in cells within the neuropile of the accessory olfactory bulb (Figures 6E and 6F), suggesting that nonneuronal cells do not express plexin. The Lateral Line System and the Vestibular System The lateral line is a collection of mechanosensory receptor cell masses (neurornasts) that are arranged in linear array in the skin of the head and body (Russell, 1976). Immuno-

staining with MAbB2 indicates that the plexin proteins are localized on cells of the neuromasts, but not on the epider- mal cells (Figures 7A and 7B). We did not detect a sufficient amount of in situ hybridization signals in the neuromasts, probably owing to technical problems. However, in epider- mis cultured for 2 days in Holtfreter's solution, MAbB2 also bound to the cells of the neu romasts, but not the epidermal cells (data not shown), indicating that the receptor cells express plexin. The lateral line nerve that innervates the neuromasts (Figures 7A and 7B) and the lateral line gan- glion neurons that are located at the dorsal (and the ven- tral) aspects of the trigeminal ganglion also express the plexin proteins (Figure 7B). The trigeminal ganglion neu- rons and their fibers are negative for immunostaining with MAbB2.

In the vestibular system, MAbB2 binds to the cells of the vestibular receptor, the vestibular nerve (the eighth cranial nerve), and the vestibular ganglion neurons (Figure 7C). The antibody also bound to the cultured vestibular epithelium (data not shown), indicating that not only the vestibular sensory neurons but also the vestibular recep- tors express plexin.

The lateral line nerve and the eighth nerve project to the lateral line nucleus and the nuclei of the eighth nerve, respectively, which are located at the dorsal part of the medulla (Nikundiwe and Nieuwenhuys, 1983). Also, it has

Neuron 1194

Figure 6. Expression of Plexin in the Accessory Olfactory System (A, B, and D) Localization of the plexin proteins, as detected by immunofluorescent staining with MAbB2. The epithelium of the vomeronasal organ (VNE in [A]), the vomeronasal nerve fibers (arrows in [A] and [B]), and the neuropile of the accessory olfactory bulb (AOB in [D]) are positively stained with MAbB2. (C, E, and F) Localization of the plexin mRNAs, as detected by in situ hybridization. (C) and (E) show dark-field micrographs; (F) shows a bright-field micrograph of (E). In situ hybridization signals are detected in the epithelial cells of the vomeronasal organ (VNE in [C]) and the target neurons for the vomeronasal nerve fibers (arrows in [E] and [F]) within the accessory olfactory bulb (AOB). Asterisks in [B] indicate the Bowman's glands. All images are at stage 51. Bars, 100 rim (A-C); 200 pm (D-F).

been reported that, at larval stage in Xenopus, the termi- nals of both the lateral line nerve and the eighth nerve overlap to form a common neuropile (Shelton, 1971). The plexin proteins are localized in the common terminal neu- ropile for the lateral line and eighth nerves (Figure 7D). The plexin mRNAs are expressed in cells within the lateral line nucleus and the nuclei of the eighth nerve (Figures 7E and 7F).

In the sensory systems described above, plexin was expressed throughout larval life and in frogs. However, the expression of plexin in the receptors and the sensory nerves of both the lateral line and vestibular systems is prominent in embryos and young tadpoles before stage 48, but weak in those at the later stages and in frogs.

Discussion

Our previous studies have shown that plexin (the antigen for the monoclonal antibody MAbB2) is a membrane pro-

tein expressed in restricted regions of Xenopus tadpole nervous tissues (Takagi et al., 1987; Ohta et al., 1992), and that it should be involved in neuronal cell interaction (Ohta et al., 1992). In this study, we isolated cDNA clones encoding the plexin protein and demonstrated that it is a novel type I membrane protein with cell adhesion prop- erties.

The plexin protein has no homology to known neuronal cell surface molecules, but possesses three internal re- peats of cysteine-rich domains in its extracellular segment. Each cysteine-rich domain is composed of 8 cysteine resi- dues, and it has a homology to the cysteine-rich domain of the c-met proto-oncogene protein product (Park et al., 1987; Chan et al., 1988), which is a receptor for the hepato- cyte growth factor (HGF) (Bottaro et al., 1991) and of the more recently identified Met protein family, c-sea (Huff et al., 1993) and Ron (Ronsin et al., 1993). The alignment of 8 cysteine residues in the cysteine-rich domains of the plexin protein and the Met protein family is similar. Re-

Plexin: A Novel Neuronal Cell Adhesion Molecule 1195

Figure 7. Expression of Plexin in the Lateral Line System and the Vestibular System (A-D) Localization of the plexin protein, as detected by immunofluorescent staining with MAbB2. The neuromasts of the lateral line receptors (LLR in [A] and [B]), the lateral line nerve fibers (arrows in [A] and [B]), and the lateral line ganglion neurons (LLG in [B]) located at the dorsal part of the trigeminal ganglion (TG) are stained with MAbB2. Neither the epidermal cells (small arrows in [A] and [B]) nor the trigeminal ganglion neurons (B) are stained with MAbB2. The antibody also binds to the vestibular receptor (VR in [C]), the eighth (the vestibular) nerve fibers (arrow in [C]), the vestibular ganglion neurons (VG in [C]), and the common terminal neuropile for the lateral line and the eighth nerves within the medulla (see text; asterisk in [D]). TB/BT in (D) indicates the MAbB2-positive tectobulbal and bulbotectal tracts, which are the efferent and afferent of the optic tectum, respectively (see Ohta et al, 1992). (E and F) Localization of the plexin mRNAs, as detected by in situ hybridization. Dark-field and light-field micrographs ([E] and IF], respectively) of a frontal section of the medulla. In situ hybridization signals are detected in the cells of the lateral line nucleus (dotted area; LLN) and of the nucleus for the eighth nerve (nVIII). Asterisks indicate the neuoropile corresponding to that in (D). Images are at stage 43 (A and B) and at stage 48 (C-F). Bars, 50 p~m (A), 100 I~m (B and C), 200 I~m (D-F).

peated cysteine residues within the extracel lu lar seg- ments of growth factor receptors are genera l ly considered

to form an essent ia l structural backbone for l igand-binding domains (UIIrich et al., 1984, 1985; Park et al., 1987; Bot- taro et al., 1991). Thus, the structural s imi lar i ty of the cyste- ine-rich domains between the plexin protein and the Met protein fami ly may suggest that plexin is a Met-related growth factor receptor. However , the structure of the

plexin protein is dif ferent f rom that of growth factor recep- tors in several aspects. First of all, the putat ive cytoplasmic segment of the plexin protein lacks the tyrosine kinase domain, which is an essent ial funct ional domain of the HGF receptor (c-Met protein; see Park et al., 1987; Chan et al., 1988; Bottaro et al., 1991), of Met-related putat ive growth factor receptors (Huff et al., 1993; Ronsin et al., 1993), and of the other growth factor receptor protein tyro-

Neuron 1196

sine kinases (for reviews, see Carpenter, 1987; Chao, 1992). Also, the piexin protein lacks a potential proteolytic processing site, which is characteristic of the Met-related growth factor receptor proteins (Chan et al., 1988; Huff et al., 1993; Ronsin et al., 1993) and the human insulin receptor precursor protein (Ullrich et al., 1985). Therefore, it seems unlikely that plexin acts as a growth factor re- ceptor.

Rather, the present finding that L cells acquire cell ad- hesion ability after transfection with the plexin cDNA indicates that plexin is a cell surface molecule with cell adhesion properties. Contact between plexin-expressing transfectants occurs via the homophilic binding of the plexin proteins and requires calcium ions, suggesting that plexin is related to the cadherin protein family, a group of calcium-dependent homophilic cell adhesion molecules (for reviews, see Takeichi et al., 1990; Takeichi, 1991). However, database searches did not reveal any similari- ties in the amino acid sequence of the plexin protein to the cadherin protein family, nor to other known cell adhesion molecules. Thus, plexin is a new molecule with calcium- dependent homophilic cell adhesion properties.

Which part(s) of the plexin protein is essential for the homotypic molecular interaction remains unknown. Fur- thermore, the plexin protein lacks the EF hand, which is a calcium-binding site (Kretsinger, 1980), the consensus sequence AXDXD, which is the site of calclium-binding in the cadherin protein family (Ozawa et al., 1990), and DXDXDGXXD, which should contribute to the divalent cat- ion-binding property of the integrin molecular family (for review, see Hynes, 1992). Thus, it is unclear how calcium (or magnesium) ions are involved in the plexin-mediated cell adhesion.

In the plexin-expressing transfectants, the plexin pro- teins accumulate at the sites of the cell contact (see Figure 4A), as in cadherin-expressing cells (Nagafuchi et al., 1987; Hatta et al., 1988). It has been reported that cadher- ins colocalize with cortical actin bundles and other cy- toskeletal proteins at cell-cell boundaries (Hirano et al., 1987; Geiger and Ginsberg, 1991), and that the interac- tions between the cytoplasmic segment of cadherins and the cytoskeletal proteins are essential for cadherin- mediated cell adhesion (Nagafuchi and Takeichi, 1988; Ozawa et al., 1990; Hirano et al., 1992). However, the present finding that the L cells transfected with the mutant plexin cDNA, in which the sequences encoding most of the cytoplasmic segment of the plexin protein has been deleted, exhibit cell adhesiveness suggests that the mode of cell adhesion mediated by plexin is different to that by cadherins. The cytoplasmic segment of the plexin protein has no homology to the proteins reported, except a motif for the tyrosine kinase phospholylation site. Further analy- ses are necessary to clarify the roles of the cytoplasmic segment of the plexin protein in cell adhesion.

The present finding that plexin mediates cell adhesion enables us to predict the involvement of the molecule in the formation and/or maintenance of neuronal cell con- tacts. As reported previously (Ohta et al., 1992), in the Xenopus neural retina, the plexin proteins were localized in cellular processes of the retinal neurons to form plexi-

form layers, and an anti-plexin antibody perturbed the for- mation of the plexiform layers. It seems possible that plexin is involved in the contacts between cellular pro- cesses of the retinal neurons, probably by means of a homotypic molecular interaction. As shown in this study, in the accessory olfactory system, plexin is expressed in both the vomeronasal nerve fibers and their target neurons within the accessory olfactory bulb (see Figure 6). Also, in the lateral line and vestibular systems, all neuronal ele- ments composing the neuron circuits, such as the recep- tors, sensory neurons, and the central targets to which these sensory neurons project, express plexin (see Figure 7). Thus, on the assumption that plexin has cell adhesion properties, it could be speculated that plexin expressed in both the pre- and postsynaptic elements plays a role in specific neuronal cell contacts. In the lateral line and vestibular systems, both the receptors and the sensory neurons derive from common placodes and are in contact from the beginning of development. In these systems, plexin may be involved in the maintenance of the contacts between the receptors and sensory neurons.

Several cell adhesion molecules acting with a homotypic molecular interaction, such as NCAM (Cervello et al., 1991), E-cadherin (Shimamura et al., 1992), and fasciclins I and II (Bastiani et al., 1987; Elkins et al., 1990; Grennin- gloh et al., 1991), are thought to be involved in axon-axon interaction to regulate axon fasciculation. We recently re- ported that plexin and the other neuronal cell surface mole- cule, neuropilin (renamed from A5; see Takagi et al., 1987, 1991; Fujisawa et al., 1989), are differentially expressed in Xenopus olfactory axon subclasses, and that the path- ways for these axon subclasses within the olfactory nerve are sorted according to the expression levels of these cell surface molecules (Satoda et al., 1995). Also, the present study shows that the lateral line ganglion neurons and the trigeminal ganglion neurons are closely associated to form a single cell mass, but their fibers (the lateral line and the trigeminal nerve fibers) form completely separated fiber bundles (see Figure 7B). The trigeminal nerve expresses neuropilin (Takagi et al., 1991 ; Hirata et al., 1993) but not plexin, whereas the lateral line nerve expresses plexin (see Figure 7), but not neuropilin (unpublished data). These findings suggest that plexin and probably neu ropilin are also involved in axon-axon contact and mediate selec- tive axon fasciculation.

This study shows that plexin mediates cell adhesion via a homophilic molecular interaction in model system. How- ever, we failed to obtain direct evidence indicating the involvement of plexin in cell adhesion of Xenopus neurons, because of technical limitations. The aggregation activity of the plexin-expressing L cells seems not as strong as that of the cadherin-expressing L cells (Nagafuchi et al., 1987; Hatta et al., 1988). Also, the deletion of the cyto- plasmic segment of the plexin protein did not affect the plexin-mediated cell adhesion. Thus, it is still possible that cell adhesion is not a major role of plexin in vivo. The homophilic molecular interaction of plexin might mediate some cell-cell signalings in vivo, as suggested for the cell adhesion molecules NCAM and N-cadherin (Doherty et al., 1991). It is also open to question whether the plexin

Plexin: A Novel Neuronal Cell Adhesion Molecule 1197

proteins interact in a heterophil ic manner. Though the

present cell adhesion assay using L cells transfected with the plexin cDNA shows the homotypic binding of the mole- cule, we cannot completely exclude the possibil i ty that L cells lack cell adhesion l igands for the plexin protein. The finding that the extracel lular segment of the plexin protein

possesses the RGD and LRE sequences, which are the putative adhesion sites of extracel lular matrix molecules (Hunter et al., 1989; for review, see Hynes, 1992), might indicate that plexin interacts with molecules other than itself. It is also possible that the regularly repeated cysteine residues within the extracel lular segment of the plexin pro- tein interact with unknown HGF-li!<e l igands and play roles

in cel l -cel l interactions. To understand the exact roles of plexin in the organiza-

tion of nervous systems, further analyses, such as the characterization of the biochemical propert ies of the plexin protein and the manipulation of p exin-gene expression in vivo, are necessary. Our more recent f indings that mole- cules homologous to Xenopus pls,xin are distr ibuted in the

chick and mouse (unpublished data) may open the way for these approaches.

Experimental Procedures

Construction and Screening of cDNA Library Total poly(A) + RNA from Xenopus tadpole brains (stage 51; according to the criteria by Nieuwkoop and Faber, 1956) was isolated by phenol extraction following oligo(dT) affinity chromatography as described (Agata et al., 1983). ~,gtl0 and ~.gtl I cDNA libraries were constructed according to Huynh et al. (1985). The ~.gtl 1 library was screened with a serum antibody generated against affinity-purified plexin proteins (Ohta et al., 1992). A cDNA insert showing the strongest immunoreac- tivity with the antibody was used as a probe for the screening of the Zgtl0 library to determine the full length of the plexin-coding region.

DNA Sequencing and Computer Analysis The cDNA inserts in the positive clones were subcloned into the EcoRI site of the Bluescript KS vector (Stratacjene) and sequenced in both directions by the dideoxy chain termination method (Sanger et al., 1977). DNA and amino acid sequences were analyzed using DNASIS programs. The plexin sequences were compared with those of other proteins registered in the Swiss-Prot protein sequence database using the GENETYX program.

Northern Blotting Total cellular RNA was isolated from Xe~opus tadpole brains (stages 51-53) using the acid guanidinium-phenol-chloroform method (Chom- czynski and Sacchi, 1987), separated on lO/o agarose-formaldehyde gels, and blotted on to Hybond-N+ membrane (Amersham). Clone 8B (see Results) was labeled by random priming (Boehringer) and used as the probe. The hybridization was performed according to the standard method (Sambrook et al., 1989).

Amino Acid Sequencing of the Plexin Protein About 100 p~g of the plexin protein was affinity purified using MAbB2 from 5000 Xenopus tadpole brains as described (Ohta et al., 1992), separated by SDS-PAGE (50/o gel), then transferred to Immobilon membranes (Millipore). Proteins comigrating at 220 kDa and 160 kDa positions, which correspond to the native plexin molecule and its deg- radation product, respectively (Ohta et al., 1992), were visualized by staining with Ponceau S (Sigma), excised, then sequenced using a 470A Applied Biosystems Protein Sequencer.

Transfection of cDNA The complete 6.2 kb cDNA insert covering the putative plexin protein with both the 5' and 3' untranslated reg!ons (0.1 and 0.3 kb, respec- tively) or the mutant plexin cDNA, in which the sequence encoding

the C-terminal 595 amino acid residues had been deleted with EcoRI, were ligated into the cloning site of the mammalian expression vector pSVL (Pharmacia), respectively. L cells were cultured in Ham's F12 and Dulbecco's modified Eagle's medium (1:1) supplemented with 10% fetal calf serum and antibiotics, then transfected with either 20 I~g pSVL vector (without insert) or pSVL-plexin plasmids by standard calcium phosphate coprecipitation (Chen and Okayama, 1987). To obtain stable transfectants, the cells were cotransfected with pSTneoB (Katoh et al., 1987) and selected for G418 resistance (400 p.g/ml) (Sigma).

Immunohistochemistry and Immunoblotting To stain living transfectants, the cells were incubated with the mono- clonal antibody MAbB2 (50 I~g/ml in culture medium) for 30 min at 37°C, fixed with 20/0 Paraformaldehyde in PBS, then reacted with fluo- rescein isothiocyanate-conjugated anti-mouse immunoglobuUn (Cap- pel). Some cultures were processed for immunohistochemistry after fixing with 2% paraformaldehyde. The cells were examined with an epifluorescence microscope (UFX-IIA, Nikon). The procedures for the immunostainings of frozen sections of Xenopus nervous tissueswith MAbB2 were as reported (-rakagi et al., 1987; Ohta et al., 1992).

The plexin proteins, which had been affinity purified from Xenopus tadpole brains at stage 53, and the transfectants were suspended in SDS-PAGE sample buffer, then separated by SDS-PAGE (5% acryl- amide; Laemmli, 1970). The proteins were blotted onto nitrocellulose filters (Towbin et al., 1979) by a semidry Blotter (Bio-Rad). The filters were blocked with 5% skimmed milk and incubated with a guinea pig anti-plexin antiserum. The filters were washed and incubated with 200x diluted biotinylated secondary antibodies against guinea pig immunoglobulin (Zymed). After washing, the filters were reacted with peroxidase-conjugated streptavidin (Amersham). The immunoreactiv- ity was developed using the ECL detection system (Amersham).

Cell Aggregation and Adhesion The transfectants were dissociated into single cells with 1 mM EDTA in Ca 2+- and Mg2+-free Hanks' solution (CMF), then suspended in 10 mM HEPES-buffered CMF (HCMF) with or without 2 mM CaCI2 (or MgCI2). The cell suspension was kept on ice to inhibit cell aggregation. Cell suspension (500 p~l; 1 x 106 cells/ml) was placed in each well of a 24-well microtiter plate precoated with 5% bovine serum albumin, then agitated at 80 rpm for 60 rain at 37°C. Aliquots were removed at 15 min intervals, and the number of particles were counted using a hemocytometer. The extent of the aggregation was represented by the index Nt/No, where Nt and No were the total particle number at incubation times t and 0, respectively (Takeichi, 1977). For the antibody blocking test, the transfectants were incubated with 0.4 mg/ml guinea pig anti-plexin antibody (Fab fragment; see Ohta et al., 1992) for 30 min at 0°C before starting the cell aggregation. To examine whether plexin mediates cell adhesion via a homophilic binding mechanism, binding of the transfectants to the immobilized plexin proteins was tested. The plexin proteins were affinity purified from Xenopus tadpole brains and spotted on nitrocellulose-coated dishes for 1 hr at 37°C in a CO2 incubator as reported (Lemmon et al., 1989; Tanaka et al., 1991). The dishes were blocked with 5% skimmed milk in PBS at 4°C overnight, washed several times with HCMF, and then 2 ml of cell suspension (t x 106 cells/ml) was added. The dishes were kept in a CO2 incubator for 30 rain at 37°C.

In Situ Hybridization In situ hybridization was performed as reported (-Iakagi et al., 1991). In brief, frozen sections of Xenopus tadpoles at stages 43-51 (18 p~m thick) were collected on polylysine-coated microscope slides and incubated with either sense or antisense ~S-labeled RNA probes that were transcribed from a plexin cDNA template (clone 8B; see Results). Hybridization proceeded for 18 hr at 42°C in a mixture consisting of 50°/0 formamide, 0.3 M NaCI, 20 mM Tris-HCI (pH 7.5), 5 x Denhardt's solution, sheared salmon sperm DNA (0.25 p.g/pl), yeast tRNA (1 ~g/ p.I), 10% dextran sulfate, 10 mM dithiothreitol, and 105 dpm/ml RNA probe. The sections were washed, dehydrated, dipped in NTB-2 nu- clear track emulsion (Eastman Kodak), and exposed for 1-5 weeks.

Neuron 1198

Acknowledgments

This work was supported by grants from the Ministry of Education, Science, and Culture, Japan, and by a grant from the Uehara Founda- tion of Japan.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC Section 1734 solely to indicate this fact.

Received November 7, 1994; revised March 31, 1995.

References

Agata, K., Yasuda, K., and Okada, T. S. (1983). Gene coding for a lens specific protein, 8-crystalline, is transcribed in nonlens tissue of chick embryos. Dev. Biol. 100, 222-226.

Bastiani, M. J., Harrelson, A. L., Snow, P. M., and Goodman, C. S. (1987). Expression of fasciclin I and II glycoproteins on subsets of axon pathways during neuronal development in the grasshopper. Cell 48, 745-755.

Bottaro, D. P., Rubin, J. S., Faletto, D. L., Chan, A. M.-L., Kmiecik, T. E., Vande Woude, G. F., and Aaronson, S. A. (1991). Identification of the hepatocyte growth factor receptor as the c-met proto-oncogene product. Science 251,802-804. Burton, P. R. (1990). Vomeronasal and olfactory nerves of adult and larval bullfrogs. II. Axon terminations and synaptic contacts in the ac- cessory olfactory bulb. J. Comp. Neurol. 292, 624-637.

Burton, P. R., Coogan, M. M., and Borror, C. A. (1990). Vomeronasal and olfactory nerves of adult and larval bullfrogs. I. Axons and the distribution of their glomeruli. J. Comp. Neurol. 292, 614-623.

Carpenter, G. (1987). Receptors for epidermal growth factor and other polypeptide mitogens. Ann. Rev. Biochem. 56, 881-914.

Cervello, M., Lemmon, V., Landreth, G., and Rutishauser, U. (1991). Phosphorylation-dependent regulation of axon fasciculation. Prec. Natl. Acad. Sci. USA 88, 10548-10552.

Chan, A. M.-L., King, H. W. S., Deakin, E. A., Tempest, P. R., Hilkens, J., Kroezen, V., Edwards, D. R., Wills, A. J., Brookes, P., and Cooper, C. S. (1988). Characterization of the mouse met proto-oncogene. On- cogene 2, 593-599. Chao, M. V. (1992). Neurotrophin receptors: a window into neuronal differentiation. Neuron 9, 583-593. Chen, C., and Okayama, H. (1987). High efficiency transformation of mammalian cells by plasmid DNA. Mol. Cell. Biol. 7, 2745-2752.

Chomczynski, P., and Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extrac- tion. Anal. Biochem. 162, 156-159. Cooper, J. A., Esch, F. S., Taylor, S. S., and Hunter, T. (1984). Phos- phorylation sites in enolase and lactate dehydrogenase utilized by tyrosine protein kinases in vivo and in vitro. J. Biol. Chem. 259, 7835- 7841. Dodd, J., and Jessell, T. M. (1988). Axon guidance and the patterning of neuronal projection in vertebrates. Science 242, 692-699.

Doherty, P., Cohen, J., and Walsh, F. S. (1990)~ Neurite outgrowth in response to transfected N-CAM changes during development and is modulated by polysialic acid. Neuron 5, 209-219.

Doherty, P., Ashton, S. V., Moore, S. E., and Walsh, F. S. (1991). Morphoregulatory activities of NCAM and N-cadherin can be ac- counted for by G protein-dependent activation of L- and N-type neu- ronal Ca 2+ channels. Cell 67, 21-33. Edelman, G. M., and Cunningham, B. A. (1990). Place-dependent cell adhesion, process retraction, and spatial signaling in neural morpho- genesis. Cold Spring Harb. Symp. Quant. Biol. 55, 303-318. Elkins, T., Zinn., K., McAIlister, L., Hoffmann, F. M., and Goodman, C. S. (1990). Genetic analysis of a Drosophila neural cell adhesion molecule: interaction of fasciclin I and Abelson tyrosine kinase muta- tions. Cell 60, 565-575. Fujisawa, H., Ohtsuki, T., Takagi, S., and Tsuji, T. (1989). An aberrant retinal pathway and visual centers in Xenopus tadpoles share a corn-

mon cell surface molecule, A5 antigen. Dev. Biol. 135, 231-240.

Geiger, B., and Ginsberg, D. (1991). The cytoplasmic domain of ad- herens-type junctions. Cell Motil. Cytoskel. 20, 1-6.

Gennarini, G., Durbec, P., Boned, A., Rougon, G., and Goridis, C. (1991). Transfected F3/F11 neuronal cell surface protein mediates intercellular adhesion and promotes neurite outgrowth. Neuron 6, 595- 606.

Goodman, C. S., and Shatz, C. J. (1993). Developmental mechanisms that generate precise patterns of neuronal connectivity. Cell 721Neu- ron 10 (Suppl.), 77-98. Grenningloh, G., Bieber, A. J., Rehm, E. J., Snow, P. M., Traquina, Z. R., Hortsch, M., Patel., N. H., and Goodman, C. S. (1990). Molecular genetics of neuronal recognition in Drosophila: evolution and function of immunoglobulin superfamily cell adhesion molecules. Cold Spring Harb. Syrup. Quant. Biol. 55, 327-340.

Grenningloh, G., Rehm, E. J., and Goodman, C. S. (1991). Genetic analysis of growth cone guidance in Drosophila: fasciclin II functions as a neuronal recognition molecule. Cell 67, 45-57.

Hatta, K., Nose, A., Nagafuchi, A., and Takeichi, M. (1988). Cloning and expression of cDNA encoding a neural calcium-dependent cell adhesion molecule: its identity in the cadherin gene family. J. Cell Biol. 106, 873-881.

Hirano, S., Nose, A., Hatta, K., Kawakami, A., and Takeichi, M. (1987). Calcium-dependent cell-cell adhesion molecule (cadherins): subclass specificities and possible involvement of actin bundles. J. Cell Biol. 105, 2501-2510.

Hirano, S., Kimoto, N., Shimoyama, Y., Hirohashi, S., and Takeichi, M. (1992). Identification of a neural a-catenin as a key regulator of cadherin function and multicellular organization. Cell 70, 293-301.

Hirata, T., Takagi, S., and Fujisawa, H. (1993). The membrane protein A5, a putative neuronal recognition molecule, promotes neurite out- growth. Neurosci. Res. 17, 159-169.

Huff, J. L., Jelinek, M. A., Borgman, C. A., Lansing, T. J., and Parsons, J. T. (1993). The protooncogene c-sea encodes a transmembrane pro- tein-tyrosine kinase related to the Met/hepatocyte growth factor/scatter factor receptor. Proc. Natl. Acad. Sci. USA 90, 6140-6144.

Hunter, T. (1982). Synthetic peptide substrates for a tyrosine protein kinase. J. Biol. Chem. 257, 4843-4848.

Hunter, D. D., Porter, B. E., Bulock, J. W., Adams, S. P., Merlie, J. P., and Sanes, J. R. (1989). Primary sequence of a motor neuron-selective adhesive site in the synaptic basal lamina protein s-laminin. Cell 59, 905-913. Huynh, T. V., Young, R. A., and Davis, R. W. (1985). Construction and screening cDNA libraries in lambda gt11 and lambda gtl0. In DNA Cloning: A Practical Approach, Volume 1, D. M. Glover, ed. (Ox- ford: IRL Press), pp. 49-78.

Hynes, R. O. (1992). Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11-25. Jessell, T. M. (1988). Adhesion molecules and the hierarchy of neural development. Neuron 1, 3-13. Katoh, K., Takahashi, Y., Hayashi, S., and Kondoh, H. (1987). Im- proved mammalian vector for high expression of G418 resistance. Cell Struct. Funct. 12, 575-580. Kozak, M. (1989). The scanning model for translation: an update. J. Cell Biol. 108, 229-241. Kretsinger, R. H. (1980). Structure and evolution of calcium-modulated proteins. CRC Crit. Rev. Biochem. 8, 119-174. Kyte, J., and Doolittle, R. F. (1982). A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157, 105-132.

Laemmli, U. K. (1970). Cleavage of structural proteins during the as- sembly of the head of bacteriophage T4. Nature 227, 680-685.

Lemmon, V., Farr, K. L., and Lagenaur, C. (1989). Ll-mediated axon outgrowth occurs via a homophilic binding mechanism. Neuron 2, 1597-1603.

Matsunaga, M, Hatta, K., Nagafuchi, A., and Takeichi, M. (1988). Guidance of optic nerve fibres by N-cadherin adhesion molecules. Nature 334, 62-64.

Plexin: A Novel Neuronal Cell Adhesion Molecule 1199

Nagafuchi, A., Shirayoshi, Y., Okazaki, K., Yasuda, K., and Takeichi, M., (1987). Transformation of cell adhesion properties by exogenously introduced E-cadherin cDNA. Nature 329, 341-343.

Nagafuchi, A., and Takeichi, M. (1988). Cell binding function of E-cadherin is regulated by the cytoplasmic domain. EMBO J. 7, 3679- 3684.

Neugebauer, K. M., Tomaselli, K. J., Lilien, J., and Reichardt, L. F. (1988). N-cadherin, N-CAM and integrins promote retinal neurite out- growth on astrocytes in vivo. J. Cell Biol. 107, 1177-1187. Nieuwkoop, P. B., and Faber, J. (1956). Normal Tadpole of Xenopus laevis. (Amsterdam: Elsevier).

Nikundiwe, A. M., and Nieuwenhuys, R. (1983). The cell masses in the brainstem of the South African clawed frog Xenopus laevis: a topo- graphical and topological analysis. J. Comp. Neurol. 213, 199-219. Nose, A., Mahajan, V. B., and Doodman, C. S. (1992). Connectin: a homophilic cell adhesion molecule expressed on a subset of muscles and the motoneurons that innervate them in Drosophila. Cell 70, 553- 567.

Ohta, K., Takagi, S., Asou, H., and Fujisawa, H. (1992). Involvement of neuronal cell surface molecule B2 in the formation of retinal plexiform layers. Neuron 9, 151-161.

Ozewa, M., Engel, J., and Kemler, R. (1990). Single amino acid substi- tutions in one Ca 2+ binding site of uvomorulin abolish the adhesive function. Cell 63, 1033-1038

Park, M., Dean, M., Kaul, K., Braun, M. J., Gonda, M. A., and Woude, G. V. (1987). Sequence of MET protooncogene cDNA has features characteristic of the tyrosine kinase family of growth-factor receptors, Proc. Natl. Acad. Sci. USA 84, 6379--6383. Patschinsky, T., Hunter, T., Esch, F. S., Cooper, J. A., and Sefton, B. M. (1982). Analysis of the sequence of amino acids surrounding sites of tyrosine phosphorylation. Proc. Natl. Acad. Sci. USA 79, 973- 977.

Ronsin, C., Muscatelli, F., Mattei, M.-G., and Breathnach, R. (1993). A novel putative receptor protein tyrosine kinase of the met family. Oncogene 8, 1195-1202.

Russell, I. J. (1976). Amphibian lateral line receptors. In Frog Neurobi- ology: A Handbook, R. Llinas and W. Precht, eds. (Springer-Verlag), pp. 513-550.

Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular Clon- ing: A Laboratory Manual, Second Edition. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory).

Sanger, F., Nicklen, S=, and Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463- 5467.

Satoda, M., Takagi, S., Ohta, K., Hirata, T., and Fujisawa, H. (1995). Differential expression of two cell surface proteins, neuropilin and plexin, in Xenopus olfactory axon subclasses. J. Neurosci. 15, 942- 955. Scalia, F. (1976). Structure of the ol,~actory and accessory olfactory systems. In Frog Neurobiology: A Hardbook, R. Llin~s and W. Precht, eds. (Springer-Verlag), pp. 213-133.

Seilheimer, B., and Schachner, M. (1988). Studies of adhesion mole- cules mediating interactions between cells of peripheral nervous sys- tem indicate a major role for L1 in mediating sensory neuron growth on Schwann cells in culture. J. Cell Biol. 107, 341-351. Shelton, P. (1971). The structure and function of the lateral line system in larval Xenopus laevis. J. Exp. Zoot. 178, 211-231.

Shimamura, K., Takahashi, T., and Takeichi, M. (1992). E-cadherin expression in a particular subset of sensory neurons. Dev. Biol. 152, 242-254.

Snow, P. M., Bieber, A. J., and Goodman, C. S. (1989). Fasciclin II1: a novel homophilic adhesion molecule in Drosophila. Cell 59, 313- 323.

Takagi, S., Tsuji, T , Amagai, T., Takamatsu, T., and Fujisawa, H. (1987). Specific cell surface labels in the visual centers of Xenopus laevis tadpole identified using monocional antibodies. Dev. Biol. 122, 90-100.

Takagi, S., Hirata, T., Agata, K., Mochii, M., Eguchi, G., and Fujisawa,

H. (1991). The A5 antigen, a candidate for the neuronal recognition molecule, has homologies to complement components and coagula- tion factors. Neuron 7, 295-307. Takeichi, M. (1977). Functional correlation between cell adhesive prop- erties and some cell surface proteins. J. Cell Biol. 75, 464-474.

Takeichi, M. (1991). Cadherin cell adhesion receptors as a morphoge- netic regulator. Science 251, 1451-1455. Takeichi, M., Inuzuka, H., Shimamura, K., Fujimori, T., and Nagafuchi, A. (1990). Cadherin subclass: differential expression and their roles in neuronal morphogenesis. Cold Spring Harb. Symp. Quant. Biol. 55, 319-325. Tanaka, H., Matsui, T., Agata, A., Tomura, M., Kubota, I., McFarland, K. C., Kohr, B., Lee, A., Phillips, H. S., and Shelton, D. L. (1991). Molecular cloning and expression of a novel adhesion molecule, SCI. Neuron 7, 535-545. Towbin, H., Staehelin, T., and Gordon, J. (1979). Electrophoretic trans- fer of proteins from acrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76, 4350-4354. UIIrich, A., Coussens, L., Hayflick, J. S., Dull, T. J., Gray, A., Tam, A. W., Lee, J., Yarden, Y., Libermann, T. A., Schlessinger, J., Down- ward, J., Mayers, E. L., Whittle, N., Waterfield, M. D., and Seeburg, P. H. (1984). Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells. Nature 309, 418-425. UIIrich, A., Bell, J. R., Chen, E. Y., Herrera, R., Petruzzelli, L. M., Dull, T. J., Gray, A., Coussens, L., Liao, Y.-C., Tsubokawa, M., Mason, A., Seeburg, P. H., Grunfeld, C., Rosen, O. M., and Ramachandran, J. (1985). Human insulin receptor and its relationship to the tyrosine kinase family of oncogenes. Nature 313, 756-761.

von Heijne, G. (1986). A new method for predicting signal sequence cleavage sites. Nucl. Acids Res. 14, 4683-4690.

GenBank Accession Number

The nucleotide sequence data reported in this paper will appear in the GSDB, DDBJ, EMBL, and NCBI nucleotide sequence databases with the accession number D38175.