MOLECULAR AND CELLULAR BIOLOGY, Dec. 2007, p. 8797–8806 Vol. 27, No. 240270-7306/07/$08.00�0 doi:10.1128/MCB.01355-07Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Haploinsufficiency of c-Met in cd44�/� Mice Identifies aCollaboration of CD44 and c-Met In Vivo�

Alexandra Matzke,1 Vardanush Sargsyan,2 Bettina Holtmann,3 Gayane Aramuni,2 Esther Asan,4Michael Sendtner,3 Giuseppina Pace,1 Norma Howells,1 Weiqi Zhang,2

Helmut Ponta,1 and Veronique Orian-Rousseau1*Forschungszentrum Karlsruhe, Institute for Toxicology and Genetics, Postfach 3640, 76021 Karlsruhe, Germany1; Department of

Neurophysiology and DFG-Research Center of Molecular Physiology of Brain, University of Gottingen, Humboldtallee 23,37073 Gottingen, Germany2; Institute for Clinical Neurobiology, University of Wurzburg, Josef Schneider Strasse 11,

97080 Wurzburg, Germany3; and Institute of Anatomy and Cell Biology, University of Wurzburg,Koellikerstrasse 6, 97070 Wurzburg, Germany4

Received 27 July 2007/Accepted 27 September 2007

Recent evidence has shown that the activation of receptor tyrosine kinases is not only dependent on bindingof their ligands but in addition requires adhesion molecules as coreceptors. We have identified CD44v6 as acoreceptor for c-Met in several tumor and primary cells. The CD44v6 ectodomain is required for c-Metactivation, whereas the cytoplasmic tail recruits ERM proteins and the cytoskeleton into a signalosomecomplex. Here we demonstrate that c-Met (and hepatocyte growth factor and Gab1) is haploinsufficient in acd44�/� background, as the cd44�/�; met�/� (and cd44�/�; hgf�/� and cd44�/�; gab1�/�) mice die at birth.They have impaired synaptic transmission in the respiratory rhythm-generating network and alterations in thephrenic nerve. These results are the first genetic data showing that CD44 and c-Met collaborate in vivo and thatthey are involved in synaptogenesis and axon myelination in the central and peripheral nervous systems.

After ligand binding, receptor tyrosine kinases (RTKs) areactivated and induce physiological events such as proliferation,migration, differentiation, or apoptosis (reviewed in reference25). For the majority of RTKs, the activation step was thoughtto be exclusively dependent on binding of the ligand, althoughfor members of the fibroblast growth factor receptor (FGF-R)family an involvement of a coreceptor, namely heparansulfatemodified proteins (HSPG), has long been proposed (39, 44). Inthis case it is believed that the ligand has to bind to the HSPGand the complex then activates the receptor. This concept of acoreceptor function for RTKs has expanded recently and ap-pears to have a much broader relevance. A variety of adhesionmolecules such as N-cadherin, N-CAM, or integrins can inter-act with FGF-Rs, vascular endothelial growth factor receptor,GFR�1, or c-Met to facilitate or enhance receptor activation(2, 5, 21, 34, 36). Particularly interesting is the family of CD44adhesion molecules. CD44s or CD44v3 isoforms collaboratewith epidermal growth factor receptors (29, 30). We haveshown that CD44v6 acts as a coreceptor for c-Met. We havedemonstrated that the expression of CD44 isoforms containingexon v6 is a prerequisite for c-Met activation by its ligandhepatocyte growth factor (HGF) in several tumor cells and inprimary cells (19). CD44v6, HGF, and c-Met form a ternarycomplex and the CD44v6 extracellular portion is required forc-Met activation, whereas the cytoplasmic tail recruits ERMproteins (ezrin, radixin, and moesin) and the cytoskeleton into

a signalosome complex that mediates signal transduction (19,20).

The experiments performed on cancer cell lines suggest acollaboration between CD44v6 and c-Met in the metastaticprocess. The expression of CD44v6 in otherwise nonmetastatictumor cells rendered these cells responsive to HGF (19) andmade them metastatic (7). Data obtained from knockout miceseem to contradict that such a collaboration occurs in animals.The c-Met (and the HGF) knockout mice are embryonic lethal(1, 26, 37). They die between embryonic day 12.5 (E12.5) andE16.5 from a placental defect. The migration of the myogenicprecursors is impaired so that organs like the tongue, thediaphragm, or the limbs are not formed. In addition, the liveris severely damaged. This is in striking contrast to the CD44knockout mice. Although the activation of c-Met in primarykeratinocytes is strictly dependent on CD44v6 (19) and limboutgrowth relies on CD44v3 heparansulfated isoforms (29),the CD44-null mice showed no overt phenotype during devel-opment. They showed only mild abnormalities in myeloid pro-genitor migration, bone marrow colonization (27), and lack ofhoming of lymphocytes to lymph nodes or to the thymus (22).

An explanation for this striking difference between theCD44 knockout mice and the c-Met and HGF knockout micecould be that the function(s) of CD44 is substituted by anotherprotein in the CD44-null mice (no other CD44-related proteinhas so far been detected). This hypothesis is strongly supportedby the data obtained for another type of “knockout” mouse inwhich CD44 was downregulated by means of CD44 antisensesequences expressed under the control of the keratinocyte K5promoter (10). Accordingly, these mice do not express CD44in the keratinocytes. Surprisingly and in clear contrast to thetotal CD44 knockout, the newborn mice have severe skin al-terations, such as a delay in wound healing, in local inflamma-

* Corresponding author. Mailing address: Forschungszentrum Karlsruhe,Institute for Toxicology and Genetics, Postfach 3640, 76021 Karlsruhe,Germany. Phone: 497247826523. Fax: 497247823354. E-mail: veronique.orian-rousseau@itg.fzk.de.

� Published ahead of print on 8 October 2007.

8797

tory responses, and in hair regrowth. Since the K5 promoter isturned on around day 10 during embryogenesis, we assumethat CD44 functions can be substituted during early embryo-genesis whereas at later times (when the K5 promoter becomesactive) CD44 can no longer be substituted and knocking downof CD44 is then detrimental for the animals. This notion isconfirmed by the fact that the only overt phenotypes observedin the CD44 total knockout mice are visible at the adult stage;for example, the maintenance of postpartum lactation is im-paired (45).

Here we test the relevance of CD44/c-Met cooperation invivo and the substitution of the CD44v6 coreceptor function inCD44 knockout mice. We performed crossing experiments be-tween CD44 knockout mice and met (or hgf or gab1) heterozy-gotes and asked whether the cd44�/�; met�/� (or hgf�/� orgab1�/�) mice show haploinsufficiency and therefore a pheno-type. hgf and gab1 crosses are included, since both knockoutmice have a phenotype similar to the c-Met knockout mouse(1, 24, 26, 37). For HGF, the ligand for c-Met, this was ex-pected. Gab1 is a docking protein (reviewed in reference 12)that mediates most of the signaling pathways from c-Met (17,40) as it recruits Grb2, the p85 subunit of PI3K, phospholipaseC1, and SHP-2 to the receptor. The similarity of the knockoutof Gab1 with c-Met underlines the importance of this interac-tion.

The c-met (or hgf or gab1) heterozygotes in a cd44�/� back-ground (and in cd44�/�, shown here) develop normally andshow no overt phenotype. In striking contrast, the heterozy-gotes in a cd44�/� background show lethality at birth with apenetrance reaching 70%. The animals cannot breathe andtheir lungs are not inflated. This defect results most likely froman impaired synaptic transmission in the brain stem respiratoryrhythm-generating network and from alterations in the phrenicnerve that innervates the diaphragm. This haploinsufficiency ofmet (or hgf or gab1) in the cd44�/� mice can only be explainedif CD44 and c-Met/HGF/Gab1 cooperate in vivo. We give herethe first genetic evidence that these molecules collaborate inembryogenesis.

MATERIALS AND METHODS

Mouse strains, genotyping. cd44�/� mice (27) were backcrossed with C57BL/6mice for 10 generations and then used for crossings with met�/�, gab1�/�, orhgf�/� mice (1, 26, 37) that had also been backcrossed with C57BL/6 mice.Genotyping using mouse tails was performed with the following primers in thePCR: for met inactivation, NEO1 (5�CTTGCGTGCAATCCATCTTGTTCAATG3�) and MET1 (5�CACTGAGCCCAGAAGAGCTAGTGG3�); for endoge-nous met, MET2 (5�GTACACTGGCTTGTACAATGTACAGTTG3�) andMET3 (5�CTTTTTCAATAGGGCATTTTGGCTGTG3�); for gab1 inactiva-tion, NEO2 (5�TTGTTTTTCGAGCTTCAAGGTTCAT3�) and GAB1 (5�CCCTTTGTGGATGGCTTCTTTGT3�); for endogenous gab1, GAB1 and GAB2(5�TTCTTGGCATGATCGTTTTTGTAA3�); for hgf inactivation, NEO1 andHGF1 (5�CCCGCAGAGGTATATTGTGTTGTCC3�); for endogenous hgf,HGF1 and HGF2 (5�CTGTTCCTGATACACCTGTTGGCAC3�); for cd44 in-activation, CD441 (5�CGCAGGTGTATTCCATGTGG3�) and NEO3 (5�ACGTTGTCACTGAAGCGGG3�); and for endogenous cd44, CD441 and CD443(5�ACTGATATGACCCTAATGGCTTCC3�).

Histology. Newborn mice were sacrificed, fixed in Bouin’s solution for 24 h andembedded in paraffin. Serial sagittal sections of 6 �m were stained with hema-toxylin and eosin. The histological analysis was performed by Frimorfo, Fribourg,Switzerland.

In situ hybridization. HGF and cMet cDNAs were a kind gift from Carmenand Walter Birchmeier, MDC, Berlin, Germany (32). The CD44 cDNA probecorresponds to positions 1 to 690 of rat cDNA (7). This region has 98%homology with the mouse sequence. Mouse brains were prepared, fixed for

16 h in 4% paraformaldehyde (PFA), and embedded in paraffin. In situhybridization on 6-�m sections was performed with a digoxigenin-labeledRNA probe synthesized with a Dig RNA labeling kit (Roche, Mannheim,Germany). The probes were used at a concentration of 1 �g/ml. Hybridizationconditions have been described previously (31). Hybridized RNA was de-tected with alkaline phosphatase-coupled digoxigenin antibodies according tothe manufacturer’s protocol (Roche).

Analysis of motoneurons in the facial nucleus and nucleus ambiguus. Wedetermined the number of motoneuron cell bodies in the facial nucleus andnucleus ambiguus of cd44�/�; met�/� double mutant and cd44�/�; met�/� con-trol mice that were delivered by cesarean section at E19. Similar analyses wereperformed with 3-week-old mice applying established techniques (18). Animalswere perfused transcardially with 4% PFA in 0.1 M phosphate-buffered saline atpH 7.4. The brain stem containing the facial nucleus and nucleus ambiguus wasdissected, and 7-�m paraffin sections were prepared. After Nissl staining, mo-toneurons were counted in every fifth section, and the raw counts were correctedfor split nuclei as has been described previously (13).

Phrenic nerves. We analyzed phrenic nerve fibers in E19 and 3-week-oldcd44�/�; met�/� mice as well as in cd44�/�; met�/� and cd44�/�; met�/� controlmice. Animals were perfused as described above. The phrenic nerves were thendissected, and in 3-week-old mice, the proximal and distal parts of the nerve wereseparately processed. The nerves were postfixed for 3 h in 0.1 M sodium caco-dylate buffer containing 4% PFA and 2% glutaraldehyde and kept overnight insodium cacodylate buffer containing only 4% PFA. After osmification and de-hydration, all samples were embedded in Spurr’s medium. From phrenic nervesof 3-week-old mice, semithin (1-�m) cross sections were cut with a diamondknife on an ultramicrotome. Sections were then stained with azur-methyleneblue. The number of intact myelinated fibers was determined from photographstaken from nerve cross sections under a Zeiss light microscope equipped with aZeiss HRC digital camera. For morphometric analysis of E19 phrenic nerveaxons, we prepared ultrathin (�80-nm) sections from the resin-embeddedphrenic nerves, floated on Formvar-coated single-slot nickel grids, contrastedwith uranyl acetate and lead citrate (23). Sections were then analyzed in a LEO912 AB transmission electron microscope (Zeiss SMT, Oberkochen, Germany).High-resolution overview images of nerve cross sections were acquired usinga slow-scan charge-coupled-device camera system and the multiple imagealignment function of the Universal TEM Imaging Platform ITEM (SoftImaging System, Munster, Germany). These images allow detailed qualitativeand quantitative analyses of entire nerve cross sections. As a measure foraxon size, the circumference of axons was determined and used to calculateaxonal diameter.

Electrophysiological recordings in brain stem slices. All electrophysiologicalanalyses were performed in a blinded manner on brain stem neurons of mice.Slices containing the pre-Botzinger complex (preBotC) from late embryoniclittermate mice (E18 to E19) were used for whole-cell recordings (46). Briefly,the bath solution in all experiments consisted of (in mM) 118 NaCl, 3 KCl, 1.5CaCl2, 1 MgCl2, 25 NaHCO3, 1 NaH2PO4, and 5 glucose at pH 7.4, which wasaerated with 95% O2 and 5% CO2 and kept at 28°C. The pipette solution forpatch-clamp recording contained (in mM) 140 K-gluconate (for glutamatergicpostsynaptic current [PSC]) or 140 KCl (for GABAergic/glycinergic PSC), 1CaCl2, 10 EGTA, 2 MgCl2, 4 Na3ATP, 0.5 Na3GTP, 10 HEPES, pH 7.3. Spon-taneous GABAergic/glycinergic and glutamatergic PSCs were recorded fromneurons of the preBotC in 10 �M 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX)or 1 �M strychnine and 1 �M bicuculline, respectively. Spontaneous miniatureGABAergic/glycinergic and glutamatergic PSCs (miniature inhibitory PSCs[mIPSCs] and miniature excitatory PSCs [mEPSCs]) were recorded as describedabove, but in the presence of 0.5 �M tetrodotoxin. Generally, signals withamplitudes of at least two times above the background noise were selected, andthe statistical significance was tested in each experiment. In all tested animals,there were no significant differences in the noise levels between different geno-types. The PSCs were amplified and filtered by a four-pole Bessel filter at acorner frequency of 2 kHz and digitalized at a sampling rate of 5 kHz using theDigiData 1200B interface (Axon Instruments). Data acquisition and analysiswere done using commercially available software (pClamp 9 and AxoGraph 4.6[Axon Instruments] and Prism 4 software [GraphPad]).

Statistical analysis. Statistical data are expressed as means � standard errorsof the mean (SEM). The statistical significance of the differences between meanswas assessed with Student t tests (InStat; GraphPad Software Inc.). The level ofsignificance was set at P of �0.05.

Statistical analysis of the motoneuron and axon counts was performed using atwo-tailed test, and P values of �0.05 were considered significant.

Statistical analysis of size distribution of axonal profiles was performed usingthe Mann-Whitney U test.

8798 MATZKE ET AL. MOL. CELL. BIOL.

RESULTS

cd44�/�; met�/� (or hgf�/� or gab1�/�) animals die at birth.In order to test whether c-Met is haploinsufficient in the con-text of CD44 knockout mice, we crossed cd44�/�; met�/� micethat were healthy and showed no phenotype with cd44�/�;met�/� mice (Table 1). cd44�/�; met�/� mice were born withthe expected Mendelian ratio (36 out of 152 progenies). How-ever, shortly after birth, 70% of these animals had breathingdifficulties and died. Similar experiments were performed withhgf as well as gab1 heterozygote mice. In the case of hgf, 208progenies were analyzed, and 60% of cd44�/�; hgf�/� micedied. In the case of the gab1 crossing (143 progenies), 69% ofthe cd44�/�; gab1�/� animals did not survive (Table 1). Incontrast, all animals that were heterozygous for cd44 and formet, hgf, or gab1 survived and developed normally.

These results imply that the three molecules used in thecrossings are involved in the same pathways. Their haploinsuf-ficiency in the context of CD44-null mice shows that thesemolecules are connected to CD44 in wild-type animals.

A possible explanation for the survival of 30% of thecd44�/�; met�/� mice could be an upregulation of c-Met in thesurviving animals. Therefore, we determined the amount ofc-Met in liver lysates by Western blot analysis. Livers wereisolated from animals at birth. From the same animals, thelungs were tested for floating on water (see below). Livers ofthe cd44�/�; met�/� mice show higher amounts of c-Met com-pared to amounts for the cd44�/�; met�/� mice, irrespective ofwhether the lungs of cd44�/�; met�/� animals sank or floatedon water (data not shown). Thus the survival of these animalsappears not to be due to an upregulation of c-Met.

The cd44�/�; met�/� animals that survived developed in allrespects normally and in particular were fertile. To evaluatewhether these animals gained an unrelated genetic alterationthat accounts for their survival, we crossed them with cd44�/�;met�/� animals and monitored again the phenotypes of prog-enies. The majority of the cd44�/�; met�/� progenies also diedat birth, similar to the first crosses (Table 1).

cd44�/�; met�/� animals die at birth from a lung defect. Toanalyze phenotypic changes that might account for the haplo-insufficiency, an overall histological analysis of whole embryos

was performed. Serial sagittal sections through entire embryoswere done, followed by hematoxylin and eosin staining. Twocontrol animals (cd44�/�; met�/�) were compared to twocd44�/�; met�/� animals that showed breathing difficulty atbirth. The gross anatomy of the most vital organs, such asheart, kidney, as well as the overall structure of the brain, didnot show any obvious abnormalities that might explain thelethal phenotype. The lungs of the two cd44�/�; met�/� mice,however, showed multifocal atelectasis that was obviouslycaused by a primary asphyxia (Fig. 1A).

We next performed a simple test to demonstrate whetherthe lungs were inflated. The lungs of newborn animals wereremoved and deposited in flasks containing water. We rea-soned that the lungs from the cd44�/�; met�/� mice that even-tually would die should sink. In 76% of the cd44�/�; met�/�

animals, the lungs sank, indicating that they were not inflated(Table 1 and Fig. 1B). All the control lungs floated. The per-centage of cd44�/�; met�/� animals affected in the floatingassay corresponded well to the percentage of animals that died(Table 1).

Nerve fiber morphology in the phrenic nerve is changed inthe cd44�/�; met�/� mice. Explanations for the lethal asphyxiain cd44�/�; met�/� mice might be either pathological alter-ations of the peripheral organs along the airway, such as lung,diaphragm, etc., or neuronal impairments in the peripheraland/or central nervous system related to respiration. The his-tological analysis revealed that the general anatomy of theintercostal muscles and the muscle fibers of the diaphragm wasnormal (data not shown). We determined the number of mo-toneuron cell bodies in the facial nucleus and nucleus am-biguus in mice at E19. We examined four animals of eachgenotype. We could not find any difference in motoneuronnumbers between cd44�/�; met�/� and control mice in thefacial nucleus or in the nucleus ambiguus. Also, the morphol-ogy of motoneurons appeared similar (see example in Fig. 2A).Similar results were obtained with 3-week-old cd44�/�; met�/�

mice (that survived) and cd44�/�; met�/� control mice.In the phrenic nerve, however, we detected differences.

FIG. 1. A lung defect in cd44�/�; met�/� mice. (A) Histology oflungs obtained shortly after birth from mice with the indicated geno-types. (B) Floating test for lungs on water.

TABLE 1. c-Met haploinsufficiency in the context of CD44

Genotype and testparameter

No. ofprogeny

No. ofcd44�/�;x�/� (%)

No.dead(%)

No. withlungs thatsank (%)

Cross of cd44�/�; x�/�

and cd44�/�; x�/�

Survivalx met 152 36 (24) 25 (70)x hgf 208 56 (27) 33 (60)x gab1 143 32 (22) 22 (69)

Lung assayx met 85 21 (25) 16 (76)x hgf 97 23 (23) 12 (52)x gab1 84 21 (25) 13 (62)

Backcross of cd44�/�; x�/�

and cd44�/�; x�/�

Survival, x met 55 13 (24) 9 (69)Lung assay, x met 63 19 (30) 8 (42)

VOL. 27, 2007 COOPERATION OF CD44 AND c-Met IN VIVO 8799

FIG. 2. Motoneurons in facial nucleus and nucleus ambiguus and nerve fibers in phrenic nerves. (A) Nissl-stained paraffin sections ofmotoneurons in the facial nucleus and nucleus ambiguus of the brain stem of E19 and 3-week-old mice with the indicated genotypes and semithin

8800 MATZKE ET AL. MOL. CELL. BIOL.

Electron microscopy of phrenic nerve cross sections of E19control mice showed that most of the axons were completelyencircled and were thus segregated from each other bySchwann cells (Fig. 2B, panels a and c). Individual Schwanncell profiles were surrounded by a basal lamina and collagen(Fig. 2B, panel c). In isolated Schwann cell profiles, 54% �4.5% ensheathed one axon (Fig. 2B, panels a and c), 40% �3.13% ensheathed 2 to 4 axons, and 5% � 1.98% ensheathed5 to 10 axons. Only 0.5% � 1.8% of isolated Schwann cellprofiles were found to ensheath bundles of axons in whichmore than 10 axons were in close apposition without interven-ing Schwann cell processes (Fig. 2B, panel a). Occasionally, weobserved that myelination was just starting (Fig. 2B, panel c).

In contrast, in cross sections of phrenic nerves of cd44�/�;met�/� mice individual Schwann cell profiles appeared to en-sheath, on average, more axons than in control mice (Fig. 2B,panel b). Significantly less Schwann cell profiles than in con-trols ensheathed only one axon (28.0% � 6.0%; P � 0.001),43.0% � 1.5% ensheathed 2 to 4 axons, and 24.0% � 5.6%ensheathed 5 to 10 axons. The number of Schwann cells asso-ciated with bundles of axons without intervening Schwann cellprocesses was increased to 5.0% � 0.6% (Fig. 2B, panels b andd). Moreover, axons that were ensheathed by a single Schwanncell profile showed comparatively large variations in size, whilein controls they usually displayed similar diameters. The num-ber of segregated axons (axons completely ensheathed bySchwann cell profiles) in phrenic nerves of cd44�/�; met�/�

mice was significantly reduced by 20% (Table 2) (192 � 15,n 4, compared to 240 � 12, n 5, in cd44�/�; met�/� orcd44�/�; met�/� mice). In addition the percentage of segre-gated axons with a diameter size between 0.75 to 1.00 �m was14.96% � 1.41% in cd44�/�; met�/� mice compared to27.41% � 4.36% in control mice (Table 2). In contrast, thepercentage of small (�0.75-�m) and large (1.00-�m) caliberaxons was only slightly increased in cd44�/�; met�/� mice(Table 2).

In summary, these data suggest that the mechanisms foraxonal segregation by Schwann cells, myelination, and axonsize distribution are impaired in cd44�/�; met�/� mice. Inter-estingly, in cd44�/�; met�/� mice that survived we could notfind any change in axon numbers or Schwann cell morphologyin the proximal or distal phrenic nerves (Fig. 2A).

Severe impairment of overall network activity in preBotCneurons in cd44�/�; met�/� mice. Another possibility thatmight cause a lethal phenotype is a failure in the central res-piratory control system, as was shown previously (15, 38). Wetherefore performed in situ hybridization experiments to testthe expression of the respective proteins in the central respi-ratory system. Indeed, our results revealed coexpression ofCD44, c-Met, and HGF in the ventral respiratory group withinthe brain stem (Fig. 3), suggesting that these molecules mightcooperate. Thus it is possible that alterations in these tissues inthe cd44�/�; met�/� mice could have an impact on the centralrespiratory system.

We therefore started a series of electrophysiological exper-

cross sections of the phrenic nerves of 3-week-old mice. Scale bars for facial motoneurons equal 500 �m, for motoneurons of the nucleusambiguous, 200 �m, and for phrenic nerves, 50 �m. (B) Digitally generated overview (a, b) and high-magnification images (c, d) of phrenic nervecross sections of E19 mice. In phrenic nerves of cd44�/�; met�/� mice (a, c), a 1:1 ratio of Schwann cell to axon is very frequent, as indicated inthe boxed area in panel a and by arrows in panel c. Occasionally, early myelination is found (white arrowhead in panel c). Bundles of axons thatare in direct apposition and surrounded by a common Schwann cell sheath (arrow in panel a) are comparatively rare. In phrenic nerves of cd44�/�;met�/� mice, a 1:1 Schwann cell-to-axon ratio is found only occasionally (arrow in panel b), while bundles are more often seen (arrowheads inpanels b and d). Scale bars for panels a and b, 5 �m; for c and d, 1 �m.

FIG. 3. In situ hybridization of brain tissue. In situ hybridization ofsagittal brain sections of mouse pups at birth with CD44, HGF, andc-Met probes. VRG indicates the ventral respiratory group. An over-view of the whole brain section is shown in the left panels, the area ofthe VRG is shown in higher magnification in the right panels. Hybrid-ization with sense CD44 mRNA is shown in the lower panels (ctrl).Hybridization with sense HGF and c-Met mRNA gave similar results.

TABLE 2. Phrenic nerves at E19

Mouse typeNo. of

segregatedaxons

Fiber distribution (%)

0.75–1 �m �0.75 �m 1 �m

cd44�/�; met�/� 240 � 12 27.41 � 4.63 53.72 � 1.32 18.85 � 7.8cd44�/�; met�/� 192 � 15 14.96 � 1.41 60.73 � 2.35 24.28 � 2.23

VOL. 27, 2007 COOPERATION OF CD44 AND c-Met IN VIVO 8801

iments to find possible failures in the central respiratory con-trol system that might cause the lethal phenotype in cd44�/�;met�/� mice. We first monitored the overall spontaneouspostsynaptic currents (sPSCs) using whole-cell recordings inpreBotC neurons. These currents represent the general activitylevel of the respiratory network. The activity in preBotC neu-rons was diminished in cd44�/�; met�/� mice compared to thatin the control littermates (Fig. 4A and B). It is interesting tonote that the activity levels in the preBotCs of individualcd44�/�; met�/� mice were heterogenous. Four of seven miceshowed an activity level that was lower than 25% of the controllevel (mice no. 1 to 4 in Fig. 4C). In three of seven mice theactivity levels ranged between 37% to 55% of the level incontrol mice (mice no. 5 to 7 in Fig. 4C).

Severe impairment of synaptic transmission in preBotCneurons in cd44�/�; met�/� mice. To further characterize thefailure in the respiratory network, we next analyzed the inhib-itory and excitatory PSCs in the cd44�/�; met�/� mice in com-parison with their littermates. The frequency of glycinergic andGABAergic miniature mIPSCs in preBotC neurons was signif-icantly decreased in cd44�/�; met�/� mice (Fig. 5A and B).In contrast, the amplitudes of mIPSCs were not changed(Fig. 5C).

The mEPSCs were measured using whole-cell recordings inpreBotC neurons. It is striking that the frequency of glutama-tergic mEPSCs in preBotC neurons was already decreased incd44�/�; met�/� mice, while they were even further decreased

in cd44�/�; met�/� mice (Fig. 6A and B). The amplitude ofmEPSCs was only moderately changed between littermateswith different genotypes (Fig. 6C). Thus, these data demon-strated that homozygous null mutation of CD44 showed asignificant effect on excitatory transmission that became evenmore severe by additional heterozygous mutation of c-Met.

In conclusion, these experiments show that the haploinsuf-ficiency of c-Met in the CD44-null background is based on adefect in synaptic transmission in the preBotC, suggesting thatin the CD44-expressing animals c-Met and CD44 cooperate insynaptogenesis.

DISCUSSION

Previous data have demonstrated that CD44v6 acts as acoreceptor for c-Met in transformed as well as primary cells(14, 19, 20). An important question remained as to whetherCD44 is essential for c-Met function in vivo. The data pre-sented here, that the majority of met�/� mice in a CD44-nullbackground did not survive in contrast to met�/� or met�/�

mice in a CD44-positive background, are the first genetic evi-dence that the coreceptor CD44 is essential for the function ofc-Met during embryogenesis in vivo.

A further conclusion drawn from our experiments is that theCD44 function for c-Met is substituted by another protein inthe CD44 knockout mice and that this protein does not func-tion as efficiently as CD44. This idea that functions of CD44

FIG. 4. Severe impairment of overall network activity in preBotC neurons in cd44�/�; met�/� mice. (A) Representative recordings of sPSCsin brain stem preBotC neurons. (B) Average frequency of sPSCs. Data are expressed as means � SEM. Statistical significance of differencesbetween means was assessed with Student t tests (InStat; GraphPad Software Inc.). The level of significance was set at P of �0.05. The numbersindicate the number of neurons/mouse. (C) The sPSC average frequency of cd44�/� mice in panel B was set to 100% and was compared to thesPSC frequency of individual cd44�/�; met�/� mice.

8802 MATZKE ET AL. MOL. CELL. BIOL.

might be substituted in the CD44 knockout mouse is welldocumented. Ablation of CD44 expression in the skin shows adrastic phenotype that is not at all observed in total CD44knockout mice (10). From this observation two conclusions canbe drawn, (i) CD44 functions are substituted in CD44 totalknockout animals and (ii) a substitution cannot occur at latertimes in embryogenesis, namely at day 10 when the expressionof CD44 in the skin is turned off (10). In addition Rhamm, areceptor for hyaluronic acid, has been shown to take over afunction of CD44 in the CD44 knockout mice. In the absenceof CD44, Rhamm is able to react with hyaluronan and triggersan enhanced inflammatory response to collagen-induced ar-thritis (16). Furthermore, we have recently identified ICAM-1as a protein that takes over the function of CD44 as a core-ceptor for c-Met in human hepatoma cells and in the liver ofCD44-null mice (our unpublished data).

That the decreased concentration of collaborating moleculesmight lead to haploinsufficiency was already successfully dem-onstrated in the case of Gab1 knockout mice (24). In gab1�/�

embryos, the migration of muscle precursor cells into the limbsand the diaphragm is strongly reduced but not completelyblocked. This is in contrast to the hgf�/� and met�/� mice andalso to the gab1�/�; met�/� mice where the phenotype is muchmore drastic. In the context of the Gab1 knockout mice, c-Metis haploinsufficient because it cannot transmit some crucialsignals (24). Another example of haploinsufficiency in signaling

pathways is the rescue of embryonic lethality of pten�/� miceby grb2 heterozygosity (4).

In general, there might be many different causes for thelethal phenotype of the met�/� mice in the CD44-null back-ground. One reason might be failures in the organogenesis ofone or several peripheral organs. The gross anatomy of themost vital organs, such as heart, lung, kidney, as well as theoverall structure of the brain, did not show any obvious abnor-malities that might explain the lethal phenotype.

Although the gross anatomy of the lung seems to be intact,the lungs of the lethal met�/� mice in the CD44-null back-ground were never inflated. As all organs and tissues along theairways were intact, there are no physical or aerodynamic rea-sons why respiratory movement and gas exchanges never takeplace in these animals. This suggests that the lethal phenotypewas caused by failures in neuronal activity.

The respiratory rhythm is generated by a brain stem respi-ratory network. Among other parts of the brain, the preBotCin the brain stem is an essential component of the centralrespiratory rhythm-generating network. The presented datashow that the overall network activity of the preBotC isstrongly compromised. The reduction of network activity variesbetween different cd44�/�; met�/� animals from 45 to 95% ofthe level in control animals. Previous studies of neurexin andneuroligin triple knockout mice (15, 38) demonstrated that areduction of network activity of 45% was necessary to cause

FIG. 5. Severe impairment of inhibitory synaptic transmission in preBotC neurons in cd44�/�; met�/� mice. (A) Representative recordings ofminiature glycinergic and GABAergic IPSCs (mIPSCs) in brain stem preBotC neurons. (B and C) Average frequency (B) and amplitude (C) ofmIPSCs. Data are expressed as means � SEM. Statistical significance of differences between means was assessed with Student t tests. The levelof significance was set at P of �0.05; n.s., not significant. The numbers indicate the number of neurons/mouse.

VOL. 27, 2007 COOPERATION OF CD44 AND c-Met IN VIVO 8803

visible irregularities in the resting ventilation activity, and onlya reduction of 75% caused a life-threatening failure of ven-tilation. Thus, the different level in the reduction of overallnetwork activity could explain the partial penetrance of thedefect in the cd44�/�; met�/� mice.

The detailed analysis of the defect in the preBotC showedthat met heterozygotes in the CD44-null background have adecrease of both synaptic excitation and inhibition. Interest-ingly, the glutamatergic synaptic excitation in the CD44 knock-out mice is already severely decreased but no change is ob-served in glycinergic and GABAergic synaptic inhibition nor inthe overall synaptic activity. Within the respiratory network,the synaptic inhibition is essential to generate a stable andregular respiratory rhythm, and therefore the overall networkactivity is mainly influenced by synaptic inhibition. Thus itseems that the impairment of the synaptic inhibition is themain cause of the decrease of the overall synaptic activity inthe cd44�/�; met�/� animals.

It is noteworthy that although there is a strong decrease inthe frequency of the miniature excitatory and inhibitory trans-mission in cd44�/�; met�/� animals, their amplitude was onlymoderately compromised. This suggests that both the excita-tory and inhibitory synaptic reductions are most likely of pre-synaptic nature. As the overall structure within the brain stemnetwork was not dramatically changed, as judged from thehistology of the region, these data would suggest either a re-

duced number of neurons and/or synapses in the brain stempreBotC region.

In addition to the disturbances in the preBotC, in cd44�/�;met�/� mice a significant reduction in the number of axonsthat were segregated for myelination and a change in their sizedistribution were observed in the phrenic nerve that innervatesthe diaphragm. Axonal signals are thought to be of crucialimportance for early stages in myelination (for an example, seereference 28). A function of HGF/c-Met in the development ofmotoneurons has been described at stages corresponding tothe period of naturally occurring cell death (between E13 andE20) (42), where they might act as motoneuronal survivalfactors (6). HGF also promotes neural induction (33) and axonoutgrowth (43). Interestingly, HGF acts preferentially on asubset of limb-innervating motoneurons within the cervicalspinal cord (42), a population of motoneurons that extendslong axons. In addition, HGF has been shown to act as achemoattractant that guides developing axons to their target(6). A reason for the change in axon morphology and for thedelay in reaching the 1:1 axon/Schwann cell ratio required formyelination (3) in the phrenic nerve might therefore be thatthe capacity of motoneurons to sustain outgrowth of function-ally adequate axons is impaired. Indeed, the phrenic nucleus isa column of motoneurons extending from the third to the sixthcervical cord segment that also makes long axons and thereforemight be preferentially affected. Also, the migration defect of

FIG. 6. Severe impairment of excitatory synaptic transmission in preBotC neurons in cd44�/� mice. (A) Representative recordings of miniatureglutamatergic EPSCs (mEPSCs) in brain stem preBotC neurons. (B and C) Average frequency (B) and amplitude (C) of miniature glutamatergicEPSCs. Data are expressed as means � SEM. Statistical significance of differences between means was assessed with Student t tests. The level ofsignificance was set at P of �0.05. The numbers indicate the number of neurons/mouse.

8804 MATZKE ET AL. MOL. CELL. BIOL.

muscle precursor cells in c-Met- or HGF-null mice (1, 26, 37)seems to be due to the loss of a function of HGF as a che-moattractant since ectopic expression of HGF in transgenicanimals leads to ectopic muscle formation (35). However, themorphology of the diaphragm and of the intercostal muscleswas normal, indicating that a defect of muscle cell migrationdoes not contribute to the haploinsufficiency in the cd44�/�;met�/� mice.

Alternatively or additionally, changes in myelination may becaused by impaired differentiation of Schwann cells. HGF canstimulate Schwann cell mitosis via c-Met (11). However, thefunction of HGF or c-Met for Schwann cell differentiation hasnever been investigated in knockout mice because HGF andc-Met knockout mice die early during development, beforeSchwann cells start to differentiate. The effects of the cd44deletion in met heterozygote mice provide the first geneticevidence that HGF/c-Met signaling plays a physiologically rel-evant role for development of functional axons that innervatethe diaphragm and possibly also other muscles that contributeto respiration.

Taken together, our data suggest that the presence of thecoreceptor CD44 is essential for the function of c-Met in thesynaptogenesis and axon/nerve fiber development in the pe-ripheral and central nervous systems.

Although the haploinsufficiency of c-Met in the CD44-nullbackground undoubtedly shows that in wild-type mice c-Metand CD44 cooperate, it does not allow any conclusion as towhether this cooperation is direct or indirect and what theunderlying mechanism is. From our previous studies using celllines, we suggest that during embryogenesis specific isoformsof CD44 containing exon v6 also act as coreceptors for c-Met.Indeed CD44 variant expression has been shown on axons andneurons in the central nervous system (9) and HGF acts viac-Met on central nervous system neurons (8, 41). The sugges-tion that CD44 acts as a coreceptor for c-Met during embry-ogenesis is further strengthened by our findings that in carci-noma cells lacking CD44v6, ICAM-1 fulfills the coreceptorfunction for c-Met, and that in CD44-null mice this new core-ceptor for c-Met substitutes for the CD44 function in liver (ourunpublished results).

ACKNOWLEDGMENTS

We are thankful to Walter Birchmeier, MDC Berlin-Buch, for fruit-ful suggestions and discussions and to Selma Huber and the membersof the animal facility for their help in animal experiments.

This work was supported by grants of the Deutsche Krebshilfe(Forderschwerpunktprogramm: Zelladhasion, Invasion und Metasta-sierung), the Deutsche Forschungsgemeinschaft (SPP 1190), the Deut-sche Forschungsgemeinschaft through the DFG-Research Center forMolecular Physiology of the Brain (W.Z.), and SFB 487 (M.S.) andSFB 581 (E.A., B.H., and M.S.).

REFERENCES

1. Bladt, F., D. Riethmacher, S. Isenmann, A. Aguzzi, and C. Birchmeier. 1995.Essential role for the c-met receptor in the migration of myogenic precursorcells into the limb bud. Nature 376:768–771.

2. Cavallaro, U., J. Niedermeyer, M. Fuxa, and G. Christofori. 2001. N-CAMmodulates tumour-cell adhesion to matrix by inducing FGF-receptor signal-ling. Nat. Cell Biol. 3:650–657.

3. Court, F. A., L. Wrabetz, and M. L. Feltri. 2006. Basal lamina: Schwann cellswrap to the rhythm of space-time. Curr. Opin. Neurobiol. 16:501–507.

4. Cully, M., A. Elia, S. H. Ong, V. Stambolic, T. Pawson, M. S. Tsao, and T. W.Mak. 2004. grb2 heterozygosity rescues embryonic lethality but not tumori-genesis in pten�/� mice. Proc. Natl. Acad. Sci. USA 101:15358–15363.

5. Dejana, E., M. G. Lampugnani, O. Martinez-Estrada, and G. Bazzoni. 2000.The molecular organization of endothelial junctions and their functional rolein vascular morphogenesis and permeability. Int. J. Dev. Biol. 44:743–748.

6. Ebens, A., K. Brose, E. D. Leonardo, M. G. Hanson, Jr., F. Bladt, C.Birchmeier, B. A. Barres, and M. Tessier-Lavigne. 1996. Hepatocyte growthfactor/scatter factor is an axonal chemoattractant and a neurotrophic factorfor spinal motor neurons. Neuron 17:1157–1172.

7. Gunthert, U., M. Hofmann, W. Rudy, S. Reber, M. Zoller, I. Hau�mann, S.Matzku, A. Wenzel, H. Ponta, and P. Herrlich. 1991. A new variant ofglycoprotein CD44 confers metastatic potential to rat carcinoma cells. Cell65:13–24.

8. Hamanoue, M., N. Takemoto, K. Matsumoto, T. Nakamura, K. Nakajima,and S. Kohsaka. 1996. Neurotrophic effect of hepatocyte growth factor oncentral nervous system neurons in vitro. J. Neurosci. Res. 43:554–564.

9. Kaaijk, P., S. T. Pals, F. Morsink, D. A. Bosch, and D. Troost. 1997. Dif-ferential expression of CD44 splice variants in the normal human centralnervous system. J. Neuroimmunol. 73:70–76.

10. Kaya, G., I. Rodriguez, J. L. Jorcano, P. Vassalli, and I. Stamenkovic. 1997.Selective suppression of CD44 in keratinocytes of mice bearing an antisenseCD44 transgene driven by a tissue-specific promoter disrupts hyaluronatemetabolism in the skin and impairs keratinocyte proliferation. Genes Dev.11:996–1007.

11. Krasnoselsky, A., M. J. Massay, M. C. DeFrances, G. Michalopoulos, R.Zarnegar, and N. Ratner. 1994. Hepatocyte growth factor is a mitogen forSchwann cells and is present in neurofibromas. J. Neurosci. 14:7284–7290.

12. Liu, Y., and L. R. Rohrschneider. 2002. The gift of Gab. FEBS Lett. 515:1–7.13. Masu, Y., E. Wolf, B. Holtmann, M. Sendtner, G. Brem, and H. Thoenen.

1993. Disruption of the CNTF gene results in motor neuron degeneration.Nature 365:27–32.

14. Matzke, A., P. Herrlich, H. Ponta, and V. Orian-Rousseau. 2005. A five-amino-acid peptide blocks Met- and Ron-dependent cell migration. CancerRes. 65:6105–6110.

15. Missler, M., W. Zhang, A. Rohlmann, G. Kattenstroth, R. E. Hammer, K.Gottmann, and T. C. Sudhof. 2003. Alpha-neurexins couple Ca2� channelsto synaptic vesicle exocytosis. Nature 423:939–948.

16. Nedvetzki, S., M. Walmsley, E. Alpert, R. O. Williams, M. Feldmann, and D.Naor. 1999. CD44 involvement in experimental collagen-induced arthritis(CIA). J. Autoimmun. 13:39–47.

17. Nguyen, L., M. Holgado-Madruga, C. Maroun, E. D. Fixman, D. Kamikura,T. Fournier, A. Charest, M. L. Tremblay, A. J. Wong, and M. Park. 1997.Association of the multisubstrate docking protein Gab1 with the hepatocytegrowth factor receptor requires a functional Grb2 binding site involvingtyrosine 1356. J. Biol. Chem. 272:20811–20819.

18. Oppenheim, R. W., S. Wiese, D. Prevette, M. Armanini, S. Wang, L. J.Houenou, B. Holtmann, R. Gotz, D. Pennica, and M. Sendtner. 2001. Car-diotrophin-1, a muscle-derived cytokine, is required for the survival of sub-populations of developing motoneurons. J. Neurosci. 21:1283–1291.

19. Orian-Rousseau, V., L. Chen, J. P. Sleeman, P. Herrlich, and H. Ponta. 2002.CD44 is required for two consecutive steps in HGF/c-Met signaling. GenesDev. 16:3074–3086.

20. Orian-Rousseau, V., H. Morrison, A. Matzke, T. Kastilan, G. Pace, P. Herrlich,and H. Ponta. 2007. Hepatocyte growth factor-induced Ras activation re-quires ERM proteins linked to both CD44v6 and F-actin. Mol. Biol. Cell18:76–83.

21. Paratcha, G., F. Ledda, and C. F. Ibanez. 2003. The neural cell adhesionmolecule NCAM is an alternative signaling receptor for GDNF familyligands. Cell 113:867–879.

22. Protin, U., T. Schweighoffer, W. Jochum, and F. Hilberg. 1999. CD44-deficientmice develop normally with changes in subpopulations and recirculation oflymphocyte subsets. J. Immunol. 163:4917–4923.

23. Reynolds, E. S. 1963. The use of lead citrate at high pH as an electron-opaque stain in electron microscopy. J. Cell Biol. 17:208–212.

24. Sachs, M., H. Brohmann, D. Zechner, T. Muller, J. Hulsken, I. Walther, U.Schaeper, C. Birchmeier, and W. Birchmeier. 2000. Essential role of Gab1for signaling by the c-Met receptor in vivo. J. Cell Biol. 150:1375–1384.

25. Schlessinger, J., and A. Ullrich. 1992. Growth factor signaling by receptortyrosine kinases. Neuron 9:383–391.

26. Schmidt, C., F. Bladt, S. Goedecke, V. Brinkmann, W. Zschiesche, M.Sharpe, E. Gherardi, and C. Birchmeier. 1995. Scatter factor/hepatocytegrowth factor is essential for liver development. Nature 373:699–702.

27. Schmits, R., J. Filmus, N. Gerwin, G. Senaldi, F. Kiefer, T. Kundig, A.Wakeham, A. Shahinian, C. Catzavelos, J. Rak, C. Furlonger, A. Zakarian,J. J. Simard, P. S. Ohashi, C. J. Paige, J. C. Gutierrez-Ramos, and T. W.Mak. 1997. CD44 regulates hematopoietic progenitor distribution, granu-loma formation, and tumorigenicity. Blood 90:2217–2233.

28. Sherman, D. L., and P. J. Brophy. 2005. Mechanisms of axon ensheathmentand myelin growth. Nat. Rev. Neurosci. 6:683–690.

29. Sherman, L., D. Wainwright, H. Ponta, and P. Herrlich. 1998. A splicevariant of CD44 expressed in the apical ectodermal ridge presents fibroblastgrowth factors to limb mesenchyme and is required for limb outgrowth.Genes Dev. 12:1058–1071.

VOL. 27, 2007 COOPERATION OF CD44 AND c-Met IN VIVO 8805

30. Sherman, L. S., T. A. Rizvi, S. Karyala, and N. Ratner. 2000. CD44 enhancesneuregulin signaling by Schwann cells. J. Cell Biol. 150:1071–1084.

31. Sonnenberg, E., A. Godecke, B. Walter, F. Bladt, and C. Birchmeier. 1991.Transient and locally restricted expression of the ros1 protooncogene duringmouse development. EMBO J. 10:3693–3702.

32. Sonnenberg, E., D. Meyer, K. M. Weidner, and C. Birchmeier. 1993. Scatterfactor/hepatocyte growth factor and its receptor, the c-met tyrosine kinase,can mediate a signal exchange between mesenchyme and epithelia duringmouse development. J. Cell Biol. 123:223–235.

33. Streit, A., C. D. Stern, C. Thery, G. W. Ireland, S. Aparicio, M. J. Sharpe,and E. Gherardi. 1995. A role for HGF/SF in neural induction and itsexpression in Hensen’s node during gastrulation. Development 121:813–824.

34. Suyama, K., I. Shapiro, M. Guttman, and R. B. Hazan. 2002. A signalingpathway leading to metastasis is controlled by N-cadherin and the FGFreceptor. Cancer Cell 2:301–314.

35. Takayama, H., W. J. La Rochelle, M. Anver, D. E. Bockman, and G. Merlino.1996. Scatter factor/hepatocyte growth factor as a regulator of skeletal mus-cle and neural crest development. Proc. Natl. Acad. Sci. USA 93:5866–5871.

36. Trusolino, L., A. Bertotti, and P. M. Comoglio. 2001. A signaling adapterfunction for alpha6beta4 integrin in the control of HGF-dependent invasivegrowth. Cell 107:643–654.

37. Uehara, Y., O. Minowa, C. Mori, K. Shiota, J. Kuno, T. Noda, and N.Kitamura. 1995. Placental defect and embryonic lethality in mice lackinghepatocyte growth factor/scatter factor. Nature 373:702–705.

38. Varoqueaux, F., G. Aramuni, R. L. Rawson, R. Mohrmann, M. Missler, K.Gottmann, W. Zhang, T. C. Sudhof, and N. Brose. 2006. Neuroligins deter-mine synapse maturation and function. Neuron 51:741–754.

39. Vlodavsky, I., H. Q. Miao, B. Medalion, P. Danagher, and D. Ron. 1996.

Involvement of heparan sulfate and related molecules in sequestration andgrowth promoting activity of fibroblast growth factor. Cancer MetastasisRev. 15:177–186.

40. Weidner, K. M., S. Di Cesare, M. Sachs, V. Brinkmann, J. Behrens, and W.Birchmeier. 1996. Interaction between Gab1 and the c-Met receptor tyrosinekinase is responsible for epithelial morphogenesis. Nature 384:173–176.

41. Wong, V., D. J. Glass, R. Arriaga, G. D. Yancopoulos, R. M. Lindsay, and G.Conn. 1997. Hepatocyte growth factor promotes motor neuron survival andsynergizes with ciliary neurotrophic factor. J. Biol. Chem. 272:5187–5191.

42. Yamamoto, Y., J. Livet, R. A. Pollock, A. Garces, V. Arce, O. deLapeyriere,and C. E. Henderson. 1997. Hepatocyte growth factor (HGF/SF) is a muscle-derived survival factor for a subpopulation of embryonic motoneurons. De-velopment 124:2903–2913.

43. Yang, X. M., and M. Park. 1993. Expression of the met/hepatocyte growthfactor/scatter factor receptor and its ligand during differentiation of murineP19 embryonal carcinoma cells. Dev. Biol. 157:308–320.

44. Yayon, A., M. Klagsbrun, J. D. Esko, P. Leder, and D. M. Ornitz. 1991. Cellsurface, heparin-like molecules are required for binding of basic fibroblastgrowth factor to its high affinity receptor. Cell 64:841–848.

45. Yu, W. H., J. F. Woessner, Jr., J. D. McNeish, and I. Stamenkovic. 2002.CD44 anchors the assembly of matrilysin/MMP-7 with heparin-binding epi-dermal growth factor precursor and ErbB4 and regulates female reproduc-tive organ remodeling. Genes Dev. 16:307–323.

46. Zhang, W., A. Rohlmann, V. Sargsyan, G. Aramuni, R. E. Hammer, T. C.Sudhof, and M. Missler. 2005. Extracellular domains of alpha-neurexinsparticipate in regulating synaptic transmission by selectively affecting N- andP/Q-type Ca2� channels. J. Neurosci. 25:4330–4342.

8806 MATZKE ET AL. MOL. CELL. BIOL.