[page 35↓]

3  Results

3.1 Yeast Two-Hybrid Screen with ErbB-2 Bait Proteins

Previous studies implicate the receptor tyrosine kinase ErbB-2 in lobulo-alveolar development of the mammary gland (Niemann et al., 1998; Jones and Stern, 1999). In search for new substrates of ErbB-2 that mediate its function in mammary alveolar morphogenesis, yeast two-hybrid screens were performed with ErbB-2 baits (Fields and Song, 1989). The classical baits for receptor tyrosine kinases are hybrid proteins, which consist of an E.coli LexA DNA-binding and dimerization domain followed by the intracellular region of the receptor tyrosine kinase (O'Neill et al., 1994; Weidner et al., 1996). However, baits spanning the full-length cytoplasmic region of the ErbB-2 receptor were highly unstable in the yeast (my own observation; data not shown). Therefore, modified yeast baits were conceived on the basis of previous studies with chimeric receptors; when C-terminal sequences of c-Met are fused to the kinase of TrkA (the nerve growth factor receptor), the hybrid receptor elicits Met-specific morphogenic responses upon activation with NGF (Sachs et al., 1996). Thus, the C-terminal docking region of a receptor is sufficient to determine its signaling specificity; this region contains tyrosine residues that, when phosphorylated, become docking sites for receptor-specific downstream effectors. In view of these results, chimeric ErbB-2 bait proteins were generated: the tyrosine kinase TprMet was fused to LexA, and the C-terminal tail of Met was replaced by the C-terminal tail of ErbB-2 (Figure 1). TprMet was chosen as heterologous bait kinase since it is strongly active when expressed in the yeast (my own observation; data not shown). TprMet is the human oncogenic counterpart of c-Met (Park et al., 1986). Tpr (translocated promoter region) codes for two putative dimerization domains, and is a movable DNA element that has spontaneously translocated into a genomic region upstream of the Met kinase; the resulting TprMet protein dimerizes in a ligand-independent fashion through its newly-acquired Tpr sequences, and is therefore a cytoplasmic, constitutively active tyrosine kinase protein with the biological activity of c-Met. It is therefore likely that the presence of a C-terminal tail of ErbB-2 in the modified TprMet-ErbB2 baits bestows ErbB-2 signaling [page 36↓]properties on the chimeric proteins (see below). Two ErbB-2 baits, TprMet-ErbB2(Y1-4) and TprMet-ErbB2(Y3-5), were generated to cover the five tyrosine residues that are putative autophosphorylation sites of ErbB-2 (Fig. 1).

Figure 1. Structure of the chimeric ErbB-2 baits used in yeast two-hybrid screens.

The baits consist of a LexA DNA-binding and dimerization region, followed by the kinase domain of TprMet; the C-terminal tail of TprMet is substituted for sequences of the ErbB-2 multiple docking region. The TprMet-ErbB2(Y1-4) bait stretches over tyrosines Y1028 (for simplicity Y1 in the scheme), Y1144 (Y2), Y1201 (Y3) and Y1226/7 (Y4) of ErbB-2, with Y1253 (Y5) being mutated to F; the TprMet-ErbB2(Y3-5) bait includes tyrosines Y1201 (Y3), Y1226/7 (Y4) and Y1253 (Y5).

TprMet-ErbB2(Y1-4) contains the four N-terminal tyrosines of c-ErbB-2 (Y1 to Y4); TprMet-ErbB2(Y3-5) includes the three most C-terminal tyrosines (Y3 to Y5). Both baits were efficiently expressed in the yeast and were constitutively phosphorylated on tyrosine residues (data not shown). In preliminary tests, these baits exhibited ErbB-2 selectivity for binding downstream signaling molecules: TprMet-ErbB2(Y1-4) interacted with Grb-2 and Shc, while TprMet-ErbB2(Y3-5) interacted with Shc (data not shown); it has already been described that Grb-2 and Shc directly interact with phosphorylated Y2 and Y4 residues of ErbB-2, respectively (Dankort et al., 1997).

Both chimeric ErbB-2 baits were used to screen the Hollenberg library (see Materials and Methods). The library consists of mouse E10.5 embryonic cDNAs that are fused to the VP16 activation domain, and has extensively been used in our group to find novel interaction [page 37↓]partners of receptor tyrosine kinases (Weidner et al., 1996; Grimm et al., 2001). In several independent screens, clones encoding the SH2 domains of known and novel interaction partners of ErbB-2 were isolated. Table 1 summarizes the results of screens for each ErbB-2 bait protein. Fig. 2 indicates the amino acid region encoded in the interacting clones. Src and PLCγ1 have previously been described as substrates of c-ErbB-2 (Fazioli et al., 1991; Muthuswamy and Muller, 1995b); their isolation in the yeast screen corroborates ErbB-2 binding specificity of the chimeric baits. Vav2, Nck and Grb10 were here identified as novel interaction partners of the ErbB-2 receptor.

Table 1. Interaction partners of ErbB-2 isolated in independent yeast two-hybrid screens using different baits.

Clone

TprMet-ErbB2(Y1-4)

TprMet-ErbB2(Y3-5)

Vav2

2

1

PLCγ1

6

7

Nck

-

1

Src

-

2

Grb10

-

34

Numbers indicate frequency of each isolated interaction partner, including overlapping clones.

Interaction of the prey proteins with the baits was confirmed by co-transfection into yeast cells, followed by analysis of yeast growth and β-galactosidase activity assays (data not shown). Sequence comparison analyses revealed that the interacting regions spanned the SH2 domain of each ErbB-2 partner (Fig. 2), strongly suggesting that interaction involved phosphotyrosine residues of the baits. This hypothesis was confirmed by using kinase-defective baits; a point mutation of lysine 243 to alanine (K243A) was introduced in the ATP-[page 38↓]binding site of the TprMet kinase, which completely disrupts the catalytic activity (Rodrigues and Park, 1993). The inactivating mutation of the bait kinases abolished interaction, confirming that the isolated SH2 domains bound to phosphotyrosine residues (Fig. 3, middle panel; compare to interaction with wild-type bait, upper panel). Furthermore, deletion of the ErbB-2 C-terminal tail also impaired binding (Fig. 3, lower panel), clearly indicating that the clones interacted with phosphotyrosines that were located on the ErbB-2 tail and not within the TprMet sequences of the baits.

Figure 2. Interaction partners of ErbB-2 in the yeast system are SH2 domains.

The overall structure of the interacting proteins is schematically shown, relevant domains are highlighted. The amino acid stretch encoded in the isolated clones is indicated by the bars below the protein structures, overlapping clones are shown. AD: acidic domain; CH: calponin-homology domain; DH: Dbl-homology domain; PH: pleckstrin-homology domain; PR: proline-rich region; SH2: Src-homology-2 domain; SH3: Src-homology-3 domain; ZF: zinc finger.

Figure 3. The SH2 domains of substrates isolated in yeast two-hybrid screens interact with phosphotyrosine residues of the ErbB-2 C-terminal tail.

Preys were co-transfected into yeast together with wild-type TprMet-ErbB2(Y3-5) (TprMet-ErbB2 for simplicity), kinase-defective (TprMet-ErbB2-Kin-) or C-terminally deleted (TprMet-Δtail) baits (schematic structures are on the right), and interaction was tested in a yeast colony-growth assay on selection plates. The various preys interacted with the wild-type TprMet-ErbB2 bait, but not with the kinase-defective or the deletion mutant without ErbB-2 C-terminal tail.

3.2 Distinct Phosphotyrosine Residues of ErbB-2 Bind to Various Interaction Partners

The phosphotyrosine residues of ErbB-2 that are direct binding sites for its interaction partners were mapped by mutational analysis of the bait proteins. Sequences of the C-terminal tail of ErbB-2 containing single tyrosines or pairs of consecutive tyrosine residues were fused to TprMet to generate bait deletion mutants (see Materials and Methods); these new baits were tested with the library clones for interaction in the yeast, which was evaluated by colony-growth on selection plates and β-galactosidase activity. Grb-2 and Shc were included in these [page 40↓]experiments as controls. The results of yeast colony-growth and β-galactosidase assays are summarized in Table 2. Nck interacted with baits containing Y3 (Y1201) and Grb10 with baits containing Y4 (Y1226/7); Src required simultaneously both phosphorylated Y3 and Y4 (Y1201 and Y1226/7); PLCγ1 interacted with every tyrosine-phosphorylated bait mutant (data not shown). Vav2 directly bound to phosphorylated Y1 (Y1028) and Y3 (Y1201) of ErbB-2 (see also Fig. 4). Interestingly, both tyrosines that bound Vav2 are located within a pYLVP motif, which apparently constitutes the consensus binding sequence for the SH2 domain of Vav2. These results again validate the reliability of this modified yeast two-hybrid approach to identify phosphotyrosine binding proteins and to map their direct binding motifs.

Table 2. Characterization of specific binding sites of ErbB-2 for its interaction partners.

Bait*

Vav2

Nck

Src

Grb10

Grb2

Shc

Y1

+

-

-

-

-

-

Y1-2

+

-

-

-

+

-

Y2

-

-

-

-

+

-

Y2-3

+

+

-

-

+

-

Y3

+

+

-

-

-

-

Y3-4

+

+

+

+

-

+

Y3-5§

+

+

-

+

-

+

Y4-5

-

-

-

+

-

+

Y4

-

-

-

+

-

+

Y5

-

-

-

-

-

+

Plus (+) indicates interaction, minus (-) indicates no interaction.
*Baits are TprMet-ErbB2 deletion mutants; numbers indicate single or pairs of tyrosine residues in the C-terminal tail of ErbB-2.
§TprMet-ErbB2(Y3-5) bait was included as internal control.

Figure 4. Vav2 directly interacts with phosphotyrosines Y1 and Y3 of c-ErbB-2.

The full-size cDNA of Vav2 was isolated from a phage library (see Materials and Methods) and inserted in VP16 yeast vector. To characterize the Vav2 binding sites on c-ErbB-2, the full-length Vav2 prey protein was tested with various ErbB-2 bait mutants for interaction in the yeast. The results from colony-growth (left panel) and β-galactosidase assays (right panel) are shown. ErbB-2 bait mutants are those from Table 2; additionally, Y2 and Y3 are mutated to phenylalanine in the TprMet-ErbB2(Y2F-3) and TprMet-ErbB2(Y2-3F) baits, respectively. TprMet-Δtail lacks an ErbB-2 C-terminal tail. Kin denotes kinase-deficient bait mutants. Hybrid TprMet bait proteins for the other ErbB receptors (EGF receptor, ErbB-3 and ErbB-4) were simultaneously tested (see below).

3.3 Vav2 Interacts with Receptor Tyrosine Kinases of the ErbB Subfamily

Next, the receptor specificity of each ErbB-2 interaction partner for other RTKs was tested in the yeast system. Src, Nck, PLCγ1 and Grb10 interacted with various receptor tyrosine kinases from unrelated subfamilies (Table 3). In contrast, Vav2 appears to preferentially bind receptors of the ErbB subfamily; Vav2 bound to baits encoding TprMet fusion proteins with [page 42↓]the C-terminal tails of EGF receptor, ErbB-3 and ErbB-4 (Table 3 and Fig. 4); it additionally interacted with a constitutively active PDGFβ receptor (Table 3).

Table 3. Interaction of the various ErbB-2 partners with different receptor tyrosine kinases.

Receptor*

PLCg1

Src

Nck

Grb10

Vav2

TprMet-ErbB2

+

+

+

+

+

TprMet-EGFR

+

+

+

+

+

TprMet-ErbB3

-

+

+

+

+

TprMet-ErbB4

-

-

-

+

+

TprMet-Δtail

-

-

-

-

-

Met

+

+

+

+

-

Insulin Receptor

+

-

+

+

-

Ret

+

+

-

+

-

Kit

-

-

-

-

-

KGF Receptor

-

-

-

-

-

Ros

+

+

-

+

-

Sea

+

-

-

-

-

PDGFβ Receptor

+

+

+

+

+

Plus (+) indicates interaction.


[page 43↓]

Together, these results point to Vav2 as a putative candidate to mediate ErbB signals in mammary gland development. First, Vav2 specifically interacts with all the ErbB receptors; second, all ErbB receptors are essential at different stages of mammary development (see section 2.5 of the Introduction). Lastly, activated Vav2 induces cytoskeletal rearrangements, which are a critical step for morphogenic events. Therefore, the potential of Vav2 to mediate ErbB-2 signals in alveolar morphogenesis was next tested in organotypic cell culture experiments.

3.4 Vav2 Induces Alveolar Morphogenesis of EpH4 Mammary Epithelial Cells

Activation of endogenous or ectopic ErbB-2 receptor instructs EpH4 mammary epithelial cells to form functional alveoli-like structures when these cells are cultured on EHS matrix (termed Matrigel; Niemann et al., 1998). To examine whether Vav2 mediates the morphogenic effect of ErbB-2 in organotypic culture, EpH4 mammary epithelial cells were stably transfected with a cDNA encoding N-terminally truncated Vav2, the oncogenic, constitutively active form of Vav2 (termed ΔN-Vav2; structures are shown in Fig. 5; see Schuebel et al., 1996); afterwards, transfectants were cultured on Matrigel to test their ability to undergo alveolar morphogenesis. ΔN-Vav2 was preferred to full-size Vav2 for the Matrigel assay in view of its functional properties: ΔN-Vav2 is constitutively active as guanine nucleotide exchange factor and induces cytoskeletal rearrangements when overexpressed in fibroblasts (Schuebel et al., 1996); in contrast, full-size Vav2 requires tyrosine phosphorylation to elicit such responses (Schuebel et al., 1998). Remarkably, overexpression of ΔN-Vav2 induced the arrangement of the transfected EpH4 cells into large alveolar structures within two days of culture on Matrigel (Fig. 5B, compare with control in A). The alveoli consisted of monolayers of tightly associated cells facing a lumen (Fig. 5D). Ultrastructural studies revealed that the alveolar cells were polarized; microvilli were scattered at the luminal side, and tight junctions were observed at apical cell-cell borders (Fig. 6B, C). Polarization of alveolar cells in the mammary gland during pregnancy and lactation has been described (Nemanic et al., 1971). Control [page 44↓]EpH4 cells formed small solid aggregates, and cells were non-polarized and were only loosely connected (Fig. 6A).

Figure 5. Constitutively active Vav2 induces alveolar morphogenesis of EpH4 mammary epithelial cells.

(B) Cells expressing ΔN-Vav2 (see schematic structures below) form large alveoli-like structures on matrigel. (D) Alveolar cells are facing a lumen (asterisk), as observed in semithin sections. (A, C)Mock-transfected controls form small aggregates. Bar: 50 μm. CH: calponin-homology domain; AD: acidic domain; DH: Dbl-homology domain; PH: pleckstrin-homology domain; ZF: zinc finger.


[page 45↓]

Figure 6. Constitutively active Vav2 induces polarization of alveolar EpH4 mammary epithelial cells.

(B, C) Microvilli (arrows) are present at the luminal side (asterisk) of alveolar EpH4 cells expressing ΔN-Vav2; apical tight junctions (arrowheads) are revealed by electron-microscopy of ultrathin sections. (A)Mock-transfected controls form small clumps of loosely associated cells. Bars: (A, B) 5 μm; (C) 0.5 μm. ECM: extracellular matrix.

Full-size Vav2 is inactive as GDP/GTP exchanger in a non-phosphorylated state (Schuebel et al., 1998). Not surprisingly, EpH4 cells overexpressing full-size Vav2 did not produce alveoli (data not shown), suggesting that full-size Vav2 may also require tyrosine-phosphorylation to become morphogenic. It has previously been shown that C-terminal fusion of a substrate to its receptor results in efficient tyrosine phosphorylation of the substrate (Schaeper et al., 2000); moreover, the insulin receptor substrate sequences are fused to the insulin receptor in D. melanogaster (Yenush et al., 1996). Therefore, the fusion protein TrkErbB2-Vav2 was generated by replacing the C-terminal tail of ErbB-2 with full-length Vav2 in a TrkErbB2 receptor(Sachs et al., 1996; structures are shown in Fig. 7). The resulting protein lacks the docking sites of ErbB-2 and hence could only signal through the coupled substrate when activated by nerve growth factor (NGF). Indeed, EpH4 cells expressing TrkErbB2-Vav2 formed large alveoli following stimulation with NGF (Fig. 7C, D). In the absence of NGF or in transfectants lacking Vav2 in the fusion protein, respectively, none or only rudimentary structures were seen (Fig. 7A, B and E, F). Taken together, these findings demonstrate that [page 46↓]Vav2 induces alveolar morphogenesis of EpH4 cells when activated either by N-terminal truncation or upon tyrosine phosphorylation by ErbB-2.

Figure 7. Tyrosine phosphorylation of full-size Vav2 by ErbB-2 activates its morphogenic potential.

Left panel: overviews by light microscopy; right panel: semithin sections. (A, B) EpH4 transfectants expressing the TrkErbB2-Vav2 hybrid (see structure below) form alveolar structures on matrigel upon NGF stimulation. (C, D) Non-stimulated transfectants and (E, F) cells expressing a TrkErbB2 protein without the C-terminal tail of ErbB-2 (TrkErbB2-Δtail) lack morphogenic activity. Bars: 50 μm.

3.5 Vav2 and ErbB-2 Can Both Directly and Indirectly Associate in Mammalian Cells

The interaction between Vav2 and ErbB-2 was also studied in cultured mammalian cells. Human epithelial kidney 293T cells were transiently co-transfected with Vav2 and wild-type or mutated TrkErbB2 receptors, and interaction after stimulation with NGF was examined by Western blotting. Exogenous Vav2 co-immunoprecipitated with a TrkErbB2 hybrid receptor in lysates of transfected 293T cells, but not with kinase-deficient receptor TrkErbB2-Kin- (Fig. 8A, upper panel). These results indicate that association of Vav2 with ErbB-2 requires tyrosine phosphorylation of the receptor, as previously observed in the yeast system. Surprisingly, Vav2 still interacted with TrkErbB2mut, which lacks the Vav2 binding sites of ErbB-2 (Y1028 and Y1201) identified in the yeast two-hybrid system. This suggests that Vav2 may bind directly and indirectly to ErbB-2. However, interaction of Vav2 with TrkErbB2 or TrkErbB2mut depends on activation of the hybrid receptors by NGF (Fig. 8A, upper panel). Increased tyrosine phosphorylation of Vav2 was observed upon binding to ErbB-2 proteins that contain an activated kinase (Fig. 8A, middle panel).

Interaction between Vav2 and ErbB-2 was also tested in EpH4 cells that are cultured under standard conditions. Subconfluent EpH4 cells were treated shortly with neuregulin-1 (see Materials and Methods) to activate the endogenous ErbB-2 receptor (Niemann et al., 1998) and then lysed. Immunoprecipitation of endogenous Vav2 from cell lysates followed by Western blotting revealed that Vav2 associated with ErbB-2 in lysates of neuregulin-treated EpH4 cells (Fig. 8B, upper left panel). In addition, increased tyrosine phosphorylation of ErbB-2 and Vav2 was observed in lysates of stimulated EpH4 cells (Fig. 8B, upper right panel). These findings indicate that interaction between Vav2 and ErbB-2 occur in EpH4 cells after activation of ErbB-2 by neuregulin. Moreover, activated ErbB-2 may further [page 48↓]phosphorylate Vav2 on tyrosine residues. The interaction between Vav2 and ErbB-2 in EpH4 cells may therefore be an essential step in neuregulin-induced morphogenesis of these cells.

Figure 8. Analysis of the interaction between Vav2 and ErbB-2 in mammalian cells
and in mammary tissue.

(A) Co-immunoprecipitation of Vav2 with various TrkErbB2 receptors in transfected 293T cells. (Top) Vav2 co-immunoprecipitates with TrkErbB2 and TrkErbB2mut, when these hybrid receptors are activated by NGF. Vav2 does not associate with kinase-defective TrkErbB2-Kin-. (Middle) Binding of Vav2 to activated TrkErbB-2 receptors results in increased tyrosine phosphorylation of Vav2. (Bottom) Control immunoblots of total cell lysates. Molecular weight markers on the left are in kDa. (B) Interaction between Vav2 and ErbB-2 in EpH4 mammary epithelial cells. (Top) Vav2 co-immunoprecipitates with ErbB-2 in lysates of neuregulin-treated cells (left panel); association of Vav2 with activated ErbB-2 results in increased tyrosine phosphorylation of Vav2 (right panel). (Bottom) Control immunoblots of total cell lysates. (C) In vivo association between Vav2 and ErbB-2 in mammary glands during pregnancy. Vav2 co-immunoprecipitates with ErbB-2 in mammary gland lysates from pregnant mice. Anti-c-Myc antibody was used as non-specific control; Vav2 immuno-precipitation is shown below.

3.6 Vav2 and ErbB-2 Are Associated in Mammary Epithelium During Pregnancy

Interaction between Vav2 and ErbB-2 was also studied in mammary glands from pregnant mice. Lysis of all mammary glands from a female mouse at day 17.5 of pregnancy was followed by immunoprecipitation of endogenous Vav2 and Western blotting as above. Indeed, Vav2 co-immunoprecipitated with ErbB-2 from the mammary lysates (Fig. 8C), suggesting the presence of endogenous complexes between Vav2 and ErbB-2 in mammary tissue during pregnancy.

Cellular localization of Vav2 and ErbB-2 was studied in mammary tissue from pregnant mice by in situ hybridization. Importantly, Vav2 and ErbB-2 are co-expressed in the epithelial cell layer that lines alveoli and ducts of mammary glands during pregnancy (Fig. 9; Schroeder et al., 1998). The co-localization of Vav2 and ErbB-2 in mammary epithelium supports the presence of interaction in vivo during pregnancy, as above suggested by biochemical analysis (see Fig. 8C); moreover, it rules out the possibility of association during the immuno-precipitation of Vav2. Taken together, these observations indicate that ErbB-2 and Vav2 can functionally associate in the mammary gland while lobulo-alveolar development takes place.


[page 50↓]

Figure 9. Vav2 and ErbB-2 are co-expressed in mammary alveolar epithelium during pregnancy.

(Top) Vav2 and ErbB-2 co-localize in alveolar (black arrowheads) and ductal (arrows) epithelia of the mammary gland at late pregnancy. (Bottom) Neuregulin (NRG) is located in the mesenchyme surrounding alveoli and ducts (white arrows); a control section (control) was hybridized with a Vav2 sense probe. Bar: 100 μm.

3.7 The Dbl-Homology Domain of Vav2 Is Required for Its Morphogenic Activity in EpH4 Cells

The structural domains of Vav2 confer the protein two interesting features: first, Vav2 acts as a GDP/GTP exchange factor for Rho/Rac GTPases via its catalytic Dbl-homology domain; second, Vav2 has the capability to participate in signal transduction cascades by binding further downstream proteins via its C-terminal SH2 and SH3 domains(Bustelo, 2000). To define the functional role of the various domains of Vav2 in alveolar morphogenesis, mutational analyses of constitutively active Vav2 were performed. ΔN-Vav2-ΔC encodes an active protein which lacks SH2 and both SH3 domains (structures are shown in Fig. 10). ΔN-Vav2-dblmut lacks the GTP/GDP exchange activity due to a deletion of seven conserved amino acids within the Dbl homology domain (336LLLKELL342), which has been reported [page 51↓]to inactivate Dbl, GRF1 and Sos1 (Hart et al., 1994; Freshney et al., 1997; Qian et al., 1998). EpH4 cells were stably transfected with plasmids encoding Vav2 mutant proteins and examined in Matrigel assays. Clearly, the Dbl homology domain is required for the biological function of Vav2 whereas the SH2/SH3 domains are dispensable (Fig. 10B, C). Overexpression of a minimal Vav2 protein which additionally lacks the pleckstrin homology domain (ΔN-Vav2-ΔCΔPH) was also sufficient to elicit alveolar morphogenesis (Fig. 10D).

Figure 10. The Dbl homology domain of Vav2 is essential to promote alveolar morphogenesis of EpH4 mammary epithelial cells.

(C) C-terminally truncated Vav2 and (D) a minimal protein encompassing the DH and ZF domains, respectively, are as efficient as constitutively active Vav2 (in A) to induce morphogenesis. (B) Catalytically inactive Vav2 carrying a mutation in the Dbl homology domain does not promote formation of alveoli. Bar: 50 μm. DH: Dbl-homology domain; PH: pleckstrin-homology domain; ZF: zinc finger.


[page 52↓]

3.8 Catalytically Inactive Vav2 Blocks Neuregulin-Mediated Morphogenesis of EpH4 Cells

EpH4 cells that express catalytically inactive Vav2 (ΔN-Vav2-dblmut) were tested for alveolar morphogenesis following neuregulin treatment. Importantly, ΔN-Vav2-dblmut interferred with the formation of alveoli when cells were treated with neuregulin: instead of large hollow structures (Fig. 11B, C), cell aggregates without lumina were observed (Fig. 11E, F), indicating that alveolar morphogenesis is prevented while cell growth still occurs. Only small cell groups were formed in absence of neuregulin treatment (Fig. 11A, D).

Figure 11. Catalytically inactive Vav2 harbouring a mutation in the Dbl homology domain blocks the morphogenic signals of neuregulin in EpH4 mammary epithelial cells.

(E, F) EpH4 cells expressing ΔN-Vav2-dblmut (see structure below) do not form hollow alveoli but rather large cell aggregates upon stimulation with neuregulin. (B, C) Mock-transfected cells form alveoli-like structures in response to neuregulin, as previously described (Niemann et al., 1998). (A, D) Non-stimulated controls. Bars: 50 μm. DH: Dbl-homology domain; PH: pleckstrin-homology domain; ZF: zinc finger.

Interestingly, ΔN-Vav2-dblmut did not interfere with tubular branching when EpH4 cells were treated with hepatocyte growth factor/scatter factor (Fig. 12D, compare to control cells in B; Niemann et al., 1998). Thus, catalytically inactive Vav2 blocks neuregulin-specific signals for alveolar morphogenesis of EpH4 cells, while it does not affect cellular responses to other morphogenic stimuli.

Figure 12. Catalytically inactive Vav2 does not interfere with branching morphogenesis
of EpH4 mammary epithelial cells.

(D) EpH4 cells expressing ΔN-Vav2-dblmut (see structure in Fig. 11) form tubular structures upon stimulation with HGF/SF. (B) Mock-transfected cells exhibit branching morphogenesis in response to HGF/SF, as previously described (Niemann et al., 1998). (A, C) Non-stimulated controls. Bars: 50 μm. DH: Dbl-homology domain; PH: pleckstrin-homology domain; ZF: zinc finger.


© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 3.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated:
19.10.2004