Importance of the major extracellular domain of CD9 and the epidermal growth factor (EGF)-like domain of heparin-binding EGF-like growth factor for up-regulation of binding and activity.

Heparin-binding epidermal growth factor (EGF)-like growth factor (HB-EGF) is a member of the EGF family of growth factors. The membrane-anchored form of HB-EGF (proHB-EGF) is mitogenically active to neighboring cells as well as being a precursor of the soluble form. In addition to its mitogenic activity, proHB-EGF has the property of binding to diphtheria toxin (DT), serving as the specific receptor for DT. Tetramembrane-spanning protein CD9, a member of the TM4 superfamily, is physically associated with proHB-EGF at the cell surface and up-regulates both mitogenic and DT binding activities of proHB-EGF. To understand this up-regulation mechanism, we studied essential regions of both CD9 and proHB-EGF for up-regulation. Immunoprecipitation experiments revealed that not only CD9 but also other TM4 proteins including CD63, CD81, and CD82 associate with proHB-EGF on the cell surface. However, these TM4 proteins did not up-regulate DT binding activity of proHB-EGF. Transfection of a series of chimeric constructs comprising CD9 and CD81 showed that the major extracellular domain of CD9 is essential for up-regulation. Assays of DT binding activity and juxtacrine mitogenic activity of the deletion mutants of proHB-EGF and chimeric molecules, derived from proHB-EGF and TGF-alpha, showed that the essential domain of proHB-EGF for up-regulation is the EGF-like domain. These results indicate that the interaction of the extracellular domains of both molecules is important for up-regulation.

Scatchard plot analysis of DT binding to proHB-EGF indicated that increased DT binding with CD9 is due to an increase in the number of effective binding sites for DT (7). However, the precise mechanism of up-regulation remains to be elucidated.
In this report we have studied the regions of proHB-EGF and CD9 responsible for up-regulation. Results showed that the second extracellular domain of CD9 and the EGF-like domain of proHB-EGF are essential for the up-regulation of proHB-EGF, indicating that interaction of these molecules at the extracellular domains is important.
Plasmid Constructions-Plasmids encoding monkey CD9 (pRcT1843) and human proHB-EGF (pRcHBEGF) were used as described previously (11,13). cDNA encoding human CD63 and CD81 were kindly provided by Dr. H. Hotta (Kobe University) and Dr. S. Levy (Stanford University), respectively. cDNA of human CD82 was obtained by polymerase chain reaction from the human B cell cDNA library. These cDNAs were inserted into the HindIII/XbaI site in the expression vector, pRc/CMV. CD9/CD81 chimeras were constructed as follows. NdeI site and NarI sites were introduced in CD9 cDNA by substituting 324 C to A and 594 A to C, respectively. The mutations introduced were synonymous; thus no amino acid substitutions occurred. Then CD9/CD81 chimeras were constructed by substituting to the corresponding polymerase chain reaction fragments of CD81 cDNA with corresponding restriction enzyme sites as illustrated in Fig. 3A.
The deletion mutants of proHB-EGF were derived from pTHG-1 (11). XhoI and BamHI sites were introduced in pTHG-1 by site-directed mutagenesis from 445 CCA 447 to GAG and from 478 ACAA 481 to GGAT, respectively, as illustrated in Fig. 4A. These alterations resulted in amino acid substitutions of P149E, T160G, and T161S. The deletions were made by digesting proHB-EGF cDNA with corresponding restriction enzyme sites and linking them with oligonucleotide. In the case of the FRM construct, synthetic nucleotide of the corresponding region of transferrin receptor was inserted. ⌬Cyto was made by introducing stop codon at the 530 base pair by site-directed mutagenesis.
HB-EGF/TGF-␣ chimeras were based on the mutant proHB-EGF, which has additional BalI, XhoI, and BamHI sites (see Ref. 11 and Fig. 4A). proHB-EGF has DraII and KpnI sites in its sequence as shown in Fig. 4. Each domain of TGF-␣ was generated by polymerase chain reaction using the corresponding primer and inserted into HB-EGF cDNA digested by corresponding restriction enzymes. To construct TGJ, synthetic nucleotide corresponding to the juxtamembrane domain of TGF-␣ was inserted into HB-EGF cDNA digested by XhoI and BamHI (Fig. 5A).
Cell Culture and Transfection-Monkey Vero cells, human HT1080D cells, mouse L cells, and their derivatives were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml of penicillin G, and 100 g/ml of streptomycin. EP170.7 cells, obtained from Dr. J. Pierce (National Institutes of Health, Bethesda, MD) were grown in RPMI 1640 medium supplemented with 10% fetal calf serum, 5% WEHI-3 cell conditioned medium, 100 units/ml of penicillin G, and 100 g/ml of streptomycin. Transfection of plasmids into recipient cells was done as described previously (18). Transfected cells were cultured for 48 h and then used for further studies. LC, LH, and LCH cells are stable transfectants of L cells expressing monkey CD9 alone, human proHB-EGF alone and both monkey CD9 and human HB-EGF, respectively (7). Stable transfectants expressing TG-E alone (L/TGE cell) or both TG-E and CD9 (L/TGE/D cell) were isolated from L cells in a selection medium containing 40 g/ml of G418 after transfection with each of the plasmids. HT1080D cells were cloned in the selection medium containing 40 g/ml of G418 after transfection with monkey CD9 into HT1080 cells. VeroHKa cells were Vero cells stably expressing proHB-EGF and CD82.
Cell Surface Biotinylation, Immunoprecipitation, and SDS-PAGE-Cell surface biotinylation was carried out as described previously (14). Briefly, cells were washed and incubated in a biotinylating solution containing 0.2 mg/ml NHS-LC-Biotin (Pierce) for 30 min at 4°C. The reaction was stopped by the addition of 40 mM of glycine. For immunoprecipitation, cells were lysed with HBS (10 mM Hepes, 150 mM NaCl, pH 7.0) containing 10 mM CHAPS, 10 g/ml chymostatin, and 20 g/ml antipain. The lysates were cleared of insoluble material by centrifugation at 40,000 ϫ g for 30 min, and the supernatant was precipitated with primary antibodies at a concentration of 5 g/ml followed by the addition of Sepharose 4B-conjugated goat anti-mouse antibodies. The gel was washed with washing buffer (HBS containing 10 mM CHAPS) then boiled with SDS-PAGE sample buffer. Material recovered from the gel was analyzed by SDS-PAGE. Samples subjected to SDS-PAGE were electrotransferred to an Immobilon membrane. The membrane was blocked with TBS containing 3% bovine serum albumin (Sigma) at room temperature for 1 h, and proteins were detected by incubation with 100 ng/ml horseradish peroxidase-streptavidin (Pierce), then analyzed with an ECL-Western blotting kit (Amersham Pharmacia Biotech).
DT Binding and Antibody Binding Assay-Binding of 125 I-labeled DT to cells was measured as described previously (12). Nonspecific binding of 125 I-DT was assessed in the presence of a 100-fold excess of unlabeled DT. Specific binding was determined by subtracting the nonspecific binding from the total binding obtained with 125 I-DT alone. The amount of DT bound to the cells was calculated from the value of the specific binding of DT. The amounts of proHB-EGF expressed on the cell surface were determined as described previously (12). Briefly, cells were incubated with 5 g/ml anti-HB-EGF antibody in binding medium at 4°C for 2 h. After washing three times, cells were incubated with 1 g/ml 125 I-secondary antibody in binding medium. Finally cells were washed with washing buffer three times, and the cell-associated radioactivity was counted. DT binding activity was expressed as the calculated value B/A, in which A is the amount of proHB-EGF expressed on the cell surface, whereas B is the amount of DT bound to the cells.
Juxtacrine Assay-Juxtacrine mitogenic activity of proHB-EGF was monitored by measuring the incorporation of [ 3 H]thymidine into DNA of EP170.7 cells as described previously (6). Stable transfectants were plated in 24-well plates and incubated for 1 day. The cells were washed twice with Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 2 M NaCl and fixed with 4% paraformaldehyde for 5 min. The fixed cells were washed twice with 10% fetal calf serum/RPMI 1640, and EP170.7 cells were added in co-culture. After incubation for 36 h, [ 3 H]thymidine (37 kBq/ml) was added to the well, and the coculture cells were incubated for 4 h. The EP170.7 cells were harvested and analyzed for incorporation of [ 3 H]thymidine into DNA.

RESULTS
Association between proHB-EGF and TM4SF Proteins-We have shown that proHB-EGF forms a complex with CD9 and integrin ␣ 3 ␤ 1 (14). Recent studies also show that integrins ␣ 3 ␤ 1 and ␣ 6 ␤ 1 associate with not only CD9 but also other TM4SF proteins including CD63 and CD81 (19). Furthermore, TM4SF proteins including CD9, CD63, CD81, and CD82 have a tendency to associate each other (20). We examined whether other members of the TM4SF associate with proHB-EGF by co-precipitation experiments using a lysate of VeroHKa cells. Vero-HKa cells are stable transfectants of Vero cells overexpressing proHB-EGF and CD82 and also endogenously express CD9, CD63, and CD81 of TM4SF. Cell lysates of surface-biotinylated VeroHKa cells were immunoprecipitated with specific antibody against CD9, CD63, CD81, or CD82, and the precipitated material was subjected to SDS-PAGE and Western blotting followed by staining of the biotinylated proteins. Anti-CD9 antibody precipitated integrin ␣ 3 ␤ 1 , proHB-EGF, and unidentified proteins including those at 95 and 48 kDa and other minor bands, as well as CD9 itself, as previously shown (14) (Fig. 1, A and B, lane 2). Antibodies against CD63, CD81, and CD82 also co-precipitated integrin ␣ 3 ␤ 1 and proHB-EGF (Fig. 1, A and B, lanes [3][4][5]. Western blotting analysis of co-precipitated material probed with anti-integrin ␣ 3 antibody confirmed that the band at 150 kDa was integrin ␣ 3 (Fig. 1C). Similarly, anti-proHB-EGF antibody confirmed that the bands at 20 -27 kDa were proHB-EGF (data not shown). However, the bands of CD63 and CD82, which generally migrate over a broad range because of their extensive glycosylation, were difficult to see in precipitates of the biotinylated cell lysate, as previously men-tioned (19). Associations among TM4SF proteins were also detected. Anti-CD63, -CD81, and -CD82 antibodies co-precipitated the 27-kDa band of CD9 (Fig. 1A, lanes [3][4][5], as confirmed by Western blotting using anti-CD9 antibody (data not shown). Western blotting analysis also revealed the co-precipitation of CD63 and CD81 with anti-CD9 antibody (data not shown). These results indicated that CD63, CD81, and CD82 also associate with proHB-EGF and that these TM4SF proteins are included in the integrin ␣ 3 ␤ 1 , CD9, and proHB-EGF complex.
CD9 associates with proHB-EGF, and it up-regulates both DT binding activity and the mitogenic activity of proHB-EGF. We tested whether other TM4SF molecules up-regulate DT binding activity of proHB-EGF, as with CD9. LH cells, stable transformants of L cells expressing human proHB-EGF, were transfected with plasmids encoding cDNA of CD9, CD63, CD81, or CD82, and the amounts of 125 I-labeled DT bound to the cell surfaces of the transfected cells were measured. Transfection efficiency was about 60% for all transfections, and the expression of transfected cDNA at the cell surface was confirmed by indirect immunofluorescence. proHB-EGF molecules expressed on cell surfaces were determined by the binding of the anti-HB-EGF antibody followed by the 125 I-labeled secondary antibody. DT binding activity was normalized from the amount of proHB-EGF molecules expressed on the cell surface as described previously (12). Although CD63, CD81, and CD82 are able to associate with proHB-EGF, these TM4SF molecules did not enhance the DT binding activity of proHB-EGF at all ( Fig. 2A). Similar results were obtained using stable transfectants of LH cells expressing each of the TM4SF molecules (data not shown).
CD63, CD81, and CD82 associate with proHB-EGF but do not up-regulate DT binding activity of proHB-EGF. These results raise the possibility that CD63, CD81, or CD82 may inhibit the effect of CD9 in a competitive manner. We examined this possibility using HT1080D cells that are stable transformants of HT1080 cells expressing CD9. HT1080D cells express low amounts of human proHB-EGF endogeneously. After transfection with CD63, CD81, or CD82 into HT1080D cells, the DT binding activity of proHB-EGF was measured. As shown in Fig. 2B, the ectopic expression of these TM4SF proteins diminished the up-regulation effect of CD9. These results indicate that CD63, CD81, and CD82 have the ability to reduce the effect of CD9. However, because TM4SF proteins including CD9 associate with each other, the possibility cannot be ruled out that the association of TM4SF proteins with proHB-EGF in the present study is indirect, i.e. merely a consequence of their interaction with CD9. Thus, their ability to inhibit the potentiation of proHB-EGF activity by CD9 might be attributable to sequestration of CD9. No more enhancement of DT binding activity was observed from further transfection with CD9, probably because CD9 was saturated in HT1080D cells CD9/CD81 Chimeric Analysis-Next we determined the domain within CD9 that is essential for up-regulation of proHB-EGF. To examine this we made chimeric constructs between CD9 and CD81 and tested which chimeric molecules up-regulated the DT binding activity of proHB-EGF. Because CD63, CD81, and CD82 associate with proHB-EGF but do not upregulate the DT binding activity ( Fig. 2A), these molecules could thus be candidates for being partners of chimeric molecules. Among these TM4SF we selected CD81 as the partner. This because CD81 was localized predominantly at the cell surface, similar to CD9 (data not shown), whereas the majority of CD63 and CD82 was localized in lysosomes and other intracellular vesicles. The schematic structures of the chimeric constructs studied here are shown in Fig. 3A. Plasmids encoding the chimeric constructs were transfected into LH cells, and the DT binding activity of the cells was measured. All the chimeric molecules were expressed on the cell surface in the expected sizes and associated with proHB-EGF as shown in co-immunoprecipitation experiments with anti-CD9 antibody or anti-CD81 antibody (data not shown). Among the chimeric constructs, DR11 and TA2 enhanced DT binding activity, whereas DR8 did not enhance it (Fig. 3B). These results indicated that the second extracellular loop of CD9 is important for up-regulation.
Domain of proHB-EGF Necessary for Up-regulation by CD9 -We also studied the domain(s) within proHB-EGF necessary for up-regulation by CD9. To examine this we made deletion mutants of proHB-EGF. proHB-EGF can be structurally divided into seven domains: pre, pro, heparin-binding, EGF-like, juxtamembrane, transmembrane, and cytoplasmic domains. Of these domains, the heparin-binding, EGF-like, juxtamembrane, or cytoplasmic domain was deleted, and the constructs were termed ⌬HBD, ⌬EGF, ⌬JxM, and ⌬Cyto, respectively. FRM is a chimeric construct in which the transmembrane domain of proHB-EGF was substituted to that of the transferrin receptor. Because the transmembrane domain is essential for anchoring of membrane proteins, a mutant form lacking transmembrane domain was not constructed. Schematic structures of these constructs are shown in Fig. 4A. It should be noted that deletion mutants shown here were made from the pseudo-wild type of proHB-EGF, which has amino acid substitutions of P149E, T160G, and T161G. The pseudowild type had DT binding activity, which was as up-regulated by CD9 as that of wild type proHB-EGF.
Plasmids encoding each proHB-EGF mutant and plasmids encoding CD9 or vectors only were introduced into L cells. These deletion mutants were expressed on the cell surface of the L cells in expected sizes and coprecipitated with CD9 (Fig.  4B). The DT binding activity of the transfected cells was next determined. The results are shown in Fig. 4C. Under these conditions DT binding activity of wild type proHB-EGF was up-regulated about 15-fold by the transfection with CD9. ⌬HBD, ⌬Cyto, and FRM were significantly up-regulated by CD9, although the up-regulation for FRM seemed to be lower than wild type. These results indicated that the heparin-binding, transmembrane, and cytoplasmic domains were not essential for up-regulation and that the remaining domains must be responsible for the up-regulation. Because DT binding activity was not observed in the cells transfected with ⌬EGF or ⌬JxM, regardless of the presence or absence of CD9, the effect of CD9 on the DT binding activity of ⌬EGF or ⌬JxM could not be determined. ⌬EGF and ⌬JxM were expressed on the cell surface in amounts similar to wild type proHB-EGF (data not shown); therefore deletion of these domains must cause a loss of DT binding activity. The loss of DT binding activity in ⌬EGF is reasonable because DT binds to the EGF-like domain of proHB-EGF (11). The failure of ⌬JxM DT binding is surprising and is probably due to the impaired access of DT to the EGFlike domain of proHB-EGF without the juxtamembrane domain. 2 HB-EGF/TGF-␣ Chimeric Analysis-Deletion mutant analysis still retained the possibility that either the EGF-like or the juxtamembrane domain was responsible for the up-regulation mechanism. To determine the domain necessary for up-regulation, we made a series of chimeric constructs between proHB-EGF and TGF-␣. TGF-␣ does not bind DT (11), and the juxtacrine mitogenic activity of TGF-␣ is not up-regulated by CD9 (6). N-terminal regions containing pre, pro, and heparin-binding, EGF-like, and juxtamembrane domains or both transmembrane and cytoplasmic domains of proHB-EGF were replaced 2 T. Takahashi and E. Mekada, manuscript in preparation. with the corresponding domains of TGF-␣, termed TGN, TGE, TGJ, or TGMC, respectively, as illustrated in Fig. 5A. TGN-JMC, the chimera in which the EGF-like domain of proTGF-␣ was replaced with that of proHB-EGF was also constructed. These chimeras were transfected into L cells with or without CD9 cDNA, and DT binding activity was determined. As shown in Fig. 5B, DT binding of TGN, TGJ, and TGMC were upregulated 5-10 fold by the expression of CD9, whereas that of wild type proHB-EGF was 10-fold. Consistent with the deletion analysis, the substitution of the N-terminal, juxtamembrane, transmembrane, or cytoplasmic domain with the corresponding domain of TGF-␣ did not affect the up-regulation properties.
TGE had no DT binding activity, as expected, and thus the effect of CD9 could not be determined. It is also noteworthy that the DT binding of TGNJMC was also up-regulated by the expression of CD9. These results indicate that the pre, pro, heparin-binding, juxtamembrane, and cytoplasmic domains are not essential and that these domains are dispensable to the corresponding domains of TGF-␣ for up-regulation by CD9, whereas the EGF-like domain of proHB-EGF is responsible for up-regulation.
Although the above studies indicated that the EGF-like domain of proHB-EGF would be essential for up-regulation by CD9, DT does not bind to the EGF-like domain of TGF-␣ and thus up-regulation of DT binding activity in TGE by CD9 could not be directly determined. To examine whether the replacement of the EGF-like domain of proHB-EGF by that of TGF-␣ results in the loss of up-regulation, juxtacrine growth factor assay was performed as described previously (6). To do this assay we used stable transfectants expressing proHB-EGF alone (LH), TGE alone (L/TGE), both proHB-EGF and CD9 (LCH), or both TGE and CD9 (L/TGE/D). As shown in Fig. 6A, LCH cells display much higher juxtacrine activity than LH cells, despite LCH cells expressing lower numbers of proHB-EGF molecules on the cell surface (Fig. 6B). In the case of TGE, two independently isolated clones of L/TGE4/D (L/TGE4/D2 and L/TGE4/D7) cells had rather lower juxtacrine activity than L/TGE4 cells (Fig. 6A), although these three cell lines expressed similar numbers of TGE molecules on the cell surface (Fig. 6B). Similar results were also obtained using other stable transfectant cell lines, L/TGE15 and L/TGE15/D2 (data not FIG. 4. Analysis of proHB-EGF deletion mutants. A, schematic representation of proHB-EGF deletion mutants. ⌬HBD, ⌬EGF, ⌬JxM, and ⌬Cyto are deletion mutants of the heparin-binding, EGF, juxtamembrane, and cytoplasmic domains, respectively. FRM is a mutant whose transmembrane domain is substituted with that of transferrin receptor. The striped box represents the region of the transferrin receptor. Bl, BalI; D, DraII; X, XhoI; Bm, BamHI; K, KpnI. B, association of CD9 with proHB-EGF mutants. LC cells were transfected with control vector or proHB-EGF mutants. After incubation for 48 h, cells were surface biotinylated and lysed with 10 mM CHAPS solution. The lysates were immunoprecipitated with anti-CD9 antibody, and the precipitates were analyzed by SDS-PAGE. The asterisk shows bands of CD9. C, up-regulation of DT binding activity of proHB-EGF deletion mutants by CD9. Plasmids encoding proHB-EGF, ⌬HBD, ⌬EGF, ⌬JxM, FRM, or ⌬Cyto were transfected with control Rc/CMV vector or plasmids encoding CD9. After incubation for 48 h, the same aliquots were subjected to determine DT binding activity. DT binding activities are expressed as relative values, which were obtained by comparing with those of wild type proHB-EGF without CD9. Data represent the means Ϯ S.D. of the results obtained from triplicate samples. shown). These results indicated that the juxtacrine growth factor activity of TG-E was not up-regulated by CD9. From our DT binding and juxtacrine mitogenic assays, we concluded that the EGF-like domain of proHB-EGF is essential for up-regulation by CD9. DISCUSSION A number of TM4SF proteins have been described in a wide variety of animal cells (16). A characteristic feature of this family of proteins is to form complexes with a variety of membrane proteins. They are thus supposed to be "adapters for membrane proteins" or "molecular facilitators" that mediate the formation of large molecular complexes and allow them to function more efficiently. CD9 is one of the best characterized TM4SF members, and a number of studies have suggested that CD9 is involved in cell signaling (21,22), cell growth (6), cell motility (23,24), cell adhesion (25)(26)(27), tumor cell metastasis (28 -30), and development and maintenance of neural system (31)(32)(33)(34). However, despite such fundamental roles, evidence to demonstrate the role of CD9 in molecular complexes is limited. One of the functions of CD9 clearly demonstrated is the upregulation activity of CD9 for proHB-EGF. We originally identified CD9 as a diphtheria toxin receptor-associated protein (15). In subsequent studies we have shown that CD9 greatly up-regulates DT binding activity and the juxtacrine activity of proHB-EGF (6,7). Here we study the molecular mechanism of up-regulation by analyzing essential regions of both CD9 and proHB-EGF for up-regulation activity. Results obtained from these studies are useful for understanding the molecular nature of the complex formation.
Previous studies indicated that the number of proHB-EGF molecules at the cell surface was not changed in the presence or absence of CD9 (7). Furthermore, direct binding of DT with CD9 has not been detected. Scatchard plot analysis of DT binding to proHB-EGF indicated that increased DT binding with CD9 is due to an increase in the number of effective binding sites for DT rather than any increased binding affinity for DT (7). Therefore, an increase in the number of effective binding sites must be attributable to protein-protein interaction between proHB-EGF and CD9. Co-precipitation studies showed that not only CD9 but also CD63, CD81, and CD82 are associated with proHB-EGF and integrin ␣ 3 ␤ 1 . However, among these TM4SF proteins only CD9 up-regulated the DT binding activity of proHB-EGF. We took advantage of the inability of CD81 to up-regulate to analyze the region necessary for up-regulation. A series of chimeric molecules was made to analyze the region of CD9 essential for up-regulation. All the chimeric molecules expressed on the cell surface, but only constructs that had a second extracellular domain of CD9 had up-regulation activity. Hence, we concluded that the second loop of CD9 is sufficient for up-regulation and that all of the transmembrane domains and the first loop between the first and the second transmembrane domains are exchangeable with CD81. Consistant with our present results, a recent report has suggested that the latter half of CD9 is necessary for up-regulation of DT sensitivity (35).
We also studied the domain of proHB-EGF necessary for up-regulation by using a series of deletion mutants of proHB-EGF and chimeric molecules between proHB-EGF and TGF-␣. DT binding assay revealed that none of the domains of proHB-EGF, except for the EGF-like domain, were essential for upregulation, whereas the chimeras that have the EGF-like domain of proHB-EGF were all up-regulated by CD9, suggesting the involvement of the EGF-like domain in up-regulation. The chimeric molecule TGE, which has the EGF-like domain of TGF-␣, does not bind to DT; thus up-regulation of DT binding was not determined by this chimera. To circumvent this difficulty, juxtacrine mitogenic assay was performed, and the results showed that juxtacrine activity of TGE is not up-regulated by CD9. Thus, the domain of proHB-EGF necessary for up-regulation is the EGF-like domain. From these results, together with data obtained from CD9/CD81 chimera, we concluded that the extracellular domains of CD9 and the EGF-like domain of proHB-EGF are important for up-regulation.
Our studies demonstrated that proHB-EGF associated with CD63, CD81, and CD82 as well as with CD9, but these TM4SF proteins and some CD9/CD81 chimera failed to up-regulate DT binding activity. Moreover, it was shown that TGE was associated with CD9 ( Fig. 4B) but that the juxtacrine activity was not up-regulated by CD9. Therefore, association of CD9 with proHB-EGF is probably essential but is not enough for upregulation. We attempted to define the domains responsible for the association of proHB-EGF and CD9 in this study. However, because all of the deletion mutants of proHB-EGF, including ⌬EGF and FRM, were co-precipitated with CD9, at more or less the same efficiency compared with wild type proHB-EGF, these studies did not allow us to define a particular domain of proHB-EGF necessary for association. These results may suggest that multiple domains, e.g. the EGF-like domain and the transmembrane domain, are involved in association, but the possibility of the artifact being due to the overexpression of CD9 and proHB-EGF cannot be ruled out.
The present study showed that not only CD9 but also other TM4SF members are associated with proHB-EGF. The role of the TM4SF members in this complex is unclear, but they would have different functions at cell-cell contact sites, because only CD9 of these TM4SF members up-regulates proHB-EGF. Subcellular localization of each TM4SF member seems to be different. Immunofluorescence studies showed that CD9 and CD81 localized mainly at the cell surface, whereas CD63 mainly localizes at lysosomes and secreted vesicles. Thus CD63 may have functions in the transport of proHB-EGF or integrin ␣ 3 ␤ 1 to or from cell surfaces. Consistent with this notion, association of CD63 with PI4 kinase has been reported, and its role for integrin internalization has been suggested (36). Diphtheria toxin is internalized with proHB-EGF, but the domains responsible for internalization have not been found. It would be intriguing to speculate that internalization of DT-bound proHB-EGF is achieved by assistance from a TM4SF member. Further study is needed to clarify the role of each TM4SF member in the complex.
In conclusion, the present domain analysis of CD9 and proHB-EGF suggests the importance of the interactions of both molecules at their extracellular domains. The precise molecular mechanism for up-regulation still remains to be clarified. Further studies, especially of the extracellular domains, would help with understanding the molecular mechanism of up-regulation and also help to create dominant negative forms of these proteins.