JOURNAL OF CELLULAR PHYSIOLOGY 145543-548 (1990)
Binding of Epidermal Growth Factor to Its Receptor Is Affected by Membrane Phospholipid Environment TAMIKO KANO-SUEOKA,* DAVID M. KING, HAROLD A. FISK, AND STEFANIE J. KLUG Department of Molecular, Cellular and Developmental Biology, University of Colorado,
Boulder, Colorado 80309-0347
Cells of epithelial origin generally require ethanolamine to grow in culture; when these cells are grown without ethanolamine, the phosphatidylethanolamine content of their membrane phospholipid becomes 1/2 to 1/3 of the normal amount, and growth stops. We have hypothesized that growth ceases because the
phospholipid environment becomes unsuitable for membrane-associated function. Using ethanolamine-requiring rat mammary cells, we have investigated the possible effect of phosphatidylethanolamine deficiency on the binding characteristics of epidermal growth factor. Apparent dissociation constant for the high-affinity sites in cells having normal membrane phospholipid was 1.7 x 1 O-'OM, whereas that of phosphatidylethanolamine-deficient cells was 2.7 X 10-'OM: the difference was small, but significant. Pretreatment with phorbol ester caused the loss of high-affinity sites in cells having normal membrane, whereas binding characteristics of epidermal growth factor became refractory to the pretreatment in phosphatidylethanolamine-deficient cells. In addition, the rate of internalization of bound epidermal growth factor in phosphatidylethanolamine-deficientcells was about 1/4 of normal cells. Further, whether cells had normal or phosphatidylethanolamine-deficient membranes seemed to affect the phosphorylation patterns of membrane proteins in response to epidermal growth factor or phorbol ester. These results suggest that membrane phospholipid environment affects the activity of the epidermal growth factor receptor. Cells of epithelial origin generally require ethanolamine (Etn) (or phosphatidylethanolamine [PEN to grow in culture, whereas fibroblasts or neurocells proliferate well in the absence of Etn (Kano-Sueoka and King, 1987a,b). The composition of membrane (plasma and total) phospholipid of Etn-requiring cells is similar to that of in vivo tissue when they are grown in the presence of Etn (Kano-Sueoka et al., 1983).However, in the absence of supplemental Etn, the content of PE in their membrane becomes 1/2 to 1/3 of the normal amount and that of phosphatidylcholine (PC)increases by 30% (Kano-Sueoka et al., 1983; Kano-Sueoka and King, 198713).In the absence of Etn, the rate of growth slows down as PE content is reduced (and PC content increases), and growth eventually stops (Kano-Sueoka et al., 1983). It is probable that proliferation ceases because the phospholipid environment becomes unsuitable for membrane-associated functions. Using a typically Etn-requiring rat mammary carcinoma cell line, 64-24, we previously demonstrated that the apparent dissociation constant (Kd) of phorbol 12,13-dibutyrate (PDB) was significantly higher in cells having PEdeficient membrane, and the growth-promoting activity of PDB, which can readily be demonstrated in cells having normal membrane phospholipid, was not observed in PE-deficient cells (Kano-Sueoka and King, 1988).These results suggested that the transduction of 0 1990 WILEY-LISS, INC.
the PDB signal subsequent to PDB binding to the cell membrane was profoundly affected by the nature of the plasma membrane. In the present study we have used the 64-24 cell line to examine the effect of PE-deficienc on the behavior of the epidermal growth factor (E F) receptor. The EGF receptor is a well-characterized cell surface receptor which consists of a single transmembrane polypeptide and possesses tyrosine kinase activity in its cytoplasmic domain (see reviews, Schlessinger, 1986; Carpenter, 1987; Carpenter and Cohen, 1990). Binding characteristics of EGF to its receptor as well as the autophosphorylation and tyrosine kinase activity of the receptor are modulated by EGF and other molecules including protein kinase C. Generally about 10%of the receptors exist in a high-affinity state which is thou ht to be responsible for the mitogenic activity of E F, whereas 90% of the receptors are in a low-affinity state. Phosphorylation of the receptor by rotein kinase C abolishes the Ipumor-promoting can activate rotein binding of EG to its receptor by phosphorylating the EGF receptor molecule
E
8
5
Received June 22, 1990; accepted September 7,1990. 'To whom reprint requestsicorrespondence should be addressed.
544
KANO-SUEOKA ET AL.
at specific sites (Friedman et al., 1984; Hunter et al., tal ones. Proteins were estimated by the method of 1984; Iwashita and Fox, 1984; McCaffrey et al., 1985; Schacterle and Pollack (1973), using the wells that were used to estimate nonspecific bindings. The ScatDavis and Czeck, 1985; Davis, 1988). The results of the present study indicate that PE- chard anal sis was carried out using a computer prodeficiency in cellular membranes affects not only the gram devecped by Dr. Noboru Sueoka (University of binding characteristics of EGF, but also the rate of Colorado, Boulder, CO). internalization of the bound EGF. Further, phosphoryIn vitro radiolabeling of membrane proteins lation atterns of membrane proteins in cells having Cells having normal and PE-deficient membrane norma or PE-deficient phospholipid as examined by in vitro radiolabeling show that these cells are likely to phospholipids were pre ared using 100 x 15 mm culture dishes as describe previously (Kano-Sueoka and respond differently to EGF or PDB. King, 1988). Pretreatment of these cells with PDB was MATERIALS AND METHODS also carried out as described above. For the pretreat1251-mouseEGF (specific activity, 105-171 pcilpg) ment with EGF, cells were incubated with 20 ng/ml was obtained from Du Pont-New England Nuclear EGF for 10 min at 37°C. Cells were washed twice with (Boston, MA) and Collaborative Research (Bedford, ice-cold PBS, scra ed, and suspended in buffer containMA), and mouse EGF, receptor grade, was obtained ing 0.01 M Tris- C1, pH 7.8, 2 mM MgClZ,20 pgiml from Collaborative Research and Upstate Biotechnol- aprotinin, and 2 mM PMSF. Cells were then ru tured o y (Lake Placid, NY). PDB was purchased from Sigma at 4°C by means of pressure homogenization at 00 psi with Nz gas in a Parr cell disruption bomb. Nuclei were (it. Louis, MO). removed by centrifuging at 10006' for 10 min and the Cells and cell culture membrane fraction was pelleted at 100,OOOgfor 60 min. A rat mammary carcinoma cell line 64-24 has been The pellet was sus ended in 20 mM HEPES buffer, pH maintained in Dulbecco's modified Eagles' medium 7.4, and recentri uged at 100,000 for 60 min and (DME) supplemented with 5% horse serum and 2.5% resuspended in the same buffer. T e membrane fracfetal calf serum. Cells having normal and PE-deficient tions were quickly frozen and stored at -70°C. Some membrane phospholipids were prepared for the EGF membrane preparations were made in the presence of binding studies as described previously (Kano-Sueoka 50 mM NaF and 1 mM Na3V05. Phosphorylation reactions contained 20 to 70 pg protein, 20 mM HEPES, and King, 1988). pH 7.4, 1 mM MnClZ,0.13 pg/ml BSA, 1-2 pM ATP, Binding of '251-EGF to cells and 2-5 pCi y3'P ATP. In some reactions 16.7 mM NaF The cells were plated at seeding densities of 2.5- and 250 pM Na3V05 were included. The phosphoryla3.0 x lo5 and 3.0-3.5 X lo5 cells/well (35 mm in diam- tion reaction was carried out for 5 min on ice or at room eter) for normal and PE-deficient cells, respectively, in tem erature. Phosphorylated membrane proteins were multi-well plates. One day later the cells were washed ana yzed by sodium dodec 1 sulfate gel electrophoresis twice with DME medium, and 0.2 ml DME containing using 7.5% acrylamide ge s. 15 mM N-2-Hydroxyethyl iperazine-N'-2-ethanesulfoRESULTS nic acid (HEPES), 1m /m bovine serum albumin, and EGF bindin to cells having normal and desired amounts of 126-EGFwere added in duplicates PE- eficient membranes or triplicates. For saturation binding experiments, the binding was carried out at 4°C for 90 min with entle 64-24 cells were grown in the presence or absence of shaking. To estimate nonspecific binding 1 p Im EGF 10 pM Etn for 10 to 15 enerations before the binding was added in addition to the radioactive EG . At the assays were performet Time course of binding of end of the incubation, the cells were washed with 1ml '251-EGF showed that the s ecific binding of 1251-EGF of ice-cold PBS six times, lysed with 0.2 ml 1N NaOH, to both types of cells (norma and PE-deficient) reached and the radioactivity was counted in a Beckman maximum at around 90 min and then gradually dey-counter. Determination of cell surface-bound and creased, and in PE-deficient cells the amount of EGF internalized lz5I-EGF was carried out similarly to the boundimg protein was about 70% of that bound by cells method of Haigler (1983). Briefly, the cells were incu- having a normal phos holi id composition (Data not bated with 1251-EGFat 37°C for varied lengths of time, shown). The binding o EG to these cells was further washed with ice-cold PBS five times, and then incu- studied by Scatchard analysis (Scatchard, 1949). Varbated in 0.5 m10.2 M acetic acid containing 0.5 M NaC1, ied amounts of lZ5I-EGFwere bound to the cells in midpH 2.5 for 8 min at 4°C with rotary shakin . The acetic to late-log hase under saturation conditions. Nonacidisalt solution (which contains extrace lular EGF) specific biniings were determined in the presence of was removed from the well, and the cells were rinsed excess nonradioactive EGF (1 pg/ml) and were meawith 0.5 ml of the acetic acid/salt solution which was sured separately for each concentration of radioactive combined with the original acidhalt extract. The re- EGF. The results indicated that the level of saturation maining radioactivity (intracellular EGF) was recov- of bound EGF was consistently higher (1.5- to 1.7-fold) ered by lysing the cells with 1m10.5 N NaOH at 70°C in normal cells than in PE-deficient cells (Fi .lA). The for 1 hr. For the pretreatment with PDB, 100 ng/ml of data were plotted according to the method o Scatchard PBD dissolved in dimethylsulfoxide (DMSO)was added (1949) to yield apparent dissociation constants (Kds) to the cells and incubated at 37°C for 10 min prior to the and the number of binding sites. The results are washin and addition of radioactive EGF. Control wells indicated in Figure 2A. Under our experimental conreceive DMSO in the same amount as the experimen- ditions about half of the receptors in both types of cells
P
B
#
f
P
P
P
a
7
fi
%!
Y
P P
5
B
%
MEMBRANE PHOSPHOLIPID AND EGF RECEPTOR
-
80
f
60
545
5
a
- 40
=E U
20
0 2
m
O
0
20
40
20
0
EGF
Fig. 1. Dose saturation binding of EGF to 64-24 cells having normal and PE-deficient membrane and the effect of preincubation with PDB. 64-24 cells having normal and PE-deficient membrane phospholipids were prepared by growing cells in the presence and absence of 10 pM Etn as described previously (Kano-Sueoka and King, 1987b) and also in “Materials and Methods.” The cells were incubated with varied amounts of lZ5I-EGFat 4°C for 90 min and processed as described in “Materials and Methods.” A: Dose saturation binding of EGF to
40 (nqlrnl)
normal and PE-deficient cells. B:Dose saturation binding of EGF to normal cells with and without the pretreatment of PDB. C: Dose saturation bindin of EGF to PE-deficient cells with and without the , cells having normal membrane pretreatment of 8DB. phospholipid; - - - -, cells having PE-deficient membrane phospholipid. Thick and thin lines represent cells treated and not treated with PDB, respectively.
EGF bound (frnol/rng protein)
Fig. 2. Scatchard plots of the dose saturation binding of EGF to 64-24 cells. Data presented in Figure 1A-C were plotted according to Scatchard (1949)in order to obtain apparent Kd values and the number of binding sites. A-C correspond to A-C of Fi re 1 The number of cellsimg protein for normal and PE-deficient cells were 8.0 x lo6 and 5.6 x 10P , respectively. ’ .
were high-affinity species, and the rest were lowaffinity species. Apparent Kd values for the hifhaffinity sites were 1.64 x lop1’ M and 2.86 X 10-1 M for normal and PE-deficient cells, respective1 difference in Kd values between these two cel types The was small (at most 2-fold), but was repeatedly observed as shown in Table 1. The number of the hi h-affinit binding sites seemed to be similar between t e two ce 1 types, although that of PE-deficient cells was considerably more variable (Table 1). Reliable estimation of the Kd values or the number of binding sites for the low-affinity species was difficult because of the small number of data points, as seen in Figure 2A. However, the analysis suggests that the Kd values a pear to be between 3.5 to 6 x 10-l’ M and the num er of sites seems to be several thousand per cell for both cell types from three experiments (data not shown). Effect of phorbol ester on EGF binding We have previously shown that the PE-deficiency has abnormal effects on cellular functions mediated by phorbol esters (Kano-Sueoka and King, 1988). Phorbol esters are known to cause the loss of the high-affinity binding sites of EGF receptors by activating protein
9’
f l P
l
kinase C (Friedman et al., 1984; Hunter et al., 1984; Davis and Czech, 1985). Therefore, it is possible that the mode of modulation of EGF bindin by phorbol esters may become abnormal under PE- eficient conditions. Indeed, a striking difference in the nature of the modulation of EGF binding by PDB was observed between the two types of cells when they were pretreated with 100 n /ml of PDB for 10 min followed by the binding of EGf. In the cells having normal membrane phospholipids the modulation of the EGF receptor took place as expected. Namely, in PDB-pretreated cells the level of saturation of the bound EGF was reduced to half or sometimes less than half as compared to the untreated cells (Fig. 1B). According to the Scatchard analysis the majority of the binding sites after PDB treatment had a Kd value of 3.7 x lo-’’ M with the number of sites being 4,200 siteskell (Fig. 2B). The data thus indicate that PDB treatment causes the loss of the high-affinity, but not of the low-affinity sites. When the same experiment was carried out using PE-deficient cells, entirely different results were obtained (Figs. 1C,2C). Name1 , the binding characteristics of EGF were similar w ether the cells were pretreated with PDB or not, indicating that under these
%
x
546
KANO-SUEOKAET AL. TABLE 1. Binding of EGF to 64-24 cells having normal and phosphatidylethanolamine-deficient membranes: analysis of high affinity binding sites' Cell PE-deficient Kd Binding sites M) (per cell)
Exp.
(X
1
2.86 2.50 2.78 2.71 f .112
2
3
Averages
'Deviations of S.E.M. fmm the averages were
~s.E.M.
't
I
I
50
2,700 7,000 8,100 5,900 1,700
+
PDB (ng/ml) Fig. 3. Dose-response effect of PDB on the binding of EGF to normal and PE-deficient 64-24 cells. 64-24 Cells were prepared as in the experiments described in Figures 1and 2. The cells were incubated for 10 min with varying amounts of PDB at 37°C and then lZ5I-EGF (30,000cmpiwell) was added and incubated further at 4'C for 100 min. The cells were processed as described in Figure 1.M, cells having normal membrane phospholipid; O---0, cells having PE-deficient membrane phospholipid.
conditions PDB hard1 modulated EGF binding properties. The effects of P B pretreatment on EGF binding to both normal and PE-deficient cells were reproducible. The dose-response effect of PDB on EGF binding in the two t es of cells further confirmed that the binding of GF in PE-deficient cells was not affected b PDB within the range of concentrations of PDB testeJ however, in normal cells the binding was reduced in a dose-de endent manner, and at 50 ng/ml the bindin was re uced to less than 50% of the level of untreate cells (Fig. 3). PE-deficient cells bound more EGF than normal cells at all doses of PDB. This is due to the fact that PDB treatment resulted in the loss of high-affinity sites in normal, but not in deficient cells as indicated by Scatchard analysis (Fig. 2; Table 1). These results therefore indicate that the modulation of the EGF receptor by PDB, at least with re ard to the binding affinity, does not take place in ce 1s having PE-deficient membranes.
6
P
B
i
5
(x 10-lo M)
Binding sites (per cell)
1.64 1.96 1.47 1.69 f .14
6,500 5,200 4,750 5,500 f 550
+ 2.2% for Y-intercept(boundlfree) and *
/%
I00
Normal Kd
2.6%for the slope.
Internalization of bound EGF Whether PE-deficiency in the membrane affects the rate and extent of the internalization of receptor-bound EGF was determined by measuring the amount of 1251-EGFin aciasalt-resistant (intracellular) and acid/ salt-sensitive (cell surface-bound) fractions. As shown in Figure 4,1251-EGFbound to the cell surface reached a lateau within 5 min after the addition of radioactive EP*F in both types of cells. The plateau level was 4- to 5-fold higher in cells having normal membrane phosholi id. Moreover, the rate of internalization of the Koung lZ5I-EGFwas markedly higher in normal cells than PE-deficient cells. In order to determine if PDB also affects the rate of internalization of bound EGF, the acidhalt-resistant and -sensitive fractions of the bound lZ5I-EGFafter 30 min incubation at 4°C were examined. Preincubation time with PDB was 10 min at 37°C. A ain, as in the case of saturation binding ex eriments escribed above the amount of 1251-EGFfouncfon the cell surface as well as inside of the cells was significantly reduced by phorbol ester treatment in normal cells, whereas in PE-deficient cells horbol ester did not affect the amount of 1251-EG found either on the surface or inside of the cells (Table 2). Phosphorylation of proteins in membrane fractions The results so far obtained suggest that events that take lace subsequent to EGF binding and interaction of E F receptor with other si aling pathways are influenced by the deficiency of E in the membranes. In order to test the above possibilities the effect of treatment of cells with PDB and EGF on phosphorylation of membrane proteins was examined in cells having normal and PE-deficient membrane. Post nuclear membrane fractions were prepared from these cells after treatment with either PDB or EGF and ghosporylated in vitro using y-32PATP as described in Materials and Methods." The results of PDB treatment and EGF treatment are shown in Figures 5 and 6, respectively. In normal cells the PDB treatment caused si nificant changes in banding patterns of phos horylate proteins as indicated by arrows. Intensity o the label in some bands was reproducibly suppressed as a result of the PDB-treatment and that of some was enhanced. In contrast, in cells having PE-deficient membranes the PDB treatment did not seem to change the phosphory-
cf
P
8
P
f!
d
547
MEMBRANE PHOSPHOLIPID AND EGF RECEPTOR
5-
-
Etn;
...’
4-
P
0
x
3-
C .-
0,
c
g F
2-
\
__---
PDB:
- -
+ +
+
-
I
+ -205 -116
- 97.4
-
66
-
45
______----A 1
I
20
“0
40
Time (min.)
Fig. 4. Time course of changes in the amount of surface-bound and intracellular Iz5I-EGFin normal and PE-deficient 64-24 cells. Ex r imental procedure was essentially the same as that describefin Figure 1, except that the binding was carried out at 37°C. Surfacebound and intracellular lz5I-EGF were determined as described in “Materials and Methods.” Pretreatment with PDB was carried out for 10 min at 37°C. ,cells having normal membrane;----, cells having PE-deficient membrane; A ,A , surface bound; 0, 0, intracellular. ~
TABLE 2. Effect of PDB on internalization of the EGF receptor in
cells having normal and PE-deficientmembranes
Fig. 5. Phosphorylation of isolated normal and PE-deficient membrane fractions in response to PDB treatment. 64-24 Cells with normal and PE-deficient membranes were repared as in Figure 1.Cells were treated with 100 ng/ml PDB in DhSO or with an equal amount of DMSO without PDB for 10 min at 37°C. Post-nuclear membrane fractions were prepared, and phosphorylation of membrane proteins was performed as described in “Materials and Methods.” Analysis of r t e i n s was carried out by SDS polyacrylamide gel electrophoresis ollowed by autoradiography. Differences in phosphorylationpatterns between PDB treated and non-treated cells are indicated by arrows.
cpm bound/mg protein
PDB treatment
Ethanolamine
Site
-
+
Surface Internal Surface Internal
850’ 4,886 1,521 9,489
771 4,691 786 3,650
’Average of duplicate samples. The results are reproducible
lation atterns. A similar result can be seen in the case of EG treatment (Fig. 6). Here again EGF reproducibly altered banding patterns of phos hor lated proteins in normal cells, whereas that of P -de icient cells seemed not to be affected by EGF. In these studies we were unable to clearly identify the EGF receptor protein, perhaps because 64-24 cells possess, according to our estimate, only around 10,000 EGF receptors per cell.
P
E F
DISCUSSION
Our previous studies indicated that the properties of
PDB binding to PE-deficient membranes were different from those of normal membranes, and the stimulation of cell proliferation by PDB, which can be demonstrated in cells having normal membranes, did not occur in PE-deficient cells (Kano-Sueoka and King, 1988). In the present study, binding characteristics of EGF in these two types of cells were investigated. The results suggest that EGF binding to the hi h-affinity receptor site is affected by properties of the p ospholipid bilayer. Analogous to the situation of PDB binding, the Kd
a
values were different by about two-fold between the two ty es of cells: i.e., the affinity of EGF to its receptor was a ittle weaker in PE-deficient cells. Although this difference was repeatedly observed and was significant, it is not certain whether this small difference can bring about any significant hysiological effects..However, a ronounced effect of P -deficiency was observed when ginding of EGF was examined after PDB pretreatment. In cells having normal membrane, the high-affinity binding was lost after PDB treatment, while in PEdeficient cells the nature of EGF binding was similar whether the cells were treated with PDB or not. These results thus suggest that a cascade of reactions initiated by PDB is interrupted at some point in PEdeficient cells. The most plausible explanation of the above (assumin that rotein kinase C is the major protein affected y PDB is that either 1)the activation of protein kinase C by PDB does not take place in normal fashion in PE-deficient cells or 2) the activation can take place, but the modulation of EGF receptor by protein kinase C does not occur. The present results do not indicate if either is the case. The effect of EGF and PDB treatment in these cells on the distribution of protein kinase C in subcellular fractions and the extent of phosphorylation of the serine, threonine, and tyrosine residues of EGF receptor molecules are currently under investigation. These studies will like1 yield information regarding the particular step w ich is sensitive to PE-deficiency. In addition to the binding characteristics per se, PE-deficiency affected that manner in which the bound EGF was processed. The
’I
E
g
P
x
548
Etn: EGF
KANO-SUEOKA ET AL.
+
+
+
-
I
I
+
-205
-116 97.4
-
more than two-fold (Schroeder et al., 1984).
-66
ACKNOWLEDGMENTS We thank Dr. Marsha R. Rosner (University of Chica 0, Chica 0, IL) for her valuable help at the initia stage o f t is work. This work was supported by NIH grant CA 30545.
-45
Fig. 6. Phosphorylation of isolated normal and PE-deficient membrane fractions in response to EGF treatment. 64-24 cells with normal and PE-deficient membranes were prepared as described in Figure 5. Cells were treated with or without 20 ngiml EGF for 10 min at 37°C. Post-nuclear membrane fractions were prepared, phospho lated, and analyzed as described in Figure 5. Arrows indicate ban% showing different extent of phosphorylation between EGF treated and nontreated cells.
manner in which phorbol ester affected the internalization was also consistent with the way PDB affected the binding affinity of EGF as shown in Figures 1 and 2. Our results however do not indicate whether change in fluidity or charge brought about by PEdeficiency influences the process of internalization, or if the internalization is in some way correlated with the activity of the EGF receptors. The method employed in the present studies to analyze membrane-associated proteins does not yield detailed information regarding phosphorylation of individual proteins. However, it appears to show that in cells having normal membrane phospholipids the phosphorylation patterns of proteins in membrane fractions changed in response to PDB or EGF treatment. This is expected since PDB and EGF activate protein kinase C (Nishizuka, 1984)and EGF receptor kinase (see review, Carpenter, 19871, respective1 , and in turn, the activated kinases will hosphory ate their respective substrates which may urther activate a network of rotein kinases or phosphatases. Interestingly, phosp orylation patterns of the proteins in PE-deficient cells do not seem to be influenced significantly by treatment with PDB or EGF. This result is in agreement with the refract0 nature of EGF binding to PDB treatment in these cel s. The results presented in this report suggest that membrane phos holipid pla s important roles in membrane-associate function. issues and organs in animals seem to have their own unique membrane hospholipid compositions which may be a resu t of adaptation for optimal activity of various tissue-specific membrane-associated functions. In cell culture, depending on culture conditions, cells (particularly of
P
P
R
7
B
i
P
f
a
LITERATURE CITED Carpenter, G. (1987) Receptors for epidermal growth factor and other polypeptide mitogens. Annu. Rev. Biochem., 56:881-1014. Car enter, G., and Cohen, S. (1990)Epidermal growth factor. J. Biol. &em., 265:7709-7712. Davis, R.J. (1988)Independent mechanisms account for the regulation by rotein kinase C of the epidermal owth factor receptor affinity anat rosine- rotein kinase activity. YBiol. Chem.,263:9462-9469. Davis, kJ., antCzech, M.P. (1985)Tumor-promotingphorbol diesters cause the phosphorylation of epidermal growth factor receptors in normal human fibroblasts at threonine-654. Roc. Natl. Acad. Sci. U.S.A., 82:1974-1978. Friedman, B., Frackelton, A.R., Jr., Ross, A.H., Conners, J.M., Fujiki, H., Sugimura, T., and Rosner, M.R. (1984) Tumor promoters block tyrosine-specific phosphorylation of the epidermal growth factor receptor. Proc. Natl. Acad. Sci. U.S.A., 8133034-3038. Haigler, H.T. (1983) Receptor-mediated endocytosis of epidermal growth factor. Methods Enzymol., 98:283-290. Hunter, T., Ling, N., and Cooper, J.A. (1984) Protein kinase-C phosphorylation of the EGF-receptor at a threonine residue close to the cytoplasmicface of the plasma membrane. Nature, 311 :480433. Iwashita, S., and Fox, C.F. (1984)Epidermal growth factor and potent phorbol tumor promoters induce epidermal growth factor receptor phosphorylation in a similar but distinctly different manner in human epidermoid carcinoma A431 cells. J. Biol. Chem., 259:25592567. Kano-Sueoka, T., and King, D. (1987a) Role of ethanolamine and phosphatidylethanolaminein proliferation of mammar epithelial cells. In: Growth and Differentiation of Mammary Epitielial Cells in Culture. J. Enami and R.G. Ham, eds. Japan Scientific Societies Press, Tokyo, pp. 39-58. Kano-Sueoka, T., and King, D.M. (198713)Phosphatidylethanolamine biosynthesis in rat mammary carcinoma cells that re uire and do not require ethanolamine for proliferation. J. Bid. Chem., 2626074-6081. Kano-Sueoka, T., and King, D.M. (1988) Effects of phosphatidylethanolamine and phosphatidylcholine in membrane phos holipid in binding of phorbol ester in rat mammary carcinoma c e h . Cancer Res., 48:1528-1532. Kano-Sueoka, T., Errick, J.E., King, D., and Walsh, L. (1983) Phosphatidylethanolamine synthesis in ethanolamine-responsive and -nonresponsive cells in culture. J. Cell. Physiol., 117109-115. McCaffrey, P.G., Friedman, B.A., and Rosner, M.R.(1984)Diacylglycerol modulates bindin and phosphorylation of the epidermal growth factor receptor. Biol. Chem., 25912502-12507. Nishizuka, Y. (1984)The role of protein kinase C in cell surface signal transduction and tumour promotion. Nature, 308:693-697. Scatchard, H. (1949) The attractions of proteins for small molecules and ions. Ann. N.Y. Acad. Sci., 51:660-672. Schacterle, H.R., and Pollack, R.L. (1973)A simplified method for the quantitative analysis of small amounts of protein in biological material. Anal. Biochem., 51:654-655. Schlessinger,J. (1986) Allosteric re ulation of the epidermal growth factor receptor kinase. J. Cell Biof., 103:2067-2072. Schroeder, F., Goetz,I., and Roberts, E. (1984) Age-related alterations in cultured human fibroblast membrane structure and function. Mech. Ageing Dev., 25365389.
5.
Binding of Epidermal Growth Factor to Its Receptor Is Affected by Membrane Phospholipid Environment TAMIKO KANO-SUEOKA,* DAVID M. KING, HAROLD A. FISK, AND STEFANIE J. KLUG Department of Molecular, Cellular and Developmental Biology, University of Colorado,
Boulder, Colorado 80309-0347
Cells of epithelial origin generally require ethanolamine to grow in culture; when these cells are grown without ethanolamine, the phosphatidylethanolamine content of their membrane phospholipid becomes 1/2 to 1/3 of the normal amount, and growth stops. We have hypothesized that growth ceases because the
phospholipid environment becomes unsuitable for membrane-associated function. Using ethanolamine-requiring rat mammary cells, we have investigated the possible effect of phosphatidylethanolamine deficiency on the binding characteristics of epidermal growth factor. Apparent dissociation constant for the high-affinity sites in cells having normal membrane phospholipid was 1.7 x 1 O-'OM, whereas that of phosphatidylethanolamine-deficient cells was 2.7 X 10-'OM: the difference was small, but significant. Pretreatment with phorbol ester caused the loss of high-affinity sites in cells having normal membrane, whereas binding characteristics of epidermal growth factor became refractory to the pretreatment in phosphatidylethanolamine-deficient cells. In addition, the rate of internalization of bound epidermal growth factor in phosphatidylethanolamine-deficientcells was about 1/4 of normal cells. Further, whether cells had normal or phosphatidylethanolamine-deficient membranes seemed to affect the phosphorylation patterns of membrane proteins in response to epidermal growth factor or phorbol ester. These results suggest that membrane phospholipid environment affects the activity of the epidermal growth factor receptor. Cells of epithelial origin generally require ethanolamine (Etn) (or phosphatidylethanolamine [PEN to grow in culture, whereas fibroblasts or neurocells proliferate well in the absence of Etn (Kano-Sueoka and King, 1987a,b). The composition of membrane (plasma and total) phospholipid of Etn-requiring cells is similar to that of in vivo tissue when they are grown in the presence of Etn (Kano-Sueoka et al., 1983).However, in the absence of supplemental Etn, the content of PE in their membrane becomes 1/2 to 1/3 of the normal amount and that of phosphatidylcholine (PC)increases by 30% (Kano-Sueoka et al., 1983; Kano-Sueoka and King, 198713).In the absence of Etn, the rate of growth slows down as PE content is reduced (and PC content increases), and growth eventually stops (Kano-Sueoka et al., 1983). It is probable that proliferation ceases because the phospholipid environment becomes unsuitable for membrane-associated functions. Using a typically Etn-requiring rat mammary carcinoma cell line, 64-24, we previously demonstrated that the apparent dissociation constant (Kd) of phorbol 12,13-dibutyrate (PDB) was significantly higher in cells having PEdeficient membrane, and the growth-promoting activity of PDB, which can readily be demonstrated in cells having normal membrane phospholipid, was not observed in PE-deficient cells (Kano-Sueoka and King, 1988).These results suggested that the transduction of 0 1990 WILEY-LISS, INC.
the PDB signal subsequent to PDB binding to the cell membrane was profoundly affected by the nature of the plasma membrane. In the present study we have used the 64-24 cell line to examine the effect of PE-deficienc on the behavior of the epidermal growth factor (E F) receptor. The EGF receptor is a well-characterized cell surface receptor which consists of a single transmembrane polypeptide and possesses tyrosine kinase activity in its cytoplasmic domain (see reviews, Schlessinger, 1986; Carpenter, 1987; Carpenter and Cohen, 1990). Binding characteristics of EGF to its receptor as well as the autophosphorylation and tyrosine kinase activity of the receptor are modulated by EGF and other molecules including protein kinase C. Generally about 10%of the receptors exist in a high-affinity state which is thou ht to be responsible for the mitogenic activity of E F, whereas 90% of the receptors are in a low-affinity state. Phosphorylation of the receptor by rotein kinase C abolishes the Ipumor-promoting can activate rotein binding of EG to its receptor by phosphorylating the EGF receptor molecule
E
8
5
Received June 22, 1990; accepted September 7,1990. 'To whom reprint requestsicorrespondence should be addressed.
544
KANO-SUEOKA ET AL.
at specific sites (Friedman et al., 1984; Hunter et al., tal ones. Proteins were estimated by the method of 1984; Iwashita and Fox, 1984; McCaffrey et al., 1985; Schacterle and Pollack (1973), using the wells that were used to estimate nonspecific bindings. The ScatDavis and Czeck, 1985; Davis, 1988). The results of the present study indicate that PE- chard anal sis was carried out using a computer prodeficiency in cellular membranes affects not only the gram devecped by Dr. Noboru Sueoka (University of binding characteristics of EGF, but also the rate of Colorado, Boulder, CO). internalization of the bound EGF. Further, phosphoryIn vitro radiolabeling of membrane proteins lation atterns of membrane proteins in cells having Cells having normal and PE-deficient membrane norma or PE-deficient phospholipid as examined by in vitro radiolabeling show that these cells are likely to phospholipids were pre ared using 100 x 15 mm culture dishes as describe previously (Kano-Sueoka and respond differently to EGF or PDB. King, 1988). Pretreatment of these cells with PDB was MATERIALS AND METHODS also carried out as described above. For the pretreat1251-mouseEGF (specific activity, 105-171 pcilpg) ment with EGF, cells were incubated with 20 ng/ml was obtained from Du Pont-New England Nuclear EGF for 10 min at 37°C. Cells were washed twice with (Boston, MA) and Collaborative Research (Bedford, ice-cold PBS, scra ed, and suspended in buffer containMA), and mouse EGF, receptor grade, was obtained ing 0.01 M Tris- C1, pH 7.8, 2 mM MgClZ,20 pgiml from Collaborative Research and Upstate Biotechnol- aprotinin, and 2 mM PMSF. Cells were then ru tured o y (Lake Placid, NY). PDB was purchased from Sigma at 4°C by means of pressure homogenization at 00 psi with Nz gas in a Parr cell disruption bomb. Nuclei were (it. Louis, MO). removed by centrifuging at 10006' for 10 min and the Cells and cell culture membrane fraction was pelleted at 100,OOOgfor 60 min. A rat mammary carcinoma cell line 64-24 has been The pellet was sus ended in 20 mM HEPES buffer, pH maintained in Dulbecco's modified Eagles' medium 7.4, and recentri uged at 100,000 for 60 min and (DME) supplemented with 5% horse serum and 2.5% resuspended in the same buffer. T e membrane fracfetal calf serum. Cells having normal and PE-deficient tions were quickly frozen and stored at -70°C. Some membrane phospholipids were prepared for the EGF membrane preparations were made in the presence of binding studies as described previously (Kano-Sueoka 50 mM NaF and 1 mM Na3V05. Phosphorylation reactions contained 20 to 70 pg protein, 20 mM HEPES, and King, 1988). pH 7.4, 1 mM MnClZ,0.13 pg/ml BSA, 1-2 pM ATP, Binding of '251-EGF to cells and 2-5 pCi y3'P ATP. In some reactions 16.7 mM NaF The cells were plated at seeding densities of 2.5- and 250 pM Na3V05 were included. The phosphoryla3.0 x lo5 and 3.0-3.5 X lo5 cells/well (35 mm in diam- tion reaction was carried out for 5 min on ice or at room eter) for normal and PE-deficient cells, respectively, in tem erature. Phosphorylated membrane proteins were multi-well plates. One day later the cells were washed ana yzed by sodium dodec 1 sulfate gel electrophoresis twice with DME medium, and 0.2 ml DME containing using 7.5% acrylamide ge s. 15 mM N-2-Hydroxyethyl iperazine-N'-2-ethanesulfoRESULTS nic acid (HEPES), 1m /m bovine serum albumin, and EGF bindin to cells having normal and desired amounts of 126-EGFwere added in duplicates PE- eficient membranes or triplicates. For saturation binding experiments, the binding was carried out at 4°C for 90 min with entle 64-24 cells were grown in the presence or absence of shaking. To estimate nonspecific binding 1 p Im EGF 10 pM Etn for 10 to 15 enerations before the binding was added in addition to the radioactive EG . At the assays were performet Time course of binding of end of the incubation, the cells were washed with 1ml '251-EGF showed that the s ecific binding of 1251-EGF of ice-cold PBS six times, lysed with 0.2 ml 1N NaOH, to both types of cells (norma and PE-deficient) reached and the radioactivity was counted in a Beckman maximum at around 90 min and then gradually dey-counter. Determination of cell surface-bound and creased, and in PE-deficient cells the amount of EGF internalized lz5I-EGF was carried out similarly to the boundimg protein was about 70% of that bound by cells method of Haigler (1983). Briefly, the cells were incu- having a normal phos holi id composition (Data not bated with 1251-EGFat 37°C for varied lengths of time, shown). The binding o EG to these cells was further washed with ice-cold PBS five times, and then incu- studied by Scatchard analysis (Scatchard, 1949). Varbated in 0.5 m10.2 M acetic acid containing 0.5 M NaC1, ied amounts of lZ5I-EGFwere bound to the cells in midpH 2.5 for 8 min at 4°C with rotary shakin . The acetic to late-log hase under saturation conditions. Nonacidisalt solution (which contains extrace lular EGF) specific biniings were determined in the presence of was removed from the well, and the cells were rinsed excess nonradioactive EGF (1 pg/ml) and were meawith 0.5 ml of the acetic acid/salt solution which was sured separately for each concentration of radioactive combined with the original acidhalt extract. The re- EGF. The results indicated that the level of saturation maining radioactivity (intracellular EGF) was recov- of bound EGF was consistently higher (1.5- to 1.7-fold) ered by lysing the cells with 1m10.5 N NaOH at 70°C in normal cells than in PE-deficient cells (Fi .lA). The for 1 hr. For the pretreatment with PDB, 100 ng/ml of data were plotted according to the method o Scatchard PBD dissolved in dimethylsulfoxide (DMSO)was added (1949) to yield apparent dissociation constants (Kds) to the cells and incubated at 37°C for 10 min prior to the and the number of binding sites. The results are washin and addition of radioactive EGF. Control wells indicated in Figure 2A. Under our experimental conreceive DMSO in the same amount as the experimen- ditions about half of the receptors in both types of cells
P
B
#
f
P
P
P
a
7
fi
%!
Y
P P
5
B
%
MEMBRANE PHOSPHOLIPID AND EGF RECEPTOR
-
80
f
60
545
5
a
- 40
=E U
20
0 2
m
O
0
20
40
20
0
EGF
Fig. 1. Dose saturation binding of EGF to 64-24 cells having normal and PE-deficient membrane and the effect of preincubation with PDB. 64-24 cells having normal and PE-deficient membrane phospholipids were prepared by growing cells in the presence and absence of 10 pM Etn as described previously (Kano-Sueoka and King, 1987b) and also in “Materials and Methods.” The cells were incubated with varied amounts of lZ5I-EGFat 4°C for 90 min and processed as described in “Materials and Methods.” A: Dose saturation binding of EGF to
40 (nqlrnl)
normal and PE-deficient cells. B:Dose saturation binding of EGF to normal cells with and without the pretreatment of PDB. C: Dose saturation bindin of EGF to PE-deficient cells with and without the , cells having normal membrane pretreatment of 8DB. phospholipid; - - - -, cells having PE-deficient membrane phospholipid. Thick and thin lines represent cells treated and not treated with PDB, respectively.
EGF bound (frnol/rng protein)
Fig. 2. Scatchard plots of the dose saturation binding of EGF to 64-24 cells. Data presented in Figure 1A-C were plotted according to Scatchard (1949)in order to obtain apparent Kd values and the number of binding sites. A-C correspond to A-C of Fi re 1 The number of cellsimg protein for normal and PE-deficient cells were 8.0 x lo6 and 5.6 x 10P , respectively. ’ .
were high-affinity species, and the rest were lowaffinity species. Apparent Kd values for the hifhaffinity sites were 1.64 x lop1’ M and 2.86 X 10-1 M for normal and PE-deficient cells, respective1 difference in Kd values between these two cel types The was small (at most 2-fold), but was repeatedly observed as shown in Table 1. The number of the hi h-affinit binding sites seemed to be similar between t e two ce 1 types, although that of PE-deficient cells was considerably more variable (Table 1). Reliable estimation of the Kd values or the number of binding sites for the low-affinity species was difficult because of the small number of data points, as seen in Figure 2A. However, the analysis suggests that the Kd values a pear to be between 3.5 to 6 x 10-l’ M and the num er of sites seems to be several thousand per cell for both cell types from three experiments (data not shown). Effect of phorbol ester on EGF binding We have previously shown that the PE-deficiency has abnormal effects on cellular functions mediated by phorbol esters (Kano-Sueoka and King, 1988). Phorbol esters are known to cause the loss of the high-affinity binding sites of EGF receptors by activating protein
9’
f l P
l
kinase C (Friedman et al., 1984; Hunter et al., 1984; Davis and Czech, 1985). Therefore, it is possible that the mode of modulation of EGF bindin by phorbol esters may become abnormal under PE- eficient conditions. Indeed, a striking difference in the nature of the modulation of EGF binding by PDB was observed between the two types of cells when they were pretreated with 100 n /ml of PDB for 10 min followed by the binding of EGf. In the cells having normal membrane phospholipids the modulation of the EGF receptor took place as expected. Namely, in PDB-pretreated cells the level of saturation of the bound EGF was reduced to half or sometimes less than half as compared to the untreated cells (Fig. 1B). According to the Scatchard analysis the majority of the binding sites after PDB treatment had a Kd value of 3.7 x lo-’’ M with the number of sites being 4,200 siteskell (Fig. 2B). The data thus indicate that PDB treatment causes the loss of the high-affinity, but not of the low-affinity sites. When the same experiment was carried out using PE-deficient cells, entirely different results were obtained (Figs. 1C,2C). Name1 , the binding characteristics of EGF were similar w ether the cells were pretreated with PDB or not, indicating that under these
%
x
546
KANO-SUEOKAET AL. TABLE 1. Binding of EGF to 64-24 cells having normal and phosphatidylethanolamine-deficient membranes: analysis of high affinity binding sites' Cell PE-deficient Kd Binding sites M) (per cell)
Exp.
(X
1
2.86 2.50 2.78 2.71 f .112
2
3
Averages
'Deviations of S.E.M. fmm the averages were
~s.E.M.
't
I
I
50
2,700 7,000 8,100 5,900 1,700
+
PDB (ng/ml) Fig. 3. Dose-response effect of PDB on the binding of EGF to normal and PE-deficient 64-24 cells. 64-24 Cells were prepared as in the experiments described in Figures 1and 2. The cells were incubated for 10 min with varying amounts of PDB at 37°C and then lZ5I-EGF (30,000cmpiwell) was added and incubated further at 4'C for 100 min. The cells were processed as described in Figure 1.M, cells having normal membrane phospholipid; O---0, cells having PE-deficient membrane phospholipid.
conditions PDB hard1 modulated EGF binding properties. The effects of P B pretreatment on EGF binding to both normal and PE-deficient cells were reproducible. The dose-response effect of PDB on EGF binding in the two t es of cells further confirmed that the binding of GF in PE-deficient cells was not affected b PDB within the range of concentrations of PDB testeJ however, in normal cells the binding was reduced in a dose-de endent manner, and at 50 ng/ml the bindin was re uced to less than 50% of the level of untreate cells (Fig. 3). PE-deficient cells bound more EGF than normal cells at all doses of PDB. This is due to the fact that PDB treatment resulted in the loss of high-affinity sites in normal, but not in deficient cells as indicated by Scatchard analysis (Fig. 2; Table 1). These results therefore indicate that the modulation of the EGF receptor by PDB, at least with re ard to the binding affinity, does not take place in ce 1s having PE-deficient membranes.
6
P
B
i
5
(x 10-lo M)
Binding sites (per cell)
1.64 1.96 1.47 1.69 f .14
6,500 5,200 4,750 5,500 f 550
+ 2.2% for Y-intercept(boundlfree) and *
/%
I00
Normal Kd
2.6%for the slope.
Internalization of bound EGF Whether PE-deficiency in the membrane affects the rate and extent of the internalization of receptor-bound EGF was determined by measuring the amount of 1251-EGFin aciasalt-resistant (intracellular) and acid/ salt-sensitive (cell surface-bound) fractions. As shown in Figure 4,1251-EGFbound to the cell surface reached a lateau within 5 min after the addition of radioactive EP*F in both types of cells. The plateau level was 4- to 5-fold higher in cells having normal membrane phosholi id. Moreover, the rate of internalization of the Koung lZ5I-EGFwas markedly higher in normal cells than PE-deficient cells. In order to determine if PDB also affects the rate of internalization of bound EGF, the acidhalt-resistant and -sensitive fractions of the bound lZ5I-EGFafter 30 min incubation at 4°C were examined. Preincubation time with PDB was 10 min at 37°C. A ain, as in the case of saturation binding ex eriments escribed above the amount of 1251-EGFfouncfon the cell surface as well as inside of the cells was significantly reduced by phorbol ester treatment in normal cells, whereas in PE-deficient cells horbol ester did not affect the amount of 1251-EG found either on the surface or inside of the cells (Table 2). Phosphorylation of proteins in membrane fractions The results so far obtained suggest that events that take lace subsequent to EGF binding and interaction of E F receptor with other si aling pathways are influenced by the deficiency of E in the membranes. In order to test the above possibilities the effect of treatment of cells with PDB and EGF on phosphorylation of membrane proteins was examined in cells having normal and PE-deficient membrane. Post nuclear membrane fractions were prepared from these cells after treatment with either PDB or EGF and ghosporylated in vitro using y-32PATP as described in Materials and Methods." The results of PDB treatment and EGF treatment are shown in Figures 5 and 6, respectively. In normal cells the PDB treatment caused si nificant changes in banding patterns of phos horylate proteins as indicated by arrows. Intensity o the label in some bands was reproducibly suppressed as a result of the PDB-treatment and that of some was enhanced. In contrast, in cells having PE-deficient membranes the PDB treatment did not seem to change the phosphory-
cf
P
8
P
f!
d
547
MEMBRANE PHOSPHOLIPID AND EGF RECEPTOR
5-
-
Etn;
...’
4-
P
0
x
3-
C .-
0,
c
g F
2-
\
__---
PDB:
- -
+ +
+
-
I
+ -205 -116
- 97.4
-
66
-
45
______----A 1
I
20
“0
40
Time (min.)
Fig. 4. Time course of changes in the amount of surface-bound and intracellular Iz5I-EGFin normal and PE-deficient 64-24 cells. Ex r imental procedure was essentially the same as that describefin Figure 1, except that the binding was carried out at 37°C. Surfacebound and intracellular lz5I-EGF were determined as described in “Materials and Methods.” Pretreatment with PDB was carried out for 10 min at 37°C. ,cells having normal membrane;----, cells having PE-deficient membrane; A ,A , surface bound; 0, 0, intracellular. ~
TABLE 2. Effect of PDB on internalization of the EGF receptor in
cells having normal and PE-deficientmembranes
Fig. 5. Phosphorylation of isolated normal and PE-deficient membrane fractions in response to PDB treatment. 64-24 Cells with normal and PE-deficient membranes were repared as in Figure 1.Cells were treated with 100 ng/ml PDB in DhSO or with an equal amount of DMSO without PDB for 10 min at 37°C. Post-nuclear membrane fractions were prepared, and phosphorylation of membrane proteins was performed as described in “Materials and Methods.” Analysis of r t e i n s was carried out by SDS polyacrylamide gel electrophoresis ollowed by autoradiography. Differences in phosphorylationpatterns between PDB treated and non-treated cells are indicated by arrows.
cpm bound/mg protein
PDB treatment
Ethanolamine
Site
-
+
Surface Internal Surface Internal
850’ 4,886 1,521 9,489
771 4,691 786 3,650
’Average of duplicate samples. The results are reproducible
lation atterns. A similar result can be seen in the case of EG treatment (Fig. 6). Here again EGF reproducibly altered banding patterns of phos hor lated proteins in normal cells, whereas that of P -de icient cells seemed not to be affected by EGF. In these studies we were unable to clearly identify the EGF receptor protein, perhaps because 64-24 cells possess, according to our estimate, only around 10,000 EGF receptors per cell.
P
E F
DISCUSSION
Our previous studies indicated that the properties of
PDB binding to PE-deficient membranes were different from those of normal membranes, and the stimulation of cell proliferation by PDB, which can be demonstrated in cells having normal membranes, did not occur in PE-deficient cells (Kano-Sueoka and King, 1988). In the present study, binding characteristics of EGF in these two types of cells were investigated. The results suggest that EGF binding to the hi h-affinity receptor site is affected by properties of the p ospholipid bilayer. Analogous to the situation of PDB binding, the Kd
a
values were different by about two-fold between the two ty es of cells: i.e., the affinity of EGF to its receptor was a ittle weaker in PE-deficient cells. Although this difference was repeatedly observed and was significant, it is not certain whether this small difference can bring about any significant hysiological effects..However, a ronounced effect of P -deficiency was observed when ginding of EGF was examined after PDB pretreatment. In cells having normal membrane, the high-affinity binding was lost after PDB treatment, while in PEdeficient cells the nature of EGF binding was similar whether the cells were treated with PDB or not. These results thus suggest that a cascade of reactions initiated by PDB is interrupted at some point in PEdeficient cells. The most plausible explanation of the above (assumin that rotein kinase C is the major protein affected y PDB is that either 1)the activation of protein kinase C by PDB does not take place in normal fashion in PE-deficient cells or 2) the activation can take place, but the modulation of EGF receptor by protein kinase C does not occur. The present results do not indicate if either is the case. The effect of EGF and PDB treatment in these cells on the distribution of protein kinase C in subcellular fractions and the extent of phosphorylation of the serine, threonine, and tyrosine residues of EGF receptor molecules are currently under investigation. These studies will like1 yield information regarding the particular step w ich is sensitive to PE-deficiency. In addition to the binding characteristics per se, PE-deficiency affected that manner in which the bound EGF was processed. The
’I
E
g
P
x
548
Etn: EGF
KANO-SUEOKA ET AL.
+
+
+
-
I
I
+
-205
-116 97.4
-
more than two-fold (Schroeder et al., 1984).
-66
ACKNOWLEDGMENTS We thank Dr. Marsha R. Rosner (University of Chica 0, Chica 0, IL) for her valuable help at the initia stage o f t is work. This work was supported by NIH grant CA 30545.
-45
Fig. 6. Phosphorylation of isolated normal and PE-deficient membrane fractions in response to EGF treatment. 64-24 cells with normal and PE-deficient membranes were prepared as described in Figure 5. Cells were treated with or without 20 ngiml EGF for 10 min at 37°C. Post-nuclear membrane fractions were prepared, phospho lated, and analyzed as described in Figure 5. Arrows indicate ban% showing different extent of phosphorylation between EGF treated and nontreated cells.
manner in which phorbol ester affected the internalization was also consistent with the way PDB affected the binding affinity of EGF as shown in Figures 1 and 2. Our results however do not indicate whether change in fluidity or charge brought about by PEdeficiency influences the process of internalization, or if the internalization is in some way correlated with the activity of the EGF receptors. The method employed in the present studies to analyze membrane-associated proteins does not yield detailed information regarding phosphorylation of individual proteins. However, it appears to show that in cells having normal membrane phospholipids the phosphorylation patterns of proteins in membrane fractions changed in response to PDB or EGF treatment. This is expected since PDB and EGF activate protein kinase C (Nishizuka, 1984)and EGF receptor kinase (see review, Carpenter, 19871, respective1 , and in turn, the activated kinases will hosphory ate their respective substrates which may urther activate a network of rotein kinases or phosphatases. Interestingly, phosp orylation patterns of the proteins in PE-deficient cells do not seem to be influenced significantly by treatment with PDB or EGF. This result is in agreement with the refract0 nature of EGF binding to PDB treatment in these cel s. The results presented in this report suggest that membrane phos holipid pla s important roles in membrane-associate function. issues and organs in animals seem to have their own unique membrane hospholipid compositions which may be a resu t of adaptation for optimal activity of various tissue-specific membrane-associated functions. In cell culture, depending on culture conditions, cells (particularly of
P
P
R
7
B
i
P
f
a
LITERATURE CITED Carpenter, G. (1987) Receptors for epidermal growth factor and other polypeptide mitogens. Annu. Rev. Biochem., 56:881-1014. Car enter, G., and Cohen, S. (1990)Epidermal growth factor. J. Biol. &em., 265:7709-7712. Davis, R.J. (1988)Independent mechanisms account for the regulation by rotein kinase C of the epidermal owth factor receptor affinity anat rosine- rotein kinase activity. YBiol. Chem.,263:9462-9469. Davis, kJ., antCzech, M.P. (1985)Tumor-promotingphorbol diesters cause the phosphorylation of epidermal growth factor receptors in normal human fibroblasts at threonine-654. Roc. Natl. Acad. Sci. U.S.A., 82:1974-1978. Friedman, B., Frackelton, A.R., Jr., Ross, A.H., Conners, J.M., Fujiki, H., Sugimura, T., and Rosner, M.R. (1984) Tumor promoters block tyrosine-specific phosphorylation of the epidermal growth factor receptor. Proc. Natl. Acad. Sci. U.S.A., 8133034-3038. Haigler, H.T. (1983) Receptor-mediated endocytosis of epidermal growth factor. Methods Enzymol., 98:283-290. Hunter, T., Ling, N., and Cooper, J.A. (1984) Protein kinase-C phosphorylation of the EGF-receptor at a threonine residue close to the cytoplasmicface of the plasma membrane. Nature, 311 :480433. Iwashita, S., and Fox, C.F. (1984)Epidermal growth factor and potent phorbol tumor promoters induce epidermal growth factor receptor phosphorylation in a similar but distinctly different manner in human epidermoid carcinoma A431 cells. J. Biol. Chem., 259:25592567. Kano-Sueoka, T., and King, D. (1987a) Role of ethanolamine and phosphatidylethanolaminein proliferation of mammar epithelial cells. In: Growth and Differentiation of Mammary Epitielial Cells in Culture. J. Enami and R.G. Ham, eds. Japan Scientific Societies Press, Tokyo, pp. 39-58. Kano-Sueoka, T., and King, D.M. (198713)Phosphatidylethanolamine biosynthesis in rat mammary carcinoma cells that re uire and do not require ethanolamine for proliferation. J. Bid. Chem., 2626074-6081. Kano-Sueoka, T., and King, D.M. (1988) Effects of phosphatidylethanolamine and phosphatidylcholine in membrane phos holipid in binding of phorbol ester in rat mammary carcinoma c e h . Cancer Res., 48:1528-1532. Kano-Sueoka, T., Errick, J.E., King, D., and Walsh, L. (1983) Phosphatidylethanolamine synthesis in ethanolamine-responsive and -nonresponsive cells in culture. J. Cell. Physiol., 117109-115. McCaffrey, P.G., Friedman, B.A., and Rosner, M.R.(1984)Diacylglycerol modulates bindin and phosphorylation of the epidermal growth factor receptor. Biol. Chem., 25912502-12507. Nishizuka, Y. (1984)The role of protein kinase C in cell surface signal transduction and tumour promotion. Nature, 308:693-697. Scatchard, H. (1949) The attractions of proteins for small molecules and ions. Ann. N.Y. Acad. Sci., 51:660-672. Schacterle, H.R., and Pollack, R.L. (1973)A simplified method for the quantitative analysis of small amounts of protein in biological material. Anal. Biochem., 51:654-655. Schlessinger,J. (1986) Allosteric re ulation of the epidermal growth factor receptor kinase. J. Cell Biof., 103:2067-2072. Schroeder, F., Goetz,I., and Roberts, E. (1984) Age-related alterations in cultured human fibroblast membrane structure and function. Mech. Ageing Dev., 25365389.
5.
