ARCHIVES
OF BIOCHEMISTRY
Vol. 292, No. 2, February
AND
BIOPHYSICS
1, pp. 4484551992
Differential in vitro Phosphorylation of Clathrin Light Chains by the Epidermal Growth Factor ReceptorAssociated Protein Tyrosine Kinase and a ~~60”~“‘“Related Spleen Tyrosine Kinase’ Marilyn
J. Mooibroek,2 Heung-Chin
Cheng,3 and Jerry H. Wang4,5
MRC Group in Signal Transduction, Department of Medical Biochemistry, Faculty of Medicine, University of Calgary, 3330 Hospital Drive N. W., Calgary, Alberta TZN 4N1, Canada
Received July 11,1991, and in revised form September 20, 1991
The epidermal growth factor (EGF) receptor-associated protein tyrosine kinase activity has been suggested to play important roles in the EGF-enhanced, clathrincoated pit-mediated receptor internalization (W. S. Chen, C. S. Lazar, M. Peonie, R. Y. Tsien, G. N. Gill, and M. G. Rosenfeld, 1987, Nature 328, 820-823) but the kinase substrate important for this process has not been identified. This study demonstrates that the EGF receptor, partially purified from A43 1 epidermoid carcinoma cells, catalyzes the phosphorylation of one of the two clathrin light chains, clathrin light chain a (LCa). The phosphorylation activity is stimulated by EGF and immunoprecipitated by an EGF receptor monoclonal antibody. The phosphorylation occurs exclusively on tyrosine residues. Amino acid composition of the major tryptic phosphopeptide of the EGF receptor-phosphorylated LCa corresponds closely to that of residues 1 to 97 of LCa. A stoichiometry of 0.2 mol phosphate/m01 LCa was attained after 60 min at 30°C and a K, value of 1.7 PM was determined for the reaction. LCa of either neuronal or nonneuronal origin could serve as a substrate. In addition to the EGF receptor tyrosine kinase, a particulate src-related protein tyrosine kinase purified from bovine spleen (C. M. E. Litwin, H.-C. Cheng, and J. H. Wang, 1991, J. Biol. Chem. 226,2557-2566) was shown in this study to also phosphorylate the light chains. However, in contrast to the EGF receptor phosphorylation, both clathrin light chains a and b were phosphorylated by the spleen kinase, suggesting that the two tyrosine kinases have differential site specificities. Given the specificity of LCa phosphorylation by the EGF receptor, we propose that LCa phosphorylation on a tyrosine residue(s) may be important in EGF-induced receptor internalization. o NW Academic Press,
Inc.
Protein phosphorylation plays a key role in cellular signal transduction, with protein tyrosine kinases being important parts of this system (1,2). Receptor-mediated endocytosis is also believed to play an active role in signal transduction, being involved in translocation of the signal generating entities (3) and signal desensitization (4). Agonists which stimulate protein tyrosine kinase activity also greatly enhance the endocytosis of the receptor-ligand complex, suggesting that a relationship between protein tyrosine phosphorylation and endocytosis exists. An increase in receptor affinity for specific binding sites in coated vesicles may explain the ligand-induced enhancement in uptake of the EGF’ receptors (5, 6). Although the notion that the EGF receptor kinase activity is involved in ligand-induced endocytosis has received general support (5, 6), there are also reports suggesting that the kinase activity does not play a role in the receptor endocytosis, Ref. (7) for example. A possible explanation is that the protein tyrosine kinase activity plays a regulatory rather than an obligatory role in this process.
1 This work was supported by grants from the Medical Research Council of Canada. ’ Recipient of an Alberta Heritage Foundation for Medical Research studentship (1984-1988). Present address: Department of Biochemistry, University of Ottawa, Ottawa KlH 8M5, Canada. 3 Recipient of an Alberta Heritage Foundation for Medical Research Postdoctoral fellowship (19881989) and a Medical Research Council of Canada postdoctoral fellowship (198991992). 4 Alberta Heritage Foundation for Medical Research scientist. ’ To whom correspondence should be addressed. ’ Abbreviations used: LCa, clathrin light chain a; LCb, clathrin light chain b; EGF, epidermal growth factor; HPLC, high-performance liquid chromatography; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; TFA, trifluoracetic acid; DMEM, Dulbecco’s modified Eagle’s medium; cdc, cell division cycle; CaM, calmodulin; PDGF, platelet-derived growth factor.
448 All
0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, Inc. rights of reproduction in any form reserved.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
The participation of cla.thrin-coated vesicles in the receptor-mediated endocytosis pathway is well known (reviewed in (8)). Coated vlesicles are characterized by a polyhedral protein coat which can dissociate into basic structural units known as triskelions (9). Each triskelion is composed of three clathrin heavy chains and three clathrin light chains of which there are two types: LCa and LCb (10). Although clathrin is considered to be a structural protein, the light chains have been suggested to play regulatory roles in coated vesicle function (10,ll). In vitro studies have revealed that both clathrin light chains contain calmodulin (11) and Ca2+ binding sites (12) and are required for tlne activity of the coated vesicle uncoating ATPase activity (13). LCb has been shown to be phosphorylated in uitro by a casein kinase II type activity (14). As well, LCb and to a much lesser extent LCa are phosphorylated in uiuo on serine residues but in this case the kinase(s) respons#ible is unknown (15). Given the potential regulation of endocytosis by protein tyrosine kinases and the regulatory characteristics of the clathrin light chains, the ability of the light chains to serve as substrates for protein tyrosine kinases in vitro was examined. Two tyrosine kinases were used, the EGF receptor-associated protein tyrosine kinase and a nonreceptor-associated protein tyrosine kinase purified from bovine spleen which is a m’ember of the src-family protein tyrosine kinases (16). Both kinases were found to phosphorylate the light chains but with different specificities. The EGF receptor kinase only phosphorylated LCa while the spleen tyrosine kinase phosphorylated both LCa and LCb. This phosphorylation pattern was seen with both neuronal and nonneuronal light chains. MATERIALS
AND
METHODS
Materials Fresh bovine tissues were obtained from a local abattoir and processed immediately. [ya’P]ATP (10 Ci/nnmol) was from ICN Biomedicals Canada Ltd. (St. Laurent, Quebec). ,4garose-bound wheat germ agglutinin was obtained from Vector Laboratories (Burlingame, CA). Reagents for SDS-polyacrylamide gel electrophoresis were from Bio-Rad Canada Ltd. (Mississauga, Ontario). Protein standards for SDS-polyacrylamide gel electrophoresis were from Pharmacia (Canada) Inc. (Dorval, Quebec). EGF and Hepes were from Boehringer-Mannheim Canada (Dorval, Quebec). All other chemicals were obtained from commercial sources (Sigma and Fisher) and were of reagent grade.
Methods Protein purification. Bovine brain and liver clathrin light chains were purified according to the procedure described previously (12). Separation of the light chains was achieved by chromatography on a phenyl column (Pierce PH.300 Aquapore; 4.6 X 220 mm) using a Waters Model 510 HPLC system. The column was preequilibrated with 0.1% trifluoroacetic acid (TFA) in water. The light chains were eluted at a flow rate of 1 ml/min with a linear gradient (l%/min) of aqueous acetonitrile containing 0.1% TFA. The EGF receptor was partially purified from A431 epidermoid carcinoma cells through wheat germ agarose chromatography essentially as described by Akiyama et al. (17). The A431 cells were grown in 750ml flasks in DMEM supplemented with 10% fetal calf serum at 5% CO,
LIGHT
CHAIN
BY TYROSINE
KINASES
449
and 95% air. Extract from the membrane fraction of the cell lysate was applied to the lectin column; the bound fraction from the lectin column was eluted from the column with N-acetylglucosamine and then dialyzed 0.2 mM phenylagainst 40 mM Hepes (pH 7.2), 1 mM dithiothreitol, methylsulfonyl fluoride, 10% glycerol, and 0.05% Triton X-100 and stored at -7O’C. Immunoprecipitation of the receptor was performed as described in (18) using an anti-EGF receptor monoclonal antibody kindly provided by Dr. M. Hollenberg (Department of Pharmacology, University of Calgary). The spleen protein tyrosine kinase was purified from bovine spleen to close to homogeneity and found to interact with a monospecific antibody prepared against a peptide containing the autophosphorylation site sequence of pp60c”” and other src-family protein kinases (16). Phosphmylation assays. The EGF receptor kinase-catalyzed protein phosphorylation reaction was carried out according to the procedure of Fava and Cohen (19). The partially purified receptor preparation (1 pg protein) was preincubated at room temperature for 30 min in the absence or presence of 200 mM EGF. The incubation mixture, final volume, 50 ~1, routinely contained 20 mM Hepes (pH 7.5), 20 mM MgCl,, 2 mM MnCl*, 1 mM CaCl*, and 10 yM Na,VO, unless otherwise indicated in the figure legends. Following preincubation, the tubes were placed on ice and the light chains (2 ag) added. The phosphorylation reaction was initiated by the addition of [y-32P]ATP (0.1 mM, 1000 cpm/pmol). The reaction was terminated after 20 min by adding 50 ~1 of twofold concentrated Laemmli (20) electrophoresis sample buffer and boiling for 3 min. The samples were analyzed on 10% SDS-PAGE gels using the buffer system of Laemmli (20) followed by autoradiography. Phosphate incorporation into the light chains was determined by excising the bands from the dried gel and counting in a scintillation counter. Phosphory lation of the clathrin light chains by the spleen protein tyrosine kinase was carried out under the same buffer conditions described for the EGF receptor-catalyzed reaction. The reaction was carried out at 30°C for 1 h unless otherwise indicated. Phosphoamino acid analysis was performed on the phosphorylated proteins after purification by HPLC (21).
RESULTS
Although it has been suggested that the protein tyrosine kinase activity associated with the EGF receptor plays an important role in the endocytosis of the receptor (5, 6, 22-26), the putative protein substrate(s) involved in the process is not known. Since clathrin light chains appear to have regulatory functions in receptor-mediated endocytosis, we have examined the possible phosphorylation of clathrin light chains by the EGF receptor. The receptor was partially purified from A431 epidermoid carcinoma cells over wheat germ agglutinin agarose. Analysis by SDS-PAGE and autoradiography of the receptor preparation, incubated in the presence of [Y-~~P]ATP, revealed a single radioactive band with an apparent molecular weight of about 170,000 (Fig. 1). EGF increased the phosphate incorporated into this band. Even upon prolonged exposure, no other radiolabeled bands were observed, indicating the absence of endogenous EGF receptor kinase substrates in this preparation. The bovine brain clathrin light chain preparations contained on the average about 30 and 70% of LCa and LCb, respectively. When these preparations were subjected to phosphorylation conditions in the presence of the EGF receptor preparation, LCa was significantly phosphorylated, with this phosphorylation being increased threefold by EGF treatment (Fig. lB, lanes 3 and 4). In the absence of added kinase, no radiolabeled bands
450
MOOIBROEK,
CHENG,
C EGF-R
FIG. 1. Phosphorylation of clathrin light chains by the EGF receptorassociated protein kinase. The phosphorylation reactions were conducted on ice for 20 min and the samples analyzed on 10% SDS gels and autoradiographed. (A) Autoradiograph of the autophosphorylation of the receptor preparation (2 pg) in the presence of 200 nM EGF. (B) Phosphorylation of the clathrin light chains (2 pg) by the EGF receptor (0.5 pg) performed in the absence and presence of 200 nM EGF. Lanes 1 and 2, Coomassie blue staining. Lanes 3 and 4, corresponding autoradiographs. (C) Phosphoamino acid analysis of LCa phosphorylated by the EGF receptor in the presence of 200 nM EGF.
were detected (not shown). The phosphate into LCa was restricted to tyrosine residues by phosphoamino acid analysis (Fig. 1C). tyrosine specificity and EGF stimulation
incorporated as determined The observed of LCa phos-
Retention
AND
WANG
phorylation strongly suggested that it was catalyzed by the EGF receptor rather than a contaminating protein kinase. To further substantiate this suggestion, attempts were made to immunoprecipitate the LCa phosphorylation activity by EGF receptor antibody. An aliquot of the partially purified EGF receptor (4 pg) was phosphorylated in the presence of [T-~‘P]ATP and then subjected to immunoprecipitation by an EGF receptor monoclonal antibody, EGFRl, in the presence of Staph A. Analysis of the supernatant and the immunoprecipitate by SDSPAGE and autoradiography showed that the 170-kDa radioactive protein was quantitatively precipitated. Although the removal of the 170-kDa protein band was correlated with the removal of the LCa phosphorylation activity from the supernatant, the enzyme activity was, however, not recovered in the immunoprecipitate, suggesting that the enzyme in the immunoprecipitate was inactive. In a control experiment where the antibody was omitted, both the 170-kDa radioactive band and the LCa kinase activity were found in the supernatant. While the EGF receptor kinase appeared to be specific for LCa, with some of the light chain preparations a low level of phosphorylation of a slightly lower molecular weight band was noted (not shown). This lower molecular weight band tended to migrate at the lower margin of the LCb band, thus suggesting that it may be derived from LCa by proteolysis. To determine conclusively if LCb was in fact a substrate for the EGF receptor kinase, the two light chains were separated by HPLC on a phenyl column (Fig. 2), and purified LCb was tested for its ability to
Time (min)
FIG. 2. Elution profile of bovine brain clathrin light chains separated by HPLC and phosphorylation of purified LCb by the EGF receptor. Clathrin light chains were eluted from a phenyl column (Pierce PH-300 Aquapore) with a linear gradient of acetonitrile containing 0.1% TFA (- - -) as described under Materials and Methods. Fractions (1 ml) were collected and analyzed on 10% SDS gels. Inset, phosphorylation of purified LCb by the EGF receptor. Lanes 1 and 2, Coomassie blue staining of 0.6 and 2 pg LCb, respectively. Lanes 3 and 4, corresponding autoradiographs.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
LIGHT
CHAIN 7679
3x105-
NH2
- 4!i.O00 2x105-
/TP-ER - 3’1.000
FIG. 3. HPLC phosphopeptide map of tryptic fragments of EGF receptor-phosphorylated bovine brain LCa. Bovine brain LCa purified by HPLC was phosphorylated by the EGF receptor in the presence of [y3ZP]ATP under standard conditions. The phosphorylated sample was subjected to tryptic digestion (substrate protein:protease = 2O:l (w/w)) in 0.1 M ammonium bicarbonate buffer, pH 7.8, for 16 h at 37°C. The HPLC mapping of the hydrolysate was carried out using a Waters reverse-phase (PBondapak C18) HPLC column. The column was developed with a linear gradient of 1% increase in solvent B per minute at a flow rate of 0.75 ml/min. Solvent system: the aqueous phase solvent, 10 mM ammonium acetate in H,O, pH 6.8, and the organic phase solvent, 70% acetonitrile and 30% the aqueous solvent. Radioactivity in collected fractions was determined by Cerenkov counting. Inset shows autoradiogram of SDS-PAGE analysis of the peak radioactive fraction.
serve as an EGF receptor substrate. In the water/acetonitrile buffer system used, LCb (retention time = 43 min) eluted ahead of LCa (retention time = 45 min). The identities of the HPLC-purified LCa and LCb were confirmed by SDS-PAGE and amino acid composition analyses (data not shown). Two concentrations of LCb were tested, one approximating the am.ount of LCa normally present in the reaction mixture (0.6 /*M) (Fig. 2, inset, lanes 1 and 3) and one corresponding to the total amount of light chains normally present in the reaction mixture (2 PM) (Fig. 3, inset, lanes 2 and 4). No phosphorylation of LCb was noted in either instance, suggesting that only LCa is a substrate for the EGF receptor. To further confirm that the peptide phosphorylated by EGF receptor in the clathrin light chain preparation is indeed LCa rather than a contaminating protein of similar molecular weight, the phosphorylated LCa was characterized in terms of its tryptic phosphopeptides. A sample of bovine brain LCa purified by HPLC was phosphorylated by the EGF receptor and then subjected to exhaustive tryptic digestion. Figure 3 shows that the tryptic hydrolysate contained one major phosphopeptide, which migrated on SDS-PAGE gel as a lo-kDa peptide (Fig. 3, inset). This tryptic phosphopeptide was purified by HPLC (Fig. 3). Its amino acid composition was determined to be in close agreement with1 that of amino acid residues 1 to 97 of LCa. As schematically shown in Fig. 4, peptide LCa l-97 is among the tryptic peptides predicted on the basis of the bovine brain LCa sequence (29) and is the longest. In addition to substantiating the notion that LCa,
-
BY TYROSINE 69
LCa(l-97)
142
-
451
KINASES 164172183
1 Iih
“11
~~~~~~ tc-
FIG. 4. Schematic representation of potential tryptic cleavage in LCa. The potential tryptic cleavage sites were shown by vertical and the tyrosine residues were labeled as Y. The tyrosine residues numbered according to the amino acid sequence of bovine brain (29).
sites lines were LCa
rather than a contaminating protein, is the substrate of the EGF receptor, the result shows that the phosphorylation by the receptor kinase is highly specific in terms of the target tyrosine residues. The phosphorylation of LCa by the EGF receptor was examined under a number of divalent cation conditions. The highest activity was noted in the presence of 20 mM Mg2+ and 2 mM Mn2+. In the presence of Mg2+ or Mn2+ alone the activity was decreased 60 and 70%, respectively. The addition of 2 mM Ca2+ to the reaction resulted in a modest decrease (15%) in the observed activity. The nature of this inhibition is uncertain; however, as the clathrin light chains have been shown to be calcium binding proteins (12), this inhibition may be due to Ca2+ binding to the substrate. Under all the conditions tested, the phosphorylation was essentially specific for LCa and was stimulated by EGF by at least threefold. Figure 5 shows the time course of EGF receptor-catalyzed phosphorylation of bovine brain clathrin light chains
20 60 g ‘C f! ifi b 8
/’
H g
/
7’ /’
/ /
15.
/
al ;ii Jz
min
2.0-
,’
0.5-
I /.y’ (1’1 /
t
0’ ,/’ /’
I 7:
/’
,P”
/
l.O-
/’
/’ l
,,,/” y
1
10
I 20
1 30
I 40
I 50
I 60
Time (mid FIG. 5. Time course of phosphorylation of bovine brain clathrin light chains by the EGF receptor. Purified bovine brain clathrin light chains (2 fig) were phorphorylated by the EGF receptor (1 pg) in the presence of 200 nM of EGF for the indicated times either on ice (0) or at 30°C (m). Inset: Lane 1, Coomassie blue staining LC sample after incubation for 20 min on ice; Lanes 2 and 3, autoradiograph of phosphorylated light chains after 20 and 60 min on ice.
452
MOOIBROEK,
CHENG,
at two different temperatures. The reactions were not complete after 60 min when the stoichiometries of phosphate incorporation were 0.22 and 0.15 mol phosphate per mole LCa at 30 and 4°C. Even after extended periods of time, phosphorylation of LCb was not detected (Fig. 5, inset), suggesting that the lack of LCb phosphorylation seen under standard reaction conditions (i.e., 20 min incubation) was not due to a slow reaction rate of LCb phosphorylation. Kinetic studies were carried out by phosphorylating the light chains at concentrations ranging from 0.69 to 6.9 PM of LCa. After SDS-PAGE, the relative concentration of LCa in the mixture was calculated from a densitometric scan of the gel. The LCa band was then cut from the dried gel and counted to determine the extent of LCa phosphorylation. An apparent Km value of 1.7 ELM was determined for the EGF-stimulated reaction. Even at the higher light chain concentrations the phosphorylation of LCb was still not observed (Fig. 6, inset), arguing that the differential phosphorylation of the light chains by the EGF receptor kinase cannot be explained by differences in kinase affinity for the light chains. Previous studies on the phosphorylation of the clathrin light chains have shown that LCb is preferentially phosphorylated by a protein serine kinase, casein kinase II (13-15). This observation has been used to suggest that casein kinase II may play regulatory roles in endocytosis and that LCb is specifically involved in this protein phosphorylation-mediated endocytosis. Thus, the observed specific phosphorylation of LCa by the EGF receptor was of particular interest. To determine if this differential
AND
WANG
phosphorylation was specific for the EGF receptor kinase or a common property of protein tyrosine kinases, many of which have shown to share substrates (27,28), the ability of a non-receptor-associated pp60”.““-related spleen protein tyrosine kinase (16) to phosphorylate the light chains was investigated. In contrast to the results observed with the EGF receptor, it was found that the spleen protein tyrosine kinase phosphorylated the two clathrin light chains equally well (Fig. 7, inset). All the phosphate was on tyrosine residues as determined by phosphoamino acid analysis (Fig. 7B); an average stoichiometry of 0.6 mol phosphate per mole light chain was obtained after 120 min (Fig. 7A). Mg2+ (20 mM) and Mn2+ (2 mM) were equally effective in supporting the reaction while the addition of Ca2+ (up to 1 mM) had no effect. At a light chain concentration of 30 PM, the enzyme was still far from saturated (result not shown), suggesting that the Km for this reaction is at least an order of magnitude higher than that observed with the EGF receptor. Examination of the amino acid sequence of neuronal light chains (29), which had been used so far in this study, revealed that the neuronal-specific insertion sequence in these light chains was relatively rich in tyrosine residues. Three of the seven tyrosine residues present in neuronal LCa and two of the three in neuronal LCb are contained within their respective insertion sequences. This raised the possibility that the tyrosine phosphorylations observed herein may be a unique property of neuronal light chains. To investigate this, the ability of nonneuronal light chains to serve as substrates for these tyrosine kinases was examined. When subjected to phosphorylation
.OQ
11 [S]
(PM)-’
FIG. 6. Kinetic parameters for the phosphorylation of bovine brain clathrin light chains by the EGF receptor. Increasing concentrations of bovine brain clathrin light chains (0.69-6.9 PM) were phosphorylated by the EGF receptor under standard phosphorylation conditions as described under Materials and Methods. The data are plotted as a double reciprocal plot (r = 0.9938). Inset, Coomassie blue staining (C) and autoradiograph (A) of phosphorylated light chains at the concentration of 6.9 pM.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
LIGHT
CHAIN
BY TYROSINE
KINASES
453
I
-PTK
10
Y
I 20
I 40
I 60
20
30 45
I 80
I 100
60
85 120150
I 120
min
I 140
TIME (mid FIG. 7. Phosphorylation of bovine brain clathrin light chains by the pp60c-“‘c-related spleen tyrosine kinase. (A) Bovine brain clathrin light chains (2 pg) were phosphorylatecl by the spleen protein tyrosine kinase for the indicated times (see Materials and Methods). Phosphate incorporation was expressed as mole phosphate per mole of light chains (average molecular weight: 25.9 kDa). Inset, autoradiograph of the phosphorylated light chains at the indicated time points. (B) Phosphoamino acid analysis of the light chains phosphorylated by the spleen kinase for 120 min.
conditions, bovine liver light chains were found to serve as substrates for both the EGF receptor and the spleen tyrosine kinases (Fig. 8). As was noted with the brain light chains, the EGF receptor phosphorylation was restricted to LCa and was stimulated by addition of EGF while the spleen tyrosine kinase recognized both liver light chains. With both kinases,, the extent of phosphorylation of the neuronal and nonneuronal light chains was similar, suggesting that the tyrosine residue(s) phosphorylated was not localized in the neuronal insert region. DISCUSSION
The potential involvement of protein tyrosine kinases in the endocytosis of a number of receptors, including the insulin, PDGF, and EGF receptors, prompted an investigation into whether coated vesicle associated proteins may act as substrates for tyrosine kinases. The results of this paper indicate that the clathrin light chains are in vitro substrates for at least two protein tyrosine kinases but that each kinase appea.rs to have a unique specificity. The spleen protein tyrosine kinase, which belongs to the src-gene family of protein tyrosine kinases, phosphorylated both clathrin light chains to a similar extent. This was noted with light chains from either neuronal or nonneuronal sources. As LCb from nonneuronal sources contains only one tyrosine, Tyr-81(29), and both neuronal and nonneuronal light chains are phosphorylated to a similar extent, this residue is believed to be the site of phosphorylation for this kinase on LCb from both sources. The analogous tyrosine in LCa (residue 89) is proposed to be the major site of phosphorylation in this light chain based on the equivalent level of phosphorylation seen be-
tween LCb and LCa. Of interest, the hydroxyl oxygen of tyrosine 89 of LCa has been proposed as one of the chelating atoms in the clathrin light chain calcium binding loop (30), raising the possibility that calcium binding may be modulated by phosphorylation. However, under the reaction conditions used in this study (20 mM Mg2+ and 2 mM Mn2+), calcium had no effect on the extent of phosphorylation of either light chain by the spleen protein tyrosine kinase. In contrast to the spleen tyrosine kinase, the EGF receptor kinase specifically phosphorylated LCa. The lack of LCb phosphorylation was shown not to be due to a slower reaction rate or to a much lower affinity of the receptor for LCb compared to LCa. These observations indicate that the two protein tyrosine kinases have distinct intrinsic in vitro substrate specificities. This is in agreement with the finding that the spleen tyrosine kinase, but not the EGF receptor, showed an unusually high activity toward a synthetic peptide derived from the cell division cycle controlling element, ~34”~“~ (14). The characteristics of phosphorylation of LCa by the EGF receptor share many similarities with those of other identified substrates of this kinase. The stoichiometry observed for LCa phosphorylation (0.2 mol phosphate/ mol LCa after 60 min at 30°C) is comparable to the stoichiometries reported for other cytoskeletal proteins (27, 28). Using a similar solubilized EGF receptor preparation, the incorporation of 0.25,0.1,1.5, and 0.28 mol phosphate per mole of fodrin, tubulin, MAP2, and tau, respectively, after 2 h at 25°C was observed (28). The stoichiometry reported for the EGF receptor substrate, ~36, was 0.15 mol/mol (19). The K, of the reaction (1.7 FM) is also
454
MOOIBROEK,
CHENG,
6
A EGF Receptor
Spleen Tyrosine Kinase
Kinase
-EGF-
q SPTK
EGF
+
-
+
-
FIG. 8. Phosphorylation of bovine liver clathrin light chains by the EGF receptor and the spleen tyrosine kinase. Phosphorylation of brain (1 pg) and liver (1 pg) light chains by the EGF receptor and spleen tyrosine kinases was carried out as outlined for the respective kinases under Materials and Methods. (A) Phosphorylation of brain and liver clathrin light chains by the EGF receptor in the presence and absence of 200 nM EGF. (B) Phosphorylation of brain and liver clathrin light chains by the spleen protein tyrosine kinase.
comparable to the values reported for the cytoskeletal proteins (0.1-3.0 PM (28)) and p36 (0.8 PM (19)). In addition, the phosphorylation appears to be highly specific, as only one major phosphopeptide was obtained in the tryptic hydrolysates from phosphorylated LCa. The physiological significance of the light chain phosphorylation by the EGF receptor remains uncertain; however, the selectivity of the reaction for LCa, the ability of both neuronal and nonneuronal LCa’s to serve as substrates, and the relatively low K,,, value of the reaction make the physiological importance of this reaction an intriguing possibility. Several studies suggest that the EGF receptor tyrosine kinase activity is required for ligand-induced high affinity endocytosis (5, 6, 26). Receptors lacking kinase activity due to mutations or to inhibition of the kinase activity by microinjection of anti-phosphotyrosine antibodies were found either to not undergo internalization (22-24, 26) or to be internalized at a rate reflective of constitutive receptor internalization (6). The kinase activity did not appear to be required for intracellular receptor sorting (6, 26). Other studies, however, observed little difference between the time courses of uptake of wild-type and kinase negative EGF receptors (37, 31) but found that, whereas the wild-type receptor was found to be degraded, the kinase negative receptor recycled, suggesting that the receptor kinase activity was important for receptor sorting.
AND
WANG
The lack of kinase-dependent EGF receptor endocytosis observed in these studies has been suggested to be due to the combined use of experimental conditions which bias the specific internalization rates to lower values and the measurement of net internalization rates versus specific internalization rates (6). While the biochemical mechanisms underlying the regulation of receptor internalization by the intrinsic tyrosine kinase activity are unknown, it appears that the phosphorylation of an exogenous protein(s) in addition to autophosphorylation is required for rapid internalization. Deletion of the major autophosphorylation sites from the carboxyl terminus of the EGF receptor had little effect on EGF-induced endocytosis (24,32). As well, though endocytosis was inhibited by microinjection of anti-phosphotyrosine antibodies, the autophosphorylation reaction was not affected, again suggesting that the phosphorylation of an exogenous protein is important for internalization. As ligand-induced internalization has been proposed to result from an increase in receptor affinity for specific binding sites in coated pits (5), it suggests that coated-pit-associated proteins may be targets for the receptor tyrosine kinase. Given the specificity of LCa phosphorylation by the EGF receptor observed in this study, it may be suggested that LCa is one such target. LCa and LCb are similar in many of their functional properties. Both are capable of binding Ca2+ (12) and calmodulin (11) and their interactions with the clathrin heavy chain are interchangeable (33). The Ca2+ binding sites of the light chains have been tentatively localized and suggested to be conserved in different biological species (30, 34). Ca2+ binding to the proteins, has been suggested to cause a change in the proteins’ conformation resulting in a change in affinity toward the uncoating ATPase (30). The two light chains are highly homologous in primary structure, with over 60% of the amino acids being identical, and in certain regions of the proteins, the structural conservation is much stronger (27). Such conserved structures appear to form the molecular basis of the common functionality of the two proteins. On the other hand, since LCa and LCb are almost always coexpressed in the cell, it suggests that each of the clathrin light chains also has unique functions. The only differential functional property noted for LCa and LCb to date is that LCb but not LCa serves as a substrate for the coated-vesicle-associated casein kinase II (14, 15). The molecular basis of this specific LCb phosphorylation has been elucidated (35) and shown to be preserved in different animal species. The observed phosphorylation of LCa but not LCb by the EGF receptor-associated tyrosine kinase provides another example of a differential phosphorylation of the two light chains. The observation that the two clathrin light chains are differentially phosphorylated by different protein kinases supports the suggestion that different signals in the regulation of endocytosis may be mediated by distinct light chains.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
ACKNOWLEDGMENT The excellent technical assistance of Jackie Andrea and Faith Hwang is gratefully acknowledged.
LIGHT
CHAIN
BY TYROSINE
KINASES
455
17. Akiyama, T., Kadooka, T., and Ogawara, H. (1985) Biochem. Biophys. Res. Commun. 131, 442-448. 18. Pallen, C. J., Valentine, K. A., Wang, J. H., and Hollenberg, (1985) Biochemistry 24,4727-4730.
M. D.
19. Fava, R. A., and Cohen, S. (1984) J. Biol. Chem. 259, 2636-2645.
REFERENCES 1. Krebs, E. G., and Beavo, J. A. (1979) Annu. Reu. Biochem. 48,923959. 2. White, M. F., and Kahn, C. R. (1986) in The Enzymes (Boyer, P. D., and Krebs, E. G., Eds.), Vol. 17, pp. 248-310, Academic Press, Orlando. 3. Khan, M. N., Baquiran, G., Brule, C., Burgess, J., Foster, B., Bergeron, J. J. M., and Posner, El. I. (1989) J. Biol. Chem. 264,12,93112,940.
4. Beguinot, L., Lyall, R. M., Willingham, M. C., and Pastan, I. (1984) Proc. Natl. Acad. Sci. USA 61, 2384-2388. 5. Lund, K. A., Opresko, K. L., Starbuck, C., Walsh, B. J., and Wiley, H. S. (1990) J. Biol. Chem. 265, 20,517-20,523.
20. Laemmli, U. K. (1970) Nature 227, 680-685. 21. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) in Methods in Enzymology (Corbin, J. D., and Hardman, J. G., Eds.), Vol. 99, pp. 387-402, Academic Press, San Diego. 22. Chen, W. S., Lazar, C. S., Poenie, M., Tsien, R. Y., Gill, G. N., and Rosenfeld, M. G. (1987) Nature 328, 820-823. 23. Chen, W. S., Lazer, C. S., Lund, K. A., Walsh, J. B., Chang, C. P., Walton, G. M., Der, C. J., Wiley, H. S., Gill, G. N., and Rosenfeld, M. G. (1989) Cell 59, 33-43. 24. Glenney, J. R., Jr., Chen, W. S., Lazar, C. S., Walton, G. M., Zokas, L. M., Rosenfeld, M. G., and Gill, G. N. (1988) Cell 52, 675-684. 25. Sorkin, A., Waters, C., Overholser, J. Biol. Chem. 266, 8355-8362.
K. A., and Carpenter,
G. (1991)
6. Wiley, H. S., Herbst, J. J., Walsh, B. J., Lauffenburger,
26. Helin, K., and Beguinot, L. (1991) J. Biol. Chem. 266,8363-8368. 27. Kadowaki, T., Koyasu, S., Nishida, E., Tobe, K., Izumi, T., Takaku, F., Sakai, H., Yahara, I., and Kasuga, M. (1987) J. Biol. Chem. 262,
7. Honegger, A. M., Dull, T. J., Felder, S., Van Obberghen, E., Bellot,
7342-7350. 28. Akiyama, T., Kadowaki,
D. A., Rosenfold, M. G., and Gill, G. IV. (1991) J. Biol. Chem. 266, 11,08311,094. F., Szapary, D., Schmidt, A., Ullrich,
A., and Schlessinger, J. (1987)
CelZ51,199-209. 8. Wileman, T., Harding, C., and Stahl, P. (1985) Biochem. J. 232, 1-14.
9. Ungewickell, E., and Branto.n, D. (1981) Nature 289, 420-422. 10. Brodsky, F. M., and Parham, P. (1983) J. Mol. Biol. 167,197-204. 11. Lisanti, M. P., Sapiro, L. S.. Moskowitz, N., Hua, E. L., Puszkin, S., and Schook, W. (1982) E,ur. J. Biochem. 125,463-470. 12. Mooibroek,
M. J., Michiel,
I>. F., and Wang, J. H. (1987) J. Biol.
Chem.262,25-28. 13. DeLuca-Flaherty, C., McKa,y, D. B., Parham, P., and Hill, B. L. (1990) Cell 62, 875-887.
14. Bar-Zvi, D., and Branton, D. (1986) J. Biol. Chem. 261,9614-9621. 15. Bar-Zvi, D., Mosley, S. T., and Branton, D. (1988) J. Biol. Chem. 263,4408-4415. 16. Litwin, C. M. E., Cheng, H.-C., and Wang, J. H. (1991) J. Biol. Chem. 266,2557-2566.
T., Nishida, E., Kadooka, T., Ogawara, H., Fukami, Y., Sakai, H., Takaku, F., and Kasuga, M. (1986) J. Biol. Chem. 261,14,797-14,803.
29. Jackson, A. P., and Parham, P. (1988) J. Biol. Chem. 262, 16,68816,695. 30. Nathke, I., Hill, B. L., Parham, P., and Brodsky, Biol. Chem. 265, 18,621-18,627. 31. Felder, S., Miller, K., Moehren, G., Ullrich, Hopkins, C. R. (1990) Cell 61,623-634.
F. M. (1990) J.
A., Schlessinger, J., and
32. Livneh, E., Prywes, R., Kashles, O., Reiss, N., Sasson, I., Mory, Y., Ullrich, A., and Schlessinger, J. (1986) J. Biol. Chem. 261, 12,49012,497. 33. Winkler,
F. K., and Stanley, K. K. (1983) EMBO J. 2, 1393-1400.
34. Silveira,
L. A., Wong, D. H., Masiarz, (1990) J. Cell Biol. 111, 1437-1449.
35. Hill, B. L., Drickamer, K., Brodsky, J. Biol. Chem. 263,5499-5501.
F. R., and Schekman,
R.
F. M., and Parham, P. (1988)
OF BIOCHEMISTRY
Vol. 292, No. 2, February
AND
BIOPHYSICS
1, pp. 4484551992
Differential in vitro Phosphorylation of Clathrin Light Chains by the Epidermal Growth Factor ReceptorAssociated Protein Tyrosine Kinase and a ~~60”~“‘“Related Spleen Tyrosine Kinase’ Marilyn
J. Mooibroek,2 Heung-Chin
Cheng,3 and Jerry H. Wang4,5
MRC Group in Signal Transduction, Department of Medical Biochemistry, Faculty of Medicine, University of Calgary, 3330 Hospital Drive N. W., Calgary, Alberta TZN 4N1, Canada
Received July 11,1991, and in revised form September 20, 1991
The epidermal growth factor (EGF) receptor-associated protein tyrosine kinase activity has been suggested to play important roles in the EGF-enhanced, clathrincoated pit-mediated receptor internalization (W. S. Chen, C. S. Lazar, M. Peonie, R. Y. Tsien, G. N. Gill, and M. G. Rosenfeld, 1987, Nature 328, 820-823) but the kinase substrate important for this process has not been identified. This study demonstrates that the EGF receptor, partially purified from A43 1 epidermoid carcinoma cells, catalyzes the phosphorylation of one of the two clathrin light chains, clathrin light chain a (LCa). The phosphorylation activity is stimulated by EGF and immunoprecipitated by an EGF receptor monoclonal antibody. The phosphorylation occurs exclusively on tyrosine residues. Amino acid composition of the major tryptic phosphopeptide of the EGF receptor-phosphorylated LCa corresponds closely to that of residues 1 to 97 of LCa. A stoichiometry of 0.2 mol phosphate/m01 LCa was attained after 60 min at 30°C and a K, value of 1.7 PM was determined for the reaction. LCa of either neuronal or nonneuronal origin could serve as a substrate. In addition to the EGF receptor tyrosine kinase, a particulate src-related protein tyrosine kinase purified from bovine spleen (C. M. E. Litwin, H.-C. Cheng, and J. H. Wang, 1991, J. Biol. Chem. 226,2557-2566) was shown in this study to also phosphorylate the light chains. However, in contrast to the EGF receptor phosphorylation, both clathrin light chains a and b were phosphorylated by the spleen kinase, suggesting that the two tyrosine kinases have differential site specificities. Given the specificity of LCa phosphorylation by the EGF receptor, we propose that LCa phosphorylation on a tyrosine residue(s) may be important in EGF-induced receptor internalization. o NW Academic Press,
Inc.
Protein phosphorylation plays a key role in cellular signal transduction, with protein tyrosine kinases being important parts of this system (1,2). Receptor-mediated endocytosis is also believed to play an active role in signal transduction, being involved in translocation of the signal generating entities (3) and signal desensitization (4). Agonists which stimulate protein tyrosine kinase activity also greatly enhance the endocytosis of the receptor-ligand complex, suggesting that a relationship between protein tyrosine phosphorylation and endocytosis exists. An increase in receptor affinity for specific binding sites in coated vesicles may explain the ligand-induced enhancement in uptake of the EGF’ receptors (5, 6). Although the notion that the EGF receptor kinase activity is involved in ligand-induced endocytosis has received general support (5, 6), there are also reports suggesting that the kinase activity does not play a role in the receptor endocytosis, Ref. (7) for example. A possible explanation is that the protein tyrosine kinase activity plays a regulatory rather than an obligatory role in this process.
1 This work was supported by grants from the Medical Research Council of Canada. ’ Recipient of an Alberta Heritage Foundation for Medical Research studentship (1984-1988). Present address: Department of Biochemistry, University of Ottawa, Ottawa KlH 8M5, Canada. 3 Recipient of an Alberta Heritage Foundation for Medical Research Postdoctoral fellowship (19881989) and a Medical Research Council of Canada postdoctoral fellowship (198991992). 4 Alberta Heritage Foundation for Medical Research scientist. ’ To whom correspondence should be addressed. ’ Abbreviations used: LCa, clathrin light chain a; LCb, clathrin light chain b; EGF, epidermal growth factor; HPLC, high-performance liquid chromatography; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; TFA, trifluoracetic acid; DMEM, Dulbecco’s modified Eagle’s medium; cdc, cell division cycle; CaM, calmodulin; PDGF, platelet-derived growth factor.
448 All
0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, Inc. rights of reproduction in any form reserved.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
The participation of cla.thrin-coated vesicles in the receptor-mediated endocytosis pathway is well known (reviewed in (8)). Coated vlesicles are characterized by a polyhedral protein coat which can dissociate into basic structural units known as triskelions (9). Each triskelion is composed of three clathrin heavy chains and three clathrin light chains of which there are two types: LCa and LCb (10). Although clathrin is considered to be a structural protein, the light chains have been suggested to play regulatory roles in coated vesicle function (10,ll). In vitro studies have revealed that both clathrin light chains contain calmodulin (11) and Ca2+ binding sites (12) and are required for tlne activity of the coated vesicle uncoating ATPase activity (13). LCb has been shown to be phosphorylated in uitro by a casein kinase II type activity (14). As well, LCb and to a much lesser extent LCa are phosphorylated in uiuo on serine residues but in this case the kinase(s) respons#ible is unknown (15). Given the potential regulation of endocytosis by protein tyrosine kinases and the regulatory characteristics of the clathrin light chains, the ability of the light chains to serve as substrates for protein tyrosine kinases in vitro was examined. Two tyrosine kinases were used, the EGF receptor-associated protein tyrosine kinase and a nonreceptor-associated protein tyrosine kinase purified from bovine spleen which is a m’ember of the src-family protein tyrosine kinases (16). Both kinases were found to phosphorylate the light chains but with different specificities. The EGF receptor kinase only phosphorylated LCa while the spleen tyrosine kinase phosphorylated both LCa and LCb. This phosphorylation pattern was seen with both neuronal and nonneuronal light chains. MATERIALS
AND
METHODS
Materials Fresh bovine tissues were obtained from a local abattoir and processed immediately. [ya’P]ATP (10 Ci/nnmol) was from ICN Biomedicals Canada Ltd. (St. Laurent, Quebec). ,4garose-bound wheat germ agglutinin was obtained from Vector Laboratories (Burlingame, CA). Reagents for SDS-polyacrylamide gel electrophoresis were from Bio-Rad Canada Ltd. (Mississauga, Ontario). Protein standards for SDS-polyacrylamide gel electrophoresis were from Pharmacia (Canada) Inc. (Dorval, Quebec). EGF and Hepes were from Boehringer-Mannheim Canada (Dorval, Quebec). All other chemicals were obtained from commercial sources (Sigma and Fisher) and were of reagent grade.
Methods Protein purification. Bovine brain and liver clathrin light chains were purified according to the procedure described previously (12). Separation of the light chains was achieved by chromatography on a phenyl column (Pierce PH.300 Aquapore; 4.6 X 220 mm) using a Waters Model 510 HPLC system. The column was preequilibrated with 0.1% trifluoroacetic acid (TFA) in water. The light chains were eluted at a flow rate of 1 ml/min with a linear gradient (l%/min) of aqueous acetonitrile containing 0.1% TFA. The EGF receptor was partially purified from A431 epidermoid carcinoma cells through wheat germ agarose chromatography essentially as described by Akiyama et al. (17). The A431 cells were grown in 750ml flasks in DMEM supplemented with 10% fetal calf serum at 5% CO,
LIGHT
CHAIN
BY TYROSINE
KINASES
449
and 95% air. Extract from the membrane fraction of the cell lysate was applied to the lectin column; the bound fraction from the lectin column was eluted from the column with N-acetylglucosamine and then dialyzed 0.2 mM phenylagainst 40 mM Hepes (pH 7.2), 1 mM dithiothreitol, methylsulfonyl fluoride, 10% glycerol, and 0.05% Triton X-100 and stored at -7O’C. Immunoprecipitation of the receptor was performed as described in (18) using an anti-EGF receptor monoclonal antibody kindly provided by Dr. M. Hollenberg (Department of Pharmacology, University of Calgary). The spleen protein tyrosine kinase was purified from bovine spleen to close to homogeneity and found to interact with a monospecific antibody prepared against a peptide containing the autophosphorylation site sequence of pp60c”” and other src-family protein kinases (16). Phosphmylation assays. The EGF receptor kinase-catalyzed protein phosphorylation reaction was carried out according to the procedure of Fava and Cohen (19). The partially purified receptor preparation (1 pg protein) was preincubated at room temperature for 30 min in the absence or presence of 200 mM EGF. The incubation mixture, final volume, 50 ~1, routinely contained 20 mM Hepes (pH 7.5), 20 mM MgCl,, 2 mM MnCl*, 1 mM CaCl*, and 10 yM Na,VO, unless otherwise indicated in the figure legends. Following preincubation, the tubes were placed on ice and the light chains (2 ag) added. The phosphorylation reaction was initiated by the addition of [y-32P]ATP (0.1 mM, 1000 cpm/pmol). The reaction was terminated after 20 min by adding 50 ~1 of twofold concentrated Laemmli (20) electrophoresis sample buffer and boiling for 3 min. The samples were analyzed on 10% SDS-PAGE gels using the buffer system of Laemmli (20) followed by autoradiography. Phosphate incorporation into the light chains was determined by excising the bands from the dried gel and counting in a scintillation counter. Phosphory lation of the clathrin light chains by the spleen protein tyrosine kinase was carried out under the same buffer conditions described for the EGF receptor-catalyzed reaction. The reaction was carried out at 30°C for 1 h unless otherwise indicated. Phosphoamino acid analysis was performed on the phosphorylated proteins after purification by HPLC (21).
RESULTS
Although it has been suggested that the protein tyrosine kinase activity associated with the EGF receptor plays an important role in the endocytosis of the receptor (5, 6, 22-26), the putative protein substrate(s) involved in the process is not known. Since clathrin light chains appear to have regulatory functions in receptor-mediated endocytosis, we have examined the possible phosphorylation of clathrin light chains by the EGF receptor. The receptor was partially purified from A431 epidermoid carcinoma cells over wheat germ agglutinin agarose. Analysis by SDS-PAGE and autoradiography of the receptor preparation, incubated in the presence of [Y-~~P]ATP, revealed a single radioactive band with an apparent molecular weight of about 170,000 (Fig. 1). EGF increased the phosphate incorporated into this band. Even upon prolonged exposure, no other radiolabeled bands were observed, indicating the absence of endogenous EGF receptor kinase substrates in this preparation. The bovine brain clathrin light chain preparations contained on the average about 30 and 70% of LCa and LCb, respectively. When these preparations were subjected to phosphorylation conditions in the presence of the EGF receptor preparation, LCa was significantly phosphorylated, with this phosphorylation being increased threefold by EGF treatment (Fig. lB, lanes 3 and 4). In the absence of added kinase, no radiolabeled bands
450
MOOIBROEK,
CHENG,
C EGF-R
FIG. 1. Phosphorylation of clathrin light chains by the EGF receptorassociated protein kinase. The phosphorylation reactions were conducted on ice for 20 min and the samples analyzed on 10% SDS gels and autoradiographed. (A) Autoradiograph of the autophosphorylation of the receptor preparation (2 pg) in the presence of 200 nM EGF. (B) Phosphorylation of the clathrin light chains (2 pg) by the EGF receptor (0.5 pg) performed in the absence and presence of 200 nM EGF. Lanes 1 and 2, Coomassie blue staining. Lanes 3 and 4, corresponding autoradiographs. (C) Phosphoamino acid analysis of LCa phosphorylated by the EGF receptor in the presence of 200 nM EGF.
were detected (not shown). The phosphate into LCa was restricted to tyrosine residues by phosphoamino acid analysis (Fig. 1C). tyrosine specificity and EGF stimulation
incorporated as determined The observed of LCa phos-
Retention
AND
WANG
phorylation strongly suggested that it was catalyzed by the EGF receptor rather than a contaminating protein kinase. To further substantiate this suggestion, attempts were made to immunoprecipitate the LCa phosphorylation activity by EGF receptor antibody. An aliquot of the partially purified EGF receptor (4 pg) was phosphorylated in the presence of [T-~‘P]ATP and then subjected to immunoprecipitation by an EGF receptor monoclonal antibody, EGFRl, in the presence of Staph A. Analysis of the supernatant and the immunoprecipitate by SDSPAGE and autoradiography showed that the 170-kDa radioactive protein was quantitatively precipitated. Although the removal of the 170-kDa protein band was correlated with the removal of the LCa phosphorylation activity from the supernatant, the enzyme activity was, however, not recovered in the immunoprecipitate, suggesting that the enzyme in the immunoprecipitate was inactive. In a control experiment where the antibody was omitted, both the 170-kDa radioactive band and the LCa kinase activity were found in the supernatant. While the EGF receptor kinase appeared to be specific for LCa, with some of the light chain preparations a low level of phosphorylation of a slightly lower molecular weight band was noted (not shown). This lower molecular weight band tended to migrate at the lower margin of the LCb band, thus suggesting that it may be derived from LCa by proteolysis. To determine conclusively if LCb was in fact a substrate for the EGF receptor kinase, the two light chains were separated by HPLC on a phenyl column (Fig. 2), and purified LCb was tested for its ability to
Time (min)
FIG. 2. Elution profile of bovine brain clathrin light chains separated by HPLC and phosphorylation of purified LCb by the EGF receptor. Clathrin light chains were eluted from a phenyl column (Pierce PH-300 Aquapore) with a linear gradient of acetonitrile containing 0.1% TFA (- - -) as described under Materials and Methods. Fractions (1 ml) were collected and analyzed on 10% SDS gels. Inset, phosphorylation of purified LCb by the EGF receptor. Lanes 1 and 2, Coomassie blue staining of 0.6 and 2 pg LCb, respectively. Lanes 3 and 4, corresponding autoradiographs.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
LIGHT
CHAIN 7679
3x105-
NH2
- 4!i.O00 2x105-
/TP-ER - 3’1.000
FIG. 3. HPLC phosphopeptide map of tryptic fragments of EGF receptor-phosphorylated bovine brain LCa. Bovine brain LCa purified by HPLC was phosphorylated by the EGF receptor in the presence of [y3ZP]ATP under standard conditions. The phosphorylated sample was subjected to tryptic digestion (substrate protein:protease = 2O:l (w/w)) in 0.1 M ammonium bicarbonate buffer, pH 7.8, for 16 h at 37°C. The HPLC mapping of the hydrolysate was carried out using a Waters reverse-phase (PBondapak C18) HPLC column. The column was developed with a linear gradient of 1% increase in solvent B per minute at a flow rate of 0.75 ml/min. Solvent system: the aqueous phase solvent, 10 mM ammonium acetate in H,O, pH 6.8, and the organic phase solvent, 70% acetonitrile and 30% the aqueous solvent. Radioactivity in collected fractions was determined by Cerenkov counting. Inset shows autoradiogram of SDS-PAGE analysis of the peak radioactive fraction.
serve as an EGF receptor substrate. In the water/acetonitrile buffer system used, LCb (retention time = 43 min) eluted ahead of LCa (retention time = 45 min). The identities of the HPLC-purified LCa and LCb were confirmed by SDS-PAGE and amino acid composition analyses (data not shown). Two concentrations of LCb were tested, one approximating the am.ount of LCa normally present in the reaction mixture (0.6 /*M) (Fig. 2, inset, lanes 1 and 3) and one corresponding to the total amount of light chains normally present in the reaction mixture (2 PM) (Fig. 3, inset, lanes 2 and 4). No phosphorylation of LCb was noted in either instance, suggesting that only LCa is a substrate for the EGF receptor. To further confirm that the peptide phosphorylated by EGF receptor in the clathrin light chain preparation is indeed LCa rather than a contaminating protein of similar molecular weight, the phosphorylated LCa was characterized in terms of its tryptic phosphopeptides. A sample of bovine brain LCa purified by HPLC was phosphorylated by the EGF receptor and then subjected to exhaustive tryptic digestion. Figure 3 shows that the tryptic hydrolysate contained one major phosphopeptide, which migrated on SDS-PAGE gel as a lo-kDa peptide (Fig. 3, inset). This tryptic phosphopeptide was purified by HPLC (Fig. 3). Its amino acid composition was determined to be in close agreement with1 that of amino acid residues 1 to 97 of LCa. As schematically shown in Fig. 4, peptide LCa l-97 is among the tryptic peptides predicted on the basis of the bovine brain LCa sequence (29) and is the longest. In addition to substantiating the notion that LCa,
-
BY TYROSINE 69
LCa(l-97)
142
-
451
KINASES 164172183
1 Iih
“11
~~~~~~ tc-
FIG. 4. Schematic representation of potential tryptic cleavage in LCa. The potential tryptic cleavage sites were shown by vertical and the tyrosine residues were labeled as Y. The tyrosine residues numbered according to the amino acid sequence of bovine brain (29).
sites lines were LCa
rather than a contaminating protein, is the substrate of the EGF receptor, the result shows that the phosphorylation by the receptor kinase is highly specific in terms of the target tyrosine residues. The phosphorylation of LCa by the EGF receptor was examined under a number of divalent cation conditions. The highest activity was noted in the presence of 20 mM Mg2+ and 2 mM Mn2+. In the presence of Mg2+ or Mn2+ alone the activity was decreased 60 and 70%, respectively. The addition of 2 mM Ca2+ to the reaction resulted in a modest decrease (15%) in the observed activity. The nature of this inhibition is uncertain; however, as the clathrin light chains have been shown to be calcium binding proteins (12), this inhibition may be due to Ca2+ binding to the substrate. Under all the conditions tested, the phosphorylation was essentially specific for LCa and was stimulated by EGF by at least threefold. Figure 5 shows the time course of EGF receptor-catalyzed phosphorylation of bovine brain clathrin light chains
20 60 g ‘C f! ifi b 8
/’
H g
/
7’ /’
/ /
15.
/
al ;ii Jz
min
2.0-
,’
0.5-
I /.y’ (1’1 /
t
0’ ,/’ /’
I 7:
/’
,P”
/
l.O-
/’
/’ l
,,,/” y
1
10
I 20
1 30
I 40
I 50
I 60
Time (mid FIG. 5. Time course of phosphorylation of bovine brain clathrin light chains by the EGF receptor. Purified bovine brain clathrin light chains (2 fig) were phorphorylated by the EGF receptor (1 pg) in the presence of 200 nM of EGF for the indicated times either on ice (0) or at 30°C (m). Inset: Lane 1, Coomassie blue staining LC sample after incubation for 20 min on ice; Lanes 2 and 3, autoradiograph of phosphorylated light chains after 20 and 60 min on ice.
452
MOOIBROEK,
CHENG,
at two different temperatures. The reactions were not complete after 60 min when the stoichiometries of phosphate incorporation were 0.22 and 0.15 mol phosphate per mole LCa at 30 and 4°C. Even after extended periods of time, phosphorylation of LCb was not detected (Fig. 5, inset), suggesting that the lack of LCb phosphorylation seen under standard reaction conditions (i.e., 20 min incubation) was not due to a slow reaction rate of LCb phosphorylation. Kinetic studies were carried out by phosphorylating the light chains at concentrations ranging from 0.69 to 6.9 PM of LCa. After SDS-PAGE, the relative concentration of LCa in the mixture was calculated from a densitometric scan of the gel. The LCa band was then cut from the dried gel and counted to determine the extent of LCa phosphorylation. An apparent Km value of 1.7 ELM was determined for the EGF-stimulated reaction. Even at the higher light chain concentrations the phosphorylation of LCb was still not observed (Fig. 6, inset), arguing that the differential phosphorylation of the light chains by the EGF receptor kinase cannot be explained by differences in kinase affinity for the light chains. Previous studies on the phosphorylation of the clathrin light chains have shown that LCb is preferentially phosphorylated by a protein serine kinase, casein kinase II (13-15). This observation has been used to suggest that casein kinase II may play regulatory roles in endocytosis and that LCb is specifically involved in this protein phosphorylation-mediated endocytosis. Thus, the observed specific phosphorylation of LCa by the EGF receptor was of particular interest. To determine if this differential
AND
WANG
phosphorylation was specific for the EGF receptor kinase or a common property of protein tyrosine kinases, many of which have shown to share substrates (27,28), the ability of a non-receptor-associated pp60”.““-related spleen protein tyrosine kinase (16) to phosphorylate the light chains was investigated. In contrast to the results observed with the EGF receptor, it was found that the spleen protein tyrosine kinase phosphorylated the two clathrin light chains equally well (Fig. 7, inset). All the phosphate was on tyrosine residues as determined by phosphoamino acid analysis (Fig. 7B); an average stoichiometry of 0.6 mol phosphate per mole light chain was obtained after 120 min (Fig. 7A). Mg2+ (20 mM) and Mn2+ (2 mM) were equally effective in supporting the reaction while the addition of Ca2+ (up to 1 mM) had no effect. At a light chain concentration of 30 PM, the enzyme was still far from saturated (result not shown), suggesting that the Km for this reaction is at least an order of magnitude higher than that observed with the EGF receptor. Examination of the amino acid sequence of neuronal light chains (29), which had been used so far in this study, revealed that the neuronal-specific insertion sequence in these light chains was relatively rich in tyrosine residues. Three of the seven tyrosine residues present in neuronal LCa and two of the three in neuronal LCb are contained within their respective insertion sequences. This raised the possibility that the tyrosine phosphorylations observed herein may be a unique property of neuronal light chains. To investigate this, the ability of nonneuronal light chains to serve as substrates for these tyrosine kinases was examined. When subjected to phosphorylation
.OQ
11 [S]
(PM)-’
FIG. 6. Kinetic parameters for the phosphorylation of bovine brain clathrin light chains by the EGF receptor. Increasing concentrations of bovine brain clathrin light chains (0.69-6.9 PM) were phosphorylated by the EGF receptor under standard phosphorylation conditions as described under Materials and Methods. The data are plotted as a double reciprocal plot (r = 0.9938). Inset, Coomassie blue staining (C) and autoradiograph (A) of phosphorylated light chains at the concentration of 6.9 pM.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
LIGHT
CHAIN
BY TYROSINE
KINASES
453
I
-PTK
10
Y
I 20
I 40
I 60
20
30 45
I 80
I 100
60
85 120150
I 120
min
I 140
TIME (mid FIG. 7. Phosphorylation of bovine brain clathrin light chains by the pp60c-“‘c-related spleen tyrosine kinase. (A) Bovine brain clathrin light chains (2 pg) were phosphorylatecl by the spleen protein tyrosine kinase for the indicated times (see Materials and Methods). Phosphate incorporation was expressed as mole phosphate per mole of light chains (average molecular weight: 25.9 kDa). Inset, autoradiograph of the phosphorylated light chains at the indicated time points. (B) Phosphoamino acid analysis of the light chains phosphorylated by the spleen kinase for 120 min.
conditions, bovine liver light chains were found to serve as substrates for both the EGF receptor and the spleen tyrosine kinases (Fig. 8). As was noted with the brain light chains, the EGF receptor phosphorylation was restricted to LCa and was stimulated by addition of EGF while the spleen tyrosine kinase recognized both liver light chains. With both kinases,, the extent of phosphorylation of the neuronal and nonneuronal light chains was similar, suggesting that the tyrosine residue(s) phosphorylated was not localized in the neuronal insert region. DISCUSSION
The potential involvement of protein tyrosine kinases in the endocytosis of a number of receptors, including the insulin, PDGF, and EGF receptors, prompted an investigation into whether coated vesicle associated proteins may act as substrates for tyrosine kinases. The results of this paper indicate that the clathrin light chains are in vitro substrates for at least two protein tyrosine kinases but that each kinase appea.rs to have a unique specificity. The spleen protein tyrosine kinase, which belongs to the src-gene family of protein tyrosine kinases, phosphorylated both clathrin light chains to a similar extent. This was noted with light chains from either neuronal or nonneuronal sources. As LCb from nonneuronal sources contains only one tyrosine, Tyr-81(29), and both neuronal and nonneuronal light chains are phosphorylated to a similar extent, this residue is believed to be the site of phosphorylation for this kinase on LCb from both sources. The analogous tyrosine in LCa (residue 89) is proposed to be the major site of phosphorylation in this light chain based on the equivalent level of phosphorylation seen be-
tween LCb and LCa. Of interest, the hydroxyl oxygen of tyrosine 89 of LCa has been proposed as one of the chelating atoms in the clathrin light chain calcium binding loop (30), raising the possibility that calcium binding may be modulated by phosphorylation. However, under the reaction conditions used in this study (20 mM Mg2+ and 2 mM Mn2+), calcium had no effect on the extent of phosphorylation of either light chain by the spleen protein tyrosine kinase. In contrast to the spleen tyrosine kinase, the EGF receptor kinase specifically phosphorylated LCa. The lack of LCb phosphorylation was shown not to be due to a slower reaction rate or to a much lower affinity of the receptor for LCb compared to LCa. These observations indicate that the two protein tyrosine kinases have distinct intrinsic in vitro substrate specificities. This is in agreement with the finding that the spleen tyrosine kinase, but not the EGF receptor, showed an unusually high activity toward a synthetic peptide derived from the cell division cycle controlling element, ~34”~“~ (14). The characteristics of phosphorylation of LCa by the EGF receptor share many similarities with those of other identified substrates of this kinase. The stoichiometry observed for LCa phosphorylation (0.2 mol phosphate/ mol LCa after 60 min at 30°C) is comparable to the stoichiometries reported for other cytoskeletal proteins (27, 28). Using a similar solubilized EGF receptor preparation, the incorporation of 0.25,0.1,1.5, and 0.28 mol phosphate per mole of fodrin, tubulin, MAP2, and tau, respectively, after 2 h at 25°C was observed (28). The stoichiometry reported for the EGF receptor substrate, ~36, was 0.15 mol/mol (19). The K, of the reaction (1.7 FM) is also
454
MOOIBROEK,
CHENG,
6
A EGF Receptor
Spleen Tyrosine Kinase
Kinase
-EGF-
q SPTK
EGF
+
-
+
-
FIG. 8. Phosphorylation of bovine liver clathrin light chains by the EGF receptor and the spleen tyrosine kinase. Phosphorylation of brain (1 pg) and liver (1 pg) light chains by the EGF receptor and spleen tyrosine kinases was carried out as outlined for the respective kinases under Materials and Methods. (A) Phosphorylation of brain and liver clathrin light chains by the EGF receptor in the presence and absence of 200 nM EGF. (B) Phosphorylation of brain and liver clathrin light chains by the spleen protein tyrosine kinase.
comparable to the values reported for the cytoskeletal proteins (0.1-3.0 PM (28)) and p36 (0.8 PM (19)). In addition, the phosphorylation appears to be highly specific, as only one major phosphopeptide was obtained in the tryptic hydrolysates from phosphorylated LCa. The physiological significance of the light chain phosphorylation by the EGF receptor remains uncertain; however, the selectivity of the reaction for LCa, the ability of both neuronal and nonneuronal LCa’s to serve as substrates, and the relatively low K,,, value of the reaction make the physiological importance of this reaction an intriguing possibility. Several studies suggest that the EGF receptor tyrosine kinase activity is required for ligand-induced high affinity endocytosis (5, 6, 26). Receptors lacking kinase activity due to mutations or to inhibition of the kinase activity by microinjection of anti-phosphotyrosine antibodies were found either to not undergo internalization (22-24, 26) or to be internalized at a rate reflective of constitutive receptor internalization (6). The kinase activity did not appear to be required for intracellular receptor sorting (6, 26). Other studies, however, observed little difference between the time courses of uptake of wild-type and kinase negative EGF receptors (37, 31) but found that, whereas the wild-type receptor was found to be degraded, the kinase negative receptor recycled, suggesting that the receptor kinase activity was important for receptor sorting.
AND
WANG
The lack of kinase-dependent EGF receptor endocytosis observed in these studies has been suggested to be due to the combined use of experimental conditions which bias the specific internalization rates to lower values and the measurement of net internalization rates versus specific internalization rates (6). While the biochemical mechanisms underlying the regulation of receptor internalization by the intrinsic tyrosine kinase activity are unknown, it appears that the phosphorylation of an exogenous protein(s) in addition to autophosphorylation is required for rapid internalization. Deletion of the major autophosphorylation sites from the carboxyl terminus of the EGF receptor had little effect on EGF-induced endocytosis (24,32). As well, though endocytosis was inhibited by microinjection of anti-phosphotyrosine antibodies, the autophosphorylation reaction was not affected, again suggesting that the phosphorylation of an exogenous protein is important for internalization. As ligand-induced internalization has been proposed to result from an increase in receptor affinity for specific binding sites in coated pits (5), it suggests that coated-pit-associated proteins may be targets for the receptor tyrosine kinase. Given the specificity of LCa phosphorylation by the EGF receptor observed in this study, it may be suggested that LCa is one such target. LCa and LCb are similar in many of their functional properties. Both are capable of binding Ca2+ (12) and calmodulin (11) and their interactions with the clathrin heavy chain are interchangeable (33). The Ca2+ binding sites of the light chains have been tentatively localized and suggested to be conserved in different biological species (30, 34). Ca2+ binding to the proteins, has been suggested to cause a change in the proteins’ conformation resulting in a change in affinity toward the uncoating ATPase (30). The two light chains are highly homologous in primary structure, with over 60% of the amino acids being identical, and in certain regions of the proteins, the structural conservation is much stronger (27). Such conserved structures appear to form the molecular basis of the common functionality of the two proteins. On the other hand, since LCa and LCb are almost always coexpressed in the cell, it suggests that each of the clathrin light chains also has unique functions. The only differential functional property noted for LCa and LCb to date is that LCb but not LCa serves as a substrate for the coated-vesicle-associated casein kinase II (14, 15). The molecular basis of this specific LCb phosphorylation has been elucidated (35) and shown to be preserved in different animal species. The observed phosphorylation of LCa but not LCb by the EGF receptor-associated tyrosine kinase provides another example of a differential phosphorylation of the two light chains. The observation that the two clathrin light chains are differentially phosphorylated by different protein kinases supports the suggestion that different signals in the regulation of endocytosis may be mediated by distinct light chains.
DIFFERENTIAL
PHOSPHORYLATION
OF CLATHRIN
ACKNOWLEDGMENT The excellent technical assistance of Jackie Andrea and Faith Hwang is gratefully acknowledged.
LIGHT
CHAIN
BY TYROSINE
KINASES
455
17. Akiyama, T., Kadooka, T., and Ogawara, H. (1985) Biochem. Biophys. Res. Commun. 131, 442-448. 18. Pallen, C. J., Valentine, K. A., Wang, J. H., and Hollenberg, (1985) Biochemistry 24,4727-4730.
M. D.
19. Fava, R. A., and Cohen, S. (1984) J. Biol. Chem. 259, 2636-2645.
REFERENCES 1. Krebs, E. G., and Beavo, J. A. (1979) Annu. Reu. Biochem. 48,923959. 2. White, M. F., and Kahn, C. R. (1986) in The Enzymes (Boyer, P. D., and Krebs, E. G., Eds.), Vol. 17, pp. 248-310, Academic Press, Orlando. 3. Khan, M. N., Baquiran, G., Brule, C., Burgess, J., Foster, B., Bergeron, J. J. M., and Posner, El. I. (1989) J. Biol. Chem. 264,12,93112,940.
4. Beguinot, L., Lyall, R. M., Willingham, M. C., and Pastan, I. (1984) Proc. Natl. Acad. Sci. USA 61, 2384-2388. 5. Lund, K. A., Opresko, K. L., Starbuck, C., Walsh, B. J., and Wiley, H. S. (1990) J. Biol. Chem. 265, 20,517-20,523.
20. Laemmli, U. K. (1970) Nature 227, 680-685. 21. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) in Methods in Enzymology (Corbin, J. D., and Hardman, J. G., Eds.), Vol. 99, pp. 387-402, Academic Press, San Diego. 22. Chen, W. S., Lazar, C. S., Poenie, M., Tsien, R. Y., Gill, G. N., and Rosenfeld, M. G. (1987) Nature 328, 820-823. 23. Chen, W. S., Lazer, C. S., Lund, K. A., Walsh, J. B., Chang, C. P., Walton, G. M., Der, C. J., Wiley, H. S., Gill, G. N., and Rosenfeld, M. G. (1989) Cell 59, 33-43. 24. Glenney, J. R., Jr., Chen, W. S., Lazar, C. S., Walton, G. M., Zokas, L. M., Rosenfeld, M. G., and Gill, G. N. (1988) Cell 52, 675-684. 25. Sorkin, A., Waters, C., Overholser, J. Biol. Chem. 266, 8355-8362.
K. A., and Carpenter,
G. (1991)
6. Wiley, H. S., Herbst, J. J., Walsh, B. J., Lauffenburger,
26. Helin, K., and Beguinot, L. (1991) J. Biol. Chem. 266,8363-8368. 27. Kadowaki, T., Koyasu, S., Nishida, E., Tobe, K., Izumi, T., Takaku, F., Sakai, H., Yahara, I., and Kasuga, M. (1987) J. Biol. Chem. 262,
7. Honegger, A. M., Dull, T. J., Felder, S., Van Obberghen, E., Bellot,
7342-7350. 28. Akiyama, T., Kadowaki,
D. A., Rosenfold, M. G., and Gill, G. IV. (1991) J. Biol. Chem. 266, 11,08311,094. F., Szapary, D., Schmidt, A., Ullrich,
A., and Schlessinger, J. (1987)
CelZ51,199-209. 8. Wileman, T., Harding, C., and Stahl, P. (1985) Biochem. J. 232, 1-14.
9. Ungewickell, E., and Branto.n, D. (1981) Nature 289, 420-422. 10. Brodsky, F. M., and Parham, P. (1983) J. Mol. Biol. 167,197-204. 11. Lisanti, M. P., Sapiro, L. S.. Moskowitz, N., Hua, E. L., Puszkin, S., and Schook, W. (1982) E,ur. J. Biochem. 125,463-470. 12. Mooibroek,
M. J., Michiel,
I>. F., and Wang, J. H. (1987) J. Biol.
Chem.262,25-28. 13. DeLuca-Flaherty, C., McKa,y, D. B., Parham, P., and Hill, B. L. (1990) Cell 62, 875-887.
14. Bar-Zvi, D., and Branton, D. (1986) J. Biol. Chem. 261,9614-9621. 15. Bar-Zvi, D., Mosley, S. T., and Branton, D. (1988) J. Biol. Chem. 263,4408-4415. 16. Litwin, C. M. E., Cheng, H.-C., and Wang, J. H. (1991) J. Biol. Chem. 266,2557-2566.
T., Nishida, E., Kadooka, T., Ogawara, H., Fukami, Y., Sakai, H., Takaku, F., and Kasuga, M. (1986) J. Biol. Chem. 261,14,797-14,803.
29. Jackson, A. P., and Parham, P. (1988) J. Biol. Chem. 262, 16,68816,695. 30. Nathke, I., Hill, B. L., Parham, P., and Brodsky, Biol. Chem. 265, 18,621-18,627. 31. Felder, S., Miller, K., Moehren, G., Ullrich, Hopkins, C. R. (1990) Cell 61,623-634.
F. M. (1990) J.
A., Schlessinger, J., and
32. Livneh, E., Prywes, R., Kashles, O., Reiss, N., Sasson, I., Mory, Y., Ullrich, A., and Schlessinger, J. (1986) J. Biol. Chem. 261, 12,49012,497. 33. Winkler,
F. K., and Stanley, K. K. (1983) EMBO J. 2, 1393-1400.
34. Silveira,
L. A., Wong, D. H., Masiarz, (1990) J. Cell Biol. 111, 1437-1449.
35. Hill, B. L., Drickamer, K., Brodsky, J. Biol. Chem. 263,5499-5501.
F. R., and Schekman,
R.
F. M., and Parham, P. (1988)
