Eur. J. Biochem. 208.91 - 100 (1992) (?] FLBS 1992

A strong protein-tyrosine kinase activity is associated with a baculovirus-expressed chicken tkl gene Thomas GARTNER, Herbert KUHNEL, Gerhard RAAB, Monika RAAB, Klaus STREBHARDT and Helga RUBSAMEN-WAIGMANN Chcmotherapeutisches Forschungsinstitut, Georg-Speyer-Haus, Frankfurt, Federal Republic of Germany (Rcccived March 31, 1992) - EJB 92 0451

We have previously described a gene named tkl (tyrosine kinase related to lck). It belongs to the family of protein-tyrosine kinases and among these it has significant homology to the lck gene (lymphoide cell kinase). The tkl gene product may represent the avian homolog of Lck, which is believed to participate in a lymphocyte-specific signal transduction pathway by association with a membrane receptor. To study the biochemical properties of the protein, a nearly complete tkl gene (isolated from a cDNA library from chicken spleen cells) was expressed in a baculovirus system. Approximately 10% of the extracted protein consisted of the soluble 51-kDa Tkl protein ( ~ 5 1 ' ~at ' ) 40 h post-infection. This protein was found to be phosphorylated on tyrosine and serine residues at a ratio of 5 : 1. As expected, glycosylation or myristoylation could not be detected. Immunocomplex kinase assays indicated strong autophosphorylation of ~ 5 1 'at ~ tyrosine ' residues and phosphorylation of exogenous substrates such as D-glyceraldehyde-3-phosphatedehydrogenase (GAPDH), histones H2b and H4, and casein. This protein-tyrosine kinase activity also exhibited a marked preference for Mn2+ compared to Mg2+.The high level expression of enzymatically active Tkl should provide an excellent tool to further study the biological functions of this class of enzymes. SYC

Protein-tyrosine kinases (PTKs) are involved in the signal transducing pathway and play an important role in cell growth and differentiation (for reviews see: Hunter and Cooper, 1985; Cantley el al., 1991). These proteins are either membranespanning receptors or cytoplasmic enzymes. Some of the latter (e.g. Lck) have been shown to be associated with mernbranc receptors which do not possess PTK activity (Rudd et al., 1988; Veilletteet al., 1988). Another PTK activity has recently been found to be involved in the control of the cell cycle. Here the product of the cdc2 gene (cell division cycle) is regulated by tyrosine-specific phosphorylation and dephosphorylation (Gould and Nurse, 1989; Krek and Nigg, 1991; Dunphy and Kumagai, 1991 ; Gautier at al., 1991; Ottilie et al., 1991). The gene tkl, which we previously isolated from a chicken spleen cDNA library (Strebhardt et al., 1987), is a member of the src family of PTKs. These proteins, such as Src, Lck, Fgr, Fyn, Hck, Lyn, Yes, and Blk (Hanks et al., 1988; Dymecki

et al., 1990), are cytoplasmic membrane associated. Members of the src family are regulated in their PTK activity by a tyrosine-specific phosphorylation. An enzyme exerting this function has recently been discovered (Nada et al., 1991; Brauninger et al., 1992). Due to significant nucleotide sequence similarity, tkl was postulated to be the avian homolog of Ick, a gene initially isolated from a murine-leukemia-virus-induced thymoma cell line (Marth et al., 1985). Recent data showed an involvement of Lck in the control of T-cell growth and differentiation via physical association between the N-terminus of the 56-kDa protein encoded by Ick (p56Ick)and the cytoplasmic domain of the CD4 and CD8 receptor, mediated by cysteine motifs (Veillette et al., 1988; Shaw et al., 1990; Turner et al., 1990). Antibody-mcdiated crosslinkiiig of CD4 resulted in stimulation of ~ 5 6 ' kinase "~ activity in murine and human T lymphocytes (Veillette et al., 1989). Chicken T lymphocytes represent the only non-mammalian immune cells where homologs of CD4 and CD8 have Correspondence to Prof. H. Riibsamen-Waigmann, Chemotherabeen detected (Chan et al., 1988; Lillchoj et al., 1988). Since peutisches Porschungsinstitut, Georg-Speyer-Haus, Paul-Ehrlich-Slr. ikl is highly homologous to lck, it is a strong candidate for a 42 -44; W-6000 Frankfurt am Main 70, Federal Rcpublic of Germany A hhreviations. aa, amino acids; AcNPV, Autograplza cal~fornica PTK interacting with the avian CD4- and CD8-like transmemnuclear polyhedrosis virus; EGFR, epidermal growth factor receptor; brane glycoproteins in a similar fashion as ~ 5 6 ' does " ~ in the GAPDH, D-glyceraldehyde-3-phosphatedehydrogenase; IR, insulin mammalian system (Veillette and Ratcliffe, 1991). receplor; nt, nucleotide(s); NaCI/P,, phosphate-buffered saline; p. i., The predicted sequence of Tkl contains the characteristic post-infection: PTK, protein-tyrosine kinase; Sf. Spodoptera SYC homology regions SH-2 and SH-3 which mediate the asfrugiperda. Enzyrrzes. Calf intestine phosphatase (EC 3.1.3.1); D-glyceralde- sociation with other proteins (Pawson, 1988; Anderson et al., 1990). Tkl also contains the highly conserved kinase activity hyde-3-phosphate dehydrogenase (EC 1.2.1.12); mung bean nuclease (EC 3.1.30.1); protein-lyrosine kinases (EC 2.7.1.112); restriction region of about 260 amino acids (aa), located within the endonucleases BamHI, EcoRI, Sad (EC 3.1.21.4); T4 DNA ligase carboxy-terminal domain (Hanks et al., 1988), and the regulat(EC 6.5.1.1). ory tyrosine at the C-terminus.

,

92 To study the biochemical properties of the Tkl protein, especially its putative PTK activity, the previously characterized tkl cDNA (Strebhardt et al., 1987) was expressed in Spodoptera frugiperda (Sf) insect cells using a recombinant baculovirus derived from Autographa californica nuclear polyhedrosis virus (AcNPV). This system is capable of producing high amounts of heterologous expressed proteins and carrying out eucaryotic post-translational protein modifications (for reviews see: Summers and Smith, 1987; Luckow and Summers, 1988). Here were report the generation of high amounts of Tkl and the characterization of its PTK activity.

MATERIALS AND METHODS Cells and viruses Spodoptera frugiperda ( S f ) cell lines IPLB-SF21(A) (Vaughn et al., 1977) or its clonal derivate Sf9 were propagated as a monolayer culture in TC 100 insect medium (Gibco/BRL) supplemented with 10% fetal bovine serum and 50 pg/ml gentamycin according to the procedure of Brown and Faulkner (1977). Viral infections were performed at a multiplicity of infection of 5 for protein production and for labeling experiments, or ofO.1 for virus production. Construction of recombinant baculovirus transfer vectors The recombinant transfer vector designated pAcYM1-1tkl was prepared as follows. The 3.4-kb EcoRI fragment of the tkl cDNA, representing nucleotides (nt) 1 - 3384 from the clone C1 described by Strebhardt et al. (1987), was isolated and blunted by mung bean nuclease treatment. BamHI linkers were added and the DNA fragment was cloned into the BanzHI site of the transfer vector pAcYMl (see Fig. 1A). The recombinant transfer vector designated pAcYM1-2tkl was prepared as follows. The 1.6-kb EcoRI-Sac1 fragment of the tkl cDNA (representing nt 1 - 1619) was recovered and blunted with mung bean nuclease. After linearization of the vector pAcYMl with BamHI and subsequent mung bean nuclease treatment, a blunt end ligation was performed. Due to this cloning procedure, the BamHI insertion site of the transfer vector was deleted. Both constructs contain the 1371-bp open reading frame of the tkl cDNA coding for 457 aa. In pAcYM1-2tk1, however, the major part of the 3' untranslated region is deleted (only 229 nt of the 3' untranslated region remained). The correct orientation of the inserts in the recombinant transfer vectors was demonstrated by restriction analysis and sequence determination (Sanger et al., 1977).

Protein analysis At the indicated times post-infection, the confluent monolayers of Sf cells propagated in 35-mm cell culture dishes were rinsed with phosphate-buffered saline (NaCI/P,: 137 mM NaC1, 10 mM Pi, pH 7.4) and subsequently lysed with 100 p1 buffer A (150 mM NaCI, 50 mM Tris/HCl, 10 mM EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, pH 7.7). The cellular debris was removed by centrifugation (15000 g, 30 min, 4°C) and the protein concentration of the lysates was measured with the Bio-Rad protein assay kit. From the protein samples, 20 pg was mixed with a protein dissociation buffer (final concentration: 0.1 M Tris/HCI pH 6.8, 7.5% 2-mercaptoethanol, 3% SDS, 3% glycerol, 0.01% bromphenol blue), boiled for 2 min and separated by SDS/ PAGE on 12% gels, using standard molecular mass protein markers as reference. The gels were stained with Coomassie blue and dried under vacuum.

In vivo protein labeling

Prior to in vivo labeling with [35S]methionine,infected cells were washed with NaCI/P, and then incubated for 1 h in a methionine-free TC 100 starvation medium to reduce the intracellular methionine pool. At the indicated times, 1 2 x lo6 cells were pulse labeled for 2 h with 20 FCi [35S]methionine (1000 Ci/mmol, 10 mCi/ml, Amersham) in methioninefree TC 100 medium; 20 pg solubilized cell protein was then separated by SDSjPAGE and analyzed by autoradiography. To detect phosphoproteins, cells were washed and incubated for 3 h in phosphate-free medium to remove intracellular pools of phosphate. At the indicated times, labeling of 1 2 x lo6 cells was performed with 1 mCi ["P]orthophosphate (40 mCi/ml, Amersham) for 3 h in TC 100 medium lacking phosphate. The cells were subsequently harvested and protein analysis was performed by SDSjPAGE and autoradiography.

Production of antibodies in rabbits The 3.4-kb cDNA of tkl was cloned into the PET vector (plasmid for expression by T7 RNA polymerase; Studier and Moffatt, 1986) and expressed under the control of the T7 RNA polymerase promotor in Escherichia coli HMS 174. The protein was gel purified by SDS/PAGE, recovered by electroelution (Biotrap, Schleicher & Schiill, FRG) and dialysed extensively against 20 mM Tris pH 7.0 at 4°C. Rabbits were immunized with 50 - 350 pg according to conventional protocols. The sera were titered by Western blot analysis.

Immunoprecipitationand kinase assays Cotransfection and selection of tkl-recombinant viruses Recombinant baculoviruses were prepared by allelic exchange between wild-type AcNPV DNA and the DNA of the recombinant transfer vectors (Fig. 1 A). Cotransfection was performed according to the procedure described by Matsuura et al. (1986), using 1 pg AcNPV DNA and various concentrations of plasmid DNA. Plaques of recombinant viruses were recognized by the lack of polyhedra, using a light microscope. After at least three plaque purification steps, high titered lysates plaque forming units/ml) of recombinant baculovirus were obtained.

For immunoprecipitation 1 - 2 x lo6 cells rinsed with NaCI/Pi were lysed in 100 pl lysis buffer (50 mM NaH2P04, 10 mM EDTA, 10 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 2 mM 6-aminohexanoic acid, 0.2% Triton X-100, pH 7.0). In order to remove the cellular debris, lysates were centrifuged for 30 min at 15000 g and 4'C. For each reaction, 7 mg protein-A-Sepharose (Pharmacia) (equilibrated for 30 min in 50 pl lysis buffer), was incubated with 5 pl polyclonal antiserum for 1 h on ice with occasional shaking. The IgG-coated Sepharose beads were washed three times with lysis buffer, finally resuspended in

93 150 pl lysis buffer containing 10-20 pg cellular protein and incubated again for 1 h on ice with occasional shaking. The beads were subsequently washed in 400 p1 each of buffers 13 (buffer 1, lysis buffer containing 40 mM NaF and 1 mg/ml bovine serum albumin; buffer 2, buffer 1 containing 1 M NaCl; buffer 3, buffer 1 containing 1% Triton X-100). Finally, the beads were equilibrated in 1 ml kinase-reaction buffer (20 mM Bistris-propane, 10 mM MnC12, pH 5.9) and collected by centrifugation. For autophosphorylation assays, I0 pI of a 1 :70 dilution of ['y32P]ATP(3000 Ci/mmol, 10 mCi/ ml, Amersham) in kinase-reaction buffer were added to the Sepharose pellet. The samples were mixed and kept on ice for 3 min. For phosphorylation assays of in vitro substrates of ~ 5 1 ' ~10 ' , yg potential substrate and 5 yCi [Y-~~PIATP, in 20 pl kinase-reaction buffer, were incubated together with the immunocomplex for 30 min at 30 "C. Autophosphorylation reactions were terminated by adding 200 p1 inhibition buffer (100 mM KH2P04, 40 mM NaF, 10 mM EDTA), the beads were pelleted and the supernatant was discarded. The samples were boiled for 2 min in 35 p1 protein dissociation buffer (0.1 M Tris/HCl pH 6.8, 7.5% 2-mercaptoethanol, 3% SDS, 3% glycerol and 0.01% bromphenol blue) to dissolve the immunocomplexes. After centrifugation, the supernatant was analyzed by SDS/PAGE. Phosphorylation assays of in vitro substrates were terminated by adding 0.25 vol. of 5 x sample buffer. The samples were boiled for 2 min centrifuged and the supernatant was analyzed by SDS/PAGE. The radioactivity in the vacuuni-dried gels was visualized using an X-ray film. To determine the amount of PO:- transferred during the autophosphorylation reaction, an aliquot of the sample was spotted on pieces (1.5 x 1.5 cm) of Whatman paper and washed three times for 30min with 10% trichloroacetic acid on ice. The radioactivity of the samples was determined using a liquid scintillation counter. After subtraction of the radioactivity determined for the background samples, the amount of PO:- incorporated into ~ 5 1 ' was ~' calculated using the radioactivity of the [y-32P]ATP-containing kinase-reaction buffer as reference. The protein kinase activity was expressed as the amount of PO:- incorporated into the protein/mass protein of the lysate used for the immunoprecipitation.

Phosphoamino acid analysis

After SDS/PAGE of the kinase reaction, X-ray films were exposed to the wet gels in order to localize the band of interest. The radioactively labeled bands were excised and the proteins were eluted electrophoretically, extensively dialyzed against doubly distilled water, and lyophilized in Eppendorf vials. The samples were digested for 3 h at 110°C with 200 p16 M HC1 under nitrogen. After the digest, 200 p1 H2O was added, the samples were lyophilized again and resuspended in 10 p1 HzO. Aliquots (0.5 - 2 pl) were separated on a thin-layer cellulose sheet (Merck, ordering number 5577) using 3 pl of a mixture containing Ser(P), Thr(P) and Tyr(P) at a concentration of 1 mg/ml in 1 mM H3P04 as reference. Separation of the amino acids was performed by 1000 V in one dimension using buffer system 2 or, alternatively, in a two-dimensional fashion using buffer systems 1 and 2. Buffer system 1, acetic acid/ formic acid/H20 (78 :25 : 897), pH 1.7; buffer system 2, acetic acid/pyridine/H,O (50: 5 : 949, pH 3.5. The phosphoamino acid markers were visualized using ninhydrin spray and the radioactivity was located by autoradiography.

Isoeleetric focussing by two-dimensional gel electrophoresis

Two-dimensional gel isoelectric focussing was performed as described by O'Farrell(l975) using the carrier ampholytes Bio-Lyte 5/7 and Bio-Lyte 3/10 (Bio-Rad). For the first dimension, a lysate from pulse-labeled Sf cells, containing 25 pg protein, was focussed in a tube gel at 400 V overnight. For the second dimension, a SDSjPAGE was performed. The Tkl spot was identified by autoradiography and its isoelectric point was determined. Reference capillary gels without protein samples were cut into l-cm pieces. 'The reference pieces were equilibrated overnight in 1 ml degassed water and the pH of the water was determined.

RESULTS Construction of the recombinant transfer vector and generation of recombinant baculoviruses

For heterologous expression of tkl, a baculovirus system based on the transfer vector pAcYM1 was used which has been shown to generate high yields of recombinant protein (Matsuura et al., 1987). This vector contains all the upstream regulatory sequences of the AcNPV polyhedrin gene up to the A of the initiating ATG codon. The baculovirus expression system is very advantageous due to the usage of the strong polyhedrin promotor which can efficiently drive the expression of foreign genes. Sites for post-translational modifications like phosphorylation. glycosylation, myristoylation and proteolytic cleavage have been shown to be recognized when foreign proteins were expressed (for reviews see: Summers and Smith, 1987; Luckow and Summers, 1988). The previously characterized tkl cDNA clone C1 (Strebhardt et al., 1987) was used to construct two recombinant transfer vectors (see Materials and Methods): The first, pAcYM1-ltkl, contained the 3.4-kb EcoRI restriction fragment of tkl cDNA clone C1 with an open reading frame of 457 aa and the 2.0-kb 3' untranslated region of the gene (Fig. 1A). The second, pAcYM1-Ztkl, contained the 1.6-kb EcoRI - Sac1 restriction fragment of tklwith the open reading frame of 457 aa and 229 nt of the 3' untranslated region. For both fragments, the putative initiation codon for the expression of the protein is located 33 nt downstream of the original location of the polyhedrin start codon (Fig. 1B). This initiation codon is presumably not used in the normal expression of tkl, since the clone is missing the coding sequence for the 52 N-terminal aa (T. Gartner, unpublished; Chow et al., 1992). However, PTK activity and substrate recognition should not be affected (see Discussion). We utilized this clone, since the complete nucleotide sequence was predicted to yield a myristoylated protein which in case of v-src led to a reduced expression in the baculovirus system (G. Raab, unpublished). Recombinant viruses were generated by cotransfections of the recombinant transfer vectors with wild-type AcNPV DNA and selection of progeny viruses exhibiting a polyhedrin-negative phenotype (frequency approximately 0.1 %). Four recombinant AcNPV-1tkl baculoviruses and two recombinant AcNPV-2tkl baculoviruses were isolated from the supernatants of six different cotransfections. Correct orientation of inserts in recombinant viruses was verified by Southern blot analysis. One of each group (AcNPV-1tkl and AcNPV-2tkl) was selected for further studies.

94

A

pUC8

transfer vector pAcYMl

EcoRl digest

f

BamHl linker

BaHl

pdyhedrin promotor

ORF 3UTR ...................................... I

Wl

sac1

1

ligation

pAcYMi-ltkl AcNPV

Ir

poiyhedrin gene

f cotransfection of Sf cells

w recombinant baculovirus AcNPV-ltkl

........ .. .. .. .. .. .. .. .

......... .. .. .. .. .. .. ..

poiyhedrin promotor

tkl

ACGGATCCGCGCCCCCTGGTGTCCTACGAGGCCATG

Fig. 1.

f start codon

95

High-level expression of Tkl protein in Sf cells

Characterization of the p51fk'protein-tyrosine kinase activity

Infection of cells with both types of recombinant baculoviruses gave rise to a protein of 51 kDa ( ~ 5 1 ' ~which ' ) appeared 24 h p.i. Its amount increased up to 40 h p.i. and then decreased slightly (Fig. 2). In TPLB-SDl(A) or Sf9 cells the expression was similar No differences in virus yield or effects on the infected cells were observed. Wild-type infected cells were harvested 24 h p.i. Pulse-labeling experiments revealed that the time course of tkl expression differed from the wild-type polyhedrin gene late in infection. In both cases synthesis started 15 - 18 h p.i. While polyhedrin production reached its maximum about 50 h p i and continued until 72 h p i , ~ 5 1 ' reached ~' a maximum about 24- 30 h p i and then decreased continously (Fig. 3). To quantitate the yield of p51tk', insect cells were transferred into [35S]methionine-containing medium, at the time of infection. The cells were harvested 40 h p i and proteins were scparated by SDS/PAGE. The radioactivity in the p51tk'band represented nearly 10% of the topal [35S]methionine-labeled cellular protein as determined by liquid scintillation counting. This result was confirmed by densitometric evaluation of Coomassie-stained gels. Generally. up to 3 0 pg recombinant protein/3S-mm cell culture dish was obtained. Compared to p51ik', the production of polyhedrin was about five times higher (Figs 2 and 3).

To characterize the en~ymaticactivity of pS1 tk' expressed in the insect cells, lysates from cultures infected with recombinant virus were immunoprecipitated with Tkl-specific rabbit serum raised against bacterially expressed p5Iik1and subjected to kinase assays using [,J-~'P]ATP(see Materials and Methods). A 51-kDa protein was found to be strongly phosphorylated which was absent when lysates from wildtype infected cells were used (Fig. 4). This reaction most likely represents an autophosphorylation of ~ 5 1 ' ~However '. the phosphorylation by a kinase tightly associated with this immune complcx cannot be definitely excluded. The time course ofkinase activity showed a maximum in lysates harvested 40 h p.i. (Fig. 4), correlating with the accumulation of ~ 5 1 ' in ~' infected Sf cells (Fig. 2). Therefore, cell lysates were harvested at this time in all further experiments. The heavy chains of the polyclonal IgG antibodies were not phosphorylated in this assay, in contrast to immunocomplex kinase assays with pp60"-"c (Collett and Erikson, 1978). Phosphoamino acid analysis of the in vitro phosphorylated ~ 5 1 ' revealed ~' that it was phosphorylated on tyrosine residues, exclusively (Fig. 5). The lysis buffer and kinasc-reaction buffer used have been shown to preserve intracellular kinases in their active state, resulting in high levels of in vitro phosphorylation activity of pp6OC-"" with Mg2+as divalent cation (Jacobs and Rubsamen,

Fig. 2. Accumulation of p51 in infected Sf cells, analysed by SDS/PAGE and Coomassie blue staining. Cells wcre infected with recombinant baculovirus AcNPV-ltkl (A) or AcNPV-2tkl (B). At the indicated times p.i., the cells were harvested and lysates were prepared; 20 pg protein/ lane was separated by SDS/PAGE. Lane WT shows a protein lysate of AcNPV-infected Sf cells; the position orpolyhedrin is indicated. Lane C shows the prolein lysate of uninfected control cclls. The p51 band, indicated by an arrow, was first detected 24 h p.i., reached its maximum about 40 h p.i., and then decreased slightly. The molecular masses of standard proteins (lane M) are indicated at the left.

Fig. 1. Generation of the recombinant baculovirus AcNPV-ltkl. The 3.4-kb EcoRJ fragment of the tkl cDNA was inserted into the BumHI site of the transfer vector pAcYM1. After cotransfection of Sf cells with the recombinant transfer vector and wild-type virus DNA, recombinant baculoviruses were obtained (A). The correct junction betwecn the vector and the 5' end of the insert was confirmed by sequence analysis. The sequence shown in (B) encompasses the A of the polyhedrin start codon, thc BamHI linker (underlined), 21 5' untranslated nucleotides of the tkl cDNA and the ATG of the truncated gene.

96

Fig. 3. Time course of protein synthesis in baculovirus-infected Sf cells. Cells were infected with AcNPV-1tkl (A) or AcNPV wild-type (B) at a multiplicity of infection of 5. At the indicated times p i . cells were labeled for 2 h with 20 pCi [35S]methionine as described in Materials and Methods; 20 pg solubilized cell protein was separated on each lane and analyzed by autoradiography. Positions of standard molecular mass markers are shown in the middle. The bands representing p51'" and polyhedrin are indicated by arrows, Lane C shows uninfected control cells labeled at 24 h p.i.

Fig. 4. Time course of ~51'"autophosphorylation activity. Sf cells were infected with AcNPV-1 tkl and harvested at the indicated times p.i. Clarified lysates containing 1 5 pg protein were immunocomplexed withpolyclonal antiserum. Kinase assays were performed with 1.4 pCi [y-32P]ATPfor 3 min on ice. The immunocoinplexes were separated by SDSIPAGE and analyzcd by autoradiography. In lane WT, a kinase assay with wild-type infected Sf cells is shown. Positions of molecular mass standards are indicated at the left.

Fig. 5. Phosphoamino acid analysis of in vitvo phosphorylated ~51'". After the kinase assay, the immunocomplex was separated by SDSi PAGE. The band representing pSltk' was eluted from the gel, and processed as described in Materials and Methods. Thc sample was spotted onto a thin-layer chromatography plate and separated by electrophoresis in two dimensions. The positions of the nonradiactive references were visualized using ninhydrin spray and are circled on the autoradiogram.

1983). It has previously been reported that the in vitro kinase 10 mM MnClz was much more efficient in supporting the activity of PTKs is dependent on the kind and the concen- reaction, resulting in the transfer of almost 2000 fmol PO:-/ tration of divalent cations. Therefore, we tested the influence mg protein used in the precipitation reaction (Fig. 6). No of various concentrations (1-200mM) of MgC12, MnCI, reaction was observed when C a Z f was used, as has been or CaC1, on the autophosphorylation activity of ~ 5 1 ' ~ ' .shown for pp60'-"'"(Riibsamen et al., 1979). In further analogy Maximum incorporation of 150 fmol POi-/mg protein was to pp60"-src,GTP was accepted as phosphate donor but was achieved at a concentration of 50 mM MgC1,. However, less efficient than ATP (data not shown).

97 phosphate bound [frnol/rng protein]

1 mM

5 rnM

10 m M

25 mM

50 mM

i00 m M

200 mM

concentration of divalent cations

Fig. 6. Autophosphorylation activity of ~51"' under various ionic conditions. Lysates of AcNPV-ltkl-infected cells were harvested at 40 h p.i.; 15 pg solubilized protein was reacted with Tkl-specific antiserum and kinase assays were performed using reaction buffers with different concentrations of MnC12 or MgC12 (1 -200 mM). The incorporation of 3zP was measured as described under Materials and Methods. The y-axis of the diagram shows the amount of transferred POi-/mass of cellular protein used for the immunoprecipitation.

Isoelectric point of ~ 5 1 ' ~and ' post-translational modifications

of the glycolytic pathway, are phosphorylated at tyrosine in Rous-sarcoma virus-transformed chicken or mouse fibroTwo-dimensional gel electrophoresis of [35S]methionineblasts (Cooper et al., 1983; Eigenbrodt et al., 1983) and can labeled proteins from p51'k'-producing cells revealed a major serve as exogenous substrates for both pp60''"'" and pp60'-"' spot at an isoelectric point of 5.8 f 0.2 representing p51tk'. in vitm (Coussens et al., 1985; Eigenbrodt et al., 1983). Other Two neighbouring aecondary spots with thc same molecular potential substrates tested in vivo and/or in vitro were either mass, which may be different modifications of the protein, phosphorylated weakly and predominantly at serine and threwere also seen. To investigate whether ~ 5 1 ' ~was ' phos- onine (aldolase, GAPDH, pyruvate kinase, yeast hexokinase) phorylated, the insect cells were labeled with [32P]orthoor not phosphorylated (triose phosphate isomerase, phosphate, After SDSjPAGE of the soluble cell proteins or phosphoglycerate kinase). after immunoprecipitation with a Tkl-specific serum followed Various potential substrates were tested: lactate dehydroby SDSIPAGE, a phosphorylated p51"' band was detected genase, enolase, GAPDH, pyruvate kinase, histones, casein, on the X-ray film, demonstrating that the protein was phos- violet phosphatase, and a,-macroglobulin. Among them, phorylated in the insect cells. A phosphoamino acid analysis GAPDH and the histones H4 (strongly) and H2b (weakly) of in vivo phosphorylated and immunoprecipitated protein were phosphorylated in immunocomplex kinase assays with revealed a major phosphorylation of tyrosine and minor phos- ~ 5 1 ' (Fig. ~ ' 7). Phosphoamino acid analysis revealed, in the phorylation of serine at a ratio of 5 : 1 (data not shown). case of GAPDH and H4, phosphorylation on tyrosine resiA putative N-glycosylation site (Asn-Xaa-Thr/Ser) is pre- dues exclusively, whereas H2b was phosphorylated at both sent in p51tk'at ad 78 - 80. In order to check for glycosylation, tyrosine and serine (Fig. 8). Compared to GAPDH and histunicamycin was employed at a concentration of 5 pg/pl. Tu- tones, casein, which is a strong in vitt-o substrate of pp60'-"' nicamycin, an inhibitor of N-linked glycosylation, had been (Coussens et al., 1985), was only weakly phosphorylated by shown to be effective in AcNPV-infected Sf cells (Stiles et al., ~ 5 1 ' at ~ 'tyrosine and serine to equal levels. When wild-type 1983) and was used to produce nonglycosylated interferon$ Iysate was used as negative control, the weak phosphorylation (Smith et al., 1983). The addition of tunicamycin resulted in of casein was observed at serine residues only (data not the disappearance of a 64-kDa AcNPV glycoprotein and to shown). Phosphorylation by ~ 5 1 ' could ~' not be detected for the appearance of a new 58-kDa protein (Dobos and Cochran, pyruvate kinase, violet phosphatase, lactate dehydrogenase, 1980). In our investigations, the inhibitor had no effect on the enolase, and a,-macroglobulin. molecular mass of ~ 5 1 ' ~ suggesting ' that the protein had no When the ~ 5 1 ' ~immunocomplex ' kinase assay was N-linked carbohydrates. This finding is consistent with reperformed with control rabbit serum as exogenous substrate, ports that intracellular PTKs are not glycosylated. Presumsome serum proteins were phosphorylated, but they were missably, due to the N-terminal truncation, no incorporation of ing when a wild-type protein lysate was used (data not shown). [3H]myrist~cacid into ~ 5 1 'could ~ ' be detected.

Substrates of p51"' in immunocomplex kinase assays

DISCUSSION

It has been been demonstrated that enolase, phosphoglycerate mutase, and lactate dehydrogenase, three enzymes

Recent data indicate that tkl probably is the avian homolog of the mammalian Ick PTK gene (Chow et al., 1992)

98

Fig. 7. Phosphorylation of in vitro substrates by p51"'. Solubilized lysates of virus-infected Sf cells containing 20 pg protein were immunoprecipitated with specific antiserum. A phosphorylation assay using 5 pCi [y-'*P]ATP and 10 pg GAPDH or a mixture of all five histoncs was performed as described in Materials and Methods. The samples were resolved by SDSjPAGE and analyzed by autoradiography. The phosphorylated GAPDH and histones (II2b and H4) are shown in the left lanes of A and B, respectively, and are indicated by arrows. The middle lanes or both A and B show the kinase assay without addition of any substrate, only the strong autophosphorylation signal of p51 appears (arrow). The right lanes in both A and B show the control assay with wild-type infected cells.

Fig. 8. Phosphoamino acid analysis of gel purified substrates after phosphorylation in an immunocomplex kinase assay. The in vitro phosphorylated substrates were eluted from the gel, processed as described undcr Materials and Methods and separated on a thin-layer chromatography plate using buffer system 2 (see Materials and Methods). Positions of ninhydrin-stained phosphoainino acid standards Ser(P), Tyr(P) and Thr(P) are circled. (Indicated by pSer, pTyr and pThr, rcspcctiveiy).

and that it is physically associated with the avian CD4 and CD8 T-cell surface antigens (Veillette and Ratcliffe, 1991). Furthermore, ~ 5 6 " was ~ implicated in mammalian lymphomagenesis (Marth et al., 1985). The role of Tkl in the avian immune system and, especially, in lymphoid neoplasia remains to be elucidated. To study its putative PTK activity, tkl was expressed in a baculovirus system utilizing the formerly isolated cDNA coding for a nearly complete Tkl protein ( ~ 5 1 ' ~ 'This ) . protein is missing the 52 N-terminal amino acids which provide a signal for N-terminal myristoylation (Giirtner et al., unpublished; Chow et al., 1992). In previous PTK expression studies with v-src, we observed that the myristoylated form of the protein was expressed less efficiently in the baculovirus system (G. Raab, unpublished). However, the enzymatic properties should not be affected since the expressed Tkl protein comprises the complete kinase domain and the amino-terminally located src homology regions (SH-2 and SH-3) which may be involved in substrate specificity and in the regulation of PTK activity (Pawson, 1988; Anderson et al., 1990). For insulin receptor (IR) and epidermal growth fiictor receptor (EGFR), both expressed in a baculovirus system, it has been demonstrated that the PTK domain alone (without the N-terminal sequences of the protein) was sufficient to exhibit PTK activity (Ellis et al., 1988; Herrera et al., 1988; Wedegaertner and Gill, 1989). The truncated IR was also able to phosphorylate exogenous substrates like histone H2b and poly(G1u-Tyr).The truncated EGFR showed similar enzymatic characteristics to the intact activated EGFK (Hsu et al. 1990). In further comparative studies with the complete Tkl protein, the function of the N-terminus will be elucidated. The efficient expression of an intracellular PTK in the baculovirus system, under the transcriptional control of thc strong polyhedrin promotor demonstrated here, provides a versatile tool for further biochemical studies. The use of the recombinant virus AcNPV-2tk1, where the major part of the 3' untranslated region of the tkl cDNA was deleted, did not result in an alterated yield of recombinant protein. It seems that the 3' part of the tkl mRNA has no effect on protein expression in this system. The synthesis of ~ 5 1 ' started ~' at the same time as that ofpolyhedrin (15- 18 h p i ) and reached a maximum between 24 and 30 h p.i. However, in clear contrast to polyhedrin, the rate of ~ 5 1 ' synthesis ~' dropped after 30 h p.i. This may be due to toxic effects of the heterologous protein on the metabolism of infected cells. Based on [3sSS]methionineincorporation and densitometric evaluation, we estimate that ~ 5 1 ' amounts ~' to 10% of the extracted cellular protein. This is equivalent to 10 pg ~ 5 1 ' ~ ' being produced/35-mm cell culture dish. The accumulation of ~ 5 1 ' is ~ at ' least ten times higher than the amount of pp60"-"'" expressed in a comparable system (Piwnica-Worms et al., 1990) and at least hundredfold higher than the expression of pp60'-"" using the same baculovirus system (G. Kaab, personal communication). In additional comparisons to pp60'-"'" expression, ~ 5 1 ' appeared ~' to show much less toxicity although it accumulated to higher levels in infected insect cell cultures. This is supported by the observation that insect cells expressing pp60'-"" were partly detached from their matrix 20 h pi., while wild-type infected and AcNPV-1tkl infected cells detached much later p.i. and in less quantity. The predicted protein, based on the nucleotide sequence characteristics, encoded by the chicken tkl gene, was postulated to be an intracellular PTK. As demonstrated here, the heterologously expressed p51tk' was phosphorylated in vivo and displayed strong phosphorylation on t y rosine in

99 immunocomplex kinase assays. This is very likely due to autophosphorylation activity. GAPDH, histones H2b and H4, and casein were phosphorylated by the p51tk'-associated tyrosine kinase activity in vitra. demonstrating that the baculovirus expression system yielded a functional protein. However, the poor phosphorylation efficiency of these substrates compared to the strong phosphorylation of p5Itkiin the immune complex assay inakes it likely that none of these proteins belongs to the physiological substrates of Tkl. Although a considerable enrichment of p51tklis achieved by immunoprecipitation, such complexes cannot be considered entirely pure. Thus, at present, we cannot definitely prove that the observed PTK activity is intrinsic. The activity could also be due to another tightly associated kinase. This second protein would then have physiological significance. It is surprising that enolase was not found to be a substrate for ~ 5 1 ' ~ since ' it is a substrate for p561Ckand for all other SPC family meinbers (Hunter et al., 1988). This observation may indicate a functional difference between Tkl and Lck. Substituting MnCI, for MgClz increased the kinase activity associated with p51tk'. This preference for MnZt is in contrast to pphO'-"'" (Riibsamen et al., 1979) and to v-abl produced in bacteria (Foulkes et al., 1985), both of which showed a marked preference for M g 2 + . A preference for Mn"+, however. has been observed for other PTKs expressed in baculovirus systems including the IR kinase domain (Herrera et al., 1988) and thc kinase domain of EGFK (Wedegacrtner and Gill, 1989). Also, the considerable decrease of the kinase activity of p51Lk',at high concentrations of divalent cations ( > 100 mM), is consistent with these results. It has been suggested that Mn2+ might have a greater catalytic activity than Mg", especially for PTKs with a low level of activity (Schicven and Martin, 1988). Our data, however, appear to confirm the idea of different ionic requirements for the kinase activity of diverse intracellular PTKs. On average, after 3 min of incubation on ice, approximately 0.1 mol phosphate was incorporated/molp51tk'.This activity is in the same range as the autophosphorylation of the cytoplasmic domain of the human 1R (Herrera et al., 1988). Due to the high enzymatic activity associated with p51tk' expressed in the baculovirus system, further work to purify this protein and to elucidate the interactions between the kinase and its substrates should provide insight into the physiological role of this protein. We thank D. H. L. Bishop ( N t R C Institute of Virology. Oxford) for providing the baculovirus cxprcssion systcni, M . Sturm for expert technical assistance, and K. Unger for critical reading of the manuscript. This work was supported by a graduate fellowship from the Deul.vc.lza Geseilschnjt f u r Cliemisches Appurntcwesen (T. G.). Thc Georg-Spe~~er-Huusis supported by the Biindesniinisterium fur G'tmridlreit and the Hes.rischiJMinistcJi-iumf u r Wi,vsmscfiafiund Kunst.

REFERENCES Andcrson. D., Koch. C. A,, Grey, L.; Ellis, C., Moran, M. F. & Pawson, T. (1990) Binding ofSTI2 domains of phospholipase C,1. GAP, and Src to activated growth factor rcceptors. Science 250. 979 -982. Brauningcr, A., Holtrich, U., Strebhardt, K. & Riibsamen-Waigmnnn. 11. (1992) Isolation and characierization of a human gene that encodes a new subclass of protein tyrosine kinascs. Gene 110, 205-21 1.

Brown, M. & Faulkncr, P. (1977) A plaque assay for nuclear polyhcdrosis viruses using a solid overlay. J . Gen. Virol. 36, 361 364. Cantley, L. C., Auger; K. R., Carpenter, C., Duckworth. B., Graziani. A., Kapeller. R. & Soltoff, S. (1991) Oncogcnes and signal transduction. Cell64.281 -302. Chan. M. M.. Chen, C. L., Ager, L. L. & Cooper, M. D. (1988) Identification of the avian homologucs of mammalian CU4 and CD8 antigens. J. Imnzunol. 140, 2133-2138. Chow, L. M. L., Ratcliffe, M. J. IT. & Veillette, A. (1992) tkl is the avian homolog of thc mammalian Iclz tyrosinc-protcin kinase gene. M o l . Ccll. Biol. 12, 1226- 1233. Collctt, M. S. & Erikson. R. L. (1978) Protein kinase activity associated with the avian sarcoma virus src gene product. Proc. Nut1 Acad. Sci. U S A 75.2021 -2024. Cooper, J. A,, Rciss, N . A., Schwartz, R. J. & Hunter, T. (1983) Three glycolytic enzymes are phosphorylated at tyrosine in cells transformed by Rous sarcoma virus. Nature 302,218-223. Coussens, P. M., Cooper, J. A., Hunter, T. & Shalloway, D. (3985) Restriction o f the in vitro and in vivo tyrosine protein kinase activities ofpp60"~""relative to pp60"-"". Mol. Cell. B i d . 5,2753 2163. Dobos. 1.' & Cochrdn, M. A . (1980) Protein synthesis in cells infected by Aictogrupha californicu nuclear polyhedrosis virus (AcNPV): thc effect of thc cytosine arabinoside. Virology 103,446-464. Dunphy, W. G. & Kumagai, A. (1991) Thc cdc2.5 protein contains an intrinsic phosphatase activity. Cell 67, 189- 196. Dymccki, S. M., Nicderhuber, J. E. & Desiderio, S. V. (1990) Specific expression of a tyrosine kinase genc. blk, in B lymphoid cells. Scimce 247, 332 - 336. Eigenbrodl, E., Fister, P., Rubsamen, H . & Priis, R. R. (1983) Influence of transformation by Kous sarcoma virus on the amount, phosphorylation and enzyme kinetic properties of enolase. EMBO J . 2. 1.565- 1570. Ellis. L.. Levitan, A,. Cobb, M . H., & Ramos, P. (1988) Efficient expression in insect cells o f a soluble; active human insulin receptor protein-tyrosine kinase domain by usc of a baculovirus vector. J . Virol. 62. 1634 - 1639. Foulkes, J . G., Chow. M., Gorka. C., Frackelton, A. R. Jr & Baltimore, D. (1 985) Purification and characterization of a protein-tyrobine kinase encoded by the Abelson murine leukemia virus. J . Biol. C'hem. 260. 8070-8077. Gautier. J., Solomon, M. J., Booher, K. N., Bazan, J. F. & Kirschncr, M. W. (1991) cdc25 is a specific tyrosine phosphatase that directly activates p34C"2. CdI 67, I97 - 21 I . Gould. K. L. & Nursc. P. (1989) Tyrosiiic phosphorylation of the fission yeast cdc2+ protein kinase regulates entry into mitosis. Nuture 342. 39-45. Hanks, S. K., Quinn, A. M. & Hunter, T. (1988) Thc protciii kinase family: conserved features and deduced phylogeny of the catalytic domains. S&ncc 241. 42- 52. Herrera, R.. Lebwohl, D., de Herreros, A. G.. Kallen, I