the product of the kin1 locus in saccharomyces cerevisiae is a serine/threonine-specific protein...

YEAST VOI.. 7: 219-228 (1991)

The Product of the KINI Locus in Saccharomyces cerevisiae is a Serine/Threonine-Specific Protein Kinase ALLEN LAMB*. MICHAEL TlBBETTSt AND CHARLOTTE 1. HAMMOND;

Dcywttnent of Moleculur Rio1og.r und Biochemistry, Wlcdeyun University. Middletown. Connecticut 06457. G.S. A .

Received 25 September. 1990

The catalytic domain (30 kDa) of all protein kinases can be aligned for maximum homology. thereby revealing both invariant and highly conserved residues. The K I N l locus from Saccharomyces cerevisiae was isolated by hybridization to a degenerate oligonucleotide encoding the conserved protein kinase domain, DVWSFG. The predicted amino acid sequence revealed significant homology to the catalytic domain of protein kinases. Using antibodies raised against a bacterial LucZ. K I N 1 fusion protein. we have identified by immunoprecipitation the yeast K I N l gene product as a 145 000 dalton protein (p 145K'9. In exponentially growing yeast cells, the K I N l protein is phosphorylated primarily on serine residues. The gene product of K I N l was shown to be a scrincithreonine-specific protein kinase in immune complexes. as determined by the transfer of label from [y-32P]ATP to either pp145K'"' or to an exogenously added substrate. u-casein. The optimal metal ion concentration in this assay was 20 mM-MnCI,. Subsequent phosphoamino acid analysis of the radiolabclled product, pp145K'"', demonstrated that this autophosphorylation was specific for serine;threonine residues. There is no apparent difference between wild-type cells and cells containing a disrupted K I N 1 gene. The biochemical characterization of protein kinases in simple eukaryotes such as yeast will aid us in determining the role of phosphorylation in cellular growth and physiology.

KEY WORDS -- - Protein kinase; Saccharomyces cerevisiae; yeast; protein phosphorylation.

INTRODUCTION

We have focused on the characterization of yeast protein kinases, homologous to the protein kinase family of viral oncogenes, in order to understand their structure and function in normal cellular pro- cesses. Yeast is an ideal system for these studies due to the ease of rigorous genetic analysis. Specific DNA sequences can be easily introduced and expressed in yeast using in vitro transformation procedures. Cloned yeast genes can be mutated in ritro and used to replace the wild-type gene by in viw recombination (for review, see Botstein and Davis. 1982).

The study of the protein kinase family has many advantages. Not only does thc family contain the largest number of viral oncogene members, but a great deal of information is available concerning

*Present addrcss: Cytogenetics Laboratory. Dcpartment of Pediatrics. University of North Carolina, Chapel Hill. North Carolina 27599. U.S.A. +Present address: Department of Biology, Univcrsity of Michigan. Ann Arbor. Michigan 48109. U.S.A.

Addressee for correspondence.

0749 503X .'9 I ;0302 19.- I0 505.00 1991 by John Wiley & Sons Ltd

the functional domains of these proteins (Bishop and Varmus, 1985; Hanks et al., 1988; Hunter and Cooper, 1985). This family is defined by its shared homology to the catalytic domain of pp60'-"' (the 30 000 dalton carboxy-terminal half of the protein that carries phosphotransferase activity) (Brugge and Darrow, 1984; Levinson et al., 1981). Within this region of v-src, the amino acid sequence is 4&82% homologous to the other oncogene protein kinases. Moreover, there is significant homology between the tyrosine- and serine!threonine-specific kinases (Barker and Dayhoff, 1982; Bishop and Varmus, 1985; Hanks et al., 1988; Hunter and Cooper, 1985). This conservation of structure and function among members of the protein kinase family suggests a shared ancestry. I t is possible that this large family of cellular oncogenes descended from progenitor genes whose identities may best be elucidated using a simple eukaryote such as yeast (Doolittle, 1981; Hanks et al., 1988). At the very least, we would expect the actual number of protein kinases in yeast to be less than in mammalian cells and therefore more accessible to the study of regulation by protein phosphorylation.

220

The KINl gene from the budding yeast Succharo- inj.ces wrrrisiae was isolated by hybridization to a pool of synthetic oligonucleotides encoding amino acids conserved among the protein kinases (DVWSFG). DNA sequence analysis of the KINI predicted protein product showed 30% identity to bovine CAMP-dependent protein kinase and 27% identity to the protein kinase encoded by the v-src oncogene within their catalytic kinase domains (Levin et ul.. 1987). Therefore, it seemed likely that the protein product of the KINl gene would be a phosphoprotein and have protein kinase activity. Several other protein kinases have been studied in ycast and they have all fallen within the serine; threonine-specific protein kinase family (for a review and individual references, see Hanks e f al., 1988). Although low levels of phosphotyrosine and tyrosine-specific protein kinase activity have been detected in S . cerevisiur cell lysates (Castellanos and Mazon, 1985; Schieven et al.. 1986), no protein tyrosine kinase gene has yet been isolated.

We report here the use ofantiserum raised against a bacterial LucZ,'KI.NI fusion protein to identify and characterize the yeast KIN1 gene product. As predicted from the nucleic acid sequence analysis, the protein encoded by the KIN1 gene is a phosphoprotein in viva and has serinejthreonine-specific protein kinase activity. It remains to be determined where this serineithreonine protein kinase is local- ized in the yeast cell and how this phosphorylation step affects yeast physiology.

A. LAMB, M. TIBBETTS A N D C. 1. HAMMOND

MATERIALS AND METHODS Enzymes und reugents

Phenylmethylsulfonyl fluoride (PMSF), isopropylthio-fLgalactoside, dithiothreitol, T4 DNA ligase and all restriction endonucleases were purchased from Bethesda Research Laboratories. Aprotinin. glass beads (425--600 microns) and u-casein were obtained from Sigma Chemical Company. [35S]Methionine. [''P]orthophosphate. and [y-32P]ATP were purchased from New England Nuclear. Pansorbin cells were obtained from Calbiochem Co.

Yeus t strains Succhuroniyces cerevisiae strain H R 125-5D

( M A T a leu2-3.112 uru3-52 his3-532hi.d t rp l - I ) was obtained from Dr Ira Herskowitz, University of California, San Francisco and used as the wild-type strain ( K I N I ) in all experiments. To disrupt the

KINl gene, 80% of the coding domain was deleted from the 5'-most SphI site to the 3'-most Xhal site (See Figure I ) . The yeast selectable marker URA3

Gene Dtsruotion of KIN1

P H p S S S X X E Hp P

I 1 I I

P Hp S

w X E Hp P

P X E Hp P URA3

P H p S P

i Hp SP P X E

U RA3 - Figure I . Gene disruption of K f N f ( k i n f - A f : : L : R A 3 ) . Eighty per cent of the nucleotides encoding the K f N / gene were deleted from the 5'-most Sphl site to the 3'-most Xhol site and replaced with the yeast selectable gene. L'RA.3. The isolated H p d - € c o R I fragment containing the deleted region of the K f N f gene replaced with the L:RA3 gene was used to transform yeast cells. The open box depicts the coding domain of the Kf,V/ gene drawn from left to right. 5' to 3' direction. The black box depicts the coding domain of the L:RA3 gene. Restriction enzyme recognition sites are indicated for €coKI(E) . H p o l ( H p ) . PsrI(P). Sphl(S) and Xbol (X ).

replaced the Sphl-Xbal deleted region of the KINI gene with the direction of transcription of URA3 opposite to that of the KINl gene. This plasmid construct (an Hpal-EcoRI fragment) was used to transform yeast cells (HR125-5D). In this gene replacement experiment, a double recombination event replaces the chromosomal locus with the disrupted gene and selectable marker (Rothstein, 1983). The deletion of the chromosomal copy of the KINI gene (k in l -Al : :URA3) was verified by Southern blotting (unpublished data). The entire KINl gene was subcloned into the single BamHI restriction endonuclease cleavage site of the multicopy plasmid YEp24 (YEp24.KINI) (gift of Dr David Levin, Johns Hopkins University). YEp24.- KINI DNA was used to transform the yeast strain. H R 125-5D.

Kf,\ / I.OCUS IN SAU'HAROMYCES CEREVISIAEIS A SERINE.THREONlNE-SPECIFIC PROTEIN KINASF. 22 1

Prcpurution of' the LacZiKIN 1 fusion protein und gcncmtion of KIN I-specific untiserum

Recombinant DNA procedures were performed using standard techniques (Maniatis et al., 1982). Restriction endonucleases were purchased from Hethesda Research Laboratories and used accord- ing to the manufacturer's recommendations.

The 1.24 kb internal XhaI fragment of the KINI gene was ligated, in frame, into the single XhaI site in the expression vector pUR288 (Ruther and Muller-Hill. 1983) (Figure 2). The pUR expression

ration of E - ~ l K l N l F m

KIN 1

5 ' L 3'

P H p S S S X X E Hp P

1 kinase domam

Xba 1 fragment ofml subcbned In frame into the vector. pUR288. 1- at the 3 end of Dgalaaosdase gene

X X

I ' k b I

Figure 2. Generation of ~-galactosidase, K I N / bacterial fusion protein. The 1.24 kb internal Xhol fragment of the KfN/ gene W A S ligated. in frame, into the single Xhal site of the expression vector. pUR288. This Xhul fragment of the KfNf gene is located 3' to the catalytic kinase domain. Correct orientation of thc Xhal fragment in the vector was confirmed by restriction cnzymc analysis. This black box depicts the location of the catalytic kinase domain. Restriction enzyme recognition sites are indicated for €coRI(E), Hpal(Hp). Psr l (P) . Sphl(S) and XhaI(X).

vectors contain the LacZgene encoding P-galactosidase. with various restriction enzyme cleavage sites in the3'end. The 1.24 kb Xbal fragment ofthe KIN1 gene is located 3' to the putative catalytic kinase domain. The orientation of the KINI 1.24 kb XbaI fragment cloned into the vector was confirmed by restriction enzyme mapping.

Escherichia coli strain JM 101 was transformed with the pUR288.KINI-XbaI gene fusion. The transformed cells were grown overnight with vigorous shaking at 30°C in minimal medium (60mM- K,HPO,, 33 mM-KH,PO,, 8 mM-(NH,),SO,. 2 mM-

Na citrate.2H20. supplemented with 0.0005% thia- mine, 0.6 pM-MgS0,. 0. I YO casamino acids and 0.5% glycerol). The following day. the cells were diluted back, grown to mid-log phase (ODMK, = 0.5). and induced by thc addition of isopropylthio-P- galactoside to a final concentration of 1 m w After 2 h induction. near the end of log phase (OD,M,= 1.0), bacterial cells from 100 ml cultures were harvested by centrifugation, 20 min 2600 x g at 4'C, and resuspended in 5 ml of Laemmli sample buffer (50 mM-Tris-HCI (pH 6.8). 2% SDS, 10% glycerol. 0.0 1 % bromophenol blue, 2% P-mercaptoethanol; Laemmli. 1970). Cells were broken by sonication, on ice, with several 10-15 s bursts. Cell debris was removed by centrifugation in a clinical centrifuge (1750 x g for 10 min at room temperature).

Bacterial proteins were boiled for 3 min and separated by SDSipolyacrylarnide gel electrophoresis (7.5%: 29.2:0.8 acrylamide: bisacrylamide) (Laemmli, 1970) on 3 m m prcparative gels. The I65 kd P-ga1actosidase:KINI fusion protein was identified by its size and increase in quantity over time with induction. To determine the location of the fusion protein following electrophoretic separ- ation, a small portion of the gel was stained (0.1 YO Coomassie Blue R-250. O.lO/o Amido Black. 30% methanol, 10% acetic acid), and then destained with several changes of 50% methanol, 10% acetic acid. This gel slice was then rehydrated in water, aligned with the remaining unstained gel and a horizontal slice of the unstained gel was cut that corresponded to the location of the fusion protein. The fusion protein was electroeluted into dialysis tubing with 1 x SDS running buffer (25 mM-Tris HCI, 191 mM- glycine, 0.1 ?An SDS (pH 8.3); Laemmli, 1970) for a minimum of 6 -8 h at 15 mA. constant amperage. and then dialysed in phosphate-buffered saline. The purity and quantity of the fusion protein was assessed by electrophoresis on an analytical SDS; polyacrylamide gel. A New Zealand white rabbit was injected with IOOpg of fusion protein, plus adjuvant. every 4 weeks for 1 year. Serum samples were collected every 4-8 weeks, and specificity to the ycast KINI protein was determined by immunoblot analysis.

In vivo labelling

For [?3]methionine labelling, yeast cells were grown overnight with vigorous shaking at 30'C in YNB medium (6.7 g/l yeast nitrogen base (Difco) supplemented with 2% glucose and appropriate amino acids). Four OD,," units of logarithmically

222 A. LAMB. M. TIBBETTS AND C . 1. HAMMOND

growing cells were collected in a clinical centrifuge ( 1 750 x g, 5 min) and resuspended in 2 ml of fresh YNB medium. A total of 67 pCi of [35S]methionine was added to each 2 ml culture and incubated with shaking for 3 h at 30°C.

For ['2P]orthophosphate labelling, yeast cells were grown overnight with vigorous shaking in a low phosphate medium containing 20 mM-KCl, 2 rn~-MgS0,-7H,O, 2 m~-caC1,-2H,O, 15 mM- (NH,),SO,, 8.5 mM-NaC1, 0.1 % yeast extract (Difco), 2% glucose plus the appropriate amino acid supplements. Logarithmically growing yeast cells were harvested, 4 OD,,, units were resuspended in 2ml of medium, and labelled with IOOpCi of ["P]orthophosphate for 3 h at 30°C.

Immunoprecipitation

Cells were pelleted in a clinical centrifuge ( 1 750 x g, 5 min) at room temperature, washed once in cold 10 mM-NaN,, and resuspended in 50 pl of NET buffer (400 mM-NaC1, 5 mM-EDTA, 50 mM- Tris-HCI (pH S), 1% NP-40) containing 2mM- PMSF, 0.1 mg/ml aprotinin and 0.2 g of glass beads. Samples were vortexed at high speed for 90s and held on ice. Four hundred and fifty microliters of NET buffer were added to the broken cells, gently mixed, and the solution was removed from the glass beads. The lysate was clarified by centrifuging for 5 min in a microfuge at room temperature. Fifty microliters of Pansorbin (10% suspension of formalin-treated staphylococci) were added to each 500 p1 of labelled lysate (4 OD units of cells) and rocked for 30 rnin at room temperature. The Pansorbin was pelleted by centrifuging for 5 min at room temperature in a microfuge and discarded. To perform immunoprecipitations, 3 p1 of KIN1 antiserum or 5 p1 of pre-immune antiserum were added to the supernatant and rocked for 1 h at room temperature. After the addition of 50 pl of Pansorbin, the incubation was continued for another 30 rnin at room temperature. The Pansorbin-bound immune complexes were collected by pelleting in a microfuge for 20s at room temperature. The immune complexes were washed by resuspending the pellet and collected by centrifuging for 20s at room temperature. The pellets were washed three times with NET buffer plus 0.3% SDS and 1.0% sodium deoxycholate and once in 0.01 M-Tris-HCI (pH 6.8) plus 1 mM-PMSF. After the final wash, pellets were resuspended in Laemmli sample buffer and boiled for 3 min to release the radiolabelled immunoprecipitated proteins. The Pansorbin was pelleted by

centrifuging for 2 min in a microfuge and the supernatant was loaded on a 7.5% SDS/polyacrylamide gel.

Protein kinase activity in immune complexes The procedure for immunoprecipitation of KINl

protein from unlabelled cell lysates was identical to that described above with the following modifications. 0.8 OD,,, units of cells were used in each assay. The Pansorbin-bound immune complexes were pelleted and washed three times in NET buffer and twice in 20mM-Hepes (pH 7.3) plus 1 mM- PMSF at room temperature. The Pansorbin-bound immune complex was then resuspended in assay buffer (20 mM-Hepes (pH 7.3), 0.5 mM-dithiothreitol, 5 pCi [Y-~~P]ATP and various combinations of cations) and incubated at 30°C for 15 min. Sodium vanadate was added to some kinase assays at the final concentration of 5 0 p ~ . The reaction was stopped by addition of 2 x Laemmli sample buffer plus 30 mM-EDTA (Laemmli, 1970).

The identification of the phosphoamino acid labelled in the protein kinase assay was determined by one-dimensional thin-layer electrophoresis (Hunter and Sefton, 1980; Snyder et al., 1983).

RESULTS Production of antibodies raised against a b-galactosidase/KIN 1 fusion protein

To initiate the study of the KINl gene product, antibodies to this protein were needed. Knowledge of the KINl gene structure (Levin et al., 1987) allowed us to fuse a segment of the coding region to the LacZ gene of E. coli to produce a p-galactosidase gene fusion. The 1.24 kb XbaI fragment of KINl (see Figure 2) was ligated, in frame, into the single XbaI site of the expression vector pUR288 (Ruther and Muller-Hill, 1983). The XbaI fragment of KINl is located 3' to the putative kinase domain and does not contain any part of the catalytic kinase domain. After transformation of this gene fusion clone into E. coli, total protein from crude lysates (Reed, 1982) was electrophoresed on SDS/polyacrylamide gels (Laemmli, 1970).

After Coomassie Blue staining of the SDS/polyacrylamide gel, we could detect a 165 kilodalton (kDa) protein present in induced bacterial cultures containing the gene fusion clone. The identification of the bacterial P-galactosidase/KINI fusion protein was based on the expected size (165 kDa) and the increase in protein level with induction time.

P f S l LO<'I;S IN SACCHAROMYCES CEREVfSlAE IS A SER1NE:THREONINE-SPECIFIC PROTEIN KINASE 223

As expected, the fusion protein was present at low levels in non-induced cultures and absent in cultures with the pUR288 vector alone (data not shown). The P-ga1actosidase:'KINI fusion protein was extracted from the polyacrylamide gels and prepared for immunization of rabbits (Tjian et ul., 1974).

The K I N I gene product is identified bv K I N 1 - .vpc~ific antiserum as u 14.5 kDa protein

To identify the protein product encoded by the KI'VI gene. we prepared in vivo [%]methionine- labelled cell lysates from the following (see Figure 3): ( I ) yeast cells containing YEp24.KINI--the entirc KIN1 gene subcloned into a plasmid containing a 2-pm replicon for high copy maintenance (20 -30 copies per cell) (lanes A. B, C ) ; (2) wild-type yeast cells (KINI) (lane D); and (3) yeast cells containing a deletion of the KINl locus (kinl-Al:: L,'RA3) (lane E).

Yeast cells deleted for the KINl gene (kinl-Al:: L'RA3) are viable and have generation times similar to wild-typc yeast cells (KINI). These cells have a normal cellular morphology, and can mate and sporulate efficiently. We have examined their ability to utilize non-fermentable carbon sources without finding any differences from wild-type yeast cells. In fact. at this time. there appears to be no obvious phenotypic defect associated with deletion of the KIA'/ gcne (unpublished results).

Antiserum to the P-ga1actosidase;KINI fusion protein immunoprecipitated two yeast proteins of Mr - 145 000 and -95 000 as determined by electrophoresis through SDSlpolyacrylamide gels (Figure 3: lanes C, D). Since the amount of the 95 kDa protein diminishes under conditions that re- duce proteolysis, we believe that the 95 kDa protein represents a breakdown product of the KINI gene product. As expected for the protein product of the KIIVl gene. the 145 kDa protein was present in large quantities when yeast cells contained the multicopy plasmid YEp24.KINI (lane C ) , lower quantities when yeast cells contained a single copy of the KINl genc (lane D), and no protein was seen in cells with the KIiVI genedeleted (lane E). The 145 kDa protein could not be immunoprecipitated with pre-immune rabbit serum from [35S]methionine-labelled cell lysatescontaining YEp24.KINI (lane B). Indeed the majority of the 3sS-labelled proteins seen in Figure 3 (lanes A -E) adhere to Pansorbin alone (lane A). The identity of the KINI gene product as the 145 kDa protein (p 145'"'') was confirmed by immunoblot- ting techniques (data not shown).

A B C D E

180- 145- 116-

84 - 58-

Figure 3 . Lysates from ["Slmethionine-labelled yeast cells immunoprecipitated with anti-KIN/ serum or pre-immune rabbit antiserum. Three different yeast cell lysates were used to identify the KIN/ gene product: ( I ) yeast cells containing the multicopy plasmid YEp24.KINI (lanes A, B, C): (2) yeast cells containing a singlechromosomalcopy ofthe KINl gene(lanc D); and (3) yeast cells containing a deletion of the KINl locus (k in l - AI;:G'RA3) (lane E). Lane A shows that the majority o f back- ground "S-labelled proteins adhere to Pansorbin alone. Lane B is a yeast cell lysate immunoprecipitated with pre-immune rabbit antiserum; lanes C--E are yeast cell lysates immunoprecipitated with anti-KIN/ serum.

The protein product of'KIN I is u phosphoprotcin

Since the KINl gene protein predicted from the DNA sequence analysis was homologous to known protein kinases (Levin et ul., !987), most of which are phosphoproteins, we wanted to determine if the KINl protein was a phosphoprotein in vivo.

224 A. LAMB, M. TIBBETTS A N D C. 1. HAMMONI)

Lysates as described above were obtained from ycast cells labelled in v i w with [32P]orthophosphate (see Figure 4) and analysed by immunoprecipitation

A B C D

pl45-

Figure 4. Lysates from [':P]orthophosphate-labelled yeast cells immunoprecipitated with anti-Kfhrlserum or pre-immune rabbit serum. Three different yeast cell lysates were used to determine if the KfNf gene product was phosphorylated in riso: ( I ) yeast cells containing the multicopy plasmid YEp24.KfNI (lanes A and D); (3) yeast cells containing a single chromosomal copy of KIN1 (lane B); and (3) ycast cells containing a deletion of the KfNl locus ( k i n / - A l : : L : R A 3 ) (lane C). In lanes A C. protein was immunoprecipitated with anti-KfNf serum; lane D, prorein was Immunoprecipitated with pre-immune rabbit serum.

with either pre-immune rabbit serum (lane D) or anti-KIN1 rabbit serum (lanes A--C). A phosphorylated 145 kDa protein (pp145) was present in large quantities when yeast cells contained the YEp24.KINI plasmid (lane A), lower quantitics when yeast cells contained a single copy of the KINl gene (lane B). and absent when yeast cells contained a deletion of the KINI gene (lane C). Moreover, ppl45 could not be immunoprecipitated from yeast cell lysates with pre-immune rabbit serum (lane D). Therefore, the protein product of the KINI gene exists as a phosphoprotein ( ~ ~ 1 4 5 ~ ' " ' ) in exponentially growing yeast cells.

To identify the amino acids of ~ ~ 1 4 5 ~ " ' that are phosphorylated in vivo, the "P-labelled band of 145 kDa was extracted from the SDSipolyacryl- amide gel. Partial acid hydrolysis followed by one- dimensional thin-layer electrophoresis (Hunter and Sefton, 1980; Snyder et al., 1983) indicated that ~ ~ 1 4 5 ~ ' " ' contained both phosphoserine and phosphothreonine residues but no phosphotyrosine residues (Figure 7; in v i w ) .

pp145K"' is u serineithreonine-specific protein kinase

Due to the striking homology of the KIN1 gene product with thecatalyticdomain of known protein kinases, a protein kinase activity for the KINI protein was expected. Yeast cell lysates as described above were immunoprecipitated with either pre- immune rabbit serum or anti-KIN1 serum. Immune complexes were incubated under a variety of ionic conditions in the presence of [y3?P]ATP and electrophoresed on an SDSipolyacrylamide gel (See Figure 5). In this phosphotransferase reaction, pp145K'"' transferred phosphate from [.I-~'PP]ATP to either itself or the exogenously added substrate, a-casein. Unlike the class of tyrosine protein kinases such as pp60'-"' (Collett and Erkison, 1978; Levinson et al., I978), the heavy chain of the immu- noglobulin was not phosphorylated. The levels of protein kinase activity corresponded to the quantity of pp145K"V' present in the cellular lysates (Figure 5; lanes A, B. C). In addition, the degree of pp145K'"' autophosphorylation always paralleled the phosphorylation of the added substrate, a-casein. Very little, if any, protein kinase activity was seen either in lysates from cells containing the KINl deletion (lane C) or in lysates from cells immunoprecipitated with pre-immune rabbit serum (lane D).

To optimize the kinase assay and determine metal ion specificity, immune complexes containing PP145K"v' were prepared and transfer of phosphate

KIVl LOCUS IN SACCHAROMYCES CEREVISIAE IS A SERINE THREONINE-SPECIFIC PROTEIN KINASE

A B C D E F G H I J p145-

225

casein-

Figure 5 . Characterization of the protein kinase activity of pp14SK"'. Yeast cell lysates were immunoprecipitated with either anti- K I N f serum (lanes A--C; E-J) or pre-immune rabbit serum (lane D). lmmunecomplexes were incubated with 20 mM-MnCI,and a final concentration of I mgiml a-casein (lanes A-D) and assayed for the ability to transfer ["P]phosphate from [y-!>P]ATP to pp145""' o r the exogenously added substrate. a-casein. The following yeast cell lysates were used: ( I ) yeast cells containing YEp24.KINI (lanes A and D); (2) yeast cellscontaining a singlechromosomalcopyof K I N f (lane B; lanes E -J); and (3) yeast cellscontaininga deletion in the K I N f locus (kinf-Al::URA.3) (lane C). Other assay conditions tested were as follows: lane E 20 mM-MnCI,, lane F 20 mu-MgCI,. lane G 20 mhi-MnCI,+ I mM-ZnCI,, lane H 20 mM-MnCI,+ 50 VM vanadate, lane I 20 mM-MgCI,+ 50 phi-vanadate, and lane J 20 mM-MnCI,+ 1 mM-ZnCI, + 50 pM-vanadate. The signal intensity obtained for lane A took only I h whereas the signal for lanes B J was obtained after 48 h

from [y-32P]ATP to pp145K'"' was measured with varying concentrations of either Mn" or MgZf metal ions (Figure 6 ) . The optimal ionic concentration was 20 mM-MnC1,. Although protein kinase activity could also be obtained with 20 mM-MgCI,, MgClz gave a two- to three-fold reduction in activity compared with MnCI,. Since orthovanadate often stimulates protein kinase activity presumably by the inhibition of phosphatases (Leis and Kaplan, 1982; Swarup ef af.. 1982a,b), we wanted to determine what effect the addition of vanadate would have on K I N l protein kinase activity (Figure 5; lanes E-J). The addition of 50 p-vanadate consistently in- creased protein kinase activity when 20 mM-MnC1, was used as the ion (lane E; -vanadate vs lane H; fvanadate), but no effect was seen on protein kinase activity when MgCI, was used (lane F;

-vanadate vs lane I; +vanadate). The addition of 1 mM-Zn2+ to assays containing either 20 mM- MgCl, or MnCI, significantly inhibited protein kinase activity, unlike the protein kinase activity of the yeast cell division cycle gene, CDC28, which requires Zn2+ ions (Reed el al., 1985) (Figure 5 ) . Moreover, the protein kinase activity of pp145""' was not effected by the addition of 2 mM-CaZf (data not shown).

To determine whether pp145K'.v' was a serinei threonine- or tyrosine-specific protein kinase. the pp145 band autophosphorylated in the immune complex kinase assay was extracted from a SDSj polyacrylamide el. A phosphoamino acid analysis

product has serine/threonine-specific protein kinase activity in vitro (Figure 7; in vifro).

of this ~ ~ 1 4 5 ~ ' " P band revealed that the K I N l gene

226 A. LAMB, M. TIBBETTS A N D C. I. HAMMOND

Mn++ Ionic Conditions Mg++ Ionic Conditions

5000 5000

4000 4000

4 3000 $ 3000 m

Y (0 Y : 2000 : 2000

1000 1000

0 0

0 5 1 0 20 30 40 50 rnM conc

0 5 1 0 20 30 4 0 50

rnM conc

Figure 6 . Determination of the optimal ionic conditions for pp145"'"' in the immune complex protein kinase assay. Yeast cell lysates from cells containing a single chromosomal copy of KIN1 were immunoprecipitated with anti-KIN1 serum. Immune complexes were incubated with varying concentrations of either Mnz+ or Mg" cations, as described in Materials and Methods, and the products of the kinase reactions run on a 7.5% SDS/polyacrylamide gel. The effect of ion concentration on the ability to transfer the phosphate from [y-"PIATP to pp 145""' was determined by scanning the autoradiogram with a Zeineh Soft Laser Densitometer (Biomed Instrument Inc.).

IN VIVO IN VlTRO

Figure 7. Phosphoamino acid analysis ofppl45""'. The 32P-labelled band, pp145x'"', either phosphorylated in vivo (see Figure4) or in vitro in the protein kinase assay (see Figure 5) was extracted from the SDS/polyacrylamide gel and subjected to one-dimensional thin-layer electrophoresis. Positions of marker phosphoamino acids are indicated.

K I V l LOCbS IN SA(‘CHAROMYCE.5 CLREVISIAE IS A SERINE THREONINE-SPECIFIC PROTEIN KINASK 227

DISCUSSION We have used antiserum raised against a bacterial Lrrc.Z-KI.NI fusion protein to identify and characterize the K I N I gene product in Saccharom~ws cww‘isiuc. The K I X I gene product is a 145 000 dalton protein as demonstrated by immunoprecipitation from radiolabelled yeast cell lysates. The 145 000 dalton protein (ppl45”.”) is phosphorylated ii7 riiw predominantly on serine residues although a low level of phosphothreonine residues w;is also detectcd.

DNA sequence analysis predicts that p~145~;’.“ is a protein kinase (Levin c t ul.. 1987). In the immune complex assay, pp 145’”’ can transfer phosphate from (./-“P]ATP to either itself or the exogenously added substrate. u-casein. Although optimal protein kinase activity requires Mn’+, MgZL ion can be substituted with a subsequent decrease i? protein kinase activity. This preference for Mn-’ as the metal ion in the protein kinase assay was not seen with the three other yeast protein kinases whose ac- t iv i ty has bcen characterized in i i t ro (CDC28 (Reed ct (11.. 1085); c ~ k . 2 ~ (Simanis and Nurse, 1986); SNFI (Celenra and Carlson. 1986)). In these cases, no prefcrcncc between the divalent cations, Mg” and Mn” . was seen. In addition, Zn” ions. which stimulate CDC2X protein kinase activity (Reed et (I/.. 1985). actually dccreasc the protein kinase activity of pp145”.\’. Added Ca” ions have no efrcct at all on thc protein kinase activity. From the phosphoamino acid analysis of the autophosphorylated band at 145 000 daltons and subsequent one-dimensional thin-layer electrophoresis, we can conclude that pp145’’.‘’ is 11 serinejthreonine- spccific protein kinase in r i m .

A notcworthy feature of the K I X I gene product is that its molecular weight (145 kDa). as determined by SDS. polyacrylamide gcl electrophoresis. is sub- stantially greater than the molecular weight predicted from the DNA sequence analysis ( I 17 kDa) (Lcvin c’f (11.. 1987). Unusual stretches of amino acids or post-translational modifications such as phosphorylation and glycosylation might contrib- ute to this 28 kDa differcncc. Thcrc are ten potcn- tial consensus sequences for I%-linked glycosylation (carbohydrate chains linked to asparagine residues) (Marshall. 1974; Struck and Lennarz, 1980) pre- dictcd from DNA sequence analysis of the K I N I gene (Lcvin e t ul.. 1987). Preliminary studies dcmonstrating that pp145””.\” can bind to the lectin (Concanavalin A) suggest that pp145K’”’ is lyco-

not change eithcr in tunicamycin-treated yeast sylated. The apparent mobility of pp145’’.” B does

cell lysates or with endoglycosidase-H digestion (unpublished data). Some specific ossibilities are suggested by these results. ppl45“’.’‘might be modified post-translationally by N-linked glycosylation but this modification is not solely responsible for the altered mobility of the protein. On the other hand. pp145K”\” might not be modified by N-linked glycosylation at all. but rather, may be modified by 0- linked glycosylation (carboyhydrate chains linked to serine/threonine residues). Finally, pp 145K’’v’ might be post-translationally modified by both N- linked and O-linked glycosylation. On SDS/polyacrylamide gels, the K I N l protein migrates as a single, discrete band with a molecular mass of 145 000 daltons. In contrast, proteins modified by N-linked glycosylation like invertase usually migrate as a broad, diffuse band extending over 10-30 kDa. We think i t is, therefore, more likely that the KIN1 protein is modified by 0-linked glycosylation than N-linked glycosylation.

To determine precisely the reason(s) for the altered electrophoretic mobility of the K I N l protein, we are in the process of identifying the i j i r im) translation product of the K I N I locus. The relative size of the protein should reflect the mobility of the K I N I protein in a SDS/polyacrylamide gel before any post-translational modifications. Further structural analysis will be required to determine the exact locations of post-translational modifications on the KINI polypeptide chain and their functional role in ViW.

To date. this K / N I protein kinase is the largest protein kinase in yeast. A general feature of most protein kinases is that thecatalytic activity is located at the carboxy-termini of the proteins. Like the S N F l gene in S . cvrevisiuc (Celenza and Carlson, 1986) and the viral ah1 oncogene (Reddy ef al.. 1983). the kinase domain of the K I N l gene product is located near the amino-terminus. I t remains to be determined if these proteins have any functional similarities. The intracellular location of this serine, threonine-specific yeast protein kinase should pro- vide an important clue to its function at the molecular level.

ACKNOWLEDGEMENTS We thank Ira Herskowitz for the HR125-5D yeast strain and David Levin for the Y E p 2 4 . K I N l plasmid. Special thanks to Gregory Payne. Randy Schatzman and Peter Kiener for advice and assist- ance and to Laura Grabel and T. C. James for valuable comments on the manuscript.

228

This research was supported by grant BMV-335A from the American Cancer Society.

A. LAMB. M. TIBBETTS A N D C. 1. H A M M O N D

REFERENCES Barker, W. C. and Dayhoff. M. 0. (1982). Viral src gene

products arc related to the catalytic chain of mammalian CAMP-dependent protein kinase. Proc. Natl. Acad. Sci. C!SA 79,28362839.

Bishop. J. M. and Varmus, H. (1985). Functions and origins of retroviral transforming genes. In Weiss, R., Teich. N., Varmus, H. and Coffin, J. (Eds). RNA Tumor Viruses. Cold Spring Harbor Laboratory. Cold Spring Harbor, New York, pp. 999-1 108.

Botstcin. D. and Davis. R. W. (1982). Principles and practicc of recombinant DNA research with yeast. In Strathern. J . N., Jones. E. W. and Broach. J. R. (Eds), Thc Molecular Biologj* of' ihr Yeasi Sacrharomyces: A4c~iaholi.sm and Gene E.ypres.sion. Cold Spring Harbor Laboratory. Cold Spring Harbor. New York. pp. 607- 636.

Bruggc. J. S. and Darrow. D. (1984). Analysis of the catalytic domain of phosphotransfcrase activity of two avian sarcoma virus-transforming protcins. J . Bid. Cheni. 259,4550455 7.

Castellanos. R. M. P. and Mazon. M. J . (1985). Identifi- cation of phosphotyrosine in ycast proteins and of a protein tyrosine kinasc associated with the plasma membrane. J . B id . Chem. 260,8240-8242.

Cclcnza. J . L. and Carlson, M. (1986). A ycast gene that is essential for release from glucose repression encodes a protein kinase. Science 233, I I75--1180.

Collctt, M. S. and Erikson. R. L. (1978). Protcin kinase activity associated with the avian sarcoma virus src genc product. Proc. Natl. Acad. Sci. USA 75, 202 I - 2024.

Doolittlc. R. (1981). Similar amino acid sequences: Chance or common ancestry? Science 214, 149-1 59.

Hanks. S. K . , Quinn, A. M. and Hunter, T. (1988). The protcin kinasc family: Conserved features and deduced phylogeny of the catalytic domains. Science 241,42-5 1 .

Hunter, T. and Cooper, J. A. (1985). Protein-tyrosine kinases. Ann. Rev. Biochem. 54,897-930.

Huntcr. T. and Sefton, B. M. (1980). The transforming gene product of Rous sarcoma virus phosphorylates tyrosine. Proc. Nuil. Acad. Sci. USA 77, I3 I 1-1 3 15.

Laemmli. U. K . (1970). Cleavage of structural proteins during thc asscmbly of the head of bacteriophage T4. Naiure 227,680-685.

Leis, J. F. and Kaplan, N. 0. (1982). An acid phosphatase in the plasma membranes of human Astrocytoma showing marked specificity toward phosphotyrosine protein. Proc. Nail . Acad. Sci. USA 79,6507-65 I 1 .

Lcvin, D. E., Hammond, C. I . , Ralston, R. 0. and Bishop, J. M. (1987). Two yeast genes that encode unusual protein kinases. Proc. Nail. Acad. Sci. USA 84,6035--6039.

I-evinson. A. D.. Courtneidgc, S. A. and Bishop, J . M. ( 198 1 ). Structural and functional domains of thc Rous

sarcoma virus transforming protein (pp60"'). Proc. Nail. Acad. Sci. USA 78, 1624-1628.

Levinson, A. D., Oppermann, H., Levintow, L., Varmus, H. E. and Bishop, J . M. (1978). Evidence that the transforming gene of avian sarcoma virus encodes a protein kinase associated with a phosphoprotein. Cell

Maniatis, T., Fritsch, E. F. and Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.

Marshall, R. D. (1974). The nature and metabolism of the carbohydrate-peptide linkages of glycoproteins. Biochem. Soc. Symp. 40, 17-26,

Reddy. E. P.. Smith, M . J. and Strinivasan, A. (1983). Nuclcotidc sequence of Abelson murine leukemia virus genome: Structural similarity of its transforming gene product to other one gene products with tyrosinc- specific kinase activity. Proc. Nail. Acad. Sci. USA 80,

Reed, S. I . (1982). Preparation of product-specific anti- sera by genc fusion: Antibodies specific for the product of the yeast-division-cycle gene CDC28. Gene 20,

Reed, S. I., Hadwiger, J. A. and Lorincz, A. T. (1985). Protein kinase activity associated with the product of the ycast cell division cycle genc CDC28. Proc. Natl. Acad. Sci. USA 82,40554059.

Rothstcin, R. J. (1983). One-step genedisruption in yeast. In Wu, R.. Grossman, L. and Moldave, K. (Eds), Methods in Enzymology. vol. 101, Recombinant DNA, Part C. Academic Press, New Y ork, pp. 202-2 1 1 .

Ruther. U. and Muller-Hill, B. (1983). Easy identification ofcDNAclones. EMBO J . 2, 1791-1794.

Schieven, G.. Thorner, J. and Martin, G. S. (1986). Protein-tyrosine kinase activity in Saccharomyces cerevisiae . Science 23 1,39@- 393.

Simanis, V. and Nurse, P. (1986). The cell cycle control gene cdc2' of fission yeast cncodes a protein kinasc potentially regulated by phosphorylation. Cell 45,

Snyder, M. A.. Bishop, J. M., Colby, W. and Levinson. A. D. (1983). Phosphorylation of tyrosine-416 is not required for the transforming properties and kinase activity of pp6W.'". Cell32,891-901.

Struck, D. and Lennarz, W. (1980). In Lennarz, W. (Ed.), The Biochemistry of Glycoproteins and Proteogl.vcans. Plenum Publishing Corp., New York, pp. 35-83.

Swarup, G., Cohen, S. and Garbcrs, D. L. (1982). Inhi- bition of membrane phosphotyrosyl-protein phospha- tasc activity by vanadate. Biochem. Biophys. Res. Commun. 107, 1104-1 109.

Swarup. G., Spccg, Jr., K . V., Cohen, S. and Garbcrs, D. L. (1982). Phosphotyrosyl-protein phosphatase of TCRC-2 cclls. J . B id . Chem. 257,7298 -7301.

Tjian. R., Stinchcomb, D. and Losick, R. (1974). Anti- body directed against Bacillus subtilis G factor purified by sodium dodecyl sulfate slab gel electrophoresis. J . Biol. Chrm. 250.8824 8828.

15,561-572.

3623-3627.

255-265.

26 1-268.

the product of the kin1 locus in saccharomyces cerevisiae is a serine/threonine-specific protein...

Documents