Priya KunapuliS, James J. OnoratoO, M. Marlene H o s e a and Jeffrey L. BenovicS .... pH 7.2, 2 m EDTA, 0.5 nm phenylmethylsulfonyl fluoride, 20 pdml.
THEJOURNAL OF B I O ~ I C CHEMISTRY AL 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 269,No. 2, Issue of January 14,pp. 1099-1105, 1994 Printed in U.S.A.
Expression, Purification, and Characterization of the G Protein-coupled Receptor KinaseGRK5* (Received for publication, July 26, 1993, and in revised form, September 24, 1993)
Priya KunapuliS, James J. OnoratoO, M. Marlene H o s e a and Jeffrey L. BenovicS From the jllepartment of Pharmacology, Jefferson Cancer Institute, Thomas Jefferson University, Philadelphia, Pennsylvania 19107, the §Department of Medicine (Nephrology Section), University of Wisconsin, Madison, Wisconsin 53792, and the Wepartment of Pharmacology, Northwestern University Medical School, Chicago, Illinois 60611
G protein-coupled receptor kinases (GRKs) such as tems, a rapid loss of responsiveness occurs following receptor loss of receprhodopsin kinase and the P-adrenergic receptorkinase activation (3,4). This rapid activation-dependent (PARK) play an important rolein agonist-specific phos- tor responsiveness appears to be mediated by specific G prophorylation and desensitizationof G protein-coupledre- tein-coupled receptor kinases (GRKs) that have the unique ceptors. GRK5 is a recently identified member of the ability torecognize and phosphorylate their receptor substrates GRK family that has greater homology with rhodopsin when they are in their active conformations(5). The p-adrenerkinase than with PARK. To further characterizethe ac- gic receptor kinase (PARK) (6, 7) and rhodopsin kinase (8, 9) tivity of GRK5, it has been overexpressed in Sf9 insect have beenimplicated as the major kinases involved in the cells and purified by successive chromatography on S- stimulus-dependent phosphorylation of the P z A R and rhodopSepharose and Mono S columns. GRK5 phosphorylates sin, respectively. The subsequent uncoupling of the receptor the &-adrenergic receptor ( P a ) , m!Z muscarinic cho- and G protein is then mediatedby arrestin proteins that spelinergic receptor, and rhodopsin in an agonist-depend- cifically bind to the phosphorylated and activated forms of the ent manner to maximal stoichiometries of-2.5,1.5, and 1 mol ofphosphatelmol of receptor, respectively, with K, receptor (3-5, 10). values of ”0.5 p~ for theP-, -16 p~ for rhodopsin, and To date, six mammalian andtwo Drosophila GRKs have been -24 forATP. Peptide phosphorylationstudies suggest identified. In addition to PARK (11)and rhodopsin kinase (121, that in contrast to PARK and rhodopsin kinase, GRK5 the other members of the GRK family include bovine PARK2 preferentially phosphorylates nonacidic peptides with(13), a human I T l l (14), GRK5 (15) and GRKG (16), and DroK, of -1.5 m ~Heparin . and dextran sulfate were foundsophila GPRK-1 and GPRK-2 (17). A comparison of the amino acid sequences of the GRKs suggests that there aretwo major to be potent inhibitors ofGRK5 with ICso values of -1 rm,thereby being at least 150-fold more potent on GFtK5 branches of the GRK family tree. While P A R K 2 and Drosophila than on PARK. GRK5 can also be activated by polyca- GPRK-1 appear to be most similar to PARK, with amino acid tions, with 10 p~ polylysine promoting an -2.6-foldac- identities of 84% and 64%, respectively, rhodopsin kinase, IT11, tivation. Overall,these studies demonstrate that GRK5 GRK5, GRK6, and Drosophila GPRK-2 have significantly lower has unique properties that distinguish it from other homology with PARK ( 3 5 4 0 % amino acid identity) andform a members of the GFtK family and thatlikely play an im- separate branchof the family tree. Thephylogenetic classificaportant role in modulating its mechanism of action. tion of the GRK family is further supportedby functional analysis of the various GRKs. In vitro studies have demonstrated that both PARK and PARK2 share a very similar substrate Many transmembrane signalling systemsconsist of specific specificity at the level of amino acid preference (they phospeptides withamino-terminal G protein-coupled receptors that transduce the binding of ex- phorylateserine-containing tracellular ligands into intracellular signalling events 2). (1,G acidic residues) (18), receptor phosphorylation( P 2 A R , m2 musprotein-coupled receptors modulate the activityof a wide vari- carinic cholinergic, and substance P receptors are good subety of effector molecules including adenylyl cyclase, cGMP strates in vitro) (18-201, and potential mechanism of cellular G protein P y subunits) (18).PARK, phosphodiesterase, phospholipase C, phospholipase Az, and activation (interaction with various ion channels. Two of the best characterizedG protein- PARK2, and Drosophila GPRK-1 also appear to be ubiquitous coupled receptors arethe&-adrenergic receptor ( P 2 A R ) , l proteins being expressed in a variety of tissues (11, 13, 17). Other than rhodopsin kinase, which is predominantly exwhich mediates catecholamine stimulation of adenylyl cyclase (31, and the visual “light receptor” rhodopsin, which mediates pressed in rod and possibly cone outer segments (4), the exphototransduction in retinal rod cells (4). In both of these sys- pression pattern of the various GRKs does not coincide with any particular G protein-coupled receptor. While PARK is a ubiquitous protein that likely plays a major role in the phos* T h i s research was supportedin part by National Institutes of Health Grants GM44944 (toJ. L. B.), HL45964 (to J. L. B.), HL31601 (to phorylation and desensitizationof the agonist-activated &AR, M. M. H.), GM47120 (to J. J. 0.1, a grant from the Wisconsin affiliate of it may well have a general role in phosphorylating and reguthe NationalKidneyFoundation (to J. J. O.),and by a predoctoral lating theactivity of multiple receptors (18-23). The expression fellowship from the American Heart Association, Southeastern Penn- and in vitrosubstrate specificity of P A R K 2 demonstrates that it sylvania Affiliate (toP. K.). The costs of publication of this article were shares many properties with PARK (13, 18-20). However, redefrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” inaccordancewith 18 cent findings suggest that P A R K 2 may also be localized in U.S.C. Section 1734 solely to indicate this fact. olfactory cilia in the rat and may play a role in regulating The abbreviations used are: p,AR, &-adrenergic receptor; PARK, odorant receptor function (24). In general, the tissue distribuP-adrenergic receptor kinase; G protein, guanine nucleotide binding protein; GRK, G protein-coupled receptor kinase; H-7, l-(5-isoquinoli- tion of human GRKG also appears to be very similar to that of levels for both in brain, skeletal muscle, nylsulfonyl)-2-methylpiperizine;m2 mAChR, m2 muscarinic choliner- PARK with the highest gic receptor; ROS, rod outer segments. of I T l l and pancreas (16). In contrast, the tissue distributions
Purification and Characterization of GRKS
and GRK5 are quite distinct from that of PARK, PARK2, and GRK6. IT11 is expressed at highest levels in the testis and at lower levels in a number of other tissues (14) while GRK5 is expressed at highest levels in the heart, placenta, and lung (15). While at present little is known about the function of GRK5, GRK6, IT11, and Drosophila GPRK-2, the similarities between these four proteins suggest that they may well share common roles in the cell. In this study, we present the results of the first successful purification and characterization of the properties of GRK5, including an analysis of its ability to phosphorylate three functionally distinct G protein-coupled receptors. The results highlight striking differences that exist in the properties of GRK5 and PARK.
tuted into chick heart phospholipid vesicles as previously described (19). Preparation of Synthetic Peptides-Peptides were synthesized on an Applied Biosystems 430A synthesizer using t-butoxycarbonyl chemistry. Peptides were purified by high performance liquidchromatography on a Dynamax C-18 reverse phase column using a 0-20% gradient of acetonitrile in 0.1% trifluoroacetic acid. Phosphoserine-containingpeptides were made in a 3-ml reaction mixture containing 20 nm Tris-HC1, pH 7.2,2 m EDTA, 6 m concentration of the peptide RRRASASAAor RRRASAAASAA, 10m MgC12,6 m [Y-~~PIATP (1cpdpmol), and 7 pg of MP-dependent protein kinase catalytic subunit. After a 6-h incubation at room temperature, 32Pincorporation was determined by spotting 10-pl aliquots onto P81 paper followed by six washes with 75 nm phosphoric acid(28). The phosphoserine-containingpeptides were then purified on a 1-ml Mono S column using a linear gradient from 0-1 M ammonium acetate as described (29).This step effectively separated the more acidic phosphorylated peptide from the nonphosphorylated pepEXPERIMENTALPROCEDURES tide. All peptide stock solutions were adjusted to pH 7.0 with 50 nm Tris Materials-The chromatography resins S-Sepharose and Mono S and then used in phosphorylation reactions. were purchased from Pharmacia LKB BiotechnologyInc. Frozen bovine Assay for GRK5-GRK5 was routinely assayed using 2-6 p d rhodopretinas were from George A. Hormel & Co. Wild type Spodoptera fru- sin (urea-treated ROS), 20 m Tris-HC1, pH 7.5, 2 m EDTA, 5 m giperda (Sf9)cells were from the American m e Culture Collection. MgC12, 0.1 nm [Y-~~PIATP (specificactivity of 700-1500 cpdpmol) in a l-(5-Isoquinolinylsulfonyl)-2-methylpiperizine (H-7), heparin (average total reaction volume of 20-40 pl. Rhodopsin phosphorylation reactions M, = 4,000-6,000),dextran sulfate (average M, = 5,000), poly-L-aspartic were incubated at 30 "C for 1-3 min in the presence of fluorescent room acid (M,= 5,000-15,000), poly-D-glutamic acid (M, = 2,000-15,000), light. When the pzAR (20 m) or the m2 mAChR (20 m) was used as spermine, spermidine, poly-L-lysine (M, = 1,000-4,000),histones (type substrate, 100 concentrations of the agonists (-)-isoproterenol or IIS), phosvitin, (-)-isoproterenol, and carbachol were from Sigma. Tis- carbachol were included, respectively. These reactions were incubated sue culture reagents were from Life Technologies and Sigma. at 30 "C for up to 60 min. All receptor phosphorylation reactions were [y-32PlATPand [12511iodopindololwere purchased from Dupont NEN. stopped with 10-20 pl of SDS sample buffer (8% SDS, 25m Tris-HC1, The catalytic subunit of the CAMP-dependent protein kinase was pur- pH 6.5, 10% glycerol, 5% P-mercaptoethanol,0.003% bromphenolblue) chased from Promega. and then electrophoresed on a 10%SDS-polyacrylamidegel by the Expression and Purification of GRKS from Sfs Cells-Previously, we method of Laemmli (30). The gels were dried and autoradiographed, isolated a baculovirus containing the GRK5 cDNA (pBacPAK-GRK5) and the 32P-labeledreceptor bands were excised and counted to deterand used it to express human GRK5 in Sf9 insect cells (15). To purify mine the picomoles of phosphate transferred. All values presented are GRK5, Sf9 cells in three 150-mm dishes (-18 x lo6 cellddish) were the means of 2-6 independent experiments. In theexperiments testing infected with 5 ml of the pBacPAK-GRK5 virus per dish. Following a various protein kinase inhibitors and activators, 100% activity repre48-h infection, the cells wererinsed with phosphate-bufferedsaline and sents the phosphorylation of rhodopsin by GRK5 in the absence of any then harvested by scraping in 3 ml of ice-cold bufferA (20 nm HEPES, compound. IC5o values represent the concentration of the compound pH 7.2, 2 m EDTA, 0.5 nm phenylmethylsulfonyl fluoride, 20 pdml that inhibited 50% of the GRK5 activity. Whensynthetic peptides were leupeptin, 20 pg/ml benzamidine, 0.02% Triton X-100) containing 250 used as substrates, the reactions were incubated at 30 "C for 30 or 60 m NaCl. The cells were lysedwith a Brinkmann tissue disrupter (30 s min. Peptide phosphorylation reactions were quenchedby the addition at maximum speed) and centrifuged at 40,000 x g for 20 min. The of trichloroacetic acid to a final concentration of 15% followed by spotsupernatant was then centrifuged at 300,000 x g for 1 h. This high ting the samples on P81 paper, washing six times in 75 m phosphoric speed supernatant was diluted to 6 ml with ice-cold bufferA and loaded acid, and counting in a scintillation counter. The counts observed in the onto a 1-ml S-Sepharose column pre-equilibrated with buffer A contain- absence of peptide were subtracted from all values before determining ing 125 m NaCl. The column was washed with 20mlof buffer A the amount of phosphate incorporated into the peptide. containing 200 m NaCl and then eluted with a 25-ml linear gradient of 200-500 m NaCl in buffer A at a flow rate of 0.5 d m i n . GRK5 RESULTS activity was assessed by measuring the ability of the various fractions to phosphorylate rhodopsin in a light-dependent fashion. The S-SephaPreviously, we demonstrated that human GRK5, expressed rose fractions containing the peak of GRK5 activity were pooled (-10 in Sf9 insect cells using the baculovirus system, was able to ml), concentrated to 2.5 ml in anAmicon concentrator, diluted to 10 ml phosphorylate rhodopsin in a light-dependent fashion (15). In with ice-cold buffer A, and reconcentrated to 5 ml. This sample was then an effort to more extensivelycharacterize the activity of GRK5, loaded on a 1-ml Mono S column which wasequilibrated with buffer A containing 100 m NaCl at a flow rate of 1 d m i n . The column was here we have purified the recombinant human GRK5 fromSf9 washed with 10 ml of buffer A containing 200 nm NaCl and theneluted cells and tested the kinase for its reactivity toward several G with a 15-ml linear gradient of 200-500 m NaCl in buffer A. The protein-coupled receptors. The ability of GRK5 to phosphoryfractions were assayed, and the peak fractions were then pooled, con- late rhodopsin was used to assay the kinase activity at the centrated, and stored at -80 "C in the presence of 10% glycerol. This various stages of the purification. A time course of expression procedure yields 20-50 pg of homogeneous GRK5. following infection of Sf9 cells with the pBacPAK-GRK5 virus Receptor Preparations-Rod outer segments (ROS) were prepared from bovineretinas as previously described(25). The ROS were resus- reveals that -48 h postinfection yields the highest expression pended in 50 n m Tris-HC1, pH8.0,5 nm EDTA, 5 M urea (0.5 d r e t i n a ) , ofGRK5 (data not shown). Thus, cells were harvested 48 h sonicated on icefor 4 min, diluted with 2 volumes of 50 m Tris HCl, pH postinfection,homogenized, and centrifuged, andthe high 7.4, and centrifuged at 100,000 x g for 45 min (26). The pellet was speed supernatant fraction was then chromatographed on an washed three times with 50 nm Tris and then resuspended in 50 m S-Sepharose column. Elution ofGRKS from the S-Sepharose Tris-HC1, pH 7.4, aliquoted, and frozen at -80 "C. Urea-treated ROS column with a linear salt gradient yielded -80% recovery of the showed negligible rhodopsinkinase activity and consisted of >90%rhodopsin as determined by Coomassie Blue staining. All procedures were kinase activity (Table I).The pooled S-Sepharosefractions were -60% pure, as judged by SDS-polyacrylamide electrophoresis performed in thedark or under dim red light. The hamster p2AR was expressed in Sf9 cells using the baculovirus and Coomassie Bluestaining (Fig. 1, lune 2 ) . GRK5 migrates at expression system and purified by affinity chromatography on an al- -66 kDa, similar to its predicted molecular mass of 67.7 kDa. prenolol-Sepharose column as described (18,27). The purified receptor The pooled S-Sepharose fractions were then chromatographed was reconstituted into phosphatidylcholine vesicles, pelleted by centrifugation, and resuspended in 20 m Tris-HC1,pH 7.2, 2 EDTA. on a Mono S column. An overall -34-fold purification and 25% The human m2 muscarinic cholinergic receptor (m2 mAChR) was ex- recovery were obtained from the purification (Table I). This pressed in Sf9 cells, purified by affinity chromatography, and reconsti- purified kinase preparation appears to be >95% homogeneous
Characterization Purification and
TABLEI Purification of human recombinant GRK.5 CRK5 was purifieda s described under "Experimental Procedures."Briefly, three 150 x 25-mm dishes of monolayer Sf9 cells were infected with the pBacPAK-GRK5 virus and harvested 48 h postinfection. After homogenization ofthe harvestedcells and several centrifugations, the high speed supernatant was chromatographed on S-Sepharose and Mono S columns. At each step of the purification. GRK5 was assayed using 2-5 p~ rhodopsin and 0.1 n" [y-"PIATP (700-1000 cpdpmol) in buffer containing 20 mM TYis-HCI, pH 7.5, 2 m w EDTA, and 5 mM M&12 a s described under "Experimental Procedures." Protein concentrations were determinedby the method of Bradford (38).The means t S.E. from four separate purifications are presented. Purification Step
1.0 Crude 1.0 S-Sepharose Mono S
100 t 0.7 0.2 f 0.02 0.03 0.01 1
4.5 t 2.0 52 f 4 84 f 4
71 2 8 36 t 3
71 t 8 25 f 2
23 2 2 1.5 2 0.2
23 2 2 34 t 3
FIG.1. Purification of GRK5 from SPg cells. Aliquots from each step of purification were electrophoresedon a 10% SDS-polyacrylamide gel and stained with Coomassie Blue. Lane 1, crude supernatant obtained after high speed centrifugation of pBacPAK-GRK5-infected Sf3 cells (10 pg of protein); lane 2, pooled peak fractions after chromatography on an S-Sepharose column (2 pgof protein); fane 3. pooled peak fractions obtained after chromatography on Mono S column (0.8 pg of protein).
8 18 12 20 From Mg" or Yn" ( mY )
FIG.2. Divalent metal cofactor requirement of CRK5. Rhodop sin was phosphorylatedby GRK5 in the presence of increawng amountn of Mg2*( 0 )or Mn'* (A)a s described under 'Experimental Procedures." Briefly, phosphorylation reactions contained 3 ~w rhodopsin and 15 nM GRK5 in 20 mM Tris-HCI, pH 7.5.2 m w EDTA, 0.1 my ly-:"PMTP (IO00 cpdpmol). and 3-25 m~ M&12 or MnCI, in a t o t a l reaction volume of 20 pl. The reactions were incubated a t 30 "C for 3 min. stopped by the addition of 10 pl of SDS sample buffer, and then electrophoresed on a 10% SDS-polyacrylamide gel and autoradiographed. ',P incorporation in the rhodopsin bands was determinedby excising the bandsfrom the gel and counting in a scintillation counter.
(Fig. 1,lane 3 ) and was used for all further characterizationsof GRK5. The initial characterization of GRK5 involved determining the optimal buffer conditions for the kinase. Similar to most other protein kinases,GRK5 prefers M$+ as thedivalent metal cofactor (Fig. 2). The optimum concentration was found to be 2-3 mM free M e + ,while higher concentrations were inhibitory to the kinase. This is very similar to the M$+ requirements for tion by GRK5 was agonist-dependent. In the absence of the PARK (18,311. The optimum Mn2+concentration was 1-2 m y appropriateagonist, GRK5 phosphorylated the P2AR, m2 however, Mn2+gave only -23% the activity of M$+. The opti- mAChR, and rhodopsin to maximal stoichiometriesof only 0.4, mum pH for GRKS was found to be in the range of 5.5 to 7.5 0.24, and 0.07 moVmol, respectively. The patternof phosphory(data not shown). aAt rhodopsin concentrationof 10 p ~ the , K,,, lation observed with GRK5 was strikingly different from that of GRK5 for ATP was -24 p~ (Table XI). This is similarto the observed with PARK. In experiments performed under identivalue previously measured for PARK (37-60 PM)(18, 31) but cal conditions, the P2AR and m2mAChR were both rapidly and significantly higher than the K , for ATP determined for rho- extensively phosphorylated (4-5 moVmol) while rhodopsin was dopsin kinase (-2 p ~ (32). ) also rapidly phosphorylated but to a stoichiometry of only -2 The ability of GRK5 to phosphorylate different G protein- moVmol under these conditions(Fig. 3R ). While previous studcoupled receptors was studied next. Our previous work had ies have demonstrated that G protein Py subunits significantly established that GRK5 phosphorylates rhodopsin in a light- enhance both the rate and stoichiometry of phosphorylation of dependent manner, albeit notwell as as PARK (15).In addition, each of these receptor substrates by PARK (18, 19), G protein in contrast to the ability of PARK to be activated by G protein P y subunits did not affectGRK5-mediatedphosphorylation P y subunits, rhodopsin phosphorylationby GRKS was not en- (Ref. 15 and data not shown). Previous kinetic studies have hanced by fly subunits (15). In this study, we have compared shown that PARK phosphorylates theP2AR with a K,,, of 0.15the ability of GRK5 and PARK to phosphorylate the p2-adre- 0.25 p~ and rhodopsin with a K,,, of 6-14 p~ (18,31).A similar nergic receptor, the m2 muscarinic cholinergic receptor, and series of kinetic studies demonstrates that GRK5 phosphorhodopsin. As shown in Fig. 3A, the P2AR serves as the best rylates the pZAR with a K , of -0.5 p~ while the K,,, for rhosubstrate for GRK5, being rapidly phosphorylatedto a stoichi- dopsin was -16 p~ (Table XI). O u r initial efforts a t determining ometry of2-2.5molof Pi/mol of receptor. Rhodopsin is also a K , for phosphorylation of the m2 mAChR by GRK5 were rapidly phosphorylated by GRK5, albeit to a stoichiometry of unsuccessful, possibly reflecting a low affinity for the kinase. only -1 moVmol under these conditions. In contrast, the m2 In an effort to assess whether GRK5 phosphorylates a parmAChR appears to be the poorest receptor substrate testedfor ticular consensus amino acid sequence, we studied theability of GRK5 since both the rateof phosphorylation and finalstoichi- GRKS to phosphorylate various peptide substrates. Previous ometry (1-1.5 moVmol) are significantlylower than thatfor the studieshavedemonstratedthat PARK preferentially phosP 2 A R (Fig. 3 A ) . However, for each receptor, the phosphorylaphorylates peptides containingacidic residues amino-terminal
Purification and Characterization of GRKS
TABLE I1 Kinetic parameters for GRK5 The kinetic parameters of GRK5 for ATP were determined by incubating 10 p~ rhodopsin, 60 I"GRK5, and 10-100 p~ [y-32P]ATP(1250 cpdpmol) for 1 min at 30 "C in buffer containing 20 m~ Tris-HC1, pH 7.5,2m~ EDTA, and 5 m~ MgC12.The kinetic parameters for the PzAR were determined by incubating 0.025-0.45p~ reconstituted p z A R and 60 I"GRK5 for 10 min at 30 "C in buffer containing 20 m~ Tris-HC1, pH 7.5,2 m~ EDTA, 5 m~ MgC12, and 0.1 m~ [Y-~~PIATP (1400c p d pmol). Thekinetic parameters for rhodopsin were determined by incubating 2.5-40p~ rhodopsin and 15 rn GRK5 for2 min at 30 "C in buffer containing 20 m~ Tris-HC1, pH 7.5,2 m~ EDTA, 5 m~ MgC12,and 0.1 m~ Iy-32PlATP (1000 cpdpmol). All reactions were stopped by the addition of SDS sample buffer, followed by electrophoresis on a 10% SDS-polyacrylamide gel, and autoradiography. The 32P incorporation was determined by cutting and counting the receptor bands in a scintillation counter. The kinetic parameters for the peptides RRRASASAA and RRRAEASAA were determined by incubating 0.05-5 mM peptide and 60 IIM GRK5 for 30 min at 30 "C in bufer containing 20 mM "isHC1, pH 7.5,2 m~ EDTA, 5 mM MgCl,, and 0.1 m~ [y-32PlATP(1000 cpdpmol). Reactions were stopped by the addition of trichloroacetic acid to a final concentration of 15%. The denatured protein was removed bycentrifugation of the samples for 10 min at 16,000x g,and the supernatants were then spotted on P81 paper and washed six times with 75 m~ phosphoric acid. The 32Pincorporation into the peptide was determined by counting the P81 paper in a scintillation counter. Substrate
ATP P Z A R
Rhodopsin RRRASASAA RRRAEASAA
23.8 * 1.3 0.54 0.15
15.6* 4.0 1500 f 500 1700 2 310
I 2.5 -
1148 2 167 580 * 142 1001 2 180 1.34f 0.4 0.50 f 0.12
to a serine or threonine, while rhodopsin kinase prefers peptides containing acidic residues on the carboxyl-terminalside of a serineor threonine (33).T w o of the best peptide substrates for PARK, RRRDDDDDSAAAand RRREEESGGG, were not phosphorylated by GRK5 when tested at a concentration of 1mM. Similarly, the best identified peptide substrates for rhodopsin kinase, RRRAAAAASEEE and RRREEESEEE, were also not substrates for GRK5. Of all of the peptide substrates tested, RRRASASAA, a poor substrate for PARK (29), was foundto be the best substrate for GRKS (Fig. 4). However, this peptide was a very poor substrate ( K , of -1.5 m ~V,,,, of -1.3 nmol/min/ to the various receptors tested with an mg) compared -100,000-fold lower Vm,/K,,, ratio (Table 11).When the first serine in the peptide RRRASASAA was changed to an acidic residue (either glutamic acid or phosphoserine), the resulting peptides were poorersubstrates (Fig. 4). This was also reflected in an -3-fold decreased Vm,JK,,, ratio for RRRAEASAA phosphorylation by GRK5 ( K , of -1.7 mM, V,,, of -0.5 nmol/min/ mg) compared to RRRASASAA (Table 11). This may be due t o the loss of the first serine inthe peptide or due to the creation of an acidic environment near the remaining serine. The peptide RRRASAAASAA was also a poorer substrate compared to RRRASASAA (Fig. 4). This might suggest that thesecond serine in these peptides prefers being nearer the basic arginine residues at the amino terminus of the peptide. However, LRRASLG (Kemptide), a basic peptide substrate for the CAMPdependent protein kinase, was a poor substrate for GRKS (Fig. 4). While these studies do not delineate any consensus sequence forGRKS phosphorylation, they do suggest that GRK5, in contrast to PARK and rhodopsin kinase, does not phosphorylate serine residues in an acidic environment. In an attemptto further characterize the substrate specificity of GRK5, a number of general protein kinase substrates such as histones, casein, and phosvitin were studied. As shown in Fig. 5, GRK5 is able to phosphorylate the acidic proteins casein and phosvitin. In contrast, GRK5 does not phosphorylate the basic protein histones (data not shown).The level of
Tim ( min )
FIG.3. Time course of phosphorylation of G protein-coupled receptors. 20 rn rhodopsin in the presence of fluorescent room light (O),20 I" p2AR in the presence of 100 p~ isoproterenol (W), and 20 rn m2 mAChRin the presence of 100 p~ carbachol (A)were incubated with 15 rn GRK5 (A) and 15 rn PARK ( B ) in buffer containing 20 m~ Tris-HC1, pH 7.5,2 m~ EDTA, 5 m~ MgC12, and 0.1 m~ [y-3zPlATP (1200cpdpmol). The reactions were incubated at 30 "C, and, at the indicated times, the reactions were stopped by the addition of SDS sample buffer. Samples were then electrophoresed on a 10% SDS-polyacrylamide gel and autoradiographed. 32P incorporation was determined by excising and counting the receptor bands.
phosphorylation of casein and phosvitin by GRK5 is comparable to that observed forseveral of the peptide substrates and again issignificantly lower than thereceptor phosphorylation. Similar results have previously been obtained for PARK (34). The ability of polyanions to modulate the activity of PARK and rhodopsin kinase has been studied previously (18,34,35). Similar studies performed with GRK5 demonstrate that heparin anddextran sulfate are very potent inhibitors of GRK5 with IC50 values of -1 and -0.6 m, respectively (Table111).Heparin and dextran sulfate are also the most potent inhibitors of PARK with IC50values of 150-2800 IIM (18,34); however, these compounds are a t least 150-fold morepotent at inhibiting GRK5 as compared to PARK. Heparin is also a potent inhibitor of casein kinase I1 (IC50of -20-60 IIM) (36), while it isa weak inhibitor of rhodopsin kinase (IC50of -200 PM) (35). Comparable results were obtained when polyglutamic and polyaspartic acid were tested as inhibitors. Polyglutamic and polyaspartic acids were potent inhibitors of GRK5 (IC50values of -25 IIM and -80 m, respectively) (Table III), weaker inhibitors of PARK (IC50 values of -2000 and -1300 I") (341, and poor inhibitors of rhodopsin kinase (ICso values of -700 p~ and -400 p ~ (35). ) These results seem to be consistent with our analysis of the
Characterization Purification and
TABLE I11 Inhibition of GRK5 by protein kinase modulators Phosphorylation reactions containing3-5 rhodopsin and 15-25 m GRK5 in 20 m Tris-HC1, pH 7.5, 2 m EDTA, 5 ~ lMgCl2, l ~ 0.1 w [-p32P]ATP(800-1000 cpdpmol) in the presence of increasing concentrations of various inhibitors were incubated at 30 "C for 3 min. The reactions were stopped by the addition of 10 pl of SDS sample buffer and then electrophoresed on a 10% SDS-polyacrylamide gel. The ICs0 values are presented as the means S.E. from 2-6 independent experiments. Compound
Heparin Dextran sulfate Polyaspartic acid Polyglutamic acid H-7
ICso for GRK5
ICso for @ARK"
1.05 -c 0.06 m 0.58 f 0.02 m 80.0 f 0.8 m 25.0 f 0.3m 170 f 5 1.1~ 5 8 f l m ~
150-1400 nM 150-2800 m 1300 m 2000 m 300 f 42 p~ 79 f 12 mM
NaCl Values were taken from Refs. 18 and 34.
Time ( min )
FIG.4.Peptide phosphorylationby GRK6. Phosphorylation reactions containing1 m peptide, pH 7.0,25 m GRK5, 20 m Tris-HC1, pH of rhodopsin kinase 7 . 5 , 2 m EDTA, 5 m MgC12. and 0.1m [y-32PlATP(800 cpdpmol) in inhibitors of PARK (34), they are activators I1 (36,37). When tested for their ability a total reaction volume of 20 pl were incubatedat 30 "C. At 30 and 60 (35) and casein kinase min, reactions were stoppedby spotting the sample on P81 paper and to modulate the activity of GRK5, spermine, spermidine, and washing six times with75 m phosphoric acid. The peptides used were: polylysine were able to activate GRK5 (Fig. SA). Spermidine 0, R R R A S A S A A ; +, RRRAEASAA; A, RRRASpASAA; of GRK5 to phosphorylate rhodopsin 1.5RRRASAAASAA; V, RRRASpAAASAA; A, LRRASLG (Kemptide). At activated the ability the 60-min time point, the stoichiometries of phosphorylation ranged fold at a n optimum concentration of 1 m, while spermine from 0.7 x for LRRASLG to 5 x moVmol for RRRASASAA. at an optimum concentration of 0.1 activated GRK5 "&fold Under these conditions, nophosphorylation of the peptidesRRRDm.This activation is similar to observed that for casein kinase DDDDSAAA, RRRAAAAASEEE, RRREEESGGG, and RRREEESEEE I1 at the same physiological concentrations of these compounds was detected. (37). Polylysine was found to be the most potent activator of GRK5 resulting in an overall -2.6-fold activation a t a n optiI I I , 1 I mumconcentration of 10 (Fig. 6A).In comparison,0.1 polylysine results in an -2-fold activation of casein kinase I1 (36), while 2 mM polylysine promotes only a n -0.4-fold activation of rhodopsin kinase (35).Although polycations are weak inhibitors of PARK,a t low concentrations they are also effective at reversing the abilityof heparin to inhibit PARK (34). Similarly, low concentrations of the various polycations were also effective at reversing the heparin inhibition of GRK5 (Fig. 6B1. At higher concentrations, the polycations activated GRK5 to a level similar to that observed in Fig. 6.4 while at still higher concentrations the polycations were inhibitory to GRKS (Fig. 6B).
DISCUSSION G protein-coupled receptorkinases play an important role in FIG.5. Phosphorylation of casein and phosvitin by GRK6. regulating receptor function by their unique ability to specifiPhosphorylation reactions were performedas described under "Experi- cally phosphorylate the activated form of various G proteinmental Procedures" using 25 pgof casein (0)or phosvitin (A)as the number of studies substrate and 20 m GRK5 in 20 m Tris-HC1, pH7.5,2 m EDTA, 5 nw coupled receptors (5). While a significant MgCl,, and 0.1IMI [y32PlATP(800 cpdpmol) in a total reaction volume have extensively characterized the properties ofPARK and of 20 pl. Reactions were incubated at 30 "C and were stopped at the rhodopsin kinase, very little information is currently available indicated timesby the addition of 10 pl of SDS sample buffer. Samples on IT11, GRK5, and GRK6, the three most recently identified wereelectrophoresed on a 10% polyacrylamidegel andautoradioGRKs. To this end, this work focused on more extensive chargraphed, andthe szPincorporation was determined. At the 60-min time point, the stoichiometries of phosphorylation were6 x lo4 moVmol for acterization of GRK5 utilizing the baculovirus expression system to overexpress and purify recombinant human GRK5. PumoVmolforphosvitin.Under similarconditions, caseinand 7.4 x no phosphorylation of histones was detected. rified GRK5 exhibited several properties that were similar to many other protein kinases including a preferential requirepeptide substrates,i.e. while PARK and rhodopsin kinase pref- ment for Mg2' as the divalent cation, a pH optimum of 5.5-7.5, erentially phosphorylate serine residues in an acidic environ- and a K,,, for ATP of -24 w. ment, GRK5 appears to be inhibited by an acidic environment. One of the crucial features of GRKs is their ability to specifiTwo other compounds were also tested for their ability to in- cally phosphorylate different G protein-coupled receptorsin an hibit GRK5. H-7, a potent inhibitorof protein kinaseC and the agonist-dependent manner. In the past few years, thep 2 A R and CAMP-and cGMP-dependent protein kinases, was found to be the m2 mAChR have been overexpressed in the baculovirus a weak inhibitor of GRK5 (ICso of -170 w) (Table 111). Simi- expression system, purified and reconstituted in phospholipid larly, NaCl also appears to inhibit the activity of GRK5 with a n vesicles, and shownto be functionally active.The availability of ICboof -60 mM, similar tothe inhibition of PARK by NaCl(l8, such purified and reconstituted G protein-coupled receptors 31). greatly facilitates detailed studies of the specificities of the Since polyanions were potent inhibitors of GRK5, the effect GRKs. Using thesepurified receptors, PARK has been shownto of polycations on GRKS activity was also studied.While poly- phosphorylate both the agonist-occupied P 2 A R and m2mAChR cations sucha s spermine, spermidine, and polylysine are weak in vitro (18, 19). Moreover, this phosphorylation is greatly ennmo ( min )
Purification and Characterization of GRK5
additional sites. At the present time, it is not clear whether the lower stoichiometry of receptor phosphorylation by GRK5 compared to PARK reflects the possibility that it recognizes a A unique setof phosphorylation sites on the receptors or simply a reduced binding of the substrates by GRK5 (e.g. GRK5 has a 2-%fold lower affinity for the P z A R as compared to PARK). to the p z A R , m2 While the present study was constrained mAChR, and rhodopsin, future substrate specificity studies will need to include a wider range of potential substrates and involve both in vivo and in vitro analysis. Previous peptidephosphorylation studies haverevealed that PARK preferentially phosphorylates Serrrhr residues in an acidic environment (either with acidic amino acids or a phosphoserine amino-terminal to the Ser/Thr), while rhodopsin kinase phosphorylates peptides with acidic amino acids carboxylterminal to theSer/"hr (29, 33). In general, peptides arevery poor substrates for PARK and rhodopsin kinase compared to the P z A R and rhodopsin. Anumber of peptides tested here were also found to be substrates for GRK5. However, in contrast to PARK and rhodopsin kinase, GRK5 appears to prefer peptides that do not contain acidic amino acids in vicinity the of a serine. 250 that the The kinetic parameters for the peptides tested indicate rate of presence of acidic residues predominantly decreases the 200 phosphorylation rather than the affinity of GRK5 for these E. peptides. The peptide studies contrast with the results ob150 The tained using several general protein kinase substrates. E acidic proteins casein and phosvitin were found to be weakly a #? 100 phosphorylated by GRK5 while the basic histones were not phosphorylated by GRK5. The results with these proteins are, 50 in fact, very similar to those obtainedfor PARK (34). Overall, the results suggest that GRK5 has a complex substrate speciI , v ficity that is not defined simply by the presence of acidic or 0 -7 -6 -5 -4 -3 -2 basic residues in thevicinity of a serine or threonine. - Log [Poiycatlon] M A number of protein kinases such as casein kinase11, PARK, FIG.6. A, activation of GRK5 by polycations. 5 w rhodopsin was and rhodopsin kinase havepreviously been shownto be moduphosphorylated by 20 m GRK5 in the absence (a)or presence of 10 lated by polyions (18,3637). The modulation of GRK5 activity polylysine ( ), 100 p d spermine (A),or 1m spermidine (W) in buffer by polyions reveals several interestingdifferences comparedto containing 20 nm Tris-HC1, pH 7.5, 2 nm EDTA, 5 m MgC12,and 0.1 nm [yS2P]ATP (1000 cpdpmol) in a total reaction volumeof 20 4. PARK and rhodopsin kinase. Although heparin has been idenReactions were incubated at 30 "C and stopped at the times indicatedby tified as the most potent inhibitor ofPARK with an IC50 of the addition of 10 pl of SDS sample buffer. Samples were then electro- 150-1400 m (34), heparin is >150-fold more potent at inhibitphoresed on a 10%polyacrylamide gel,and, following autoradiography, ing GRK5 (IC50of -1 m). Heparin is also a potent inhibitorof S2Pincorporation was determined by excising and countingthe rhodop- casein kinaseI1 (IC50of -20-60 m) (36); however, it is only a sin bands. B, reversal of heparin inhibition by polycations. 5 w rhodopsin was phosphorylated by 20 m GRK5 in the presence of 1 I ~ M weak inhibitorof rhodopsin kinase (IC50of -200 p"35). These used as a "speheparin and the indicated concentrations of polylysine ( 0 ,spermine results demonstratethat heparin should not be (V),or spermidine(0) in buffer containing20 nm Tris-HC1, pH7.5,2 m cific" inhibitor of PARK (7). Several other polyanions such as EDTA, 5 nm MgC12, and 0.1 nm [-pS2P1ATP(1000cpdpmol) in a total polyaspartic andpolyglutamic acid were alsofound to be much reaction volume of 20 pl. Reactions were incubated at 30 "C for 5 min and stopped by the addition of 10 pl of SDS sample buffer. Samples were more potent inhibitors (16-80-fold) of GRK5 as compared to a tremendousvariance then electrophoresed on a 10% polyacrylamide gel, and, following au- PARK. Thesestudiesdemonstrate toradiography, 32Pincorporation was determined by excising and count- (-200,000-fold) in the abilityof polyanions to inhibit different ing the rhodopsinbands. 100% activityrepresentsrhodopsinphosmembers of the GRK family and suggestthat specific inhibitors phorylation by GRK5 in the absence of heparin. The GRK5 activity in the presence of 1m heparin and in the absence of any activator was of these kinases could be engineered. Previous studies have shownthat polyamines like spermine 49%. and spermidine activate casein kinase I1 2-%fold and may hanced in the presence of G protein Py subunits (18, 19). serve as physiological modulators of this kinase inreticulocytes P A R K 2 , which has -85% amino acid identity with P A R K , (36,37). GRK5 is also significantly activated by the polycations shows a similar substrate specificity and activation by G pro- polylysine, spermine, and spermidine. Furthermore, the extent tein Py subunits in vitro (18, 19). Since virtually nothing was of activation of GRK5 by these polycations appears to be prosince optimal concentrationsof sperknown about the substratespecificity of GRK5, we compared portional to their charge, the abilityof GRK5 and PARK to phosphorylate the& A R , m2 midine(three positive charges)activates GRK5 -1.5-f0ld, mAChR, and rhodopsin. It is clear that GRK5 is significantly while spermine (four positive charges) activates GRK5 -1.8less active towardall three receptor substrates as compared to fold. Polylysine, a 14-16-amino acid peptide,is the most potent PARK (Fig. 3). This ismost evident in both the rate and extent activator of GRK5 among thecompounds tested, promoting an of phosphorylation of the m2mAChR by GRK5 (compare Fig. 3, -2.6-fold activation at an optimal concentrationof 10 w. RhoA and B ) . In addition, unlike PARK, GRK5 phosphorylation of dopsin kinase has also been shown to be weakly activated by the P z A R and m2mAChR appears tobe biphasic.This suggests these compounds (35). While polycations do not activatePARK, that there may be a primary site or sites that are rapidly at lower concentrations they are able to partially reverse the phosphorylated by GRK5 followed by slower phosphorylationof inhibition of PARK by heparin, most likely by directly binding
Purification and Characterization of GRK5
to heparin andthereby preventing its inhibition (34).A similar 12. Lorenz, W., Inglese, J., Palczewski,K, Onorato, J. J., Caron, M. G . & Lefkowitz, R.J. (1991) Pmc. Natl. Acad. Sci. U. S.A. 88,87158719 phenomenon was observed forGRK5, where spermine, spermi- 13. Benovic, J. L., Onorato, J. J., Arriza,J. L., Stone, W. C., Lohse, M., Jenkins, N. dine, and polylysine effectively reversed the ability of heparin A,, Gilbert, D. J., Copeland, N.G., Camn. M. G . & Lefkowitz, R.J. (1991) J. Biol. Chem. 286, 14939-14946 to inhibit GRK5. It is thusconceivable that in vivo polyamines Ambrose, C., James, M., Barnes, G . ,L h , C., Bates,G., Altherr, M., Duyao, M., or polycationic surfaces may play a role as physiological modu- 14. Groot, N.,Church, D., Wasmuth, J. J., Lehrach, H., Housman,D., Buckler, lators of GRK5. A, Gusella, J. F. & MacDonald, M. E. (1992)Humnn Mol. Genet.1,697-703 In conclusion, these studies have demonstrated that while 15. Kunapuli, P. & Benovic, J. L. (1993)Prcc. Natl. Acad. Sci. U. S.A. 90,55885592 GRK5 has the ability to phosphorylate multiple G protein- 16. Benovic, J. L. & Gomez, J. (1993) J. Biol. Chem. 268, 19521-19527 coupled receptors, it also has several unique properties com- 17. Cassill, J.A.,Whitney, M., Joazeim, A. P., Becker,A. & Zuker, C. S . (1991)Proc. Natl. Acad. Sci. U. S. A. 88, 11067-11070 pared to PARK and rhodopsin kinase. These include its preferKim, C. M., Dion, S . B., Onorato, J. J. & Benovic, J. L. (1993)Receptor 3,39-55 encefor phosphorylating nonacidic peptides as well as the 18. 19. Richardson, R. M., Kim, C. M., Benovic, J. L. & Hosey, M. M. (1993) J. Biol. ability of polyanions to serve as potent inhibitors. Future studChem. 268, 13650-13656 ies will involve more rigorouslyaddressing the substratespeci- 20. Kwatra, M. M., Schwinn, D. A,, Schreurs, J., Blank,J. L., Kim, C. M., Benovic, J. L., Krause, J. E., Caron, M. G. & Lefkowitz, R.J. (1993) J. Biol. Chem. ficity of GRK5 as well as potential modes of regulation. The 268,9161-9164 identification of heparin and dextran sulfate as extremely po- 21. Benovic, J. L., Regan, J. W., Mataui, H., Mayor, F., Jr., Cotecchia, S . , LeebLundberg, L. M. F., Camn. M. G. & Lefkowitz, R. J. (1987) J. Bwl. Chem. tent inhibitors of GRK5 should prove useful in further eluci262,17251-17253 dating the role of GRK5 in theregulation of G protein-coupled 22. Strasser, R. H., Benovic, J. L., Caron, M. G . & Lefkowitz, R. J. (1986) Proc. receptors. Natl. Acad. Sci. U. S. A. 83, 6362-6366 Acknowledgment-We thank Judy Rasienski for preparing the purified and reconstituted m2 mAChFt. REFERENCES 1. Dohlman, H. G.,Thorner, J., Caron,M. G . & Lefkowitz, R.J. (1991)Annu. Rev. Biochem. Bo, 653-688 2. O'Dowd, B. F., Lefkowitz, R. J. & Caron, M. G. (1989) Annu. Rev. Neurosci. 2, 6743 3. Gomez, J. & Benovic, J. L. (1992) in Molecular Biology of Receptors and Zhmsporters: Receptors (Friedlander, M. & Muekler, M., eds) pp. 1-34, Academic Press, San Diego 4. Hargrave, P. A. & McDowell, J. H. (1992) in Molecular Biology ofReceptors and "sporterst Receptors (Friedlander, M. & Muekler, M., eds) pp. 49-98, Academic Press, San Diego 5. Palczewski, K. & Benovic, J. L. (1991) %rids Biochem. Sci. 16,387-391 6. Benovic. J. L.. Strasser. R. H.. Caron. M. G . & Lefkowitz. R. J. (1986) Proc. Natl.'&ad.'Sci. U. A. 2797-2801 7. Lohse, M. J . , Benovic, J. L., Caron. M. G . & Lefkowitz, R. J. (1990) J. Biol. Chem. 266,32023209 8. Kuhn, H. & Dreyer, W. J. (1972) FEES Lett. 20, 1-6 9. Bownds, D., Dawes, J., Miller, J. & Stahlman, M. (1972) Nature 237, 125-127 10. Lohse, M. J.,Benovic, J. L., Codina, J., Caron,M. G. & Lefkowitz, R.J. (1990) Science 248, 1547-1550 11. Benovic, J. L., DeBlasi, A., Stone, W. C., Carun, M. G. & Lefkowitz, R.J. (1989) Science 246,235-246 ~~
23. Mayor, F. M., Jr., Benovic, J. L., Camn, M. G.& Lefkowitz, R.J. (1987) J. Biol. Chem. 282,64684471 24. Dawson, T.M., Arriza, J. L., Jaworsky, D. E., Borisy, F. F., Attramadal, H., Lefkowitz, R. J. & Ronnett, G . V. (1993) Science 250,825429 25. Shiehi, H. & Somers, R.L. (1978) J. Biol. Chem. 263, 7040-7046 Wilden, U.& Kuhn, H. (1982) Biochemistry 21,30143022 26. 27. Benovic, J. L., Shorr, R. G., Caron, M. G . & Lefkowitz, R. J. (1984) Biochemistry 23,4510-4518 28. Cook, P. F., Neville, M.E., Vrana, K E., Hartl, F. T. & Roskoski, R.,Jr. (1982) Biochemistry 21,5794-5799 29. Chen, C.-Y., Dion, S . B., Kim, C. M. & Benovic, J. L. (1993)J. Biol. Chem. 268, 7825-7831 30. Laemmli, U.K. (1970) Nature 227, 680-685 31. Benovic, J. L., Mayor, F., Jr., Staniszewski,C., Letkowitz, R. J. & Caron, M. G . (1987) J. Biol. Chem. 282,9026-9032 32. Palczewski, K., McDowell, J. H. & Hargrave, P. A. (1988) J. Bwl. Chem. 263, 14067-14073 33. Onorato, J. J., Palczewski, K., Regan, J., Caron, M. G., Lefkowitz, R. J. & Benovic, J. L. (1991)Biochemistry SO, 511-125 34. Benovic, J. L., Stone, W. C., Caron, M. G . & Lefkowitz, R. J. (1989) J. B i d . Chem. 284,67074710 35. Palczewski, K., Arendt, A., McDowell, H. J. & Hargrave, P. A. (1989) Biochemistry 28,8764-8770 36. Lin, W. J., n a z o n , P. T. & Traugh, J. A. (1991) J. Biol. Chem. 286,5664-5669 37. Hathaway, G . M. & Traugh, J. A. (1984) J. Biol. Chem. 259, 7011-7015 38. Bradford, M. M. (1976) Anal. Biochem. 72,248-254