Substrate Specificity of a Multifunctional Calmodulin-dependent ...

Vol. 260, No. 27, Issue of November 25, pp, 14471-14476,1985 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

0 1985by The American Society of Biological Chemists, Inc.

Substrate Specificity of a Multifunctional Calmodulin-dependent Protein Kinase* (Received for publication, January 29, 1985)

Richard B. Pearson$, James R. Woodgettjjq, Philip Cohenjj, and Bruce E. KempS From the $Department of Medicine, University of Melbourne, Repatriation General Hospital, Heidelberg, Victoria 3081, Australia and the SDepartment of Biochemistry, University of Dundee, Dundee, Tayside, Scotland, DD 14HN

The substrate specificity of the multifunctional cal- dant in thebrain, termed synapsin-I kinase-I1 (3) or calmodmodulin-dependent protein kinase from skeletal mus- ulin-dependent protein kinase-I1 (4),has many properties in cle has been studied using a series of synthetic peptide common with an enzyme found in skeletal muscle that cataanalogs. The enzyme phosphorylated a synthetic pep- lyzes the phosphorylation of glycogensynthase at serine 7 (5). tide corresponding to theNHz-terminal 10 residues of This latter enzyme has been termed the multifunctional calglycogen synthase, Pro-Leu-Ser-Arg-Thr-Leu-Ser-modulin-dependentprotein kinase to distinguish it from phosVal-Ser-Ser-NH2, stoichiometrically at Ser-7,the phorylase b kinase and themyosin light chain kinases and to same residue phosphorylated in the parent protein. The highlight its broader specificity (4).In view of the possibility , of that the multifunctional calmodulin-dependent protein kisynthetic peptide was phosphorylated with a V 12.5 gmol*min-’*mg-’ and an apparent K , of 7.5 p M nase(s) may play a pivotal role in the actions of calcium in a compared to values of 1.2 pmol*min-’*mg-’ and 3.1 wide variety of tissues, we have sought to characterize its p ~ respectively, , for glycogen synthase. Similarly, a synthetic peptide corresponding to the NHz-terminal substrate specificity using synthetic peptides. 23 residues of smooth muscle myosin light chain was EXPERIMENTAL PROCEDURES~ readily phosphorylated on Ser-19 with a K , of 4 p~ and a V,,, of 5.4 pmol*min”*mg-l. The importance of RESULTS the arginine 3 residues NH2-terminal to the phosphorylated serine in each of these peptides was evident Phosphorylation of Synthetic Peptide-The peptide Profrom experiments in which this arginine was substi- Leu-Ser-Arg-Thr-Leu-Ser-Val-Ser-Ser-amide, corresponding tuted by either leucine or alanine, as well as from to the NHz-terminal 10 residues of skeletal muscle glycogen experiments in which its position in the myosin light synthase, was readily phosphorylated by the multifunctional chain sequence was varied. Positioning arginine 16 at calmodulin-dependent protein kinase (Fig. 1).The stoichiresidues 14 or 17 abolished phosphorylation, while location at residue 15 not only decreased V,, 14-fold ometry of phosphorylation of the synthetic peptide was folbut switched the major site of phosphorylation from lowed using reverse-phase high pressure liquid chromatograSer-19 to Thr-18. It is concluded that the sequence phy (Fig. 2). The phospho and dephospho forms of the peptide Arg-X-Y-Ser(Thr) represents theminimum specificity eluted at 15.8 and 17 min, respectively. The 32P-phosphoryldeterminant for the multifunctional calmodulin-de- ated synthetic peptide electrophoresed as a single spot on pendent protein kinases. Studies with various syn- high voltagepaper electrophoresis.When subjected to partial thetic peptide substrates and their analogs revealed acid hydrolysis (5.7 M HC1, 110“C, 2 h) and high voltage that thespecificity determinants of the multifunctional paper electrophoresis at pH1.9, 95% of the 32Pradioactivity calmodulin-dependent protein kinasewere distinct was associated with phosphoserine and 5%with phosphothrefrom several other “arginine-requiring“ protein ki- onine. The results suggested that the synthetic peptide was nases. phosphorylated stoichiometricallyon a single serine residue. Peptide Phosphorylation Site-Since the glycogen synthase peptide (residues 1-10) contained 4 serine residues, it was necessary to determine which of these was phosphorylated by It is now recognized that many tissues contain one or more the multifunctional calmoddin-dependent protein kinase. calcium/calmodulin-dependentprotein kinases. The most The siteof phosphorylation was assigned using two independthoroughly studied examples of these enzymes include phos- ent approaches. The 32P-phosphorylatedpeptide was subphorylase b kinase (1)and themyosin light chain kinases (2). jected to tryptic digestion and purified by highpressure liquid Both have relatively narrow substrate specificities. Another chromatography (Fig. 3). The 32P-labeledtryptic peptide had calcium/calmodulin-dependent protein kinase that is abun- a composition consistent with the expected peptide Thr-LeuSer-Val-Ser-Ser (Ser, 2.84; Leu, 1.17; Thr, 0.90; and Val, l.lO), * This research was supported bygrants from the National Health and Medical Research Council of Australia and the HeartFoundation of Australia (to B.E. K.) and by the British Medical Research Council, British Diabetic Association, and Royal Society (to P. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This articlemusttherefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ll Recipient of a postgraduate research studentship from the British Medical Research Council. Present address: the Salk Institute, La Jolla, CA.

Portions of this paper (including “Experimental Procedures,” Figs. 1-7, and Tables 1-3) are presented in miniprint at the end of this paper. The abbreviation used is: HPLC, high pressure liquid chromatography. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 85M-279, cite the authors, and include a check or money order for $3.60 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

14471

14472

Calmodulin-dependent Protein Kinase Substrate Specificity

indicating that Ser-3 was not phosphorylated. The 32P-labeled tryptic peptide was subjected to three cycles of manual Edman degradation with electrophoresis after each cycle (13). After the third step of Edman degradation, 32Pradioactivity no longer migrated with the cationic peptide but instead migrated. as ananion -inthe position expected for inorganic phosphate (Fig. 4). The loss of radioactivity from the tryptic phosphopeptide after the third cycle (Thr-Leu-Se;-Val-Ser-Ser-NH2) indicated that Ser-7 was the phosphorylation sitein the glycogen synthase peptide. The phosphorylation site was independently assigned using enzymic digestion of the 32P-phosphorylated peptide with thermolysin. The thermolytic peptides were separated using high voltage paper electrophoresis (pH 3.5, 1000 V, 45 min). The 32P-labeledpeptides were identified by autoradiography, eluted with 5.5 M acetic acid, and subjected to amino acid analysis. Two 32P-labeledthermolytic peptides were obtained, one with a composition consistent with the structure Leu-Ser (Table 3) and thesecond with the structureLeu-Ser-Val-SerSer (contaminated with some aspartic acid and proline). The results established that Ser-7 was the site of phosphorylation in thesynthetic peptide, the same site phosphorylated in the parent protein ( 5 ) . Comparison of Peptide and Protein Substrates-It was of interest to compare the kinetics of phosphorylation of the glycogen synthase peptide (residues 1-10) with the parent protein. Phosphorylation of the peptide was linear with respect to time (upto 7 min) andenzyme concentration (Fig. 5 , A and B ) . Peptide phosphorylation was dependent on peptide concentration, and the apparent K, and V,, values were approximately 7.5 p~ and 12.5 pmol. min-'. mg-' (Fig. 6). The apparent K, of the peptide was approximately 2-fold higher than for glycogen synthase, while the V,,, was approximately 10-fold higher (Table 4). Theseresults indicate thatthe glycogen synthase peptide is phosphorylated with comparable kinetic constants to the parent protein. Phosphorylation of the glycogen synthase peptide was dependent onthe presence of calmodulin with a Koaof approximately 80 nM under the conditions tested (Fig. 7). Since the multifunctional calmodulin-dependent protein kinase has been shown to phosphorylate a variety of proteins including smooth muscle myosin light chains, it was of interest to evaluate the capacity of synthetic peptide analogs of this protein to act assubstrates for the enzyme. The smooth muscle myosin light chains were phosphorylated with an apparent K , of 41 ptM and V,,, of 3.4 pmol. min" .mg-'. The synthetic peptide analog of the myosin light chains corresponding to residues 1-23 was phosphorylated with an apparent K,,, of 4 p~ and a V,, of 5.4 pmol. min-' .mg-' (Table 4). The Role of Arginine-Glycogen synthase is also phospho-

rylated on Ser-7 by phosphorylase b kinase (14), and studies with synthetic peptides have indicated that Arg-4 acts as an important specificity determinant for this enzyme (15, 16). Wewere therefore prompted to test the influence of this residue on the specificity of the multifunctional calmodulindependent protein kinase. When Arg-4 was substituted by leucine, the rate of phosphorylation of the glycogen synthase peptide was negligible (Table 5). The role of arginine was further investigated using synthetic peptide analogs of the smooth muscle myosin light chain. Substitution of Arg-16 in this peptide with alanine abolished peptide phosphorylation (Table 5, peptides 5 and 6). The results obtained with peptide analogs of both protein substrates indicated that an arginine NHz-terminal to the phosphorylatable serine was important for phosphorylation by the multifunctional calmodulin-dependent protein kinase. Although the Ala- and Leu-substituted peptides were not substrates, the Lys-4 analog, Pro-Leu-Ser-Lys-Thr-Leu-SerVal-Ala-Ala,of glycogen synthase was readily phosphorylated. The apparent K, was 184 p ~ approximately , 10-fold higher than the parent peptide (Table 5, peptides 17 and 18), and the Vmaxwas reduced by 33%. The Lys-4 analog was phosphorylated exclusively on Ser-7 (results notshown). These results indicate that a positive charge is essential for phosphorylation and that the guanidino group is necessary for optimum kinetics. Influence of Arginine Location-The importance of the location of the essential argininewas investigated further using a series of synthetic peptides corresponding to the smooth muscle myosin light chain sequence, residues 11-23. The effect of placing arginine at positions 14-17 is shown in Table 6. Optimum kinetics of peptide phosphorylation were obtained when arginine was located at residue 16. When arginine was moved to position 15, the kinetics of phosphorylation were much poorer, the major effect being a 14-fold decrease in Vmu. Placing arginine either at position 17 or 14 abolished peptide phosphorylation (Table 6). These resultsindicate that the essential arginine has an optimal location at the third residue on the NH2-terminal side of the phosphorylated serine (Ser-19). The multifunctional calmodulin-dependent protein kinase was shown to phosphorylate Ser-19 in themyosin light chain peptide Lys-Lys-Arg-Ala-Ala-Arg-Ala-Thr-Ser-AsnVal-Phe-Ala (Table 7). This peptide contained only a single threonine and a single serine. Partial acid hydrolysis (5.7 M HC1, 110 "C, 2 h) of the phosphorylated peptide indicated that 99% of the phosphate was esterified to Ser-19 and only 1%to Thr-18. The peptide analog Lys-Lys-Arg-Ala-Arg-AlaAla-Thr-Ser-Asn-Val-Phe-Ala with Arg at position 15 was phosphorylated predominantly on Thr-18(73%)and only 27% on Ser-19. The substantial switch in specificity between

TABLE4 Kinetics of phosphorylation of protein and peptide substrates Peptide and protein phosphorylation was determined as described under "Experimental Procedures." Kinetic constants (fS.E.) were estimated by fitting the data to the Michaelis-Menten equation using the method of least squares. Substrate K, VPM pmol .min" .mg" Glycogen synthase 3.1 f 0.2 1.17 f 0.04 12.5 -C 0.4 7.5 f 0.6 Peptide 1, glycogen synthase 1-10 41 f 4.3 3.4 f 0.14 Myosin light chains (chicken gizzard) 5.38 f 0.18 4.1 f 0.4 Peptide 3, myosin light chain 1-23 (chicken gizzard)

TABLE5 Influence of arginine on the kinetics of phosphorylation Peptide phosphorylation was determined as described under "Experimental Procedures."Kinetic constants (rtS.E.) were estimated by fitting the data to the Michaelis-Menten equation using the method of least squares. Peptide Sequence K, V,, p~

pmol. min" .mg"

7.5 f 0.6 12.5 f 0.4 PLSRTLSVSS Negligible PLSLTLSVSS" 14.7 rf: 1.6 12.4 f 0.4 PLSRTLSVAA 184 f 23 8.2 f 0.5 18 PLSKTLSVAA 5 KKRAARATSNVFA 45.6 f 1.1 2.58 f 0.02 6 KKRAAAATSNVEAb Negligible Less than 0.1 pmol. min-'. mg-l at 2 mM peptide. 0.04 pmol .min" .mg-' at 2 mM. 1 2 17

Calmodulin-dependent Protein Kinase Substrate Specificity TABLE6 Influence of arginine locution on the kinetics of peptide phosphorylation Peptide phosphorylation was determinedas described under “Experimental Procedures.” Kinetic constants (fS.E.1 were estimated by fitting the data to the Michaelis-Mentenequation using the method of least squares. Sequence

equence

Peptide

K,

V,.

p~

pmol. min”. mg”

KKRAARATSNVFA 45.6 f 1.1 2.58 f 0.02 KKRRAAATSNVFA“ Negligible 8 KKRARAATSNVFA 53.9 f 5.1 0.18 f 0.00 9 KKRAAARTSNVFA“ Negligible a No phosphorylation detectedat 2 mM peptide. TABLE7 Effect of bask residue location on the site of phosphorylation Peptides were phosphorylated as described under “Experimental Procedures.” The phosphorylatedpeptides were subjectto partial acid hydrolysis and high voltage paper electrophoresis at pH 1.9 as described under “Experimental Procedures.’’

equence

Ser-lS(P) Thr-l8(P) %

%

KKRPQRATSNVFS 2 98 ‘KKRAARATSNVFA 1 99 KKRAAAATSNVFA“ 9 91 KKRRAAATSNVFA“ 0 0 KKRARAATSNVFA 73 27 KKRAAARTSNVFA” 0 0 a Peptides not phosphorylatedor poorly phosphorylated. 4 5 6 7 8 9

TABLE 8 Influence of other features on the kinetics of peptide phosphorylation Peptide phosphorylationwas determined as described under “Experimental Procedures.” Kinetic constants (&%E.)were estimatedby fitting the data to the Michaelis-Mentenequation using the method of least squares. Peptide

4 KKRPQRATSNVFS 5 KKRAARATSNVFA RPQRATSNVFS 10 RATSNVFS 11 Negligible RRATSNVFA” 12 Negligible 13 RKASGPPV” 17 PLSRTLSVAA 19 PLSRRLSVAA 20 PLRRTLSVAA Less than0.1 pmol .min” .mg-’

TABLE9 Comparison of model synthetic substrates, Peptide phosphorylation was determinedas described under “Experimental Procedures.” Kineticconstants (fS.E.) were estimated by fitting the data to the Michaelis-Menten equation using the method of least squares. Peptide K, Vp~

5 7

Sequence Peptide

14473

K,

V-

p~

pmol. min“ .mg”

13.3 f 0.6 45.6 f 1.1 244 f 39 487 f 58

6.54 f 0.14 2.58 f 0.02 1.32 f 0.04 1.32 f 0.04

14.7 f 1.6 37.1 f 0.2 3.5 f 0.2 at 2.8 mM.

12.4 f 0.4 9.6 f 0.2 11.3 f 0.3

pmol.min-l.mg-l

14 LRRASLG 468 f 85 0.40 f 0.03 15 KAKQISVRGLAG 50 f 12 0.07 f 0.00 4 KKRPQRATSNVFS 13.3 f 0.6 6.54 f 0.14 16 VKRISGL“ Negligible 1 PLSRTLSVSS 7.5 f 0.6 12.5 f 0.4 “No detectable phosphorylation at peptide concentration up to 2.24 mM.

that lack of phosphorylation of the peptide Arg-Arg-Ala-ThrSer-Asn-Val-Phe-Alawas due to the presence of 2 adjacent arginine residues. This idea wastested directly using glycogen synthase peptide analogs containing adjacent arginines. The and Propeptides Pro-Leu-Ser-Arg-Arg-Leu-Ser-Val-Ala-Ala Leu-Arg-Arg-Thr-Leu-Ser-Val-Ala-Ala were surprising good substrates with K, values of 37 and 3.5 p ~ respectively, , and Vmx values comparable to the parentpeptide (Table 8, peptides 17, 19, and 20). In both of these peptide analogs, phosphorylation occurred exclusivelyat Ser-7 (results not shown). Thus, it is clear that thepresence of adjacent arginine residues alone does not account for, the failure of Arg-Arg-Ala-ThrSer-Asn-Val-Phe-Alato act as a substrate. The substitution of Pro-Gln with Ala-Ala resulted in a3.4-fold increase in the apparent K , and a60% decreasein the V,,,. While relatively modest, these changes nevertheless indicate that neutral residues can also influence the kinetics of peptide phosphorylation by the multifunctional calmodulin-dependent protein k’mase. Comparison with OtherProtein Kinase Substrates-In view of the requirement of the multifunctional calmodulin-dependent protein kinase for arginine, it was of interet to test model substrates for other protein kinases containing this residue. The CAMP-dependent protein kinase peptide substrate Leu-Arg-Arg-Ala-Ser-Leu-Gly was a poor substrate for the multifunctional calmodulin-dependentprotein kinase with a K,,, of460 p~ and a V,,, 0.4 pmol. min-l. mg” (Table 9). Similarly,the peptide Arg-Lys-Ala-Ser-Gly-Pro-Pro-Val, corresponding to thesite on histone H1 phosphorylated by cyclic AMP-dependent protein kinase, was not a substrate. The phosphorylase b-related peptide, Lys-Ala-Lys-Gln-Ile-SerVal-Arg-Gly-Leu-Ala-Gly, was also a poor substrate with a very low V,,, of 0.07 pmol.min-le mg-I and a K, of 50 p ~ . The histone H4 kinase peptide substrate, Val-Lys-Arg-IleSer-Gly-Leu,was not phosphorylated. The myosin light chain kinase peptide substrate, Lys-Lys-Arg-Pro-Gln-Arg-Ala-ThrSer-Asn-Val-Phe-Ser,was the only modelprotein kinase substrate phosphorylated with kinetic parameters similar to those obtained with the glycogen synthase peptide. These results highlight the different substrate specificities of these protein kinases, even though they all require arginine as animportant specificity determinant.

Ser-19 and Thr-18 further highlights the importance of the location of the basic residue determinant. Influence of Other Structural Features-Investigation of the capacity of the multifunctional calmodulin-dependentprotein kinase to phosphorylate a variety of synthetic peptides revealed that there were other structural requirements besides that for arginine. The myosin light chain peptide Lys-LysArg-Pro-Gln-Arg-Ala-Thr-Ser-Asn-Val-Phe-Ser was phosphorylated with an apparent K,,, 18-fold lowerthan theanalog peptide Arg-Pro-Gln-Arg-Ala-Thr-Ser-Asn-Val-Phe-Ala (TaDISCUSSION ble 8).These results suggest that residues besides the essential Arg-16 can influence the kinetics of peptide phosphorylation. In this study, we have demonstrated that the synthetic This is further supported by the finding that thepeptide Arg- decapeptide Pro-Leu-Ser-Arg-Thr-Leu-Ser-Val-Ser-SerArg-Ala-Thr-Ser-Asn-Val-Phe-Ala was not a substrate, NH2, corresponding to the NHz-terminal 10 residues of glywhereas the peptide Arg-Pro-Gln-Arg-Ala-Thr-Ser-Asn-Valcogen synthase, is an excellent substrate for the multifuncPhe-Ser, containing 2 arginines, was a better substrate(Table tional calmodulin-dependent protein kinase. The V-/K, 8). The peptide Arg-Ala-Thr-Ser-Asn-Val-Phe-Ser contain- value obtained with this peptide as substrate was 1.67 coming a single arginine residue was phosphorylated, suggesting pared to 0.38 for glycogen synthase. This peptide was phos-

14474

Calmodulin-dependent Protein Kinase Substrate Specificity

Asn-Val-Phe-Ala, and Arg-Lys-ala-Ser-Gly-Pro-Pro-Val. The reason why these latterpeptides are such poor substrates is not clear. The multifunctional calmodulin-dependent protein kinase is also influenced by changes in nonbasic residues. In the peptide Lys-Lys-Arg-Pro-Gln-Arg-Ala-Thr-Ser-Asn-ValPhe-Ser, substitutionof Pro-Gln with Ala-Ala caused a 4-fold increase in apparent K,. Again, it is not clear whether this effect is separate from the interaction of the essential arginine with the enzyme or is modulating it. Perhaps the most surprising feature to have emerged from 13-Ala-14-Ala-15-Arg-16-Ala-17-Thr-18-Ser-19-Asn-ValPhe-Ala was placed at position 15, Thr-18 rather thanSer-19 studies on the substrate specificity of protein kinases is the became the major target for phosphorylation. Placement of number that utilize arginine as a primary specificity deterthe essential arginine at position 14 or 17 in this sequence minant. These include phosphorylase b kinase (18), CAMPprevented peptide phosphorylation, indicating that the loca- dependent protein kinase (17, 19), cGMP-dependent protein tion of this arginine is critical. We have found that altering kinase (20), the calcium phospholipid-dependent protein kithe location of Arg-16 also causes smooth muscle myosin light nase,* smooth muscle myosin light chain kinase (21, 23), the chain kinase to phosphorylate Thr-18 rather than Ser-19 (23). histone H4 kinase (22), and the multifunctional calmodulinHowever, this enzyme does not have as stringent a require- dependent protein kinase. In many instances, there is rement as the multifunctional calmodulin-dependent protein markably little overlap in specificity using model synthetic kinase because it will still phosphorylate peptides when the substrates despite the fact that these enzymes share a common specificity determinant. We wouldexpect the multifunctional essential arginine is moved to position 14 or 17 (23). On the basis of phosphorylation site sequence information, calmodulin-dependent protein kinasesisolated from liver and Payne et aZ. (24) proposed that theliver calmodulin-dependent brain (24-26) to have similar specificity requirements as the protein kinase probably required an arginine 3 residues on skeletal muscle counterpart described in this paper. the NH2-terminal side of the phosphorylated serine. The Acknowledgments-We are indebted to Michele Zorzi of the Howresults described herein provide strong support for the idea that Arg-X-Y-Ser(Thr) appears to represent the minimum ard Florey Institute for amino acid analysis, Susan Vienet for purification of synthetic peptides, and Klara Kasperowicz for typing the specificity determinant of the related skeletal muscle multi- manuscript. functional calmodulin-dependent protein kinase. However, REFERENCES other factors influence the rate of peptide phosphorylation, 1. Cohen, P. (1982) Nature 296,613-620 and in thecase of the myosin light chain peptide Lys-ll-Lys2. Adelstein, R. S., and Eisenberg, E. (1980) Annu. Reo. Biochem. 49, 9210K.C 12-Arg-13-Pro-Gln-Arg-Ala-Thr-Ser-Asn-Val-Phe, the adja3. McGuinness, T. L., Lai, Y., Greengard, P., Woodgett, J. R., and Cohen, P. cent lysine residues had a strong influence. Their deletion (1983) FEBS Lett. 163,231-236 caused almost a 20-fold increase in the apparent K,. The 4. Woodgett, J. R., Cohen, P., Yamauchi, T., and Fujisawa, H. (1984) FEBS rIptt.17n. 49-54 - - -~ reason for this effect is not known, but it illustrates an 5. W&&ett. J. R.. Davison. M. T.. and Cohen., P. (1983) . . Eur. J. Biochem. important problem regarding the interpretationof structure/ 136,481-487 Kemp, B. E., Pearson, R. B., and House, C. (1982) J. BioL Chem. 2 5 7 , 6. function studies with synthetic peptide substrates. Had our 13349-13353 study been restricted to theanalogs of the myosin light chain 7. Klee, C. B., and Krinks, M. H. (1978) Biochemistry 17,120-126 R. S.,and Merrifield,R.B. (1975) Anal. Bwchem. 65,241-272 sequence, it would have been reasonable to conclude that the 8. Hodges, 9. Stewart, J. M., and Young, J. P. (1966) Solid Phase Peptide Synthesis, pp. multifunctional calmodulin-dependent protein kinase has a 44 and 66, W. H. Freeman and Co.,San Francisco requirement for additional basic residues including Lys-11 10. Kemp, B. E. (1979) J. Biol. Chem. 254,2638-2642 11. Deleted in proof and Lys-12. However, since the glycogen synthase peptide 12. Glass, D.B., Masaracchia, R.A., Feramisco, J. R., and Kemp, B. E. (1978) Anal. Biochern. 87,566-575 contains only a single arginine and is anexcellent substrate, 13. Roskowski, R. (1983) Methods Enrymol. 99,3-6 the simplest hypothesis is that the multifunctional calmodu- 14. Rylatt, D. B.,and Cohen, P. (1979) FEBS Lett. 98,71-75 (1981) ColdSpringHarborConf. Celt lin-dependent protein kinase has an absolute requirement for 15. Kemp,B. E., andJohn,M.J. Proliferation 8,331-342 only 1 arginine. This interpretation would then dictate that 16. Chan, K. F. J., Hurst, M. O., and Graves, D. J. (1982) J.Biol. Chem. 257, 3655-3659 Lys-11 and Lys-12 facilitate interaction of the essential argi17. Kemp, B. E., Bylund, D. B., Huang, T. S., and Krebs, E. G. (1975) Proc. nine with the enzyme. A similar suggestion has been made NatL Acad. Sci. U. S. A. 72.3448-3452 for the CAMP-dependent protein kinase (lo), where multiple 18. Tessmer, G. W., SkusterlJ. R., Tabatabai, L. B., and Graves, D. J. (1977) J. Biol. Chem. 252,5666-5671 basic residues are commonly found onthe NHZ-terminal side 19. Zetteravist. 0..Ramarsson. U.. Humble. E.. Bereland. L.. and Enestrom. , of the phosphorylation site sequence, and yet several exoge- 20. L. (f976j Bbchek. Biophys. kes. Comhui, 70,-696-703 Glass, D. B., and Krebs, E.G. (1982) J. Biol. Chem. 257,1196-1200 nous substrates, such as @-caseinB (17) and denatured lyso- 21. KemD. B. E.. Pearson, R. B., and House, C. (1983) Pmc. Natl. Acad. Sci. U.-S.A. 83,7471-7475 zyme (27), contain only a single arginine. 22. Eckols, T. K., Thompson, R. E., and Masaracchia, R. A. (1983) Eur. J. The multifunctional calmodulin-dependent protein kinase Biochem. 134,249-254 Kemp, B. E., and Pearson, R. B.(1985) J. Biol. Chern. 260,3355-3359 is not strongly influenced by the presence of adjacent argi- 23. 24. Payne, M. E., Schworer, C. M., and Soderling, T. R. (1983) J. Bid. Chem. nines at thesite of the recognized arginine. This was assessed 258,2376-2382 Z., De Paoli-Roach, A. A., and Roach, P. J. (1982) J. Biol. Chem. using the glycogen synthase peptide analogs Pro-Leu-Ser- 25. Ahmad, 257,8348-8355 Arg-Arg-Leu-Ser-Val-Ala-Ala and Pro-Leu-Arg-Arg-Thr- 26. Kennedy, M. B., McGuinness,T., and Greengard,P. (1983) J. Neurosci. 3, 81a-m Leu-Ser-Val-Ala-Ala. Other peptides with the structure Z- 27. Bylund, D. B., and Krebs, E. G. (1975) J. Bwl. Chem. 250,218-225 Arg-X-Y-Ser(Thr), where Z and X were basic residues, were poor substrates.These included Leu-Arg-Arg-Ala-Ser-Leu- Turner, R. S., Kemp, B. E., Su, H.-de, and Kuo, J. F. (1985) J. Gly, Val-Lys-Arg-Ile-Ser-Gly-Leu, Arg-Arg-Ala-Thr-Ser- Biol. Chem 260,11503-11507 phorylated at Ser-7, the samesite phosphorylated in the parent protein. Similarly, a synthetic peptide corresponding to the NHz-terminal 23 residues of myosin light chain was readily phosphorylated by the multifunctional calmodulindependent protein kinase with a V,,/K,,, of 1.31. The importance of arginine 3 residues NHz-terminal to thephosphorylatable serine in both of these peptides was evident from experiments in which this arginine was either substituted by alanine or leucine, or changed in location. When the essential arginine at position 16 in the sequence Lys-11-Lys-12-Arg-

V""

~

~~~~

~~~~~

~~

Calmodulin-dependent Substrate Kinase Specificity Protein

0.9. 1 .08

14475

14476

Calmodulin-dependentProtein Kinase Substrate Specificity