‘rHE JOURNAL OF BIOLOGICAL CHEMISTRY (L: 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No , 11, Issue of ’ April 15, pp. 7087-7091, 1991 Printed in U.S.A.

Identification and Primary Structure of Calmodulin Binding Domains in the Phosphorylase Kinase Holoenzyme*

(Received for publication, October 31, 1990)

Peter James$, Philip Cohenf, and Ernest0 CarafoliTlII From the IlLaboratory of Biochemistry, Swiss Federal Institute of Technology (ETH), 8092 Zurich, Switzerland and the §Medical Research Council Protein Phosphorylation Unit, Department of Biochemistry, Medical Sciences Institute, University of Dundee, . . Dundee DIl1 4HN, Scotland

Fast twitch skeletal muscle phosphorylase kinase was isolated and incubated with a radioactive, bifunc- tional, photoactivable, and cleavable cross-linker con- jugated to calmodulin. Incubation of the holoenzyme only resulted in the labeling of the a-subunit in the presence of Ca2+. After cleavage with CNBr (and sub- digestion with Asp-N protease), a sequence was iden- tified (residues 1069-1087) in the a-subunit which had the predominant basic character and the propensity to form an amphiphilic helix like other calmodulin bind- ing domains. If cross-linked calmodulin was incubated with the isolated subunits of phosphorylase kinase, radioactivity was recovered in seven CNBr peptides: three came from the a-subunits, one of them corre- sponding to the sequence labeled in the holoenzyme. Three came from the @-subunit, and one came from the y-subunit. The latter contained the two adjacent cal- modulin binding domains recently identified in the y- subunit (Dasgupta, M., Honeycutt, T., and Blumenthal, D. K. (1988) J. Biol. Chern. 264,17156-17163).

Phosphorylase kinase, a key regulatory enzyme in the path- way of glycogenolysis, is controlled by calcium ions and by phosphorylation (reviewed in Refs. 1 and 2). It has a molecular mass of about 1300 kDa and a complex subunit composition: (a@y6), (3). The active site is located on the y-subunit (4), which interacts with a regulatory 6-subunit ( 5 , 6). The latter is identical to calmodulin (7, 8) and has the peculiarity of remaining associated with phosphorylase kinase even in the absence of Cay+. The site of interaction of the &subunit with the y-subunit has been identified recently and found to consist of two regions, one spanning residues 283-331 and the other spanning residues 332-371 (9). This dual contact of calmod- ulin with a target protein is reminiscent of adenylate cyclase from the bacterium Bordetella pertussis (10) and may explain the extremely tight binding of CaM’ to the y-subunit, even in

* This work was supported by Grant 31-25285.88 from the Swiss National Science Foundation and by the Medical Research Council and the Royal Society, London. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Present address: Mass Spectrometry Facility, University of Cal- ifornia, San Francisco, CA 94143.

11 To whom correspondence should be addressed: Laboratory of Hiochemistry, Swiss Federal Institute of Technology (ETH), 8092 Zurich, Switzerland.

I The abbreviations used are: CaM, calmodulin; C,,E,, dodecyl octaoxyethylene glycol monoether; DJ, Denny-Jaffe reagent; DTT, dithiothreitol; Hepes 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate; TBAH, tetrabutylammonium hy- droxide; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; I’TH, phenylthiohydantoin; HPLC, high performance liquid chro- matography.

the absence of Ca2+. The interaction of Caz+ with the y- subunit is almost certainly responsible for conferring Cay+ sensitivity to the phosphorylase kinase reaction (reviewed in Ref. 2).

Phosphorylase kinase can bind four additional molecules of calmodulin, in this case in a Ca2+-dependent manner (3, 5). These four molecules, termed the 6’-subunits, enhance cata- lytic activity a further 5-fold (3), and cross-linking studies using dimethylsuberimidate have indicated that they are in close proximity to both the a- and P-subunits ( 5 ) . Interaction between the &subunit and the a- and/or P-subunits is also indicated by experiments in which phosphorylase kinase is subjected to limited proteolysis. Brief tryptic digestion cleaves the CY- and @-subunits. The trypsin-treated enzyme no longer binds to calmodulin-Sepharose (5, ll), indicating that the ability to interact with the 6”subunit has been destroyed. Brief chymotryptic digestion cleaves the a-subunit relatively selectively and prevents the enzyme from binding to calmod- ulin-Sepharose in the presence of Ca2+ under conditions where the native enzyme is retained (5).

The purpose of the present study was to identify the binding site(s) for the 6”subunits by reacting the holoenzyme with a photoactivatable, cleavable, radioactive cross-linker conju- gated to calmodulin. Although calmodulin-binding sites were found on the dissociated a - and P-subunits, when the experi- ment was carried out on the holoenzyme, just a single site was detected on the a-subunit.

EXPERIMENTAL PROCEDURES

Materials-Asp-N specific endoprotease was purchased from Boeh- ringer Diagnostics, Mannheim, Germany. The catalytic subunit of CAMP-dependent protein kinase from bovine heart (catalogue num- ber P2645) was from Sigma. Radioactively labeled Denny-Jaffe re- agent was obtained from Du Pont-New England Nuclear, and the unlabeled reagent was a kind gift from CIBA-GEIGY, Basle, Swit- zerland. The HPLC solvents were from May and Baker, Dagenham, England, and the chemicals used for amino acid analysis and sequenc- ing were from Applied Biosystems, Foster City, CA.

Purification of Proteins-Phosphorylase kinase was isolated as described in Ref. 12 and used without further purification for studies on the intact holoenzyme. For studies on the isolated subunits, the kinase was purified further by calmodulin affinity chromatography ( 3 ) , and the subunits were separated by HPLC as described in Ref. 13. The isolated fractions were pooled, and the detergent C,,E, was added to a final concentration of 9.1 mg/ml before removing the organic solvent under a stream of dry nitrogen. The function of the detergent was to keep the subunits soluble while removing the organic solvent. It was used instead of SDS as suggested in Ref. 13. The kinase and the isolated subunits were dialyzed against 50 mM sodium glycerophosphate, 10 mM Hepes, 1 mM Mg’+ and 2 mM EGTA in 0.1 mg/ml CILEH, pH 6.8 (buffer A). Calmodulin was isolated from bovine brain (14) and modified with the Denny-Jaffe reagent (15-17) in the Ca’+-bound form.

Cross-linking-Photolabeling of the intact enzyme and the suh- units was carried out using a 1:l molar ratio of holoenzyme (or

7087

7088 Calmodulin Binding to Phosphorylase Kinase

suhunit) to CaM. The efficiency of the coupling of the reagent to CaM was about 85%. When CaM was coupled to cold reagent higher amounts of the latter could be used, increasing the coupling efficiency to more than 95%. Since the radioactive DJ-CaM was diluted with cold Dd-CaM 1000-fold, the amount of unmodified CaM was by necessity lower than 5%. For HPLC and sequencing work, the iodi- nated Denny-Jaffe CaM (specific activity 10 Ci/mmol) was diluted into a 1000-fold molar excess of the cold reagent. Cross-linking was performed by preincubating the DJ-CaM with the kinase for 5 min. The kinase and its subunits were all diluted to yield a final protein concentration of 1 mg/ml. Ca'+, when present, was added to yield a final free concentration of 100 p ~ . The solution was then placed 10 cm in front of a mercury arc lamp with a saturated copper sulfate solution (with a light path of about 2 cm) to filter out light of <300 nm. The irradiation lasted 2 min, the optimal time being determined empirically by varying the length of irradiation and determining the incorporation of radioactivity into the protein. A 100-fold excess of unmodified CaM over radioactive DJ-CaM completely inhibited the incorporation of radioactivity. The yield of DJ-CaM kinase was 10- 15% based on the incorporation of radioactivity. Cleavage of calmod- d i n from the kinase was achieved by 2 h incubation at room temper- ature with three additions of 100 mM sodium dithionite (final con- centration 300 mM), followed by dialysis overnight (Spectropor tub- ing, Spectrum Medical Industries, Los Angeles, CA, M, cutoff, 3000) against 10 mM Hepes buffer, pH 7.5, a t 4 "C with 100 mM dithionite. The products were analyzed by SDS-polyacrylamide gel electropho- resis (7%); and the gels were stained with Coomassie Blue, dried, and subjected to autoradiography for 4 days a t -70 "C. A major problem occurred with the solubility of the cr-subunit after cross-linking. The cy-hand formed large aggregates which remained trapped between the stacking and running gels. To prevent this, EGTA (2 mM) was included in the running and sample buffers. Immediately after pho- tolysis, sample buffer was added and the solution heated at 60 "C for 5 min prior to loading onto the gel.

Preparation of Labeled Protein and Digestion-The holoenzyme or its isolated subunits labeled as described above were precipitated using trichloroacetic acid (final concentration lo%), redissolved in 6 M guanidinium chloride, and reduced with 10 mM DTT for 2 h under nitrogen in the dark. Alkylation was carried out for 2 h in the dark using a 1.2-fold molar excess of freshly recrystallized iodoacetic acid over total SH in the solution. The reaction was terminated by the addition of a 5-fold excess of DTT. The solution was dialyzed briefly against water before precipitation with trichloroacetic acid. The pellet obtained was dissolved in 70% formic acid in water by gentle soni- cation. CNBr was added in 50-fold excess over methionine in the protein, and the digestion was carried out for 24 h in the dark under nitrogen. The volume of the solution was reduced by 90% under a stream of dry nitrogen, and trifluoroacetic acid was added to a final concentration of 50%. The solution was incubated for a further 5 h a t 37 "C.

( i ) The solutions containing the individual subunits were then reduced in volume to 100 pl, and the pH was adjusted to 10.5 with TBAH prior to separation by HPLC, as described under "Separation, Subdigestion, and Sequencing."

(ii) The CNBr cleavage of the labeled holoenzyme was prepared in the same manner, except that instead of injection onto the HPLC system, the solution was diluted into buffer A and the pH adjusted to 6.8. Ca'+ was added to a final free concentration of 200 p M and the solution passed over a CaM affinity column equilibriated in the same solution. The column was then washed with 10 bed volumes of the solution (now made 6 M in guanidinium chloride) before being eluted with 1 mM EDTA in 6 M guanidinium chloride. The eluted peak of radioactivity was pooled and the pH adjusted to 10.5 before injection onto the HPLC column.

Separation, Subdigestion, and Sequencing-Aliquots of the digest of the subunits each containing the same total amount of radioactivity were separated on a Du Pont-New England Nuclear poly(F) column using a gradient of 0-100% 1-propanol in a buffer of 0.1% TBAH, pH 10.5. The peaks of radioactivity were pooled as- indicated in Fig. 2 and rechromatographed on a Brownlee C-4, 300 A, 30 X 2.1 mm, 7 p reverse C Y phase column. Gradients were developed from 0-80% acetonitrile in 0.1% trifluoroacetic acid, and peptide peaks collected by hand in microcentrifuge tubes for radioactivity counting. The peaks containing radioactivity were then used directly for sequencing.

The peptide isolated from the digest of the labeled holoenzyme by CaM affinity chromatography was also run on the poly(F) column under the alkaline conditions described for the isolated subunits. 20% of the peptide isolated was used for sequencing. The rest was subdig-

ested using Asp-N protease to define further the site of labeling. Digestion was carried out using a ratio of protease to peptide of approximately 1:50 w/w for 2 h a t 37 "C as described by the suppliers. Then the pH was adjusted to 2.0 before separation by HPLC, using a C-8, 300 A, 100 X 2.1 mm, 7 p reversed phase column using a gradient of 0-80% 1-propanol in 0.1% trifluoroacetic acid.

All sequencing was carried out using an Applied Biosystems 470A sequencer with on-line PTH detection and a 420 derivatizer with on- line phenylthiocarbanoyl detection for amino acid analysis.

RESULTS

Cross-linking of the Holoenzyme with the Derivatized Cal- modulin-In cross-linking experiments using a 1:1 molar ratio of holoenzyme to DJ-CaM, only the a-subunit became labeled; but when a large molar excess of CaM (up to 100-fold) was used, labeling of all the subunits of the enzyme was observed also. Labeling of other subunits was obtained also when using old preparations of the enzyme, or fresh preparations that had been submitted to a freeze-thawing procedure. Because exogenous CaM only exchanges extremely slowly with the integral CaM (the &subunit) bound to the y-subunit (5), and the latter does not bind a second external CaM molecule, it was decided not to use holoenzyme preparations in which the y-subunit became labeled with DJ-CaM. Since the @-subunit only became labeled when the y-subunit was also labeled, it is likely that this was due to partial dissociation of the holoenzyme and did not represent a native binding site.

Calmodulin binds the Denny-Jaffe reagent most probably a t Lys-75 (18)'which is in line with results obtained using N- hydroxysuccinimide derivatives. Lys-75 is apparently not es- sential for the interaction of CaM with some enzyme targets, e.g. the Ca2+ pump of the plasma membrane (14). The as- sumption was made that this is also the case for the interac- tion of CaM with the subunits of phosphorylase kinase: al- though plausible, the assumption has not been verified di- rectly, and a caveat on the results to be presented here thus seems in order.

Identification and Primary Structure of the CaM Binding Site($ in the Isolated Subunits-After incubation with DJ- CaM, the isolated subunits were cleaved with CNBr, and aliquots of the digests containing the same total amount of radioactivity were separated as indicated under "Experimen- tal Procedures." Three labeled peptides were isolated from both the a - and @-subunits, and a single labeled peptide from the y-subunit. In the radioactivity traces of the HPLC frac- tionation of the digest of the isolated subunits (Fig. l ) , the peaks containing radioactivity were pooled as indicated by horizontal bars. Of the six peptides of the a - and @-subunits (Pep 1-6 in Fig. 1, A and B ) , five corresponded to possible CaM binding sites predicted from the cDNA-deduced se- quences (19) based on the abundance of basic and hydrophobic amino acids (Table I ) . The sixth peptide ( p e p 1 in Fig. lA), located in the a-subunit, however was not predicted. Only the N-terminal 10 amino acids of the 5-kDa fragment were se- quenced. From the cDNA-derived sequence one can see that this fragment contains a region of interspersed basic and hydrophobic residues. This would be rather unique among CaM binding domains in CaM-modulated proteins in that it would contain predominantly histidine as the basic residue, and it would be unusually rich in prolines. The single labeled peptide obtained from the y-subunit (pep 7 in Fig. 1) con- tained the two adjacent domains which jointly form the Ca2+- independent CaM binding site on the integral &subunit (9).

Identification and Primary Structure of the CaM Binding Site in the Holoenzyme-When incubated with DJ-CaM un- der standard photolysis conditions (see "Experimental Pro-

'' P. James, unpublished work.

Calmodulin Binding to Phosphorylase Kinase

30000 1 7089

B

20000

10000

0 0 1 0 2 0 30 4 0 50

HPLC fraction number

0 1 0 20 30 4 0 50

HPLC fraction number

30000

Pep 7

20000

10000

0 0 1 0 2 0 30 4 0 50

C

HPLC fraction number

FIG. 1. Radioactivity traces of the HPLC separation of CNBr digests of the isolated subunits labeled with DJ-CaM. The isolated subunits were labeled with DJ-CaM and the products cleaved with dithionite and dialyzed to remove any free radioactivity. The protein subsequently was akyiated and digested with CNRr. An equal amount of radioactivity was removed from each digest, and the pH was adjusted to 10.5 before injection onto a poly(F) column (Du Pont). A gradient was developed over 40 min from 0-100% 1-propanol in a buffer of 0.1% TBAH, pH 10.5. The peaks of radioactivity were pooled as indicated on the figure by the horizontal bars and subsequently purified as detailed under “Experimental Procedures.” The resultant sequences are given the in Table I and are indicated on the graph by a number corresponding to a sequence in the table. A , a-subunit; B, @-subunit; C, y-subunit.

cedures”), the holoenzyme only became labeled in the a- subunit in the presence of Ca2+ (Fig. 2). After cleavage with CNBr and separation of the resulting peptides as described, radioactivity was recovered in one peak (Fig. 3), 20% of which was used for sequencing, yielding the following structure: X P S F L S P G T S V T P S S G S , w h e r e X w a s a n u n i d e n - tified residue. The sequence corresponded to that of the peptide labeled with DJ-CaM in the isolated a-subunit (pep 2 in Fig. l), i.e. it was contained in a sequence spanning residues 1037 through 1114. To define further the site of interaction with DJ-CaM, the remaining 80% of the CNBr peak was subdigested with Asp-N protease. After separation of the new digest, as indicated under “Experimental Proce- dures,” two of the peptides obtained were found to contain radioactivity (Fig. 4). They were collected and sequenced. Peak 1 contained the undigested starting CNBr peptide; peak 2 yielded the following sequence: D R V P I G F Y Q K-

V W K L Q K C H . . ., which corresponds to amino acids 1080-1098 of the a-subunit. Although the peptide was not sequenced through to the end, amino acid analysis indicated that the peptide composition matched that of residues 1080- 1114, i.e. the last third of the CNBr peptide. The N terminus of the sequence must have arisen from the deamidation of the asparagine (see the cDNA-deduced sequence, 20) to aspartic acid, thus permitting the attack by the Asp-N protease.

DlSCUSSlON

Phosphorylase kinase from fast twitch skeletal muscle is a hexadecamer containing four copies of each of the four sub- units: a (138 kDa), p (125 kDa), y (45 kDa), and 6 (17 kDa) (3,8,20,21). Because y is the catalytic subunit, 6 (calmodulin) is the Ca2+-binding subunit, and the (3 is the component whose phosporylation at an N-terminal serine is primarily respon-

Calmodulin Binding to Phosphorylase Kinase

A B

Y( - DJa+CaM

1 21".

/" c

-3000 g x -.

-2000 c I;.

f 1000 j

I

1

v) 0 N

am" 1 2 3 1 2

FIG;. 2. The binding of DJ-CaM to the holoenzyme. T h e hol- ornzymr ( 1 mg/ml in h f l e r A, see "Experimental I'rocedures") was inrul)nted for 5 min with an equimolnr amount of D-J-CaM prior to photolysis. Immctli;~tely alter photolysis all samples were I)rought to 2 mM 13;'rA ;Ind incul)nted with sample buffer for 54 min at (3) "C. I'nnrI A shows (:oomassie Brilliant Hlue-stained gels of the prepnra- t ions. In Inrw 1 thr cross-linking wras performed in the presence o f 2 mM EGTA, in hnr 2 in the presence of 1 0 0 mM Ca". Imnr 1 and 2 gels were overloaded with enzvme purified as indicated in Ref. 12, h u t omitting the I'innl inn exchange step and the calmodulin column treatment. /,nnr .'I is n highly purilietl preparat ion of the kinase; the rnzyme \vas treated with the calmodulin column procedure and was not cross-linked. I t is presented in the Iiprlre as a standard. The minor 100-kI)t1 band in I m r .'I is an impuritiv. termed component S in lief. 12. I'nncl H shows autoradiograms of Innrs I and 2 o f pond ;I. / A / ( \ + ( 'OM refers t o the suhunit attached t o CaM via the I1.J reagent. I N is the result of partial cleavage of the o-sulwnit from the ('akl by the DTT in the 1,aemmli sample I~uffer acting on the 11.1 reagrnt.

sible for activation by cyclic AMP-dependent protein kinase (reviewed in Refs. 1 and 2), the role of the a-component has been the most difficult of the four subunits to comprehend. T h e tr-subunit is phosphorylated on 1 serine residue toward its C terminus hy cyclic AMP-dependent protein kinase in uitro and in vivo which can enhance activity up t o 2-fold, provided that the &subunit is also phosphorylated (reviewed in Refs. 1 and 2). T h e tu-subunit also can be phosphorylated extensively by phosphorylase kinase itself (autophosphoryla- tion), hut there is no evidence that autophosphorylation oc- curs in viuo (reviewed in Kef. 2 ) . The present finding that D.7- CaM only cross-links to the tu-subunit in the holoenzyme now

c"----r"--T" - ~. ~. 10 20 30 40 50'

rime (min)

FIG. 4. HPI,C separation of the Asp-N digest of the CaM binding domain isolated as shown in Pig. 3. &O' , o l the p r p t i t l v isolnted in Fig. :1 wns sul)digestetl using Asp-S with n ratio of' protease to peptide of approximately 1 5 0 w/w for 2 h at 37 *('. Then the was ndjusted t o 2.0 t~efnre separation hy Hi'IA', using n C-8, :{(IO A,

2.1 X 1 0 0 mm, 7 p reversed phase column with a Kradient o f 0.80' ; I-propanol in 0.1'; trifluoroacet ic acid nver .'IO min. 'l'wn p e p t id(. peaks containing radioactititv were collected and sequenced. I ' w r k I represents the undigested starting ('Nlir peptide ( p r p 2 ) . anti the sequence ofpcnk 2 is given in the text. The other t 'V ahwrtling peak. contained no radioactivity.

strongly suggests that one role of the tr-suhunit is to interact with the 6'-subunit. This conclusion is not only consistent with the effects of limited proteolysis (see the introduction), but also with the situation in slow twitch skeletal muscle and cardiac muscle, where the tr-subunit is replaced hy a slightly smaller n-subunit (22, 23). The tr,-suhunit has the same S- terminal sequence as the tu-subunit (2.11, and like the n- subunit is absent in I-strain mice which lack both phospho- rylase kinase activity (25, 26) and mRNA ( 2 7 ) in fast twitch, slow twitch, and cardiac muscle. These observations suggest t h a t n and tr ' may he produced by alternative splicing of the same gene. Since cardiac phosphorylase kinase does not bind to calrnodulin-Sepharose in the presence of 0.2 mxf Ca.", unlike the skeletal muscle enzyme ( 2 8 ) . this again suggests that the h-subunit interacts with the tu-subunit. I t will clearly be of considerable interest to elucidate the structural differ- ence between the ( u - and tu'-suhunits.

When one inspects the cDNA sequence corresponding t o the sole C N R r peptide Iaheled with D.J-CaM, three distinct

Calmodulin Binding to Phosphorylase Kinase 7091

regions are apparent, and are shown below as domains A, B, and C.

A. MSPSFLSPGTSVTPSSGSFPGHHTS(lO37-1061) B. KDSRQGQWQRRRRLDGALNRVPIGYOKVWKVLQKCHGL

(1062-1 100)

C. SVEGFVLPSSTTRE (1101-1114)

Region A is not charged (except for the histidine) and is not predicted to have any preferred structure. I t is thus probably not part of the binding domain. Region B appears to consist of two halves separated by a flexible region (Pro-1083). The second half, where radioactivity was found, has an evident basic character and shows the propensity to form the amphi- philic helix so characteristic of CaM binding domains (under- lined). The first half is also very basic and possibly could be involved in the binding of CaM, although no radioactivity was found associated with it. The rest of the CNBr peptide, domain C, contains several acidic residues and is unlikely to take part in forming the binding domain proper.

It is interesting that the peptide labeled with DJ-CaM on the holoenzyme is located just C-terminal to the region on the a-subunit which is phosphorylated multiply. The phos- phorylation sites include the serine labeled by cyclic AMP- dependent protein kinase, several autophosphorylation sites, and four additional serines which are phosphorylated partially i n vivo by as yet unidentified protein kinases (29). The pres- ence of a calmodulin binding domain just C-terminal to a phosphorylation site(s) is turning out to be a common phe- nomenon in CaM-regulated protein kinases. The region con- taining the phosphorylation site(s) has been proposed to act (in its dephosphorylated state) as an inhibitory "pseudosub- strate" domain which binds to the active site of the protein kinase. Inhibition may be relieved either by the binding of CaM at its nearby site, or by phosphorylation of the pseudo- substrate domain itself. In some cases, e.g. the plasma mem- brane Ca2+ pump, the pseudosubstrate domain cannot be phosphorylated and can only be removed from the active site by CaM. This may be the case for phosphorylase kinase as well, because phosphorylation by cyclic AMP-dependent pro- tein kinase does not seem to prevent interaction of the holo- enzyme with the 6-subunit3 and autophosphorylation of the a-subunit does not seem to occur i n uiuo. However a difference from other calmodulin-dependent protein kinases is that the calmodulin binding site and the active site are located on different subunits.

Although only the GaM binding site on the a-subunit appears to be operational in the holoenzyme, several further potential CaM binding sites on both the a- and (3-subunits were revealed by cross-linking the isolated subunits with DJ- CaM (Table I), raising the question of whether one or more of these might have physiological significance. One possibility is that troponin C may interact with phosphorylase kinase at one of these regions, i.e. at a site distinct from that which binds the 6"subunit. Troponin C, like the 6'-subunit, is a powerful activator of phosphorylase kinase (30, 31), and sev- eral lines of evidence suggest that activation by troponin C may be important in synchronizing glycogenolysis and muscle contraction (31; reviewed in Ref. 2). Preliminary cross-linking experiments with dimethylsuberimidate have indicated the formation of cross-linked complexes between troponin C and the (3-subunit, but not between troponin C and the a-subunit

'' P. James, unpublished observations.

(32). It will be of interest to see if DJ-troponin-C cross-links to the holoenzyme at a site on the @-subunit, rather than the a-subunit.

The integral CaM subunit ( 6 ) confers Ca2' sensitivity to the catalytic (y) subunit and remains bound to it even in the absence of Ca", whereas the y6 complex is inhibited by the two larger subunits (a and (3) (33). The isolation of both ay6 and y6 complexes by partial dissociation with LiBr indicates that the a-subunit interacts with the y-subunit (34), and the external CaM ( 6 ) may activate phosphorylase kinase by neu- tralizing inhibitory effects exerted on the y-subunit by the a- and/or @-subunit. Thus the calcium signal may be acting through three pathways: directly through the &subunit and indirectly via 6' and troponin C. Phosporylation of the (3- subunit by cyclic AMP-dependent protein kinase during ad- renergic stimulation may relieve the inhibitory effects of the @-subunit in an analogous manner. Clearly, the structural complexity of phosphorylase kinase reflects its key position at the point of convergence of the neuronal and hormonal pathways which regulate glycogenolysis.

Aknowledgrnent-We are indebted to Rolf Moser for help in pre- paring the illustrations.

REFERENCES 1. Pickett-Gies, C. A,, and Walsh, D. A. (1986) in The Enzymes (Boyer, P .

D., and Krebs, E. G., eds) Vol. 17, pp. 396-459, Academic Press, Orlando, FL

2. Cohen, P. (1988) in Mol. Aspects Cell. Regul. 5, 123-144 3. Shenolikar, S., Cohen, P. T. W., Cohen, P., Nairn, A. C., and Perry, S. V.

4. Skuster, J. R., Chan, K. F. J., and Graves, D. J. (1980) J. Biol. Chem. 2 5 5 ,

5. Picton, C., Klee, C. B., and Cohen, P. (1980) Eur. J . Biochem. 1 1 1 , 553-

6. Chan, K.-F. J., and Graves, D. G. (1982) J. Biol. Chem. 257,5956-5961 7. Cohen, P., Burchell, A,, Foulkes, J. G. , Cohen, P. T. W., Vanaman, T. C.,

8. Grand, R. J., Shenolikar, S., and Cohen, P. (1981) Eur. J. Biochem. 1 1 3 ,

(1979) Eur. J. Biochem. 100,329-337

2203-2210

561

and Nairn, A. C. (1978) FEBS Lett. 92,287-293

R.SQ-RG7 9. Dasgupta, M., Honeycutt, T., and Blumenthal, D. K. (1988) J. Bid. Chem.

""" "1.

10. Landant, D. (1988) J. Biol. Chem 263,2612-2618 11. DePaoli-Roach. A. A.. Gihhs. J. B.. and Roach. P. J. (1979) FEES Lett.

264,17156-17163

105,321-324 12. Cohen, P. (1973) Eur. J . Eiochem. 3 4 , 1-14 13. Crahh, J. W., and Heilmeyer, L. M. G., Jr. (1984) J . Biol. Chem. 2 5 9 ,

6346-6350 14. Guerini, D. (1986) Relationship between Structural Principles and Func-

tional Properties of the Ca'+-binding Protein Calmodulin. Swiss Federal Institute of Technology (ETH) Dissertation 8113

15. James, P., Maeda, M., Fischer, R.,. Verma, A. K., Krebs, J., Penniston, J.,T., and Carafoll, E. (1988) J. B~ol. Chem. 263,2905-2910

16. Jaffe, C. L., Lis., H., and Sharon, N. (1980) Biochemistry 19,4423-4429 17. Denny, J. B., and Blohel, G. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,

18. Vanaman, T. C. (1987) Methods Enzymol. 139,417-433 19. Buschmeier, B., Meyer, H. E., and Mayr, G. W. (1987) J . Bid. Chem. 262,

9454-9462 20. Zander, N . F., Meyer, H . E. Hoffmann-Posorski, E., Crabb, J. W., Heil-

meyer, and Kilimann, M. W. (1988) Proc. Natl. Acad. Sei. U. S. A. 85,

21. Kilimann, M. W., Zander, N. F., Kuhn, C. C., Crabh, J. W., Meyer, H. E., 2929-2933

and Heilmeyer, L. B. G. (1988) Proc. Natl. Acad. Sci. U. S. A. 85,9381- 9385

~~~. .. ~

5286-5290

22. Jennissen, H. P., and Heilmeyer, L. M. G. (1975) FEES Lett. 4 2 , 77-80 23. Burchell, A,, Cohen, P. T. W., and Cohen, P. (1976) FEES Lett. 67, 17-22 24. Crahh, J. W., Harm, W. R., Johnson, C. M., Sotiroudis, J. S., Kuhn, C. C.,

25. Cohen, P . T. W., Burchell, A., and Cohen, P. (1976) Eur. J . Biochem. 6 6 , and Heilmeyer, L. M. G. (1990) Electrophoresis 1 1 , 133-140

? A 7 - ? W

26. Cohen, P. T. W., LeMarchand-Brustel, Y., and Cohen, P. (1981) Eur. J .

27. Bender, P. K., and Lalley, P. A. (1989) Proc. Natl. Acad. Sci. U. S. A. 86,

"" yy"

Biochem. 115,619-625

9996-1000

B i d . Chem. 255, 11102-11105

J. Biochem. 118,367-376

28. Sharma, R. K., Tam, S. W., Waisman, D. M., and Wang, J. H. (1980) J .

29. Meyer, H. E., Meyer, G. F., Dirks, H., and Heilmeyer, L. M. G. (1990) Eur.

30. Cohen, P., Picton, C., and Klee, C. B. (1979) FEBS Lett. 104,25-30 31. Cohen, P. (1980) Eur. J. Biochem. 1 1 1,563-574 32. Picton, C., Shenolikar, S., Grand, R., and Cohen, P. (1983) Methods

33. Paudel, H. K., and Carlson, G. M. (1987) J . Bid. Chem. 2 6 2 , 11912-11915 34. Kee, S. M., and Graves, D. J. (1982) J. Bid. Chem. 2 6 1 , 4732-4737

Enzymol. 102,219-227

1